The present invention is directed towards a method of re- generating deactivated methane oxidation catalysts, in par ^¬ ticular methane oxidation catalysts employed cleaning of exhaust gas from lean burn gas engines at reduced tempera ^¬ tures . Natural gas (compressed or liquefied) is a clean and af ^¬ fordable fuel. The discovery of large amounts of shale gas as well as the large natural resources has contributed to make natural gas a competitor of traditional oil products. Unlike oil, natural gas prices have been stable the past decade which make it a competitive alternative. In addi ^¬ tion, natural gas has very low sulphur content and offers opportunities for complying with most stringent emission regulations. It is today a serious candidate for powering engine - so-called gas engines.

Operation of lean burn gas engines has a major drawback referred as methane slip. It results from incomplete combus ^¬ tion of the fuel injected in the pre-chamber or cylinders of the engine. The remaining methane in the exhaust is therefore up to several thousand of ppm.

Given the limited temperature of the exhaust gases, removal of the methane slip is a difficult task. The stable chemi ^¬ cal configuration of methane prevents oxidation unless tem- perature is increased significantly and therefore demands use of state-of-the-art catalyst. State-of-the-art lean burn methane oxidation catalysts are typically composed of pure palladium on high surface oxides or compositions of palladium and/or platinum and/or rhodium on high surface oxides.

The high activity of the catalyst at low temperature is vi ^¬ tal. However, it makes the catalyst prone to adsorb a wide range of poisons. If the natural gas is mostly free from poisons, lubrication oils and other oil used in the engine are still sources of potential poisons. The most critical are sulphur oxides formed from sulphur compounds. The sul ^¬ phur level originating from the mechanical oil is far below legislation limits, typically resulting in ppm or ppb levels of SOx in the flue gas. It is, however, sufficient for reducing the life time of the catalyst significantly. Deac ^¬ tivation of the catalyst by sulphur adsorption, allows the operation of the catalyst only within a given activity range. When deactivation has proceeded to the limit of this range, the catalyst must be replaced or reactivated.

Traditional regeneration techniques involve significant temperature increase which damages the catalyst; in partic ^¬ ular have negative effect upon low temperature activity of the catalyst.

Present attempts to remove methane slip in oxygen contain ^¬ ing exhaust gas at low temperature levels i.e. between 400 and 700°C combined with attempts for thermal catalyst re ^¬ generation in oxygen containing gas in the temperature range 600 - 850C° have not been successful. We have found that deactivated methane oxidation catalysts can periodically regenerated in situ by means of a reducing agent being passed over the catalyst during regeneration. In this case the regeneration can take place within the temperature range of normal operation or below. The nature of the reducing agent to be used maybe tailored to the spe ^¬ cific catalyst. Methanol, natural gas, very low sulphur diesel, alcohols, ethers, hydrogen, carbon monoxide or am ^¬ monia are suitable candidates.

Essentially suitable temperature levels for regeneration would be from 150°C to 750°C.

Thus, the invention provides a method for the regeneration of a deactivated methane oxidation catalyst comprising the step of bringing the deactivated catalyst in contact with a reducing agent, selected from the group of methanol, natu ^¬ ral gas, very low sulphur diesel, alcohols, ethers, hydro ^¬ gen, carbon monoxide and ammonia or mixtures thereof.

These regeneration agents enable regeneration at lower temperatures, hence preserves the catalyst low temperature ac ^¬ tivity. The catalysts will normally be in the oxidised state or be poisoned by sulphur oxides forming sulphate or sulphite compounds on the catalyst surface that further can migrate to the carrier oxides. The reducing agent enables removal of the sulphur oxides, sulphates and sulphites as gaseous hydrogen sulphide and sulphur dioxide. As methane is already available, the reducing agent can be generated in a methane or natural gas burner running rich at lambda below 1.00 where CO, H2 and reactive hydrocarbons can be formed.

Thus, a preferred reducing agent is hydrogen and/or carbon monoxide.

The hydrogen reducing agent can also be prepared by one of methane or hydrocarbon reforming, methanol reforming or cracking, or ammonia cracking.

Reducing conditions may also be accomplished by utilizing the thermal inertia in the methane oxidation reactor and catalyst. In this case the reducing agent such as methanol, other alcohols, ethers, ammonia or mixtures thereof are fed into the methane oxidation reactor and the thermal inertia is used to convert the reducing agent into H2 and CO which then regenerate the catalyst. The formation of H2 and CO from other reducing agents is catalysed by the methane oxi ^¬ dation catalyst or a separate catalyst also placed in the methane oxidation rector. The formation of H2 and CO from the other reducing agents is highly endothermic and thus the methane oxidation reactor must be designed with suffi ^¬ cient thermal mass to provide enough heat for generating enough H2 and CO to regenerate the methane oxidation cata- lyst.

Alternatively natural gas, other hydrocarbons including very low sulphur diesel mixed with water or steam may also constitute a reducing agent for the preparation of H2 and/or CO. In further an embodiment, methane containing flue gas is by-passed the deactivated methane oxidation catalyst during the regeneration and brought into contact with a regenerat ^¬ ed methane oxidation catalyst

In another embodiment, the H2 and/or CO reducing agent is prepared by burning natural gas at lambda below 1.00, typi ^¬ cally between 0.85 and 0.97. Suitable methane oxidation catalysts comprise palladium and/or platinum and/or rhodium.

Preferred catalysts comprise palladium on yttria and/or lanthanum stabilized alumina, optionally coated on cordier- ite or corrugated metal sheets.

In an embodiment, the oxidation catalyst can be arranged in the methane oxidation reactor in form of a monolith with a plurality of straight channels.

In these embodiments, a part of the channels can under re ^¬ generation with the reducing agent, while the remaining channels are active in the oxidation of methane. It may be preferred that the reducing agent is passed in counter current flow direction to the flow direction of a methane containing gas .

Handling of the regeneration off gas containing carbon mon- oxide, hydrogen and unburned hydrocarbons with a minor ele ^¬ vation of sulfur compounds as S02 and H2S can be problemat ^¬ ic . On a ship it may be allowed to discharge the off gas to the environment for a short regeneration time. But for some sea areas, vehicles and stationary units the regeneration off gas must be properly treated in order to avoid exhaust of particularly CO and H2S.

Thus it might be preferred off gas from the regeneration of the deactivated methane oxidation catalyst is recycled and mixed with the engine exhaust gas and passed to a regener ^¬ ated methane oxidation catalyst.

Alternatively, the off gas can be recycled to a gas engine connected to the methane oxidation reactor.

Different modes of operation of the method according to the invention are shown in the drawings.

Operation of the method in general is shown in Fig. la and lb. It consists of two stages, namely the slip removal

(Fig. la) and the regeneration (Fig. lb) stage. Surplus of oxygen in the flue gas ensures combustion and the tempera ^¬ ture is increased when passing through the methane slip catalyst. During this phase, the deactivation proceeds with sulphur oxides deposition. The flue gas cleaned from me ^¬ thane slip is passed through the turbo charger or recycled to the engine as EGR (exhaust gas recirculation) . The cata ^¬ lyst is regenerated when the catalyst activity is close to the operation limit or optionally periodically after a pre- determined time. The flue gas is directed to the turbo- charger and not cleaned, hence there will be some methane slip emission during the regeneration time. During the re- generation, a reducing agent is injected upstream of the methane slip catalyst and the flow over the catalyst re ^¬ moves sulphur poisoning as ¾S or SO ₂ . The sulphur contami ^¬ nated reducing agent is mixed with the flue gas and exits through the turbo-charger.

If the regeneration stage is lengthy, the emissions of me ^¬ thane might reach unacceptable levels. Fig. 2 shows a system with two parallel methane slip cata ^¬ lyst units for ensuring that slip is always removed. One of the units is in operation and the other is under regenera ^¬ tion. Fig. 3 presents a system where the additional fuel is con ^¬ verted catalytically to hydrogen prior to injection as re ^¬ ducing agent, instead of expensive storage of hydrogen. This conversion could be methane or hydrocarbon reforming, methanol reforming or cracking or ammonia cracking.

Alternatively, it could be a dehydration reaction for pre ^¬ paring an ether (as reducing agent) from alcohol.

In Fig. 4 a fraction of flue agent form the engine is passed to the catalyst unit during regeneration for combus- tion of a fraction of the additional fuel ( methane or nat ^¬ ural gas) in a burner, thereby increasing the temperature during regeneration. This enables to increase/ control the temperature during the regeneration phase. Although some oxygen containing flue gas is injected during the regenera- tion phase, the regeneration occurs in reducing atmosphere, hence the reducing agent is in excess compared to oxygen to maintain fuel rich conditions. As already mentioned above the reducing agent can be gener ^¬ ated in a methane or natural gas burner running rich at lambda <1.0 where CO, H2 and reactive hydrocarbons can be formed. This embodiment is shown in Fig.5

Fig. 6 shows a further alternative, wherein an oxidation reactor is divided in two or more compartments where the reducing gas is directed to one compartment while the other compartments are used for methane oxidation. Simple dampen- ers can be used to control the flow. Due to the fact that the required reducing gas volume to regenerate the catalyst is modest, the emissions of residual CO and H2 are kept to a minimum.

In order to minimize the slip of exhaust gas into the at ^¬ mosphere the reducing gas can be redirected to the main gas flow at the catalyst inlet using flaps also at the catalyst outlet and a recycle, as shown in Fig.7.

Fig. 8 shows a device moving across the catalyst surface injecting reducing gas and thus continuously regenerating one catalyst segment at the time. Due to the fact that the required reducing gas volume to regenerate the catalyst is modest, the emissions of residual CO and H2 are kept to a minimum. Ceiling flaps can be used to avoid mixing reducing and oxidizing gas close to the catalyst surface. Openings in the distributor top segment can be used to direct excess reducing agent away from the catalyst surface.

In order to minimize the slip of CO and H2 into the atmos ^¬ phere the reducing gas can be collected at the catalyst outlet and returned to the main flow where the reducing agent H2 and CO will be oxidized when passing the catalyst bed together with the oxygen rich main flow as shown in Fig.9. The collector follows the movement of the regenera- tive gas distributor. Alternatively, the reducing agent can be injected in counter current direction towards the main gas flow. The reducing gas will have to be injected a high ^¬ er pressure level compared to the main flow. Reducing gas will in this layout (not shown) automatically mix with oxy- gen rich gas at the main flow inlet and oxidize when pass ^¬ ing the catalyst in the direction of the main gas flow of methane containing gas .

Fig. 10 shows a rotating methane oxidation reactor. The re- ducing agent can be collected at outlet of reactor and dis ^¬ tributed into methane containing bulk gas flow (not shown) . The reducing agent can in this layout also be injected in counter current direction towards the main gas flow of me ^¬ thane containing gas .

In the case of a two-rector layout as shown in Fig. 11, for example a gas engine operating with two parallel rectors, one reactor can be bypassed during regeneration. Burner off gas from rich (lambda below 1.00) combustion of methane containing the H2 and CO reducing agent is fed counter current to the engine exhaust gas flow into the reactor with the catalyst that shall be regenerated. The regeneration off gas is subsequently passed to the other reactor that is in normal engine exhaust methane slip operation. This re- quires that the burner is operating at higher pressure than the pressure of the main gas flow of methane containing gas . Fig.12 shows a possible layout for gas engine applications operating with only one reactor. The methane oxidation reactor is bypassed during the regeneration of the methane oxidation catalyst (CH4 slip cat.) and burner off gas con ^¬ taining the H2 and CO containing reducing agent to the tur- bocharger air inlet after passing the methane oxidation catalyst. The reducing gas is thus redirected back to the engine .

Fig.13 shows a layout suitable for gas engine applications operating with two reactors. In this layout one methane ox ^¬ idation reactor is regenerated while the methane containing engine exhaust gas is passed to the other reactor. Flue gas containing H2 and CO reducing agent from a burner operating at lambda below 1.00 is passed to a turbocharger air inlet after having passed through the methane oxidation catalyst for regeneration and is thus redirected back to the engine. Examples

Example 1

In this example it is shown that heating of a Pd based cat- alyst to 600 °C does not restore the catalytic activity for methane oxidation.

A quartz U-tube reactor with an inner diameter of 4 mm was filled with 200 mg of a catalyst consisting of 2 wt% Pd on La-promoted alumina, which was granulated to a size range of 150-300 ym. The reactor exit gas was determined by on- line infrared spectrometry, using a GasmetDX4000 FTIR spec ^¬ trometer. The catalyst was heated in a mixture containing 880 ppm CH ₄ , 10% 0 ₂ , and 5% H ₂ 0 in inert gas to 600 °C for 1 h, without added S02 to the feed gas. After this initial heating the catalyst was cooled down to 450 °C and the con ^¬ version of methane with time on stream was determined, us ^¬ ing a total flow of 200 Nml/min (8.93 mmol/min) . It is known that Pd catalysts deactivate during methane oxidation (P. Gelin, L. Urfels, M. Primet, E. Tena, Catal . Today 83 (2003) 45, R. Burch, P.K. Loader, F.J. Urbano, Catal. Today

27 (1996) 243) and a decreasing conversion with time on stream is expected.

Table 1 lists the measured CH ₄ conversion levels at differ ^¬ ent times on stream after initial heating to 600 °C, and additional 1 h heating to 600 °C, after the conversion had dropped to about 66%. It is clear that heating to 600 °C after deactivation restores a catalytic activity for a short period, but after about 3 h after the regeneration, the conversion level is already below the level measured immediately before the regeneration, which indicates that no stable activity is obtained after heating to 600 °C.

Table 1. Measured conversions of methane at different times on stream after initial heated and attempted regeneration by heating at 600 °C. Time on Initial Time on Heated at 600

stream stream °C after de ^¬ Conversion%

activation

Fresh cata ^¬ Heated

lyst catalyst Conversion%

0 h 100 0 h 97.7

1 h 97.4 1 h 78.9

3 h 86.7 3 h 59.3

13 h 66.1 13 h

Example 2.

This example shows that heating the catalyst to 600 °C af ^¬ ter deactivation due to S O2 exposure does not restore the catalytic activity. After the measurement described in Ex ^¬ ample 1, approximately 10 ppm S O2 was added to the feed gas, keeping temperature at 450 °C and the total flow at 200 Nml/min. Addition of S O2 results in an accelerated de ^¬ crease of the methane conversion, and after a total of 3.5 hours of exposure to S O2 , the conversion had decreased to 43%, and the conversion remained at 43 % after removal of the S O2 from the feed gas. Table 2 lists the measured con- versions of methane after 1 h heating at 600 °C; both the heating treatment and the activity measurements were done using the feed gas as mentioned in Example 1 without S O2 added to the feed. After heating, the methane conversion is increased for a short period of time. After only 1.5 hours, the conversion level is almost back to the level it had be- fore the heating, indicating that heating to 600 °C does not result in a stable regeneration.

Table 2. Measured conversion of methane, using feed gas without added SO ₂ , after deactivation by SO ₂ , and attempted regeneration by heating to 600 °C in feed gas without added

S0 ₂ .

Example 3.

This example shows that a Pd catalyst can be regenerated in a gas containing ¾ and CO, corresponding to the product gas of a fuel-rich combustion of methane. Starting from the catalyst after the measurements in Example 2, 10 ppm SO ₂ was added to the feed once again and the methane conversion was reduced to about 35%. The conversion remained constant after removal of the SO ₂ from the feed gas. The catalyst was then exposed to a gas mixture consisting of 3% ¾, 3% CO, 8.5% C0 ₂ , 20% H ₂ 0 in N ₂ (65.5%) - the estimated composi ^¬ tion of the exhaust gas from a fuel-rich combustion of natural gas at about 10% air deficiency (A «0.9) [ North

American Combustion Handbook, Vol I: Combustion, Fuels, Stoichiometry, Heat Transfer, Fluid Flow., 3rd Ed, North American Mfg. Co., Cleveland, OH, n.d.3] - for 30 min. at 450 °C at a flow rate of 200 Nml/min. Immediately after this exposure to a fuel-rich exhaust gas, the methane con ^¬ version was measured at 450 °C and a flow of 200 Nml/min feed consisting of 880 ppm CH ₄ , 10% 0 ₂ , and 5% H ₂ 0 in inert gas (Table 3) . Clearly, the methane conversion is 100 % im ^¬ mediately after the treatment, and shows a deactivation be ^¬ havior, like the catalyst before regeneration (Example 1) .

Table 3. Measured conversions of methane at before and af ^¬ ter exposure to 3% H ₂ , 3% CO, 8.5% C0 ₂ , 20% H ₂ 0 in N ₂

(65.5%) at 450 °C.

Time on stream Exposure to 3% H ₂ , 3%

CO, 8.5% C0 ₂ , 20% H ₂ 0

in N ₂ (65.5%) at 450

°C.

Before treatment 34.8

0 h 100

1.5 h 100

2 h 100

4 h 99.4 Example 4

This example shows that the regeneration by exposure to the fuel-rich exhaust gas can be repeated without affecting the catalyst performance. In this example, the flow of the 880 ppm CH ₄ /10% 0 ₂ ,/5% ¾0/inert gas mixture was increased to 300 Nml/min. The catalyst was deactivated by adding 10 ppm S O2 and 20 ppm S O2 to the feed, and regenerated after each deactivation by exposure to a fuel-rich exhaust gas with the composition given in Example 3. In table 4, it is seen that the conversion of methane is restored after regenera ^¬ tion, and that the performance of the catalyst does not de ^¬ teriorate after repeated cycles of deactivation and regen- eration.

Table 4 Measured conversions of methane (flow 300 Nml/min) at before and after exposure to 3% H ₂ , 3% CO, 8.5% CO2 , 20% H2O in 2 (65.5%) at 450 °C, in sequential deactivation- regeneration cycles using different concentrations of S O2 .

Time on stream 10 ppm S0 ₂ 20 ppm S0 ₂

Before regen ^¬ 52.3 16.8 eration

After regeneration

0 h 100 100

1.5 h 97.6 95.6

2 h 94.0 88.6

4 h 84.3 76.3