"A METHOD FOR DRIFT COMPENSATED MEASUREMENT OF GAS CONCENTRATION, AND A PHOTOACOUSTICAL GAS SENSOR"
The present invention relates generally to measurement or detection of gas in a room, and more particularly the invention relates to a photoacoustical gas sensor and a measuring method wherein measures have been taken to compensate for drift at measurement end or in a light source.
Initially it is referred to the applicant's own previous patent application no. 1997.5447 which in principle relates to a dynamic pressure sensor, but which application also discloses a photoacoustical gas detection sensor based on such a dynamic pressure sensor. The measurement principle is that pulsed light from a light source passes through a (external) room to be investigated with regard to a certain gas, and falls toward a wavelength matched window in a wall surrounding a closed chamber. The light wavelengths match or cover the absorption line spectrum of said certain gas, so that presence of the gas in the room will influence the intensity of the light that reaches and passes through the window. The same gas that is to be detected/measured as to concentration in the "external" room, is contained also in the closed chamber. The gas in the chamber will then be heated/cooled intermittently at the same rate as the light pulses, and the chamber pressure will vary at the same rate, and to a degree that is a direct function of the concentration of the gas searched for in the "external" room. The dynamic pressure variations that arise in the gas in the chamber, can be sensed/measured by a dynamic pressure sensor.
The dynamic pressure sensor used in the above mentioned NO 1997.5447 is of a type having a diaphragm with a mechano-electric transducer. The closed chamber is preferably divided in two by the diaphragm, and a narrow restriction or opening along parts of the diaphragm edge, makes sure that the gas is in both chamber parts, and that pressure equalization is achieved for sufficiently low sound frequencies, so that frequencies can only be measured above a certain "corner frequency", the position of which depends on chamber volume, restriction size and diaphragm volume displacement.
It turns out that also the gas sensor type mentioned hereabove may experience drift, i.e. that the measurement signal that is achieved, exhibits a slow
variation over time due to e.g. collection of contamination deposits on sensitive surfaces, variation in the light source etc. The goal of the present invention is to provide a solution to the drift problem for this type of gas sensor.
From US 5,159,411 is previously known a method for investigating a gas mixture contained in a closed chamber, where another gas in the mixture masks important absorption lines of a certain gas in the mixture, which certain gas is of interest. Therefore, a third gas is added to the mixture in the chamber, to achieve so-called "kinetic" cooling. Pulsed laser light is transmitted through the chamber, which chamber has two windows and an inside microphone, and sound picked up by the microphone is measured, and the intensity of the remaining light leaving the chamber, is measured by means of a light detector. One sees that even if one further gas is added to a gas mixture to be measured, this is a measurement setup in which no search is made for a gas in a room by means of a shut-up amount of the same gas in a special chamber. This means that the measurement problems and the measurement methodology are quite different than in the present invention. The drift problem in a sensor of the type mentioned above in connection with NO 1997.5447, cannot be solved using US 5,159,411 as a basis. A gas sensor of somewhat more cognate type is known from US 5,616,826, which utilizes a closed chamber with another gas than the one to be measured, however with a similar absorption spectrum. This sensor is custom-designed for measuring unstable gases like e.g. ozone. However, the drift problem is not approached at all in this US patent.
In accordance with the present invention, a solution is provided to the problem that has been sketched above. In accordance with the invention there is provided a method for drift compensated detection or concentration measurement of a gas in a room, where light is directed from at least one light source through at least part of the room toward and through a window constituting part of a wall surrounding a closed chamber containing the same gas as the one to be detected or measured, the light is pulsed in order to heat the gas in the chamber intermittently by absorption or scattering of light energy, at least one of the light sources emitting light having wavelengths matching the absorption line spectrum of
the gas, and the chamber containing a diaphragm with a mechano-electric transducer for picking up sound generated by the gas through the above mentioned intermittent heating. The method in characterized in that 5 - a special reference gas not expected to appear naturally in the room, and having a different absorption spectrum, however with lines in the same wavelength range as the gas to be detected or measured, is added in advance to the chamber together with the measurement gas, and that the light is sent into the closed chamber with two different values of at least ιo one of the parameters spectral content, pulse frequency and phase, whereby measurement gas and reference gas will provide detectably different signal contributions from the transducer for the two different parameter values.
In accordance with the invention, there is further provided a photo- i5 acoustical sensor for drift compensated detection of a gas or measurement of concentration of gas in a room, said gas sensor comprising at least one pulsed light source for transmitting light through at least part of the room and through a window constituting part of a wall surrounding a closed chamber containing the same gas as the one to be detected or measured, and which chamber also 20 contains a diaphragm having a mechano-electric transducer for picking up sound generated by the contained gas when/if it is heated intermittently by absorption of light energy, the light source(s) emitting light having wavelengths covering the absorption line spectrum of the gas.
The gas sensor of the invention is characterized in that 25 - the chamber contains also a reference gas not expected to appear naturally in the room, and which reference gas has a different absorption spectrum, however with lines in the same wavelength range as the gas to be detected or measured, and that the light source(s), light path defining means and the closed chamber are 3o arranged in such a manner that at least one of the parameters spectral content, pulse frequency and phase can be provided with two different values, in order that measurement gas and reference gas in the chamber
may provide detectably different signal contributions from the transducer for the two different parameter values.
Advantageous embodiments of the method and the photoacoustical gas sensor in accordance with the invention, appear from the appended dependent patent claims 2-7 and 9-17.
In the following, the invention shall be illuminated more closely through a mention of exemplary embodiments, and it is at the same time referred to the appended drawings, where figs. 1a, b, c show absorption spectra for carbon dioxide gas, methanol vapor and ethanol vapor respectively; fig. 2 shows an absorption spectrum of laughing gas Ν2O, which gas can be used for reference gas; fig. 3 shows schematically a measurement set-up that can be used in certain embodiments of the invention, using two light sources and one window in a measurement cell; fig. 4 shows schematically an alternative set-up of apparatus that can be used in another embodiment of the invention, using two light sources directing light toward respective parts of the measurement cell through separate windows; fig. 5 shows schematically a further embodiment in accordance with the invention, using one light source and a two-part measurement cell with separate windows; and fig. 6 shows schematically the most preferred embodiment of the invention, in which there is used one light source and one window in the measurement cell, and wherein the light source is pulsed variably. As mentioned above, the invention relates to photoacoustical sensors, and in particular sensors having a gas in a closed chamber together with a microphone element, i.e. the same gas that it is interesting to detect in "the surroundings". Such gas sensors will be subject to drift, in similarity with other types of gas sensors. Often some of the drift will be due to a deposit on the sensitive surfaces. One example of this, is that in a kitchen with a gas stove, where it is clearly of interest to install a gas sensor, most surfaces will little by little be covered by fatty smoke. Photoacoustical sensors will also be influenced by a deposit on the
surfaces. The signal from such photoacoustical gas sensors will also be subject to drift if the light source used, exhibits drift.
A solution to the two types of drift problem mentioned above, will be to provide a reference signal. The idea of the invention is then to fill additional gas into the chamber. This gas must have absorption lines in the same wavelength range as the gas that it is desirable to investigate. It is an important premise that the reference gas used as an additional gas in a chamber, does not appear in a natural manner in the measurement area where the sensor is mounted. In figs. 1a, b, c appear the most important absorption lines for carbon dioxide gas (CO2), methanol vapor and ethanol vapor, respectively. In various connections it may be of importance to monitor these gases. One will note, see fig. 1a, that carbon dioxide exhibits a relatively wide peak in its absorption spectrum, in a wave number range from about 2.300 to about 2.280 to about 2.400. (Wave number defined as 2π/λ, λ wavelength). s For methanol, see fig. 1b, one finds a sharp peak close to wave number
1.000, and a lesser peak close to wave number 3.000. For ethanol, see fig. 1c, the highest peak is found near wave number 1.000, as well as a rather distinct peak near wave number 3.000, Laughing gas, Ν2O, will be a usable reference gas for all of the three gases mentioned. This is because laughing gas has a distinct 0 peak in its absorption spectrum near wave number 2.200, and none of the gases mentioned above exhibit any peak close to wave number 2.200 in their absorption spectra. There are of course environments where laughing gas may appear, and where such a reference gas will clearly not be favorable, but it is assumed that laughing gas quite often will be a usable reference gas for the three measurement 5 gases mentioned above.
The electronic signal processing unit that is used for processing signals from the microphone or transducer element in the measurement chamber, must in some manner be able to distinguish between the signal from the measurement gas and from the reference gas (i.e. the additional gas). In other words, it is an o important point that when the light entering the measurement chamber has such characteristics as to be absorbed specially by the measurement gas, the light vibration that is picked up, e.g. with a frequency according to the flash frequency of this light, will provide a signal that is separable from the signal belonging to the
light absorbed specially by the reference gas, and which is then flashed using another frequency. E.g. signal amplitudes associated with the two flash frequencies/pitches that are picked up, may be possible to distinguish. Then, when the signal from the reference gas is independent of gas presence in the room to be investigated, since no such reference gas will ever appear in the room in a significant concentration, possible changes in the signal from the reference gas will only be due to typical drift factors. The same drift factors will influence the signal from the measurement gas, but the reference signal can then be used for drift correction. Practical examples of arrangements that can be used, will be illuminated in the following:
In fig. 3 appears schematically a measurement cell filled with measurement gas (M) and reference gas (R). The measurement ceil is in this case depicted as a two-part measurement cell wherein a diaphragm constitutes a partition dividing the cell in two chambers, while a restriction (a narrow opening) ensures that both chambers contain the same gas. In this type of arrangement it is not necessary for the invention that the measurement cell is divided in two, in principle it would be sufficient with only one chamber having a transducer element (or microphone) of some type therein. Light is transmitted toward the chamber window through the room to be measured/investigated regarding the measurement gas, from two separate light sources. The light sources are provided with respective, specially adapted filters, and both light sources are flashed by switching on/off (i.e. pulsing of the light source itself), or by "chopping" using an electro-optical or mechanical chopper. The filters are specially matched in order to absorb just those absorption lines that are of interest for the measurement gas and the reference gas, respectively, and the filters may then e.g. be constituted by light transmission cells containing just one respective of the two gases in question. The light from the light source having an M filter (i.e. a filter containing the measurement gas or a filter having corresponding absoφtion lines) can then only measure R gas, and light from the light source having an R filter, is only able to measure M gas. Disregarding drift factors, this means that light from the light source with the M filter will constantly give rise to a constant R signal, since R gas will never appear in the
room, which means a constant R gas state (=0) in the room, and the R gas amount in the measurement cell will be constant.
Contrary thereto (still when disregarding drift factors), the light from the source with an R filter will be able to respond (in an ordinary manner) to variable occurrence of M gas in the room, i.e. it will be attenuated according to the amount of M gas, and the light reaching the measurement cell will then be attenuated just at those wavelengths where the M gas inside the cell is able to absorb it. Less light will then be absorbed by the M gas inside the cell, i.e. a weaker sound signal will be produced at the frequency in question. As previously mentioned, the two light sources are preferably flashed with different frequencies. In the case just described, two characteristic sound frequencies or pitches can be listened to inside the measurement cell, one pitch will stay at a constant intensity, whereas the other will be subject to intensity variation in an inverse relation to the amount of measurement gas (measurement gas concentration) in the room outside the cell. Correction with respect to drift can now be made by means of the inherently constant R signal, which signal will only exhibit a slow variation due to drift. The correction may be of subjective or relative type, depending on the signal nature.
The flashing series of the two lamps can be run simultaneously, since the two different flash frequencies nevertheless provide characteristic and separate pitches inside the measurement cell simultaneously, which pitches can be distinguished through signal processing that there is no need to illuminate further here. However, it is also possible to alternate between the two light sources, so that pulse trains are sent in an alternating manner. In the set-up shown in fig. 3, the two light sources must not necessarily flash at different frequencies. One and the same frequency may also be used, but the flashing must then be made in such a manner that the two light pulse trains are in mutually opposite phases. When such opposite phase flashing is used, and when none of the gases are present in the room, both pressure signals may possibly be equal in the measurement cell, i.e. the sensor unit will hardly detect any noticeable vibration at all, since both halves of a period will give the same amplitude. But when the gas searched for, appears, an imbalance will appear
between the two period halves, which imbalance will give a clearly detectable difference.
The chopping/flashing is typically executed using frequencies in the range from about 1 Hz to about 10 kHz. Another measurement arrangement is shown schematically in fig. 4, where the measurement cell is divided in two chambers by means of a diaphragm having a restriction, so that the same gases are present in both chambers. Additionally, the diaphragm has a mechano-electric transducer able to deliver electric signals in accordance with the movements of the diaphragm. Each respective chamber has a window, and the two light sources are arranged in such a manner that they transmit light to respective chambers in the measurement cell. For the rest, the same technique is used as in the embodiment shown in fig. 3, with adapted M and R filters together with the light sources. Both light paths toward the measurement cell pass through the room to be investigated/measured. (To ensure this, the two light sources may possibly be arranged close to each other, and transmit light beams rather close to each other, toward a not shown receiver unit that guides the light beams to the respective measurement cell windows by means of mirrors or possibly fiber optics.)
In this case the two light sources can be flashed at the same frequency and in phase with each other, or two different frequencies can be used. Also in this case, i.e. in the case with same frequency and phase for light toward the two chambers, it is so that a good resultant signal is achieved when an imbalance arises between the two pressure signals toward the diaphragm from two sides, when the measurement gas starts to appear in the measurement room, while when no measurement gas is present, a balance can be set up with equally strong pressure signals from both sides, i.e. a very low output signal from the sensor due to such a balance.
It has been presupposed that the restriction and the diaphragm give the opportunity for measuring sufficiently low frequencies, i.e. that the chopping frequency/frequencies lie above the corner frequency of the diaphragm.
When different frequencies are used, detectable signals having correspondingly different frequencies will be measurable by themselves, i.e. the diaphragm will respond to pressure frequencies from both sides.
An advantage of using the same frequency, is, as explained above, that there will be a small signal when there is no gas in the measurement room between light sources and measurement cell, while the signal will grow significantly when there is a rise in the concentration of the gas searched for. However, it is to be noted that as long as two light sources are used, there will still be sensitivity to drift in the very light sources. But, since as mentioned above, it is possible to use two light sources flashing in phase with each other, it is obviously possible to use only one single flashing light source, together with an arrangement of mirrors or e.g. fiber optics. Then, two light paths must be defined out from the same light source, M and R filters must be placed in respective light paths, and the light must be sent toward the measurement cell from two sides (mostly such as shown in fig. 4), see fig. 5, which shows such an arrangement with only one light source, schematically. In this case the set-up is also rather similar to the arrangement shown in fig. 4, and the operation is analogous to the "same frequency and phase" version mentioned above.
However, it is not very practical to have light enter from two sides toward the measurement cell, and the use of mirrors will also complicate the system to a certain degree. It is further a clear premise for these methods, that all surfaces are contaminated to approximately the same degree. Referring to fig. 6, finally a seemingly promising method shall be mentioned, which method is assumed to be the preferred method. In this case the distinction between the signal from the two gases (M, R) is made by utilizing the fact that the gases have different relaxation times. A light source of the same type as previously mentioned, can be used, the light path will be simple, from the flashing light source through the room where the gas of interest shall be detected, and direct to the measurement cell, and a filter is not necessary. A microcontroller (not shown) controls the light frequency, and now and then it will change into another frequency than the normal frequency, in order to provide a reference signal. The two frequencies that are used, are adapted to the relaxation times of the two gases, i.e. the relaxation times associated with light excitation of certain excitation modi for the gas molecules. If e.g. CO2 is the measurement gas, one such relaxation time of interest, is known to be about 8 μs. Laughing gas (Ν2O) may, as previously mentioned, be an interesting reference gas in connection with CO2, and
a corresponding relaxation time for Ν2O is about 0,8 μs, which is lower by a factor of 10. The two pulse frequencies selected may then typically exhibit the same ratio, i.e. one frequency may be selected ten times as high as the other one. Pulse frequencies of interest are generally situated preferably in the range of about 10 Hz to a few kilohertz, and the actual frequencies that are used, will depend on the dimensions and the gas pressure in the measurement cell. The different relaxation times of the measurement gas and the reference gas will then cause different ratios between the signals from the two flash frequencies.
This preferred method will be reasonable with regard to costs, and at the same time it will provide a possibility for letting a sensor calibrate itself, i.e. make an automatic resetting when information is submitted, e.g. by using a push button, to the effect that the measurement gas concentration is zero in the measurement room. Thereby, all drift can be compensated for completely. The sensor will also be able to tell when it has become so dirty that it cannot be used any more (self test). This means that when the reference gas signal becomes very weak due to e.g. fouling, a special signal is produced which means that cleaning must be undertaken.