Note: Descriptions are shown in the official language in which they were submitted.
CA 02746235 2016-04-06
- 1 -
OPTICAL FIBER POLARIMETRIC CHEMICAL SENSOR WITH
MODULATED INJECTION OF SAMPLE FLUID
FIELD
This application relates to optical waveguide sensors, and more particularly
to optical fiber
polarimetric sensors for chemical analysis based on capillary gas
chromatography.
BACKGROUND
The traditional methods for capillary gas chromatography involve injecting a
sample for
analysis into a carrier gas. The sample is carried by the carrier gas along a
capillary having
an inner wall onto which the sample is partitioned, leading to a slower
migration of the
analyte vapors relative to the carrier gas. The partitioning involves a
portion of the sample
(which can be referred to as the partitioned portion) that bonds to the
capillary and that is
then released in a continuous process on the molecular scale. In the case
where the
capillary is coated with a fluid film, which is more common, the bonding
occurs by absorption
in the fluid film. Alternatively, the bonding can take place by adsorption on
a solid surface.
The migration rate (v) of a given analyte and the flow rate (u) of the carrier
gas are related
by: v = pu, where p is the retention ratio. p is the probability of an analyte
to be in the carrier
gas (1¨p being the probability of absorption). The retention ratio typically
varies with the
nature of the analyte, so that each analyte has a characteristic migration
rate in a given
sample. To facilitate understanding, reference is made to Figs. 1A and 1B
which
schematically illustrate the concentrations of two different analytes
migrating along the
length of a capillary at two different moments.
Each analyte can thus be thought of as travelling in the form of a distinct
packet, or zone of
higher concentration, having a characteristic migration rate. Typically, each
packet has a
sharp zone distribution at the inlet of the capillary, and this zone gradually
broadens as the
packet travels along the capillary. The zones associated to different analytes
also become
progressively more spaced due to the characteristic migration rates of the
analytes, which
then make them more distinct.
CA 02746235 2016-04-06
- 2 -
A change in the response of a suitable sensor (such as a thermal conductivity
sensor for
instance) placed at the exit of the capillary can indicate the passage of an
analyte. The
characteristics of the capillary and the flow rate of the carrier gas being
known, a detection at
a given moment can be associated with a migration rate characteristic of a
specific analyte.
U.S. Patent No. 7,403,673 teaches a new approach to chemical sensors. This
approach
involves guiding light in a birefringent optical waveguide that has a light
propagation volume
(such as a core) positioned adjacent to a capillary. The propagation volume
and the capillary
are close enough so that an analyte absorbed in the stationary phase can
interact with the
evanescent field of the guided light by altering the polarization state of the
light. Information
on the fluid to analyze is obtained from the detected variations in the
polarization state of the
light by measuring the light power transmitted through an optical polarizer
placed at the
output of the waveguide. This approach involves using a birefringent optical
waveguide that
has two different refractive indexes defining the birefringence B and the
polarization beat
length Lb. For a given light wavelength A, both parameters are related by: Lb
=
B
The beat length Lb is the distance along the birefringent optical waveguide
that corresponds
to a phase shift of 27-c between the two polarization modes of the light, and
it is thus the
length along the waveguide for which a polarization state of the light is
recovered.
In the case of an optical fiber polarimetric chemical sensor where, for
instance, linearly-
polarized light is injected with its polarization direction parallel to one of
the polarization axes
of the optical waveguide, the presence of locally absorbed vapor in the
capillary, which is
adjacent to the propagation volume, transfers some of the light to the other
polarization axis,
and can thus be said to constitute a coupling point between the polarization
axes. The new
polarization state, which can be elliptical for instance, then evolves towards
the optical fiber
output where it can be analyzed with a polarizer. When a single light
wavelength is used, as
analytes are moving at speed (migration rate) v and as polarization states
reproduce
themselves at each distance equal to the beat length Lb, the light power
transmitted through
an output polarizer will oscillate at an oscillation frequency or beat
frequency fb given by: ¨V .
L b
CA 02746235 2016-04-06
- 3 -
The transmittance of the optical waveguide, including the output polarizer,
can be given with
a good approximation by:
r
I(t) 1
= E K, cos t
Io 2 J=1 \, Lb
where co is a phase term that can be discarded. The summation is performed
over the
analytes present in the sample fluid. The Fourier transform of the detected
signal /(t) shows
spectral peaks having locations that correspond to the specific migration
rates of the
analytes.
In the above equation Ki is the strength of the polarization mode coupling
caused by the
presence of the analyte j. This parameter is related to the concentration of
each analyte and
to its distribution in the capillary fiber. It will be understood that for
very small quantity of
analytes the mode coupling can be very small, so that the amplitude of the
signal detected at
the oscillation frequency fb can be too weak to be detected in the Fourier
spectrum of /(t).
As a result, there remains room for improvements, particularly for increasing
the sensitivity of
such chemical sensors.
SUMMARY
The sensitivity of the former sensor is limited by the "single-pulse" nature
of the injection of
the sample into the carrier gas. Indeed, at any given time there is only one
coupling zone per
analyte along the length of the birefringent waveguide, so there is a limit to
the polarization
mode coupling caused by the partitioned molecules of the given molecule type
via the
evanescent field of the guided light when the analyte concentration is low.
Henceforth, the
strength of the signal to be measured, that is the amount of light that has
changed its
polarization state can be limited, thus affecting the limit of detection of
the sensor.
This limit of detection can be enhanced by increasing the value of K. One way
to achieve this
goal is to use periodically-varying (or multiple-pulse) sample injection
instead of a point
(single-pulse) injection. If the variations in the injection are done at an
injection frequency fi
selected to be equal to the polarization oscillation frequency fb, successive
analyte pulses
CA 02746235 2016-04-06
- 4 -
will be separated in the passage (typically a capillary) by a distance equal
to the beat length,
or to an integer multiple thereof. In this manner, the later analyte pulses in
the passage will
cause polarization couplings that will add to the amplitude of the
polarization coupling
caused by the earlier analyte pulses in the passage, given the beat length of
the birefringent
optical waveguide. The amplitude of the signal detected at the polarization
oscillation
frequency fb will be increased for the given analyte, thus facilitating its
detection from the
Fourier spectrum of the measured signal.
In accordance with one aspect, there is provided a method of analyzing a
sample fluid
comprising : injecting light in a propagation volume of a birefringent optical
waveguide
having a beat length; injecting and circulating a sample fluid along a passage
located
adjacent to the propagation volume, with a partitioned portion of the sample
fluid interacting
with an evanescent wave of the injected light, thereby affecting the
polarization state of the
light thereof; modulating the injection of the sample fluid over time in a
manner that a
plurality of zones of higher concentration of the sample along the passage are
spaced one
from each other by integer multiples of the beat length.
In accordance with another aspect, there is provided a chemical sensor
including a
birefringent optical waveguide having a beat length, a propagation volume, and
a passage
defined by a partitioning material, located adjacent to the propagation volume
for sample
fluid conveyed in the passage to interact with an evanescent wave of light
propagating in the
propagation volume and thereby affect the polarization state of the light, a
light source for
injecting light into the propagation volume, an optical detector for detecting
a periodical
variation of the polarization state of the light at an oscillation frequency,
caused by the flow
of the sample in the passage, and a modulator for injecting the sample into
the passage at a
concentration varying periodically with an injection frequency.
Because the use of a liquid film is more typical in the case of capillary
analysis, the
expression "absorbed" will be used herein as encompassing the expression
"adsorbed".
Further features and combinations thereof concerning the present improvements
will appear
to those skilled in the art following a reading of the instant disclosure.
CA 02746235 2016-04-06
- 5 -
BRIEF DESCRIPTION OF THE DRAWINGS
In the drawings,
Fig. 1 is a view that illustrates the propagation of two different analytes in
a capillary;
Fig. 2 is a view showing the dependence of IRI on the period, three cases are
considered
corresponding to three different values of the velocity of an analyte;
Fig. 3 is a schematic view of a first embodiment of a sensor;
Fig. 4 shows the temporal evolution of the concentration distribution of a
time-varying
injection (below) compared to the temporal evolution of a single-pulse
injection (above);
Fig. 5 is a schematic view of an alternative embodiment, where the
acceleration effect of the
carrier gas is compensated by a temperature gradient; and
Fig. 6 is a schematic view of another alternative embodiment, where the
acceleration effect
of the carrier gas is compensated by pressure control.
DETAILED DESCRIPTION
When the injection is periodical with an injection frequency fi adjusted to
correspond to the
oscillation frequency fb (or f(oscillation)), successive analyte pulses, or
zones of higher
sample concentration, can be separated from each other in the passage by a
distance A, or
pitch, being an integer multiple of the beat length (the integer multiple
being one or more), in
accordance with the following resonance condition:
1
_______________________________________________ = f (injection )= ¨ ppu = u
= f (oscillation)
(1)
T(injection) A Lb
In other words, a first injected pulse of the sample, having a limited volume,
begins to travel
along the passage. The investigated analyte being present in the passage, the
absorbed
molecules thereof cause polarization mode coupling, i.e., a transfer of a
portion of light from
a first polarization mode to a second polarization mode. However, this
coupling can be small
since there is a limited amount of molecules of the investigated analyte in
the sample. This
CA 02746235 2016-04-06
- 6 -
first injected pulse travels along the passage of a birefringent optical
waveguide over a given
distance, and the signal being so minute, the expected mode coupling may be
hardly
detectable. However, a second sample pulse is injected as the first injected
pulse reaches a
distance equal to the beat length. This second injected pulse also causes a
transfer of light
to the second polarization mode, and since its injection is precisely timed,
this additional
signal is in phase with the signal caused by the first pulse, thus enhancing
the amplitude of
the detected signal. If the birefringent optical waveguide is sufficiently
long, the pulses
subsequently injected can all cause corresponding polarization couplings, and
the detected
signal at the output end of the birefringent optical waveguide will then be
given by the sum of
each individual timed sample pulse signal contribution, all of these
contributions being in
phase. Henceforth, the collective signal sum stemming from the contributions
of the
individual pulses present in the passage can be detected in cases where the
contribution of
any given individual pulse would be too weak to allow reliable detection. The
sensitivity of
the sensor is thus enhanced.
Injection and diffusion
The injection can take the form of a series of pulses, each having at the
entrance of the
passage a concentration distribution given by fo(z). Each pulse then moves at
a velocity v
and diffuses, taking the form f(zJ) determined by the diffusion equation (the
so-called mass-
balance equation). At some stage, the number of pulses present simultaneously
in the
passage can reach a maximum value M = L/vT, L being the fiber length, and the
overall
concentration distribution can read as:
C(z,t) = 1nm-01 f (z,t ¨nT).
(2)
where T is the time delay between the injection of successive pulses, that is,
the reciprocal
of the injection frequency.
For the sake of simplicity, we consider the specific case of a Gaussian
initial pulse shape:
( z2
f (z ,0) = Ao exp --w2.
(3)
\, r r 01
CA 02746235 2016-04-06
- 7 -
This makes the model analytically tractable while having no impact on the main
conclusions.
The diffusion equation admits such a Gaussian solution. This means that each
of the
injected pulses maintains its Gaussian shape as it moves and diffuses along
the length of
the passage. The concentration distribution of the nth pulse spreads out and
its amplitude
decreases according to:
¨1/2
Wõ = WO i+ 4Deff _______________________________ tn
(4)
wo2
and
A
A n n=
(5)
+ 4Deff tn1/2
W02
where
tn = t ¨ nT (6)
corresponds to the time elapsed after the injection of the nth pulse. In Eqs.
(4) and (5), Deff
represents the effective diffusion coefficient of the analyte vapor. The
velocity v and the
effective diffusion coefficient Deff of any given analyte are respectively
related to the velocity
u and the diffusion constant D of the carrier gas through its probability of
non-absorption p:
v = pu and De = pD.
Sensor response
In presence of a single moving coupling zone, the capillary fiber sensor can
be characterized
by a periodic time variation in the light intensity /(t) transmitted through
an output polarizer
(and thus in the second polarization axis):
/(t) 127rB
= + K COS [¨õ ¨14)1 (7)
/0 2
with the normalized modulation amplitude K given by
CA 02746235 2016-04-06
- 8 -
K
(¨dC exp AP z} dz (8)
dz
where
= 27i B , B being the fiber birefringence. K depends on the fiber design and
on the
A
properties of the stationary phase for the analyte to be detected.
It can be shown that the periodic injection of Gaussian pulses with a period
Twill give rise to
a modulation amplitude that reads as:
K KAP ISI
(9)
where
S = /7-z- eLAP" A0 W0 r exp(- rn2)
(10)
n=0
with the phase delay (1) AOT = 2nvT/Lb and rn Al3W112 = TEWb1 Lb, Lb standing
for the
polarization beat length. In presence of a single pulse, the amplitude K-
decays exponentially
with time. The periodic injection converts this decay to a small periodic
variation of period T
through the time dependence of Wn. For our purpose, it is sufficient to
evaluate the sum S at
a time t corresponding to an integer multiple of the period T. The sum then
becomes a
geometric sum that can be evaluated analytically to yield:
S = -Fr e"4 ) AoWc, exp(- ro2)R (11)
where
o
ro=W
7r -- (12)
L b
and
1- 2' (13) R= (13)
1 - y
y being defined as
CA 02746235 2016-04-06
- 9 -
[ (47-c 2DeffT
y -. exp + illd ____________________________________________________________
(14)
E
b 1
The resonance principle that leads to an increase of the sensitivity can be
expressed in
mathematical form through Eqs. (11) to (14). For given injection conditions A,
and We, the
importance of the sensor's response depends on the injection period T, and
more
particularly on the phase delay (I) --- ApvT = 27cvT/Lb between each pulse
contribution. In
particular, IRI takes its maximum value when 0 = 27c. This corresponds to the
case where
the period T is chosen so as to make the pitch A = vT of the analyte equal to
the beat length
Lb, hence synchronizing all of the mode coupling contributions of the pulses.
To better illustrate the resonance principle, Fig. 2 shows the dependence of
IRI on the
injection period T. Cases corresponding to three different values of the
velocity v are
considered. The other parameters are: L = 10 m, Lb = 4 cm and Deff = 0.03
cm2/s, and they
can be considered as typical. With these values for the parameters, the
injection period Tres
leading to the resonance condition is equal to 2.0 s, 0.4 s and 0.16 s,
respectively.
The higher the speed of the sample fluid, the sharper the resonance curves are
and so for
the maximum value of IRI. This is due mostly to the decrease of the period
Tres = Lb/v leading
to the resonance condition as the speed increases. This implies that each
pulse has not
enough time for diffusing appreciably before the next one is injected, so that
the contribution
of each pulse to the mode coupling is more important.
One can also notice the presence of secondary resonances. They correspond to
the cases
where vT = q Lb with q = 2, 3, 4,.... In those cases, the number M of pulses
present along
the length of the fiber is lower, but the main reason for the lower values of
IRI is that each
pulse spreads out more before the injection of the next pulse.
In practice, the fiber length and the beat length can be such that the number
M of pulses is
very high, so that 7m+1 ,---, 0. Moreover, for the typically small values of
the diffusion coefficient,
the quantity y is well approximated by the first two terms of its Taylor
expansion. The
maximum and minimum values of R are then approximately given by:
CA 02746235 2016-04-06
- 10 -
V Lb U Lb
Rmax
(15)
4g2 Deff 47-/-2 D
and
-2- (16)
Eq. (15) implies that the maximum gain in sensitivity does not depend on the
analyte under
analysis. It is primarily determined by the velocity u of the carrier gas,
which can be easily
modified by changing the pressure conditions.
Finally, it is worth mentioning that the dependence on the initial conditions
is only through
Eqs. (11) and (12) and that R does not depend on those conditions.
Referring now to Fig. 3, an example of a sensor 10 is illustrated. In this
example, the optical
waveguide 12 is an optical fiber 12a. The optical fiber 12a has a passage 14
for receiving
the fluid to analyze. In this embodiment, the passage 14 is a capillary 14a
positioned
adjacent to a core 16 which acts as a propagation volume in which the light
emitted from a
light source 26 is guided. To favor the partition effect, the inner surface of
the capillary 14a is
coated with a partitioning material such as a liquid of high viscosity like
polydimethylsiloxane,
which is more commonly known as PDMS. In this particular embodiment, the
partition effect
takes place by absorption of a portion of the analytes by the partitioning
material (here a film
of PDMS) rather than by the inner surface of the capillary 14a. Further, in
this embodiment,
the light can be monochromatic, though it can alternatively be broadband, as
will be detailed
below. The light source 26 emits linearly-polarized light that is then
injected in the core 16 of
the fiber 12a. The polarization direction of the light can be parallel to
either of the two
polarization axes of the birefringent optical fiber 12a. Alternatively, the
state of polarization of
the light can be measured in order to determine any subsequent change in its
polarization
state that would be caused by the presence of analytes in the fluid. After
travelling along the
optical fiber 12a, and having interacted with the partitioned analytes via its
evanescent field,
the light exits from the output end of the fiber 12a. In the case of
monochromatic light
injected in a linearly polarized state with its direction parallel to a first
polarization axis of the
fiber 12a, the optical detector 28 can be positioned downstream an optical
polarizer 18
having its axis oriented along the second orthogonal polarization axis of the
fiber 12a. The
CA 02746235 2016-04-06
- 11 -
detection of light intensity polarized parallel to the second polarization
axis is thus an
indication of a coupling effect caused by the interaction. The optical
detector 28 can be
replaced by a spectrophotometer in cases where the light source 26 emits
broadband light.
The power of the light emitted from the source 26 can be continuously measured
in a
manner to compensate for power fluctuations. In an alternative embodiment, the
polarizer 18
can be replaced by a polarization beamsplitter to measure separately the power
of the light
polarized along each of both orthogonal polarization axes. This can also serve
in
compensating for power fluctuations of the light source 26 or for the optical
losses in the
system.
In this embodiment, the passage 14 in the optical fiber 12a is used to channel
the sample
fluid (gas) to analyze. The gas can be pressurized with a pump 20 to the
desired pressure
and then transferred to a modulator 22 that injects the gas in the passage 14
according to a
periodic modulation of the concentration. The modulator 22 can, for example,
use the effect
of cold trapping in a capillary containing a stationary phase, or it can
operate by periodical
insertion of a sample vector gas in the carrier gas flux. Known devices can be
used in this
purpose, such as devices used in the GCxGC technique for instance. Other means
of
providing a varying rate of injection can be used as well.
The optical fiber 12a can be placed in an oven 24 to better control the speed
of the carrier
gas, particularly when an increase of the migration rates of the analytes is
desired. The
modulator 22 can be positioned either inside or outside of the oven 24.
The operation of the sensor 10 can be controlled by a data acquisition system
36, which can
also control the modulator 22, the pump 20 (flow rate and pressure), and the
optical detector
28, for instance.
In practice, the injection of satisfactorily timed distinct pulses of sample
gas can be
challenging, so that the injection can be modulated in a sinusoidal-like
manner, for instance.
Figs. 4A to 4D illustrate an example of such a modulation, by depicting the
longitudinal
distribution of an analyte concentration at different times t over one period
of modulation of
the injection. In each successive figure, the solid curve plotted in the lower
graph shows the
CA 02746235 2016-04-06
- 12 -
concentration distribution resulting from a sinusoidally-varying injection
whereas the dashed
curve plotted in the upper graph holds for a single-pulse injection. Z
represents the distance
along the capillary 14a. In this example, provided for the sole purpose of
illustration, the
injection period is 2 s, the velocity is 2 cm/s and the effective diffusion
coefficient Deff is 2
cm2/s. Figs. 4A to 4D progress successively from t = 9 s; t = 9.67 s; t = 10.3
s; and t = 11 s.
Typically, the injection frequency can be established as a function of a
predetermined
analyte for which the sensor is adapted to detect. This can be done by first
determining the
beat frequency for a specific analyte and test conditions, such as by testing
the sensor with a
sample of known analyte concentration, and then operating the sensor with a
sample
injection frequency set to the beat frequency before testing the presence of
the analyte in
actual samples.
Alternatively, or additionally to establishing the injection frequency
beforehand, one can scan
several injection frequencies, either by discrete steps or in continuous
manner, for detecting
the presence of peaks in the sensor's response associated with a given variety
of molecule
types. If a signal is obtained at a given injection frequency, one can then
fine tune the
injection frequency to attempt at strengthening the amplitude of the signal
and to clearly
establish the injection frequency at which a resonance is observed.
Unfortunately, the
scanning of the injection frequency may reveal as time consuming with some
practical
embodiments. Another way to look for unpredicted resonances would be to keep
the
injection frequency constant and then to vary the speed (u) of the carrier
gas, such as by
varying the pressure differential.
Another way of obtaining data is to measure the transmission spectrum of the
fiber after the
optical polarizer 18 for injection of broadband light. This can allow the
detection of more than
one analyte simultaneously. In fact, it can be noted that for multiple
analytes injected at a
same frequency, as the beat length depends on the wavelength of light, there
will be
resonances in all cases where the pitch A associated to an analyte equals the
beat length
Lb. A resonance peak can thus be expected in the transmitted light spectrum
for each
analyte present. A numerical simulation has demonstrated, for instance, that
for p = 0.50 and
0.52, u = 85 cm/s, and Ap = 0.0256 cm-1, resonance peaks can be expected at
wavelengths
of 1.297 pm and 1.349 pm, respectively. The numerical simulation also showed
that the
CA 02746235 2016-04-06
- 13 -
resolution was greater when the capillary 14a was longer. Accordingly, the
injection can be
modulated with more than one frequency
The above description set the groundwork of the theory, but it will be noted
that it was
assumed up to now that the speed of the carrier gas, or the migration speed of
any given
analyte, would be constant along the entire length of the capillary 14a. In
practice, it is likely
that the speed will increase when approaching the output of the capillary due
to the effect of
decompression of the carrier gas.
In fact, the evolution (gradient) of the speed u of the carrier gas with the
distance z along the
capillary is described by the following equation, as previously presented for
instance in T. M.
Nahir and K. M. Morales (2000) "Constant holdup times in gas chromatography by
programming of column temperature and inlet pressure", Analytical Chemistry,
vol. 72, pp.
4667-4670:
( 2\ z ( pout 2
r 2
pm 1 pout2 1 \ T
U(Z) _____________________________________________________ Y2
1 __________________________________________________________________________
(17)
16Lq p. L
where pin and pout are the inlet and outlet pressures, respectively, n is the
viscosity of the gas
while L and r stand for the length and radius of the capillary 14a,
respectively.
Referring back to Eq. (1), it can be noted that as the gas flows through the
capillary 14a, an
increase of its speed u results in a corresponding increase of A as a function
of z for a given
injection frequency. Any variation of A with z will limit the sensitivity of
the sensor 10 since it
would broaden the frequency peak associated with a given analyte.
These limits can be at least partially overcome in several ways, three of
which are described
below.
Compensation of the Effect of Acceleration by Temperature Variation
A first way to compensate for the effect of the carrier gas acceleration is to
lower the
temperature (F) along the capillary 14a to increase absorption of the analytes
and therefore
to decrease the retention ratio p. A complete compensation for u(z) is sought,
namely:
CA 02746235 2016-04-06
- 14 -
AF (4) oc 1
(18)
u(z)
The decompression leading to the acceleration of the carrier gas will remain
present with this
embodiment, but the analyte will be subjected to higher absorption as it flows
through the
capillary 14a. Its migration rate can remain stable by decreasing relative to
the increasing
carrier speed. The variations of the carrier speed as a function of
temperature change can
also be taken into account to achieve higher precision.
Since p will diminish when reducing F, we can expect the acceleration to be
less and less
important as the temperature lowers, as compared with an embodiment where the
temperature would be homogeneous.
It will be noted that this type of compensation is of the first order, and it
can be optimized for
a specific analyte only. The variations of p with F are likely to depend on
the nature of the
analyte, in accordance with Arrhenius law. Nonetheless, it can be practical
for
monochromatic sensors adapted for the detection or quantification of a single
analyte.
Fig. 5 shows an embodiment for compensation by temperature variation. In this
embodiment, a coil 30 or another temperature control device imposes a
temperature
gradient along the length of the optical waveguide 12. It can be used either
to reduce the
temperature of the optical waveguide 12 progressively toward its exit end or
to increase the
temperature of the optical waveguide 12 toward its input end.
Reduction of Acceleration by Pressure Control
A way to reduce the speed gradient u(z) of the carrier gas along the capillary
14a consists in
diminishing the pressure differential between the inlet and outlet of the
capillary 14a and by
increasing its length. In the latter case, one could connect a post column to
the outlet of the
capillary 14a.
By way of example, Fig. 6 illustrates an embodiment for the sensor 10 in which
the outlet of
the capillary 14a is connected to a post column 32. Typical values for the
lengths of the
capillary 14a and of the post column 32 are 5 m and 16 m, respectively.
CA 02746235 2016-04-06
- 15 -
Compensation of the Effect of Acceleration by Lowering Birefringence
An additional way to compensate for the effect of the acceleration of the
carrier gas is to
impose a change in the birefringence of the optical fiber 12a along its
length, in such a
manner that the beat length Lb varies with z according to the increasing pitch
A of the
analyte. The variation can be the following:
Lb(Z) CC U(Z)
(19)
Such a variation of the fiber birefringence can be achieved in several ways. A
first way would
be to coil the birefringent optical fiber 12a in a spiral, such as around a
conical cylinder.
Alternative ways include designing the optical fiber 12a in such a manner that
a variation of
pressure or of temperature would have a satisfactory effect on the
birefringence
characteristics along its length.
Of course, two or more ways to compensate for the acceleration effect can be
combined in
some embodiments to get better results.
It will be understood that the embodiments shown in Figs. 3, 5 and 6 are
exemplary only,
and many alternative embodiments can be realized. For instance, U.S. Patent
No. 7,403,673
illustrates different forms of birefringent waveguides that can be used to
channel the sample
fluid, and several optical assemblies which allow to inject both light and
fluid in the optical
waveguide.
The examples described above and illustrated are intended to be exemplary
only. The scope
is indicated by the appended claims.
