Electrochemical hydrogen sensor

Abstract

A process for determining a content of hydrogen in a fluid medium includes contacting the fluid medium with a sensor. The sensor has a housing enclosing a chamber containing an ionic liquid electrolyte, a window which is permeable to hydrogen and positioned in an opening in the housing, and electrodes in contact with the ionic liquid electrolyte in the chamber. Hydrogen is allowed to pass through the window from the fluid medium into the electrolyte and the sensor is heated. Temperature and pressure of the fluid medium is determined and electrical potential is applied to the electrodes. The method also includes measuring current flow. The sensor can be used to observe hydrogen concentration by voltammetry. The method and sensor may be used for measuring downhole hydrogen content, monitoring fiber-optic cables for damage by hydrogen, corrosion monitoring, and in small-scale process plants where hydrogen is part of a gas stream.

Claims

1. A process for determining a content of hydrogen in a subterranean fluid medium, the process comprising: contacting the subterranean fluid medium with a sensor which comprises a housing enclosing a chamber containing a quantity of electrolyte which is an ionic liquid electrolyte, a window which is permeable to hydrogen positioned in an opening in the housing, a plurality of electrodes in contact with the ionic liquid electrolyte in the chamber, and a sensing unit enclosed within the chamber and configured to monitor at least one of a temperature or a pressure of the subterranean fluid medium; allowing hydrogen to pass through the window from the subterranean fluid medium into the electrolyte; heating the sensor to a sensor temperature above a prevailing temperature of the subterranean fluid medium where the prevailing temperature is at least 100° C.; determining at least one of the temperature or the pressure of the subterranean fluid medium with the sensing unit; applying electrical potential to the plurality of electrodes; measuring a current flow; and determining the content of hydrogen based on the current flow and at least one of the temperature or the pressure of the subterranean fluid medium.

2. The process according to claim 1, further comprising allowing a hydrogen concentration across the window to reach an equilibrium.

3. The process according to claim 2, wherein heating the sensor reduces an amount of time it takes for the hydrogen concentration across the window to reach the equilibrium.

4. The process according to claim 1, wherein the sensor is heated by the subterranean fluid medium.

5. The process according to claim 1, wherein the sensing unit is a combined temperature and pressure sensor.

6. The process according to claim 1, further comprising maintaining the sensor temperature at a fixed temperature.

7. The process according to claim 1, wherein the window comprises a hydrogen permeable metal membrane that is selectively permeable to hydrogen.

8. The process according to claim 1, wherein the window comprises a hydrogen permeable membrane on a porous support, the membrane comprising a sheet of metal that is selectively permeable to hydrogen.

9. The process according to claim 8, wherein the sheet of metal that is selectively permeable to hydrogen is not more than 200 μm thick.

10. The process according to claim 8, wherein the sheet of metal that is selectively permeable to hydrogen is overlaid with a protective layer.

11. The process according to claim 1, wherein the window comprises a hydrogen permeable membrane on a porous support, the membrane comprising a polymer film and a layer of metal that is selectively permeable to hydrogen deposited on the polymer film.

12. The process according to claim 1, wherein the ionic liquid electrolyte fills the chamber and a concentration of water in the ionic liquid electrolyte is not more than 1% by weight.

13. The process according to claim 1, wherein an electrode of the plurality of electrodes comprises palladium nanoparticles.

14. The process according to claim 1, wherein the sensor is part of a downhole tool.

15. The process according to claim 1, wherein the sensor is connected to a protective enclosure containing a fiber-optic cable.

16. The process according to claim 1, wherein the subterranean fluid medium is a flow stream which is a feedstock, product, intermediate product or exhaust of at least one chemical reaction which consumes or generates hydrogen.

17. The process according to claim 16, wherein the at least one chemical reaction is any one of: partial oxidation of methane to syngas; Fischer-Tropsch conversion of syngas to alkanes and water; methanol synthesis; hydrocracking of hydrocarbons; water gas shift reaction; or hydrodesulfurization.

18. The process according to claim 1, wherein the process is carried out in a well that is equipped for carrying out upgrading of hydrocarbon by reaction with hydrogen at a downhole location, and wherein the subterranean fluid medium is a flow stream which is a feedstock to the well or produced therefrom.

19. The process according to claim 1, wherein the process is carried out with the sensor located in a fuel cell at a position to detect a leakage of hydrogen in the fuel cell.

20. The process according to claim 19, wherein the fuel cell is a polymer electrolyte membrane fuel cell that comprises a cathode exposed to a cathode gas stream containing hydrogen and an anode exposed to an anode gas stream containing oxygen and the sensor is located at a position to detect a leakage of hydrogen from the cathode gas stream into the anode gas stream.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a schematic cross-sectional view of a hydrogen sensor on line A-A of FIG. 2;

(2) FIG. 2 is a cross-sectional view on line B-B of FIG. 1;

(3) FIG. 3 is an enlarged cross-section of part of the hydrogen permeable window of the sensor of FIG. 1;

(4) FIGS. 4 and 5 show results of cyclic voltammetry with microelectrodes in Example 1;

(5) FIG. 6 shows results of square wave voltammetry with microelectrodes in Example 1;

(6) FIGS. 7 and 8 show results of cyclic voltammetry with macrodisc electrodes in Example 2;

(7) FIG. 9 shows results of square wave voltammetry with macrodisc electrodes in Example 2;

(8) FIG. 10 is a schematic cross-section of a downhole tool in a wellbore;

(9) FIG. 11 is a cross-section of a sensor attached to the flowline of the tool of FIG. 10;

(10) FIG. 12 is a schematic cross-sectional view, similar to FIG. 1, for a sensor used as a corrosion monitor;

(11) FIG. 13 is a schematic cross-section of a hydrogen sensor monitoring a fiber optic cable;

(12) FIG. 14 is a block diagram of a small-scale gas-to-liquid unit producing alkanes;

(13) FIG. 15 is a block diagram of a small-scale gas-to-liquid unit producing methanol;

(14) FIG. 16 is a block diagram of a unit for cracking wax produced by a Fischer-Tropsch reactor;

(15) FIG. 17 is a schematic of equipment for modifying heavy oil while it is still below ground; and

(16) FIG. 18 is a schematic of a PEM fuel cell.

DETAILED DESCRIPTION

(17) FIG. 1 shows an embodiment of hydrogen sensor in cross-section. The sensor is shaped as a bulkhead compartment. It has a body 10 with a cylindrical dish shape including a flat base. The body 10 is made from an electrically insulating material and in this embodiment this material is a machinable ceramic. A suitable ceramic is a material made by Corning Inc., Corning, N.Y. and sold under their trademark MACOR. The manufacturer describes it as fluorphlogopite mica (approximately 55%) in a borosilicate glass matrix. Other possible materials are high temperature polymers such as polyetheretherketone (PEEK).

(18) Three circular electrodes are provided on the base of the body 10 and connections 12 to each of these electrodes extend through the material of the body 10. These electrodes are a working electrode 14, a counter electrode 16 and a reference electrode 18 arranged in a layout shown by FIG. 2 which puts the counter and reference electrodes at similar distances from the working electrode 14. Suitable materials for the working electrode 14 include platinum and palladium. The reference electrode 18 may be silver and the counter electrode 16 may be platinum. All three electrodes 14, 16, 18 are sufficiently large to be classed as macroelectrodes. In this embodiment, the working electrode 14 has a surface area of approximately 1 mm.sup.2, while the counter electrode 16 is larger than the working electrode 14 and has a surface area of approximately 3 mm.sup.2 and the reference electrode 18 has a surface area of approximately 0.2 mm.sup.2.

(19) When the sensor is used to detect low concentrations, such as in leak detection and the dissolved hydrogen content in the ionic liquid is less than about 10.sup.−4 molar, then the working electrode can be composed of palladium nanoparticles, such as those described by Liang et al., “Controllable synthesis of hollow hierarchical palladium nanostructures with enhanced activity for proton/hydrogen sensing”, Journal of Physical Chemistry C., 2008, vol. 112, pp. 338-344.

(20) Silver in an ionic liquid is not a standard reference electrode and may not maintain an absolute value of potential. Nevertheless, it is sufficiently stable to allow electrochemical voltammetry and it is sometimes referred to as a pseudo reference. If desired, a metallocene such as ferrocene may be dissolved in the ionic liquid so that the electrochemical signal includes the reversible oxidation of the metallocene as an internal standard.

(21) A window 20 which is permeable to hydrogen is sealed to the sensor body 10 so as to fit in the mouth of the dish shape. In this embodiment the hydrogen-permeable window 20 is parallel to the base of the body 10 but spaced from it. As shown by the enlarged view in FIG. 3, this window 20 has a thin membrane 22 which is selectively permeable to hydrogen supported on a disc 24 of porous inert material such as alumina. This disc 24 is not specific in its porosity and would allow passage of other gases, if any other gases happened to be present.

(22) The membrane 22 is a thin sheet of palladium metal or an alloy of palladium, such as palladium alloyed with approximately 40% copper. The sheet may possibly be 0.01-100 μm in thickness. The supporting disc 24 may have a thickness of 0.1-10 mm, depending on the expected pressure of the test fluid in contact with window 20.

(23) In an alternative construction the membrane 22 is formed by palladium sputtered onto a polymer membrane or porous polymer support which may be a polyetherimide, polyether ether ketone (PEEK), polysulfone or polytetrafluoroethylene (PTFE). This polymer membrane is placed on the inert porous support 24.

(24) The membrane 22 is coated with a thin polymer film 25, for instance a film of poly(para-phenylterephthalamide) which is available under the trade name Kevlar or polybenzimidazole (PBI) to protect it from mechanical damage or chemical attack. The polymer coating 25 may be impregnated with one or more scavenger reagents to removing compounds harmful to the palladium membrane. For example zinc oxide can be dispersed into the protective polymer coating to capture hydrogen sulphide by the formation of zinc sulphide and water.

(25) As indicated in FIG. 2, a combined temperature and pressure sensor 26 is also provided on the base of the body 10 to measure temperature and pressure within the enclosed interior chamber 28 bounded by the body 10 and the window 20.

(26) This chamber 28 is filled with an electrolyte which is an ionic liquid. In this embodiment the electrolyte is liquid at ambient temperatures of 20-25° C. and remains liquid down to −10° C. Ionic liquids are described in a number of documents. A discussion by Silvester et al. is at Zeitschrift fur physikalische Chemie, 2006, vol. 220, pp. 1247-1274.

(27) Some examples of ionic liquids which may be used are 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, 1-butyl-3-methylimidazolium trifluoromethanesulfonate, 1-butyl-3-methylimidazolium tetrafluoroborate, 1-butyl-3-methylimidazolium hexafluorophosphate, 1-hexyl-3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate, N-butyl-N-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide, 1-hexyl-3-methylimidazolium chloride, 1-butyl-3-methylimidazolium nitrate, Hexyltriethylammonium bis(trifluoromethylsulfonyl)imide, and Tetradecyltrihexylphosphonium bis(trifluoromethylsulfonyl)imide.

(28) It will be appreciated that the ionic liquids listed above have cations which are organic molecules. They contain alkyl groups and/or an organic ring. In some of the above liquids the anions incorporate carbon atoms as trifluoromethyl groups. Ionic liquids which contain bis(trifluoromethylsulfonyl)imide or hexafluorophosphate anions are generally insoluble in water while the solubility of water in them is low, such as less than 0.5% possibly less than 0.1% by weight. Thus, ionic liquids with these anions have a relatively low atmospheric water uptake. The ionic liquid used in the chamber 28 may be one of the above liquids with a bis(trifluoromethylsulfonyl)imide or hexafluorophosphate anion.

(29) The electrodes 14, 16, 18 are connected by connections 12 to a potentiostat 32 for applying potential to the electrodes and measuring current flow. The potentiostat 32 is connected to a controlling computer 34 which receives and records values of current flow at the applied potential.

(30) In use the sensor is brought into contact with the fluid medium to be examined so that this is in contact with the exterior face of the hydrogen-permeable window 20. Hydrogen dissolved in this fluid medium can diffuse through the membrane and the support into solution in the ionic liquid electrolyte. Hydrogen can pass through the membrane in both directions and an equilibrium will be reached with concentration in the electrolyte proportional to the concentration in the fluid medium which is being examined. Electrical potential is applied to the electrodes and the current flow associated with electrochemical oxidation of hydrogen
H.sub.2.fwdarw.2H.sup.++2e.sup.−
is measured. This current flow is a measure of the hydrogen concentration in the ionic liquid electrolyte and can be used to determine the concentration in the fluid medium which is being examined after calibration of the sensor to determine the relationship between equilibrium concentrations employed in the external fluid medium being examined and in the ionic liquid in the interior chamber 28 of the sensor.

Example 1

(31) In this Example 1, experiments were carried out using an ionic liquid and hydrogen gas, without a membrane, in order to demonstrate the electrochemistry.

(32) Metal wire electrodes were sealed into a small trough in the base of a plastic dish provided with inlets and outlets for the flow of gas. The working electrode was a platinum disc of approximately 0.5 mm diameter and was shown by electrochemical calibration to have a diameter of 530 μm. The reference electrode was a silver wire of approximately 0.5 mm diameter and the counter electrode was a platinum wire also of approximately 0.5 mm diameter.

(33) The ionic liquid 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (abbreviation [C.sub.4mim][NTf.sub.2]) from QUILL, Queens University, Belfast, UK, was additionally dried and stirred under vacuum at 70° C. for 24 hours. 30 μL of this ionic liquid was pipetted into the trough to cover the three electrodes and an air-tight lid was placed on the dish. Hydrogen gas was passed through from inlet to outlet, so that the ionic liquid was in direct contact with the hydrogen. This was continued for about ten minutes to ensure complete equilibrium between the gas and liquid phases before cyclic voltammetry was carried out. (However, it was noted that the same electrochemical results were obtained if hydrogen was passed through for less than a minute, indicating that equilibrium between the gas and liquid phases was achieved in this short time).

(34) The electrodes were connected to a μ-autolab II potentiostat (Ecochemie, Netherlands) which was used to carry out cyclic voltammetry measurements at room temperature of 22° C., varying the potential applied to the working electrode, relative to the silver reference electrode, between −0.5 volt and +1.1 volt and recording the current flow as the potential was varied. The scan rate was 100 mVsec.sup.−1. This procedure was then repeated using mixtures of 1%, 10% and 50% hydrogen in nitrogen. Results are shown in FIG. 4. The voltammetric wave can be seen to contain an oxidation peak at about +0.7 volt. A plot of the peak current against hydrogen concentration is shown in FIG. 5 and shows a linear relation of electric current to hydrogen concentration.

(35) Square wave voltammetry was also carried out over the same voltage range and the results are shown in FIG. 6. At each of the hydrogen concentrations, a large and sharp peak at about +0.7 volt relative to the silver reference was observed, which, even at the lowest concentration of hydrogen (1%), was seen to be separated from the baseline (blank ionic liquid with no exposure to hydrogen) shown as a dotted line.

Example 2

(36) Again, experiments were carried out without a membrane, in order to demonstrate the electrochemistry. A 1.6 mm diameter Pt macrodisk working electrode, a reference electrode of 0.5 mm diameter silver wire, and a Pt coil counter electrode were placed in a glass vial and covered with about 1.5 ml of ionic liquid (same liquid as previous example) open to the atmosphere. Hydrogen gas or mixtures of hydrogen and nitrogen were bubbled directly into the ionic liquid for about 15 minutes to ensure complete equilibration. Cyclic and square wave voltammetry were then carried out as in the previous example.

(37) Cyclic voltammetry results are shown in FIG. 7. A broader oxidation peak than in FIG. 4 was observed, with a reduction peak that was separated from the oxidation peak by ca. 0.5 volt at 100% hydrogen, suggesting more electrochemical irreversibility on the larger electrode. Again the peak current is proportional to hydrogen concentration as shown by FIG. 8 which is a plot of current against hydrogen concentration. The square wave voltammetry results in FIG. 9 show a large peak at approx. +0.3 volt (vs. Ag) which increased progressively with increasing concentrations of hydrogen.

(38) The above procedure was repeated using the same ionic liquid which had been saturated with water. The saturation quantity of water in this ionic liquid was under 0.5% by weight. However, no significant differences in the current responses were observed.

Wellbore Tool

(39) One application of the hydrogen sensor is in a tool which is temporarily placed in a wellbore. The hydrogen sensor can then be operated to determine the concentration of hydrogen in wellbore fluid.

(40) FIG. 10 is a schematic view of a tool 40 temporarily lowered into a wellbore 42 by a wireline cable 44. The tool incorporates a pump 46 which draws in fluid from outside the tool through inlet 47 along a flowline 48 to an outlet 50. The tool 40 may also have means (not shown) to collect samples of the wellbore fluid and store them under pressure until the tool is returned to the surface.

(41) A hydrogen sensor 52 constructed as in FIGS. 1 to 3 is located on the flowline 48 so that the hydrogen-permeable window 20 of the sensor is exposed to the fluid in the flowline 48. FIG. 11 shows this in more detail.

(42) The sensor 52 and the pump 46 are connected to an electronics package 54 which incorporates a potentiostat and a controlling computer which operates the potentiostat to make measurements and which records the results. The electronics package may communicate results to the surface via the wireline cable 44.

(43) For use, when the tool 40 is at a chosen position in the wellbore, the pump 44 is used to draw fluid along the flow line 48 hydrogen passes in both directions through the permeable window of the sensor 52 and hydrogen concentration in the ionic liquid in the sensor equilibrates with hydrogen concentration in the fluid in the flowline 48. The sensor 52 is operated to determine peak current flow during a voltammetric sweep and this is used to determine the hydrogen concentration in the wellbore fluid around the tool 40.

(44) Pressure inside the sensor 52 is matched to pressure in the flow line 48 by means of a communicating passage 56 which includes a floating piston 58 or other physical barrier which prevents mixing of liquids but allows the hydraulic pressure in the flow line 48 to be transmitted to the ionic liquid in the sensor 52.

(45) It will be appreciated that the diffusivity of hydrogen in the palladium membrane of the window 20 (and hence the rate at which hydrogen passes through the membrane) does not affect the equilibrium measurement of hydrogen concentration. Any changes in the transport properties of the membrane caused by changes in temperature, or by physical fouling (for example accumulation of solid particles on the hydrogen-permeable window) or even reaction of some of the palladium metal with hydrogen sulfide or other sulfur-bearing compounds in the borehole fluids do not change the equilibrium response of the sensor, although the time taken to reach equilibrium may be increased. Equilibrium hydrogen concentration in the ionic liquid will be observed as constant oxidation current in a succession of measurements.

(46) If desired, the tool 40 may include an electrical heating coil (not shown) around the sensor body. This may be used to raise the temperature of the sensor to a constant temperature which is above the temperature in the wellbore. This will reduce the viscosity of the ionic liquid, thereby shortening the time to reach an equilibrium hydrogen concentration, and will enable the measurement of hydrogen concentration to be made at constant temperature, which can simplify the calibration of the sensor.

(47) The tool 40 may include other sensors for examining fluid in the flowline 48. In particular it may contain a sensor for hydrogen sulphide and thiols. Measurement of the concentration of hydrogen in wellbore fluids may be augmented by a measurement of the concentration of hydrogen sulfide if it is suspected that the hydrogen results from corrosion of coupon s by hydrogen sulphide. One example of a suitable wellbore hydrogen sulfide sensor has been disclosed in U.S. Pat. No. 6,939,717 and U.K. Patent No. GB 2,397,651.

Example 3

(48) To exemplify use in a downhole tool, a sensor constructed as in FIGS. 1 to 3 was used to determine the hydrogen concentration in a sample of a light oil produced from a permeable formation at a temperature of 130° C. and a total pressure of 650 bar (65 MPa). The ionic liquid used in the sensor was 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide, commonly abbreviated to [hmim][NTf.sub.2]. The sensor was placed in contact with the oil at a temperature of 130° C. and a pressure of 650 bar.

(49) Cyclic voltammetry was used to determine the concentration of hydrogen dissolved in the ionic liquid. Measurements were repeated until the peak current from the cyclic voltammetry measurement reached a constant value. This value of peak current corresponded to a dissolved hydrogen concentration of 0.0557 moles per kilogram of ionic liquid. Using data presented by Kumelan et al. in Journal of Chemical & Engineering Data, 2006, vol. 51, p. 1364, this hydrogen concentration of 0.0557 moles per kilogram of ionic liquid at a temperature of 130° C. was calculated to correspond to a partial pressure of hydrogen gas of 30 bar. Using extrapolated data presented by Gomes de Azevedo et al. in Journal of Chemical Thermodynamics, 2005, vol. 37, pp. 888-899, the density of [hmim][NTf2] at a temperature of 130° C. and a total pressure of 650 bar was taken to be 1390.2 grams per liter, and therefore the mole fraction of hydrogen dissolved in the ionic liquid is 0.0243. For a typical light oil with a reported mean molar mass of 250 g/mole, a partial pressure of hydrogen of 30 bar has been reported in the same paper by Gomes de Azevedo et al. to correspond to a dissolved hydrogen concentration of 0.1176 moles per kilogram of oil at a temperature of 130° C.

(50) A concentration of 0.1176 moles per kilogram of oil corresponds to a hydrogen content of 235 mg per kilogram or 235 ppm. This concentration can be used to assess the impact on the metals used in the completion and tubing (e.g., stress corrosion cracking) and on other components, such as optical fibers used for temperature measurements and telemetry when the oil is produced.

Corrosion Monitoring

(51) The generation of hydrogen is a common feature of corrosion reactions involving metals and the hydrogen so generated can further damage some metals by causing stress cracking. The rate of hydrogen production by corrosion reactions and the concentration of hydrogen dissolved in metals are valuable parameters in determining corrosion rates and the potential for the failure of metal components by hydrogen-induced stress cracking.

(52) FIG. 12 shows a sensor used to monitor corrosion of a test piece, a so-called coupon. The sensor is generally as shown in FIGS. 1 to 3, but the hydrogen-permeable window 20 has been replaced by a disc-shaped coupon 60 of metal to be tested for corrosion. A palladium membrane 62 is brazed to the coupon 60. In this example the metal coupon was duplex stainless steel 22Cr, which contains 22 weight percent chromium, 13 weight percent nickel, 5 weight percent manganese and 3 weight percent molybdenum. The metal coupon 60 was 4 cm in diameter and 4 mm in thickness and was brazed to a palladium disk 62 of 100 μm thickness. The ionic liquid in the electrochemical cell was 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([hmim][NTf2]).

(53) The sensor was placed in the completion of a producing well together with an electronics package 64 to carry out cyclic voltammetry measurements at regular intervals and transmit the results to the surface. The well was producing both oil and water and the average values of the hydrostatic pressure and temperature of the fluids in the vicinity of the completion were 490 bar and 118° C., respectively.

(54) The hydrogen sensor initially detected a concentration of 0.0021 moles of hydrogen per kilogram of ionic liquid. The metal coupon was exposed to wellbore fluids and corrosion led to the production of hydrogen, some of which dissolved in the coupon and migrated through the coupon and the palladium membrane into the ionic liquid.

(55) An issue with any method of measuring the hydrogen generated by a surface (corrosion) reaction is where the measurement is made and what fraction of the hydrogen generated reaches the measurement transducer. In the particular geometry presented in this example, the hydrogen is measured after dissolution in the coupon and diffusion through it. Some fraction of the hydrogen generated by the corrosion reaction can be expected to dissolve in the metal of the coupon while the remainder will dissolve in the fluid in contact with the coupon at the exterior of the sensor and be swept away. It may not be possible to assume equilibrium distribution of hydrogen between the water and the metal but experiments can be done to determine the partition of hydrogen between the metal and water as a function of the Reynolds number of the flow of the water.

(56) Over a period of several months of operation the hydrogen sensor with attached stainless steel coupon showed a gradual increase in the hydrogen concentration in the ionic liquid from 0.0021 to 0.0028 moles of hydrogen per kilogram of ionic liquid. This corresponded to an increase in the partial pressure of hydrogen in the sensor from 1.20 bar to 1.58 bar. This showed that there was also an increase in concentrations of hydrogen dissolved in the stainless steel coupon and the water adjacent to it. The increased hydrogen contents of the metal coupon and the water in contact with the coupon are indicative of a corrosion reaction at the surface of the steel coupon and that the metals of similar composition in the well are likely to be undergoing similar corrosion reactions.

(57) If it suspected that hydrogen is being generated at other locations in the well to such an extent that there is an increasing hydrogen concentration in the produced fluids at the exterior of the sensor, then a second hydrogen sensor can be deployed in the wellbore, e.g., on a production logging tool, to monitor the concentration of hydrogen in this fluid.

(a) Monitoring the Performance of Wellbore and Subsea Optical Fibre Systems

(58) The glass used in optical fibres deployed in wellbores to measure temperature can interact with hydrogen (see for instance Stone J in Journal of Lightwave Technology, 1987, vol. 5 p. 712) and reduce the transmissivity of the glass in the near-infrared and visible spectral regions. Interactions which have been identified are the dissolution of hydrogen in the glass and the reaction of hydrogen with certain Si—O or Ge—O defects to form Si—OH and Ge—OH groups. The latter are a permanent change. Both the dissolved hydrogen and the Si—OH and Ge—OH groups absorb radiation in the near-infrared spectral region and hence reduce transmissivity.

(59) A valuable component of predicting the performance of an optical fibre system in a wellbore or subsea environment is to measure the partial pressure (or concentration) of hydrogen, irrespective of its origin.

(60) FIG. 13 shows a schematic of a small hydrogen sensor 70 fitted to a metal tube 72 containing an optical fibre 74. The fibre 74 is deployed in a producing oil well and is used as a temperature sensor connected to a temperature detection system located on the surface to determine the temperature profile of the well. The optical fibre is silica (diameter 125 μm) with a polyamide coating (125 μm thickness) and the tube 72 is a stainless steel tube (6.25 mm external diameter, 3.00 mm internal diameter) containing a silicone oil. Fitting the sensor 70 to the tube 72 can be done by placing the sensor into a modified T-piece fitting, such as the well-known Swagelok fittings used for high pressure tubing connections: details of this are not shown. The electrochemical sensor is generally of the construction shown in FIGS. 1 to 3. The ionic liquid in the interior chamber of the sensor is exposed to the silicone oil in the tube containing the optical fibre by means of a hydrogen-permeable window 20 which is a palladium membrane on an inert, porous support. The surface of the palladium membrane in contact with the silicone oil may be coated with a protective polymer layer 25 as in FIG. 3.

Example 5

(61) The hydrogen sensor 70 was located at a position in the well where the ambient temperature was 182° C. as determined by the fibre optic distributed temperature sensor. The ionic liquid in the hydrogen sensor was 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([hmim][NTf2]). The working and counter electrodes were platinum; a silver pseudo reference electrode was used. At this temperature the sensor measured a hydrogen concentration of 0.0050 moles of hydrogen per kilogram of ionic liquid (10.0 ppm), which corresponds to a partial pressure of hydrogen gas of 2.30 bar. Using data on the solubility of hydrogen gas in silicone oil from Berezkin et al., Russian Chemical Bulletin, 1999, vol. 48, pp. 914-916, the corresponding concentration of hydrogen gas dissolved in the silicone oil was calculated to be 0.0178 moles of hydrogen per liter of silicone oil or 42 ppm and using data from Lee et al., Journal of Chemical Physics, 1962, vol. 36, pp. 1062-1071, the concentration of hydrogen dissolved in the glass was estimated to be 4.95×10.sup.−7 moles of hydrogen per cm.sup.3 of glass or 0.37 ppm.

Chemical Processes

(62) There are a number of chemical processes which consume or generate hydrogen or in which hydrogen is an intermediate product so that the process both generates and consumes hydrogen. A sensor as described above may be incorporated in a process plant carrying out such a process.

(63) Examples of such processes include processes for converting natural gas, which usually is predominantly methane, into liquid fuels and chemical products via the formation of synthesis gas (a mixture of hydrogen and carbon monoxide). The synthesis gas (frequently termed syngas) is generated within the gas-to-liquid process by reforming or by the partial oxidation of natural gas (predominantly methane). For example, the partial oxidation of methane (CH.sub.4) to produce syngas can be described by:
CH.sub.4+0.50.sub.2.fwdarw.2H.sub.2+CO

(64) The syngas so formed can be further reacted to generate useful chemical products. One such reaction is the Fischer-Tropsch reaction, which can be described in simplified form by the conversion of syngas to alkanes and water
2nH.sub.2+nCO.fwdarw.C.sub.nH.sub.2n+nH.sub.2O
where C.sub.nH.sub.2n represents the mixture of alkanes formed by the Fischer-Tropsch reaction. A critical aspect of the Fischer-Tropsch reaction is the H.sub.2/CO ratio and the effects of variation of this ratio on the products from the Fischer-Tropsch reaction are well known (see for instance Bartholomew and Farrauto “Fundamentals of Industrial Catalytic Processes”, 2nd Edition, 2006, Wiley-Interscience, pp. 412-415). A second reaction involving the conversion of syngas to chemical products is the well-known methanol synthesis reaction discussed in the same book, at pp. 382-398.
CO+2H.sub.2.fwdarw.CH.sub.3OH

(65) Methanol can also be synthesised using a mixture of hydrogen and carbon dioxide (CO.sub.2)
CO.sub.2+3H.sub.2.fwdarw.CH.sub.3OH+H.sub.2O
noting that water is also produced by the reaction. The methanol synthesis reaction is therefore one method of using natural gas containing significant concentrations of CO.sub.2 which may otherwise have little or no economic value.

(66) Chemical processes which create and/or consume hydrogen may be carried out in large-scale fixed industrial plant. The sensors disclosed herein may be used in such process plant. However, it is also known to carry out such chemical processes on a smaller industrial scale and the process plant may then be dimensioned and constructed to be transportable by road vehicle so that the plant can be taken to a natural gas source and taken elsewhere if the flow of natural gas diminishes.

(67) While large scale chemical process plants are able to absorb the cost of expensive process analytical equipment, such as gas chromatographs and mass spectrometers, and skilled operators to monitor the concentration of hydrogen, the economic viability of small scale systems may make it very desirable that the process plant can be manufactured and operated economically, with minimal (or even no) operator personnel and low cost components. The sensors disclosed herein may be advantageous in smaller scale process plant utilised at sites where analytical laboratories and/or trained laboratory staff are not available.

(68) Process plant for conversion of natural gas to another product and/or for upgrading the natural gas stream by converting part of the gas stream of another product, is disclosed in U.S. Patent Publication Nos. US2008/0262110, US2009/0299795, and US2010/0000153. A process plant dimensioned to be transportable by road vehicle to a site of use, such as a wellbore penetrating a small natural gas reservoir, may be configured to have length no greater than 17 meters possibly no greater than 13 meters and width no greater than 3 meters. As suggested by the documents above, the length and width may fit within the footprint of a shipping container approximately 8 feet (2.5 meters) wide and commonly up to 40 feet (12.2 meters) long, although container lengths up to 53 feet (16.2 meters) are in use. Such a transportable unit may be configured to have a vertical height during transport of not more than 8, 9.5 or 10.5 feet (2.5, 2.9 or 3.2 meters) and may be constructed with a total weight not exceeding 30,000 kg and possibly not exceeding 20,000 kg.

(69) FIG. 14 is a block diagram which illustrates use of a hydrogen sensor, of the type shown in FIGS. 1 to 3, in a small-scale gas-to-liquid unit converting natural gas to a mixture of alkanes. The sensor is used to determine the concentration of hydrogen gas being generated by a reformer and entering a Fischer-Tropsch reactor. The principal features of this gas-to-liquid unit are a gas pretreatment unit 80, a syngas generator 81, a heat exchanger 82 to lower the temperature of the syngas, a compressor 83 to raise the pressure of the syngas and a Fischer-Tropsch reactor 84. Natural gas enters at 85 and the products exit the Fischer-Tropsch reactor at 87. Other details of the unit are omitted for clarity. The reactor 84 is designed for low temperature Fischer-Tropsch synthesis using a cobalt metal catalyst. Further design parameters are that the inlet temperature and pressure of the syngas are maintained at values of 210° C. and 28.9 bar, respectively.

(70) A hydrogen sensor 86 is located on the syngas flow line immediately before it enters the Fischer-Tropsch reactor 84, i.e., after the syngas has been cooled and compressed. The sensor 86 has the construction shown by FIGS. 1 to 3 except for a different protective layer 25 over the palladium membrane 24. The palladium membrane 24 is highly selective to the transport of hydrogen but its surface may become poisoned by the high concentration of CO in the syngas. The palladium membrane can be protected from contamination by carbon monoxide and other components in the syngas, such as residual water vapour, by the use of a protective layer 25 of silica, possibly 10-100 nm in thickness, deposited on the palladium using chemical vapour deposition. For example, Lee et al., Journal of Membrane Science, 2004, vol. 213, pp. 117-126 have demonstrated that thin films of silica (20-30 nm) deposited on alumina can exhibit a highly selective permeance to hydrogen relative to other gases such as CO and CO.sub.2 at temperatures up to 600° C. Alternatively, a thin barrier coating of silicon carbide can be applied to the palladium membrane using a plasma-enhanced chemical vapour deposition technique, as for example described by Zambov et al, Journal of Vacuum Science & Technology A, 2006, vol. 24, pp. 1706-1713.

(71) The ionic liquid used in the sensor 86 is 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([hmim][NTf.sub.2]) and the working and counter electrodes are composed of platinum, while the pseudo reference electrode is composed of silver. The hydrogen sensor is small, which enables equilibrium between hydrogen in the syngas stream and hydrogen dissolved in the ionic liquid to be established quickly for the determination of the hydrogen concentration. Under the operating conditions the electrochemical measurement determined the dissolved hydrogen concentration to be 0.0453 moles of hydrogen per kilogram of ionic liquid, which equated to 90.6 ppm. The dissolved hydrogen concentration is obtained with a partial pressure of hydrogen in the gas stream of 19.5 bar, which gives a hydrogen/carbon monoxide ratio of 2.07, assuming that syngas behaves as an ideal gas under the operating temperature and pressure conditions.

(72) The concentration of carbon monoxide in the syngas stream can, if desired, be independently monitored using a small infrared gas sensor similar to those described in U.S. Pat. No. 6,995,360, but operating over the wavelength range 4.44-5.00 μm. A second electrochemical hydrogen sensor 88 and a second optical carbon monoxide sensor may if desired be fitted to the flowline 87 downstream of the Fischer-Tropsch reactor 84 to determine the concentrations of unreacted hydrogen and carbon monoxide and hence monitor the efficiency of the process.

(b) Monitoring the Concentration of Hydrogen in a Portable Methanol Synthesis Reactor

(73) FIG. 15 is a block diagram of a small-scale process plant for conversion of natural gas, via syngas to methanol. Parts of the plant correspond to those in FIG. 14 and are shown with the same reference numerals. A reactor 90 converts the syngas to methanol which flows out at 94 This conversion differs from the Fischer-Tropsch process of FIG. 14 in several respects: (i) methanol synthesis is normally carried out at higher total pressures than Fischer-Tropsch synthesis, commonly in the pressure range 50-100 bar; (ii) the conversion of syngas to methanol is low, such as less than 15% conversion in a single reactor stage and therefore after the gas stream leaves leaving the reactor 90 the methanol product is condensed by cooling and separated to flow out at 94 while the unreacted gas mixture, which contains hydrogen, carbon monoxide and carbon dioxide is recycled along line 92 to pass through the reactor again (an alternative is to use a sequence of reactors); (iii) any CO.sub.2 present in the gas stream either as a result of its introduction into the syngas or its occurrence in the original natural gas, is converted into methanol.

(74) Hydrogen sensors 86, 88 are used to monitor hydrogen concentrations at the inlet and outlet to reactor 90. The hydrogen sensors may if desired be accompanied by high temperature/high pressure CO and CO.sub.2 sensors, as for example described in U.S. Pat. No. 6,995,360. The CO.sub.2 is detected in the spectral range 4.17-4.35 μm while CO is detected in the spectral range 4.55-4.76 μm.

(75) The ionic liquid used in the sensors 86, 88 is 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide and the electrode materials are platinum for the working and counter electrodes and silver for the reference electrode. The hydrogen-permeable window separating the ionic liquid from the gas streams has a thin palladium membrane on a porous ceramic support. The membrane can be coated with a suitable barrier material, such as silica or silicon carbide, to protect the palladium membrane from gases such as CO and H.sub.2O.

(76) Design parameters for the methanol synthesis reactor shown in FIG. 15 are an inlet temperature of 225° C., an inlet pressure of 55 bar and an internal operating temperature of the reactor of 244° C. The hydrogen/carbon ratio (H/C) in the syngas stream is commonly represented by the ratio H.sub.2/(2CO+3CO.sub.2), which should be in the range 1.00-1.05 for efficient conversion.

(77) A set of further expected parameters was calculated using literature data. The hydrogen sensor 86 on the inlet flow line of the methanol reactor 90 measures a concentration of 0.093 moles of hydrogen per kilogram of ionic liquid (186 ppm), which corresponds to a partial pressure of hydrogen gas of 38.5 bar at 225° C. The infrared gas detector measures the concentrations of CO and CO.sub.2 to be 0.304 and 0.093 moles/liter, respectively, which correspond to partial pressures of 12.6 and 3.9 bar, assuming that the gases behave as pure components. If it is assumed that hydrogen in syngas behaves as a pure component, then its concentration is 0.92 moles/liter and therefore the H/C ratio is 1.04.

(78) The temperature and pressure of the gas exiting the methanol reactor are 244° C. and 45.5 bar, respectively, with 3.4 bar of the decrease in pressure due to frictional losses. The hydrogen sensor in the exit gas stream at 244° C. measures a concentration of 0.080 moles of hydrogen per kilogram of ionic liquid (160 ppm), which corresponds to a partial pressure of hydrogen gas of 31.7 bar. Assuming the hydrogen gas behaves as a pure component, its concentration in the exit gas stream is 0.73 moles/liter. The infrared gas sensor measures exit concentrations of CO and CO.sub.2 of 0.266 and 0.055 moles/liter, respectively, which, assuming the gases behave as pure components, correspond to partial pressures of 11.45 and 2.35 bar. The known temperature and pressure changes associated with the methanol reactor allow the outlet concentrations of hydrogen, CO and CO.sub.2 to be calculated at inlet conditions. Under these conditions, the changes in the concentrations of CO and CO.sub.2 are 5.9% and 36.5%, respectively, giving a total carbon conversion of 13.1% of the total inlet carbon concentration. The corresponding decrease in the hydrogen concentration on passing through the methanol reactor is 15.0%.

(79) The gas hourly space velocity (GHSV) in the reactor under standard temperature and pressure conditions is 8000 hr.sup.−1. Computing from this gives a rate of methanol synthesis of 0.51 kg per liter of catalyst per hour and a rate of water production of 0.19 kg per liter of catalyst per hour at the operating conditions of the reactor.

(c) Monitoring the Hydrogen Consumption in a Portable Hydrocarbon Cracking Reactor

(80) The low temperature Fischer-Tropsch synthesis of hydrocarbons gives rise to a wide distribution of chain lengths such that the product contains hydrocarbons which are both liquid and solid hydrocarbons at ambient temperature. The solid hydrocarbon, commonly termed wax, may contain linear hydrocarbons in the range C.sub.22 to C.sub.70 and this product frequently accounts for approximately two thirds of the hydrocarbons generated in Fischer-Tropsch synthesis. The wax may be converted to liquid products, such as diesel fuels, by hydrocracking: see for instance Leckel in Energy & Fuels, 2009, vol. 23, pp. 2342-2358.

(81) The cracking reaction results in a decrease in the average hydrocarbon number of the alkane chain and an increase in the average hydrogen:carbon ratio. A measurement of the rate of consumption of hydrogen can therefore be used in predicting the average carbon number of the hydrocarbon product of the cracking reaction.

(82) FIG. 16 is a block diagram of process plant used to hydrocrack Fischer-Tropsch wax. Molten wax and hydrogen are supplied along respective lines 101 and 102 to a reactor vessel 104. The molten wax (average hydrocarbon chain length n.sub.i) and hydrogen gas enter the reactor vessel 104 at mass flow rates of M′ and M′.sub.H, respectively, where they are mixed in an inert packed bed 105 before being passed into a catalytic bed 106 of platinum metal deposited on an amorphous silica-alumina (SiO.sub.2—Al.sub.2O.sub.3) substrate. The cracked hydrocarbon (average hydrocarbon chain length n.sub.f), exits the reactor 104 through a heat exchanger 108. Choke 110 is used to control the pressure drop across the catalytic reactor. The cracked hydrocarbon flowing from choke 110 contains liquid hydrocarbon with hydrocarbon gases and unreacted hydrogen. Liquid product is separated from gas and is removed along outlet line 112 while the gases flow along line 116. The pressure at which the liquid and gas are separated is controlled by choke 110 and the pressure at which the gas passes into line 116 is controlled by a further choke 114. The uncondensed hydrocarbons and unreacted hydrogen from this process may be used as fuel for the burner of the steam reformer 84 in FIG. 14 that generates the syngas for the Fischer-Tropsch synthesis.

(83) Hydrogen sensors 118, 119 of the type shown in FIGS. 1 to 3 are fitted to the hydrogen inlet line 102 and the outlet gas line 116.

(84) Design parameters for the hydrocracking reactor are: mass flow rate of wax M′=266 kg/hour, average initial wax hydrocarbon number n.sub.i=18.02, inlet hydrogen gas pressure=32 bar, reactor inlet temperature=96° C., inlet mass flow rate of hydrogen M′.sub.H=2.66 kg/hour, operating temperature of hydrocracking reactor=340° C., weight hourly space velocity (WHSV)=2.6 kg wax per hour per kg of catalyst, heat exchanger outlet temperature=68° C., and outlet gas line pressure=3 bar

(85) A low hydrogen/wax mass flow ratio (M′.sub.H/M′=0.01) is used to minimize the use of hydrogen since the unreacted hydrogen is not separated from the produced hydrocarbon vapour for recycling into the hydrocracking reactor. The hydrogen gas sensors 118, 119 are used to measure the concentration of hydrogen in the gas streams in the inlet and outlet lines of the hydrocracking reactor. The ionic liquid used in the gas sensor is 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([hmim][NTf.sub.2]) and the working and counter electrodes are composed of platinum while the reference electrode is composed of silver. The electrochemical cell of the sensor contacts the hydrogen in the analyte by means of a palladium membrane that is coated with a suitable material, such as silica, to minimize contamination by gases such as carbon monoxide and hydrogen sulfide.

(86) Using literature data and the above design parameters, further expected values were calculated. The input hydrogen sensor 118 measures a hydrogen concentration of 0.0516 moles of hydrogen per kilogram of ionic liquid (103.2 ppm), which corresponds to an equilibrium partial pressure of hydrogen gas of 31.1 bar. This pressure is close to the input gas pressure of 32 bar. Using a sensor to determine the input hydrogen partial pressure (concentration) facilitates using a hydrogen-containing gas mixture if pure hydrogen is not available or too costly. The output hydrogen gas sensor 119 measures a concentration of 0.0019 moles of hydrogen per kilogram of ionic liquid (3.8 ppm), which corresponds to an equilibrium partial pressure of hydrogen gas of 1.29 bar. The combined partial pressure of the low molecular weight hydrocarbons (primarily propane to octane) is approximately 1.7 bar. Assuming that the hydrogen gas behaves as a pure component, the concentration of hydrogen is 0.0454 moles per liter, which compares to an input gas concentration of 0.998 moles per liter. From the flow rate of gas on the outlet gas line the mass flow rate of hydrogen gas is 1.78 kg/hour and therefore the rate of consumption of hydrogen M′.sub.HC is 0.88 kg/hour, which is approximately 33% consumption. The ratio M′.sub.HC/M′ is equal to 0.0033 and using the equation:

(87) $\frac{{\stackrel{\_}{n}}_f}{{\stackrel{\_}{n}}_i}=\frac{M^{\prime }}{M^{\prime }+\left(7{\stackrel{\_}{n}}_i+1\right)M_{HC}^{\prime }}$
the ratio n.sub.f/n.sub.i is calculated to be 0.705 giving n.sub.f=12.71. The hydrocracking of the Fischer-Tropsch wax therefore reduces its average carbon number from 18.02 to 12.71.

(d) Monitoring In Situ Hydrocarbon Upgrading Processes

(88) FIG. 17 shows a schematic of a hypothetical in situ upgrading process for heavy oil being produced from a well. The well has been completed with casing 120, the casing has been perforated as indicated at 122 and the oil-bearing formation around the well has been hydraulically fractured. The resultant fractures are filled with proppant. As indicated by heavy arrows 123, the heavy oil flows through the perforations 122 into a porous bed 124 of catalyst particles downhole within the well. The catalyst bed 124 is enclosed by a metal mesh 126 which is shaped so that the lower part of the bed surrounds a conical cavity 128.

(89) The catalyst bed is heated to a temperature in the range 300-400° C. This may be done with electrical heating powered through cable from the surface or with a liquid, such as silicone oil, that is heated at surface and circulated down to the catalyst bed 124 and back again through pipes in the well. Prior to perforation, the portion of casing which will surround the catalyst bed is provided with a layer of a thermally insulating material on its interior, so that there is thermal insulation between the mesh and the casing. The temperature of the catalyst bed is monitored and maintained by an array of thermocouples used to control the electrical heating or the supply of heated silicone oil.

(90) Natural gas, piped from a natural gas source elsewhere, is consumed in a process for upgrading the heavy oil. Part of this natural gas is burnt to heat silicone oil or generate electricity. The heated oil or electricity is used to heat the catalyst bed. Another part of this natural gas is used as fuel to generate steam. The exhaust gases from burning natural gas are cooled to condense out water, so that the exhaust gas is a mixture of carbon dioxide and nitrogen.

(91) A further part of this natural gas is used as feed for a process plant that generates synthesis gas by reacting the natural gas, which is largely methane, with steam in the presence of a catalyst (e.g., nickel supported on a mixed silica-alumina substrate) at a temperature in the range 850-950° C.:
CH.sub.4+H.sub.2O.fwdarw.3H.sub.2+CO

(92) The synthesis gas so formed is further reacted with steam in the presence of a second catalyst (e.g., a mixture of iron(III) and chromium(III) oxides) to drive the so-called water gas shift reaction:
CO+H.sub.2O.fwdarw.CO.sub.2+H.sub.2

(93) The resulting hydrogen and carbon dioxide flowing along line 130 are mixed with the nitrogen and carbon dioxide exhaust gases flowing along line 132 and piped to a compressor 134 and thence pumped down to the cavity 128 from which they enter the heated catalyst bed 124. The hydrogen reacts with the heavy oil in the presence of the catalysts in the bed 124 to remove sulfur from the oil and to reduce the molecular weight of the oil's constituent hydrocarbons. The carbon dioxide can dissolve in the heavy oil and reduce its viscosity while the nitrogen aids in lifting the produced hydrocarbon to surface. The catalysts may be molybdenum sulfide for removal of sulfur by hydrodesulfurisation reaction with hydrogen to form hydrogen sulfide together with a solid acid catalyst, such as a mixture of amorphous silica and alumina and a lanthanide metal-exchanged zeolite, for reaction of hydrogen with the heavy oil. When required, the catalyst bed 124 in the containing mesh 126 can be removed to surface for catalyst regeneration or replacement.

(94) The produced oil and gases are separated at surface in a separator 136. The liquid is piped along line 137 and the gases are piped along line 138 which leads to a membrane separation unit (not shown) where unreacted hydrogen is extracted for recycling or combustion. The pressure in the flow line 138 is controlled by a choke 148.

(95) Hydrogen sensors 142, 144 of the type shown in FIGS. 1 to 3 are provided on the hydrogen inlet line 130 and the gas outlet line 138 in order to monitor the efficiency of the upgrading process. Further, a hydrogen sulfide sensor 146 on the outlet gas outlet line 138 can be used to determine the concentration of hydrogen sulfide generated by the hydrodesulfurisation process. The hydrogen sulfide sensor may be as described in U.S. Pat. No. 6,939,717.

(96) Design parameters of the heavy oil production and upgrading process are: pre-compressor gas pressure=2.1 bar, pre-compressor gas temperature=153° C., inlet (post-compressor) gas pressure=158 bar, mean formation depth=1380 m, API gravity of untreated heavy oil=15, density of oil=1.035 tonnes/m.sup.3, production rate of oil=700 barrel of oil/day=4800 kg/hour, initial sulfur content of heavy oil=1.70 weight percent, post-separator flow line gas pressure=17.6 bar, and post-separator temperature=102° C.

(97) A number of expected values were computed from the design parameters and literature data. The input hydrogen sensor located before the compressor measures a hydrogen concentration of 3.3×10.sup.−3 moles of hydrogen per kilogram of ionic liquid (0.67 ppm), which corresponds to a partial pressure of 0.168 bar at a temperature of 153° C. The remaining pressure of the gas is due to carbon dioxide and nitrogen from the combustion and the reforming/water gas shift reactions. Assuming the hydrogen in the gas mixture behaves as a pure component, the concentration of hydrogen in the pre-compressed gas is 0.0047 moles per liter of gas. The mass flow rate of hydrogen into the catalytic bed 124 in the well is 96.0 kg per hour.

(98) The output hydrogen sensor 144 located after the oil-gas separator measures a hydrogen concentration of 0.0020 moles of hydrogen per kilogram of ionic liquid (4.0 ppm), which corresponds to a partial pressure of hydrogen of 1.20 bar at a temperature of 102° C. Again, assuming the hydrogen in the gas mixture behaves as a pure component, the concentration of hydrogen in the produced gas beyond the separator is 0.0169 moles per liter of gas. Calculating from the flow rate of gas in the gas line 138 beyond the oil-gas separator, the mass flow rate of hydrogen exiting the well is 80.69 kg per hour and therefore the total consumption of hydrogen in the cracking and hydrodesulfurisation reactor is 15.31 kg per hour.

(99) The hydrogen sulfide sensor measures a hydrogen sulfide concentration of 1.11×10.sup.−3 moles per liter of gas (38 ppm w/v) at a temperature of 102° C. The mass flow rate of hydrogen sulfide from the well is 39.88 kg per hour and therefore the mass flow rate of sulfur released from the heavy oil by hydrodesulfurisation is 37.53 kg per hour. Calculating from the mass flow rate of heavy oil and its sulfur content yield a total mass flow rate of sulfur of 81.60 kg per hour and therefore 46.0% of the sulfur in the heavy oil is removed as hydrogen sulfide. The mass rate of hydrogen consumption of the hydrodesulfurisation process is 2.35 kg per hour and therefore the mass rate of hydrogen consumption by the heavy oil cracking treatment is 12.96 kg per hour. The ratio M′.sub.HC/M′, as defined in the previous section, is 0.0027.

Detection of a Hydrogen Leak in a Proton Exchange Membrane Fuel Cell

(100) Another application of the hydrogen sensors disclosed here is in the context of fuel cells which generate electricity from an electrochemical reaction of hydrogen and oxygen. A hydrogen sensor may be used to detect hydrogen leakage.

(101) There has been prolonged interest in the use of hydrogen in fuel cells, particularly polymer electrolyte membrane or proton exchange membrane (PEM) fuel cells. The design and operation of PEM fuel cells is well known. A detailed description is provided in Barbir, F., PEM Fuel Cells: Theory and Practice, 2005, Elsevier Academic Press, Amsterdam.

(102) At the fuel cell anode, hydrogen is ionised:
H.sub.2.fwdarw.2H.sup.++2e.sup.−
and the protons diffuse through a polymer membrane, commonly an ionomer such as Nafion which is a copolymer of tetrafluoroethylene and sulfonated tetrafluoroethylene, to the cathode where they react with oxygen:
2H.sup.++0.5O.sub.2+2e.sup.−.fwdarw.H.sub.2O

(103) The use of hydrogen, even at pressures close to atmospheric, in fuel cells raises potential safety issues from leaks either into the atmosphere or into the cathode region. The latter occurrence may be particularly hazardous since the cathode contains both oxygen and a catalyst, commonly high surface area platinum metal, that may facilitate an explosion.

(104) A significant source of hydrogen leaks in polymer electrolyte membrane (PEM) fuel cells is the polymer membrane itself. The leakage of hydrogen gas through the polymer membrane is termed crossover and as reported by Wu et al. in Journal of Power Sources, 2008, 184, pp. 104-119 can reduce the efficiency of the fuel cell and also degrade the mechanical properties of the membrane. Inaba et al. in Electrochimica Acta, 2006, 51, pp. 5746-5753 have reported that hydrogen crossover can cause excessive heating of the cathode membrane assembly and the membrane and moreover the formation of hydroxyl radicals can cause chemical attack on the membrane and the formation of pinholes. The formation of such pinholes can result in even greater hydrogen crossover with the result that potentially explosive levels of hydrogen can leak into the cathode. Monitoring the cathode region for hydrogen which has leaked into it is made difficult by the presence of larger amounts of oxygen and other materials. A sensor of the kind disclosed here is useful because it provides selective entry of hydrogen into the ionic liquid.

(105) FIG. 18 shows a schematic of a PEM fuel cell. The cell has an anode 150 and cathode 152 separated by a polymer membrane 154. A gas diffusion layer 156 is provided on the surfaces of the anode and cathode exposed to gas in the anode chamber and cathode chamber respectively. A gas diffusion layer is an electrode support layer, which may be composed of carbon to ensure good thermal and electrical conductivity, which gives mechanical support to the membrane and an electrode and enables the hydrogen and oxygen to reach the electrode in a uniform manner. A gas stream containing hydrogen is supplied along input line 162 to the anode chamber 160 and this gas stream with depleted hydrogen content leaves the anode chamber at 164. A gas stream containing oxygen is supplied along line 168 to the cathode chamber 166 and leaves with depleted oxygen content at 169.

(106) A small hydrogen sensor 170 of the type shown in FIGS. 1-3 is located in the cathode chamber 166 of the fuel cell close to the membrane surface and close to the exit 169 of the cathode gas stream in order to detect the rate of hydrogen permeation across the entire membrane. The diameter of the hydrogen sensor may possibly be less than 5 mm. The body of the sensor consists of a material that is electrically insulating, such as ceramic or polymer, and is chemically stable in the presence of oxygen and water vapour at temperatures up to 120° C. The palladium membrane of the sensor 170 is 50 μm in thickness and coated with a thin (200 nm) coating of silica. The ionic liquid used in the electrochemical cell is 1-hexyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([hmim][NTf.sub.2]) and the working electrode may be a disk of platinum metal but for greater sensitivity in detecting trace levels of hydrogen palladium nanoparticles can be used as described by Liang et al., “Controllable synthesis of hollow hierarchical palladium nanostructures with enhanced activity for proton/hydrogen sensing”, Journal of Physical Chemistry C, 2008, vol. 112, pp. 338-344. The counter electrode is composed of platinum, while the pseudo reference electrode is composed of silver.

(107) The fuel cell is operated under the following conditions: operating temperature=90° C., gas flow rate in anode compartment=1.2 liter/min, relative humidity of gas in anode compartment=100%, total anode gas pressure=3.0 bar, gas flow rate in cathode compartment=1.2 liter/min, relative humidity of gas in cathode compartment=100%, total cathode gas pressure=1.5 bar, area of membrane 154 made of Nafion=100 cm.sup.2, and thickness of Nafion membrane 154=50 μm.

(108) The hydrogen sensor measures a concentration of 5.74×10.sup.−6 moles of hydrogen per kilogram of ionic liquid (0.015 ppm), which corresponds to an equilibrium partial pressure of 0.0035 bar in the gas stream exiting the cathode. Assuming the oxygen, hydrogen and water vapour in the mixture behave as pure components, the concentration of hydrogen in the exit gas of the cathode is 1.18×10.sup.−4 moles per liter.

(109) The concentration of hydrogen C.sub.H2 in the gas exiting the cathode is related to the flux J.sub.H2 across the hydrogen-permeable membrane of the sensor by

(110) $C_{H2}=\frac{J_{H2}}{Q_C}$
where Q.sub.C is the total flow rate of gas through the cathode compartment. With Q.sub.C=1.2 liter/min the flux J.sub.H2 is 1.41×10.sup.−4 moles per min across the entire membrane, which yields a specific flux of 2.4×10.sup.−8 moles/cm.sup.2/s. The sensor can be used to monitor the hydrogen crossover flux through the membrane and indicate when it is above an acceptable level that is determined by considerations of safety or performance.

(111) In this embodiment a single hydrogen sensor 170 is used, but it will be appreciated that an array of hydrogen sensors distributed throughout the cathode chamber 166 and hence distributed over the area of the cathode 152 could be provided if desired.

(112) It will be appreciated that the example embodiments described in detail above can be modified and varied within the scope of the concepts which they exemplify. Features referred to above or shown in individual embodiments above may be used together in any combination as well as those which have been shown and described specifically. Accordingly, all such modifications are intended to be included within the scope of this disclosure as defined in the following claims.