WO1999030142A1

WO1999030142A1 - Determination of concentration of a compound in a multiple component fluid

Info

Publication number: WO1999030142A1
Application number: PCT/US1998/026097
Authority: WO
Inventors: Lee-Shin Tsai; Bradley J. Hernlem; Charles C. Huxsoll
Original assignee: The United States Of America, As Represented By The Secretary Of Agriculture
Priority date: 1997-12-09
Filing date: 1998-12-09
Publication date: 1999-06-17
Also published as: AU1810299A; WO1999030142A8

Abstract

Methods and apparatus are provided for accurately determining the concentration of a selected compound or compounds in a multiple component fluid, by reducing or eliminating the interference of non-selected compounds in the fluid. In the method of the invention, the concentration of one or more compounds in a test sample fluid which contains at least two compounds is determined using at least two amperometric or potentiometric membrane sensors which have different specificities and which have been calibrated to determine the individual response-to-concentration ratio of each sensor to the selected compound and non-selected or interfering compound. The concentration of the selected compound is determined by simultaneously measuring the response of each membrane sensor to the test sample and computing the concentration of the selected compound using the simultaneously measured responses to the test sample fluid and the calibrated responses of the sensors.

Description

DETERMINATION OF CONCENTRATION OF A COMPOUND IN A .MULTIPLE

COMPONENT FLUID

BACKGROUND OF THE INVE.NTION

1 . Field of the Invention

The present invention relates generally to sensor technology. More particularly, the invention relates to methods and apparatus for accurately determining the concentration of a selected compound or compounds in a fluid test sample, by reducing or eliminating interference of the non-selected compounds in the fluid.

2. Description of the Art

The determination of concentrations or levels of substances using sensors such as amperometric membrane sensors is an established technology. The technology allows real-time monitoring of levels rather than the alternative of titration or colorimetric methods which require collection of samples and addition of reagents. Typically, an amperometric membrane sensor comprises (a) a noble metal cathode, e.g., gold or platinum, covered by or located immediately behind a selective permeable membrane and (b) a reference electrode for controlling the electrochemical potential on the cathode . This can be accomplished, for example, with an Ag/AgCl reference electrode bathed in a buffered Cl ^" solution along with the cathode. The cathode and reference electrode are typically encased together in a plastic housing. The sensor is essentially a - miniature amperometric titration system. The cathode is polarized by an applied voltage. When an oxidant approaches the cathode, a reduction reaction at the cathode generates a current which is recorded by the ammeter. The current generated is proportional to the concentration of oxidants in the solution. Devices and membrane sensors are commercially available.

By selection of hydrophilic or hydrophobic membrane material a degree of selectivity can be achieved for a selected compound or compounds. Despite this however, the selective properties of such membranes are not perfect, and they are subject to a certain amount of interference from other non- selected compounds. In general, the response, S, of a membrane sensor as a function (f) of the concentrations { C_{l t} C₂ . ■ ■ C_n) of the individual compounds (1 to n) in a mixture can be represented by the equation (A) .

S = f ( C_{l f} C₂ , C₃ C_n) (A)

Each different sensor, having a different membrane, will exhibit a different response function. If there are no interactions between compounds in reactions at the sensor surface, then the response function can be assumed to be a linear combination of functions with only a single compound dependency as shown in equation (B) .

n S = ∑ f_j ( C_j ) (B)

J=l

However, nonspecific responses of a membrane sensor, that is , the response of a sensor to one or more compounds other than the selected compound to be measured which interfere with the measurement of the selected compound, can reduce the accuracy of measurement of the selected compound. There is a need for methods and apparatus which minimize or eliminate the nonspecific responses (interferences) so that a selected compound or compounds can be accurately measured in systems containing two or more compounds .

Exemplary of the use of this technology and the need for accurate monitoring of compounds in systems containing multiple components is the monitoring of chlorine dioxide for food processing. Chlorine dioxide is a valuable disinfectant used in the food processing, drinking water, and wastewater treatment industries for controlling the levels of pathogenic organisms . It is a gaseous soluble chemical which has advantages over other chlorine-containing disinfectants in that it does not produce trihalomethane (THM) disinfection byproducts . Recent decisions by the Food and Drug Administration have permitted the use of chlorine dioxide and acidified sodium chlorite for the control of microbes in poultry processing. The regulations specify particular allowable concentrations of these chemicals in relation to other chlorine compounds .

There is a particular need for methods and apparatus for the monitoring of chlorine dioxide in food processing water, for example, poultry chiller water. Chlorination of chiller water has steadily increased in recent years. Today, conservatively speaking, more than 60% of the "Ready-to-Cook" poultry products in the United States are chilled in water that contains some chlorine. The recent reports that 25% of poultry are contaminated by Salmonella and almost 100% are contaminated by Campylobacter is expected to further increase chlorine and chlorine dioxide uses .

Chlorine dioxide has many advantages over chlorine : it is more efficient than chlorine as a bactericide in poultry chiller water; it does not form THM; it does not form mutagens; it has low residuals of chlorine dioxide, chlorite, and chlorate at the level of use . Chlorine dioxide also has other attractive properties for processors: it is less corrosive to stainless steel equipment; the efficacy is not affected by the pH of the treated solution; it does not change the pH of the treated solution; it is more accepted than chlorine internationally. But chlorine dioxide must be generated on site because it is explosive above 10% concentration.

The biggest disadvantage of chlorine dioxide is the lack of an effective monitoring method. Failure to monitor chlorine dioxide levels is not only dangerous, it will also cause overdose or under-dose. Under-dose results in ineffective disinfection. When the water is overdosed, the escaped chlorine dioxide gas subjects workers to health risks. This is the ultimate reason that many processors who have tried chlorine dioxide are hesitant to use it .

Using the monitoring system described herein, the concentration of chlorine dioxide can be more accurately determined, so that it can be used by the food processing industry.

The use of amperometric membrane sensors for sensing levels of oxidized chlorine compounds such as chlorine dioxide and chlorine (hypochlorite) is an established technology. Chlorine dioxide is an uncharged species and diffuses preferentially through hydrophobic membranes such as polytetrafluoroethylene (PTFE) while hydrophilic membrane material is required to admit hypochlorite, the typical form of chlorine in aqueous solution. The selective properties of the membranes are not perfect, however, and sensors having hydrophobic or hydrophilic membranes are subject to a certain amount of interference. A system is needed to compensate for the nonspecific responses of a membrane sensor so that more accurate measurements of concentrations of chlorine, chlorine dioxide, other oxychloro species or other oxidants or reducible compounds can be obtained in systems containing multiple components.

SUMMARY OF THE INVENTION

The present invention is directed towards methods and apparatus for accurately determining the concentration of a selected compound or compounds in a multiple component fluid, by reducing or eliminating interference of the non-selected compounds in the fluid.

In the method of the invention, the concentration of one or more compounds in a test sample fluid which contains at least two compounds is determined using at least two amperometric or potentiometric membrane sensors which have different specificities. Each membrane sensor is calibrated to determine the individual response-to-concentration ratio of a sensor to the selected compound and non-selected or interfering compound. The calibration provides the response functions of each membrane sensor for each compound.

The concentration of a selected compound in a test sample fluid which contains at least two compounds is determined by simultaneously measuring the response of each membrane sensor to the test sample and computing the concentration of the selected compound or compounds using the simultaneously measured responses to the test fluid and the calibrated responses of the sensors, to thereby compensate for the responses due to the non-selected compound or compounds. In this way, the interference of the non-selected compounds is reduced or eliminated, and the concentration of the selected compound is more accurately measured.

The apparatus of the invention useful to carry out the methods of the invention comprises at least two amperometric or potentiometric membrane sensors having different specificities, wherein the sensors are positioned to simultaneously measure responses to the test fluid. The membrane sensors are picked which have a response to the selected compound or compounds to be measured or monitored as described in detail below. The sensors are connected to a circuit or computer to utilize calibration factors to compute and produce signals which are representative of the concentrations of selected compound or compounds without the interference of the other compounds . These signals may then be read out on individual displays or used to provide actuation for a control mechanism.

In accordance with this discovery, it is an object of the invention to provide methods and apparatus for more accurately determining the concentration of a compound in a fluid which contains a plurality of compounds, by reducing or eliminating the contribution of interfering compounds in the measurement of the concentration of a selected compound or compounds .

.Another object of the invention is the provision of effective methods to compensate for the nonspecific responses of a membrane sensor so that accurate measurements of compound concentrations can be obtained in systems containing multiple compounds .

A further object of the invention is to provide an apparatus comprising at least two membrane sensors for accurately determining the concentration of a selected compound in systems containing multiple compounds .

A still further object of the invention is the provision of a method and apparatus for the accurately monitoring chlorine dioxide in food processing water such as poultry chiller water.

Other objects and advantages of the invention will become readily apparent from the ensuing description. BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 displays the calibration data showing the chlorine (Cl₂) sensor responses to chlorine.

FIG. 2 displays the calibration data showing the chlorine dioxide (C10₂) sensor responses to chlorine.

FIG. 3 displays the calibration data showing the chlorine dioxide sensor responses to chlorine dioxide .

FIG. 4 displays the calibration data showing the chlorine sensor responses to chlorine dioxide .

FIG. 5 shows the calculated concentration of chlorine based on sensor responses using three different tests versus the concentration of chlorine determined by titration.

FIG. 6 shows the calculated concentration of chlorine dioxide based on sensor responses using three different tests versus the concentration of chlorine dioxide determined by titration.

FIG. 7 shows the chlorine sensor response versus concentration determined by titration.

DETAILED DESCRIPTION OF THE INVENTION

The methods and apparatus of the invention use membrane sensors and provide a means for accurately determining the concentration of a selected compound or compounds in a test sample fluid which contains multiple compounds, by reducing or eliminating the interference of non-selected compounds in the fluid.

In brief, in the method of the invention, the concentration of one or more compounds in a test sample fluid which contains at least two compounds is determined using at least two amperometric or potentiometric membrane sensors which have different specificities. Each membrane sensor is calibrated to determine the individual response-to-concentration ratios of a sensor to the selected compound and non-selected or interfering compound. This is done by measuring the response-to- concentration ratio of each membrane sensor to a compound in the absence of the other compounds , for example, by measuring the response-to-concentration ratio using pure fluids of (a) the compound whose concentration is to be determined and (b) one or more other compounds in the sample whose response to the membrane sensors is to be reduced or eliminated. The calibration provides the response functions of each membrane sensor for each compound.

The concentration of a selected compound in a test sample fluid which contains at least two compounds is determined by simultaneously measuring the response of each membrane sensor to the test sample and computing the concentration of the selected compound using computational methods which compensate for the responses due to the non-selected compound or compounds. In this way, the interference of the non-selected compounds is reduced or eliminated and the concentration of the selected compound is more accurately measured.

The apparatus of the invention useful to carry out the methods of the invention comprises at least two amperometric or potentiometric membrane sensors having different specificities, wherein the sensors are positioned to simultaneously measure responses to the test sample fluid. The membrane sensors are picked which have a response to the selected compound or compounds to be measured or monitored as described in detail below. The sensors are connected to a circuit or computer to utilize calibration factors to compute and produce signals which are representative of the concentrations of selected compound or compounds, that is, the responses of the sensors are subjected to computation methods to determine the concentration of the selected compound or compounds while eliminating or reducing the interference of the other compounds. These signals may then be read out on individual displays or used to provide actuation for a control mechanism.

The test sample fluid refers to the sample containing a compound whose concentration is to be determined using a membrane sensor in the presence of at least a second compound which has a response to the sensor which interferes with the measurement of the selected compound. A test sample can have the physical attributes of gases, liquids, or a solid wherein the solid can be made soluble in a liquid, and can be of any size or volume, including for example, a moving stream of liquid. The test sample can contain additional components other than the selected compound and interfering compound as long as the other components do not interfere with the sensor measurement of the selected compound. Examples of test samples include, but are not limited to: food processing water, including poultry chiller water, drinking water, and environmental samples such as ground water or waste water, and gaseous samples such as chlorine and/or chlorine dioxide in a gaseous fluid.

A selected compound whose concentration is to be determined by the methods and apparatus of the invention can be any compound that provides a response to a membrane sensor as a function of concentration. For the purposes of this invention, the term "selected compound" includes, but is not limited to, electrochemically oxidizable or reducible compounds, including electrochemically oxidizable or reducible gases and liquids . Exemplary compounds include hydrogen peroxide, ozone, chlorine and its oxides, including chlorine dioxide, chlorite, chlorate, hypochlorite, and hypochlorous acid.

The membrane sensors can be amperometric or potentiometric membrane sensors . Exemplary of amperometric membrane sensors are those which utilize a noble metal cathode and a reference electrode anode such as composed of silver coated with silver chloride . The sensor may also incorporate a counter electrode as known in the art .

The membranes on the sensors are comprised of materials permeable to one or more of the compounds, and preferably have relative selectivity for one or more compounds relative to the others in the test sample fluid.

The sensors are connected to a means to produce signals which are representative of the concentrations of selected compound or compounds, that is, the responses of the sensors are subjected to computation methods to determine the concentration of the selected compound or compounds while eliminating or reducing the interference of the other compounds. Examples of such means, but not limited to, are a circuit or computer, such as a digital computer utilizing sensor signals which have been digitized, or an analog computer. These means can include means to utilize calibration and compensation factors to compute and to store the responses of the sensors. These signals can then be read out on individual displays or used to provide actuation for a control mechanism. For example, the computed concentrations of a compound or compounds in the test sample can be used to provide feedback signals for a control loop such as to maintain a suitable or desired level of one or more of the compounds in the test sample system.

Devices and membrane sensors are commercially available. U.S. Patents No. 5,232,600 describes hydrophobic membranes. U.S. Patent Nos. 5,286,382 and 5,217,802 describe hydrophobic polymeric membrane composites. U.S. Patent Nos. 5,827,980; 4,166,775; 4,057,478; and 4,364,810 describe electrochemical gas detector systems. U.S. Patent No. 5,472,590 describes an ion sensor having an ion selectivity. U.S. Patent No. 5,807,305 describes silver/silver chloride electrodes and composite electrodes. U.S. Patent No. 4,773,969 describes a chlorine ion-selective electrode. U.S. Patent No. 4,581,121 describes a free chlorine gas analyzer. U.S. Patent No. 4,176,032 describes a chlorine dioxide analyzer. U.S. Patent No. 5,393,399 describes an amperometric measuring device having an electrochemical sensor. U.S. Patent No. 5,620,579 describes an apparatus for reduction of bias in amperometric sensors . The foregoing patents are incorporated herein by reference.

A critical feature of the invention is that at least two membrane sensors are positioned in the test sample fluid which contains at least two compounds to simultaneously measure or monitor the test fluid sample containing the selected compound or compounds. If desired, the sensors may be exposed to the same sample fluid with a controlled flow rate of solution past the sensors and with a controlled pressure or head of fluid at all sensors .

A temperature sensor as known in the art such as a resistance temperature detector may be employed to provide compensation for temperature effects on sensor response or incorporate variation in response ratios due to temperature. If desired, an offset correction of a sensor may be computed and applied as known in the art. This is discussed in detail in Example 1, below.

For the sake of simplicity, the method and apparatus are first described for the case wherein two compounds , denoted as the first and second compound, are in the test sample fluid. This case is also described in detail in Example 1, below.

In this case, two membrane sensors, denoted as the first and second membrane sensor, are used which have different specificities for the two compounds in the test sample. The first membrane sensor has a response to the first compound as a function of concentration of the first compound wherein the response-to-concentration ratio is a ₁ and has a response to the second compound as a function of concentration of the second compound wherein the response-to-concentration ratio is a₁₂ ■ The second membrane sensor has a response to the second compound as a function of concentration of the second compound wherein the response-to-concentration ratio is a₂ and either does not have a response to the first compound or has a response to the first compound as a function of concentration of the first compound wherein the response-to-concentration ratio is a₂ι and wherein the ratio of a_u/a_ι2 is not equal to the ratio of a₂ι/a₂₂.

Each membrane sensor is calibrated by separately measuring the response signal to the first compound and the response signal to the second compound. Assuming that the response of each membrane sensor is linear with respect to concentration of either compound and independent of each other, then the following equations describe the response signal, S, of each membrane sensor to concentration, C, of compounds 1 and 2.

£₂ = a₂ιCι + a₂₂^₂ ( 2 )

The response parameters a.^ , wherein i denotes the sensor and j denotes the compound, is determined from the calibration of each membrane sensor to each compound in the absence of the other compound. A calibrant fluid for calibrating a membrane sensor is a fluid which contains a known concentration of a selected compound in the absence of the second compound. Typically, this is done using a pure solution of the compound in fluid. It is preferred that the concentration of the selected compound in the calibrant fluid be in the range of that expected to be in the test sample .

The concentration, C_{l t} of the first compound in the test sample is determined by simultaneously measuring the response Si of the first sensor and the response S₂ of the second sensor and calculating the concentration of the first compound using:

c, = an

(3)

The concentration C₂ of the second compound in the test sample is determined using:

(4 )

The invention is also useful to determine the concentration of one or more compounds in a fluid which contains more than two compounds .

In this case, a plurality of membrane sensors are used, and matrix algebra can be used to solve the response equations . [See, for example, Elementary Differential Equations wi th Linear Algebra , by Albert Rabenstein, Academic Press, Inc., New York, New York, pages 117-119 (1982).] The number of sensors used are at least as many as the number of compounds in the test fluid, excluding the non-interfering substances as defined above, and the matrix of the coefficients (ai-*) of the response equations must have a determinant that does not equal zero. The first membrane sensor has a response to a first compound as a function of concentration of the first compound wherein the response-to-concentration ratio is a _l and has a response to a second compound as a function of concentration of the second compound wherein the response-to-concentration ratio is a₁₂ and has further response to further compounds as a function of concentration of each additional compound where the response- to-concentration ratios are expressed as ai_j for all integer values of j between 3 and n, the total number of compounds.

The additional membrane sensors have responses to the first, second, and additional compounds as functions of the concentrations of the compounds wherein a response to concentration ratio a^ is defined for each sensor i and compound j such that determinant of the matrix A, denoted as det A, of all response-to-concentration ratios a^ for all sensors and compounds is not equal to zero. The response equations are:

<3ιιCι + aι_{2 2} + . . . + aι_nC_n = Si a₂ιCι + a₂₂C₂ + . . . + a_2nC = S₂

a_nιCι + a_{n2 2} + . . . + a_nnC_n = S_n The matrix A is

a , , a ,₂ a ,„ a 21 a ₂ ^* fl 2, -i α „ α ,,„

The sensors are calibrated by measuring the response of the ith sensor to known concentration of the j'th compound in fluid absent the additional compounds (calibrant fluid) and calculating the response-to-concentration ratio a _j of each sensor i for each compound j .

The concentration of a selected compound in the test sample fluid denoted as Cj is dete.rmined by simultaneously measuring the response Si of all the sensors Si from 1-n and calculating the concentration Cj using:

(5)

wherein the matrix Bj is identical to matrix A with the jth column replaced by the matrix of the sensors signals, S, wherein

and det B is the determinant of matrix B .

For example to determine the answer for C₂ the matrix B₂ is constructed as follows:

«ii

B a ₂ \ a n \ s. a .

The foregoing is illustrative of cases where the responses are linear. Nonlinear responses can be corrected using standard mathematical techniques for solving systems of nonlinear equations.

EXAMPLE 1

The following example, which shows the method and apparatus of the invention for use to accurately determine chlorine dioxide and chlorine in a test sample, is intended only to further illustrate the invention and is not intended to limit the scope of the invention which is defined by the claims .

Material and Methods . The membrane sensors used were the chlorine dioxide sensor and the residual chlorine sensor commercially available from Analytical Technology, Inc. (Oaks, PA 19456) . The chlorine dioxide sensor having a hydrophobic membrane was that supplied with a Model A15/65 Chlorine Dioxide Monitor by Analytical Technology, Inc. It is a polarographic membrane sensor which measures chlorine dioxide directly. The chlorine sensor had a hydrophilic membrane. The membrane sensors were housed in a acrylic plastic flow chamber.

The experimental apparatus was a jacketed glass 1.5 L vessel held at a constant temperature at about 22°C with a circulating water bath. Distilled deionized water was circulated, using a peristaltic pump, from the glass vessel through the sensor flow chamber described above and back to the vessel. The glass vessel and sensor flow chamber were connected with TYGON and silicone tubing. Each sensor was inserted into an acrylic plastic block with the membrane exposed to the vessel . Concentrated chlorine or chlorine dioxide solution (in the range of 3000 ppm) was introduced into the glass vessel with a syringe. With each addition of chlorine or chlorine dioxide the responses on the sensor monitors were recorded.

Sample solution was pumped through the flow chamber at a flow rate of 850 mL/min. and across each sensor membrane with a constant head of pressure and flow. A custom signal conditioning circuit was constructed to apply polarization voltage to each sensor and measure the current response of each sensor. The signals were digitized and recorded using a 16-bit AD converter board (PCI-MIO-16XE-10 by National Instruments) . A program written in LabView language logged the sensor response data and calculated the concentrations of chlorine and chlorine dioxide in real time using equations (3) and (4) described above, to thereby provide more accurate chlorine and chlorine dioxide concentration values .

A series of experiments were carried out to determine the constants (an), (a₂ι) , (a_i2)and (a₂₂) in equations (1) and (2) above using pure systems containing either chlorine or chlorine dioxide alone in water. Samples of concentrated chlorine or chlorine dioxide were added to the glass vessel as needed to adjust the concentration of these compounds.

Stock solutions of Cl₂ (hypochlorite) were prepared fresh by bubbling Cl₂ gas into a 6 g/L NaOH solution stirred at 4°C until the pH of the solution fell below 8.5. Stock solutions of chlorine dioxide were prepared by passing chlorine dioxide gas into aqueous solution.

Actual concentrations of chlorine and chlorine dioxide were determined by amperometric titration with standardized phenylarsine oxide (0.00564 N) using Standard Methods ( Standard Methods for the Examination of Water and Wastewater, 17th edition, AP.HA-AWWA-WPCF, 1989) . Calibration. To determine the response-to-concentration ratio of each sensor to chlorine and chlorine dioxide individually, a series of measurements was made using solutions containing different concentrations of either component separately.

The chlorine amperometric membrane sensor was calibrated by:

(1) measuring the response signal, Si, of chlorine in aqueous solutions containing known concentrations, C_{l t} of chlorine only, and calculating the value of ₁/C₁ to obtain ( a₁ ) (see Table 1 and FIG. 1) ;

(2) measuring the response signal, Si, of chlorine dioxide in aqueous solutions of known concentrations, C₂, of chlorine dioxide only, and calculating the value of Sι/C₂ to obtain (a₁₂)

(see Table 1 and FIG. 4) .

The chlorine dioxide amperometric membrane sensor was calibrated by:

(1) measuring the response signal, S₂, of chlorine in aqueous solutions containing known, concentrations, C_{l t} of chlorine only, and calculating the value of S₂/ C to obtain (a₂ι) (see Table 2 and FIG. 2) ;

(2) measuring the response signal, S₂, of chlorine dioxide in aqueous solutions of known concentrations, C₂, of chlorine dioxide only, and calculating the value of S₂/ C₂ to obtain ( a₂₂ )

(see Table 2 and FIG. 3) .

The amount of chlorine in the test sample was determined by:

(1) simultaneously measuring the response signal, Si, of the chlorine amperometric membrane sensor (first sensor) to the test solution and response signal, S₂, of the chlorine dioxide amperometric membrane sensor (second sensor) to the test solution, and (2) calculating the concentration of chlorine, C_λ .

The amount of chlorine dioxide in the test sample was determined by:

(1) measuring the response signal, Si, of the chlorine amperometric membrane sensor (first sensor) and response signal, S₂ , of the chlorine dioxide amperometric membrane sensor (second sensor) , and

(2) calculating the concentration of chlorine dioxide, C₂ :

To test the algorithm, a series of measurements were made such than an aliquot of chlorine stock solution was added to a recirculating deionized water solution followed by an injection of chlorine dioxide stock solution after the sensor response for chlorine had stabilized. Fresh thermally equilibrated water was used for each pair of measurements .

The resultant data were handled in three different ways, termed "uncompensated" , "test 1" and "test 2" according to the values of the calibration parameters applied. For test 1, it was assumed that because in pure water both sensors exhibit essentially zero offset, the ratio of signal-to-concentration should be used for determination of the an, a₁₂ , a_{21 l} and a₂₂ parameters. Therefore, these were obtained from the average values in Tables 1 and 2. In reality, as observed in FIGS. 1 through 4, the sensor responses followed trends with non-zero intercepts . In test 2 these parameters were taken from the least squares linear slope through the data and the intercepts were lumped into the offset terms, X₀. The actual values of the parameters used are tabulated in Table 3. FIGS . 5 and 6 show the results of these comparisons. FIG. 7 shows the plot of the chlorine sensor response versus titrated chlorine for the single compound cases in this experiment for comparison with the FIG. 1 calibration data.

Examining FIG. 5 we see the data for each case presented as a pair of vertical points at each specific chlorine concentration. This is because a measurement was taken after chlorine was added and then after chlorine dioxide was added. The concentration of chlorine remained the same in each case but the chlorine dioxide interference is indicated by the higher level point. In the uncompensated case, only the single compound calibration parameters, an, and a₂₂ were applied and severe interference by chlorine dioxide is observed. Application of the compensation algorithm using the test 1 and 2 values listed in Table 3 show significant improvement in reducing the chlorine dioxide interference so that the compensated response values approach the expected trend indicated by the line .

The data in FIG. 6 are displayed similarly except in this case the single compound chlorine points are lumped at 0 ppm chlorine dioxide . There is little difference between the uncompensated and compensated test cases for the chlorine dioxide sensor because of its relatively superior specificity. Test 1 results mirror the uncompensated case because the same offset terms were used in each case . The test 2 parameters result in a trend which parallels the expected trend but with a positive offset. This could have been eliminated by making further adjustments in the offset terms. Variations in the 0 ppm data are a manifestation of the "memory" effect of these sensors. This effect is the result of the delay, particularly in the case of the chlorine dioxide sensor to return to zero signal when removed from the presence of chlorine dioxide. It is this effect that is probably also responsible for the non-zero intercept behavior seen in the calibration plots.

Despite the non-ideal behavior of these sensors, this data demonstrates the ability of the invention to significantly reduce the interference observed when these sensors are used individually in mixed solutions of chlorine and chlorine dioxide .

Table 1: Cl₂ Calibration Data ppm CI2 ci₂ sensor cιo2 sensor ci₂ sensor ratio cιo2 sensor response response μ^V/ppm ratio μA μA μ^ ppm

5 ₃₈ 0 86 0 071 0 160 0 013

7 76 1 17 0 105 0 151 0 014

8 96 1 26 0 116 0 141 O OP

4 69 0 87 0 079 0 186 0 017

3 00 0 69 0 058 0230 0 019

8 26 1 15 0 100 0 139 0012 ave 0 168 0015

Tble 2: cιo2 Calibration Data ppm C102 sensor C102 sensor ci₂ sensor _C102 sensor response response ratio ratio μA μA μA/ppm μA/ppm

2.14 0.72 5.24 0.336 2.45

2.51 0.84 6.14 0.335 2.45

6.02 1.38 11.54 0.229 1.92

-.48 1.42 11.84 0.219 1.83

3.27 1.04 7.20 0.318 2.20

1.54 0.63 4.53 0.409 2.94

5.85 1.39 11.35 0.238 1.94

7.47 1.50 13.50 0.201 1.81

Table 3: Test Algorithm Parameters parameter uncompensated Test 1 Test 2

X0,1 0.006 0.006 0.400

X0,2 0.002 0.002 2.178 all 0.168 0.168 0.095 a21 0.000 0.015 0.025 a21 0.000 0.286 0.155 a22 2.190 2.190 1.529

It is understood that the foregoing detailed description is given merely by way of illustration and that modification and variations may be made within, without departing from the spirit and scope of the invention.

Claims

What is claimed is :

1. A method for determining the concentration of a compound in a fluid which contains at least two compounds, using amperometric or potentiometric membrane sensors, which comprises :

(a) measuring response to said fluid with a first membrane sensor, said sensor having been calibrated by measuring its response to concentration of a first compound in the absence of other compounds, and by measuring its response to concentration of a second compound in the absence of other compounds ;

(b) simultaneously measuring response to said fluid with a second membrane sensor having a different specificity than said first sensor, said second sensor having been calibrated by measuring its response to said first compound in the absence of other compounds, and by measuring its response to said second compound in the absence of other compounds; and

(c) determining the concentration of said first compound using said simultaneously measured responses and said calibration responses .

2. A method using at least two membrane sensors for accurately determining the concentration of a compound in a test sample fluid which contains at least two compounds, which comprises :

(a) providing a first membrane sensor which (i) has a response to a first compound as a function of concentration of said first compound wherein the response-to-concentration ratio is an and wherein an is a value calculated as the response of said first sensor to known concentration of said first compound in absence of a second compound, and (ii) has a response to a second compound as a function of concentration of said second compound wherein the response-to-concentration ratio is a╬╣ , and a_x2 is a value calculated as the response of said first sensor to known concentration of said second compound in the absence of said first compound;

(b) providing a second membrane sensor which (i) has a response to said second compound as a function of concentration of said second compound wherein the response-to-concentration ratio is a ₂ , and a₂2 is a value calculated as the response of said second sensor to known concentration of said second compound in the absence of said first compound and wherein said sensor either does not have a response to said first compound or has a further response to said first compound as a function of concentration of said first compound wherein the response- to-concentration ratio is a₂₁ and a₂₁ is a value calculated as the response of said second sensor to known concentration of said first compound in the absence of said second compound, wherein the ratio of an/a₁₂ is not equal to the ratio of _2╬╣/a₂₂; (c) determining a concentration d of said first compound in said test sample fluid containing said first and second compounds by simultaneously measuring response S of said first sensor and response S₂ of said second sensor and calculating said concentration using:

3. The method of claim 2 which further includes : (d) determining a concentration C₂ of said second compound in said test sample using:

4. A method using at least two membrane sensors for accurately determining the concentration of a compound in a test sample fluid which contains at least two compounds, which comprises :

(a) providing a first membrane sensor which has a response to a first compound as a function of concentration of said first compound wherein the response-to-concentration ratio is a- and has a response to a second compound as a function of concentration of said second compound wherein the response-to- concentration ratio is a╬╣₂;

(b) providing a second membrane sensor which (i) has a response to said second compound as a function of concentration of said second compound wherein the response-to-concentration ratio is a₂₂ and either does not have a response to said first compound or has a response to said first compound as a function of concentration of said first compound wherein the response- to-concentration ratio is a₂₁ and wherein the ratio of a_l /a╧ç₂ is not equal to the ratio of a₂₁/a₂₂ ;

(c) calibrating said first membrane sensor by (i) measuring the response of said first sensor to known concentration of said first compound in a fluid in the absence of said second compound and calculating a_╬╗l as a ratio of said response to said concentration, and (ii) measuring the response of said first sensor to known concentration of said second compound in aqueous solution in the absence of said first compound and calculating a₁₂ as a ratio of said response to said concentration;

(d) calibrating said second membrane sensor by (i) measuring the response of said second sensor to known concentration of said second compound in said fluid in the absence of said first compound and calculating a₂₂ as a ratio of said response to said concentration, and (ii) if said second membrane sensor has a response to said first compound, further measuring the response of said second sensor to known concentration of said first compound in said fluid in the absence of said second compound and calculating a₂₁ as a ratio of said response to said concentration;

(e) determining a concentration C of said first compound in a test sample fluid containing said first and second compounds by simultaneously measuring a response _x of said first sensor and a response S₂ of said second sensor and calculating said concentration using:

5. The method of claim 4 which further includes : (f) determining a concentration C₂ of said second compound in said test sample fluid using:

s₂ - a _{2 l}

6. A method using a plurality of membrane sensors for accurately determining the concentration of a compound in a test sample fluid which contains a plurality of compounds, which comprises :

(a) providing a first membrane sensor which has a response to a first compound as a function of concentration of said first compound wherein the response-to-concentration ratio is an and has a response to a second compound as a function of concentration of said second compound wherein the response-to- concentration ratio is a₁₂ and has a response to further compounds as a function of concentration of said further compounds wherein the response-to-concentration ratios are a_Xj for all integer values of j between 3 and n, the total number of compounds ;

(b) providing further membrane sensors which have responses to said first, second, and further compounds as functions of the concentrations of said compounds wherein a response-to- concentration ratio for each sensor i and compound j is i_j , wherein matrix A is defined as the matrix of all response-to- concentration ratios aij for all sensors and compounds, and wherein the determinant of the matrix A, det A, is not equal to zero;

(c) calibrating said sensors by (i) measuring the response of each of said sensors to known concentration of said first compound in the absence of said second and further compounds and calculating an as a ratio of said response to said concentration for each sensor i, and (ii) measuring the response of each of said sensors to known concentration of said second compound and each of said further compound alone, and calculating the response-to-concentration ratio a┬▒_j as the response of sensor i to the concentration of compound j ;

(d) determining a concentration of a compound C_j in said test sample fluid by simultaneously measuring the response i of all the sensors i from l-n and calculating the concentration C_j using:

B_j ╧é⁼ l Γëñ j Γëñ n ╬¼ A wherein the matrix Bj is identical to matrix A with the j'th column replaced by the matrix of the sensors signals, S, wherein

and det B is the determinant of matrix B.

7. A method using a plurality of membrane sensors for accurately determining the concentration of a compound in a test sample fluid which contains a plurality of compounds, which comprises :

(a) providing a first membrane sensor which (i) has a response to a first compound as a function of concentration of said first compound wherein the response-to-concentration ratio is a_xl and wherein an is a value calculated as the response of said first sensor to known concentration of said first compound in absence of other compounds, (ii) has a response to a second compound as a function of concentration of said second compound wherein the response-to-concentration ratio is a₁₂ and wherein a_i is a value calculated as the response of said first sensor to known concentration of said second compound in the absence of other compounds, and (iii) has a response to a further compound as a function of concentration of said further compound wherein the response-to-concentration ratio is a_Xj for all integer values of j between 3 and n, the total number of compounds, wherein a_Xj is a value calculated as the response of said first sensor to known concentration of said further compound in the absence of other compounds ; (b) providing further membrane sensors which (i) have responses to said first, second, and further compounds as functions of the concentrations of said compounds wherein a response-to-concentration ratio for each sensor i and compound j is a _j , wherein matrix A is defined as the matrix of all response-to-concentration ratios a^ for all sensors and compounds, wherein the determinant of the matrix A, det A, is not equal to zero, and wherein ai_j is a value calculated as the response of the ith sensor to known concentration of the j'th compound in the absence of other compounds ;

(c) determining a concentration of a compound _j in said test sample fluid by simultaneously measuring the response Si of all the sensors Si from 1-n and calculating the concentration _j using:

det B,

C = ΓÇöΓÇöΓÇö 1 < j < n ^; det ,4 '

wherein the matrix B_j is identical to matrix A with the j^'th column replaced by the matrix of the sensors signals, S, wherein

5, S i

S =

and det B is the determinant of matrix B.

8. The method of claim 7 wherein said test fluid is provided to said sensors at a controlled flow rate and a controlled pressure.

9. In an apparatus for determining concentration levels of first and second compounds in a fluid with amperometric or potentiometric membrane sensors, the improvement comprising:

(a) a first sensor positioned to measure responses to said fluid; said sensor having been calibrated by measuring its response to said first compound in the absence of other compounds, and by measuring its response to said second compound in the absence of other compounds ;

(b) a second sensor having a different specificity than said first sensor, wherein said second sensor is positioned to simultaneously measure responses to said fluid as said first sensor; said second sensor having been calibrated by measuring its response to said first compound in the absence of other compounds, and by measuring its response to said second compound in the absence of other compounds ;

(c) computation means having stored therein responses from said calibrations of said first and second sensors; said computation means connected to said first and second sensors to produce signals in response to the combination of (i) said stored responses and (ii) outputs from said first and second sensors generated by said fluid.