Note: Descriptions are shown in the official language in which they were submitted.
CA 02395563 2002-08-08
NOVEL DEVICE AND METHOD FOR GAS
ANALYSIS
1, Introduction and Review of the Prior Art
This invention relates to the technical field of gas detection, diagnosis, and
analysis;
that is to say methods of analysis and detection of one or more gases present
in an
atmosphere, or gas components of a gas mixture, and apparatus for use in
performing such methods. This invention is of particular significance in
sensors,
sensor arrays and electronic nose fabrications.
Gas sensors are used for atmospheric monitoring in general, e.g. ~ coal mines,
offshore installations and industrial production facilities. Gas sensors are
also used
to control combustion processes in engine exhaust systems, etc., for both
economical and environm~tal reasons. In simple terms, a gas sensor performs as
the nose of a robot or an electronic control system. The requirement for
monitoring
toxic gases in the environment has steadily increased in recent years as
safety and
health professionals have become increasingly aware of the dangers posed by
these
substances. Greater awareness has further prompted government regulations to
address environmental monitoring and related issues. Although such monitoring
serves to protect the environment as a whole, the safety of people in the
workplace
continues to be of most vital concern. In this regard, most toxic gases have
various
levels or limits, set by industry associations or regulatory agencies.
Typically,
several levels are defined for each type of gas. For example, a threshold
limit value
sets the maximum allowable level of a gas that a Person may be exposed to for
an
eight-hour period, five days a week. The short-term exposure limit gives the
maximum exposure that a person may be exposed to for a fifteen minute period
not
to exceed four occurrences per 8-how work day. The permissible exposure limit
is
the maximum limit a person may be exposed to the gas for any time period.
A~ere~nce to these standards requires a toxic gas detector capable of accurate
detection of the toxic gases of interest. Further, as these various exposwe
limits
may span a large range of concentrations, the toxic gas detector must
accwately
measure the concentration over a wide range of concentrations. Thus Presently,
the
1
CA 02395563 2002-08-08
technical demands for devices which can selectively detect a certain gas or
diagnose
a prevailing gas or odder is rapidly increasing. Such devices, usually
referred to as
electronic noses, are now employed for security checks and identifications,
drug
defections, chemical assessments, food quality control, chemical and
biochemical
process control, ... etc.
A reliable and precise gas diagnosis is usually carried out by sophisticated
spectroscopy or other analytical systems. However, for the said applications
compactness, rigidity, simplicity, and low price are of importance. The
present
invention intends to provide a novel device and method for fast gas analysis.
The
device is compact, simple, rigid, stable and cost effective. The device can be
fabricated based on practically all the available types of gas sensors; and
plwality of
such devices can form reliable electronic noses.
According to the prior art, a selective gas detection, analysis and gas
diagnosis have
been subjects of marry investigations:
1. According to US Patent No. 4,911,892, a selective sensitivity to a pre-
selected
gas is achieved in a resistive semiconductor gas sensor by the means of
surface
decoration with noble metals such as platinum and palladium.
2. According to US Patent No. 4,347,732, selective detection of gases of
certain
molecular size range has been afforded by incorporation of molecular sieves
onto a
Zn0 based resistive gas sensor.
3. According to US Patent No. 6,284,545, a filter made of a high surface area
substrate impregnated with a silver salt or copper salt, incorporated onto a
gas
sensor is effective in reducing the cross-sensitivity to certain gas species.
4. US Patent 6,263,723, discloses a multilayer gas sensor element having
properties
capable of detecting methane and carbon monoxide selectively with only 1
sensor
by improving the selectivity of the semiconductor gas sensor.
2
CA 02395563 2002-08-08
5. CA Patent Application No. 2326210 discloses an electrochemical gas sensor
based on a polymer solid electrolyte which selectively detects only the
hydrogen
concentration in the atmosphere.
6. US Patent No. 6,353,225 discloses a method for the selective detection of
gases
by an optical gas sensor, in which the emission spectrum of a laser diode is
varied
by temperature variation to match the characteristic absorption line of the
target gas.
7. Canadian Patent No. CA 1248176 discloses a gas analysis method based on the
differential temperature variation of the detection sensitivities and
transient
responses of a resistive gas sensor for various gases. The same method has
pr~~t~ ~~~ selectivity by application on sensor arrays.
8. US Patent No. 399,122 discloses a selective sensitivity field effect
transistor
gas sensor, the selectivity of which arises from the selective gas absorbent
material
deposited on the gale electrode.
9. Chul Han Kwon et al. have described fabrication of a multi-layered
resistive gas
sensor which achieves considerable selective sensing via catalytic filtering
technology in: Sensors and Actuators B, 65,1-3, p. 327-330 (2000).
10. Different methods for selectivity enhancement in resistive gas sensors are
also
the subjects of the following articles:
a. A. Cirera, A. Vila A. Dibguez, A. Cabot, A. Cornet, and J. R. Morante,
Sensors and Actuators B 64,1-3, p. 65-69, (2000).
b. I. Simon, N. Biirsan, M~ Bauer, and U. Weimar, Sensors and Actuators B,
73, p. 1-26 (2001).
c. D: S. Lee, d: K. dung, J. -W. Lim, J. -S. Huh, D: D. Lee, Sensors and
Actuators: B, 77, p. 22&236 (2001).
Furthermore, higher levels of selectivity in gas sensing and diagnosis is
achieved by
using sensor arrays rather than a single sensor element. These are
exhaustively
covered in more recent patent applications and related technical journals. As
an
example, Rajnish K. Sharma et al. have presented resistive sensor arrays which
are selectively sensitive to carbon monoxide and hydrogen: Sensors and
Actuators
3
CA 02395563 2002-08-08
B: 72, p. 160-166 (2(101). Also, Canadian patent applications CA 2314237, CA
2215332, and CA 2264839 have presented many examples from the prior art
regarding electronic noses.
2. Summary of the Invention
As described in the prior art, the transient response of a gas sensor to one
target gas
is different from the same for another. The difference in the said transient
responses
has been used in the prior art (e.g. see, Canadian Patent No. CA
1248176) for selective gas detection, but not only those response are depend~t
on the nature of the target gas but also simultaneously on the concentration
of the
said gas. In this situation, the information regarding the nature and
concentration of
the target gas is intricately woven and extraction of diagnostic data from the
said
responses is difficult.
The present invention is based on the fact that the diffusion constant of a
target gas,
in air or in another gas, depends strongly on its molecular structure. For
example, if
the open end of an air filled capillary tube (pipe) of lrnown length is
inserted into a
chamber containing air polluted with a target gas of low concentration level,
the
time at which the concentration of the target gas at the other end of the tube
reaches
a predetermined portion (e.g. 20%) of that in the chamber d~ends strongly on
the
nature of the target gas, while the said time is, as shown later in the
present
disclosure, almost independent from the gas concentration level. The idea is
that, by
recording the progess of the diffusion process in such a tube and comparing
the
result with the results of previous experiences, the target gas can be
identified.
Another important advantage of the present invention over the other diagnostic
gas
sensors and methods of prior art is the fact that the diffusion theory is well
established (see e.g. J. Crank, KThe Mathematics of Diffusion", Ozford
University Press, Ozford, 1975) and almost accurate simulations of the
diffusion
process for one or more lrnown gases in a tube are possible. This advantage
further
facilitates comparison of the recorded data with the results of computer
simulations.
4
CA 02395563 2002-08-08
Yet another advantage of the present invention is the simplicity and stability
of the
process and the parts of the device that determine the transient response; the
structiue of the tube and the said diffusion process would not alter by aging
of the
system, while the transient responses relied upon in the prior art are related
to
complex electrochemical solid / gas interactions and prone to changes caused
by
aging. The method is also versatile, as it can be applied to almost all types
of gas
sensors; but those with fast responses and compact embodiments (e.g. see,
Canadian patent appUcation CA 2,267,881) are preferred. Ideally the response
of
the sensor employed must be much faster than the said diffusion process. In
this
case, the significant features of the transient response of the invented
device are
independent from the instabilities encountered with the sensor employed.
The instant invention uses one or more of the said tubes, each fiunished with
one or
more gas sensors to record the progress of the diffusion process of the target
gas
molecules along the said tube(s). The recorded or stored data regarding the
progress
of the said diffusion process are then used for the diagnosis of a target gas
or
analysis of a gas mixture. This is achieved by a comparison of the said data
with the
recordings from the previous experiences, calibration recordings, or
mathematical
simulation results. The basic idea described is applicable for various
important
technical needs. Dii~erent embodiments of the present invention are employed
for
fulfillment of its various aspects:
I. It is an object of this invention to diagnose a target gas by recording the
progress of its diffusion process against time in an air filled capillary tube
via one or more gas sensors installed appropriately in the tube.
2. It is another object of this invention to diagnose a target gas by
recording the
progress of its diffusion process against time in a capillary tube filled with
a
predetermined gas or gas mixture via one or more gas sensors appropriately
installed in the tube.
3. It is another object of this invention to measure the diffusion constant of
a
known gas in air or in an another gas or gas mixture by recording the
5
CA 02395563 2002-08-08
progess of its diffusion process against time in an air or gas filled
capillary
tube via one or more gas sensors installed appropriately in the tube.
4. It is another object of this invention to carry out one or more of the
functions
given in 1, 2, and 3, at an elevated or reduced temperature, by provision of
the appropriate heating or cooling apparatus for the said tube(s),
respectively.
5. It is yet another object of the present invention to carry out one or more
of
the functions presented in 1-4 at an elevated or reduced pressure.
6. It is another object of the present invention to perform the functions
described in 1-5 for two or more target gases simultaneously, where there
are more than one target gases present.
7. It is yet another object of the present invention to perform the functions
described in 1-6 by using plurality of identical or different (in geometry or
material) tubes furnished appropriately with plurality of identical or
different gas sensors.
8. It is yet another aspect of the present invention to present methods for
simple and fast extraction of the useful diagnostic and analytical data from
the transient responses recorded while performing any of the functions
related to the aspects described in I -7.
3. Brief Description of the Drawings
The above objects of the invention will become clearer by reference to the
attached
drawings in which:
Fig. I:
Schematic illustrations of 4 different embodiments of the device invented.
6
CA 02395563 2002-08-08
(a). One gas sensor is installed at the closed ~d of the diffusion pipe;
target gas
diffusion takes place from the open end. The diffusion starts upon the removal
of
the gas impermeable lid of the tube, or by insertion of the open end into a
target gas
contaminated camber.
(b). One gas sensor is installed at a predetermined distance (L) from an open
~d of
a long tube. The target gas diffuses from the open end of the pipe, as
described for
(a).
(c). The tail of the pipe shown in (b) is connected to an "exhaling" device
which
provides a flow of pure air or a known pure gas or gas mixture, in the
opposite
direction of the said difl'usion process, for facilitating a rapid recovery of
the device
from one test for starting another. The diffusion of the target gas starts
after the
flow is stopped.
(d). One gas sensor is installed in the diffusion pipe, preferably at its mid
point and
the diffusion takes place from both ends of the pipe simultaneously, the
diffusion
process starts as described in case of (a) .
Table 1: Diffusion constants of three organic vapors in air, as given by CRC
Handbook, and as resulted by the analysis of the transient response of the
prototype
fabricated comprising a quartz glass diffusion pipe of 4 mm bore and 70 mm
length
and Zn0 resistive gas s~sor.
Fig. 2:
(a). Analytically predicted normalized transient responses of the prototype
fabricated comprising a diffusion pipe of 4 mm bore and 70 mm length and a
fast
resistive gas sensor, simulated for methanol (A), ethanol (B), and butanol
(C).
(b). Traces of the first time derivatives of A (A'), B (B'), and C (C'); bars
1, 2, and
3 indicate the positions of the deflection points on A, B, and C,
respectively.
Fig. 3:
Normalized transient responses of the prototype fabricated comprising a quartz
glass diffusion pipe of 4 mm bore and 70 mm length and a Zn0 resistive gas
sensor,
for methanol (A), ethanol (B), and butanol (C) and their respective first time
derivatives A', B' and C'; the deflection points of A, B, and C are indicated
by the
bars 1, 2 and 3, respectively. The thickness of the bar 2 indicates the range
of
7
CA 02395563 2002-08-08
deflection times obtained when concentration of ethanol varied in the range of
500
- 5000 PPM in air. The trace S is the transiea~t response of the gas sensor
used
(without difl'usion pipe) given for the purpose of comparison.
4. Detailed Description of the Invention
Semiconductor gas sensors, particularly resistive semiconductor gas sensors
are
compact, fast and cost ei~ective. They can also be integrated with the signal
processing circuits required, but the information related to the nature of the
target
gas and its concentration is intricately woven in their responses, and
generally, the
diagnostic data extraction from the recorded responses of these sensors is
difficult.
A novel device is disclosed wherein the diagnostic information regarding a
target
gas is readily extractable from its transient response to the said target gas.
In its
most basic structure the invented device consists of a resistive gas sensor
and a
capillary tube. However, different embodiments of the invention can be
fabricated
by using other types of gas sensors, including but not limited to
semiconductor,
capacitive, optical, polymer, electrochemical, catalytic, capacitive,
electrothermal
types; the priority is inclined towards the sensors of compact sizes and fast
responses.
The fact that the transient responses of many common gas sensors vary from one
target gas to another has been used for gas diagnosis in the prior art. In
this
invention, however, the transient response employed for gas diagnosis, in
fact, is
not a feature of the gas sensor employed but is the result of the selective
retardation
imposed by the diffusion process through a specific diffusion pipe. The latter
process is physical, simple and stable in nature over the lifetime of the
device; while
the transient responses in the prior devices are determined by complicated
electro-
chemical processes taking place at their sensitive surfaces and can change
considerably during its lifetime.
The invention will be discussed in greater details for a few non-limiting
example
embodiments applied for some example applications:
8
CA 02395563 2002-08-08
Example 1:
In the example embodiment, presented schematically in Fig. l a, the device is
comprised a gas sensor (e.g. in the prototype fabricated a thick-film zinl:
oxide
resistive gas sensor was used), and a "diffusion pipe" of known diameter and
length
(e.g. in the prototype fabricated a quartz glass capillary of 4 mm internal
diameter
and 70 mm length was used). The gas sensor located at the closed end of the
pipe is
connected to a digital recorder. The pipe is filled, and in equilibrium, with
clean air
at atmospheric pressure. An impermeable lid physically isolates the pipe from
the
surrounding atmosphere which is polluted with a target gas of constant
concentration Co. (The existence of the said lid is not necessary in most of
the
embodiments and applications of the invention, as is described in the other
examples below.) In the following paragraphs a very brief quantitative
analysis of
the device is presented in a more technical language.
Upon the removal of the lid, at time (t) = 0, target gas molecules diffuse
through the
air along the pipe before affecting the gas sensor. The target gas
concentration
along the pipe is a function of time and distance from the open end of the
pipe,
C(x,t). The diffusion constant of the target gas in air, D, is related to its
molecular
structure. The quantitative relationship between D and the molecular
parameters of
the diffusing gas has been thoroughly discussed in the literature; a
comprehensive
account of the subject is presented by RE. Treybal, in KMass Transfer
Operation" 3'~ Edition, McGraw-Hill (1980). The effective target gas
concentration experienced by the gas sensor, C(L, t), and hence its transient
response, G(t), would depend on the nature of the target gas. The transient
responses of an ideal prototype, comprising the above given tube and an ideal
gas
sensor, for three different target gasses are simulated bellow. It is shown
that TG
can be diagnosed by a simple mathematical operation on G(t). For the sake of
the
simplicity of the calculations involved, it is assumed that the response of
the gas
sensor to the prevailing target gas is a step fimction the amplitude of which
is
proportional to the effective gas concentration. The amplitude normalized G(t)
and
C(L, t) would then be identical. In the prototype fabricated the response of
the gas
sensor employed was much faster than the target gas diffusion process through
the
pipe used, and the approximation applied proved to be valid.
9
CA 02395563 2002-08-08
C(x, t) was obtained by solving the diffusion equation for the geometry of the
diffusion pipe employed. For a pipe with an internal diameter of a few
millimeters,
Fick's law is applicable. However, Knudsen diffusion equation should be
considered for pipes of much smaller bores (see e.g., J. Szekly et al., KGas-
Solid
Reaction", Academic Press, 1979). The following boundary conditions were
applied:
dC(x, t) = 0 at x = L all t
dx
C(x,t)=Co at x=0 all t
Normalized solutions, obtained for three different values of D, are presented
in
Fig.2a. The D values inserted were those of CH30H, CZH30H, and C4H90H vapor
in air, as given in CRC Handbook of Materials Science Vo1.14 (1984), which are
also presented at the first row of Table 1. An algebraic solution was not
possible,
solutions were calculated numerically and results are presented in Fig.2. The
functions traced in Fig.2a are the transient responses of the prototype device
predicted for methanol (A), ethanol (B), and butanol (C), respectively.
The G(t) functions produced, as shown in Fig.2a, have deflection points. The
calculations carried out indicate that the deflection always occurred before
the
corresponding G(t) had gained 20% of its maximum value. The deflection time,
td,
is independent from Co and is related to the length of the pipe and the
diffusion
constant of the target gas in air (to is almost proportional to L~2/D). Hence
in
theory, to is a characteristic parameter of the prevailing target gas
molecule. In
Fig.2.b the predicted deflection times for methanol, ethanol and butanol
target gases
are depicted by the bars 1, 2, and 3, respectively. A', B', and C' curves in
Fig. 2b,
are the first time derivatives of A, B, and C, respectively, the maximum
points of
which coincide with the deflection times predicted.
The analytical results were verified experimentally. The sample device
fabricated
consisted of a 70 mm long, 4 mm bore quartz tube and a thick film zinc oxide
resistive gas sensor. The normalized transient response of the gas sensor
employed
(without the tube), to 4000 PPM of methanol is give in Fig.3 as trace (S). The
recorded transient responses of the fabricated device for methanol, ethanol
and
butanol are traced in Fig. 3 as A, B, and C, respectively. The respective
deflection
CA 02395563 2002-08-08
times of the traces A, B, and C are depicted as bars 1, 2, and 3. In order to
show that
the deflection time is independent from the target gas concentration, the
ethanol
concentration was varied from 500 to 5000 PPM. The width of the bar related to
ethanol, bar 2 in Fig.3, indicates the variation range of the deflection times
measured for various concentrations.
The minor differences between the experimental results (Fig.3) and the
predictions
of the simulation (Fig.2) are caused by the deviations of the response of the
gas
sensor used from the assumed step fimction. It was discovered that by using a
gas
sensor of faster response the transient response of the prototype fits more
accurately
to those predicted analytically. Appropriate corrections could be applied for
more
elaborate diagnostic tests, but even without arty correction, the simple
prototype
fabricated could reliably differentiate among the three alcohols mentioned,
regardless of their concentration. However, in real applications diagnosis is
carried
out by comparison of the recorded transient response with those of the
previous
experiences, which would make the said corrections unnecessary.
Another technically important point is that by using the above described
method of
gas diagnosis, the transient response of the device is needed to be traced
only up to
its deflection point, i.e. when the first time derivative of the response
passes through
a maximum. Hence, the time required for each diagnostic test is almost equal
to td;
the extension of the curves beyond this point in Fig.3 is for the purpose of
comparison with the simulation results given in Fig. 2.
It was discovered both analytically and experimentally that the differential
diffusion
retardation caused by the diffusion pipe was more profound in longer tubes; it
was
shown that the deflection time is almost proportional to the square of the
length of
the pipe. H~ce the diagnostic power of a device of longer pipe is higher where
it is
a slower device and its diagnostic tests will take longer.
It was discovered both analytically and experim~tally that the performance of
the
device is insensitive to the diameter of the pipe employed. This is valid for
pipes of
down to a few t~th of a millimeter bore. Further smaller bore diameter pipe,
although different in the analytical description of its characteristic
behavior, in
11
CA 02395563 2002-08-08
practice should indicate higher selectivity because of the comparable size of
the
pipe diameter and the "molecular mean free pass" of the target gases of
interest
which should cause even stronger diffusion discrimination among the said
target
gas molecules. That is to say, the diffusion time difference between methanol
and
S ethanol vapors would be higher in a 10 micrometer bore pipe than in a 1 mm
bore
one. Moreover, smaller diameter pipes would render compact devices,
particularly
when plurality of pipes and plurality of gas sensors are to form an array
sensor for
electronic nose applications.
It was discovered both analytically and experimentally that the device
operation
was independent from the material of the diffusion pipe employed when the pipe
diameter was in the millimeter range. Various ceramics, glass, metal and
polymer
pipes tested rendered similar results. However for much smaller diameter
pipes, and
particularly in the case of very low target gas concentrations, the
interaction
between the pipe wall and the diffusing target gas molecule becomes
considerably
important. This also can add to the selective sensitivity of the sensor in
special
conditions.
In the above described example embodiment of the instant invention, the simple
arrangement employed included only a proper attachment of a glass capillary to
a
non-selective, non diagnostic bare and simple resistive gas sensor. It was
shown
both analytically and experimentally that the combined device has very
reliable gas
diagnostic features.
Since the theoretical basis of the present invention was also briefly
described in the
Example 1, in the following examples however, the device structure and the
example application will be presented only.
Eaample 2:
The same embodiment of the invention as that used in Example 1 is employed
here
to diagnose and measure the concentration levels of an unknown target gas in
air, in
a closed chamber. The device was in equilibrium with the pure air outside the
chamber; the open end of the tube was inserted to the chamber when the
recording
of the response of the sensor was started simultaneously. The transient
response and
12
CA 02395563 2002-08-08
its first time derivative, i.e. G(t) and dG(t) / dt, were both traced on a CRT
screen.
The traces resembled those given in Fig.3 after amplitude normalization. It
was
observed that in about 30 seconds after the beginning of the recording, the
derivative trace passed a maximum point and started descending. This is the
deflection time of the transient response, and the measurement was complete.
Two
distinct features of the recorded transient response could easily be extracted
and
compared with the previous experiences of the device:
a The deflection time; measured as 29.5 seconds; where the best match among
the previous calibration recordings was the average deflection time obtained
for ethanol vapor. The diagnosis of the target gas was correct.
b. The value of the transient response (before normalization) at the
deflection
time was directly related to the concentration of ethanol vapor in the
chamber via the previous calibration charts drawn for ethanol vapor;
resulting 2400 PPM for the ethanol concentration. The calibration charts
were almost linear which fiwther simplified the concentration calculation
process.
In this experiment using the device and the method invented, the prevailing
target
gas was identified and its concentration in air was measured simultaneously.
All
this was achieved by using a general resistive gas sensor and a capillary
tube.
'The diagnostic decision regarding the nature of the gas in this example was
based
on a single readout from the "time axis", while the concentration information
was
read from the "amplitude axis"; this isolation of information regarding the
nature
and concentration of the target gas is of technical significance. The
fimctional steps
of recording, detection of the deflection point, deflection time readout,
amplitude
readout, and comparisons of the read figures with the calibration tables, all
can be
carried out online with a personal computer.
Example 3:
The embodiment described in Example 1, was applied for selective detection and
diagnosis of hydrogen in various gas mixtures. Owing to the fact that the
diffusion
constant of hydrog~ is much smaller than the other target gases of interest,
its
13
CA 02395563 2002-08-08
respective deflection point in the transient response of the device was
distinct and
easy to identify. Using the prototype device described above the deflection
point
related to hydrogen occurred at about 7 seconds. Using the same experimental
set
up as described for Example 2, the same prototype was able to detect hydrogen
as
the target gas in many gas mixtures. That is to say, the device performed as a
reliable hydrogen detector at the presence of other target gases such as
methane and
butane.
Example 4:
The prototype used in this experiment had a structure similar to that of
Examples 1-
3, different in that a micro-heater had been attached to the diffusion pipe so
that it
could provide a constant elevated temperature along the diffusion pipe.
Experiments
of Example 1 were repeated for pipe temperatures of 80,120, and 160 °C.
The
deflection time obtained for a particular gas reduced as the pipe temperature
increased, which is an indication of the dependence of the diffusion constant
of the
target gas on temperature of the diffusion medium. This dependence, best
described
as D(T), is also of diagnostic value. That is to say, by application of two or
more
devices, the pipes of which are at different temperatures, simultaneously for
performing arty of the tests of Examples 1- 3, the recorded data would have
resulted
the diffusion constant of the target gases) in three different temperatures,
making
the identification process more elaborate.
Example 5:
A different prototype was fabricated by using a 25 cm long and 4mm internal
diameter tube. The same resistive gas sensor as of the previous prototype was
located inside the pipe at a point 70 mm away from the open end. The prototype
is
schematically presented in Fig.lb. The open end of the pipe was inserted to
the gas
chamber for any of the above described tests. As the pipe is considered
infinitely
long in the calculations and in practice, whether the far end of the pipe is
open or
closed was of no significance in the analytical predictions and experimental
results.
The prototype was equally successful in performing the tests given in
Examplesl-4.
A theoretical analysis of this device was carried out by solution of the
diffusion
equation assuming that the diffusing species observe the tube as an infinitely
long
one. Hence the boundary conditions of the problem differed from those given in
14
CA 02395563 2002-08-08
Example 1. An exact solution of the differential equation is possible in this
case
(they are error functions), which makes the theoretical predictions easier.
The
mathematical analysis of the transit response of the device was also possible;
it
was carried out by a summation of the solutions obtained for each of the gas
components individually. The analytically predicted transient response for a
known
mixture of two different target gases in air, both at low concentration
levels, was
verified experimentally. The same method of analytical prediction was applied
to
the case of closed end prototypes described in previous examples; it was
equally
successful, but the numerical calculations involved required longer
calculation
times. The fact that the transient response of the invented sensor is
analytically
predictable for mad gas mixtures of interest is unique among the compact gas
sensors available, and is of great technical significance.
Eaalmple 6:
The recovery times of the prototypes illustrated in Fig. l a-b were long. That
is to
say, after completion of a test, it took a few minutes for the device to get
ready for
another measurement; because this time was necessary for the traces of the
target
gas to diffuse out of the pipe. Another embodiment of the present invention,
shown
schematically in Fig. l c, presented a much shorter recovery time. In this
prototype
the prototype described in Example 5 was connected to a pure air container of
positive relative pressure by a flexible capillary (e.g. pvc tubing) and a
controlled
electric valve. The valve is closed during the test, but it opens a weak
stream of
pure air briefly prior to each test The flow is in opposite direction of the
target gas
diffusion, forcing the target gas molecules out. That is to say, the prototype
had an
"exhaling" mechanism which shortened the recovery time.
Moreover, the prototype described did not need the impermeable lid mentioned
in
Example 1, recording started as the valve abruptly stopped the flow of the
pure air.
The recorder was synchronized with the valve as its closure initiated the
recording.
The amount of air flown was negligible compared to our test chamber and did
not
alter the conca~tration of the target gas in the latter. This prototype could
provide
multiple read outs from the atmosphere of a chamber when the open end of its
pipe
was inserted into the chamber. In other words, it could provide a continuous
monitoring of a chamber atmosphere, e.g. once in each minute or two, while no
CA 02395563 2002-08-08
disconnection from the chamber was required. The same ''exhaling" concept
could
be applied in conjunction with all of the above described prototypes. In case
of the
closed end diffusion pipes (as in Example 1 ) the exhaling stream was provided
through a small hole at the close end.
Example 7:
Yet another version of the prototype described in Example 6 was fabricated
which
was different in using a stream of a pure gas instead of pure air for
"exhaling". Such
a device facilitated the measurement of the diffusion coeffici~t of one gas in
another. As an example diffusion constant of hydrogen in nitrogen could
readily be
measured with this prototype. Also, by altering the present embodiment
according
to the technical feature described in Example 4, these measwements could be
carried out at differ~t temperatures. This prototype was modified by
installing two
or more sensors along the tube, so that more data readout was possible, each
related
to a different effective pipe length, at each test. The latter prototype
afforded a more
~~ra#e measurement of diffusion constant.
Assuming a sealed chamber, its pressure could be altered. The pipe is
physically in
equilibrium with the chamber immediately after the "exhaling", and the
diffusion
coefficients mentioned could be measured in different pressures. Hence the
modified prototype provided a precision instrument for the measurement of the
diffusion coe~cient of a gas in another gas or gas mixture, at predetermined
temperature and pressure.
Ezample 8:
The prototype described in Example 7 was used as an oxygen sensor. The stream
of
nitrogen or argon was used for device recovery. Oxygen could be detected and
its
partial pressure be measured based on the comparison of the transient response
obtained with the results of the previous experiences. (The electrical
conductance of
the sensor decreased as the oxygen conc~tration increased; a trend opposite to
all
of the above given target gas examples)
16
CA 02395563 2002-08-08
Ea~~rle 9:
The device described in Example 6, was applied to chambers containing air with
differ~t mixtures of three target gases of interest. The transient response
recorded
for each target gas mixture was distinctly different from the others. This
afforded
the identification of the gas mixture by a comparison of the recorded trace
with the
previous experiences and /or the results of computer simulations of the
transient
response for gas mixtures.
Example 10:
Using more than one of the disclosed devices, each being different in one or
more
structural features simultaneously for probing of an unlrnown gas mixture
would
provide a vast source of information regarding the gas mixture tested which
could
in turn afford a more accurate diagnosis and analysis of the mixture. The said
structural differences may include one or more of the: Gas sensor type used or
its
operating temperature, pipe effective length, pipe temperature, exhaling gas,
pipe
diameter or cross-sectional shape, pipe material, etc.
In this Example three devices with respective diffusion pipe lengths of 5, 10
and 20
cm, were applied simultaneously to the same experimental conditions as
described
in Example 9. The reliability and accuracy in determination of the gas
mixtures
increased.
17
