WO1993000581A1

WO1993000581A1 - Gas sensor

Info

Publication number: WO1993000581A1
Application number: PCT/GB1992/001148
Authority: WO
Inventors: Carl John Sofield; Patrick Timothy Moseley
Original assignee: United Kingdom Atomic Energy Authority
Priority date: 1991-06-26
Filing date: 1992-06-24
Publication date: 1993-01-07
Also published as: GB2271643B; GB2271643A; AU2153392A; GB9324390D0; GB9114038D0

Abstract

Semiconductor gas sensors of modified sensitivity or selectivity or thermal stability are prepared by chemicoradiolytic activation of the sensor: this is achieved by irradiation or bombardment of the sensor, with ionising radiation or high energy particles, in the presence of an interactive material in its gaseous or liquid form.

Description

Gas Sensor

This invention relates to a gas sensor and more especially to a semiconductor gas sensor and its method of manufacture.

By the term semiconductor gas sensor is meant a gas sensor which comprises a body of a semiconductor material typically a metal oxide, which is deposited on an electrically insulating substrate and is in contact with at least two electrodes. The semiconductor material is such that its electrical resistance changes in the presence of gases to be detected. This change in resistance may be an increase or a decrease. The change in resistance of the semiconductor material on exposure to a gas can be measured via the electrodes using a suitable amplifier and detector. The electrodes are typically referred to as resistance electrodes. Semiconductor gas sensors which operate in this manner may be referred to as resistance effect gas sensors.

Japanese laid-open patent application No. 61-104070 describes a method of vacuum depositing a thin film of material containing oxygen in which the material is irradiated with an electron beam during its deposition in the presence of a gaseous atmosphere containing oxygen as a constituent element. The object of the invention is to introduce oxygen into the material being deposited without causing a deterioration in the crystallinity of the thin film. It is stated that thin films composed of substances containing oxygen or oxides may be used, amongst other things, in various gas sensors.

According to the present invention there is provided a semi-conductor gas sensor comprising, an insulating substrate, a body of semiconductor material formed on the insulating substrate chemico-radiolytically activated in situ, and at least two electrodes in electrical contact with the chemico-radiolytically activated semiconductor material.

Also according to the present invention there is provided a method of manufacturing a semiconductor gas sensor including the operations of contacting an active surface of a body of semiconductor material forming part of the semiconductor gas sensor with an interactive material capable of interacting with the semiconductor material under the influence of ionising radiation or high energy particles to alter the electrical properties of the semiconductor material, and exposing the surface of the body of semiconductor material to the ionising radiation or high energy particles thereby to activate the semiconductor material chemico-radiolytically and alter the electrical properties thereof. The semiconducting material may be an oxide, for example ZnO, SnO₂, ln₂O₃, MoO₃, Ga₂O₃, Nb₂O₅, WO₃, TiO₂, Fe₂O₃, Ta₂O₅, bronzes such as Bi₆Fe₄Nb₆O₃₀, niobates such as K₂Nb₆O₁₆, pyrochlores such as Bi₂Sn₂O₇, tantalates such as KTaWO₆, perovskite such as SrSnO₃ or tungsten oxides such as Na_0.1Nb_0.1W_0.9O₃. The semiconducting material may be a non-oxide such as CdS. The semiconductor may be an n type, p type or a mixture of semiconductor materials or it may be an organic semiconductor such as a semiconducting polymer. The substrate may be of any material which is electrically insulating such as alumina.

By the term chemico-radiolytically activated is meant a modification or change in the electrical properties which occurs at the active surface of a semiconductor material forming part of a semiconductor gas sensor as a consequence of that surface being treated with ionising radiation or high energy particles in the presence of an interactive material in its gaseous or liquid form.

By the term ionising radiation is meant radiation which induces ionisation on impinging upon a semiconductor material for example γ radiation, β radiation or UV light. By the term high energy particles is meant ions, atoms, nuclei or molecules which have sufficient energy to induce ionisation on impacting upon a semiconductor material for example H⁺, Cl⁶⁺ or α particles.

By the term interactive material is meant any material which in the presence of ionising radiation or high energy particles interacts physically or chemically with the active surface of a semiconductor material forming part of a semiconductor gas sensor during chemico-radiolytic activation.

Suitable interactive materials include for example N₂, O₂, Air, H₂O, S and sulphur-containing materials, hydrocarbons and other organic materials, catalytic metals such as Pd and Pt in organometallic compounds and any material which can be used to produce a mixed metallic oxide. These interactive materials may be used in their gaseous or liquid forms and may be chemisorbed or physisorbed onto the active surface of the semiconductor material before or during its chemico-radiolytic activation. It is believed that the ionising radiation or high energy particles may ionise the active surface of the semiconductor material or the interactive material or both and that the interactive material in gaseous or liquid form interacts physically or chemically with the active surface of the semiconductor material to produce a chemico-radiolytically activated surface. The chemico-radiolytic activation of a semiconductor gas sensor may be carried out at ambient, sub-ambient or elevated temperatures , ambient or elevated pressures or in a partial vacuum. The semiconductor gas sensor and the interactive material in gaseous or liquid form may be at the same or different temperatures during chemico- radiolytic activation.

When high energy particles are used their energy may be between 1 keV up to 40 MeV, most preferably between 2 MeV and 30 MeV. When ionising radiation is used, its energy will depend on the nature of the semiconductor material, but should be sufficient to induce ionisation of the semiconductor material in the active surface region of the semiconductor material. The fluence of high energy particles may be up to 1×10²⁰ particles cm^-2 and is typically in the range 1×10¹² - 1×10¹⁶ particles cm^-2.

The semiconductor gas sensor according to the present invention has superior properties compared to known semiconductor gas sensors. Depending on the nature of the semiconducting material, the nature of the interactive material in gaseous or liquid form and the nature of the ionising radiation or high energy particles used, the activated semiconductor gas sensor may have increased sensitivity, increased selectivity, and/or improved high temperature stability. Chemico-radiolytic activation may increase or decrease the background resistance in air of an activated gas sensor.

The semiconductor gas sensor according to the present invention may be used to detect any material which induces a change in the resistivity of the chemico-radiolytically activated semiconductor material. Gaseous materials which may be detected include for example; O₂, NH₃, C₂H₄, CO, H₂ and CH₄. In some circumstances semiconductor gas sensors may need to be operated at elevated temperatures for example 400-600ºC or greater. The semiconductor gas sensor can be brought to these operating temperatures by a resistance heating element which is in contact with the electrically insulating substrate.

As a consequence of irradiation with ionising radiation or bombardment with high energy particles the chemico-radiolytically activated semiconductor material may have reduced resistivity compared to that before chemico-radiolytic activation. This may be as a consequence of radiation damage to the bulk semiconductor material and/or as a consequence of chemico-radiolytic activation. If reduced resistivity is observed it is preferred that the chemico-radiolytic activation is followed by a thermal annealing stage. The thermal annealing stage requires the chemico-radiolytically activated gas sensor to be held at an elevated temperature e.g. 800°C until the resistivity of the sensor reaches at least 90% of the resistivity before chemico-radiolytic activation and more preferably 100%.

When high energy particles are used it is not necessary that they be implanted into the semiconductor material. However it may be that ion or particle implantation during or after chemico-radiolytic activation may have a beneficial effect.

The invention will now be described, by way of example only, with reference to the accompanying drawings in which:

Figure 1 shows a schematic longitudinal sectional view of a semiconductor gas sensor suitable for chemico-radiolytic activation;

Figure 2 shows graphically the variation of Log resistance (Ω) with elapsed time (minutes) for the exposure of the semiconductor gas sensor of Figure 1 , before ( - - - - ) , and after

, chemico-radiolytic activation, to a variety of gases at an operational temperature of 450°C;

Figure 3 shows graphically the variation of the resistance in air minus the resistance in 1% CO divided by the resistance in air of the semiconductor gas sensor of Figure 1 (●) and the same semiconductor gas sensor after chemico-radiolytic activation (■) with temperature (°C) for the detection of 1% carbon monoxide in air;

Figure 4 shows graphically the variation of resistance (kohms) with elapsed time (minutes) for the exposure of tin oxide gas sensors, (a) irradiated with 4 MeV protons in an air atmosphere; and (b) irradiated with 4 MeV protons in a nitrogen atmosphere, to one pulse of 1% methane in air and one pulse of 1% carbon monoxide in air at an operational temperature of 400°C;

Figure 5 shows graphically the variation of resistance (kohms) with elapsed time (minutes) for the exposure of tin oxide gas sensors (a) comprising 3 coats of tin oxide from a sol and (b) 1 coat of tin oxide from a sol, to one pulse of 1% methane in air and one pulse of 1% carbon monoxide in air at an operational temperature of 450°C;

Figure 6 shows graphically the variation of resistance (kohms) with elapsed time (minutes) for the exposure of the tin oxide gas sensors (a) and (b) of Figure 5, after treatment with 4 MeV protons in an air atmosphere, to one pulse of 1% methane in air and one pulse of 1% carbon monoxide in air at an operational temperature of 450ºC;

Figure 7 shows graphically the variation of resistance (kohms) with elapsed time (minutes) for the exposure of a tin oxide gas sensor comprising 3 coats of tin oxide from a sol (a) before; and (b) after treatment with 4 MeV protons in air, to one pulse of 1% methane in air and one pulse of 1% carbon monoxide in air at an operational temperature of 600°C;

Figure 8 shows graphically the variation of resistance (kohms) with elapsed time (minutes) for the exposure of a tin oxide gas sensor comprising 1 coat of tin oxide from a sol (a) before; and (b) after treatment with 4 MeV protons in air, to one pulse of 1% methane in air and one pulse of 1% carbon monoxide in air at an operational temperature of 600°C;

Figure 9 shows graphically the variation of Log resistance (ohms) with elapsed time (minutes) for the exposure of gas sensors 1, 2, 3 and 4 of Table 3 to a variety of gases at an operational temperature of 450°C;

Figure 10 shows graphically the variation of Log resistance (ohms) with elapsed time (minutes) for the exposure of gas sensors 3 and 4 of Table 3 (a) before; and (b) after ageing at 860°C and 780°C for 60 hours in air for sensor 3 and 4 respectively, to a variety of gases at an operational temperature of 600°C;

Figure 11 shows graphically the variation of Log resistance (ohms) with elapsed time (minutes) for the exposure of gas sensor 1 of Table 3 to a variety of gases at an operational temperature of 600°C; and

Figure 12 shows graphically the variation of Log resistance (ohms) with elapsed time (minutes) for the exposure of gas sensor 2 of Table 3 to a variety of gases at an operational temperature of 600°C. Referring to Figure 1 a semiconductor gas sensor 11, suitable for chemico-radiolytic activation includes an electrically insulating substrate 12 which acts as a support for a layer 13 of a semiconductor material, interdigitated electrodes 14 and a resistance heating element 15. During chemico-radiolytic activation the semiconductor gas sensor of Figure 1 is irradiated or bombarded such that the ionising radiation or high energy particles impinge or impact upon an active surface 16 of the layer 13 of semiconductor material.

The invention will now be further described by way of the following examples: Example 1

Alumina substrates in the form of thin wafers were coated on one side with a Pt based coating to form a resistance heater element. The reverse side of each wafter was coated with a Pt based coating to form a series of interdigitated resistance electrodes. SnO₂ semiconductor material was then sputter coated onto the substrates to a thickness of approximately 200nm in order to cover the interdigitated electrodes whilst maintaining the resistance heater element side of the substrate free from SnO₂ semiconductor material.

Semiconductor gas sensors as prepared above were chemico-radiolytically activated by treatment with either 4 MeV protons in air at a flux of up to 10¹⁶ protons per cm² or 30 MeV Cl⁶⁺ ions in a vacuum (10^-6 Torr) at a fluence of up to 5 x 10¹⁴ ions per cm². After treatment the resistivity of the sensors in air was measured, if this was less than 90% of that recorded before treatment the sensors were annealed at 800°C until the resistivity was more than 90% of the resistivity of the untreated sensors. The chemico- radiolytically activated gas sensors were used at 450°C to detect various gases by monitoring the change in resistivity of the modified sensors on exposure to the gas. The responses for various gases are illustrated in Figure 2 and the effect of activation or sensitivity to a number of gases are listed in Table 1. The results in relation to CO₂ are an indication of O₂ sensitivity as the sensors were exposed to nominally 100% CO₂. Thermal Stability Test

An unactivated and a chemico-radiolytically activated gas sensor (4 MeV H⁺ in air) were tested for their thermal stability by monitoring the resistivity of the sensors in 1% CO at temperatures up to 600°C. The activated sensor had improved stability as indicated by Figure 3.

Example 2 An alumina substrate in the form of a thin wafer was coated on one side with a Pt based coating to form a resistance heater element. A gold based coating was deposited on the reverse side of the wafer to form an interdigitated electrode array. Tin oxide was then deposited by radio frequency (RF) sputtering to produce films of thickness of approximately 1μm in order to cover the interdigitated electrodes whilst maintaining the resistance heater element side of the substrate free from SnO₂ semiconductor material.

The tin oxide gas sensors as prepared above were chemico-radiolytically activated by treatment with either 4 MeV protons in an air atmosphere at a dose of 3 × 10¹⁶ particles cm^-2 or 4 MeV protons in a nitrogen atmosphere at a dose of 3 × 10¹⁶ particles cm^-2. The gas response behaviour of the activated sensors was measured at 400°C, 450°C, 500°C, 550-C and 600°C for 1% methane in air and 1% CO in air. The activation increased the background resistance (i.e. in air) of the sensors. Activation with protons in a nitrogen atmosphere for one sample increased the background resistance by approximately 21% from 3.48 MΩ to 4.2 MΩ and for another sample with protons in air by approximately 37% from 2.70 MΩ to 3.70 MΩ.

The gas response behaviour as measured at 400°C is illustrated in Figure 4. At 400°C the normal response to methane (a resistance decrease) was not seen; instead the sensor irradiated in air (a) shows little net response, and that irradiated in nitrogen (b) shows a small resistance increase. The sample irradiated in nitrogen shows a significantly greater response to carbon monoxide than the sample irradiated in air. This difference was maintained at operating temperatures up to 500°C.

Three samples were analysed by X-ray photoelectron spectroscopy (XPS); the resultant data is listed in Table 2. The data shows that no N is present in the as prepared gas sensor before activation; but after activation with 4 MeV protons in air there is a 4.6% N at the surface which is doubled by activation in an atmosphere of N₂.

Low energy electron loss spectra (EELS) were also recorded. The spectrum of the as prepared gas sensor was characteristic of SnO₂ but after activation with 4 MeV protons in N₂ there was a broadening of the spectra which indicated that the structure had been modified. The XPS analysis indicated a N_1s peak of binding energy 405.7eV. These results indicate that N has reacted with and been incorporated into the surface of the gas sensor after irradiation in air or N₂.

Example 3 An alumina substrate was prepared as in Example 2 with the exception that the tin oxide was deposited from a tin oxide sol, of concentration 100 gl^-1. by dipping the substrate into the sol, drying, and firing in air at 600°C. Specimens were prepared with either 1 coat or 3 coats of tin oxide and each coat was estimated to be greater than lμm thick on the basis of a comparison with results of deposition onto a glass substrate.

Specimens were irradiated with a beam of 4 MeV protons to a dose of 3 × 10¹⁶ particles cm^-2 in an atmosphere of air.

The gas response of these sensors was measured as for Example 2. The results at 450°C for sol-gel 3 coats and sol-gel 1 coat, with and without activation are shown in Figure 5 and 6. The results at 600°C for sol-gel 3 coats and sol-gel 1 coat are shown in Figure 7 and 8 respectively.

As can be seen from the Figures the resistance of the 1 coat was higher than the 3 coat; this relationship was repeated at temperatures up to 600°C. The activated 3 coat sensor exhibited higher resistance over the non-irradiated sensor at all temperatures up to 600°C. This was also true for the 1 coat sensor up to 500°C but at 600°C the difference was virtually lost.

It is clear that sensor prepared from sols behave in the same general way as those prepared from RF sputtering; there were shifts in the base line resistance after treatment when measured in air and in the relative size of the response on exposure to gases. The results achieved with the sol-gel method were generally more reproduceable than those achieved with RF sputtering; this is possibly due to the greater control over coating thickness which was possible with the sol-gel method.

Example 4

Alumina substrates were preared as in Example 2 to produce a tin oxide gas sensor. The SnO₂ films after deposition were black due to a small deficiency of oxygen which was readily restored by firing in air at up to 850°C for 20 minutes. These sensors were then chemico-radiolytically activated with various particle beams, doses and exposure atmospheres using a 6 MV tandem accelerator. The resistance (ohms) in air was measured at 450°C for each sensor and the results are listed in Table 3.

The gas response of these sensors was determined as described in Example 2, for methane, carbon monoxide, hydrogen, ethene and ammonia. These gases were used as 1% mixtures in dried air with the exception of ethene which was used as a 0.7% mixture in dried air and hydrogen which was used as a 100 ppm or 1% mixture in dried air. Intermediate purges of dried air were also used. In addition nominally 100% CO₂ (which is equivalent to 10^-6 atmospheres of O₂), and nominally 100 ppm N₂ in Argon were also included to test for O₂ sensitivity. For each test the gas response cycle was repeated to ensure that no irreversible modification of the gas sensing surface had occurred during the first exposure. The gas response results are illustrated in Figure 9, 10, 11 and 12.

Considering the results in Table 3 it is clear that the nature of the irradiating particles, the dose and the irradiation atmosphere have a marked effect on the resistance of the sensor in air at 450°C. For example with irradiation with 4 MeV protons the resistance at 450°C in air increased in the order air>Argon>vacuum 10^-6 Torr. Although not shown in these results irradiation in N₂ produces a higher resistance at 450°C in air than irradiation in air. In general irradiation led to a decrease in the resistance in air of all the sensors listed in Table 3. All of the irradiated sensors retained gas sensitivity with the exception of sensor 4 which received a high dose of chlorine ions (see Figure 9). This loss was reversed to some extent by thermal treatment at 780°C for 60 hours in air (see Figure 10). Also the background resistance of sensor 3 was significantly increased after thermal treatment at 860°C for 60 hours in air (see Figure 10). The pattern of gas sensitivity of both sensors 3 and 4 did not return to that of the unirradiated sensor when used at the same temperature.

The gas response data also illustrates that chemicoradiolytic activation can improve the selectivity of gas sensors. For sensor 1 irradiation with 4 MeV protons in air increased its sensitivity to oxygen, carbon monoxide, hydrogen and ammonia when measured at 600°C but had little effect on the sensitivity to 1% CH₄, 100% carbon dioxide and 0.7% C₂H₄ (see Figures 11 and 12).

The activation has caused a decrease in resistance and for certain combinations of conditions an increase in gas response which may not apply to all gases equally and hence the activation improves selectivity.

A general observation from the results is that chemico-radiolytic activation is capable of substantially modifying the gas sensing properties of semiconductor gas sensors such as those made from tin oxide.

Table 2

X-ray Photoelectron Spectroscopy Analysis (Atom %) of the surface of tin oxide sensors

H⁺ Treatment H⁺ Treatment

Element As Sputtered in Air in Nitrogen

Sn 28 13.4 14.9

O 51.6 50.1 46.9

N 0.0 4.6 9.2

C 20.2 14.1 17.6

Pb 0.0 2.8 2.8

Mg 0.0 1.8 1.5

Si 0.0 4.9 2.1

Al 0.0 7.6 4.5

Cu 0.0 0.2 0.0

TABLE 3

Sensor Type of Energy Dose Atmosphere Resistance in air Reference irradiation of ion particles (cm^-2) (ohms) at 450°C Number

1 Nil 1.8 x 10⁵

2 H⁺ 4 MeV 10¹⁶ air 1.3 x 10⁵

3 ^{3 5}C l ⁶⁺ 30 MeV 5 × 10¹³ vacuum, 10^-6 Torr 1.6 x 10⁴

4 ^{3 5}C l ⁶⁺ 30 MeV 5 × 10" vacuum, 10^-6 Torr 1.6 x 10³

5 Nil 5 x 10⁵

6 N⁺ 1 keV 10¹⁶ nitrogen, 10^-4 mbar 4 x 10⁴

7 Ar+ 1 keV 10¹⁶ argon, 10^-4 mbar 5 x 10⁴

8 H⁺ 4 MeV 10¹⁶ air 1.6 x 10⁵

9 H⁺ 4 MeV 10¹⁶ vacuum, 10^-6 Torr 3.5 x 10⁴

10 H⁺ 4 MeV 10¹⁶ argon 5 x 10⁴

Claims

1. A semiconductor gas sensor comprising, an insulating substrate, a body of semiconductor material formed on the insulating substrate chemico-radiolytically activated in situ and at least two electrodes in electrical contact with the chemico-radiolytically activated semiconductor material.

2. A semiconductor gas sensor as claimed in Claim 1 wherein the semiconductor material is a semiconducting oxide.

3. A semiconductor gas sensor as claimed in Claim 2 wherein the semiconducting oxide is a tin oxide.

4. A semiconductor gas sensor as claimed in any of the preceding claims wherein the chemico-radiolytically activated semiconductor material incorporates nitrogen atoms.

5. A method of manufacturing a semiconductor gas sensor including the operations of contacting an active surface of a body of semiconductor material forming part of the semiconductor gas sensor with an interactive material capable of interacting with the semiconductor material under the influence of ionising radiation or high energy particles to alter the electrical properties of the semiconductor material, and exposing the surface of the body of semiconductor material to ionising radiation or high energy particles thereby to chemico-radiolytically activate the semiconductor material and alter the electrical properties thereof.

6. A method as claimed in Claim 5 wherein the ionising radiation is α, β or γ radiation.

7. A method as claimed in Claim 5 wherein the high energy particles are either protons, chlorine ions, argon ions, or nitrogen ions.

8. A method as claimed in any of one of Claims 5 to 7 wherein the interactive material is air or nitrogen.

9. A method as claimed in any one of Claims 5 to 7 wherein the interactive material is oxygen, water, sulphur, a sulphur containing compound, or catalytic metals in organometallic compounds.

10. A method as claimed in any one of Claims 5 to 9 wherein the semiconductor material is derived from radio frequency sputtering.

11. A method as claimed in any one of Claims 5 to 9 wherein the semiconductor material is derived from a sol.

12. A method as claimed in any one of Claims 5 to 11 wherein the semiconductor gas sensor is thermally annealed after chemico-radiolytic activation.

13. A method as claimed in Claim 12 wherein the annealing is undertaken at a temperature of 800°C or less.

14. A method as claimed in either Claim 5 or Claim 7 wherein the energy of the particles is between 1 keV and 40 MeV.

15. A method as claimed in Claim 14 wherein the energy of the particles is between 2 MeV and 30 MeV.

16. A method as claimed in Claims 14 or 15 wherein the fluence of high energy particles is less than 1 × 10²⁰ particles cm^-2.

17. A method as claimed in Claim 16 wherein the fluence of high energy particles is in the range 1 × 10¹² - 1 × 10¹⁶ particles cm^-2.

18. A semiconductor gas sensor obtainable by the method of any one of Claims 5 to 17.

19. The use of a semiconductor gas sensor as claimed in any one of Claims 1 to 4 or Claim 18 for the detection of gaseous O₂, NH₃, C₂H₄, CH₄, CO, or H₂.