Patent Application
20090132206
The present invention generally relates to fluid analytical systems and methods for measuring impurities in a gas background. It more particularly concerns a system and a method of eliminating interference, stabilizing and linearizing on line plasma or discharge type emission based detector for impurities measurement in noble gases.
In the air separation industry or in the field of the electronic industries, many different types of high purity gases must be analyzed for quality or process control. To achieve this, there are several types of analytical instruments available on the market using different types of detectors.
Some of these detectors are specific to an impurity to be measured and their levels of susceptibility to other impurities present in the sample can be acceptable since they do not affect dramatically the accuracy of the measurement. An example of such a detector known in the art is the O2 fuel cell detector used to measure O2 impurities in gas samples. The presence of H2, N2, H2O or CH4 in about the same level of the O2 to be measured does not interfere with the O2 identification and quantification.
The measurement of Hydrocarbons by a standard configuration flame ionization detector (FID) is another example of this. The FID will not be really affected by other traces of impurities in presence in the sample like O2, N2, CO, CO2 or moisture. Thus, in such detector configurations no special consideration has to be taken to protect them against such undesirable interference.
Typical online analytical systems rely on detector that uses the chemical or physical properties of the impurity to be measured to generate an electrical signal. However, when the impurity to be measured is inert at ambient condition, such detector could not be used. A good example of this is the measurement of traces of N2 in Argon or Helium. The inertness of N2 makes it difficult to be measured at low levels. Researchers have chosen the spectroscopic emission method in the early age of the industrial analytical equipments to measure it. This method is generally known as plasma emission detection and is used with sample background gases that are easy to ionize with the help of electrical or electro magnetic field generator. An early spectroscopic emission system for the N2 in Argon measurement is described in “Spectroscopic analysis of gas mixtures”, O. P. Bochkova and E. Y A. Shreyder, Leningrad State University Academic Press, 1965 SF-4101. This concept, introduced more than 50 years ago, is still in use today.
The basic concept of this system is to get a discharge through the gas to create plasma, i.e. a gaseous ionized zone. The molecules of the background gas get excited and excite the gas molecules of N2 impurities which in turn emit various spectral emission wavelengths characteristic of N2. A common application of this concept is the measurement of N2 in Argon or Helium sample. Such technology used in this system is known as emission spectroscopy. The molecular species in the vapour phase is excited to emit light, then, the spectrum and the intensity of the emitted light are analyzed to determine the concentration of impurity in the sample.
The excitation of an atom or a molecule to a level that can emit photons requires energy greater than or equal to the excitation energy of a given level. This energy can be supplied by conversion of the kinetic energy of electrons (electrons temperature), ions or atoms colliding, absorption of light quanta and collisions of second kind. However, the later refers to radiation-less transfer of excitation energy from other particles.
In the art, almost all possible ways to create plasma and to excite the sample background (Ar or He) have been experimented for the determination of N2. The excitation ranges from direct DC or AC current discharge through the gas, the simplest form of excitation, to micro-wave induced discharge or plasma, passing by the low frequency silent electric discharge and VHF-RF frequency range. The operating pressures are also variable, ranging from low pressure, i.e. rough vacuum operating emission cell, to the atmospheric pressure. Depending on excitation mode, some discharges are of a glow type, like in low pressure, and others are of a silent electric discharge type, also called dielectric barrier discharge (DBD) or streamer discharge. This type of discharge occurs at atmospheric pressure. The power coupling technique used in the above spectroscopic emission systems, also differs. The direct coupling method is referred to a configuration where the metal electrodes are in contact with the gas to be ionized.
Such systems are described in U.S. Pat. Nos. 3,032,654; 5,412,467; 4,148,612 and 5,168,323, and in the following papers: “Spectroscopic analysis of gas mixtures”, O. P. Bochkova and E. Y A. Shreyder, Leningrad State University Academic Press, 1965 SF-4101: “Excitation de l'azote, oxygène et hydrogène en impuretés dans des décharges de gaz rares He, Ne et argon”, A. Ricard, J. Lefebvre, Revue de physique appliquée, 10 mai 1975 tome 10; “Detector for Trace Amount of Nitrogen in Helium”, R. J. Walker, Cryogenics, 1986, vol. 26 May; “A DC Microplasma on a chip employed as an optical emission detector for gas chromatography”, Jan C. T. Eijkel et Al., Analytical Chemistry vol. 72, No. 11 Jun. 1, 2000; “An Optical Emission Study on Expanding Low Temperature Cascade Arc Plasmas”, Q. S. Yu and H. K. Yasuda, Plasma chemistry and plasma processing Vol. 18, No. 4, 1998, and “Emission Spectrometric Method and Analyzer for Trace of Nitrogen in Argon”, Homer Fay, Paul H. Mohr, and Serard A Cook, Linde Co. Division of Union Carbide, Analytical Chemistry, Vol. 34, No. 10 Sep. 1962.
In all of these abovementioned references, it can be seen that for the N2 measurement, the emission wavelength typically used for low level measurement is 337.1 nm. However, 391 nm is sometimes used when sample background is Helium because of its strong intensity compared to other N2 emission lines or bands. This leads to the fact that in all the commercially available systems known today using emission spectroscopy to measure N2 in Ar, spectral emission at 337.1 nm is used to identify and quantify N2. When background is Helium, emission at 391 nm is preferred.
The wavelength of interest is filtered out by means of interference filters, grating or monochromator and its intensity is transformed into an electrical signal by any suitable photo electrical device. The electrical signal is then processed to provide the final N2 level.
Most of these systems use a single emission cell with a single optical measurement channel. It is a general belief that measuring N2 emission of an excited Argon or Helium plasma will give a linear signal from sub ppm to a few hundred ppm and no special means is generally used to correct any nonlinearity.
However, even if the above-described method is generally accepted by people involved in the art for N2 measurement, this method could become unreliable under certain conditions that are found in almost all industrial field applications. Indeed, the ideal situation would be to have a dry binary mixture of N2 in Ar or He. In such case, the measurement will not be affected by any other impurities. But even in such conditions, non-linearity occurs, as it will be described and demonstrated thereinafter, thereby resulting in substantial measurement errors.
As another example, in air separation industries, the N2 level in Argon products should be measured. Based on a particular Ar production process, many other impurities could also be present in the Ar and sometimes at a higher level than N2. Typical impurities generally found are: H2, O2, N2, CH4, CO and H2O. Furthermore, at the truck loading station, when the Ar is transferred into the tanker to be shipped to the customer, ambient air and moisture contamination is a to frequent problem. There is also the fact that, at the present time, customers of such produced high purity Argon are demanding more rigorous quality control. As consequence, customers are also asking for the measurement of other impurities like CO2 and NMHC (Non Methane Hydrocarbons), leading to believe that these impurities could also be present as an impurity in the high purity Argon.
In these situations, the presence of other impurities will have an effect on the system's performance. N2 and Ar are inert gases and normally do not react with other molecules at ambient conditions. However, inside a plasma emission system, gases under ionization state (or plasma) are a very reactive and aggressive medium. Many call it the forth state of matter, as described in “Gaze lonizate (ionized Gases)”, M. M. Badareu and Popescu, Dunod Edition. It has been known for a long time that in an ionized gas or plasma, many chemical reactions can occur even with inert gases. Plasma contains molecule radicals and atoms but also ions and free electrons which result from the coupling of energy with matters in the gaseous state. This application field relates to plasma chemistry and plasma processing.
In this field, many research works have been conducted with plasma systems similar to the ones used in analytical systems today used. Most of them use DBD or Dielectric Barrier Discharge (also called silent dielectric discharge) because of its simplicity. Thus, it is relatively inexpensive to manufacture compared to systems relying on other modes of excitation. RF excited discharge or pulse and direct current generating discharge or plasma could also be used. In this field of application, the “reactor” property of plasma is used to trigger various chemical reactions. People involved in this art have documented and reported many species, free radical and their reaction rates. They use emission spectroscopy to diagnostic or investigate plasma reactor operations.
Since most of plasma reactors used is very similar, not to say the same, as the emission cell used in the commercially available analytical instruments, they should trigger different types of chemical reactions if there is different types of impurities flowing in it. Despite this fundamental evidence it seems that, in the analytical instruments used today, no special considerations are taken to overcome interference problems.
As mentioned above, even if the previously described systems perform relatively well in the case of a dry binary mixture, except for the system linearity, it can be seen from the above that, in the real word, many other impurities could be present in the gas to be analysed, thereby leading to inaccurate measurement of the impurities to be measured. Moreover, as also mentioned above, the methods presently used today could become unreliable under particular conditions surrounding the emission system and additional impurities that could be present in the gas under analysis.
It would therefore be desirable to provide an improved method and an improved system for measuring impurities such as N2 in rare and noble gases that would be very accurate and reliable, whatever the surrounding conditions and the additional impurities that could be present in the gas under analyse. Moreover, it would be even more desirable to provide such a method that would provide accurate measurements even at sub-ppb and up to 10,000 ppm levels while providing a long-term stability.
An object of the present invention is to provide a method and a system of eliminating interference for impurities measurement in rare and noble gases that satisfies the above-mentioned needs.
Accordingly, the method of the present invention mainly relies on the use of a combination of particularly designed means serially connected for cancelling interferences. Proper means for correcting linearity issues are also advantageously implemented.
More particularly, the invention provides a method of eliminating interference for impurities measurement in noble gases, comprising steps of:
In a preferred embodiment, the gas sample comprises argon and the impurity to be measured comprises nitrogen. In a further preferred embodiment, in the step f), the specific emission wavelength is 357.69 nm.
In another preferred embodiment, the method further comprises, before step e), a step of adding oxygen to the gas sample at the outlet of the impurities trap.
In another preferred embodiment, each of the trapping means are adapted to trap at least one of H2O, O2, CH4, CO, CO2, H2, organic compounds and non-methane hydrocarbon impurities.
In another preferred embodiment, the method further comprises a step of isolating the plasma cell from surroundings.
In another preferred embodiment, the method further comprises, before step g), steps of:
Preferably, the method further comprises steps of:
Advantageously, the method further comprises a step of compensating a non-linearity of the emission signal. This step can be achieved by adjusting a driving power of the plasma cell according to the concentration of the impurity to be measured, or by applying a correcting algorithm dependent on the concentration of the impurity to be measured.
In another preferred embodiment, the method further comprises steps of:
The method of the present invention, when implemented with most of the above-mentioned preferred features, is particularly advantageous since it offers long-term stability while providing very accurate and reliable results, even at subs ppb and up to 10,000 ppm levels, whatever the surrounding conditions and the additional impurities that could be present in the gas under analyse.
Another object of the present invention is to provide a system of eliminating interference for impurities measurement in a gas sample having interfering impurities and an impurity to be measured therein. The system comprises:
a plurality of trapping means serially connected to define an impurities trap having an inlet and an outlet, each of the trapping means being adapted to trap a specific one of the interfering impurities without affecting the impurity to be measured; means for adding moisture to the gas sample, the means being serially connected to the outlet of the impurities trap;
a plasma cell of a plasma emission system serially connected to the means for adding moisture for generating an emission light representative of a concentration of the impurity to be measured of the gas sample; and
processing means operatively connected to the emission cell for processing the emission light and providing the concentration of the impurity to be measured.
In a preferred embodiment, the plurality of trapping means comprises a molecular sieve for trapping H2O impurities, a reduced copper based catalyst for trapping oxygen impurities, a palladium based catalyst for trapping hydrogen impurities, a nickel based catalyst for trapping CO, CO2 and non-methane hydrocarbon impurities, and an activated carbon based catalyst for trapping organic compounds.
In another preferred embodiment, the system is further provided with means for adding oxygen to the gas sample, the means being serially connected to the outlet of the impurities trap.
In another further preferred embodiment, the system is further provided with an insulating enclosure mounted around the plasma cell for insulating the plasma cell from surroundings.
In another further preferred embodiment, the processing means are provided with collecting means for collecting the emission light at a specific emission wavelength not subjected to spectral interference to provide an emission signal representative of a concentration of the impurity to be measured. The processing means are further provided with a microprocessor for processing the emission signal and providing the concentration of the impurity to be measured.
Preferably, the processing means is further provided with background collecting means for collecting a background emission light generated by the plasma cell at an emission wavelength neighbouring the specific emission wavelength to provide a background emission signal. The microprocessor is adapted for subtracting the background emission signal from the emission signal to provide a net emission signal representative of the net emission of the impurity to be measured, thereby cancelling background variations.
Still preferably, the processing means are further provided with a correcting algorithm for compensating a non-linearity of the emission signal according to the concentration of the impurity to be measured.
In a further preferred embodiment, the system is advantageously provided with a reference plasma cell connected in parallel with the emission cell for generating a reference emission light representative of a concentration of the impurity to be measured. The processing means are adapted to process the reference emission light for providing a level of the concentration of the impurity to be measured. The system further has adjusting means for adjusting a driving power of the plasma cell according to the level of the concentration of the impurity to be measured.
These and other objects and advantages of the invention will become apparent upon reading the detailed description and upon referring to the drawings in which:
While the invention will be described in conjunction with example embodiments, it will be understood that it is not intended to limit the scope of the invention to such embodiments. On the contrary, it is intended to cover all alternatives, modifications and equivalents as may be included as defined by the present description.
As previously mentioned, having a look at different emission spectrums generated by plasma source using different excitation modes reveals that emission line wavelengths for N2 impurities are the same.
Below, are reported the results of our investigations that were conducted to enlighten the problems encountered when attempting to measure N2 in presence of some other impurities, with emission spectroscopy. Then, a new method and a new system overcoming these problems will be described.
These investigations were conducted with a prior art plasma emission system used to measure N2 in Argon. But it is worth mentioning that the following discussion could also be applied to N2 in He, Ne or Krypton. The emission system tested here was working at atmospheric pressure and excitation frequency ranging from 60 Hz to 4 MHZ. Excitation field could also be modulated. Emission spectrum looks almost the same for all excitation frequencies except for the intensity and noise. Lower excitation frequency dielectric barrier discharge has low emission intensity and is noisier due to spatial movement of streamers. Higher frequency looks more like a glow discharge type than streamer type emission, intensities being higher and noise being lower. Nevertheless, in almost all cases emission wavelength for a particular impurity is the same.
Referring to
The above-described Figures are not subject to any interference and represent the ideal conditions. However, hereinbelow is described what could happen in real applications where there is not only N2 impurities in presence into the emission cell system.
Interference Problems:
a) Moisture interference: As previously mentioned, in air separation plants, one of the common impurities is moisture. Moisture could be present at various levels, depending on surrounding conditions such as where the sample is coming from. Typically, at truck loading stations, the moisture level could be higher due to atmospheric contamination during sample manipulation.
In the attempt of finding the influence of moisture on Nitrogen determination, some moisture was added to the previous sample of binary mixture of 5 ppm N2 in Argon.
Referring now to
Reference is now made to
It should be noted that the interference caused by moisture is more of a physical/chemical type than of a spectral type in the sense that there is no emission line caused by moisture close to the N2 emission line. However, since that, in a plasma system, there is a competition between various species in presence for the energy available, which is mostly supplied by the Argon in the metastable state, it results that the impurities having the lowest ionization potential will be favoured because they react faster with the Argon metastable, thereby reducing the excitation efficiency for the N2. Moisture has a lower ionization potential. This is shown by the emission spectrum. Indeed, an increase of moisture results in a higher intensity of OH emission while N2 emissions are going down.
As previously described, in “Development of a Detector for Ultratrace Nitrogen in Argon using Low-Pressure Capillary Glow Discharge Molecular Emission Spectrophotometry”, Hiroshi Ogino et Al., Anal-Chem. Vol 69, No 17, September 1997, Ogino et al. have also noticed the same type of interference caused by moisture, i.e. decreased sensitivity and higher background emission in presence of moisture. They also report that moisture does not cause any spectral interference.
Therefore, it can be concluded without any doubt that moisture in a sample of N2 in Argon will result in physical/chemical interference that will lead to measurement errors, regardless of the excitation mode which is used. The presence of one impurity affects the other and vice versa.
b) O2 interference: The effect of O2 impurity on N2 measurement in an Argon sample has also been investigated. When some O2 is added to a mixture of 5 ppm N2 in Argon, emission of NO between 200 and 300 nm appears, as illustrated in
c) H2 interference: Referring now to
d) CH4 interference: Referring now to
CH4 is generally present in cryogenically produced Argon. It normally comes from the air incoming to the plant and it is not stopped by the process.
Some Argon plants use a warm Argon cycle to remove the Oxygen from the Argon. In this process there is generally a Palladium catalyst operating at high temperature, where H2 is added to the crude Argon process stream. CH4 could be generated in this catalyst bed due to the presence of other impurities in the process stream. The warm Argon cycle process generates high temperature steam. In the case of process upset that leads to unstable operation of the Palladium catalyst bed, impurities like moisture, Oxygen, Hydrogen and Methane could end up in the pure Argon column and then in the final product. In both cases, the N2 level is measured by on line emission spectroscopy.
In the electronic field industries like the so-called “wafer fab”, the use of a getter alloy base purifier is mandatory to get impurity levels below the single digit ppb. The first impurity to break through is the N2. So, the monitoring of N2 level is a good indication of the end of life of such purifier. However, when such purifier is close to its end of life, it is not uncommon that H2 be released from the getter alloy. The monitoring range for N2 level is generally below 10 ppb. The H2 released will make the N2 analyzer read higher, leading to replace the expensive alloy before its time.
e) Multi-component case interference: In “Analyse qualitative d'impuretés dans l'hélium par spectroscopic d'émission”, A. Ricard and J. Lefebvre, Analysis, 1978V.6, No. 7, pp. 299 to 305, the authors have identified the emission lines for H2, O2, N2, CH4 in a Helium plasma at atmospheric pressure. They also analyze various impurity effects on the other ones. They came to the conclusion that it is not possible to do an accurate measure of N2 when there are other impurities in the sample. This is caused by the fact that these impurities do not have the same efficiency to de-excite the Helium metastable. They suggest the use of a chromatographic column to separate each of the impurities or to use a binary mixture. As previously explained, the later case is not possible in the industrial field.
The same problem complexity was also well explained in U.S. Pat. No. 4,801,209 entitled “Process and apparatus for analyzing a gaseous mixture and a visible emission spectrum generator therefor” and granted to Waldow. In this patent, Waldow clearly indicated that the emission spectrum emanating from a discharge within a mixture of gases is the sum of all photons being emitted with the result that the combined spectrum is not a linear sum of the individual spectrums.
The relative intensity of the spectral lines associated with the individual gases is not preserved due to what is known as the matrix effect. Deconvolution or separation of the individual spectrums from a combined spectrum is thus more of an art than a science. Complexity increases rapidly in multi-gas mixtures. The multi-component complexity problem is also discussed in “Spectroscopic analysis of gas mixtures”, O. P. Bochkova and E. Y A. Shreyder, Leningrad State University Academic Press, 1965 SF-4101, at pages 149 and 199. They limited their analysis to a ternary mixture and found a very limited range of application. It should also be mentioned that their analysis was performed with high levels of impurities.
f) Specific interference of Argon in Helium: In “Analyse qualitative d'impuretés dans l'hélium par spectroscopie d'émission”, A. Ricard and J. Lefebvre, Analysis, 1978 V.6, No. 7, pp. 299 to 305, the authors use the N2 emission line at 391 nm. However, further to our advanced experiments, it was demonstrated that there is another source of strong interference when using this wavelength in Helium background. Indeed, only a minute diffusion of Argon into the Helium sample quenches almost completely the N2 emission at 391 nm. The Argon comes back from ambient air diffusion. So, using this emission line for N2 measurement in Helium in the industrial applications will be very risky, even if it is the most sensitive one.
g) Other interference sources affecting N2 Quantification: Complete investigations on the combined effect of various impurities in a sample, i.e. when all the above-mentioned impurities are present in various levels in the sample, on the N2 emission were not performed. However, the effect of the atmospheric pressure on the N2 emission of a 100 ppb N2 in Argon mixture was tested. The test conducted has shown that a variation of 30 inches of water shifts the reading anywhere between 20 to 30 ppb. A variation of 30 inches of water for the atmospheric pressure is about the variation between a sunny day and a day where there is a thunder storm. Since some applications requires sub ppm and ppb measurement, these parameters become important for long term stability.
Moreover, the plasma pressure directly affects electron temperature. Our investigations have also shown another source of interference that was not reported in the prior art. Indeed, in various plasma emission cells and generator combinations for analytical applications, the plasma or discharge power is relatively low, generally less than 10 watts. Some generators are coupled to the load (i.e. the discharge cell) inductively, i.e. type H discharge, the other capacitively, i.e. type E discharge. See, for example, “Practical Spectroscopy of High Frequency Discharges”, Sergie A. Kazantsev, Vyacheslav I. Khutarshchikov, Gunter H. Guthohrlein, Laurentius Windhoiz, Plenum Press. 1998, ISBN 0-306-45676-1, pp. 7. In such arrangement the excitation signal is AC.
When the power coupling technique is capacitive, the system is generally called a dielectric barrier discharge since there is at least one dielectric barrier between coupling electrodes. In the type H or E system, the field potential is relatively high. Our experiments have shown that, in some cases, the discharge occurs not only in the plasma cell, but some corona discharge can also co-exist at the electrode surrounded by ambient air. The electrodes are normally bounded on the cell wall. The corona discharge occurring on the electrode edge, the emitted light also reaches the photo electric measuring device. This generates spurious variation in the measuring signal. It becomes worse when the relative humidity of ambient air is high. In a system designed to measure low levels of impurities and requiring long term stability, this phenomenon is not acceptable since the generated interference may lead to wrong alarms. The phenomenon was more evident with the capacitive coupling, i.e. type E discharge, since the field potential is higher. With DC or AC direct discharge, the potential is much lower since there is no dielectric. There is DC or AC current flowing between the electrodes. Sometimes the potential could only be about 300 VDC, depending on chamber and electrode design. In this case, there is not enough potential gradients to get corona effect and ionization of the surrounding atmospheric air.
Our investigations have also shown that, for some emission based N2 analyzers used in the field, the inner wall of the cell was coated with a brown color deposit. Such deposit mainly occurred in equipment installed in Air Separation plant truck loading stations. Such deposit was removed from the emission cell, and analyzes revealed that this deposit was mainly composed of Carbon that came from the cracking of some organic compound or Hydrocarbon other than CH4, released by the existing sampling system material or the polluted tanker. Here, by sampling system we mean tubing valves, fittings, filters, etc. There still are many existing sites today that use rotary valves for sample stream selection. Many of these valves have different polymer material and various sealing material and lubricant. At some sites, the use of plastic or polymer flexible lines are common. Nylon for example is a source of Hydrocarbon pollution. Valves having some grease or carbon powder to reduce the friction of the seal around the valve stem are also another source of Hydrocarbons. Some organic compounds simply come from the finger prints left on some parts of the sampling system during its assembly. Such interference results in a lost of sensitivity, since the cell slowly becomes opaque to light transmission.
Another source of carbon contamination is the material used to build the analyzer. Stainless steel tubing, depending on operating conditions, could also release some contamination. We found that the plasma chamber used in our laboratory has such deposit after 5 years of use. The source of deposit comes from the material itself or the so-called “wetted parts” of the system. We also found that the material used to make the emission chamber may have some amount of impurities in it. Normally, a very high quality quartz material is used. However, after many years of use and under the plasma, some parts of the wall's chamber do have some deposit, even with plasma maintained with a very high purity gas. This source of contamination is variable, based on material lot, and happens over a relatively long period of time. Generally, we can see this side effect beginning to appear after two to three years.
However, process analyzer should ideally be reliable for a much longer period of time. So, it would be a real benefit to be able to keep the Inner wall of the emission chamber “clean” or deposit free, in order to maintain long term sensitivity thereof.
h) non linearity phenomenon: As previously mentioned all commercially available N2 analyzers on the market do not have any linearity correction means. There is a reason for this. In fact, over a small range of measurement, the signal is quite linear, like shown in
Since the present invention, as it will be explained later, uses a method that traps impurities except the N2, and also uses a differential signal, i.e. background signal subtracted from N2 emission, one may think that these features will fix the non-linearity problem. This is true, but only for a limited operating range of N2 in the sample. Furthermore, based on noble gas type, the non-linearity behaviour changes a lot.
As described in “Spectroscopic analysis of gas mixtures”, O. P. Bochkova and E. Y A. Shreyder, Leningrad State University Academic Press, 1965 SF-4101, and also in several physic's books related to plasma physic, the amount of concentration of excited atoms or molecules depends on the current or plasma power. It also teaches that the species in presence in a plasma are in competition for the available energy. So, in a binary mixture the ionization efficiency is not the same with a small amount of impurities compared to a higher amount.
It would therefore be highly desirable to provide some means to correct linearity issues in order to get a linear reading over a wide range of N2 level, i.e. up to 10,000 ppm.
Prior Art Solutions:
Today, many users have recognized the detrimental effect from other impurities in an Argon sample on the N2 measurement. They also recognized the need to take some action to correct such problems. However, as it will be explained thereinafter, each of the various attempts proposed in the art to cancel interference is always a partial solution.
a) Reference is now made to the following prior art documents: U.S. Pat. Nos. 3,032,554; 5,168,323 and 5,009,099; “Spectroscopic analysis of gas mixtures”, O. P. Bochkova and E. Y A. Shreyder, Leningrad State University Academic Press, 1965 SF-4101; “Emission Spectrometric Method and Analyzer for Trace of Nitrogen in Argon”. Homer Fay, Paul H. Mohr, and Serard A. Cook, Linde Co. Division of Union Carbide, Analytical Chemistry, Vol. 34, No. 10 Sep. 1962; “Development of an Automated Speciation Analyzer”, Bernd Rosenkranz et Al., American Laboratory vol. 31, number 20, October 1999; “A Modular for the Automatic Subtraction of Continuous Background in Optical Spectrometers”, Optics and Spectroscopy, February 1964, Vol. 16, No. 2, pp. 184-186, G. M. Svishchev; “A Method of Background Correction for Direct Reading Optical Emission Spectroscopic Analysis Using Offset Exit Slits”, Analytical Chemistry vol. 41 No. 2, February 1969, pp. 396-398, John A. Levs, and “A Dynamic Background Correction System for Direct Reading Spectrometry” R. K. Skogerboe et Al., Applied Spectroscopy, vol. 30, No. 5, pp. 495-500, September-October 1976. In these documents, attempts to cancel out the background signal variation effect by subtracting background signal from the impurity emission lines are described. Such correction results in a more stable signal, i.e. less drift. In all these cases, even if the used designs slightly differ from each others, the tests conducted have shown more stable and accurate results by doing background corrections.
However, background correction cannot cancel alone all the detrimental effect: of interference. For example, in the case of spectral interference, background compensation or correction is useless, Furthermore, some physical/chemical effects on a particular emission line of interest, as 337 nm for N2 for example, are not cancelled by background correction. An example of this is the effect of moisture on the N2 emission lines as previously described.
b) Other users have felt the need to reduce the moisture interference on N2 measurement. For example, in “Development of a Detector for Ultratrace Nitrogen in Argon using Low-Pressure Capillary Glow Discharge Molecular Emission Spectrophotometry”, Hiroshi Ogino et Al., Anal-Chem. Vol 69, No 17, September 1997, Ogino et Al. use a desiccant to trap moisture from the Argon sample. They use a molecular sieve 5A (60/80 mesh). However, with this embodiment, the N2 molecule will be delayed by passing through this trap since 5A molecular sieve base trap acts like a chromatographic column. Any temperature variation of the trap will cause reading variations. Increasing the trap temperature will release some N2 and vice versa. This will thus cause signal wandering.
Moreover, the use of a desiccant trap will not effectively resolve the interference caused by other impurities like CH4, O2 and H2. In fact, the trap may accentuate the problem. The 5A molecular sieve will also partially trap and delay the passage of H2, O2 and CH4. Flow or temperature variation of the trap will also affect the adsorption behaviour thereof. It results that, when there is flow or temperature variation, O2, H2 and CH4 could be released or adsorbed into the trap. It results in more drift or sudden change on N2 reading.
c) Other users suggest, like in U.S. Pat. No. 5,168,323 previously cited, the use of a computational method to compensate for interference. In the described method, they also take into consideration the signal background. The intensity of various spectral lines of interest is measured. The effective line intensity generated from one particular impurity is computationally determined from linear combinations of the various measured intensities. However, this method has serious drawbacks. First, as it was shown with moisture, the interference effects are not linear. Moreover, it appears that, based on plasma excitation mode and the level of impurities in presence, the various physical/chemical reactions and their rate are not constant. Furthermore, the suggested method does not take into account some parameters that also affect the plasma like the operating pressure and quenching. Quenching occurs when a molecule having been excited to a higher level, is de-excited to a lower level by collision with other molecules before it has spontaneously emitted photons at its characteristic wavelength.
Furthermore, it should also be noted that the background signal does not vary equally in the emission spectrum. Measuring the background emission at only one point for background correction will not work for all impurities because some of them are not subject to the same signal background variation. At the best of our knowledge, the system proposed in the above-mentioned prior art reference was not really commercialized. Another problem of this proposed system resides in the fact that the possibility that the Argon sample can also have CO and CO2 impurities therein is not taken into consideration. It appears that, in theory, by taking into account all the parameters that can affect the N2 emission into the plasma, one can compensate the effect of these parameters on N2 readings. However, in practice, it would require much more than a simple linear combination and it would become really complex and costly to build such system.
d) In U.S. Pat. No. 4,801,209 previously mentioned, Waldow described a method to compensate the interference on N2 measurement caused by the variation of Ar/O2 ratio. The author recognizes the fact that the interference is non-linear and proposes an array of non-linear equations (second order form) with several coefficients that must be determined experimentally by generating various mixtures of N2/Ar/O2 ratio. It is a long and laborious work that must be repeated for each system built. Another drawback of this compensation or correction method is that only three gases are taken into consideration, i.e. N2, Ar and O2. Furthermore, the proposed method runs correctly for a fairly high amount of N2 in Argon, i.e. from 20 to 2000 ppm. However, for sub ppm measurements, which is the impurity's level that the method of the present invention aims to measure, this prior art method would not be able to provide satisfactory results. Moreover, the method suggested in this patent is for a limited and specific range of sample background, i.e. specific range of Ar and O2 in the background. No other impurities are taken into consideration. The author also reports the effect of atmospheric pressure variation on the system response. However, no compensation means is proposed.
e) In U.S. Pat. No. 5,570,179, Weckstrom described a method wherein there can be a computational base system to correct the effect of CO2 and O2 and some other gases on N2 measurement. Separate sensors like infrared for CO2 measurement, paramagnetic for O2 measurement and other gas sensors are connected to a microprocessor system. In this method, no equation or relationship is reported. The manufacturing of such system would be complex and expensive for a limited performance.
f) In U.S. Pat. No. 4,898,465, Crawford et Al. use a cold cathode gas discharge excited with DC current as emission cell. They report interference on N2 caused by O2 and CO2. They also report signal variation caused by system pressure. They suggest the use of a control loop to adjust the power or the current of the discharge based on species to be analyzed. However, this method would become impractical with multiple impurities in the sample. There is no way to know which species are responsible for signal variation.
After all the efforts done by various workers in the attempt of resolving interference problems on N2 measurement, there still are unresolved interference problems. As previously mentioned, the use of a trap for moisture doesn't resolve the problem caused by H2, O2 and CH4. Moreover, the use of computational model becomes extremely complex and laborious so only partial solutions were described.
Paradoxically, it should be mentioned that, in U.S. Pat. Nos. 5,412,467; 5,831,728 and 6,043,881, no interference at all were reported. The same type of discharge that was used by many other workers listed above was however used.
In U.S. Pat. Nos. 5,412,467 and 5,831,728, the authors state that measurement of N2, H2O and CH4 could be done in the same spectroscopic cell and that the impurities do not interfere with each other. However, the tests we conducted and all other prior art references we described thereinabove provide different results. As far as we know, such multi-impurities systems with parallel measurement of these impurities were not really commercialized. Moreover, in these references. no data is really available. This leads to believe that few parameters in their experiment have been neglected. In U.S. Pat. No. 6,043,881, the use of emission cells serially connected is suggested. This configuration is suggested for multi impurity measurements including N2. The measurement of H2O, CH4 and N2 is also suggested, even at sub ppm levels. However, due to the nature of plasma discharge, the impurity level in the second cell will not be the same as in the first one. This is due to various chemical reactions occurring with various impurities. So, the first emission cell modifies the nature of the sample flowing into the second cell. Our experiments have shown that the levels of N2 are not the same in both cells. This was easily tested by turning OFF and ON the plasma in the first cell. N2 reading was going down while the plasma was ON in the first cell.
Actually, in all commercially available instruments that measure N2 in Ar and He, they use dielectric barrier discharge at atmospheric pressure. It makes sense to use these systems since the emission spectrum is almost the same as more complex systems based on microwave, radio frequency or running with vacuum pump. However, at the best of our knowledge, none of them use means to compensate or cancel out the interference caused by other impurities in the sample, or simply caused by the ambient operating condition. Nevertheless, these systems are performing relatively well in the case of a binary mixture, but the signal is not linear.
Thus, as already mentioned above, it would be a real benefit to provide an improved method and an improved system to do the measurement of N2 in rare gases using emission spectroscopy that will have a high sensitivity without the interference and linearity problems cited above.
Present Invention:
Accordingly, the present invention proposes a method for on line measurement of N2 in noble and rare gases based on emission spectroscopy which provides very stable, sensitive and interference free results.
In brief, the method must eliminate the following interferences on N2 measurement:
Based on our investigations and the available prior art, it appears that it is very difficult and complex to cancel out the influence of typical impurities found in noble gases on the N2 measurement.
The concept of the method of the present invention is based on a combination of means instead of relying only on computational methods, baseline correction or trapping impurities alone.
Therefore, the method according to the invention, also called method of eliminating interference for impurities measurement in noble gases, comprises the steps of:
a) providing a gas sample having interfering impurities and an impurity to be measured therein;
b) providing a plurality of trapping means serially connected to define an impurities trap having an inlet and an outlet, each of the trapping means being adapted to trap a specific one of the interfering impurities without affecting the impurity to be measured;
c) introducing the gas sample at the inlet of the impurities trap for removing the interfering impurities;
d) adding moisture to the gas sample at the outlet of the impurities trap;
e) introducing the gas sample in a plasma cell of a plasma emission system;
f) collecting an emission light generated by the plasma cell at a specific emission wavelength not subjected to spectral interference for providing an emission signal representative of a concentration of the impurity to be measured of the gas sample; and
g) measuring the concentration of the impurity to be measured according to the emission signal provided in step f).
Our observations reveal that there is interference coming from: H2, O2, CH4, CO, CO2, NMHC, organic compound and H2O. In the worst case, each of these impurities may be present in the sample at the same time than N2 to be measured. Our investigations have shown that most of these impurities could be trapped with a specially designed multi-layer trap. This trap has to be specially designed to avoid any interference with the N2 impurities. Thus, according to the concept of the method of the present invention, each of the impurities will be considered one by one. It should be mentioned that the material used in various trap layers must not affect the N2 impurity to be measured.
Generally, the method proposed in the present invention partially relies on trapping the unwanted impurities. According to this concept, a multi-stage or layer (also called bed) of specific trapping material will be described hereinafter. The hereinbelow described system has been tested and has provided good results. However, ft is important to note that other trapping technologies could be envisaged as long as they do not affect the N2 impurities. So, it is to be understood that the following description is not limitative and is given as an exemplary embodiment only.
The first impurity to be removed is H2O. For this purpose, a molecular sieve with a pore diameter less than 4 Å angstroms is advantageously selected. In fact, our investigations have shown that the best results were obtained with 3A molecular sieves. The 3A sieves do not retain the N2 molecule, it just goes through. However, like previously mentioned, the effects of H2O on N2 intensity and background emission are more apparent when there is a slight change from dry sample to few ppm (about 2) of moisture. Even if our 3A based trap is efficient to remove H2O without affecting N2 in the sample, a slight change in its operating condition leads to a change in H2O adsorption. It results that there could be a small change of H2O at the outlet due to a change in operating condition, which would lead to some interference on N2. Paradoxically, even if this change in H2O level is small, it happens in the range wherein the H2O effect has a more apparent impact on N2 and background signal. Thus, to resolve this issue, some water is advantageously added to the dehydrated Argon sample. This way, the plasma operating point is biased in terms of water level. The quantity of H2O added is high enough to set the operating point where a small change of H2O level at the outlet of the H2O trap, does not affect at all the N2 reading. The added level of H2O is called the saturation point. Here, when this saturation point is reached, there is no more effect on N2 emission when there is a slight change in H2O level. The interference caused by H2O in the sample was thus completely eliminated, without affecting the N2 impurities.
The second impurity to consider is Oxygen. Oxygen is easily and advantageously trapped by a reduced Copper based catalyst. Here again, the trapping material must not affect the N2 impurities. Most Copper based catalysts are based on Copper finely dispersed on a support. The support is designed for an increase surface area to maximise the contact with the flowing gases. The idea is to select a Copper based catalyst having a pore diameter that will have little effect on N2 while trapping the O2 impurities. Furthermore, it should be mentioned that it can be advantageous to heat the Copper catalyst at some point, since it increases its capacity to trap O2 and speeds up the passage of N2 there through. This point is relatively important when the sample is Helium since Helium has a much lower molecular weight than N2. Thus, increasing trap temperature has a real benefit on N2 impurities in the sense that N2 has much less tendency to be delayed by the supporting material of the catalyst. This has efficiently resolved the problems caused by the O2 impurities on N2 measurement. A suitable catalyst which can advantageously be used is the R3-11, R3-12 or R3-15 from BASF Company. However, it is worth mentioning that other suitable catalysts could also be envisaged.
For Hydrogen Impurities, a catalyst having high affinity for H2 is selected. In this case, a Palladium based catalyst is advantageously chosen. Here again, the catalyst support material should leave the N2 impurities unaffected. This is done by selecting the proper pore size and operating temperature like done for the O2 impurities. Palladium has a very high adsorption capacity for H2. It could adsorb up to 900 times its own H2 volume. With this specific catalyst the H2 impurities are efficiently removes from the sample while the N2 are leaved unaffected.
Now, concerning CO, CO2 and NMHC impurities, such impurities are advantageously effectively removed from the sample by a Nickel based catalyst, like reported in U.S. Pat. No. 4,713,224. Here, special care has to be taken to avoid affecting the N2 passage through this catalyst. Our experiments have shown 5 that, by adjusting the mesh size of the catalyst, impurities were still effectively removed while N2 impurities remained unaffected. Again, operating temperature should preferably be constant. It should be noted that a Nickel catalyst is also able to sorb H2. A Palladium based catalyst for trapping H2 is however preferred. In fact, Nickel catalyst will effectively remove from the sample the H2O, O2, CO, CO2, H2 and NMHC impurities. It will also reduce CH4 content. However, as previously mentioned, in the present invention, the use of more specific catalysts for individual impurity removal is preferred. Thus, this way, it is easier to estimate the system performance based on catalyst manufacturer data.
With respect to the organic compounds, the method of the present invention advantageously relies on a trap based on an activated carbon. Activated Carbon has been used for many years to remove impurities from different types of liquid or gas fluids. Based on the activation method and the raw material used, the activated carbon could be made more specific to some impurities. The use of layers of proper activated carbon will advantageously efficiently remove NMHC, organic compound and will also contribute to reduce the level of CH4. The choice is governed by the effect of this layer on N2 impurities. Our experiments have shown that a deeply dehydrated coconut charcoal type activated carbon gives excellent results even if other arrangements could be considered. It should however be noted that the carbon bed temperature must preferably be kept constant.
Referring now to
The system 10 also has means 28 for adding moisture to the gas sample. These means 28 are serially connected to the outlet 26 of the impurities trap 22. The system 10 is also provided with a plasma cell 30 of a plasma emission system 32 serially connected to the means 28 for adding moisture for generating an emission light representative of a concentration of the impurity to be measured of the gas sample. The system 10 also has processing means 34 operatively connected to the emission cell 30 for processing the emission light and providing the concentration of the impurity to be measured. Preferably, the processing means 34 is provided with collecting means 36 for collecting the emission light at a specific emission wavelength not subjected to spectral interference, preferably 357.69 nm for N2 measurements, to provide an emission signal representative of a concentration of the impurity to be measured. The processing means 34 further has a microprocessor 38 for processing the emission signal and providing the concentration of the impurity to be measured.
As illustrated, once assembled, properly activated and conditioned, each of these above-described traps are then serially connected. Each of these traps is advantageously kept at constant temperature and the sample flow is also preferably kept constant, thanks to flow control means (not shown). This is relatively important since, if sample flow varies, the addition of H2O will also change. The H2O is preferably added with a permeation tube device. Since the permeation rate is constant under controlled ambient conditions, a flow variation will change the dilution factor through the tube.
When the sample gas comes out of the last trapping means 20, there is only N2 and very little CH4, if any. Then, water or moisture is added to the gas sample in order to minimize the effect of slight change of its level on the N2 emission like previously mentioned. This cancels out the detrimental effect of moisture on N2 measurement since all impurities, i.e. H2O, O2, H2, CO, CO2, NMHC and some other organic compounds are trapped. There is only the CH4 who can still cause problems because it is not trapped at 100%. CH4, as previously mentioned, results in spectral interference on N2 emission at 337.1 nm. For this reason, the 357.69 nm N2 emission line is advantageously selected. As illustrated in
However, there could be a slight change in background emission caused by CH4. To eliminate this source of interference only caused by CH4 (the other impurities are trapped and water level is constant because of the controlled permeation rate of the H2O addition module), the baseline or background emission is advantageously measured in the neighbourhood of the 357.69 nm N2 emission. This background emission is then subtracted from the N2 (357.69 nm) emission by the processing means. This way, the background variation interference on the N2 (357.69 nm) is advantageously cancelled. Background measurements at 345 nm and 365 nm have also been tested and work well. Thus, in order to implement the above-described background subtraction, the processing means 34 is advantageously further provided with background collecting means 40 for collecting a background emission light generated by the plasma cell 30 at an emission wavelength neighbouring the specific emission wavelength to provide a background emission signal. The microprocessor 38 is adapted for subtracting the background emission signal from the emission signal to provide a net emission signal representative of the net emission of the impurity to be measured, thereby cancelling background variations.
The permeation tube, which is advantageously used for the addition of H2O is preferably a sample Teflon™ tubing surrounded by water. The temperature of the tube is advantageously kept constant, so its permeation rate. In such configuration, a gas phase space must be kept to avoid mechanical failure due to pressure build up. By filling the void space with pure O2 or a mix of O2 and sample background and adjusting the length of permeation membrane, an amount Of O2 will also be added to the sample gas. This small amount of O2 into the plasma will react with any carbon deposit by “etching” it. It results that the cell wall stays clean even if Hydrocarbons are voluntarily added at the cell inlet. This solution is effective and economic to keep the plasma cell clean over an extended period of use. A separate permeation tube for oxygen addition could also be used for O2 addition.
Now, still referring to
Our experiments have shown that this solution is very effective to eliminate ionization of surrounding atmosphere even with relatively high field potential.
As previously described, the last problem to be solved is the linearity. As shown in “Spectroscopic analysis of gas mixtures”, O. P. Bochkova and E. Y A. Shreyder, Leningrad State University Academic Press, 1965 SF-4101, and in “Analyse qualitative d'impuretés dans l'hélium par spectroscopic d'émission”, A. Ricard et J. Lefebvre, Analysis, 1978 V.6, No. 7, pp. 299 to 305, dependent on operating conditions, the N2 measurement in rare gas is not linear. The measurement of N2 in He seems to be the worst case, mainly at low levels. In Argon, the problem arises at high levels of N2. To correct such non-linearity, a computational method or a correcting algorithm embedded in the microprocessor 30 could advantageously be used. The linearization method consists in recording the system signal in function of Nitrogen levels, then a polynomial curve fitting equation could advantageously be used. A simple look up table with a linear evaluation between each point could also be implemented. Such software routine is advantageously built into the microprocessor.
Such software method is performing well but does not resolve the loss of sensitivity when N2 level is increasing. Here we mean by the expression sensitivity the increase in signal intensity per unit of increase in solute concentration.
According to this, another correcting means could be applied in order to achieve linearity and constant sensitivity over a wide range of operation. In such configuration an instrument could advantageously be re-ranged to still provide sub ppb sensitivity.
The system illustrated in
Based on the level of N2, the microprocessor 38 readjusts the plasma power of the measurement cell 30, thereby compensating for the non-linearity. So, the N2 measurement signal coming from the measurement cell is a straight line from 0.25 ppb up to 10,000 ppm. The frequency of the plasma could also be changed to increase the emission rate of N2.
With reference to
It is also important to mention that long term stability is a major issue with on line measurement systems.
It should be mentioned that, in a preferred embodiment of the present invention, each trapping means is respectively particularly designed to trap 5 ppm of each gas continuously for a period of one year. It takes a few grams of each different material except for the molecular sieve 3A and the copper based catalyst where we used a much larger quantity in case of atmospheric contamination. Of course, for a particular application, one can envisage other configurations.
In conclusion and according to the experiments that were performed, the on line measurement of N2 in noble gases based on emission spectroscopy using the method of the present invention results in interference free, extremely stable, low noise and linear measures. The proposed method relies on a combination of a plurality of means that are particularly arranged together to correct linearity issues and to cancel interference. According to the best mode of the invention, this combination of means advantageously comprises:
1) Trapping the impurities that interfere with the N2 measurement without affecting the N2 impurities.
2) Selecting an emission wavelength that is not subject to spectral interference from NH emission at 336 and 337 nm, i.e. 357.69 nm. It should however be noted that other N2 emission wavelengths could also be used if all impurities are completely eliminated with the trapping means.
3) Adding moisture to the sample gas just before the plasma chamber in order to stabilize the effect of a slight moisture variation. Indeed, a slight moisture variation in a dry sample results in stronger interference compared to the same variation on a “moisturized” sample.
4) Cancelling the background variation by measuring the background signal in the neighbourhood of the N2 emission wavelength used and subtracting this background signal off the N2 emission wavelength. So the net N2 emission is advantageously used.
5) Measuring plasma pressure and compensate the N2 emission based on the measured pressure.
6) Ensuring that surrounding cell atmosphere will not become ionized. This can be done by sealing the electrode or by enclosing the plasma cell in proper potting compound, thereby eliminating all air surrounding the system.
7) Adding a small amount of O2 to the sample gas at the inlet of the plasma emission system, to eliminate carbon deposit inside the plasma cell in order to keep it clean and avoid lost of sensitivity over the time.
8) Compensating signal non-linearity through a simple software algorithm or by adjusting measurement cell plasma power based on N2 level. Plasma power or frequency could advantageously be controlled to do so.
Although preferred embodiments of the present invention have been described in detail herein and illustrated in the accompanying drawings, it is to be understood that the invention is not limited to these precise embodiments and that various changes and modifications may be effected therein without departing from the scope or spirit of the present invention. For example, the method of the present invention has been described for N2 in Argon. It is however worth mentioning that the present invention could also be applied to N2 in Helium, Neon or Krypton, and could even be applied to other impurities measurements.
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/CA07/00314 | 2/28/2007 | WO | 00 | 8/26/2008 |
| Number | Date | Country | |
|---|---|---|---|
| 60776921 | Feb 2006 | US |
