Application of Thermoplate Cooled Reactors in Sulfur Recovery

Basics

Reactors with internal cooling in sulfur recovery are not very common. But there are good reasons why to apply them. The prime incentive is that they allow very high sulfur recovery rate (SRR) with a rather simple plant configuration, as shown in the SMARTSULF process flow diagram below (Fig 1). The process as shown can reach up to 99.9% sulfur recovery rate. To achieve such values the more conventional sulfur recovery processes require much more complex plants.

Fig. 1: Process flow diagram of SMARTSULF

Process description

The acid gas to the sulfur recovery unit is burned sub-stoichiometrically exactly as in any Claus plant. Downstream follows a waste heat boiler, a sulfur separator and a reheat to the 1st reactor. This reactor is different from the conventional ones.

The fundamental idea of SMARTSULF is removing reaction heat of the Claus reaction directly in the catalyst bed rather than in a downstream heat exchanger. This controls the temperature at the outlet of the catalyst bed within a narrow range. The heat exchanger applied is a thermoplate stack with large clearances. The space between the thermoplates is filled with catalyst which is that way efficiently cooled. As this type of heat exchanger is not yet so well known it will be discussed in more detail below. But first the process needs to be described.

The first converter of a Claus plant is always a compromise between two competing targets: Reaction temperature has to be high enough for maximum COS and CS2 hydrolysis and as low as possible for a favorable equilibrium of the Claus reaction and thus maximum conversion. The internally cooled reactor can solve this conflict: The top layer of the catalyst is left without cooling. The feed temperature to this section is typically 220°C to 240°C and reaction heats it up to appr 320°C. That is the temperature required for COS and CS2 hydrolysis. The second section downstream in the same reactor is cooled and a fixed outlet temperature slightly above the sulfur dewpoint can be reached by evaporating boiler feed water. This combination of an adiabatic and a cooled section in one reactor reaches conversion rates comparable to a two-stage Claus plant.

Downstream of the 1st reactor follows a sulfur condenser and then a 2nd identical reactor which, however, is operated at lower temperature. This shifts the chemical equilibrium towards more sulfur formation. Actually the outlet temperature is chosen in the range 100°C to 125°C, i.e. it may be even below the sulfur solidification point.
During the operation below the sulfur dew point the sulfur produced accumulates on the catalyst and deactivates it slowly. Therefore, it has to be regenerated and that is done by switching it into the position of the 1st reactor. There at the high temperature of up to appr 320°C the sulfur is desorbed and the catalyst thus regenerated. The former 1st reactor is switched at the same time into the position as the cold 2nd reactor. This procedure is repeated typically once every 24 hours.

The treated gas finally is sent to the incinerator and then to stack.


Discussion of SMARTSULF versus more conventional sulfur recovery processes

Conventional sulfur recovery processes with better than 99.0% recovery rate have one disadvantage in common: They add tailgas clean-up processes to the Claus process. That makes the whole system rather complex and thus prone to much maintenance, increased downtime, higher capital and operating cost. SMARTSULF takes a new approach as it combines the catalytic Claus process with a SubDewPoint tail gas treatment in just 2 catalytic reactors.

The process principle of sub-dew-point tail gas treatment is well-proven and is known e.g. from the CBA and SULFREEN processes. Different from these older processes SMARTSULF requires only 2 catalytic converters which are cooled internally. This allows a rather low temperature at the outlet of the reactors. As a direct consequence the sulfur recovery rate rises which is shown in Fig 2.

Fig 2: Sulfur recovery rate as a function of outlet temperature

Thermodynamics clearly drive the SRR up as the temperature goes down. Lower limits are the water dewpoint of the gas which must not be transcended, and kinetics which might render the reaction too slow. This thermodynamic effect has been known for very long time. But in the conventional processes it is not possible to operate the outlet temperature below ca 130°C. The reason is that sulfur is a very good insulator. Therefore the outlet temperature of the heat exchanger upstream of a SULFREEN or CBA reactor has to be kept above the sulfur solidification temperature, i.e. it cannot be lower than ca 125°C. Due to the Claus reaction the temperature in the reactor rises a bit so that the outlet temperature from the reactor is close to or even above 130°C. This corresponds to an optimum SRR of ca 99.5%. Fluctuations in feed gas composition and/or flow plus fluctuations of the controllers reduce the SRR so that in practice the long-term average of SRR in these plants is typically 99.0% to 99.2%.

In SMARTSULF the heat exchanger in the converter reduces the temperature from the typical inlet temperature of 190°C to the outlet of down to ca 100°C. In the catalyst bed the sulfur formed is adsorbed on the catalyst faster than on the heat exchanger surface. This allows to reduce the outlet temperature of the converter to lower values, even below the sulfur solidification point and that results in sulfur recovery rates up to 99.85%. Again this value is under ideal conditions. The long-term average is typically by 0.2% to 0.3% lower, as in the conventional sub-dew-point processes.

Such high values for the SRR could be achieved in the past only by much more complicated processes, e.g. a complete Claus plant with a downstream hydrogenation plus amine scrubber, as BSR Amine, SCOT or equivalent for tail gas treatment. As the 2-reactor process contains much less equipment and process steps, it is substantially cheaper and – being less complex – it tends to be more reliable as well.

The first commercial scale plant using this 2-reactor process was started up in December 1995 at Nynäs Refinery in Sweden. Two gas streams from an amine unit and a sour water stripper had to be treated. The plant proved to be very reliable, easy to operate and cheap in maintenance. The on-stream factor was always better than 99.5%/a. The customer claims that even now after more than 15 years in operation this unit is the most reliable in the whole refinery. It showed sulfur recovery rates as calculated, the optimum values reaching up to 99.85%, with aged catalyst. At low-load operation the sulfur recovery rate dropped by only 0.1 % at a turn-down ratio of 6:1. Fig 3 shows a photo of the plant.

Fig 3: Sulfur Recovery Unit at NYNÄS refinery, Sweden

Based on these positive results the process was further improved over the years in some details and now a 2nd generation is available. It is even lower in capital cost and some minor children’s diseases were eliminated. Essentially three modifications were introduced:

1. For internal cooling a special type of plate heat exchanger, the thermoplate exchanger is now applied rather than the expensive coiled tubular heat exchangers used earlier. This reactor system is patented.

2. Preferably closed loop steam systems with air coolers to condense the steam raised are used for cooling each of the two reactors. This system is very similar to the coolers of car engines.

3. The two switch-over valves required in the process are reliable 4-way valves with a rotating cock. Patent for these valves is pending.

The new generation of catalytic reactors in SMARTSULF applies thermoplate heat exchangers and is therefore called “Thermoplate Reactor”. It avoids all the disadvantages of tubular reactors named above.

The catalyst is embedded between the thermoplates, while the cooling medium flows within the thermoplates, usually boiler feed water. Several thermoplates are combined to form a heat exchanger module. A number of such modules are then combined to provide the total heat exchanger surface required.

Thermoplates for Internally Cooled Catalytic Reactors

Internally cooled catalytic reactors have proved out in many applications. They are applied primarily for selective reactions, where a rigorous temperature control is required, or in reactions where the chemical equilibrium is strongly temperature dependant. The Claus reaction for sulfur recovery is of this second kind. So far mostly tubular heat exchangers were used for reactor cooling, mostly with the catalyst in the tubes. For a few cases also spiral wound tubular heat exchangers were applied with the tubes submerged in the catalyst. However, these types of reactors had a number of features which are disadvantageous. Primarily often the heat exchangers’ buildable geometry forced conditions on the catalytic reactions which were not optimal. For example the straight tube reactors had to be built slim and high in order not to have too high thermal stress on the tube sheets. That means necessarily high pressure drop, high linear gas velocity and mechanical stress on the lower catalyst particles. The spiral wound exchangers avoid these disadvantages to some degree. But they require many manufacturing steps and high skill and therefore typically are rather expensive. All these features of tubular reactors are rather unfavorable for sulfur recovery plants.

The catalyst is embedded between the thermoplates, while the cooling medium flows within the thermoplates, usually boiler feed water. Several thermoplates are combined to form a heat exchanger module. A number of such modules are then combined to provide the total heat exchanger surface required.

The basic element of the heat exchangers is the thermoplate. The principle is shown in Fig 4.



Fig 4: Principle of a thermoplate heat exchanger

A thermoplate consists of two equally thick metal sheets which are combined by point welding (see Fig. 5 and 6). At their edges these sheets are welded together by a resistance roller seam. These welding steps are done by robots which facilitates production of a lot of exchanger surface at low cost. Both the point welds and the seam welds of the thermoplates are gap free. After welding the plates are expanded by injecting high pressure liquid between the metal sheets. This opens the channels shown schematically above in Fig 4.

Usually the welding points form rectangles. By the expansion results the typical form of a stitch cushion as can be seen in Fig. 6. Several thermoplates aligned in parallel form a thermoplate stack (Fig 7) which is then incorporated in a shell (Fig. 8).

Fig 5: Point welding of Thermoplates

Fig. 6: Thermoplate

Fig. 7: Thermoplates with steamheader

Fig. 8: Thermoplate exchanger in shell

A lot more than thousand of such thermoplate heat exchangers have been built and installed in plants world-wide. The services included extremely difficult ones, as condensers for COCl2 which is both highly toxic and corrosive when in contact with water. The sizes delivered reached up to several thousand m² per unit. In conclusion this type of heat exchanger is certainly a mature and tested technology.

Thermoplate heat exchangers are compact, light weight, low pressure drop and thermally efficient. These features make these fully welded plate heat exchangers suitable for sulfur recovery converters. Thermoplate heat exchangers are compact, light weight, low pressure drop and thermally efficient. These features make these fully welded plate heat exchangers suitable for sulfur recovery converters.

Inner and outer fluid spaces are completely separated by weld seams. Contrary to other types of plate heat exchangers there is no contact between adjacent sheets. Rather the thermoplates are self-contained and pass no forces on to the next thermoplate. Therefore catalyst can be inserted between the plates without mechanical stress on the particles.

The thermoplate heat exchangers have a high heat transfer rate due to the high turbulence on both sides of the thermoplate. Therefore less exchanger surface has to be installed in comparison to tubular heat exchangers, though the difference is not very big.

The distance between plates, file height, pitch of the point welds, dimensions and number of thermoplates can be varied in a wide range. Therefore thermoplate heat exchangers can be optimally tailored to each application.

Since thermoplate heat exchangers can be put together in modules one can realize virtually any diameter and height of the heat exchanger. This in turn facilitates to adjust the heat exchanger according to the optimal operating conditions of the catalyst, so that the process conditions are best suited with respect to linear gas velocity, space velocity, pressure drop and temperature.

Mechanical considerations

For the dimensional accuracy of the reactor shell the tolerances for roundness according to DIN are adequate, i.e. no change to the normal manufacturing process is required there.

Thermoplate modules are installed in the reactor with one fixed end and the other end free to float. In sulfur recovery reactors the lower end is the fixed one. Thermal expansion is taken care of by a compensator in the pipe connecting the steam header to the pipe out of the reactor. Thus mechanical forces on the thermoplates are kept to a minimum.

The forces due to the weight of the heat exchanger, the catalyst and the dynamic forces of the pressure drop in operation are transmitted by carrier beams to the reactor shell and the dished head ends. This keeps mechanical forces away from the thermoplates themselves so that they do not interfere with their dimensional accuracy.

Filling with catalyst and pressure drop distribution

For catalytic reactors an even distribution of pressure drop over the cross section of the reactor is very important. To achieve this is rather easy in thermoplate reactors. The catalyst is distributed over the heat exchanger’s cross section as one would do in a fixed bed without a heat exchanger. This ensures both even distribution of the catalyst and even density of the packing. As the distance between adjacent thermoplates is rather big in comparison to the catalyst particles there is no risk of bridge building, even if there is not much care taken in filling in the catalyst.

For discharging the spent catalyst it is either vacuumed out or the carrier grid is unscrewed so that the catalyst particles flow out through the bottom of the reactor.

The switch-over procedure in SMARTSULF

Once the sub-dew-point reactor has been fully charged with sulfur it has to be switched to the position of the hot reactor for regeneration. The switch-over is started fully automatically after a predetermined flow of gas to the SRU has been reached. The complete procedure is then controlled and executed by the reactor sequence control program. No operator action is required in any of the steps.

In SMARTSULF the switch-over of the process gas is done by two 4-way valves (Fig 9) and in that respect again takes an approach different from the former technology. The valves are as simple as possible, containing only the cock which separates the gas streams. That makes them robust, easy to maintain and to operate. Only one drive is required to switch the two valves which are connected by a drive shaft. This also ensures that all ways are switched at exactly the same time.

Fig. 9: Sketch of 4-way valve

For switch-over the blocking element is rotated by 90°. In the position shown the gas streams 1 plus 4 are connected and 2 plus 3. After switch-over the gas streams 1 plus 2 are connected and 3 plus 4.


In conventional SubDewPoint processes, as in SULFREEN or CBA on/off valves are used for switching over gas streams. Eight such valves are required for two sub-dew-point reactors (see Fig 10 below), twelve for three reactors. These on/off valves have a number of disadvantages: They are expensive items and cause considerable pressure drop. Much care must be taken to prevent blocking by solid sulfur. And while closed they cause dead pipes with zero flow, where corrosion may attack due to ammonium salt deposits and sulfur plus water condensation.

Fig 10: CBA sub-dew-point process with its on/off valves (source: /1/)

In Fig 10 the blue valves are closed in this operating step, the white ones are open. After switch-over the white valves are closed, the blue ones open. Necessarily the pipes containing the closed valves are dead. That means that these pipes cool down even though they may be heated. Ammonium salts carried in by the process gas deposit there, as shown in the next photo. Such deposits necessarily lead to corrosion and subsequently to costly repairs.

Fig 11: Pipe from a conventional sub-dew-point process plant with ammonium salt deposits

To solve these problems of the older processes SMARTSULF uses two 4-way valves. This allows then a piping concept by which during all operating modes gas flows through all the process pipes continuously. There are no dead pipes at any time in the process. This eliminates or at least greatly reduces the problems of corrosion. The valves and of course the pipes are heated which safely prevents sulfur solidification and water condensation.

For these valves a patent is pending.

Comparison with other tailgas treatment processes

In Hydrocarbon Processing /4/ various tailgas treatment processes are compared with their respective respective CO2 footprint quantified. Among these processes is also SMARTSULF. The essential results of this paper are summarized in the below table 1 and the diagram Fig 12.

Tab. 1: Comparison of the CO2 and SO2 footprints of various Claus tailgas treatment processes

When going down the list from top to bottom one can easily distinguish the contributions of different forms of energy to the emissions. For calculation of these values please refer to the a.m. paper /4/.

CO2 footprint versus SO2 emission

With respect to the CO2 footprint and the SO2 emission the processes are a lot different. SCOT is the most efficient process to reduce SO2 emission. But for a tailgas treatment downstream of a 100 t/d, 2-stage Claus plant SCOT emits ca 241 kg/h of CO2 , while the other processes generate CO2 credits. In comparison the second best, SMARTSULF, emits 33 kg/h of SO2 versus 17 kg/h of SCOT. But these ca 16 kg/h less SO2 come at a cost of (241+1039) = 1 280 kg/h of CO2 . It is rather questionable whether this small SO2 amount recovered is worth the ca 75 times higher CO2 flow emitted. In Fig 12 below these values are combined for comparison.


In the current discussion of the negative effects of climate change CO2 emission is one of the main culprits. So it is worthwhile to consider not only the SO2 emission but also the CO2 emission.

Fig 12: comparison of CO2 emission versus SO2 emission.

Conclusions

Thermoplate reactors are a new type of internally cooled catalytic reactor using thermoplates to remove the heat of reaction. This type of heat exchanger has been built already a lot more than thousand times and proved out in many applications. The SMARTSULF sub-dew-point process applies thermoplate reactors in sulfur recovery which allows to reach up to 99.85% sulfur recovery rate in a simple 2-reactor system. The process proved to be relatively low cost, reliable and easy to operate. A turn-down ratio of 1/6 can be realized with only 0.1% loss of recovery efficiency. Of the first unit, started up in 1995 the customer still claims: It is the most reliable plant in the whole refinery. The unplanned down-time in all these years never exceeded 48 hours per year being a lot less most of the time.

The paper showed that the energy efficiency of SMARTSULF is higher than in other types of tailgas treatment. The CO2 footprints of the most common tailgas treatment processes vary by a factor of ca 5. The positive effect for the environment of this high CO2 emission versus so little gain in SO2 emission is rather dubious. Authorities should therefore re-think the rules which so far essentially neglect CO2 footprint of Claus tailgas treatment. In view of the climate change caused by CO2 emission any unnecessary CO2 should be avoided. Better suited Claus tailgas treatment can contribute to that goal.

Literature

/1/ Hydrocarbon Processing, Gas Processing Handbook 2009

/2/ http://www.world-nuclear.org/uploadedImages/org/education/IAEA%202000(1).gif

/3/ J. Mertens, “Greenhouse gas legislation and its impact on the refining industry”, Hydrocarbon Engineering, vol 14 no.11, vol15 no 1

/4/ M. Heisel, M. Rameshni, “Minimize Carbon Footprint from Claus tailgas units”, Hydrocarbon Processing, vol. 90, no. 2, p. 71ff