Alan J. Fougere

Falmouth Scientific, Inc.

P.O. Box 315

1140 Route 28A

Cataumet, MA 02534 USA

afougere@falmouth.com

I. INTRODUCTION

A serious limitation of the long-term stability of CTDs is the use of electrode-type conductivity sensors. Since the electrodes must be exposed directly to seawater they cannot be protected reliably from marine fouling. The small dimensions of these cells makes them particularly vulnerable to even minute amounts of fouling. In 1989 FSI was founded and commenced development of new methods to accurately measure the conductivity of fluids. This endeavor has been very successful, with now over 12,000 conductivity sensors built, calibrated and shipped. A brief review of the essential elements of high-precision conductivity measurements is presented below as a lead-in to an innovative sensor now in Field Testing by FSI and others.

Conductivity (reciprocal resistivity) is an intrinsic property of seawater from which salinity and density may be derived. Oceanographic sensors measure conductance directly; the conductivity measured is dependent on the "cell constant" of the sensor, often referred to as cell "k factor". Therefore, conductivity is found from the conductance measured by the sensor using a scale factor or "k factor". The "k factor" represents the physical geometry of the sensors. If the sensor had perfect geometry then "k = 1", or conductivity = conductance. In all types of cells, the length/area ratio corresponds to the physical dimensions of the cell's "hard parts". Another critical factor in a sensor’s "geometry" are electrodes which alter or disturb the normal flow of current through the sensor. For all sensors the sample volume which determines the cell constant is determined through wet analytical techniques. The precision of these wet analytical techniques limits the user’s ability to accurately determine "electrical conductivity". In oceanography all wet calibrations are determined from the radiometric measurements of ISAPO standard seawater and the sensor under test. As such absolute calibration is limited by the precision of ISAPO water and the determination of its "absolute conductivity" at a measured temperature.

The mathematical determination of conductivity derives from the familiar relationship:

R = rL/A*Ee

where:

R = resistance = 1/conductance

r = resistivity = 1/conductivity

L = length of sampled water volume

A = cross-sectional area of samples water volume

Ee= Electrode Effects

Note the equal importance of sample geometry and measured resistance in the determination of conductivity.

The existing FSI Inductive Conductivity Sensor consists of two coaxially mounted toroidal transformers mounted in an insulated housing, Figure 1.

Fig. 1. Present FSI Inductive Sensor Showing Field Lines.

For the simple feed-forward inductive sensor the electrical theory of operation is as follows:

An alternating voltage in the audio frequency range is applied to transformer T1, Figure 2. This induces a voltage in the seawater circuit. The resulting current (Iw) is directly proportional to the conductivity of the seawater and is measured by the second transformer T2.

The simplified model has two obvious sources of error. The first is the forward voltage drop errors in transformer T1, and the second is the current transformation errors in T2. FSI has developed electronic techniques to eliminate both these sources of error.

II. FSI INDUCTIVE CONDUCTIVITY SENSOR ELECTRONICS

Earlier feed-forward applications of inductively-coupled conductivity sensors and electronics were not particularly successful for a number of reasons, as follows:

1. The toroidal transformers in these sensors had to be pressure protected (Brown 1968) to eliminate the effects of pressure on the electrical parameters of the transformers. The required electrically insulated pressure housing dramatically increased the thermal mass of the sensor, which in turn resulted in substantial thermal contamination of the seawater being measured. The pressure housing also resulted in a small hole through the center, which restricted the seawater path, thus reducing the sensitivity and the sensor’s signal to noise ratio.

2. The major problem with these sensors was the instability of the voltage ratios caused by the combination of the finite electrical resistance (Rw1 and Rw2 in Figure 3) of the transformer windings and the variability of the inductance of these windings. These variations are unpredictable and are influenced by pressure and temperature effects as well as previous magnetic history (magnetic hysteresis) of the magnetic core. The current induced in the seawater circuit is directly proportional to the product of conductivity and the voltage induced in the seawater. Hence changes in the voltage ratio were indistinguishable from changes in seawater conductivity.

Figure 3 details the present approach, which eliminates the previous measurement error sources. This approach uses classic negative feed-back techniques to reduce the effect of these error sources to negligible levels. The operation of this circuit is as follows. The input voltage Eg is applied to one side of drive winding W1. Assume for the moment that the output of amplifier A1 is zero; the applied voltage across W1 will be Eg. The induced voltage in W1 will be equal to the applied voltage less the voltage drop across Rw1. Since the sense winding W2 is wound with exactly the same number of turns as W1, and since essentially zero current flows through Rw2, the voltage across the sense winding will be exactly equal to the induced voltage in W1 and W2. Hence the input to the amplifier A1 will be equal to the applied voltage minus the induced voltage. In other words, if the output of A1 is zero, the input is equal to the error voltage. However, negative feed-back will reduce the error by a factor equal to the gain of A1 by applying an additional voltage across W1 to reduce the error voltage so that the induced voltage is essentially equal to the applied voltage.

A similar technique is used in the output toroid (W3 and W4) to essentially eliminate the effect of resistance of W4. Since there is no current flowing in W3 the input to the very high gain amplifier A2 will only be zero when the product of the current and the turns on W4 is exactly equal to the current in the seawater circuit regardless of the resistance of Rw4. The amplifier A2 is designed for extremely high gain, which reduces the output toroid current ratio error.

This now permits the use of toroidal transformers unprotected from pressure, thus dramatically reducing the thermal mass, complexity and cost, while essentially eliminating the errors caused by the instability in the voltage ratio. The FSI sensor’s relatively large diameter .96" (24 mm) and short length 1.73" (44 mm) of the hole forming the seawater path, and the relatively low overall volume of the housing result in very low thermal inertia compared with the existing electrode types. These effects are discussed in detail by Lueck. The present implementation of the sensor uses 99.9% pure alumina oxide housing to mount the two cores; this is the same material as used in the MKIIIB CTD conductivity sensor. The use of alumina oxide allows the use of temperature and pressure correction coefficients as defined by Millard & Fofonoff. The ceramic sensor eliminates isolation problems typical of coatings used in previous implementations of inductive measurement sensors, and allows the sensor to be cleaned with a bottle brush without affecting the conductivity calibration.

The 0.96" (24 mm) inside diameter of the center hole of an inductively coupled sensor further improves the long-term stability compared with the 3 to 4 mm inside diameter typical of existing electrode-type conductivity cells. The stability of the inductive cell is limited only by the stability of the cell geometry. The sensor also exhibits significantly shorter flushing length than a MKIIIB CTD cell, flushing length being directly proportional to the length to diameter ratio of the sensor.

FSI introduced the commercial version of this sensor in 1991. A picture of our standard large geometry sensor is shown in Figure 4. Since its introduction, FSI has built over 2000 ceramic sensors. The company also introduced a PEEK injection molded version of the sensor in 1992; since its introduction over 10,000 units have been produced. It is the combination of improved materials, up-graded electronic techniques, and unique internal calibration techniques which has resulted in the wide-range acceptance and use of FSI inductive sensors.

Fig. 4. Current FSI Thin-Wall Ceramic Inductive

Conductivity Sensor (New Alace Float Sensor Background)

III. NEW NON-EXTERNAL FIELD INDUCTIVE CONDUCTIVITY SENSOR

During the last several years FSI has been developing sensors for long-term deployments in the ocean. Long-term deployment of the sensor in the ocean can result in dramatic change in the sensor geometry in areas where biological activity is high. It is also evident that the use of pumps or artificial means of sensor aspiration can not be effectively employed due to power limitations, or, in the case of a profiling system, the need for sensible flushing lengths. Present inductive sensors require anti-foulants applied to the center bore hole, which places stringent requirements on the mechanical geometry of such anti-foulants. These requirements include very low ablation rates to prevent cell geometry changes in the high-sensitivity center bore region. Alternately, some success has been achieved using internal field electrode sensors with the ends protected using impregnated tubes which leach toxins. These sensors have had limited success due to the basic instability of the sensor to both biological and mineral fouling of the electrodes; the very small internal geometry has also hindered success due to very poor flushing and the enhanced sensitivity to fouling. It has also recently been reported by users that leakage of the biological toxin onto the electrodes has resulted in premature failure of the sensor electrodes resulting in additional sensor instability. FSI, in conjunction with Brown, conceived and is now testing a new sensor geometry which allows for the use of end tube leaching protection, without the disadvantages of a sensor which does not freely flush or has unstable calibration due to degradation of electrodes from both mineral or organic fouling of the electrodes. The new sensor has been deployed on classical "profiling type" CTDs with a high degree of success. During the last year FSI has developed several methods for protecting this sensor from biological measurement degradation. These methods are now undergoing extensive testing to determine appropriate methodologies for specific deployment applications. The new sensor has been designated the Non External Inductive Conductivity Sensor (NXIC). The following sections describe its theory of operation.

The new sensor, outlined in Figures 5 and 6, and represented schematically in Figure 7, consists of two ceramic tubes mounted parallel to each other. The ends of each tube are fixed to ceramic boxes. Each tube has a pair of toroidal transformers fitted coaxially over them.

The theoretical operation of the new sensor is a follows:

Resistors R1 and R2 represent the resistances from A to B via the upper tube and lower tubes respectively. Resistor R3 represents the resistance from A to B via the path external to the structure. The voltages E1 and E2 represent the voltages induced by the transformers T1 and T2. I1 and I2 are the resulting currents flowing in the upper and lower tubes.

It can be shown that if the ratio R1/E1 is equal to the ratio R2/E2, then I1 will be equal to I2. If the directions of I1 and I2 are opposite, then the difference which flows externally will be zero. This means that the external effects will be zero. Since the seawater in the two tubes has the same conductivity and the dimensions are the same, then the two resistances R1 and R2 are the same. Transformers T1 and T2 have identical windings and are connected to the same voltage, hence E1 and E2 are assured of being equal.

The proof is as follows. If we assume that the potential at A is zero then the following equations apply.

Eb = E1 * R2||R3 / ( R1 + R2||R3) + E2 * R1||R3 / ( R2 + R1||R3)

Where:

R2||R3 is the Parallel Resistance Theorem = R2 * R3 / ( R2 + R3)

Rewriting:

Eb = E1 * R2*R3 / (R1*R2 + R1*R3 + R2*R3) + E2 * R1*R3 / (R1*R2 + R1*R3 + R2*R3)

Eb = E1 / R1 * K + E2/R2 *K

Where:

K = R1*R2*R3 / (R1*R2 + R1*R3 + R2*R3)

If E1 = -E2 and R1 = R2 then it is obvious that the above equations equate to zero and that the voltage across R3 (i.e., the external path) is zero.

In practice it may not be possible to insure that the two tubes are identical. In this case the ratio of E1 to E2 can be adjusted to compensate for the inequality of R1 and R2 to maintain zero external field.

In Figure 8 above the field lines are illustrated. Note that the two cores act as charge pumps pulling the current lines from one end reservoir down the opposite tube. This situation is very similar in operation to guard electrodes used on the open ends of a tube electrode sensor. At any point outside the end reservoirs of the NXIC sensor the potential across the sensor appears as zero. This results in an outside impedance path which is very high compared to the internal "pumped" path, resulting in the elimination of the external field effects.

Fig. 10. Top View (Prototype) FSI NXIC Sensor

Showing Inlet and Measurement Tubes

The new sensor geometry is unique in that it has no external electrical field, while continuing to use "inductive" drive and current sensing, eliminating the need for un-stable electrodes. The unit will have essentially zero external field and hence its calibration is not affected by the proximity of external objects. The unit is a free-flushing sensor. A comparative flushing length of various sensors is detailed in the following table including the new geometry sensor:

~ Corrected for Square Form

We feel that the new sensor has all of the essential features to address a wide range of oceanographic requirements, due to the essential features:

1. Very stable internal geometry
2. No electrodes to maintain, foul, or alter calibration
3. Free flushing
4. The ability to add biocides around the ends to reduce internal fouling
5. The ability to add shutters and biocide injection when used in high-fouling environments
6. Can attain very high level of precision and is not limited by signal to noise ratios, through the use of advanced magnetic materials.
7. Can be built for low cost using ceramics, or very low cost using injection molding techniques developed by FSI.

IV. NXIC BIOLOGICAL PROTECTION

Long-term, in situ sensors suffer from fouling, the major problem in coastal programs. High biological growth can severely alter sensor geometry. After application, antifoulant paints also can alter the geometry through degradation of the paint itself. Presently tri-butyltin (TBT) is the most effective antifoulant material but has strict environmental controls for permitting and use. Some states will allow limited use of TBT with strict guidelines. With antifoulant treatment, salinity sensors presently can operate for up to 3 months in coastal waters during some seasons and 2.5 years in deep water before severe accuracy degradation occurs.

There are few studies available on antifoulants and their effects on sensors and data. According to Atkinson and Woody (1999), one such study was conducted in the Netherlands to select a suitable sensor to measure conductivity and temperature for continuous, in situ monitoring of coastal waters (van Oort, et al., 1998). Upon selection of the sensor, further studies were conducted on the effects of three different antifoulant paints on the sensor. Bondit B21C6 (based on ammonium hydroxide), Seajet 033 (30-60% cupricoxide and 10-30% xylene), and Jotun HSE 3410 (0-1% tri-butyltin and 30-60% cupricoxide) were used, with one sensor free of antifoulant used for visual comparison only. The field tests were conducted in brackish, somewhat stagnant water. Conductivity data was recorded from the three test cases with reference conductivity measured weekly using a hand-held sensor of WTW, model LF196. The results showed Jotun HSE 3412 to be the most effective, and it was found that the use of antifoulant paint increased the maintenance-free period for about a week. Van Oort, et al. (1998) did not feel that this time increase was enough to justify the use of the TBT antifoulant.

An antifoulant is composed of a biocide, carrier, and binder. The biocide discourages the organisms from colonizing on the treated surface, while the binder holds it in place. FSI has initialized a project with Cape Cod Research Inc., (CCR) staff, utilizing their expertise with antifoulants. Mr. R. Scott Morris, of CCR, has suggested two broad-spectrum biocide materials. The two biocides under test are hydrogen peroxide in a water-based polyurethane binder (Trade name NoFoul EP2OOO) and C9 in a PVC binder.

The C9 bio-toxin is a novel bipolymer-based delivery system developed by Cape Cod Research, Inc. to afford the controlled release of bioactive materials. The system comprises a siloxy-ceramer-epoxy terpolymer co-reacted with a bipolymer. It is non-toxic and VOC compliant. Release of bioactive materials is controlled by hydrolysis of the terpolymer, which can be tailored by varying the number of hydrolytic linkages in the terpolymer. This liquid delivery system is blended with EPA- approved biocides, supplied as powders, to form a workable paste. The paste is applied as a coating or set into a mold. Cross-linking occurs as the material dries to form a durable composite-like material with excellent adhesive and cohesive properties. When immersed in water, the polymeric portion of the composite erodes, releasing the bioactive filler.

External surfaces of the NXIC sensor can be painted with EP-2000, a water-based, VOC compliant, antifouling paint manufactured by E paint Co. The paint contains no TBT or copper. Its primary active ingredients are a photoactive zinc oxide, which generates hydrogen peroxide under sunlight, and a proprietary algaecide.

A. Methods Under Test:

1) Time Release Bio-toxin

This version of the sensor will have all the same features of the previous version with the addition of a closure mechanism. The ends of the sensor will have guillotine-style doors, which will close when not sampling, to allow the bio-toxin levels to elevate and reduce the rate at which the bio-toxin dissolves, as shown in Figure 11.

            3) Anodic Generation of Chlorine with Closure Mechanism

This version of the sensor will have an anode made of platinum or titanium plated with platinum on the inside surface of the bottom door and a cathode made of 316L stainless steel on the inside surface of the top door. The doors will periodically close and a DC voltage will be applied to the anode and cathode. This voltage will create chlorine, which will kill any organisms in the sensor. The outer surfaces of the sensor will be painted with EP-2000 antifouling paint.

Fig. 11. NXIC Guillotine Closure

B. Field Operations and Data Telemetry

To accelerate the testing of the NXIC sensor and various approaches to anti-fouling, two to three sensors are being integrated into a buoy in the existing telemetering station operated by the Geochemical and Environmental Research Group (GERG), and the Texas Automated Buoy Program (TABS), offshore Galveston Texas. The field test of the NXIC sensors will take place at the TABS Buoy Site B, which is located off Galveston, Texas (280 58.97' N; 94’54.97' W). The field experiment will be operated for twelve months, with the final three months devoted to post-calibration and analysis of the sensors and to report preparation. The test samples are planned to be deployed in mid summer 2000 with test results completed by spring 2001.

V. CONCLUSION

The NXIC Sensor offers oceanography a new and unique conductivity sensor that can be employed in high biofouling areas with minimal impact on sensor performance. Several innovative protection techniques are under investigation as to the suitability of each to specific deployment applications. The NXIC sensor development has attained the goals and objectives expected for profiling operations; we anticipate that through novel fouling protection approaches presented herein this new sensor will play a major role in the need for improved Global Ocean Observations in high fouling areas.