Ammonia Market
Over 90% of the world’s ammonia production capacity of 160 million t/y in 2001 is based on steam reforming of natural gas or (in India) naphtha. Almost all the rest, some 10 million t/y, is based on gasification of either coal or heavy oil.
The worldwide production of ammonia is by most measures the largest of any bulk chemical. The principle use of ammonia is as nitrogenous fertilizer for agriculture.
Typical plant sizes today are 1500-2000 t/d. Process licensors are currently revealing plans for plants up to 4000 or 5000 t/d size (Davey, Wurzel, and Filippi 2003; Parkinson 2001).
Ammonia synthesis takes place at high pressure over a catalyst that is usually iron, although one process uses ruthenium according to the reaction:
N2+3H25 2NH3 — 92 MJ/kmol — N2 (7-1)
A typical specification for ammonia synthesis gas is (Mundo and Weber 1982):
1:3 (For some modern processes nitrogen excess is
required)
<30 ppmv (As sum of total oxygen containing species)
(in principle as for CO and C02, but it can be washed out with product ammonia in the synthesis unit itself)
<1 ppmv
(These are also poisons. Appl [1999] gives an upper limit of 0.1 ppm for chlorine)
<2%
minimum
Ammonia plants based on gasification technologies normally surpass these specifications when a liquid nitrogen wash is the last stage of purification.
Most ammonia plants are built in conjunction with urea plants, the C02 from the ammonia plant being used directly for urea production. According to the reaction
2 NH3 + C02 ^ NH2CONH2+H20 (7-2)
one mole of C02 is required for every two moles of ammonia. Typical requirements for the C02 are as follows:
C02 >98.5 mol%
H2S + COS <2 mg/Nm3 H2 <0.15 mol%
Methanol <10 ppmv
Design Considerations
To process a raw synthesis gas to conform to this specification a number of different tasks must be completed:
• Tar and volatiles removal (if present in raw gas)
• Desulfurization
• CO shift (CO + H20 = > H2+C02)
• C02 removal
• Final removal of carbon oxides and water
• Adjustment of N2:H2 ratio
The order in which these tasks are performed may be modified to the extent that there is a choice between performing the desulfurization before or after the CO shift. Whereas older plants may have used a copper liquor wash for CO removal, most modern gasification-based ammonia plants combine the tasks of CO removal and nitrogen addition in a liquid nitrogen wash. In addition, the gas needs to be compressed to the operating pressure of the synthesis, which can vary between 90 and 180 bar.
Prior to developing a block flow diagram, two (or perhaps three) key decisions have to be taken.
First, a decision on the overall pressure profile of the plant needs to be determined. In general, there is an energy advantage to running the gasifier at the highest possible pressure, since the energy required to compress the synthesis gas is considerably more than that required for compression of the feed materials. However, there are three variables that one needs to consider:
• The maximum pressure of the gasifier selected. With oil feed, Texaco has numerous references with 80 bar gasification. Shell can also operate SGP at this pressure, but generally recommends operation at closer to 60 bar to reduce organic acid formation in the system. For coal gasifiers, current commercial experience is limited to 40 bar with one single unit at about 60 bar.
• The oxygen compression system. Here the choice is between compression of gaseous oxygen and liquid oxygen pumping. Many 900 t/d or larger ammonia plants operating with a gasifier at 60 bar have centrifugal compressors. Plants significantly smaller than this would require the use of a reciprocating machine with installed spare. Most of the 80 bar plants use liquid oxygen pumping. The overall compression energy demand for liquid pumping (i. e., air plus oxygen plus nitrogen) is some 3-7% higher than for the equivalent compression system. This cancels some of the syngas compression advantages of an 80-bar system. There are some safety arguments claimed in favor of liquid pumping, but the excellent record of oxygen turbo-compressors in this service does not argue against their use. All in all, there is not that much to choose between the two systems.
• The synthesis pressure. In the 1970s and 1980s, typical operating pressures in the ammonia synthesis were 220 bar. Today, most plants using conventional magnetite catalyst are designed to operate at 130-150 bar, occasionally up to 180 bar. Kellogg now uses a ruthenium catalyst operating at about 90 bar. The energy demand for the synthesis does not change much over the range 130-180 bar, since syngas compression gains made by operating at the lower pressure tend to be compensated for by an increased refrigeration demand. Furthermore, at higher pressures the volume of catalyst can be reduced, making the converter smaller and, despite the increased design pressure, also cheaper. Often, therefore, the optimization consists of selecting pressure levels for the gasifier and synthesis loop, which minimize the number of casings on the synthesis gas compressor but exploit that minimum number to the maximum extent.
• An additional factor to be aware of, even if it will not substantially influence the choice of operating pressure, is that with increased pressures physical solvents for desulfurization and C02 removal, gas losses through coabsorption of CO and H2 will tend to increase. In the range up to 80 bar, however, this remains within acceptable limits.
The second important decision is the selection of the acid-gas removal system and
its integration with the CO shift.
• When reviewing alternative gas treatment systems, the one immutable parameter is the specification of the syngas. For ammonia production from gasification it is not only a matter of eliminating carbon oxides, as is the case downstream of a steam reformer. One needs to remove other components, some of which are general to all gasification systems such as ammonia and HCN, others of which may be feedstock or gasifier specific, such as the hydrocarbons produced by low — temperature gasifiers. The system, which has proved itself capable of producing on-spec gas behind practically all gasification processes, is Rectisol, which uses cold methanol as a wash liquor. As a low-temperature process Rectisol is expensive. However, in the ammonia environment this is not as serious as in other applications, since a number of synergies can be achieved. All ammonia syntheses use some refrigeration to condense the ammonia from the loop gas, and the product ammonia is often stored in low-pressure tanks at a temperature of-33°С. The integration of the refrigeration systems for Rectisol and ammonia synthesis allows some savings compared with the stand-alone case.
Similarly, it is possible to integrate the refrigeration demand of Rectisol and the liquid nitrogen wash, which operates at a temperature of -196°C. An additional advantage of a physical wash is that C02 required for urea production can in part be recovered under pressure, thus saving energy in C02 compression.
• One solution that is typical for use in conjunction with a syngas cooler is an immediate desulfurization of the raw gas. After desulfurization the gas is “clean” but also dry. In order to perform the CO shift reaction, it is necessary to saturate the gas with water vapor in a saturator tower using water preheated by the exit gas of the shift converter. C02 removal takes place in a separate step. The gas has thus to be cooled and reheated twice in the process of acid-gas removal and CO shift. The necessary heat-exchange equipment causes considerable expense and pressure drop, a fact that has to be counted as a disadvantage of this system.
The raw gas from gasification of a typical refinery residue can contain about 1% H2S and say 4% C02. A simple desulfurization of this raw gas will provide a sour gas of about 20% H2S, sufficient for direct treatment in an oxygen-blown SRU. Concentration of the H2S content up to 50% does not require much additional expense. Furthermore, the C02 formed in the shift is free from sulfur and can be used for urea production or emitted to the atmosphere without further treatment.
• When operating with a quench reactor, the gas emerges saturated with water vapor at about 240°C. This temperature is too low to be able to generate high-pressure steam, so it makes sense to utilize the water vapor immediately in a raw gas shift. In addition to saving the saturator tower, this has a number of minor advantages, for example, that COS in the raw gas is also converted to H2S, and that HCN is hydrogenated to ammonia (BASF undated). The sour gase components, H2S and C02, are then removed all in a single step. Attractive as this appears, it is not without its difficulties. In this case the “natural” sour gas has an H2S:C02 ratio of about 1:50. This requires considerable expense to concentrate the H2S to an acceptable level for a Claus furnace and to clean the C02 before emission to the atmosphere.
Ultimately, however, there is not much to choose between these two systems. A detailed study performed in 1971 (Becker etal.) compared a 60 bar scheme using syngas cooling and clean gas shift with a 90 bar scheme using quench and raw gas shift. It showed very little difference between the two routes. There was a slight energy advantage for the syngas-cooler route, and the difference in investment amounted to only 2%—lower than the estimating tolerance. This result continues to be valid to this day, as can be judged by the commercial and operating success of both schemes.
In order to elucidate these matters in more detail, we provide a worked example of a lOOOt/d ammonia plant based on heavy oil feed and using a syngas cooler and clean gas shift. Based on the selected gasifier pressure of 70 bar and an overall pressure drop of 15 bar over the gas treatment train, the syngas compressor suction would be 55 bar. This would allow a synthesis loop at 135 bar. For the sake of the example, it is assumed that a liquid oxygen pump will be applied.
Before entering into a detailed description of the block diagram in Figure 7-2, there are a number of further design considerations that need reviewing:
|
NITROGEN Figure 7-2. Residue-Based Ammonia Plant |
• Oxygen quality. Considering the fact that the final ammonia synthesis gas contains 25 mol% nitrogen, it is worth reviewing the extent to which this could and should be brought into the system with the oxygen. The energy required for oxygen production shows a flat minimum between about 90% and 95% purity. On the other hand, an increase of nitrogen in the oxygen decreases the cold gas efficiency or yield of H2+CO per kg residue. A detailed study will show an optimum at around 95% 02. This conclusion is also valid for an IGCC application where nitrogen in the gas turbine burner is in any case required for NOx suppression. For many other chemical applications (e. g., methanol), however, this would not be true, since N2 is an unwanted inert in the synthesis loop.
• Steam system. The high-pressure steam must for safety reasons be capable of entering the gasification reactor under all conditions, and it is logical to generate steam in the syngas cooler at the same selected pressure. For our example plant, we will use 100 bar as the saturation pressure of the high-pressure steam and allow a pressure drop of 8 bar across the superheater and controls. Provision will be made for 25-bar medium-pressure steam and 10-bar low-pressure steam.
• Compressor drivers. With a syngas cooler installed after the gasifier, there is usually sufficient high pressure steam to satisfy the demand of the CO shift, the syngas compressor, and the nitrogen compressor. An external energy source is required for the air compressor and (in this case) the nitrogen circulator required for oxygen evaporation in the ASU. In the event of using an oxygen compressor, this would substitute for the latter. An external energy source is also required for a refrigeration compressor. Generally, two alternatives are available, electric power (provided the grid is stable enough to cope with starting a 12 MW electric motor) and steam, which is generated in an auxiliary boiler on site. This is an economic decision, but it should be borne in mind that start-up steam is required in any case. Furthermore, having a substantial boiler in operation all the time can help in stabilizing the overall steam system.
• Refrigeration system. The ammonia synthesis will require refrigeration capacity at typically about 0°C. For final product cooling to atmospheric storage and for the Rectisol acid-gas removal unit, additional refrigeration capacity at about -33°С is required. Ammonia compression and absorption systems are available for this duty. Absorption systems are generally more expensive in capital cost, but they can operate on low-level waste heat that might otherwise remain unrecovered. Studies performed in the context of ammonia plants like our example have shown that it is more economical to use waste heat in boiler feed-water preheat than in absorption refrigeration. On the other hand, if the waste heat really is “waste,” then it pays to invest in an absorption system. Not only have both systems been used in various locations, but they have also been built in combination, where an absorption system was used to bring the refrigerant to 0°C and a booster compressor was used for the -33°С duty.
Looking at the results of this discussion in Figure 7-2 together with the material
balance in Table 7-1 we see the following:
Oxygen and nitrogen are manufactured in the ASU, where the compressors are all driven by condensing steam turbines. The oxygen is pumped in the liquid phase to a pressure of 80 bar and evaporated with gaseous nitrogen, which returns the cryogenic cold to the cold box. The vacuum residue is gasified in the partial oxidation reactor with oxygen and steam at 70 bar and about 1300°C. The raw gas from the reactor contains soot and ash, which is removed in a water wash.
The raw gas, freed of solid matter is cooled down to about -30°C in the Rectisol unit, where it is washed with cold methanol to a residual total sulfur content of less than lOOppbv. The sulfur-free gas is then heated up and saturated with water at about 220°C in a saturator tower in the CO shift. Additional steam is added that reacts over the catalyst with carbon monoxide to form hydrogen and C02. The gas at the outlet of the CO shift has a CO slip of about 3.2% and a C02 content of about 34%. This gas reenters the Rectisol unit and is washed again with cold methanol, this time at about -60°C. The C02 content is reduced to about lOppmv. The resulting gas is a raw hydrogen with about 92% H2 and about 5% CO, the rest being nitrogen, argon, and methane. This gas is cooled down to about -196°C and washed with liquid nitrogen.
Simultaneously, the amount of nitrogen required for the ammonia synthesis is added. The gas is then compressed to the pressure required for the synthesis loop.
The mass balance in Table 7-1, based on gasifying 32t/h visbreaker residue in the partial oxidation unit to produce lOOOt/d ammonia, illustrates this gas treatment scheme. The feed quality is as described in Table 4-10.