Authors:
NENAD ZECEVIC, R. KAMENSKI, Petrokemija Plc, Kutina, Croatia
JOHN ROBERT BRIGHTLING, H. AGARWAL, Johnson Matthey Plc, Billingham, UK
This article was first published at Nitrogen and Syngas Conference, 5-8 March 2013, Berlin, Germany
This paper will look at how the operators of an ammonia plant selected their LTS catalysts and the criteria that they used when comparing catalysts, including legislation and operational constraints. This case study will present the design conditions and will compare its actual performance with previous charges in respect of both methanol formation and the water gas shift reaction.
INTRODUCTION
One of the most expensive catalysts in ammonia production is the low temperature shift catalyst (LTS catalyst). If proper care is not taken, the life time of this catalyst can be significantly shorter in comparison with the other catalysts which will cause, in turn, very high operational costs. The performance of the LTS catalyst can trigger different operational problems and lead to poor economics and even shutdown of the plant. Besides achieving good LTS catalyst activity, a very important factor in performance of the same is good selectivity. Plant operators have to look at reactions other than the water gas shift that can occur across the LTS catalyst. From an operational, economic, legislation as well as environmental stand point of view the most important reaction which must be avoided is the methanol synthesis reaction. Due to this, the operators must choose the best LTS catalyst on the market to be on the safe side regarding the operational, economic, legislation and environmental requirements associated with this methanol make. In the case of proper LTS catalyst choice, the operator mitigates many plant and environmental problems by reducing the amount of methanol produced in the LTS converter.
PETROKEMIJA LTS CATALYST SELECTION CRITERIA
In the shift conversion step, carbon monoxide reacts with steam to form equivalent amounts of hydrogen and carbon dioxide. Since the shift conversion reaction is reversible one in which reaction rate is favoured by high temperatures and the equilibrium by low temperatures, two stages are provided each utilizing a different type of catalyst, as illustrated in Figure 1. Initially, the process gas from the secondary reformer which contains from 12.0% to 15.0% of CO (dry gas basis) is passed through a bed of iron oxide/chromium oxide catalyst at 350°C to 380°C and then over a copper oxide/zinc oxide catalyst at approximately 200°C to 220°C both of them at alumina carrier. The final residual CO content of the gas is 0.2% to 0.4%. If the reaction is near equilibrium, a decrease in temperature will likely improve conversion, however if it is not near equilibrium a temperature decrease will reduce the shifting reaction. Conversely, if the reaction is near equilibrium, an increase in temperature will result in a loss of conversion. The temperature conditions selected for the shift converter are based on a higher temperature for the high temperature section to take advantage of the higher reaction rate for the high CO content of the inlet gas and comparatively lower temperature for the low temperature section to take advantage of a favourable equilibrium conditions for the lower CO concentrations in this section of the converter. To achieve the best performance in activity and life time of the both shift catalyst different manufacturers vary the concentrations of the iron, chromium, copper, zinc and alumina. To enhance catalyst properties, producers include different additives, usually alkali metals to suppress methanol synthesis reaction.
Fig. 1: Two stage process for water gas shift conversion
Copper is the active component in the LTS catalyst for the water gas shift and methanol synthesis reaction, since the both reactions can take place simultaneously across the LTS catalyst, according to the following reactions:
In ammonia production plants excessive formation of methanol can lead to:
1. loss of ammonia production due to H₂ and CO consumption,
2. contamination of CO₂ that can cause problems for downstream users,
3. contamination of process condensate with ammonia, amine and methanol.
From the above mentioned reactions it is easy to calculate that for each tonne of methanol made across the LTS converter due to poor selectivity; approximately 1.1 tonnes of ammonia is lost. For safety reasons the produced CO2 must be free of the methanol and if the proper LTS catalyst is not chosen there is a danger of CO2 contamination, which must be solved with alternative methanol removal systems. Depending of the operating and capital costs, this can be some kind of scrubber or catalytic oxidation reactor.
During the catalyst reactions in the reforming catalyst sections and HTS catalyst section it is also produced trace levels of ammonia. This ammonia can react with the methanol across the LTS catalyst to produce amines, mainly methylamine. A higher methanol concentration therefore results in a higher amine concentration in the LTS catalyst effluent. Very high concentration of ammonia, amine and methanol at the inlet of the process condensate stripping unit may cause very high contamination of the stripped process condensate and, when using stripping system with low pressure steam, eventually can be vented to the atmosphere. In the case of very high contamination of the process condensate, the contaminants can have a negative influence in further process streams during the subsequent treatment of the process condensate with ion exchange resins and also have a negative influence on environmental pollution of air. In many parts of the world there are also new environmental rules which regulate the amount of VOC content emitted from ammonia plants. In the case of Petrokemija the national legislation requirements from the Regulation of Limit Emission Levels of the Pollutants in the Air from the Stationary Sources do not recognized methanol as a specific pollutant. Instead it is classified in a general manner according to the article 22, similar to overall carbon content in organic matter, which has a limit of 50mgm-3 or a mass flow of 500g per hour. The limit for methanol in the process condensate is 50mgl-1as a result of its impact on the functional groups of the ion exchange resins that are used for the treatment of the process condensate from the stripper column.
Due to the above mentioned operational, economic, legislation and environmental constraints Petrokemija must consider to choose the best LTS catalyst on the market, which will ensure as high as possible activity and in the same time as much as possible low selectivity against methanol synthesis reaction.
DISADVANTAGES OF THE CURRENT PROCESS CONDENSATE STRIPPING SYSTEM AND POSSIBILITIES FOR IMPROVEMENTS IN METHANOL REDUCTION
Figure 2 shows the current system for the LTS converter and the related equipment for the LTS catalyst reduction, with the process condensatestream to the process condensate stripping system and subsequent CO2 stream to downstream users. The main process stream treated in the process condensate stripping system is the CO shift converter effluent condensate. Treatment of condensate is carried out by steam stripping in a packed tower which operates at close to atmospheric pressure. The combined process condensate stream at 116°C enters the top of the stripper and flows down through two packed sections where, on contact with steam, the bulk of the ammonia, carbon dioxide, methanol and dissolved gases are stripped out. According to the design basis, condensate feed is expected to have up to 1000 ppm of NH3, 3000 ppm of CO2 and from 200 to 2000 ppm of CH3OH depending of activity and selectivity of the LTS catalyst in the ammonia unit. The high activity of the LTS catalyst results in the production of relatively large amounts of CH3OH in the process condensate and at times levels of up to 2000 ppm can be expected. The residual CH3OH content of the treated process condensate is expected to vary depending upon the amount in the stripper feed. On average this is expected to be in the order of 100ppm. The overhead vapours from the stripper column, containing mainly steam, NH3, CO2 and CH3OH are vented to atmosphere. Stripped process condensate is pumped from the bottom of the column to the process condensate treatment facility that uses ion exchange resins.
Fig. 2: LTS process loop with related equipment for LTS catalyst reduction and process condensate stripping system
In attempt to determine how to handle methanol made across LTS catalysts, ammonia plant operators recognize that methanol formation not only leads to environmental pollution, but can also affect the operability and profitability of the plant. Because of these factors ammonia plant operators can consider a number of solutions to prevention of methanol contamination of the environment and process systems. These solutions concentrate on either the destruction of the methanol formed or the control of by-product formation:
1. high pressure process condensate stripper – destruction of methanol in the primary reformer,
2. condensation of the stripper overheads followed by a pump to inject it into primary reformer,
3. lowering the LTS converter inlet temperature,
4. increasing the steam to gas ratio to the LTS converter,
5. lower the plant operating pressure,
6. adjusting the temperature in the condensate separator,
7. reducing the catalyst volume which decreases contact time,
8. simply waiting for the catalyst to lose methanol activity,
9. install the best LTS catalyst on the market which can achive best activity and selectivity against formation of the methanol with the longest possible life time.
The first two proposed methods are effective at removing methanol from the process condensate, but they nothing to solve the problem of amine production and CO2 contamination which directly affects end consumers. They also have relatively high investment costs.
Formation of methanol across the LTS convertor is controlled by kinetics rather than equilibrium. Reduction in the operating temperature will therefore lead to a decrease in methanol production. The operator must therefore keep the inlet temperature to the LTS converter as low as possible, taking into account minimum temperature limitations due to the dew point of the process gas stream. The minimum operating inlet temperature should be at least 20°C above the process gas stream, which in practice is usually between 193°C and 202°C. The temperature margin between the operating temperature and dew point should be carefully controlled to prevent condensation on the catalyst which can be detrimental to its performance. Lower operating temperatures also impact on the activity if the catalyst. Increasing the steam to gas ratio at the inlet of the LTS converter can reduce the methanol make depending on the life and operating temperature of the LTS catalyst. Increasing the mass flow rate of steam to the primary reformer reduces methanol make by reducing the CO partial pressure and average operating temperature of the LTS converter. However, increased steam to gas ratio can lead to higher plant fuel requirements and increased front end pressure drop. A higher operating pressure will also increase the amount of methanol make across an LTS catalyst. However, the operating pressure is usually fixed and cannot be changed to affect methanol make.
Decreasing the contact time between the process gas and the LTS catalyst (increasing the space velocity) will also decrease the amount of methanol make. In plants with only one LTS converter a reduction in catalyst volume will also lead to reduction in catalyst life.
Changes to the operating temperature and pressure of the process condensate separator can affect the amount of methanol in the process condensate and CO2 exit from the stripper column as illustrated in Figure 3. If the temperature in the condensate separator is above 120°C, 50 to 75% of the methanol can go overhead with the vapour going to the CO2 absorber column, while in the case of the lower temperature the main content of the methanol will stay in the condensate. However, the overall content of the methanol will stay the same.
Methanol formation is always highest at the start of run and decays quickly with time.
Fig. 3: Dependence of the methanol vapour/liquid concentrations with the temperature
Ammonia plants that include a low process condensate stripping system as part of the initial designdo not have many options for adjusting process conditions, because most plants have been designed to operate at a fixed pressure and steam to gas ratio to minimize energy consumption and maximize production. Changes such as those outlined above can adversely affect both energy consumption and plant rate. The best solution in such cases is using the best low methanol variant LTS catalyst on the market which will reduce the methanol make and in the same time to achieve the minimum CO slippage with longest life time.
THERMODYNAMICS AND EQUILIBRIUM
Thermodynamically, low operating temperatures are favoured for any exothermic reaction. The reaction rate, however, decreases as the operating temperature decreases. Beyond a threshold temperature, the water gas shift and methanol synthesis reactions will cease to occur. Figure 4 shows the relative activity of LTS catalyst at different temperatures. Therefore, a judicial choice between the kinetic viability and thermodynamic limitations are necessary to balance the catalyst performance with reactor operating conditions.
Fig. 4: Relative activity versus temperature dependence for the water gas shift reaction
The temperature dependence of the equilibrium constant Kp for the water gas shift reaction can be defined with the two following equations:
Where Z = (1000/T)-1 with T being the absolute temperature (in Kelvin).
The variation of the heat of reaction may be calculated with the following formula:
To keep the temperature low the heat of reaction must be removed in an appropriate way, and to achieve a sufficient reaction rate effective catalyst have to be applied.The introduction of the copper-zinc based low temperature shift catalyst in 1963made it possible to take advantage of the lower equilibrium CO concentrations at temperature around 200°C, shown for various steam to gas ratios in figure 5. Low CO concentrations even at very low steam to gas ratios are attainable with a sufficiently active LTS catalyst.
Fig. 5: Equilibrium CO concentrations in LTS catalyst in the case of average process gas stream at the inlet of LTS converter during the production of ammonia
One significant difference between the methanol synthesis and water gas shift reactions is the impact of the partial pressure of water, which acts as an oxidant agent. The oxidation potential in a shift reactor is higher than in methanol reactor. The nature of a catalyst surface reflects its exposed feed composition. According to literature, depending upon the reaction conditions, the morphology of Cu particles may change their shape reversibly. This indicates that catalyst surfaces change dynamically according to the reaction atmosphere. Taking into account all dynamic features, it is possible to design a Cu/Zn/Al catalyst suitable for the water gas shift reaction with minimum activity toward the methanol synthesis reaction. Alkali metals have been shown to impact the relative rates of water gas shift, methanol synthesis and higher alcohol synthesis reactions; their impact is in general ion specific in the following order:
Cs > Rb > K > Na ≈ Li
In LTS catalyst manufacture, if the Cu/ZnO catalyst is correctly doped with Cs it has a strong effect on the relative water gas shift and methanol synthesis reaction rates. The optimum concentration of Cs to minimise the relative methanol synthesis reaction rate is a function of where it is deposited on the Cu crystal face.
Experimental and industrial data show that shift selectivity increases with time on-stream. This is due to crystalline changes in the structure of the catalystunlike the water gas shift reaction, where catalyst activity depends on the structure of the copper crystals present on the catalyst surface, methanol formation depends only on the total copper surface area of the catalyst. The copper surface area in LTS catalysts sinters rapidly at the start of life as a consequence of the operating temperature and the presence of steam. Since methanol formation in an LTS reactor is kinetically limited, such changes in copper area have a direct impact on the levels of methanol formed over the catalyst. As a result, methanol by-product formation, which is highest at the start of catalyst life, falls rapidly within the first 6 months of operation as the copper surface area of the catalyst falls by sintering. Therefore, if the methanol formation rate can be controlled at start-of-run conditions, it will remain in control throughout the life of the catalyst. As stated earlier, this can be done by changing operating process conditions for the catalyst or changing the whole catalyst.
CASE STUDY – OPERATION WITH LTS CATALYSTS KATALCOJM 83-3X IN PETROKEMIJA AMMONIA PLANT AND ACHIEVED SAVINGS
Due to the constraints previously mentioned in this paper, Petrokemija decide not to invest in very costly equipment or changes in plant operating conditions. Instead the plant chose a lower cost alternative by installing a more selective LTS catalyst.
From the operation, legislation, environmental and economic stand point of view Petrokemija wanted to achieve following benefits from the LTS catalyst:
1. minimum CO slip,
2. low and stable pressure drop,
3. robustness to operational upsets,
4. minimum by-products ( methanol and amine),
5. long life and economic performance,
6. good self guarding for sulphur and chloride poisons,
7. fast procedure of reduction and start-up.
In satisfying all these requirements Petrokemija recognized that the KATALCOJM 83-3X series of LTS catalysts from Johnson Matthey was best solution. From 1998 until 2004, the LTS ran with a combined loading of KATALCOJM 83-3 (70%) and CCE C18-HCS (30%). In February 2005 this was replaced by a full charge KATALCOJM 83-3X which operated with very satisfactory results until it was changed out during a planned shutdown in December 2011 (nearly 7 years later). Due to the performance of the previous charge, KATALCOJM 83-3X was installed again during the last overhaul and successfully reduced in February 2012. Again, its performance to date has met expectations.
KATALCOJM 83-3 catalyst is a high activity low temperature shift catalyst with long life, durability to upset conditions and has excellent resistance to the poisons as it is aself-guardcatalyst (no specialty guard layer required). KATALCOJM 83-3 catalyst is a mixture of CuO and ZnO on alumina carrier in the shape of pellet 5.2 x 3.0 mm and with a bulk density of 1380 kgm-3. KATALCOJM 83-3X catalyst contains an additional optimized combination of alkali metal promoters to suppress the methanol formation. Methanol levels are reduced to less than 15% of those achieved with KATALCOJM 83-3 catalyst. The promoters also boost poisons pick-up, resulting in the highest poison capacity of any commercially available low temperature shift catalyst. The catalyst is made in the form of pellets with dimension of each pellet as 5.2 x 3.0 mm and with a bulk density of 1360 kgm-3.
In the Petrokemija ammonia plant, the shift conversion step consists of two vertical single bed reactors and the necessary exchangers to remove heat from the process gas as it passes from the high temperature converter to the low temperature converter. Flow of the process gas is downward through the catalyst bed. After the high temperatures shift converter, partially shifted gas must be cooled before entering the second stage. A portion of the heat is used to produce high pressure steam in the heat exchanger, with the remainder utilized to heat the methanator feed. A trim cooler enables the low temperature shift inlet temperature to be further controlled. The shift converter effluent is cooled prior to entering the CO2 removal system. The cooling scheme consists of quenching the shift effluent to the dew point by injection of recycled process condensate and utilization of the available heat to regenerate the Benfield solution used for CO2 removal and to preheat the boiler feed water. The water condensed from the process gas at this point provides the quench water for the shift effluent with the net condensate going to the condensate stripper. This is illustrated in Figure 2.
The low temperature shift converter contains one bed of catalyst with a volume of 108.8 m3. As the catalyst is self-guarding, the whole of the bed provides shift activity throughout the life of the charge, with the top portion of the bed retaining trace amounts of sulphur and chloride poisons that pass through the high temperature shift converter.
During operation KATALCOJM 83-3/CCE C18-HCS and KATALCOJM 83-3X catalyst were monitored via three main factors:
1. CO slip
2. pressure drop through the bed of catalyst
3. concentration of the methanol in the condensate at the inlet to the column of the stripper condensate
The process conditions for all three charges of LTS catalyst wereas follows:
1. process gas flow at the inlet of LTS converter = 195000 Nm-3h-1 to 200000 Nm3h-1
2. steam to gas ratio at the inlet of primary reformer = 3.5
3. pressure at the inlet of LTS converter = 28.0barₐbs to 29.0barₐbs
4. inlet temperature to the LTS converter = 200°C to 202°C
Figure 6 shows the CO concentration in the outlet process stream of the LTS converter. In all three cases, the start of run CO slip is very low and meets equilibrium. The CO slip remained below 0.2% throughout the life of both of the first two charges, as predicted by Johnson Matthey. The same performance is expected for the most recent charge.
Fig. 6: CO concentrations exit the LTS converter for three charges of LTS catalysts, 1998 to 2012
Figure 7 shows pressure drop during the life time of the catalyst. In the case of ordinary catalyst KATALCOJM 83-3/ CCE C18-HCSthere was an increase inpressure drop which had a negative influence on the energy demand at the front end of the plant.
However, with a full charge of KATALCOJM 83-3X,the pressure drop stayed approximately the same during the whole life time. This guaranteed very high potentially in energy savings, particularly in the operation of the air and natural gas compressors.
Fig. 7: Pressure drop through the catalyst bed for three charges of the LTS catalysts, 1998 to 2012
Figure 8 shows the methanol concentration profile through the life time of all three charges. In all three cases the expected higher methanol content at the beginning of the run is seen clearly. The decrease in activity for methanol formation during the early part of the catalyst life is clearly seen in the decrease in methanol formed over time. It is also very obvious that the concentration of the methanol from KATALCOJM 83-3X which uses an alkali promoter to suppress the methanol synthesis reaction is less than 15% of the methanol made with ordinary LTS catalyst. This data proves the main advantage of the KATALCOJM 83-3X catalyst.
Fig. 8: Methanol concentrations in process condensate at the inlet of stripper process column for three charges of LTS catalysts, 1998 to 2012
The last installed charge of LTS catalyst KATALCOJM 83-3X has been in operation since February 2012. The performance of the charge has been continuously monitored via the concentration of the methanol in the raw CO2 from the top of the stripper column in the Benfield system by one of the end users of the CO2 (Figure 9). The medium for the washing the synthesis gas in the absorption column is a hot solution of potassium carbonate with the mass concentration of 29.0 to 30.0% and activator LRS 10 with a mass concentration around 3.0%. The purity of CO2 is always higher than 99.0%. The temperature of the process gas at the outlet of the vessel for the process condensate separation is below 120°C. The concentration of methanol in the raw CO2 is between 40 ppm and 80 ppm, which is 20 to 25% of the overall methanol concentration from the outlet of the LTS converter. The archived result is consistent with literature data, which says that when operating in a temperature regime below 120°C, 50 to 70% of the overall methanol content remains in the process condensate.
Fig. 9: Methanol concentrations in the raw CO2 at the outlet of stripper CO2 column in the Benfield system for KATALCOJM 83-3X, February 2012 to December 2012
FINANCIAL SAVINGS ACHIEVED
From the methanol synthesis reactions (2& 3) it is clear that every molecule of methanol consumes 2 molecules of H2 and 1 molecule of CO. This means that each 1.0 tonne of methanol equates a loss of about 187 kg of H2 and 1.1 tonnes of ammonia.
The value of pressure drop is also very significant and from the industrial experience it can be concluded that lowering the pressure drop for 1 bar saves around €120.000 per year. In the case of Petrokemija with yearly production of 450.000 tonnes of liquid anhydrous ammonia and steam to gas ratio of 3.5 at the inlet of primary reformer, the savings resulting from the installation of the charge of KATALCOJM 83-3X equate to approximately €450.000. This estimate is based on savings due to less methanol production improving ammonia production, and does not include the impact of the lower pressure drop in the front end of the plant or the lower methanol content in the synthesis loop that acts as an inert gas.
CONCLUSIONS
One of the main problems during the water gas shift reaction step in ammonia production is the simultaneously reaction to make methanol. The impact of the methanol synthesis reaction during the ammonia production is variable and the main constraints are the impact on operability, profitability, legislation and environmental pollution in water and air. Due to this, ammonia operators must take certain measures to satisfy all relevant requirements. One of the best measures is to install the right LTS catalyst which will insure the maximum operability and reliability of the plant, energy savings, meeting the legislation requirements and prevention of the all environmental issues.
In solving of all imposed constraints, Petrokemija recognized the benefit of using the improved LTS catalyst, KATALCOJM 83-3X from Johnson Matthey. In the KATALCOJM 83-3X catalyst, Petrokemija recognized the outstanding catalyst life time and low methanol by-product formation, in conjunction with very good poisons resistance and activity. These features enable the maximum period of using the catalyst which fits the plants turnaround cycles, maximizes hydrogen production and address environmental concerns.
Parallel to using the improved LTS catalyst Petrokemija are developing a new project to installa medium pressure stripper column for process condensate to achieve even better energy savings and minimum environmental pollution according to the legislation demands.
ABBREVIATIONS
HTS – High Temperature Shift
LTS – Low Temperature Shift
N – Normal conditions at 101325 Pa and 273.15 K
VOC – Volatile Organic Carbon
References
EFMA (2000). “Production of Ammonia”, Best Available Techniques for Pollution Prevention and Control in the European Fertilizer Industry
Yara, Brunsbüttel Comments on D2 (2004). “TWG’s comments on the second draft BREF”, personal communication.
European Commission (1997). “Pilot Document for Ammonia Production”.
Regulation of Limit Emission Levels of Pollutants in the Air from the Stationary Sources, NN 124/2012.
Jim Richardson, Jayanta Bandyopadhyay and Yeping Cai, LTS catalyst Design: Maximizing Shift Converter Performance, Ammonia Technical Manual, AIChE 2002.
J.D. Grunwaldt, A.M. Molenbroek, N.Y.Topsoe, H.Topsoe and B.S.Clausen, J.Catalyst, 194, 452 (2000).
J.G. Nunan, C.E. Bogdan, K. Klier, K.J. Smith, C.W.Young and R.G.Herman, J.Catalyst, 113, 410 (1988).
Nakamura, M.Campbell and C.T.Campbell, J.Chem., Soc.Faraday Trans., 86, 2725 (1990)
M.V. Twigg, Catalyst Handbook, Wolfe Publishing Ltd. (1989).
Max Appl, Ammonia – Principles and industrial Practice, Wiley-VCH, (1999).
Anderson, P.V. Broadhurst, D. Cairns, F.E. Lynch and C.Park, The science of catalysis Part 3: Water gas shift and methanation, Nitrogen and Syngas, 312, July – August 2011.
Pach, Causes and consequences of by-product formation in ammonia plants, Ammonia Technical Manual, AIChE 2009.
P.V. Broadhurst, I. Johnston, C. Park, Development of a Kinetic Model for Methanol Formation over Low Temperature Shift Catalysts, Nitrogen Conference, 2006.
