 |
Description  |
|
|
The present invention relates to fiber-optic sensors and more particularly
to methods of their manufacture.
Fiber-optic sensors are popular for use in detecting changes in
temperature, pressure and strain. Fiber-optic sensors, in which the fiber
itself acts as the transducer, are of interest in the context of advanced
composite materials. This is due to the face that the fiber is generally
compatible with both thermoset and thermoplastic composites. This makes
the fiber-optic sensor particularly useful when embedded in the composite
material to function as an `in situ` measurement device.
One type of fiber-optic sensor is generally known as the Fiber Fabry-Perot
sensor (hereinafter referred to a the FFP sensor). As is well known to
those in the art, the term `Fabry-Perot` refers to the use of a first
semireflective mirror which is positioned in a light path upstream of a
second mirror wherein the second mirror is either fully or semireflective.
With this arrangement, an interference pattern is established by the light
reflected off the first and second mirrors.
Conventional Fabry-Perot sensors are generally made from a single or
multi-mode optical fiber. In some, the core region of the fiber is a glass
or plastic material. The core region is formed by doping the central
portion of the fiber such that it has a higher index of refraction than
the surrounding glass (which is known as the `cladding`). Thus, light
launched down the core of the fiber will be confined to the core, and
propagate with little loss. Surrounding the cladding is a buffer which is
generally a soft acrylate or polyamide coating which has a lower elastic
modulus than the glass fiber in order to provide mechanical strength and
protection to the fragile glass fiber.
Thus, the fabrication of an FFP cavity basically involves the formation of
a pair of reflective surfaces (usually semireflective) at spaced locations
along the fiber. The reflective surfaces are generally formed by a slice
(hereinafter referred to as a `semireflective splice) bordered by media
of different refractive indices, for example, glass to doped glass, glass
to air, glass to dielectric material, or glass to metal. There are a
number of conventional methods to form semireflective splices, including:
i) chemically precipitating silver over the entire fiber endface and fusion
splicing the coated end with another end to form a continuous fiber
(referred to as the `Silver` approach);
ii) sputtering titanium oxide over the entire fiber endface and fusion
splicing the sputtered end with another end to form a continuous fiber
(referred to as the `Titanium` approach);
iii) forming an air bubble in the core region of a fusion splice (referred
to as the `Air Gap` approach);
iv) holding apart the ends two fibers to form an air gap therebetween and
then supporting the fibers by means of a large hollow core fiber (referred
to as the `Hollow Core` approach).
Although these techniques are capable of producing a functional FFP sensor,
these conventional techniques have several shortcomings.
For example, both the Air Gap and Hollow Core approaches require an
external support structure (which can be in the form of a hollow core
fiber) to provide the necessary mechanical strength. This external
structure is undesirable for a sensor to be embedded within a composite
material.
Using the spaced semireflective mirrors means that the second splice must
match the lower reflectively of the first semireflective splice. As a
result, an additional length of fiber extending past the sensing region is
required. Since both semireflective mirrors have a low reflectivity
(typically 4%) the majority of light entering the sensing region
propagates through and must be prevented from reflecting back from outside
the sensing region. This is usually accomplished by inserting the lead-out
fiber endface in index matching fluid or gel, or by fracturing the end of
the fiber, all at a greater expense.
One use of the FFP sensor as described herein above is in fiber-optic
strain rosettes. Strain rosettes are intended to measure the in-plane
strain tensor at a point in a host structure. They may be surface adhered
or composite embedded. Two types of fiber-optic strain rosettes have been
described previously, using localized polarimeters and Michelson
interferometers as explained in:
i) the article entitled `Structurally Integrated Fiber Optic Strain
Rosettes` authored by Dr. Raymond Measures et. al. published in the
proceedings: `Fiber Optics, Smart Structures and Skins`, SPIE Vol. 986,
pp. 32-42, 1988;
ii) the article entitled `Localized Fiber Optic Strain Sensor Embedded in
Composite Materials` published in the proceedings: `Fiber Optics, Smart
Structures and Skins II`, SKIE Vol. 1170,pp. 495-504, 1989;
The former sensor lacks the strain sensitivity required to make
measurements with a 3 millimeter gauge length with a 1 microstrain
resolution. The latter suffers from the need to maintain common-mode
strain-rejection (that is, lead-in/lead-out insensitivity) of four optical
fiber leads both inside and outside the composite structure and phase
continuity across the connectors.
There remains a need for an improved fiber optical device. It is an object
of the present invention to provide just such a device.
Briefly stated, the invention involves, an optical fiber device comprising:
an optical fiber;
the fiber having a first portion with an endface, the endface having a core
and a peripheral area around the core, the endface having defined thereon
a first region adjacent the core and a second region adjacent the
peripheral area;
a layer of reflective material located on the endface and confined to the
first region, thereby forming a localized reflector thereon.
In another aspect of the present invention, there is provided a method of
making an optical fiber device comprising the steps of:
providing an first optical fiber portion with a core and a peripheral area
around the core;
forming an endface on the first portion with the core and the peripheral
area exposed thereon;
providing a first region on the endface and adjacent the core and a second
region adjacent the peripheral area;
providing a layer of reflective material on the endface and confining the
layer to the first region, thereby to form a localized reflector thereon.
Several preferred embodiments will now be described with reference to the
appended drawings in which:
FIG. 1 is a schematic cross sectional view of an optical fiber;
FIG. 2 is a schematic view of an FFP sensor;
FIG. 3 is a schematic view showing the light path through the fiber in the
FFP sensor illustrated in FIG. 2;
FIG. 4 is a schematic view showing an FFP operating system;
FIG. 5 is a schematic view showing the `relevant power budget` for the FFP
sensor illustrated in FIG. 2;
FIG. 5A is a schematic view showing multiple reflections in a fiber.
FIG. 6 is a micrograph of a mirror deposited on the endface of a fiber in
connection with the FFP sensor illustrated in FIG. 2;
FIG. 7 is a schematic vie of a mechanical mask technique in connection with
the fabrication of a localized mirror 2;
FIG. 8a) to g) are schematic views of successive steps in a photodeposition
technique in connection with the fabrication of a localized mirror;
FIG. 9 is a schematic view of a strain rosette making use of the FFP sensor
illustrated in FIG. 2;
FIG. 10 is a schematic vie of another strain rosette;
FIG. 11 is a schematic view of a core of a fiber optic element;
FIG. 12 is a schematic view of a sensor arrangement;
FIG. 13 is a schematic view of another sensor arrangement;
FIG. 14 is a schematic view of yet another sensor arrangement;
FIG. 14a is a graph depicting wavelength selectivity of a number of
components of the sensor arrangement of FIG. 15;
FIG. 15 is a schematic view of yet another sensor arrangement;
FIG. 16 is a schematic view of yet another sensor arrangement.
Referring to the FIGS. 1 to 3, there is provided a fiber Fabry-Perot (FFP)
sensor 10 which makes use of a fiber 12. The fiber 12 is of the type shown
in FIG. 1, is commercially available and known to have a core 14 and a
peripheral area around the core and integrally formed therewith from
dielectric material and what is referred to as a `bow-tie` configuration.
The term `bow-tie` arises from the use of two stress regions 16a, 16b
formed from doping the region outside of the glass core 14. These stress
regions set up a stress field across the core 14 which tends to confine
the polarization of the launched light to a single polarization mode (or
plane) along the length of the core 14.
FIG. 2 illustrates the FFP sensor 10, which has a fiber optic core 14 and a
peripheral area surrounding the core. The core 14 has a splice shown at 18
where a semireflective surface is formed on or about the core 14, namely
in a first region, to form a semireflective mirror in a manner to be
descried. It should be pointed out that the first region is adjacent the
core and may or may not have a size precisely equal to the core as will be
described. It should also be pointed out that the semireflective mirror is
confined to the first region thereby leaving a second region, adjacent the
peripheral area of the endface, uncoated. A sensing region 20 is provided
down stream of the splice 18 and terminates at a perpendicular endface 22
that is coated with a mirrored layer 24.
In use, monochromatic light is launched down the core 14 of the fiber 12
along path A until it reaches the semireflective mirror where a small
portion (in the order of 10 percent) of the incident light is reflected
back into the core 14 along path B to form a reference component. The
remainder (minus losses) continues through the sensing region 20 along
path C, reflects off the endface mirror 22 and returns to the
semireflective splice 18 along path D. A portion of the reflected light in
path D passes through the semireflective mirror to form a sensing
component which interferes with the reference component to produce a
sinusoidal intensity modulation (which will be referred to below as
`modulated light`) that is a function of strain, pressure and temperature.
Referring to FIG. 4, an FFP sensor system is shown at 30 which includes a
light source in the form of a laser diode is coupled to one end of an
optical fiber 32a. At the other end of the fiber 32a is joined to one port
of a (50:50) 2.times.2 coupler 34. The coupler 34 has three other ports,
the first of which is joined to one end of an optical fiber 32b. The other
end of the optical fiber 32b which is joined to an FFP sensor 10. The
third port of the coupler 34 is joined to one end of a fiber 32c which is
terminated by an index matching gel. The fourth port of the coupler 34 is
joined to one end of an optical fiber 32d the other end of which is joined
to a photodetector 38. An appropriate electronics system 40 is coupled to
the photodetector 38 to analyze signals carried by the fiber 32d.
All components should be spliced or coupled on axis to maintain a linear
State of Polarization throughout the system. This will help to present
signal fading.
In use, light from the laser diode is launched along the optical fiber 32a
through the coupler 34, and on to the FFP sensor to produce the reference
and sensing components. The modulated light then returns through the fiber
32b to the 2.times.2 coupler 34 which directs a portion of the light to
the photodetector 38. The system 40 is then used to interpret the optical
signal.
The steps to fabricate an FFP sensor 10 are as follows:
1) The protective buffer layer is removed from the ends of two optical
fibers. Enough of the buffer is removed to accommodate the chuck of the
fusion splicer to be used and the gauge length desired;
2) Each fiber 12 is cleaved to obtain a flat, perpendicular endface 22;
3) The core 14 of one or both endfaces is coated with a localized
reflective layer, that is, localized around the core 14 of the fiber 12
leaving the remainder of the fiber 12 uncoated. This is done by one of a
number of techniques as will be described in more detail below;
4) The endfaces are fusion spliced together. This is done by placing both
endfaces in an electric arc which softens the glass. Computer controlled
micropositioners are then be used to force the two fiber endfaces
together. Since only the core 14 of each fiber 12 is coated with the
semireflective material, the remainder of the fiber 12 is pristine. This
means that the resulting weld is substantially as mechanically strong as
the pristine fiber 12. Fusion splicers are commercially available with
adjustable arc current and time settings, as well as the positioning of
the optical fiber 12;
5) One end of one fiber 12 is cleaved to form a sensing region 20, the
length of which is determined by the sensing application, strain and
spatial resolution requirements;
6) A reflective layer is deposited on the cleaved end of the fiber 12 to
form the endface 22 mirror; and
7) The fused fibers are recoated with a suitable buffer material. For some
sensor applications, for example those which involve embedding an FFP
sensor in (or attaching an FFP sensor to) a host structure, this coating
should be sufficiently thin to couple the FFP sensor with changes in
strain and temperature in the host structure.
During the fusion process the peripheral area of one endface is bonded to
the peripheral area of the other endface. Whether significant degree of
bonding occurs in the core region will depend on the type of material
being used to form the semireflective mirror.
It should be borne in mind that when fusing the ends of the fiber 12, the
fiber's core and polarization axes must be aligned to ensure that the
light will propagate through the sensing region 20 with a substantially
constant linear state of polarization (S.O.P.). The stress regions in the
fiber 12 define the two polarization axes. These axes must be aligned
prior to fusion splicing in order to preserve linear polarization through
the semireflective splice 18. This is accomplished by several known
methods, including:
i) monitoring the state of polarization of light propagating through the
two fibers 12 prior to fusion splicing;
ii) visually aligning the stress regions visible at the fiber 12 endface
22; or
iii) visually aligning the stress regions as viewed perpendicularly through
the side of the fiber.
However, the first method cannot be used if the fiber core 14 is coated
with a highly reflective layer (on the order of at least 50%) prior to
fusion splicing since the coating covers the core 14. The second method is
also not possible if the entire endface 22 of the fiber 12 is coated since
this technique causes the stress members to be obscured.
Unlike prior FFP sensors, the present technique allows the mechanical
strength and optical characteristics to be optimized independently by
controlling:
a) Fusion splice 18 current, time and force settings
Prior techniques as indicated all require a reduction in the fusion
splicing arc current and duration, to achieve a semireflective splice.
This technique allows for the introduction of a reflective surface which,
if coated over the entire endface of the fiber, would no form an adequate
bond between the two fibers. However, by localizing the mirror to the
first region, the remaining fiber endface in the second region can be
responsible for the mechanical strength of the semireflective splice.
Thus the fusion splicing parameters as applied by the fusion splicer
manufacturer for standard optical fiber splicing, can be used to ensure
optimal mechanical strength. The characteristics of the reflective surface
can then be chosen to obtain the desired reflectivity and wavelength
response without simultaneously affecting the mechanical strength of the
semireflective splice.
The strength of the `semireflective` splice may be optimized by maximizing
the glass to glass contact. The core of a single-mode fiber is typically 4
to 10 microns in diameter. Thus the core region comprises only 0.1% to
0.6% of the total cross sectional surface area of an optical fiber
(assuming a 125 micron outside diameter). By limiting the mirror to the
core region, the vast majority (that is greater than 99%) of the fiber
endface is available for fusion splicing. Even a 15 micrometer diameter
mirror represents only 1.4% of the endface surface area of the fiber.
b) Characteristics of the material in the semireflective coating
Since the choice of the reflective material has no bearing on the
mechanical strength of the semireflective splice, a single layer metallic
or dielectric coating, for example, can be chosen where one wishes a
wavelength independent, broad band reflector. Multilayer metallic and/or
dielectric coatings can be chosen where a narrow band, wavelength
dependent reflector is desired.
At the semireflective splice 18, and the endface 22 mirror, three
parameters describe the optical quality of the mirror.
R--reflectance
T--transmittance
L--loss,
where R+T+L=1. This is referred to as the `relevant power budget`.
FIG. 5 illustrates the cavity and helps in discussing the relevant power
budget. Experimental evidence has shown that multiple reflections within
the cavity can be ignored for high loss semireflective mirrors. In order
to maximize the power budget of the sensor 10, the reference and sensing
components must be equal in power and as large as possible. For one
specific example of an FFP sensor 10, following optical properties may be
typical: R2=95% L2=5% T2=0%; while L1 (loss at the semireflective splice)
can be as high as 80%. These high losses permit the use of a fully
reflective mirror and ensure a single reflection at the fully reflective
mirror rather than multiple reflections typical of a high finesse
application, as for example shown in dashed lines in FIG. 5a. This single
reflection is made possible by the fact that a sufficiently small
percentage of the incident light is reflected back from the semireflective
mirror into the sensing cavity (in order of 1%). Of course, these
properties may vary with the type of FFP sensor 10 being made.
Since the relatively small surface area of this core is not required to
contribute to the mechanical strength of the semireflective splice, a
relatively thick reflective layer of material can be deposited over the
endface.
However the layer (after fusing) should not be excessively thick, that is
on the order of 40 microns or more, as this may:
i) significantly degrade the quality of the semireflective mirror, that is
by preventing light from propagating through the mirror to establish the
requisite optical interference between the sensing and reference
components; and
ii) significantly reduce the mechanical strength of the splice and require
additional extrinsic support as a required, for example with the `Hollow
Core` and `Air Gap` approaches.
The present technique is significant in that it eliminates the need for
additional strength members or secondary fibers with index matched ends.
Deposition of a thin reflective layer can be accomplished by various vacuum
deposition techniques (evaporation, sputtering E-gun (Electron beam gun)
MBE (Molecular Beam Epitaxy), CVD (Chemical Vapour Deposition) or through
other techniques including chemical deposition. Suitable mirror materials
include single and multi-layer metals (Ni, Co, Al, Ag, Au, etc), as well
as single and multi-layer dielectrics (TiO.sub.2, MgF.sub.2, etc.).
Increased mirror adhesion to the fiber can be achieved by heating the
fiber during deposition.
Among others, the localized semireflective mirror may be formed using the
following procedures.
1) MECHANICAL MASK
A localized semireflective mirror can be fabricated by depositing a thin
reflective film through a mechanical mask or aperture, wherein the
dimensions of the mask correspond to the dimensions of the deposited
mirror. FIG. 6 is a micrograph of a mirror deposited through a 15
micrometer diameter stainless steel mask.
As shown in FIG. 7, a mask 50 is butted against the endface 52a of the
optical fiber 52. The mask 50 has an opening 50a aligned with the core of
the fiber. FIG. 7 illustrates the positioning of the mask prior to butting
the fiber against the mask. Alignment of the fiber with the mask is
accomplished by launching light down the fiber and detecting the light
through the mask by a detector 54 placed directly under the mask. The
position of the fiber is then adjusted in the X and Y directions until the
light intensity as measured by the photodetector, is maximized. Alignment
in the Z direction is accomplished by visually aligning the fiber with a
low power stereo microscope. A series of iterations in the X, Y and Z
directions are then required to attain optimal alignment of the fiber with
the mask.
An alternate approach is to launch light through the aperture and detect
the amount of light launched into the adjacent fiber as the adjacent fiber
is brought into alignment with the mask. Either of these approaches can be
automated with a microcomputer and commercially available motorized
micropositioners.
The aperture is usually larger than the core of the fiber. Thus, optical
alignment of the fiber with the mask will occur when the fiber core is
located anywhere inside the mask and not just when the core is centered
over the mask. Therefore, as the diameter of the aperture decreases, the
concentricity of the mechanical alignment increases.
Apertures can be fabricated in several ways including the following
proposed methods:
1) Drilling a hole with a laser, electron beam or similar device into a
material such as stainless steel;
2) Chemically etching a material such as a single crystal silicon wafer;
3) Mechanically punching a hole through a thin membrane;
4) Using a series of four orthogonal moving knife edges or a camera type
iris to define mask area.
2) PHOTORESIST MASK
A mask similar to the mechanical masks described above can also be created
by coating the end of a fiber with a layer of photoresist material and
exposing the photoresist material to light. Two approaches are possible
using either a positive or negative photoresist material.
A technique using a positive photoresist material is shown in FIG. 8,
wherein a thin layer of photoresist material is first applied to the
endface of a cleaved fiber as shown in FIG. 8b. The core region is exposed
by launching light down the fiber thus ensuring only the core region is
exposed as shown in FIG. 8c. Alternatively, the core region of the fiber
can be exposed from the outside using a mechanical mask to block off
unwanted light, or by focusing light onto the core region. The photoresist
material is then developed, washing away photoresist material from the
core region, leaving the core exposed as shown in FIG. 8d. A reflective
metal or dielectric layer is then coated over the entire endface of the
fiber 8e. The photoresist material is then dissolved thus lifting the
deposited reflective layer from the cladding region. A localised mirror
over the core region is then achieved as shown in FIG. 8f.
Another approach is available using negative photoresist material. A bare
cleaved fiber is first coated over its entire cross section with the
desired reflective coating. A layer of the negative photoresist material
is then applied over the coated fiber endface. The photoresist material is
then exposed over the core region through a mechanical mask placed over
the fiber endface, or by focusing the light directly onto the core region.
The photoresist material is then developed leaving only the core region
coated with photoresist material. An acid or other etchant is then used to
dissolve the reflective coating over the cladding region. The remaining
photoresist material is then dissolved leaving a localised mirror over the
core region of the fiber endface.
In yet another approach, the photoresist may be applied to a fiber endface
coated with a single or multiple layer of dielectric material. In this
case, the core can be exposed as described for the positive photoresist
technique mentioned above where the wavelength of exposing light is chosen
such that the dielectric layer is transmissive and not reflective at that
wavelength.
STEREO LITHOGRAPHY
There are various stereo lithography or similar light assisted
photodeposition techniques may also provide a means to coat only the core
region of the fiber, by placing a cleaved fiber end in the moulding fluid
and launching light down the fiber, a solid `plastic` layer will form over
the core region of the fiber. A thin layer of material can thus be coated
over the entire end of the fiber. A solvent can then be used to dissolve
the solid moulding fluid, lifting the coating off the core region. A mask
over the fiber with an aperture in the core region similar to FIG. 8d
results. The steps of the photoresist technique can then be used from this
point to achieve a localized mirror.
3) LIGHT ASSISTED PHOTODEPOSITION
There are many light assisted photodeposition techniques where light is
used as a catalyst to stimulate deposition of a reflective layer over the
core of a fiber. Exposing the core region of the fiber with light is
accomplished by either launching light down the fiber, or exposing the
core region directly from outside the fiber through a mask, or by focusing
the light onto the core region.
The FFP sensor 10 is well suited to forming a strain rosette, since it
combines the high strain sensitivity of the Michelson interferometer with
the single fiber self-referencing property of the polarimeter. Testing has
shown that the FFP strain rosette device matches or eventually surpasses
the performance of conventional resistive foil electrical strain rosettes.
In one example shown in FIG. 9, a strain rosette 80 was formed with three
FFP sensors 82, 84 and 86 arranged in the fashion illustrated in FIG. 9.
By keeping the sensing region 82a, 84a, 86a of each FFP sensor, straight,
the standard strain-rosette reduction formulae can be employed to
interpret the readings obtained therefrom. Accordingly, as illustrated in
FIG. 9, the two outer FFP sensors 82, 84 are curved just before the
sensing region. This was be doped by plastically deforming the fibers
using the heating arc of a commercially available fusion splicer.
First, the FFP sensor was mounted in one of the vacuum chucks of the
splicer so that the electrodes were centered a few fiber diameters past
the semireflective splice. An arc was applied to heat the fiber and let it
`droop` under its own weight. The fiber is then moved through the chuck
approximately one fiber diameter and the heating arc is reapplied. This
procedure was repeated until a 30 to 40 degree (arc length of 2 mm approx)
curve is produced. The typical double-pass loss was measured for these
curves to be approximately 16 dB. The use of high N.A. (numerical
aperture) fiber could lower this loss.
While the sensing regions in the arms of the strain rosette 80 are linear,
they may equally be configured in a curve as shown for example in FIG. 10.
In that figure, a strain rosette 90 has three arms 92, 94 and 96 each of
which, as bore is made up of one FFP sensor. In this case, however, the
two outside arms 92 and 96 have sensing regions which are curved with a
substantially constant radius of curvature.
A particular feature of the rosette 90 is that the sensing region of each
FFP sensor has a relatively large radius of curvature which results in
relative low light loss through the sensing regions, as opposed to the
light loss in the tighter bends of the arms 82, 84 of the rosette 80. In
addition, the rosette 90 occupies a smaller area (footprint) than the
rosette 80.
Of course, the strain-component reduction formulae will be different for
the curved sensing and the linear sensing regions. In addition, the fiber
may be heated directly through the splice if desired.
While the figures illustrate a strain rosette with three arms, it will be
understood that any reasonable number of arms may be used depending the
characteristics of the host structure and the type of analysis being made.
While the above embodiments make use of a mirror that is confined to the
first region having a size corresponding to that of the core, the first
region (and thus the mirror contained within it) may be smaller or larger
than the core depending on the area of light to be reflected. For example,
a larger mirror, that is one that extends into the cladding region may be
useful in capturing the tail portion of the evanescent field (and hence
greater reflectivity).
Alternatively, a smaller mirror or a number of discrete mirrors may be
useful if it is not desirable to capture all of the launched light in the
core. The first region may be formed of a number of sectors, each of which
is aligned with, and has the geometry of, the spatial modes being carried
by the core. One example of this is to produce a modal filter or modal
reflector. This can be seen for example in FIG. 11 which illustrates a
typical optical fiber core 100 with a number of modes, namely the main
gaussian mode 102 and a higher order mode 104a, 104b. By coating the fiber
endface with a reflective material in the region of one or more of the
higher order modes, light will be permitted to pass along the lowest order
gaussian mode while preventing light to pass through the coated higher
order mode. Of course, this technique could also be used to prevent light
from passing through the lowest order gaussian mode by coating the
appropriate region on the fiber endface. Similarly, other higher order
spatial modes can be filtered or reflected by choosing appropriately
shaped masks which match the geometry of the spatial modes. Similarly, a
number of discrete reflecting surfaces can be configured in such a fashion
to act as a Fresnel lens or reflector.
Although the above embodiments make use of a single-mode glass fiber, it
will be understood that the present technique may be used on other fibers,
such as multi-mode or multi-core fibers, and those made from materials
other than glass, including plastic and other transparent material. In
addition to its use with high birefringent optical fibers, the present
technique may also be used with low birefringent and other generic optical
fibers. In addition, while the above technique has been discussed in
connection with a low finesse (single reflection) Fabry-Perot sensor 10,
the technique may equally be applied to high finesse (multi-refection)
applications, such as amplitude modulators, spectral filters and laser
diode locking retroreflectors.
However, one should bear in mind that a high finesse cavity is required for
these applications. Referring to FIG. 5, a high finesse cavity will have
losses that are significantly lower than those of the low finess | | |
