Containerized mass-produced miniature ammonia plants using off-the-shelf components for distributed production
Advanced ammonia energy systems
Containerized mass-produced miniature ammonia plants for low-cost ammonia
Christophe Pochari, Pochari Technologies, Bodega Bay, California, USA.
Author email: email@example.com
Author phone number: 707 848 0235
An NH3 synthesizer built in the year 1909. Ammonia started off very small, will it go back to being small?
Every technology introduced, or proposed one for that matter, must have a sound raison d’etre, it cannot simply be proposed and substantiated only because it can be labeled “innovative” or “new”. Technology are tantamount to genomes, or biological replicators, and the market is tantamount to the ecology, every technology is subject to a natural and competitive struggle, where survival of the fittest prevails. It must stand on its own two feet, without government subsidies, and without media hype. A technology cannot be justified merely because of a vague or indirect rationale, be it “climate change” or sustainability. Climate change or sustainability are only buzz words, political ideas with little to no tangible concreteness behind them, they can be comforting to the anxious, but they are not in and of themselves a rationalization for technology development, and they are definitely not a compelling impetus to the inventor, whose sagacity pierces past the fog of propaganda. A technology must have a competitive advantage over its peers, without such it is merely a distraction. Pochari Technologies has conceptualized and developed a formulation for such an idea based on sound reasoning: that miniaturized production can be cheaper and more convenient, and more importantly, it can draw from a unique and idiosyncratic form of energy. We are wholly indifferent to the prevailing attitudes by governments or policymakers, whose ideologically motivated technology and science need not bother the genuine engineer and inventor, who is incessantly searching for more efficient, cheaper, and most importantly, more powerful technologies, purblind to whatever fad or fashion characterizes his epoch.
We envisage a scenario where users of anhydrous ammonia can purchase modular micro plants online from generic manufacturers. These plants can even be built in a kit configuration, where owners can perform the assembly themselves to reduce cost. The cost of these plants will be under $100,000, and potentially as low as $50,000 per unit. They will be shipped to their location in standard I S O containers, connected to a direct current power source, and run continuously for up to 24 hours per day, to meet all their consumption requirements. These plants will have production capacities of 100 to 300 tons per annum, enough to fertilize 1350 acres. This can be achieved all while producing the compound for less than the present trading price of over a thousand per ton. If power costs are below 7 cents per kWh, a price of ammonia below $600 per ton can be achieved.
This paper provides a descriptive overview of what could be achieved if designers are creative and willing to deviate from convention, it is not prescriptive, as there can be no strict “law” on how to design any plant technology, there are many degrees of freedom allotted to the designer, granting him the latitude to make use of a plethora of design options, each tailored to a specific objective, determining where he should channel engineering effort. There are undoubtedly parameters that cannot be exceeded or requirements that cannot be violated or ignored, but there is still immense room to develop technology that can achieve the basic function but that could be arrived at in many different ways. What this paper provides is a vision of what can be achieved if a philosophy of design where creativity and a lack of conservatism is employed.
Performing an extensive cost, material, and fabrication analysis, by contacting numerous suppliers and procuring accurate price quotes, we estimate that small ammonia plants taking advantage of mass-produced largely off-the-shelf components, can achieve capital expenditure below $500 per ton of annual ammonia output.
A number of mature technologies from disparate sectors can be concatenated to make this concept viable. None of these technologies are singularly required for miniaturized production to be realized, but they make it easier and more effective.
These technologies are:
#1 Ionic liquid compressors
#2 Microporous insulation
#3 High activity Wüstite catalysts
#4 High surface area heat exchangers
Through the confluence of these four technologies, reducing the size of the present-day high-pressure ammonia synthesis process to as little as a hundred tons per year is technically viable all while maintaining nearly identical efficiency of the large scale plants, and even reducing the capital expenditure and volumetric footprint of the system.
CAD model of a 300 TPY micro-plant with reactor insulation not shown and without flow circuit optimization
A chemical synthesis system comprises at its heart the active catalytic module, which is usually what is termed a “packed bed reactor”, its name redolent of a gravel-like catalyst solidly loaded into a pressure vessel. Gravel-like catalysts are most often used as they have degrees of stability and strength and do not flake or generate powder which can contaminate reactant products and damage downstream rotary components. The ideal catalyst configuration from a purely catalytic perspective is a fine powder, but a fine powder would be extremely cumbersome to contain and prevented from mixing with the reactant products. The major requirements of a chemical reactor are first and foremost the containment of the reactant products, which often being gaseous or liquid and pressurized, must be prevented from leaking or draining from the vessel. The second requirement is to provide adequate thermal management, since most industrial catalytic reactions do not occur at sufficiently high activities at room temperature, an input of heat, often a great deal, is required to sustain the process. This places unique exigencies on the plant designer, as he must find a way to introduce this heat while effectively and efficiently transferring it into the catalyst containment area. In other cases, the molecular combination of the reactant products emits heat, which must be removed continuously from the reactor, this also places design exigencies, such as designing the reactor with sufficient thermal conductivity, or maintaining an adequate surface to volume ratio. The third requirement of a chemical reactor is housing and stabilizing the catalyst. A catalyst is potentially fragile, which must be prevented from slashing around inside the reactor and incurring damage to its surface morphology. A catalyst is subjected to a fluid, a gas or liquid, which is passing through the catalyst often at high velocity and considerable pressure, if improperly support, the catalyst will move along with the fluid, requiring the designer to place a solid metal grate to contain it from falling into the open ends of the pressure vessel. The catalyst is typically configured in a pellet form, with relatively large gravel stone-like catalyst clumps, which are usually formed from crushing the hardened catalyst product after manufacturing. The term “Packed bed reactors” typically denotes the same thing as a “fixed bed reactor” most industrial catalytic reactors make use of packed bed reactor configurations, including those that employ liquid reactant products.
Depending on whether the reaction is exothermic or endo-thermic, in a fixed/packed bed reactor, the catalyst will either be inserted in a series of tubes fixed inside a larger “mother” container or the entire container itself may house the catalyst in totality. The former is typically employed in endothermic reactions where heat has to be provided, or highly exothermic reactions, where heat must be purged to prevent run-away temperature. The former case includes steam methane reforming while the latter includes the WW2 era Fischer-Tropsch process.
Ammonia is considered only a mildly exothermic reaction, so heat transfer is not a significant design criterion, allowing modern ammonia plants to be designed with a large catalyst basket that occupies the entirety of the reactor vessel, with heat exchanging occurring between catalyst baskets in a stacked configuration.
The overall architecture of ammonia plants has remained unchanged for a century largely attributable to the fact that there is little conceivable reason that anyone could significantly improve upon the process. Most research focus is absorbed by catalyst improvement, which has yielded almost nothing. Little to no research focuses on changing the core reactor technology, as most designers find that there is little presently known to man that could advantage the process. Present ammonia plants make use of enormous radial compressors using high-blade count centrifugal impellers which operate at speeds of below 15,000 rpm connected directly to steam or gas turbines.
In conclusion, there is little hope to perform any changes that are salutary to the technology unless a radically different path is taken. This path we believe is modularization.
The prevailing dogma in the chemical production industry, and especially ammonia, is that bigger is usually better. Increasing the size of the plant is commonly understood to bring reductions in cost through economies of scale. Those proposing to miniaturize plants are violating this rule and thereby subject to a great deal of scrutiny and must demonstrate the viability of down-scaling by marshaling forward evidence and data.
Larger scale plants mean a bigger upfront purchase, giving more bargaining power to negotiate better deals with suppliers on raw material, labor, and other fixed and soft costs alike. For someone to make a credible argument in favor of miniature production, they not only need to back up their claim that this solution is competitive with existing systems but also viable from a financial and safety standpoint.
A number of challenges arise when downscaling and attempting to integrate with spasmodic renewable power.
Firstly, downscaling imposes a thermal penalty, which will be mentioned in greater detail further along in the video.
Secondly, new and improved compressor technologies that can operate at part load will need to be developed.
Thirdly, a novel solution is obliged in order for the Haber-Bosch process to be made computable with a variable output hydrogen source. This will be in the form of an electric heating system, which would be installed in the ammonia catalyst basket, to prevent thermal fatigue damage to the catalyst to attenuate sudden temperature fluctuation caused by cycling the plant on and off, according to the demands imposed by the solar or wind facility. This would entail an electric heater of about 150 watts to be integrated inside the main converter, keeping the catalyst basket at a suitable operating temperature. The electric heater will be made from an electric coil of high resistivity, and suitable to operate in the hot and corrosive ammonia atmosphere. Pochari Technologies has considered patenting this novel design but decided instead to adhere to our open-source philosophy.
The picture above highlights the main ammonia monolith converter, with 12 centimeters of microporous insulation and the electric heating coil. This insulated pressure vessel is then doubled up with a smaller version for the chilling and preheating exchanger forming the entire synthesis loop. For design simplicity, Pochari Technologies’s has chosen a dual converter design.
When the electrolyzer is unable to produce enough hydrogen, the reciprocating compressor is simply shut off, along with the flow of hydrogen and nitrogen. The main converter is closed shut, leaving the singas mixture inside at previous operating pressure. All of the components other than the catalyst would not be sensitive to thermal cycling, as they are merely metallic components that can readily handle sharp thermal gradients. Minimizing this thermal fluctuation would only be needed to prevent negative evolution and modification of the fine crystal structure of the iron oxide catalyst.
Previous efforts at developing miniaturized plants
Despite the fact the first working ammonia plant was a desktop size apparatus, miniature plants have been somewhat of a pariah in the industry.
In 1909, Carl Bosch constructed a miniature desktop size reactor apparatus and successful produced 90 grams of ammonia per hour.
Contrary to popular belief, miniature ammonia reactors are not a novel technology by any means, numerous miniature high pressure reactors have been designed and operated for solar thermochemical energy storage mostly by a team at Australia National University. In 1974, P. O Carden first proposed using ammonia synthesis and decomposition for energy storage. In 1979 Carden published a paper titled Ammonia dissociation for solar thermochemical absorbers in the International Journal of Energy Research where he lays out a concept for designing relatively small reformers paired with synthesis reactors to form a reversible thermal storage system.
Therefore, miniature ammonia synthesis was first proposed for a wholly disparate application, exploiting the reversible ammonia synthesis and decomposition reaction to store energy for concentrated solar power plants, while have little to nothing to do with traditional fertilizer production, highlighting the endurance of the classic production model. While there has not been commercial use of Carden’s concept, considerable research and modeling has occurred since his paper on the subject, many of them highlighting the design exigencies of down scaling ammonia converters.
If one searches Google scholar using the keywords “solar thermochemical ammonia synthesis”, over 16,000 results appear.
In a paper entitled Ammonia synthesis for producing supercritical steam in the context of solar thermochemical energy storage by Chen et al, they construct a tubular insulated reactor using an Inconel pipe with an inside diameter of 18.88mm and an outside diameter of 26.7mm. The reactor could be operated at a pressure of 30 MPa and a temperature of 700°C.
The catalyst used was not specified, but the conversion rate per pass was 16 percent at 30 MPa and 580°C, suggesting they used an older generation of magnetite catalyst.
The reactor built by Chen et al was insulated with 20 cm of mineral wool insulation.
Aside form the above authors, sundry other papers discuss the use of ammonia synthesis for energy storage, but few details on reactor construction and design are provided.
The literature for thermochemical energy storage using ammonia can be drawn upon and serve as a blueprint for prospective designers of small scale plants for fertilizer production, providing the theoretical foundation for miniature production.
The requirements of thermochemical energy storage closely parallel that of small fertilizer plants. Many energy storage applications require very small scales, such as 1 kW or less, such reactors are subject to the same thermal and compression challenges that reactors for distributed fertilizer production would encounter. Research into thermochemical ammonia energy storage can provide the intellectual foundation for a future generation of ammonia plant designers aiding them in transitioning the industry away from centralized production towards a distributed model.
As is well known, ammonia reactors are high-pressure devices and thus must be designed to highly stringent design criteria to ensure the safety of operators and personnel surrounding the vicinity and hazard zone. In spite of this, there is a tendancy to overestimate the difficulty and engineering challenges associated with constructing and maintaining these pressure systems. The maximum pressure encountered in an ammonia plant is approximately 30 MPa, or 4,350 psi, Inconel 625 has a yield strength of 472 MPa at 500 C, a reactor vessel with an inside diameter of 104mm and an outside diameter of 134mm will experience 171 MPa of von mises stress as a operating pressure of 33 MPa, rendering a safety factor of 2.76. Inconel 625’s yield strength only declines 20% from room temperature to 500 C.
FEA analysis using Altair Simsolid for a 13 liter micro-ammonia converter, a maximum stress of 80 MPa is found at 20 MPa of operating pressure, using Inconel 625, the safety factor based on yield strength is 5x. The approximate yield strength at a 0.2% offset for Inconel 625 is 420 MPa. The approximate weight of the above reactor is 250 kg. While Inconel 625 is ideal from a strength and reliability perspective, it is not the most desirable option from a manufacturing perspective due to its high Rockwell hardness. Stainless steel 304 has a sufficiently yield stress at 450 C to warrant its use and its manufacturing is much smoother due to its softer nature. Since welds weaken the morphology of the part, the ideal manufacturing process is to take a solid round bar, bore the center volume similar to barrel rifling, and use a lathe to remove the inner-portion between the two flanges, leaving a solid component free of welds with extremely high strength resulting in stellar safety.
Approximate size of a 13 liter reactor vessel using a an average size adult male as a reference, the reactor pictured above is capable of producing 1.26 tons of ammonia per day at Wüstite catalyst activity levels at 200 bar without any insulation installed, with the reactor wrapped in microporous insulation, an additional 10 centimeters of diameter would be added. As seen by the picture, the flanges are integral with the reactor to form a solid monolith. The images illustrates the relatively small size of the reactor, permitting manufacturing using a large lathe and rifling bore.
The cost at volume for the converter should converge to below $5000 if stainless steel is used, and around $10,000 for Inconel.
In the current industry standard, a 600 TPD plant is considered “small”, the existing high-pressure Haber-Bosch technology claims a scaling factor of 0.64 with respect to plant CAPEX. The dominant cost contributor is the feed compressor, which accounts for over 50% of the CAPEX of some plants.
Truly small-scale production would entail a plant as small as a hundred tons per year, designed to fit inside a 20-foot ISO container. The plant would be designed to make use of only commercially available componentry, minimizing custom fabrication to lower cost. The reactor vessel would take advantage of seamless forged stainless steel or Inconel piping rather than custom fabrication of a reactor vessel.
Besides economies of scale providing cost savings, large plants are also touted to be more efficient, but upon further examination, this doesn’t necessarily need to be the case if design changes are made to minimize changes in thermal flux encountered when down-scaling.
Since the development of the modern “Haber-Bosch” industry, the notion of small plants has been the subject of ridicule in the industry, no serious efforts or successful attempts at designing and manufacturing a small plant have been made to this date. Previous attempts have proven excessively energy intensive or inordinately expensive to be considered viable contenders to the current production paradigm. Prior efforts at bypassing classical high pressure systems such as using electrochemical , plasma, or novel catalysis, have usually been met with failure.
Critics of small-scale production usually lay criticism by adducing the reduced efficiency brought about by down-scaling.
When discussing the “energy efficiency” of an ammonia plant, there must be a clear understanding of what definition is being used. The actual synthesis gas reaction is an exothermic process, releasing 46 kJ/mol, or 0.75 kWh/kg of synthesis gas formed. The energy needed to maintain a synthesis gas operating temperature of 450 C in contrast is only 0.406 kWh/kg of synthesis gas, this means an excess heat of 0.304 kWh/kg is emitted during the formation of 1 kg of ammonia, this excess heat in large plants can be recovered with a steam turbine to cover a substantial portion of the compression energy. Ammonia is a liquid at 150/300 bar at below 130 C, the temperature difference of 320 C from the 450 C operating temperature represents 0.30 kWh/kg-synthesis gas.
Unfortunately, despite the thermophysical calculations suggesting at 300 bar, ammonia is a liquid at 130 C, in the real world, it has to be cooled down further to around 26 C at an operating pressure of 292 bar. For lower pressure systems, the feed gas must be cooled to below ambient temperature. At an operating temperature of 450 C, this corresponds to an hourly heat rejection requirement of 1.4 kWh/kg-NH3.
For a 10 kg-hr plant, a high pressure liquid ammonia cooled heat exchanger approximately with a high specific surface area would reject 60 kWh using 60 5mm diameter non-pressure bearing tubes.
The thermal conductivity of hydrogen at 500 C is over 0.3 W-mK, for a heat exchanger tube size of 5mm, a specific heat flux of 66 kWh/m2 is achieved in a convenient heat exchanger configuration. The specific surface area of the heat exchanger is 159 m2/m3. Since the small-scale plant deviates in many respects from the established design doctrine of mega-scale plants, one distinguishing feature is the heat exchanger and catalyst bed separation. In classic plants, the converter accommodates the heat exchangers as well as the catalyst basket, for a miniature plant, it is more convenient to design the components as separate modules. For the ammonia liquefaction heat exchanger, the cooling and heating exchanger can be made in two separate pressure vessels. A single tubular heat exchanger is used to cool the exit stream of ammonia to the condensation temperature, then the remainder of the un-reacted product is pass through the heating heat exchanger using the warmed thermal medium, this can include hydrogen gas, gaseous ammonia, helium, or liquids such as glycol if residence time is kept very short.
It should be emphasized that since the objective of miniaturized production is to reduce cost as well as to accommodate distributed production, in order to adhere to the philosophy of COTS (commercial off the shelf), the designer should try his best to design the plant around the available parts, this entails doing everything possible in keeping the design features in consonance to the components and supplies that are procurable without excessive customization. This is in contrast to designing it in such an elaborate fashion as to make the plant dictate the type of parts and components required. In other words, since an ammonia plants consist of first and foremost a high pressure compressor (diaphragm type, for helium, oxygen, CO2, argon etc), and then it merely makes use of pressure vessels, threaded fittings, vortex flow meters, the designer is wise to give up design freedom to harmonize it with the availability of mass produced components.
Thermodynamics and kinetics
Using high activity Wüstite catalysts at a space velocity of 30,000 hours at 30 MPa bar, the conversion per pass is as high as 30%, requiring a total of 1 kW/kg to reheat the gas upon cooling and separation 3.3 times to achieve full conversion. In the absence of sophisticated insulation technology which is discussed in detail below, small-scale production suffers an energy penalty due to requiring a substantial portion of this energy budget to be expended to maintain reactor temperature. This could amount to nearly 1-1.5 kWh/kg-NH3, a prohibitive figure which has served to strongly dissuade designers and engineers from exploring the option of down-scale ammonia reactors.
While losses from synthesis gas heating can be considerable for a small uninsulated reactor, the fact remains that the preponderance of the energy consumption of ammonia production is not found in the formation of the product, it is dominated by hydrogen production with compression trailing far behind. Over 90% of the total energy consumed in the process is related to produced hydrogen. Natural gas reforming is a highly endothermic process, requiring 206 kJ/mol of CH4, the reaction is: CH4 + H2O ↔ CO + 3 H2. ∆h = +206.16 kJ/mol CH4, or 10.72 kWh per kg of hydrogen produced, or roughly 32% of the lower heating value of the final hydrogen product.
Electrolysis on the other hand involves a net energy loss, since water is a highly oxidized molecule. The total energy required for water splitting is lower in the vapor phase (241.8 kJ/mol) than in the liquid phase (285.83 kJ/mol). Since commercial electrolyzers almost exclusively split water in the liquid phase, the number of 285.83 kJ/mol is used, this translates to 4.4 kWh/kg H2O, since H2O is 11.19% hydrogen, the minimum energy consumption to split water is 39.32 kWh/kg. This number could be more conveniently arrived at by simply using the lower heating value released from burning hydrogen, producing hydrogen is equivalent to the lower heating value of oxidation to obey the 1st law of thermodynamics.
Since 178 kg of hydrogen is required to produce 1 ton of ammonia (assuming no losses to the atmosphere), the minimum energy consumption is 7000 kWh/ton-NH3. Nitrogen must then be compressed from 1 bar to roughly 1 MPa to facilitate the pressure swing adsorption process to produce highly pure nitrogen. Using the best available data, a 4 stage inter-cooled piston compressor with a mechanical efficiency of 80%, an inter-stage pressure drop of 0.35 bar, and a volumetric flow rate of 1 m/s would consume around 450 kW of power. Since nitrogen is 78% of atmospheric air, 3,510 kg of nitrogen would be compressed to 1 MPa using 450 kW, or around 0.128 kWh/kg N2. For one ton of ammonia, 822 kg of nitrogen is required, so the PSA system would use 98 kWh/t-NH3.
The synthesis gas compression emerges as the second most energy-intensive step in the ammonia synthesis process. For centrifugal compressors installed on large-scale plants, the compressors draw 25,800 kilowatts of power for a 2,200 MPTD plant operating at 200 bar, translating into roughly 0.28 kWh/kg of synthesis gas. In traditional ammonia plants, the hydrogen exits the steam reformer at up to 5 MPa, while the nitrogen is produced from cryogenic distillation can be unlimited since evaporation provides effectively provides free compression, if PSA units are employed, the pressure is usually around 1 MPa.
For small-scale reciprocating compressors, the efficiency is surprisingly close, with the best-in-class reciprocating compressors using around 0.20 kWh/kg according to the calculator below.
Since the principal bulk of energy consumed in ammonia synthesis pertains to the production of the hydrogen feedstock, scale offers little to no opportunity to reduce energy demand other than trivial differences in heat transfer. Therefor we can conclude that there is no intrinsic feature of mass scale production that facilitates a marked improvement in process efficiency if heat transfer is held constant. In fact, large scale production typically makes use of centrifugal compressors, which have lower efficiency than optimized near isothermal reciprocating compressors such as novel ionic liquid compressors mentioned later in greater detail.
To better elucidate the advantages and disadvantages of small-scale production we need to thoroughly investigate how scale affects the efficiency of the actual reaction itself. Since ammonia reactors must sustain high temperatures, a larger reactor will experience a reduced thermal loss to the surrounding environment due to the square-cube law. This obliges the designer of a small-scale plant to employ low-thermal conductivity heat resistant insulation to counterpoise the additional thermal flux resulting from a higher surface to volume inherent to a small reactor vessel. Other than changes in thermal flux, a small penalty will be encountered due to the phenomenon of roughness induced turbulent pressure drop. Since a small plant uses hosing/piping that are smaller in diameter, the roughness of these internal passageways will cause a pressure drop across the circuit,and since the ratio of surface to volume is greater, more gas will be exposed to the rough outer surface and will become turbulent causing its velocity to increase and pressure to decrease. This can be mitigated by using extra-smooth internals, or keeping the GHSV at a lower number. The best way to reduce pressure drop is by increasing the internal volume of the flow-path by using larger diameter hoses and or keeping the distance between compression and final condensation as short as possible, this can be achieved by intelligent placement of components in the module by using the least circuitous orientation.
To minimize thermal losses, Pochari Technologies has designed a novel insulation system utilizing micro-porous insulation that encapsulates the high-temperature ammonia reactor module. Using micro-porous insulation produced by Luyang Energy-Saving Materials Ltd which retails for $5-10/kg, thermal flux can be brought down close to zero depending on the desired thickness.
Microporous insulation, which makes use of miniature silica fibers, can maintain a low thermal conductivity of just 0.035 W-mK at temperatures as high as 800 C, nearly twice what is encountered in a Haber-Bosch reactor. Unifrax (Tonawanda, NY) offers a flexible microporous insulation called “Excelfrax 1800 Flexliner”, at 400 C, the thermal conductivity is only 0.026 W-mK. Using 1 centimeter of insulation wrapped around the vessel, the heat flux for a 10 kg-hr (87 TPY) reactor would only amount to 120 watts, corresponding to 1.7% of the total hourly heat flux. For the 50 kg per hour reactor (250 kg net weight), the hourly heat flux with 10 centimeters of microporous insulation is 140 watts.
With this insulation technology, it’s increasingly difficult to argue that down-scaling is somehow intrinsically involves an energetic penalty.
For a wind-driven hundred-ton per year plant, a reactor vessel with an internal volume of only 5 liters is required using high activity Wüstite catalyst. The surface area of the exterior of the reactor is 0.2 m2, with a total heat flux of 161,000 watts without insulation.
The conclusion is that the argument that heat losses handicap small-scale reactor development is falsified.
A large-scale ammonia converter has a diameter of up to 3 meters, this size reactor has a surface-to-volume ratio of roughly 1.72, the miniature reactor has a surface-to-volume ratio of 35, or roughly 20x higher while the volume of the reactor is 5500x larger. For the large-scale reactor with 50 cubic meters of catalyst volume, the total thermal budget is 113,000 kWh, the heat flux through the reactor wall is 9800 kWh without any insulation.
For the small scale plant, the reactor vessel is ideally constructed out of 304 stainless steel forged pipe using the Mannesmann process. Inconel steel pipe is cut to length and two thick flanges are welded on. The two flanges are held together by 14 threaded rods holding both flanges at each end. The diameter and length of the pipe is chosen based on the volumetric capacity of catalyst that is required to produce the desired amount of ammonia hourly.
For a ten kilogram micro plant, around 30 kg of SS304 would be required at a material cost of $105 assuming a price for SS304 of $3.5/kg.
The excessive cost of hydrocarbon-based ammonia
The ammonia industry is in need of disruption in order to adapt to a new environment of increasing renewable penetration. In today’s ammonia industry, large producers have monopolized the market and charge far in excess of what natural gas’s cost as the principle feed stock would predict. For example, as of November February 2022, the price of anhydrous ammonia is $1500/ton, while the price of natural gas is only $5/thousand-cubic feet, or roughly $0.26/kg. There is approximately 19 kg of natural gas present in a 1000 cubic feet or a million BTU of natural gas. The price of producing hydrogen from natural gas is only $1-1.5/kg, approximately 3 kg of methane is required to produce 1 kg of hydrogen using steam methane reforming, therefore, the price of ammonia should be still below $500/ton.
A CAPEX for a “greenfield” Haber-Bosch plant of <$1500/TPY is typical, amortized over as little as 15 years yields less than $100/ton. There is very little information available on the internet on the typical lifespan of an ammonia plant, but the available evidence seems to suggest the lifespan of the Haber-Bosch plant is impressive. Phoenix Equipment in New Jersey acts as a broker for chemical plants online, of the 9 anhydrous ammonia plants currently for sale on their website, the average age of the plants is 43 years old, with many still in operating condition. Ammonia plants feature only one dynamic system, that is a system with moving parts: the synthesis gas compressor.
The rest of the plant consists of pressure vessels, pipes, valves, and heat exchangers. These components are not subject to dynamic stresses, but rather thermal and corrosive stresses. With proper material selection, corrosion and thermal stresses can be mitigated afforded a long lived plant.
While the labor intensity of an anhydrous ammonia plant is substantial, especially for very large plants, the net labor contribution to production cost is relatively minor. According to a 2013 thesis, the total labor intensity of anhydrous ammonia plants translates into around one man hour per ton. In a developed country, one man hour could be over $30-50/hr, adding a non-trivial cost to production of at least $30/ton.
Distributed plants, by virtue of their simplicity and their high automation potential reduce labor costs to effectively zero.
Hydrogen production costs from methane:
Methane contains 22% hydrogen by mass, the raw stochiometric ratio would predict 4.54 kg of methane is required to produce 1 kg of hydrogen. But since the carbon monoxide is converted into additional hydrogen through the water gas shift reaction, additional “free” hydrogen is extracted with only water vapor needed as an energy source.
For reference, Kapsom Industrial Solutions in Nanjing built a 600,000 TPY plant for $650,000,000, this includes hydrogen gas production and purification, which translates into CAPEX of $1083/TPY, amortized over 20 years equates to $54/ton, including 5% annual maintenance, adding another $54/ton. Since the facility is in Asia, labor costs are expected to be somewhat lower.
Existing ammonia plants often include a steam methane reformer in their construction, and a thus a considerable portion of the maintenance cost is related to the steam methane reforming system.
From these numbers, a fair estimate is $75/ton CAPEX, ($1500/TPY 20 yr period), 5% annual CAPEX allocated for maintenance, yielding another $75, plus $35 for labor, yielding a total of $185, less than 14% the cost of the current anhydrous spot price. This is congruent with findings that in 1999, Russian ammonia prices fell to as little as $50.8/ton according to CEIC data, suggesting that CAPEX is actually a much smaller contributor than estimated from construction costs. Russian ammonia prices remained below $200/ton from 1999 to 2004, with natural gas prices averaging $4/MBTU during this same period, strongly suggesting CAPEX must be below $50/ton for production to have been sustainable at $200/ton with natural gas prices of $4/MBTU. Our preliminary CAPEX estimates may be too conservative, the data from Russia suggests our estimates are substantially higher than real-world production cost data, this incongruity can be reconciled by increasing the useful life the plants.
Most Russian plants are quite old, many were built during the Soviet Union and remain operational today, the plants have been fully paid-off, with maintenance and personnel costs dominating the OPEX portion of production costs.
Steam methane reformer CAPEX
Steam methane reformer technology, much like alkaline electrolysis, is dependent to a large extent on the market price of alloys with a high nickel fraction, such Inconel, Incoloy, Hastelloy, and stainless steel. Steam methane reformers may operate as high as 1000 C and use nickel based catalysts usually supported on alumina. HK-40, 22% Ni, is the industry standard alloy. SMR reformer tubes are typically centrifugally cast as opposed to forged due to specific grain structure optimization requirements.
From an engineering perspective, our micro wind/solar driven plant must not only successfully downscale the existing Haber-Bosch process, but it must also demonstrate a successful alternative to steam methane reforming, one of the most mature chemical plant technologies.
A fair comparison between methane fueled large scale Haber-Bosch with electrolysis fueled downscale Haber-Bosch is required.
In order to perform this comparison, data on the CAPEX of steam methane reformers is required.
Chengdu Tcwy New Energy Technology Co sells a 50,000 Nm3/hr steam methane reformer for $50,000,000, the price per kg over a 15 years period would be only $0.093/kg of hydrogen produced. In February 2020, Air Products began construction of a $500 million hydrogen production facility in Saudi Arabia with a production capacity of 413,525 kg of hydrogen per day, yielding a per kg cost of $0.22 over fifteen years. Adding maintenance at 5% of initial CAPEX, an additional $0.16 can be added. In 2016, Air Products built a 289,000 kg per day plant for $400 million, yielding $0.25/kg over fifteen years. With this estimate of benchmark OPEX and CAPEX, with a natural gas price of roughly $0.30/kg, $0.90/kg would be the raw material cost for the reformer, with another $0.30 for CAPEX, and another $0.16-0.25 for OPEX, yielding a hydrogen price of <$1.5-2/kg, or $267/ton of ammonia. Are these numbers realistic?
In 2019 S&P Global Platts estimated the average bulk resale value of hydrogen was $0.79/kg based on the price of methane reforming from the Netherlands with a purity of 99.9%.
Since we have validated our crude cost estimates based on adducing the price of the raw material, and the maintenance and acquisition cost of the plant, we can be more confident that the numbers used to compare with alkaline solar/wind electrolysis are realistic market prices.
Projected uses of ammonia
Global ammonia production capacity was estimated to be 235 million tons in 2019, which is projected to rise to 289 million tons by 2030.
To summarize, ammonia prices remain excessively expensive despite relatively low natural gas prices. The U.S “Henry hub” price of $5/MBTU pales in comparison to European prices of $40/MBTU. Despite U.S natural gas prices being at a 12 year high, natural gas is still far cheaper than all hydrocarbons except for coal. Per kg, oil at $80/bl is $0.57/kg, the current spot price for coal is around $150/ton, which is based on the GC Newcastle coal futures market.
For the fertilizer market, this elevated price is tolerable as the cost contribution of fertilizer to the total cost of producing the major agricultural products, such as soybeans, wheat, and corn is only roughly 10-15%, and this is including potassium and phosphorus fertilizers, not just nitrogen. What this illustrates is the ammonia market has relative price elasticity since farmers can easily absorb higher prices as they simply pass on the cost to the end-consumer.
While agriculture currently accounts for the vast majority of ammonia consumption, the future demand for ammonia may increasingly shift to fuel and energy use as the world tries to find substitutes for hydrocarbons in large part to reduce CO2 emissions as well as securing the long term sustainability of energy intensive society.
There has been much speculation in the small ammonia energy community that this caustic fertilizer would someday join the club of liquid hydrocarbons and perform the heavy lifting as the fuel of the future to power ships, heavy-duty vehicles, and provide an energy-dense storage medium to transport energy long distances by tanker.
Notwithstanding the fact that currently as no application as a fuel outside of highly niche applications, one of the only known applications of ammonia as a commercial fuel is to power alkaline fuel cells to generate electricity for rural cell towers especially in South Africa. Ammonia is chosen not based on intrinsic energetic, emission, and overall performance characteristics as a fuel, but for an extraneous reason: thieves do not dare steal it for fear of the caustic gas!
The theft of rural cell tower fuel is an endemic issue in certain parts of Africa, engineers have forced to employ unconventional tactics to combat it, one of them is by using a rather foul fuel that people dare not approach or have little use for.
In the cell tower market, Companies like GenCell in Israel market a nickel based reformer and alkaline fuel cell package.
For ammonia to become competitive with liquid hydrocarbons, it must first reach price parity with existing fuels. This may seem strange considering current ammonia is produced from these very hydrocarbons, but these paradox illustrates the unique market conditions in the ammonia industry which contribute to the commodities price premium over its feedstock.
The Herfindahl-Hirschman (HHI) index is commonly used to evaluate the degree of industrial concentration of a given sector, ammonia ranks at 2378, with three producers, Koch, Yara, and CF Industries claiming 75% of the entire market, leaving little room for new entrants. A Herfindahl-Hirschman index of over 2,500 is considered very high by the Department of Justice and may prompt anti-thrust investigation.
It appears the lumber and ammonia industry are case studies of a high degree of industrial consolidation leading to non-competitive behavior.
The prospects of ammonia falling in price back to pre-Covid levels is not certain.
In a paper titled “Market concentration and pricing behavior in the fertilizer industry: a global approach” by Manuel A. Hernandez, the ammonia industry’s level of industrial concentration is analyzed and the conclusion is drawn that the industry is far from perfection equilibrium competition where products sell at production cost plus a small margin.
On an energy basis, an anhydrous ammonia price of $700/ton is equivalent to a gasoline price of $4/gallon. While this may not seem that high considering in Europe the average price of gasoline and diesel is over $5/gallon, in most parts of the world without high fuel taxes, gasoline and diesel is presently around $3/gallon at current <$100 oil prices, requiring ammonia to be priced at $500/ton or less. Since ammonia is an energy dense fuel, its use will be more attractive for heavy-duty vehicle applications.
A possible major use will also be energy storage and rural power generation to replace diesel generators.
At the present time, the price of >$1500 per ton is prohibitive for energy storage or transportation, suggesting that its use as a fuel is far from guaranteed unless ammonia producers are willing to sell the commodity at cost.
For electrolysis based ammonia, the costs are usually much higher than steam reforming of natural gas, since in most parts of the world, commercial electricity is generated for at least 10 cents per kWh. Moreover, the CAPEX of present day alkaline electrolysis units is in excess of a $1000/kW.
In order for ammonia to compete with hydrocarbon fuels, the hydrogen produced from the electrolyzer should ideally be under $2/kg, or around $350-400/ton. Thankfully, this number is readily achievable as long as electrolyzer plants can be produced closer to material costs.
For the wind/solar driven plant, the CAPEX of the electrolyzer stack and its ancillary components has to be included in the total CAPEX of the ammonia plant. The cost of pressure-swing absorption nitrogen production also has to be included, while the initial CAPEX of a PSA system can be substantial, the levelized production cost per kg of nitrogen is negligible.
PSA and compressor system costs
When synthesis gas production cost is excluded, the single biggest contributor is the compression system, both in terms of operation costs, energy consumption, and in the initial CAPEX of the compressor set. The current market price of PSA nitrogen generators is around $43/TPY-NH3. For a 100 Nm3/hr unit, the selling price is $57,900 for a new Chinese made unit, prices for Western units were not sourced. The price was ascertained from a RFQ on Alibaba.com from Suzhou Hengda Purification Equipment Co Ltd. For a smaller-scale nitrogen generator, YUCHENGTECH offers a 3 Nm3/hr 99.99% purity food-grade nitrogen generator for $2600, this would translate into roughly $66/ton for the first year of production. Note that a PSA nitrogen generator is simply a container consisting of the carbon molecular sieve with an air compressor to pump air in and out of the tank. Chizhou Shanli Molecular Sieve Co Ltd produces the SLUHP-100 molecular sieve with a purity of 99.999% at a price of $9.5/kg. The productivity of the this particular sieve is 95 cubic meters of nitrogen per ton of sieve per hour. For a 100 ton per year plant, around 0.37 cubic meter of sieve would be required, with a density of 650-690 kg/m3, around 250 kg of sieve would be required at a cost of only $2,400.
Compressor CAPEX was inferred by sampling the selling price of diaphragm compressors, and used to roughly approximate the future cost of mass-produced ionic liquid compressors. Since ionic liquid compressors are not currently manufactured by more than one company, one cannot ascertain a realistic CAPEX figure. Extrapolation from existing compressor architecture can provide a more realistic and conservative estimation without overestimating the price. Since ionic liquid compressors are only manufactured by Linde for hydrogen fueling stations, no cost estimate could be ascertained. Ionic liquid compression remains a highly niche technology and likely enormously expensive.
The major challenge for synthesis gas compression is preventing lube oil contamination which can downstream catalyst failure. Under ideal circumstances, cheap mass-produced oil-lubricated compressors could be used with coalescers installed, which have a proven track record in numerous industrial applications. Oil mist coalescers can remove lube oil particles in the sub-micron range down to 0.01 PPM levels, but they are unable to filter out vapor, so most compressors use synthetic oils with minimal to no vapor pressure. Outlet concentrations for these high-efficiency liquid/gas coalescers are as low as 0.003 ppm (weight). Pall Corporation markets the Pall SepraSol™ coalescer filter for gas oil mist separation. The coalescer felt assembly can be placed in a pressure vessel, the high pressure singas mixture leaving the compressor is then pumped in the coalescer module ideally at lower velocities, as this makes coalescing more effectual, the cleaned gas can then be passed into the main ammonia converter.
Wool felt is able to filter out 100% of micron size oil particles, wool felt is oleophilic, making it highly effective for oil retention. Charcol-filled canisters use used to absorb oil vapor in most industrial gas compression systems. If such an option proves impractical for whatever reason (excessive oil accumulation), diaphragm compressors can be used.
Non-diaphragm compressors, that is reciprocating oil-lubricated compressors are very low cost devices. For example, Shanghai Guoxia Compressor Co., Ltd manufactures a 150 bar 15.9 Nm3/hr compressor for ¥17890 ($2780) per unit. This compressor is the GSW265B model, it’s a 3 piston compressor, with a flow rate of 265 liters per minute, it uses a 380 volt 50 Hz 5.5 KW electric drive motor. The compressor emits 73 dB sound during normal operation and weighs only 160 kg, it’s approximate size is 94 x 70 cm x 54 cm, the RPM is 1500. At it’s rated power consumption, compressing the hydrogen from the electrolyzer at 15 bar to 150 bar would consume only 0.28 kW/kg, nitrogen from atmospheric pressure to 150 would be 0.27 kWh/kg. For one ton of ammonia to 150 bar, only 277 kWh of electric power is needed. To illustrate the minuscule cost of the oil-lubricated compressor, the flow rate of this $2800 compressor would permit a production rate of 168 tons per year, costing only $16/TPY. For non-oil lubricated reciprocating compressors that make use of molybdenum disulfide filled PTFE seals (piston rings) the lifespan is usually 4000-6000 hours
For diaphragm technology, costs per flow rate are typically much higher. The reason for this is that the diaphragm compressor has a much smaller working volume since the maximum volume is limited by the movement of the flexible plastic diaphragm membrane as opposed to the entire cylinder in a standard compressor. This means for equivalent compressor weight, the diaphragm system has a much lower volumetric flow capacity. Despite this, at larger scales, diaphragm compressors still account for negligible CAPEX contribution.
The source for the diaphragm compressor price was chosen based on a quote provided by Sichuan New Tianyuan Technologies Co., Ltd, a prominent manufacture of diaphragm gas compressors in China. One of their medium-flow models is the GL2-75/10-350, the intake pressure is 10 bar, the discharge/outlet pressure is 350 bar, and the flow rate is 75 Nm3/hr. The compressor is equipped with an explosion-proof electric motor rated for 30kW, it also features water cooling with water consumption in the closed-loop circulation cooling system of 2.5 tons per hour. The estimated outer dimensions of the unit is 1950 x 1400 x 1350 mm, the estimated curb weight is 1500kg. The unit price is a $39,600 per unit, the price was established through communication with the manufacturer directly, accurate price quotes are rarely available online for diaphragm compressors due to their specialized nature. The flow rate at 1 MPa syngas inlet pressure with an average density of 10 kg/m3 corresponds to a CAPEX of only $6/TPY. The CAPEX is highly sensitive to the inlet pressure, for example, if the inlet pressure was atmospheric, the compressor would only be able to compressor 78 kg/hr of syngas, since 1 kg of NH3 syngas at a stochiometric ratio of 0.82 kg N2 and 0.178 kg H2 has an average density of 1.04 kg/m3. The pressure of the PSA unit is around 1 MPa, and since the electrolyzer can operate as high as 30 bar, the plant designer can leverage the elevated inlet pressure to maximize the productivity of the more expensive diaphragm compressor.
Diaphragm compressors can be categorized into two sections, the life limited diaphragm and the long-lasting hydraulic piston section. The hydraulic section consists of a cylinder with a piston which pressurizes hydraulic fluid to drive the metal diaphragm which compresses the gas on the other end. The standard life span for diaphragm compressors diaphragms in MTBO (mean time before overhaul), is limited to around 5000 hours for standard metal diaphragm, usually constructed out of multiple layers of stainless steel sheet for maximum suppleness. The diaphragm sheets are subject to intense fatigue stresses due to the cumulative stress of the sheet bending on the periphery of its mounting perimeter flange. The diaphragm sheets eventually fail right on the outskirts of the flange clamping location. Failure usually begins with micro-cracks, which can cause oil-gas leakage and downstream contamination of warning sensors are not installed. A diaphragm compressor is the worst possible design from a longevity and maintenance standpoint other than that it is the only oil free and high purity compression architecture available until ionic-liquid supersedes it if coalescers and their attendant hassle are not desired.
PDC machines , a much higher end U.S manufacturer, claims a life of 6000 to up to 8000 hours for continuous use. The main determining factor in the life of the diaphragm is the pressure differential between the gas and the hydraulic fluid, if the differential is small or non-existent, diaphragm life can be extended significantly. Since the diaphragm itself only contributes a small portion of the overhaul cost, most of the life-limited components would be expected to have a similar longevity profile as an oil-lubricated compressor, and especially since the piston and cylinder are submerged in hydraulic fluid, one could argue they would experience considerably reduced friction compared to conventional compressors whose piston is often oil-free and exposed to pressurized and corrosive gases. For the this analysis, for the sake of simplicity, we assign a MTBO of 4000 hours for a conservative estimate. We also assume the overhaul consists of a piston ring replacement, check valve, and connecting rod bushing replacement at 10,000 hours. We also assume the overhaul is equal to 50% of the new manufacturing cost, this adds $3/TPY-NH3. Personal conversation with David Yang from Sichuan New Tianyuan Technologies indicated MBTO of 5-8 years depending on working environment, as humidity and salt air can reduce this number dramatically. Beijing Zhongding Hengsheng Gas Equipment Co., Ltd advertsizes their diaphragms to last 5000 hours and the main air valve, which are made by Western firms listed as Dott. Ing. Mario Cozzani Srl or HOERBIGER, is rated for 8000 hours before replacement or overhaul is due.
For more detailed information on electrolyzer capital costs, click here https://pocharitechnologies.com/2021/06/15/reduced-capex-alkaline-electrolyzers-using-commercial-off-the-shelf-component-cots-design-philosophy/
While current electrolyzer stacks from European companies such as Proton, Nel, Hydrogenics, ITM power, Hyprovide BV, etc usually feature stack prices in excess of $500/kW, advanced COTS manufacturing strategies can bring this down to close to the raw-material minimum. An electrolyzer price of $100/kW is chosen for this analysis. In 2006, General Electric estimated at high production volumes, an alkaline electrolyzer could be built $100/kW, very close to the cost of manufacturing lithium ion batteries, which use far more expensive raw materials per kW of capacity. Current electrolyzer costs in China are already below $200/kW according to Bloomberg energy.
At a $100/kW with a LHV efficiency of 91% (available at current densities below 200 miliamps/cm2), translates to $4400/kg-H2/hr. Thus, the CAPEX is $90/TPY-NH3. Using the current electrolyzer prices in China of $200/kW, this increases to $180.
The total syngas production portion of CAPEX amounts to $200/TPY, this excludes the ammonia converter, high-pressure heat exchanger, main catalyst, and syngas purification unit. The total for these respective systems is in the range of $50-100. Ammonia synthesis catalysts can be purchased online, including a large catalyst manufacturer called Lianing Haitai Sci-Tech Development Co., Ltd. They sell the catalyst, a traditional lower-yield magnetite based product, for around $15/kg for bulk orders such as for a plant catalyst replacement, and for small orders, they are willing to sell as little as 1 kg for $200.
Solar farm CAPEX
The economic competitiveness of photovoltaic and wind-sourced ammonia will become increasingly evident as hydrocarbon prices rise. Coal accounts for 80% of ammonia produced in China, as of the month of August 2021, prices for Chinese coal reached a record high of $160/ton, the previous high was $130 in 2010. Approximately 3 kg of coal is required to produce 1 kg of hydrogen, so at the current price of coal in China, hydrogen should cost $0.45/kg excluding gasifier, desulfurization, and various ancillary equipment CAPEX. Ammonia’s hydrogen cost is thus $216/ton. For the U.S market, in 2021, natural gas prices nearly doubled from a low of $2.5/MBTU in April to $5/MBTUin November. U.S natural gas is significantly more expensive than Chinese coal.
With production from inexhaustible water and air, distributed production becomes feasible, overturning the current business model away from centralized production and improving resilience against volatile hydrocarbon energy markets. Distributed solar ammonia can be thought of as a hedge against the uncertain future of conventional energy sources.
The beauty of this technology is instead of being dependent on an inherently volatile commodity (natural gas), which for the most part, is an exhaustible resource, one can pivot towards reliance on an inexhaustible resource: polysilicon. Hydrocarbons are a type of resource than conform to a “Hubbert’s peak” model, that is they are expected gradually increase in extraction difficulty over time. In contrast, distributed production is only reliant on polysilicon as a commodity, which will continue to go down in price with increased production since silica is effectively inexhaustible, around 46% of the earth’s crust.
If course, polysilicon requires energy, of which most presently derives from hydrocarbons, and more importantly, silica must be reduced with carbon to generate pure silicon metal. Ultimately, silica will have to be produced from carbon that is generated in a closed loop cycle through plasma decomposition of methane yielding carbon black, which can be produced from carbon dioxide and hydrogen. At the present moment, such a closed carbon cycle remains impractical as coal remains abundant.
To produce silicon metal from silica sand, around 1 ton of carbon is required per ton of silicon. The reaction for carbothermal reduction of silica is: SiO2 + 2C → 2CO + Si
Since silicon has a molar mass of 28 and carbon 12, around 1.16 tons of silicon could be yielded under ideal conditions.
The carbothermal reduction of silica uses a heavy carbon electrode to heat the silica and carbon mixture 1500 and 1700 C.
The specific energy inputs for the process are in the order of 12-20 kWh/t-Si, amounting to $900/ton at a power cost of 6 cents per kWh.
Once the low purity metallic silicon is produced, it has to be further purified using the Siemens process where it is converted to silane and subsequently decomposed, consuming an additional 100 kWh/kg.
In 2009, the Siemens process reactors consumed 200 kWh/kg, this number has dropped to about 80 and is expected to fall below 70, with optimized heat recovery, the number could be reduced to 43 kWh/kg.
With this data, we can roughly calculated the EROI (energy return on energy invested) of a photovoltaic plant.
Electricity consumption in cutting and placing the solar cells on the backboard are minimal, the bulk of energy consumption is found in the trichlorosilane cycle.
From the above chemical formula, we conclude we need about 0.9 tons of carbon per ton of polysilicon. Since coal is 90% carbon, the actual tonnage is increased to just a little over one.
A typical monocrystalline silicon wafer used in a photovoltaic panel is 180 microns thick and 156x156mm. The power density of a monocrystalline panel is roughly 200 watts per square meter, so five square meters are required for a kilowatt.
The total amount of polysilicon for a 1 kilowatt panel is thus 2 kg, requiring 2 kg of coal for reduction.
From silica to polysilicon, the entire process consumed 120 kWh for every kilogram produced, to be conservative, an additional 50-100 kW is assumed to be required for fabrication of the panels and their sub-components. This includes producing the glass for the cover and the aluminum frame.
The EROI can be simply calculated by diving the panel’s yearly power output by the per kilowatt embedded energy.
Real world data seem to indicate the energy required to produce solar panels is considerably greater than the silica reduction and Siemens reactor process alone indicate. A few estimates place the number as high as 10,000 kWh/kWp for mono-crystalline panels dated back to 2009, this would mean the panel would have to produce power for over 7 years in a high insolation geography just to pay off the initial energy investment.
Note that according to the above USDA statistic for April 30 2022, there are effectively no sellers offering a price below $1500/ton, so the standard deviation (SD) in offer prices is very narrow. This means it’s unlikely for some farmers to be able to snatch up good deals even if they are savvy buyers. Large purchase volumes may permit a slight discount to be had, but this commodity is relativity inflexible and globally uniform.
The incongruity between ammonia’s selling price and its constituent raw material feedstock value can be explained by Porter’s 5 forces, producers of a value-added chemical have more bargaining power than their constituent customers since the demand elasticity is very low. Note that the 2020-2022 North American lumber price crisis illustrates this exact phenomenon. For high barrier to entry production sectors, producers retain nearly all bargaining power. Only distributed production can break this monopoly. Another important factor to note is that most commodities have a fixed production/extraction cost in addition to the prevailing market value. Value and cost can diverge significantly. For example, in 2008, polysilicon sold for $400/kg, while in 2020, it was less than $20/kg, a factor of 20x difference without any underlying change in the technological or material basis of the process. To summarize, just because a synthetic chemical sells for X price does not in any indicate it cost X price to produce or manufacture, many other variables are at play, and often, market failures are to blame. Technology, proper implemented, can sometimes help rectify these market failures or imbalances. Distributed production using solar and wind, if properly designed, can offer a more competitive production system.
With the availability of low-cost photovoltaic energy without the need for balancing grid demand fluctuation, the need for storage is eliminated, this opens up the possibility of producing ultra-low cost hydrogen even below the price of methane reforming. Photovoltaic technology is chosen not necessary for its “greenness” or lack of carbon dioxide emissions per se, it’s chosen because photovoltaic technology is an inherently superior form of power generation technology than even ultra-high energy-dense sources such as nuclear. Despite uranium fission being 2000x more energy density than solar, the price of pressurized water reactors in the United States is nearly $7000/kw and about $3000/kw in China. Amortized over 40 years, the LCOE is 2 cents/kWh, excluding maintenance and fueling.
For a photovoltaic array in a high-irradiance geography at the current CAPEX of $300/kW for Chinese made modules, the LCOE over a 20 year life is 0.83 cents/kWh, not including maintenance and degradation. The actual life-span of a solar installation in a benign environment free of hurricanes, hail, and severe winds is likely in excess of 25 years. This number is quite astounding, the fact that harmless, easy to install and manufacture solar energy is equal if not cheaper than the most technologically advanced and energy dense form of energy available on earth: that is nuclear fission with moderated light-water reactor architecture.
Anticipated usage profile of distributed ammonia production.
Until ammonia begins its use as a low-emission fuel in the near future, the fertilizer market is the biggest opportunity for disruption. The size of the ammonia industry is nearly $100 billion annually, with production volumes as high as 200 million tons, nearly 20% of the total production of crude oil in comparison.
Low-cost photovoltaics have completely transformed the energy dynamic. Global benchmark capital expenditure (CAPEX) for utility-scale solar photovoltaics (PV) has been decreasing over the years. According to Madhumitha Jaganmohan, a researcher at Statista, between 2010 and 2020, utility scale PV costs decreased by 2.92 U.S. dollars per watt. In 2020, the CAPEX for utility-scale solar PV reached a low 0.61 U.S. dollars per watt. DNV GL, a Danish accredited registrar and classification society predicts utility scale PV costs to fall to $0.45/watt in 2030. Assuming this module was installed in a geography featuring high annual irradiance, for example, 1700-1800 kWh/kWp, which is achievable in many propitious geographies, as much 2030 kWh/kWp available in Lancaster California. At an optimistic irradiance, once could expect nearly 1,800,000 kWh per year amortized over 15 years minus 0.8%/anum degradation yielding a price per kWh-DC of less than 2.2 cents, congruous with findings of major solar projects around the world under producing power for below 3 cents per kWh.
While some Chinese solar plants can had for 30 cents per watt, a more realistic estimate is 60 cents per watt installed.
Since the electrolyzer uses DC power, up to 10% of the power can be saved since AC rectification is redundant, DC/AC inverters are rarely over 90% efficient. A conservative estimate of 1600 kWh/kWp is used for North America, at this irradiance level, for $500/kW, the cost per TPY is $2600.
Most hydrogen experts agree that below or at this price, green hydrogen becomes very attractive provided electrolyzer CAPEX can be further reduced, we have concluded using COTS electrolyzer technology CAPEX’s below $50/kW are achievable. Assuming an aggressive levelized solar price of 2 cents and electrolyzer power consumption of 4 kWh/m3, the cost to produce 1 kg of hydrogen is only 72 cents, or $128 per ton of ammonia. Note that photovolatic prices along with electrolyzer prices will continue to fall since their raw material feedstock is not in short supply, polysilicon is still trading far above the theoretical production costs, and alkaline water electrolyzers are 100x more expensive than bare material costs due to low production volumes.
Wind power for distributed ammonia
No technology is a panacea, in geographies limited by poor irradiance, a greater reliance on wind will be required. In those geographies that have both poor wind and insolation, transportation from more propitious geographies is unfortunately still required unless higher production costs can be tolerated. A large pipeline infrastructure already exists, transporting ammonia by pipeline is virtually free. Where pipelines are not accessible, short-distance trucking can be acceptable. The point of distributed production is not necessarily just to minimize transport, the point of distributed production is to design ammonia plants that can be more easily installed and integrated in any geography that is rich in wind and solar resources. Distributed production is designed as a convenient, modular, transportable package that can be connected directly to wind and solar farms. Transportation from these sites to the final end-use destination still has to be performed in some cases. For example, imagine a single 120 kW turbine located atop a hill where the wind velocity is highest, the 20-foot ammonia-plant container is then located beneath it to yield 60 tons a year, enough to cover 550 acres at 200 lbs of nitrogen per acre. Existing ammonia technology requires a massive wind or solar farm to be built to provide the gargantuan energy needed to fuel a 1000 TPD plant for example.
Hydrocarbon-based modern ammonia plants are getting larger and larger, with some pushing as high as 3,300 TPD. This type of plant is far too large to be used by even by the biggest farmers and is not practical to construct in very rural geographies where many wind or solar farms may be located.
Medium-sized wind turbines can offer a competitive solution for regions with wind speeds greater than 6 meters per second.
The Hebei Hopevwin New Energy Technology Co Ltd turbine model FD42-400 rated for 400 kW produces 1,450,000 kWh per year at a wind speed of 8 m/s, and 1,069,000 at 6.5 m/s. This turbines weighs 25,000 kg, has a tower height of 50 meters, and is built to the IEC IIIA design class, the cost is approximately $495,000 per piece. The advertised life of the turbine is 20 years and its extreme wind rating is 52 m/s.
In high wind geographies, for example, Cherry and Grant County Nebraska, the 50th percentile wind velocity at 50 meters of altitude is 8.2 meters per second.
A smaller version by the same manufacturer, the FD25-120 120 kW turbines sells for $180,000 and produces 492,000 kWh at 8 m/s and 374,000 kWh at 6.5 m/s.
Amortized over 15 years with an average wind speed in these counties of Nebraska, the turbine would produce a levelized cost of electricity of $0.024/kWh, or 2.4 cents, yielding an ammonia price of $192/ton.
In a less windy region where speeds are in the order of 6-6.5 m/s, for example, Iroquois County, Illinois with a mean of 6.4 m/s, the levelized cost of power would be 3.2 cents per kWh, the cost per ton of ammonia would thus be $257. If the wind speed is to fall to 5 meters per second the LCOE would reach 5.5 cents per kWh, or $500 per ton of ammonia. Wind speeds of 5 meters per second at 50 meters of height are available throughout most of the U.S Midwest.
Previous efforts to develop distributed ammonia production technology
Aside from Pochari Technologies, a few other experts has realized the potential of small-scale ammonia production using advanced technology. Richard Strait of Rice University has given a talk on his concept for the miniaturization of ammonia production.
“The small self-contained ammonia manufacturing box (AMB) came out of the need for farmers in remote locations containing modest amounts of natural gas and water to produce enough fertilizer for their local needs. Presently ammonia plants are getting larger and larger. The future however, is to make small highly automated plants that are portable, use readily available inputs and produce the products where they are needed. To this end the new small-scale plant will include all the latest leading edge technologies that apply such as micro processing technology (MPT), process intensification (PI) and automated safety systems. The ammonia plant of the future is contained on one rail car, shop fabricated and contains all the latest technology. To date the “economy of scale” has been the way to reduce the cost of an ammonia or methanol plant but we are at or near the limit of what is possible. Moreover very few companies need such large plants with their large price tags. The future is to make small highly automated plants that are portable, use readily available inputs and produce the products where they are needed. To this end the new small scale plant will include all the latest leading edge technologies that apply such as Micro Processing Technology (MPT), Process Intensification (PI) and automated safety systems. The ammonia plant of the future is contained on one rail car, shop fabricated and containing all the latest technology. The methanol plant of the future could also be a standalone railcar. The DME plant of the future could also be a standalone railcar”
See: Ammonia Manufacturing Box: Small Ammonia Plant Design Using Micro Processing Technology and Process Intensification Technology
Pochari Technologies views the present situation in a similar manner. We differ in that we are confident that rather than methane reforming, photovoltaic-based hydrogen production will actually surpass methane in cost-competitiveness.
The important question is why downscale?
When one begins inquiring into the possibility of down-scaling, a surprising number of benefits appear, often seemingly contradictory to established industry dogma, what you could call “antinomies”.
#1 Ability to integrate with intermittent renewables that are often sparsely scattered across diverse and rural geographies.
Additionally, if containerized ammonia plants are built at wind or solar farms, significant cost savings can be realized by dispensing with grid connections and long-distance power lines.
#2 Potential to reduce CAPEX with the use of ubiquitous components.
Large is not always better, scaling presents unique manufacturing and installation exigencies, components become more difficult and technically demanding to fabricate raising costs, leaving a smaller pool of suppliers to choose from. A small-scale ammonia plant be easily assembled using parts (reactor tubes, fittings, valves etc) purchased online from Alibaba.com for wholesale prices.
#3 Breaking the monopoly of centralized legacy producers, Yara, Koch, CF, etc
This enables farmers to produce ammonia for less than the price of current commercially acquired anhydrous from large suppliers. It also provides the agriculture industry with the ability to hedge against future price uncertainty as well as provide the future option to reduce emissions if necessary.
Small scale containerized plants can be moved easily by truck, increasing their long-term value, versatility, and asset recoverability. The plant can be sold quickly since geography is no longer a constraint, for example, if economic conditions change for the worse in a local region, the plant can be moved where conditions are more propitious.
The plant could also be installed nearby hydropower facilities to exploit the ultra-cheap electricity available at such installations.
The plant could also be used in offshore wind facilities since the critical components are sealed from the elements inside the ISO container.
The limits of scaling
“Sometimes individual components bigger actually make them more expensive per unit of output. Example: A fixed-tube reactor, when the shell diameter is under 12 ft, can be built in numerous shell and tube heat exchanger shop around the world. We have identified almost 100 shops in North America alone that are capable of building up to 12 ft diameter reactors. When the diameter gets bigger than 18 feet, as is the case with Shell’s reactor, the number of shops that can handle the fabrication (and the weight) drops to a handful and the cost per tube (a measure of unit productivity) goes up, not down. We have received quotes for reactors under 12 ft in diameter and compared them to quotes for reactors in the 18+ ft range. The cost per unit of reactor volume actually increased with the larger reactor (18 ft+ diameter). Additional shipping costs and erection costs also follow extremely large and heavy components”
The above is a comment by a small company that manufactures Fischer-Tropsch synthetic fuel reactors. Interestingly, they came to the same conclusion we did. An ammonia converter 3 meters in diameter is markedly more difficult to manufacture than one that is only 100 millimeters in diameter. In fact, a 100-millimeter reactor can be fabricated by any facility that has access to either welded tube, or ideally, seamless pipe produced through the mature Mannesmann (Rotary piercing) process. One important fact to highlight is a modular plant based on small-scale ubiquitous commercial off-the-shelf components can be built in a factory taking advantage of lower cost labor. A large ammonia plant has to be built on-site with much higher labor costs due to the higher skill level.
The picture above illustrates the immense size and concomitant engineering and fabrication difficulty of a large scale ammonia converter which must be forged from a very large hot steel ingot. The cost per cubic meter of the above large scale pressure vessel is going to be substantially larger than the easily machined 13 liter reactors for our plants.
Cold or hot rolling of steel plates using slip rollers is an alternative and cheaper manufacturing option, but requiring welds, they can sometimes feature compromised strength. In the case of ammonia vessels, it is not the weld in and of themselves that are weaker, welds in fact usually exceed the tensile strength of the surrounding material, it is rather the carbides and proness to hydrogen attack that makes welds pregnable. Even using slip rolling, larger pressure vessels require hugely powerful slip rollers, which are far more expensive than smaller ones, even on a per-unit output basis.
High purity synthesis gas from electrolytic production
One of the major advantages of electrolysis-based hydrogen production is the high purity and lack of permanent catalyst poisons present in the gas stream. During the production of hydrogen from natural feedstocks such as coal or natural gas, significant quantities of sulfur are present that must be removed to avoid permanent de-activation of the catalyst.
Sulfur is not the only element that acts as a poison toward the iron catalyst, phosphorous, arsenic, and chlorine are also classified as permanent poisons. While carbon monoxide is not classified as a permanent catalyst poison, it poses a tremendous challenge upon designers. Synthesis gas from methane reforming contains trace amounts of carbon monoxide even after passing through the water-gas shift reactor, requiring an additional mechanization reactor to remove these residual impurities. In China where coal is a major hydrogen feedstock for ammonia synthesis due to the lack of natural gas, short catalyst lifespan has been a common feature for many medium-size ammonia plants. If subjected to even trace concentrations of sulfur, the iron catalyst will rapidly deactivate. The beauty of electrolysis-based hydrogen production is that only one impurity is required for removal: oxygen, while oxygen is still a severe catalyst poison, it is classified as only a temporary catalyst poison since it can be reduced by passing hydrogen through the catalyst bed. Additionally, oxygen is highly electronegative and has an affinity for virtually every element, allowing straightforward capture and removal using iron particles for example.
The importance of improving reciprocating compressor technology
Synthesis gas compression, not kinetics or thermodynamics, limit the down-scaling of conventional high-pressure ammonia synthesis. Ammonia synthesis stands out from other similar temperature catalytic processes insofar as the operating pressure are many times greater than what is typically found in corresponding chemical production operations. This obviously means compressor design, maintenance, repair, and operation dominate the overall design criteria and parameters of an ammonia plant.
Therefore, a significant improvement in compressor technology has great implications for improving the 100-year-old Haber-Bosch process. The reason for this is that larger plants make use of more compact and thus lower cost and longer-lasting centrifugal multi-stage syngas compressors, while for smaller plants, these centrifugal compressors become too inefficient due to tip losses, turbulence-induced parasitic losses due to high Mach numbers, and the boundary layer effect. Additionally, centrifugal compressors have a very low-pressure rise per stage, this is exacerbated especially due to the low density of H2/N2 mixtures at a 3:1 ratio, yielding a gas with a mean density of 0.2 kg/m3. Since the operating principle of a centrifugal compressor is based on imparting kinetic energy into the gas medium rather than squeezing it in a fixed volume chamber, this makes centrifugal compressors very sensitive to gas density so there is a clear limit to how light a gas can become before it can no longer be compressed. For example, pure hydrogen cannot be compressed without the addition of higher molar mass dilutant gases.
Despite having lower footprints compared to reciprocating systems, centrifugal compressors still have relatively large footprints due to the multi-staging required for to achieve pressures over 150 bar. This, unfortunately, means that for small plants of 100 TPY, which is the size of Pochari Technologies’ CP100 plant, a radial compressor is not suitable.
Moreover, centrifugal compressors require very high RPMs to achieve a sufficient pressure rise, therefore, operation with a cyclical power source is next to impossible.
Reciprocating compressors are no panacea either. For spasmodic renewable integration, they also face serious challenges.
Traditional reciprocating compressors, either based on conventional or diaphragm architecture, suffer from high capital costs, bulkiness, the poor time between overhauls, and excessive failure rates. Diaphragm compressors, the backbone of high purity gas compression, are not suitable for small-scale ammonia production based on spasmodic renewables, for constant flow coal/MSW gasification systems they are marginal at best.
Diaphragm compressors feature a unique weakness that effectively outlaws their uses for intermittent ammonia production on a small scale.
The problem is mainly imputable to their lack of reliability and longevity for cyclical operations. Since distributed ammonia uses solar or wind energy, the plant must be able to operate at partial load capacity, with frequent starts-ups and shutdowns. According to Klaus Hoff of Neuman & Esser Group, diaphragm compressors suffer from premature diaphragm failure attributable to sharp thermal gradients experienced when reaching operating temperature from static temperature. The diaphragms are constructed from thin smooth stainless steel disks which flex only a small amount up and down driven by hydraulic oil, to minimize the risk of gas/oil cross contamination, three disks are usually used simultaneously. While the disk is extremely reliable during steady-state operation, it suffers from premature thermal-induced fatigue failures if cycled excessively. The hydrogen compression industry has experienced this issue for diaphragm units used for fuel cell vehicle fueling stations. Compressors operating in a frequent startup and shutdown regime can last as little as six months before replacement of the diaphragm is required. During a regime of continuous use, a diaphragm life of 5-10 years is achievable. 5-10 years is not 8760 hours each year obviously, but a few thousand, the maximum life is only 5000-8000, which is barely half to a full year, and most Chinese diaphragms will only last 5000, requiring two annual maintenance schedules for a 24/hour operating plant. Despite this, diaphragm compressors, regardless of their usage profile, suffer from an additional inherent weakness. Because the metal diaphragms are very thin to provide the elasticity required for a sufficient moment, the diaphragms are sensitive to particle erosion from the gas stream, if the feed gas consists of even the smallest solid particles, the diaphragm will be pitted and fail prematurely, diaphragm life as a consequence is typically limited to 5000 hours, so even in a continuous regime using coal gasification, this compression technology is simply not satisfactory.
Since 5000 hours is 0.57 years of continuous operation, using classic diaphragm compressors would require a dedicated team to arrive at the site, if the site is remote, adding considerable transport cost, to perform the task of replacing the diaphragm, which itself is a time-consuming and laborious process. Replacing the diaphragm requires the personnel to unbolt the diaphragm head, lift is it either by hand if the compressor is relatively small (100-300 TPY capacity), and if it is over 300 TPY, it may require a hoist. Before the head can be removed, the crew must uninstall the hydraulic and gas fittings and carefully reinsert the new diaphragm and then bolt everything back together, this could easily take an entire day and cost thousands. The cost of the diaphragm itself is not necessarily the issue, but rather the labor and downtime that it imposes. What is called for is steady-state operation which can endure all throughout the year with minimal degradation and maintenance requirements, for this to be satisfied, a new architecture is demanded.
A small scale ammonia production requires highly reliable near 24-7 operation with minimal human supervision, downtime must be kept to a minimum, in order to achieve this, the major components need to be designed for utmost reliability. Since the reactor, purification system, and electrolyzer have effectively no moving parts, the compressor arises as by far the as the weakest link in the chain and the component most receptive to innovation and improvement.
Clearly a better solution is required or small scale ammonia will fail to compete with centralized production that take advantage of maintenance and friction free-centrifugal compressors.
One such solution is the metal hydride compressor, using hydrogen occluding materials like Lanthanum-Nickel alloys (LaNi5), which has a hydrogen storage capacity of as high as 6% by wt. Metal hydrides are an attractive option as it would entail the elimination of any moving parts other than a valve. Unfortunately, hydrides require temperature modulation for releasing the occluded hydrogen and worst yet, at higher pressures, especially the pressures needed for ammonia synthesis, the hydrogen is hard to release from the metal. Metal hydrides, although an elegant concept theoretically, are unlikely to be a competitive and viable solution.
What is needed is a new and improved low-friction compression technology, of which without, small scale production will flounder. At first glance, electrochemical hydrogen compression may seem as the obvious option, but the energy consumption and cost of of electrochemical hydrogen compression is far too high to be competitive. The state-of-the-art electrochemical compression technology, while featuring outstanding compactness, use about 6.6 kWh/kg-H2 to compress to 400 bar. In addition, the CAPEX of these units is even higher than PEM fuel cells, these units cost over $5000/kW, immediately ruling out their use. Their high cost is attributable to their use of expensive membrane-electrode assemblies (MEAs) and proton-exchange-membranes (PEMs), made of expensive and specialized materials that do not adhere to our COTS methodology.
We believe the solution is none other than liquid-piston gas compression, using non-miscible ionic liquids. One of the major advantages of using an ionic liquid is twofold, eliminating oil contamination of the synthesis gas and achieving superb tribological performance, permitting nearly friction-free compression dramatically increasing TBO and the attendant levelized CAPEX. Ionic liquids also feature very low compressibility, reducing mechanical losses. In addition, most ionic liquids have high viscosity, enabling leak-free design using standard seals, they also have low corrosivity against stainless steel. One of the central advantages of this compression technology is its drastically higher compactness. Ionic liquid piston compressors thanks to the heat capacity of the working fluid could achieve near isothermal compression, reaching close to 100% efficiency. Ionic compressors also enjoy much longer operating intervals before scheduled maintenance, in the order of 500 days, a factor of 10x over mechanical piston compressors in large part due to their reduced number of moving parts, only around 8 vs over 500 in mechanical compressors. The ideal ionic liquid is immiscible with the gas, meaning that the ionic liquid will not end up in the synthesis gas and eventually in the reactor. Moreover, the ionic liquid does not possess a significant vapor pressure, in fact, the chosen ionic liquid has effectively zero vapor pressure.
Linde Kryotechnik AG first invented the ionic compressor in 2001, with the patent already expired, it offers tremendous potential for commercializing in the ammonia production industry as its expected that China will begin manufacturing this technology for the conventional compression markets to replace diaphragm systems.
One of the principal reasons this technology has enjoyed little in the way of commercialization in surrounding industries is due to the fact that the in the past two decades, the entirety of the intellectual property is still presently held by Linde, making it difficult for competitors in the West to offer this technology without fearing patent infringement. This is quickly changing, however, since not only are the patents close or already expired, but there is increasing latitude for foreign manufacturers to offer slightly different designs while still using the principle of the ionic separator/piston. China has recently begun developing its own ionic liquid compressors for hydrogen refueling, it is expected that this compressor technology will be highly disruptive if it lives up to Linde’s stated performance.
The Linde IC90 ionic liquid compressor designed for hydrogen refueling stations is 25x more compact than a standard reciprocating compressor. A 900 bar 370 Nm3/hr ionic liquid hydrogen compressor has a height of 1.1 meters, while a 100 Nm3/hr 440 bar reciprocating piston compressor has a height of 3.5 meters, adjusting for pressure and mass flow, the improvement is nearly 30x.
Illustration of the Linde ionic liquid compressor.
From this image, one can see the parallels between this architecture and the diaphragm compressor, in that the hydraulic fluid provides the linear force required to compressor the gas. The principal advantage of the ionic liquid design is the elimination of the failure prone diaphragm and its considerably higher volumetric flow capacity, even higher than traditional reciprocating compressors. This is due to two factors, the elimination of the connecting rod which allows a much stroke to bore ratio, which can be achieved also because of the heat capacity of the ionic liquid. The elimination of the connecting rod and crankshaft not only reduces cost, it removes friction and life limited components.
A standard radial hydraulic pump provides the compression energy which is then driven by an electric motor.
The relative simplicity of the module is highlighted in the above image.
The above image highlights the tremendous power density of the ionic liquid compressor.
The Linde IC90 draws only 75 kW to compress a mass flow of 370 Nm3 of hydrogen per hour at an inlet pressure of 5 bar to over 900 bar.
The working principle of an ionic compressor is straightforward. One can visualize the ionic liquid as a separation medium, substituting the metallic piston assembly. In the Linde design, invented by Robert Adler in the early 2000s, the mechanical force needed to push the ionic liquid to compress the gas derives from a conventional hydraulic fluid, the ionic liquid is not compressed with a hydraulic a pump. Rather, a hydraulic piston beneath a layer of ionic liquid above it forces the entire assembly to move upward. The ionic liquid itself is not a working fluid, but rather a separation medium to avoid contamination of the product gases, reducing friction, and providing near-isothermal operation by absorbing heat from the compressed gases. The ionic liquid can then be recirculated and cooled to maintain a constant temperature.
Pochari Technologies’ has chosen an alternative design. Two separate chambers are present, one for a metal piston to displace the ionic liquid into the gas compression chamber. The metal piston in the ionic section is lubricated by the ionic liquid itself, hydraulic fluid is not present to avoid contamination and mechanical losses from the hydraulic pump. It has been demonstrated that ionic liquid, especially those that are hydrophobic (the ones intended for use), possess extraordinary tribological properties, exceeding that of conventional mineral and synthetic oils.
The key requirements for the ionic liquids are a low vapor pressure, a low solubility, and a high decomposition temperature. Some ionic liquids have decomposition temperatures as high as 400 C.
Below is a discussion from a 2017 thesis authored by Nasrin Arjomand Kermani on ionic liquids being considered as candidates for a Linde-style compression system:
“For lubricating well-known systems, such as steel on steel, as well as on difficult to lubricate systems, such as steel on aluminium, ionic liquid lubricants have been shown to outperform commercially available lubricants, such as fully formulated engine oil”
“Even without any additives, ILs have shown remarkably better lubrication and anti-wear properties than normally used synthetic oils and are promising alternatives with considerably lower development costs. Furthermore, ionic liquids have superior durability as lubricants”. “Long-term experiments have shown that ionic liquids even during long-term operation do not lose or change their properties. Therefore from today’s perspective, a change of ionic liquid must at the earliest after about 10,000 hours of operation, while conventional lubricating oils an oil change after about 3,000 hours is required”. As previously mentioned, ionic liquid-based liquid piston compressors have been successfully implemented for hydrogen fueling stations, this technology has the potential to spill over into the domain of Haber-Bosch technology to further process-intensify, enabling the desired down-scaling. “A suitable ionic liquid is selected as the most reliable replacement for the solid piston in the conventional reciprocating compressors. Ionic liquids are room temperature salts that have very low vapor pressures. The ability to tune the physicochemical properties of ionic liquids by varying the cation-anion combinations is the feature of these liquids that make them as promising candidates to replace the solid piston. The ionic liquid: 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide is recommended as the best candidate that can be safely used as a replacement for the solid piston in the conventional reciprocating compressors for compressing hydrogen in the hydrogen refueling stations. In addition, the corrosion behavior of various commercially available stainless steels and nickel-based alloys as possible construction materials for the components which are in direct contact with the selected ionic liquids is evaluated. The results show a very high corrosion resistance and high stability for all of the alloys tested in any of the five selected ionic liquids. The stainless steel alloy, AISI 316L, with a high corrosion resistance and the lowest cost is selected as a material for all the components in direct contact with the ionic liquid, in the designed ionic liquid hydrogen compressor.”
Kermani et al tested numerous ionic liquids to find one with suitable comptability with available stainless steel alloys. It was found that AISI 347 showed the highest corrosion corrosion resistance when exposed to each of the the ionic liquids tested. The lowest corrosion observed was when using trihexyltetradecylphosphonium bis (trifluoromethylsulfonyl) imide.
It’s interesting to find researchers working on Linde’s largely proprietary technology, the researchers Kermani et al appear not to be funded by Linde, suggesting the patent either has been not been able to protect the basic concept of using ionic liquids as a compression medium (too broad a scope), or that Linde is beginning to promote the technology as to expand its market share. Either way, it is a perfect opportunity to develop this novel compression system to replace diaphragm compressors in the high purity gas industry (medical oxygen, semiconductor hydrogen, helium etc).
Kermani writes in her 2017 thesis, “So far, this technology has been limited to the monopoly of a single supplier”.
Only a handful of Linde ionic compressors are in operation today, since fuel cell vehicles appear not to be taking anytime soon, perhaps it will be micro ammonia synthesis that will finally make this niche technology more mainstream and spill over into different compression applications, such as helium, medical oxygen, and general gas compression. As for users of gas compression systems have not considered this attractive cannot be answered, but what is clear, is that ionic liquid compression is far more attractive than the traditional diaphragm architecture.
The potential to reintroduce the Claude process via ionic liquid compression
Regardless of whether a mechanical or liquid piston is used, hydrogen compression work input is logarithmic as a function of pressure, this is also the case for highly non-ideal gases like hydrogen, meaning that compression energy demand is non-linear, with decreasing amounts of energy required at higher pressures, this makes the use of medium pressure alkaline electrolyzers very attractive.
For example, only 1.2 times more power is required to compress to 900-1000 bar vs 200 bar. If the hydrogen is already compressed to 50 bar in the electrolyzer, only 0.48 kWh/kg is required to reach 300 bar, this increases slightly to 0.65 kWh/kg if the electrolyzer pressure is reduced to 30 bar, a number that is very reasonable with the current maturity of alkaline technology. Thus, what can be inferred is that it is not the energy consumption of compression per se that is the limiting factor of downscaling, rather it’s the footprint, and attendant CAPEX of low-speed reciprocating compressors that disincentives decentralization of production. Therefore, Pochari Technologies logically concludes the ideal avenue to approach down-scaling is improving compression technology. The availability of much-improved compression technology opens up another avenue of process intensification.
When most people think of improving ammonia synthesis, they imagine some new catalyst or entirely new process has to be invented, since the 100 year old process is virtually exhausted.
It may be possible to construct an ultra high pressure nearly single pass plant using ionic liquid compressors.
The equilibrium yield of ammonia at 400 C is close to 70% at 900 bar.
The concept of ultra-high pressure ammonia synthesis was originally developed by Georges Claude in 1917. The principle was sound, the equilibrium favored very high pressures at moderate temperatures, this was evident as Claude’s reactor was far more compact, measured in TPD (tons per day) per cubic meter of volume.
The original Claude reactor operating at 1000 bar achieved 120 TPD/m3 vs only 25 for the Haber-Bosch reactors of the time. Unfortunately, this higher density came at a cost. The major disadvantage of Claude’s reactor was that it featured a significantly shorter catalyst life, while Haber-Bosch catalysts lasted 2 years, Claude’s lasted only 3 months. In addition, the alloys at the time were much weaker than ones currently available, requiring the reactor to be massively overbuilt offsetting the cost saving. General Electric invested in the Claude and attempted to build a reactor but experienced too many technical issues and eventually abandoned it, this sealed the fate for the Claude process. The downsides of the higher pressure system could simply not be circumvented by engineering paving the way for the domination of the Haber-Bosch design.
With newer more stable and durable catalysts, higher strength alloys, and optimized compressor technologies, it may be possible to get close to these pressures while maintaining acceptable catalyst life. The trade-off that must be made is between catalyst life, reactor compactness, and conversion per pass. When Claude was trying to achieve ultra-high pressure ammonia synthesis, compressor technology was much less capable than it is today, ionic liquid compressors may enable the resurrection of the Claude process to enable miniature ammonia plants of even smaller footprints than the already small footprints achievable with high activity iron namely Wüstite. Considering the combination of logarithmic compression energy, the development of compact and high-pressure ionic compressors, and longer-lasting catalysts, the Claude design warrants serious consideration in the 21st century.
Ammonia plant failures
G. P. Williams from Allied Chemical Corp perform a detailed analysis on the nature of ammonia plant failures.
During the 1975-1976 survey period, 45% of the downtime and 55% of the shutdowns were attributed to major equipment failures, with the synthesis gas compressor and primary reformer tubes, risers, and manifolds being the most troublesome major equipment items. The average downtime experienced in the ammonia industry is 16%.
with major equipment failures and preventive maintenance accounting for 85 to 95% of the downtime. Primary reforming and major compressors and turbines are the leading contributors of major equipment downtime accounting for 63% of the downtime.
Compressor failures usually follow the following profiles. Severe vibrations on compressor and/or turbine, rotor failed, compressor thrust collar broke, coupling bolts failed, compressor seal failed, broken blades on burbine, reduction gear bearing failed, compressor O-ring failed, etc.
The synthesis gas compressor leads the “top 6” list, and it has the highest percentage (25%) of the major equipment downtime ever attributed to a single major equipment item. The “average” ammonia plant had one shutdown due to the syngas’ compressor every 14 months during the 1971-1972 survey period; once every 12 months during the 1973-1974 period; and once every 9’/2 months during the 1975-1976 period. Downtime days per shutdown averaged about four. Syngas compressor problems are also widespread; that is, two-thirds of the plants had at least one shutdown during the 1975-1976 period, with one plant reporting 10 shutdowns over the two-year period.
Since we have covered the key requirements of synthesis gas compression that makes the Haber-Bosch process unique among chemical production methods, we will cover the topics of reaction kinetics as determined by scale, and its design implication.
The kinetics of the actual reaction, of which the most important is the rate-limiting step of nitrogen dissociation, is not affected by the gross volume of the catalyst, it is however strongly affected by the pore size of the active catalyst the packing density.
The second variable is often mistaken for kinetics. While heat transfer may strongly affect the reaction rate, for the most part in ammonia synthesis, heat transfer mainly determines how much leftover waste heat is produced along with the equilibrium temperature within the reactor. Heat transfer could be argued to be the only reaction variable affected by scale, smaller reactors have higher surface-to-volume ratios, thus will produce less useable heat since they would radiate more heat into their surroundings. Approximately 1.5 kWh of thermal energy is released upon the formation of 1 kilogram of anhydrous ammonia while only 0.45 kWh of heat is needed to heat and maintain 1 kg of synthesis gas to 500 degrees Celsius, this means the reaction is moderately exothermic and more heat is produced than is needed to maintain desired reaction temperature. In industrial-scale plants, this surplus of heat is harnessed to generate steam to cover most of the compression energy requirements. Obviously, unless insulated to compensate for the difference in the surface to volume ratio, smaller-scale reactors would purge more heat, therefore leaving less heat left over to generate the steam to cover compression energy. For a miniature plant, the added hassle of installing a boiler and steam turbine is unattractive, and as with all turbomachinery, small scale results in abysmal efficiency, so a steam turbine would be useless, therefore additional heat loss is not a design consideration. Additionally, since the small-scale plant uses much more efficient ionic liquid piston compressors rather than less efficient centrifugal compressors, compression energy has less importance to the overall design criteria.
Another important design variable is found in the design and optimization of gas-separation heat exchangers. At 300 bar, ammonia must be cooled to 130 degrees C in order to bring the gas phase into the liquid state to permit separation, once liquefaction is achieved, the ammonia is removed and the unreacted gases are reheated to the operating temperature. Smaller high surface area high heat transfer exchangers would be more compact and efficient than on larger scale plants since they would perform their heat exchanging function with greater rapidity. This is a major advantage of a small-scale plant.
The second variable is material usage, all else equal, a smaller reactor vessel will require more material per unit mass/volume of catalyst than a larger reactor. This difference in material usage is minimal and does not greatly affect plant economics.
Existing ammonia plants make use of what’s called a “converter”, the converter is the heart of the ammonia plant. It is essentially a large pressure vessel that contains the catalyst “basket” where the core reaction takes place. The basket is designed to permit the replacement of the catalyst without replacing the converter. The converter features a removable cap that is bolted to the rim of the pressure vessel, this cap can be lifted exposing the catalyst basket. Many modern ammonia converters make us of active cooling or what’s termed a “cold wall” converter, where cold ammonia is flushed alongside the wall of the converter to maintain a temperature lower than the synthesis gas inside the catalyst basket, this serves to reduce nitriding as well as HTHA and preserve the life of the reactor. For the miniature plant, the designer must choose between a cooled and non-cooled design, this ultimately boil down to the budget and specific requirements on plant longevity. Ammonia converters are designed to withstand the full pressure of the synthesis gas, up to 350 bar for conventional plants, but increasingly plants are using higher activity Wüstite catalysts to permit lowing of the pressure to 150 bar to reduce compression energy expenditure.
The state-of-the-art ammonia converters, such as the S-200 Topsoe use the SA-542 Grade-B Class-4 vanadium/molybdenum steel for the converter wall construction, while this is still a very high-strength alloy, ammonia converters are not made out of a high-cost nickel/chromium alloy such as Inconel. The ultimate tensile strength of ASME SA542 Grade B Class 4 IS 860 MPa, the yield strength if 585 MPa.
Converter material selection is limited by nitriding and hydrogen attack.
Nitriding is caused by atomic nitrogen diffusing into the metal, if the solubility limit is exceeded, which is around 6% for iron, precipitation of nitrate into the grain structure takes place. This can eventually cause structural failure in the form of cracks. Nitriding is a very gradual process, and the ensuing evidence of nitriding often occurs much later in the plant’s lifecycle. Nitriding is classified as a high temperatures corrosion process taking place at temperatures greater than 370°C, nitriding is considered an issue since ammonia synthesis operating temperatures almost always exceed this threshold.Nitriding causes the formation of a thin, heterogeneous and brittle layer inside the reactor. Nitriding rarely exceed a few fractions of a millimeters per year in the innards of the converter, the solubility of 304 stainless steel is 12% by mass at 530 C, meaning that nitriding will occur once this limit is exceeded nitrogen if nitriding propitious conditions are present. Thankfully for ammonia plant designers, nitriding need not be overly concerning, since the reaction rate is so low and only at best 18-20% of the ammonia is converted per pass, the low concentration of ammonia keeps the nitriding rate to acceptable and manageable levels.
The cost of the converter construction material is substantial for large scale plants as the manufacturing technology required to fabricate a multi-ton cylinder is considerably greater than what is found on a miniaturized plant. For scale plants, compared to the total annual output of ammonia, the material usage and intensity is significant. In some cases, where the catalyst cost is high while steel prices may be lower, the catalyst cost may exceed the converter costs.
The weight versus catalyst volume of benchmark ammonia converters is estimated using three-dimensional CAD modeling of a 2D image. The image is from a patent by AMMONIA CASALE S.A. EP0287765
To confirm the veracity of the inferred 3D estimation, we compare to a ATB Riva Calzoni reactor below:
Far the inferred 3D reactor, the catalyst volume is 30 cubic meters, the dead-weight of the reactor is 252,000 kg.
Assuming an alloy price of $1500/ton, we arrive at $378,000 for material costs alone. Fabrication is difficult to estimate as no literature to date has been published on ammonia converter fabrication costs, we estimate a conservative material fraction of 25%, yielding $1,512,000 for the converter. The catalyst cost is $1,215,000 assuming $15/kg for bulk orders. The cost estimate for the SA-542 steel is conservative, with estimates from Alibaba ranging from $600-1200/ton.
Assuming a lower-end catalyst productivity of of 0.26 kg NH3 for a lower pressure plant, (data from Kapsom industrial), the reactor produces 20 tons per hour, 176,000 tons per year. The reactor CAPEX is thus only $15.35/TPY in the first year. Since the converter is the single biggest and thus expensive item in the ammonia plant, and yet the material+fabrication cost is very minimal, this confirms the high cost of onsite labor, site preparation, and assembly/erection in the overall cost of conventional site-built ammonia plants. Permitting, land acquisition, and various other ancillary factors contribute a large portion of the CAPEX of chemical plants. Note that land costs can be dramatically reduced for containerized plants since stacking is possible up to many stories high, if production in urban areas is desired.
Let’s now compare the mass-catalyst ratio for a reactor sized for 40 kg per hour assuming linear scaling. For a 40 kg-hr reactor with a catalyst activity of 1 kg NH3/kg-cat/hr for simplicity, the diameter is reduced to 0.38 meters, with a mass of 252 kg and a catalyst volume of 0.03 cubic meters. Note the mass/volume ratio remains constant as a function of scaling, the only scaling variable that is subject to non-linear modification is the surface/volume ratio, caused by the square-cube law. As the surface-to-volume strongly affects thermal flux, a small reactor faces a heat loss penalty which is tackled by the use of low-thermal conductivity but high temperature tolerant materials such as rock-wool or microporous insulation.
In Table 4.14, corrosion rates are plotted for various componentry of typical Topsoe style ammonia converters. The highest nitriding rate is found in the lining of the main converter vessel, with a corrosion rate of 9.8 mpy, the lowest corrosion rate is found in the wire meshes.
In Table 4.15, taken from High-Temperature Corrosion And Materials Applications by George Y. Lai, shows nitriding depths after 1 and 3 years of operation. The lowest nitriding rate is found for Inconel alloy 804, 50.4% nickel at only 1.2 mills. Pure nickel experiences virtually no nitriding and could be used if the price permitted.
Table 4.13 shows overall corrosion rates for various alloys in an ammonia converter and plant ammonia line which was exposed to 99% NH3, the converter is only exposed to a concentration of NH3 of 150-20%. Stainless steels, such as 446, 304, 316, and 309, suffered a regime of severe nitridation attack, corrosion rates of 100 mpy or more were observered. Moran et al found that Type 316 was more vulnerable to nitriding than Type 304.
Early ammonia synthesis converters made by BASF constructed out of carbon steel during the 1910s experienced catastrophic failures induced by high-temperature hydrogen attack. Material science has made great strides since the days of Carl Bosch, the modern designer now has entire catalogs of alloys to choose from to satisfy his needs. The “scary” aspect of running very hot and high-temperature hydrogen gas need not dissuade the designer, with proper material selection, this material compatibility challenge can be “ironed out” pun intended. Ammonia is by no means the only engineering sector that is forced to handle undesirable pressures due to fundamental chemistry requirements. The fuel cell industry, water jet cutting industry, and LDPE (low-density polyethylene synthesis industry, which operates at 350 C and up to 50,000 psi), not notwithstanding synthetic diamond synthesis (up to 120,000 bar and 3000 C), have all learned to deal with very high pressures through scrupulous engineering, designing with material fatigue, corrosion, and stress constraints as a core consideration. These extreme conditions forced designers to make the best use of finite element analysis, active monitoring of the material through stress and strain gauges, and periodic inspection. The designer of the small-scale ammonia plant need only to design his components to handle 20 MPa, which in the grand scheme of things, is not the worst of high-pressure engineering.
Hydrogen embrittlement, chemical and atomic
Hydrogen-induced embrittlement can be classified into two categories.
Chemical and non-chemical.
Non-chemical hydrogen embrittlement is what usually comes to mind, this is a low-temperature phenomenon where atomic hydrogen diffuses into the crystalline lattice and engenders a perturbation of the grain structure ultimately leading to detrimental effects on its strength.
The second mechanism is chemical, whereby sufficiently high temperature can cause chemical reactions between carbon and hydrogen.
Non-chemical hydrogen embrittlement is typically not the mechanism that is referred to when discussing hydrogen-induced material failure experienced in ammonia vessels. Rather, chemical embrittlement, usually called “high-temperature hydrogen” (HTHA) attack takes place above 400 C, often the two mechanisms are confused. Non-chemical Hydrogen embrittlement does not occur above 150 C in ferrous metals, that is most metals are immune to atomic hydrogen diffusion at above this threshold. so this form of hydrogen-induced corrosion need not be a concern for the ammonia plant designer. Most carbon-containing steel alloys can handle hydrogen at lower temperatures up to pressures as high as 2000 bar without experiencing severe embrittlement, once this pressure threshold is reached, non-chemical embrittlement becomes a severe issue. The American Petroleum Institute publishes the “Nelson curve” which helps designers choose appropriate materials depending on the temperature, pressure, and hydrogen concentration.
Early ammonia converters: material challenges.
Early ammonia synthesis converters made by BASF, constructed out of carbon steel during the 1910s, experienced catastrophic failures induced by high-temperature hydrogen attack.
Failures were so common that engineers had to place the reactor vessels in concrete bombs or explosion-proof chambers to prevent damage to surrounding equipment or harm to personnel. It was only after Carl Bosch made use of chromium steel that the process could be commercialized. Increased resistance to hydrogen attack is realized by adding chromium, tungsten, molybdenum, or vanadium to the steel, either by themselves or in combination. A vanadium concentration above 1% does not markedly improve hydrogen resistance, but increasing chromium content does not show this diminishing return until around 2%. With 2 percent or more chromium, the rate of hydrogen decarburization and nitriding is severely retarded, and is confined to the area near the surface. Ferrous alloys tested by Vanick containing 7.7% chromium, performed only marginally better than at 2%.
In the paper: Hydrogen Attack on Metals at High Temperatures and Pressures,
Corrosion and Material Protect, by J Schuyten, 1947. ASTM A387, Grade 11, U N S number K11789, is a low carbon, 1.25 chromium, 0.5 Molybdenum Steel, this alloy was found to be highly resistant to hydrogen attack under the most severe conditions encountered in nh3 synthesis and varients of this alloy have been to date.
A seminal text in the history of ammonia technology was published by the National Bureau of Standards in 1927, by J S Vanick titled: Deterioration of steels in the synthesis of ammonia. The authors tested a great number of alloys and established that vanadium and chromium were the most effective at halting severe material degradation. The proper alloying serves to form an outside protective later, as most corrosion-resistant metals do, preventing hydrogen and nitrogen from inflicting further damage to the interior of the material.
Material science has made great strides since the days of Carl Bosch, the modern designer now has entire catalogs of alloys to choose from.
The “scary” aspect of running very hot and high-temperature hydrogen gas need not dissuade the designer, with proper material selection, this material compatibility challenge can be “ironed out”, pun intended.
The designer of the small-scale ammonia plant need only to design his components to handle 20 to 30 megapascals, which in the grand scheme of things, is not the worst of high-pressure engineering.
For high-temperature hydrogen attack, which is the principal concern for material selection, stainless steel due to its absence of carbon, performs spectacularly.
“HTHA is not considered in the case of austenitic stainless steels due to the absence of carbides. In these steels, carbon content is very low and solubility of carbon is very high. Moreover, heat treatment consists of hyper-quenching to solubilize carbon and then avoid the formation of carbides. As no carbides exist, HTHA is not relevant”.
Hydrogen-induced disbonding and embrittlement of steels used in petrochemical refining, S. Pillot, L. Coudreuse, in Gaseous Hydrogen Embrittlement of Materials in Energy Technologies: The Problem, its Characterisation and Effects on Particular Alloy Classes, 2012
Chemical embrittlement or high-temperature hydrogen attack in carbon steel is caused by a mechanism where atomic hydrogen penetrates into the carbon steel forming internal methane pockets, hydrogen initially reacts with carbides present in the steel, such as cementine and martensite, once the atomic hydrogen reacts with the carbide it eventually forms into methane. The internal methane pockets cause the formation of fissures and microcracks, eventually compromising the structural integrity of the material by inducing internal grain boundary separation. Welded regions are more prone to hydrogen attack as welding causes the formation of carbides. High-temperature attack is effectively a process of decarburization, where carbon leaves the iron-containing compound and is picked up by the hydrogen.
Hydrogen-induced material weakening is a serious concern mainly because most metals are somewhat hydrogen occluding, the small size of the hydrogen atom makes it inevitable that a certain amount of it will eventually penetrate deep into the material.
The Tesoro Anacortes Refinery explosion in 2011 was caused by high-temperature hydrogen attack weakening the heat exchanger wall housing steel, causing it to rupture. To this date, ammonia converters failing in a catastrophic manner imputable to hydrogen attack has been documented only once, while for nitriding induced failure, no case has been documented. This illustrates the success of material selection and the safety factors employed in choosing the appropriate wall thickness despite the immensely harsh and antagonistic conditions experience during ammonia synthesis.
In 2006, the Mosaic Faustina Ammonia Plant in St James Lousiana experienced a catastrophic failure in the synthesis loop boiler feedwater heater causing the separation of the inlet channel separated from the heat exchanger housing, the failure was traced to hydrogen-induced cracking.
Since the early experiences at BASF, carbon steel was abandoned as a structural material in favor of ultra-low carbon alloy steels. Many early reactors made use of a design where the load bearing pressure vessel walls made use of low-carbon steel but a carbon steel layer was placed on the inside to bear the brunt of the hydrogen attack and decarburization. A novel design in the future may be the use of a ceramic liner on the insides of the reactor wall and heat exchanger tubes, shielding the ferrous material from the hydrogen. With the availability of low carbon high nickel alloys hydrogen attack is less of a concern than during the initial development stage.
Material usage and component selection
One of the principal justifications for scaling is the non linear relationship between capacity and capital expenditure. In the beginning of the paper, we mentioned that the scaling factor is estimated to be 0.64 for ammonia plants, meaning that if the capacity of the plant doubles, the capital costs only increase 64%. This is the principal impetus driving the ever growing scale of Haber-Bosch synthesis.
But is this necessarily always the case?
Upon closer examination, this “law” is not inexorable nor much of a law at all. The primary cost contributor is the cost of the material and concomitant fabrication of these materials into components. The cost of the material is a largely fixed cost, that is it has little to no elasticity at all, larger purchase volume does not affect the price of nickel or iron for example. But obviously, the cost of fabrication/production is a variable cost, and greatly so, which is largely influenced by the volume and speed at which the production facility churns out a particular part or components. The speed at which a component is fabricated has the second largest influence on the cost. For example, if a part has to be machined with a 5 axis CNC machine, and the machine sells for $90,000, has a useful life 20,000 hours, and costs $50/hr to operate, if the part requires only 1 hour to machine, the machine can produce a part for $50 + $4.5, if the same machine required 8 hours, the cost of the part would increase to $440, even if the cost of the metal itself remains the same.
Interestingly, because of the small size of the plant, the major components that require machining can be produced rapidly and at very high volumes. It bears repeating that the dogma of bigger = better is entirely overturned here since small components translated into much easier manufacturing and the potential to tap into vast suppliers of mass produced components.
Since a small ammonia plant does not require a high flow volume, unlike large scale plants which use flanged connections, small scale units would make use of threaded fittings rated for high pressure and constructed entirely from 304 stainless steel like the ones in the image above.
If purchased from retailers such as Grainger Industrial Supply or McMaster-Carr, the price of all these individuals components would amount to several tens of thousands for a 100-300 TPY plant.
On the other hand, if these fittings are either bought in bulk from industrial suppliers in China such as Sailuoke Fluid Equipment Inc. and Zhejiang Fangdun Instrument Valve Co.,Ltd, among many others who make ultra-high pressure fittings and valves for fluid transfer industries.
Most of these components picture above can be procured for very competitive prices on Alibaba.com and shipped anywhere in the world by container, or if not too heavy, by air freight which presently due to the continued reverberations of Covid is about $8-9/kg for Hong Kong to San Francisco.
It makes no sense whatsoever to custom fabricate fancy components when these high pressure stainless components can be readily used in low-flow rate plants.
For a small reactor weighing only a quarter of a ton, more expensive alloys such as stainless steel or even Inconel can be used without much of a concern for cost as the relative contribution of reactor mass is trivial. The use of these high nickel alloys compared to the relatively low-quality iron alloys used in traditional plants allows these systems to potentially be used in corrosive environments, although this should not be a concern since the sensitive components can be hermetically sealed from the elements in the ISO container.
For large-scale plants, the volume may be less important than the conversion rate per loop, hence the choice to operate at lower space velocities. For small-scale plants, what is desired is maximal footprint reduction, hence the choice to increase space velocity at the expense of conversion. The higher converter material loading is tolerated as a necessary design compromise that is unavoidable by down-scaling.
The additional material loading from down-scaling pales in comparison to the total system CAPEX, principally dominated by the prime-mover, such as solar farm or wind farm.
For industrial-scale plants, space velocity is largely a function of the pressure, the common space velocity of a low-pressure loop set at 5000–10000 hrs, a medium-pressure loop at 15000–30000hrs, and a high-pressure loop at as high as 60000 hrs. For our smaller-scale plant, we have chosen higher pressures, thus high space velocity and high catalyst productivity. One of the principal advantages of a higher pressure (>30 MPa) system is the ability to more easily separate the synthesis gas from the liquid ammonia. A higher pressure is almost always salutary to the process, but not a high space velocity, since high space velocities induce a greater pressure loss. Higher pressure itself does contribute to pressure loss, as the reactant gases can penetrate further into the porous catalyst, but by far the greatest determinant of pressure drop is space velocity, which itself is the single biggest factor affecting yield after pressure. It was found using data from the high-activity Wüstite catalyst, that there was little need to increase the power density through “process intensification” of the reactor, the ancillary componentry should therefore be the main target of efforts to process intensify. Moreover, even if there was a need to “process intensify”, there is little ability to even do this outside of improving the auxiliary systems, which themselves occupy the most space as the reactor vessel and chilling and reheating exchanger are relatively compact.
Since compressor CAPEX, footprint, and OPEX is a major limiting factor, by using ionic-liquid compressors, we allow the plant to raise pressure while keeping compressor footprint low, enabling a raising of the reactor output. Hence, there exists a compromise between compressor footprint and reactor footprint, ionic liquid compression thus plays a central role in permitting the process intensification and consequent miniaturization of the ammonia plant. For example, a 40 kg per hour ammonia plant uses only 6.9 liters of catalyst at a high space velocity of 40,000 hrs.
What can be concluded in the attempt to down-scale ammonia synthesis technology is that reaction kinetics remain unaffected while increased thermal transfer does not impose major design considerations other than minor reactor insulation. What is highlighted is that it is not kinetics, but rather compression, that remains on the obstacle and explains the preference to build ever-larger plants. It is therefore paramount to focus much of the research and develop efforts on compression technology, rather than catalyst technology. Over 100,000 different material combinations have been evaluated as ammonia synthesis catalysts, none have proven sufficiently superior to iron to pose as a possible contender to replace iron. The odds of discovering a new and improved catalyst are slim to none, in fact infinitesimally small, as nearly every conceivable elemental combination of the periodic table has been tried. Only cobalt and molybdenum bi-metallic nitrides could be possible contenders to replace iron. In 2001, Norskov et al predicted that cobalt/molybdenum bi-metallic catalysts could attain high activity for ammonia synthesis due to their excessively high and low nitrogen binding energies. Molybdenum binds nitrogen very strongly, while cobalt binds very weakly, combining the two proves to yield a highly effective catalyst. The reason for this is that we want a high absorption energy for performing nitrogen dissociation, when N2 is bound to the catalyst surface, it breaks the bond, forming atomic nitrogen, but when ammonia is formed, it is desired to have a certain percentage of catalyst sites producing only weak nitrogen absorption, with a low surface coverage for atomic nitrogen. A cobalt/molybdenum bi-metallic alloy could perhaps achieve activities as high as ruthenium, which is around 2-4 times higher than iron at lower pressures, the issue is the cost of such a catalyst could possibly offset the reduced reactor volume and catalyst mass. For decomposition, (the reverse of synthesis), its been demonstrated that Cantor alloys (high entropy alloys), based on Co/Mo nanoparticles achieve decomposition turnover frequency 20x higher than straight metallic ruthenium. One cannot compare and extrapolate synthesis and dissociation in a simplistic framework of reversing the catalysts in order. For example, ruthenium is 10x more active than the second most active catalyst decomposition: Ni, while iron has very low activity, but ruthenium is only 4x more active than its second-best: Iron, while nickel has a very low activity for synthesis. The mass reaction rate (rm) of ruthenium catalyst is higher by 3.9 times and the TOF is higher by about 3.4 times than that of Wüstite.
For ammonia decomposition, the catalytic activity is as follows:
Ru > Rh > Ni > Co > Ir > Fe >> Pt > Cr > Pd > Cu >> Pb (McCullough and Chiang 2020)
For ammonia synthesis, it looks very different.
Fe > Os > U > Mo > Ru > Mn > Ce. Ni has the lowest activity, followed by Co (Huazhang Liu, 2013)
Rambeaua and Jorti studied the activity of pure osmium powder, comparing the activity with Fe and Ru. They found the activity of osmium was 100x higher than iron at 200 C, but 100x lower than ruthenium at the same temperature. If the temperature is raised to 400 C, the activities reverse, and iron is marginally more active than ruthenium, and 100x more active than osmium. What can be deduced from this is that ammonia kinetics are highly thermally dependent, with the same elements exhibited dramatically altered performance at different temperatures. This is the main reason the huge efforts by catalyst researchers to find a “low-temperature” catalyst has yielded rather abysmal results. One must remember ammonia synthesis is a reversible reaction, since we know decomposition favors low pressure and high temperature, we can assume low temperature synthesis would be ideal, whether or not pressure is held constant. The reality is the reaction is extremely slow at low temperatures, despite theoretical yields (conversion) is very high, so designers were forced to increase temperature to speed up the kinetics, one this was done, a precarious situation was reached, if pressure was lowered, a dramatic loss of activity is experienced, as we are now essentially decomposing a substantial portion of the freshly formed ammonia. Therefor, high pressure will never be overcome unless entirely new mechanisms are discovered, such as plasma synthesis.
What can be observed is that many of the most actives elements for decomposition do not show any activity for synthesis, and vice versa.
Iron is twenty cents per kilogram, while molybdenum is $30/kg, and cobalt is $79/kg, with an average of $54/kg, yielding a 272x price difference for the raw material alone, not including catalyst processing and deposition costs. In conclusion, alternative catalysts will require much higher activities than currently found to justify their higher costs.
Early ammonia synthesis catalysts were based on magnetite, reduced with hydrogen after melting with alkali promoters to produce the desired iron crystal structure. The later catalyst made use of the Wüstite crystal structure.
What’s remarkable is the ammonia synthesis catalyst is considered the most durable catalyst among all industrial catalysts, showing virtually no structural and morphological change after 20 years of extreme conditions of pressure and temperature. As long as the catalyst is not exposed to oxygen, a severe poison (less than 0.5 PPM O2 can reduce activity by 25%), due to the formation of iron oxide, the Wüstite or Magnetite-based catalyst is highly durable, active, and cheap, making its substitution and replacement a very tall order indeed. Oxygen removal is typical done after the methanization step using a copper-ammonium salt. An iron-based oxygen “scavenger” is also an effective option. The complete oxidation of 1 g of iron can remove 300 cm3 of oxygen in standard conditions. The PSA generator produces a purity of 99.999%, with up to 99.9995% being achievable at lower flow rates using the carbon-molecular sieve. The remaining 5-10 PPM is removed using an oxygen absorber.
In 1994, Huazhang Liu and other Chinese researchers invented a new high-activity ammonia synthesis catalyst based on Wüstite as precursor in place of magnetite, the oxidic promoters being substantially the same. It was shown that the reaction rate of the new catalyst is 30- 90% higher than that of the traditional one and that such an activity difference increases when temperature decreases. Resistance to sintering and mechanical strength are reported to be the same of the traditional catalyst, while resistance to CO impurities is higher. It was also reported that the new catalyst is being used successfully in Chinese ammonia plants. The overall ammonia productivity increased significantly, for example, the figure quoted above from older magnetite catalysts from the 1960s shows 19.6% conversion at a GHSV of 29,000, while the Chinese Wüstite catalyst demonstrates the conversion of up to 33% at a GHSV of 30,000 and a pressure of 30 MPa. The Wüstite-based catalysts exhibit about 60% higher catalytic activity. It is considered the most advanced commercial iron catalyst with the highest activity and low production cost in the world, and it is competitive with ruthenium-based ammonia synthesis catalysts. The Chinese catalyst was licensed to German and Swiss companies and is marketed in the West as AmoMax® 10 by Clariant. China continues to dominate Wüstite-based ammonia synthesis catalyst manufacturing, Shaoxing Shangyu Catalyst Co., Ltd is a leading manufacturer of the A-301, A110-2, and ZA-5 series catalysts.
The use of the higher activity Wüstite usually encourages industrial operators to reduce operating pressure, in fact, some plants using the Wüstite-based catalyst in China run at only 15 MPa. In our case, we want footprint reduction, so pressure is most likely going to be kept at above 20 MPa while GHSV is kept commensurately with this pressure. Virtually all ammonia synthesis catalysts are doubly promoted with alkali metals, but aluminum oxide is also used. Donors can be classified into two categories, structural donors and electron donors. In the case of ammonia synthesis, potassium performs the function of an electron donor, donating electrons to the iron facilitating the breakage of the diatomic nitrogen bond. Calcium serves as a structural promoter, blocking certain catalyst sites and exposing more active sites, what is termed [110, 111], referring to the morphology of the crystal structure and its number of faces, site 111 is thought to be over 500 times more active than site 110. The alkali promoters are nearly always present as an oxide, such as K2O or CaO. Promoters not only accelerate the turn over frequency (TOF), but they serve to extend the active life of the catalyst, without promoters, magnetite-based catalysts rapidly deactivate.
Table 5 compares a ruthenium calcium promoted catalyst against the Chinese wustite catalyst. The ruthenium catalyst at the same pressure is approximately 4x more active than the wustite catalyst at 10 MPa.
The image below is an activity chart for the wustite catalyst from a paper titled Wüstite-based catalyst for ammonia synthesis: Structure, property and performance by Huazhang Liu and Wenfeng Han from the Institute of Industrial Catalysis, Zhejiang University of Technology. From this chart, the highest activity can be seen to when the Fe2+ and Fe3+ ratio is 6, at this ratio, the reaction rate or turnover frequency (TOF) is 95 millimoles per gram of catalyst per hour, or roughly 1.618 kg NH3 per kg of catalyst per hour at 15 MPa at an intermediate space velocity of 30,000 hrs.
Below is another chart on catalyst activity taken from a paper by Anders Nielsen of the Haldor Topsøe Research Laboratory in 1953 plotting the activity of a traditional magnetite catalyst versus the gas hourly space velocity at 33 MPa and 450 C. At a space velocity of 30,000 hours, despite twice the pressure, the activity of the old magnetite catalyst is only 0.98 kg-NH3/kg-cat-hr.
The chart illustrates the very strong effect that pressure has on the yield, even at very high space velocity, the conversion per pass is still 20% at a space velocity of just below 30,000. Using the wustite catalyst, the designer must make a trade-off between catalyst yield, longevity, reactor volume, and compressor size. For small scale plants, a more moderate pressure desirable from the perspective of component sourcing, since the cost of high-pressure rated components is proportional to their mass, which is then directly linearly proportional to the pressure they must withstand. Another advantage of lowering the pressure is minimizing high-temperature hydrogen attack as well as maximizing the catalyst longevity.
Pressure drop is an important design consideration due to the phenomenon of intraparticle diffusion. Pressure drop is a design concern due to the requirement to repressurize the feed gases during each loop, consuming additional power. Pressure drop is a direct function of the void volume of the catalyst particle, large particles feature higher void volumes permitting higher diffusion, especially at higher pressures, thus, smaller particles dramatically reduce gas diffusion and hence pressure drop. It was found that particles less than 0.9mm effectively eliminate intraparticle diffusion. Spherical-shaped particles are ideal, they minimize both pressure drop as well as maximize activity. Smaller particles also possess much higher activities than larger particles. For example, for particles 0.6-1.2mm in diameter, the relative TOF is 300, it drops to 61 for a particle size of 6-9mm. The durability of the wustite catalyst was found to be significantly improved over the magnetite-based catalyst. The attrition rate for magnetite was 1.5% over 6.5 years while only 0.5% for wustite. If the traditional magnetite has a TOF of 4.8 kg-NH3/liter-catalyst at a GHSV 40,000 with 19% conversion, we can assume this can be increased to 5.76 kg/liter-catalyst per hour assuming a 20% increase. This means for example a hypothetical 40 kg-hr plant would require only 7 liters of volume, the size of 14 small handheld water bottles for reference.
Coal and solid-waste as a strong contender to electrolysis
As we discussed earlier, photovoltaic is surprisingly competitive with hydrocarbon-based hydrogen production, but as long as irradiance is very high.
Of course in spite of this, many regions simply do not have propitious insolation for photovoltaic, or in some cases, nor is the space available.
Since are marketing this technology purely as a cost-saving solution for existing users of fertilizer, rather than a long-term “carbon mitigation” strategy, our modeling and analysis place little to no consideration for the emission profile of the hydrogen source. CO2 has yet to be demonstrated to be the sole cause of the observed warming, and until this can be demonstrated with unequivocal certitude, placing too much faith in “green” or “carbon mitigation” is unwise considering the economic and societal revelations that would ensue. The single most important criteria are how much we estimate it can be produced for by optimizing today’s technology, the conclusion is that coal and biomass are by far the cheapest source, if purification concerns can be alleviated, and they cannot yet be for certain, so we are stuck we electrolysis if we are to approach miniature production form the most conservative direction.
At a coal price of $200/ton, hydrogen can be theoretically produced for less than 60 cents per kg, yielding an ammonia price of $110/ton excluding amortized equipment CAPEX.
One of the principal disadvantages of an electrolysis-based plant is the high upfront CAPEX of the solar or wind farm required to produce the electricity needed to power the electrolyzer. Solar farm prices as mentioned above around in the range of $500-600/kW including the inverter, but this cost may be too high for a prospective user.
Natural gas reforming, being an endothermic reaction, is highly conducive to miniaturization, in fact, a plethora of papers published discuss designing microchannel high heat transfer steam methane reformers. The main problem with steam methane reforming is the issue that natural gas is not highly portable, requiring plants to be connecting a natural supply line.
If electric-based production proves too expensive in the initial phase, coal and solid waste may emerge as attractive contenders.
Municipal solid waste can be gasified providing a virtually free source of hydrogen, and coal, if it can be procured for market price rather than retail price, can be cleanly gassified in the reactor as opposed to being combusted in powerplants.
The reaction for the water-gas shaft for coal is: C + H2O = H2 + CO
The molar mass of carbon is 12, hydrogen is 2, so 6 kg of coal produce 1 kg of hydrogen.
Then the carbon monoxide can pass through an additional water gas shift process.
CO + H2O → CO2 + H2 –
An additional 1 kg of hydrogen is produced from each kg of coal gasified.
The net yield of hydrogen is comparable to natural gas, approximately 3 kg of coal required to produce 1 kg of hydrogen. Despite the lower hydrogen fraction of coal, the carbon can be continuously partially oxided into carbon monoxide yielding hydrogen from water requiring no energy input.
Despite the stoichiometry predicting a three to one mass ratio, coal gasifier typically don’t achieve better than 5 kg per kg H2, due to the formation of phenolic compounds which are resistant to oxidation.
With a coal or MSW-based hydrogen source, impurities, especially permanent poisons, such as carbon monoxide, become a greater challenge. While ammonia catalysts do not require purities lower than the mid to high part per billion range, absorbent technologies that offer part per trillion range are commercial available, and will significantly improve catalyst yield and lifespan.
Mechanization and water gas shift reaction can remove impurities down to the high PPM levels, the relatively pure gas can then be purified down to parts per trillion levels using specialized absorbents such as Entegris MegaTorr® and GateKeeper® which use micron size membranes. The lifetime of the MegaTorr membranes is around one year of continuous use, depending on inlet impurity concentrations. These high purity filters are not membrane filters that rely on differential molecular size, rather they employ specialized substrates that facilitate physicochemical bonding between the gas impurities that desire to be removed. Another filtration technology are metal-organic frameworks.
The VICI® METRONICS purifier can reduce O2, CO2, CO, H2O, sulfur compounds, and non-methane hydrocarbons from 50 parts per million down to less than 1 part per billion.
Water vapor purification works via physiorption, using zeolites such as clinoptilolite, chabazite, mordenite, and erionite. To remove oxygen, carbon monoxide, and carbon dioxide, reduced metal powder is dispersed inside the zeolite-containing cartridge. Zeolites are an aluminosilicate compound, with a molecular formula of Na2Al2Si2O8. Their cost is surprisingly low, with small diameter zeolite sieves selling for $150-200/ton.
The ammonia industry is not unique in requiring high purity gases, especially in the semiconductor manufacturing sector, which requires ultra-high purity gases, ideally in the parts per trillion range, which cannot be achieved even by cryogenic distillation, applications including photolithography, epitaxy (crystal growing), dry etching, and gas chromatography, these end-use scenarios sometimes use cartridge type in-line filters to achieve the required purity rather than procure the gas in bottles already purified.
A typical gas purifier cartridge with the partitioned zeolite compound visible.
Gas purification is a sine qua non for ammonia synthesis, of which without, the process would be impossible as the catalyst would de-activate within minutes of use. While we discussing detail the possibility of using coal, natural gas, and solid waste as hydrogen sources, all of these products generate “dirty” hydrogen, which needs extensive cleaning. We highly recommend using the electrolyzer, especially simple modular and off-the-shelf alkaline electrolyzers to produce ultra high purity hydrogen, where the only impurity is oxygen which is readily removed.
In spite of the additional purification hassle, and despite the process not being fully, “green”, the techno economics of biomass, solid waste, and coal-derived hydrogen are very challenging to beat, even with the lowest cost photovoltaic sources.
World coal reserves are estimated to be over 400 years, as the world moves away from burning coal directly generating NOx, particulates, heavy metals, and dioxins/furans, coal can be more elegantly used to produce valuable and clean fuels such as hydrogen, ammonia, or even hydrazine.
A cost of $200/ton is considered high in the coal market, as coal burning plants continue to go offline, the price should fall making its use as an ammonia hydrogen feedstock increasingly important especially as natural gas prices continue to climb, reaching nearly $7/MMBTU as of May 2022.
The average annual sale prices of coal at mines in the U.S rank of coal in 2020, in dollars per short ton (2,000 pounds) was:
- bituminous: $50.05
- subbituminous: $14.43
- lignite: $22.16
- anthracite: $98.6
Thanks to the use of advanced insulation technology and its effective integration into the synthesis reactor column, the annoying phenomenon of heat loss that has plagued small-scale production has been solved. This opens up the opportunity to exploit widely available commercial off-the-shelf equipment and components for use in moderate pressure ammonia synthesis which boasts low variable cost due to high production volume. The use of a “component-centric” design philosophy, as well as standardization of small-scale ammonia equipment, results in minimal customization, further exploiting the cost advantage of mass-produced components. The improved use of modularity, off-site manufacturing enabled by containerization, and improvements in compressor technology culminate in a viable and competitive fertilizer and fuel production package.
The anticipated CAPEX excluding any power source for the miniature ammonia plant is only around $15-50,000 for a 100 tons per year system excluding solar farm and electrolyzer costs. This number is very realistic, since we know that industrial-scale plants are in the range of $1000-1500/TPY, but since the miniature plant can use off the shelf components and requires no specialized equipment for erection and site preparation, and is much less labor-intensive, this price is probably overly conservative and could be reduced to $20,000 under an aggressive high volume manufacturing regime. From a basic material usage perspective, a technology is being efficiently produced if it sells for only a small margin above material costs. For example, raw material costs comprise 80% of the cost of modern lithium-ion batteries. For a 100 TPY plant, the weight of the unit is
Including a solar farm at $500/kW and low-cost alkaline electrolyzers at $50/kw, the total CAPEX is projected to be below $3000/TPY. At a price of $800/ton, the payback time is only 3.75 years, over a twenty year period, the levelized cost per ton of ammonia could be as little as a $135 using solar energy.
Small scale ammonia is thus poised to take significant market share away from methane and coal. The additional factor which is the uncertain future of natural gas and coal make the development of electrolytic based hydrogen to ammonia a near requirement in the coming decades. It is likely that in the future numerous companies will offer compact, containerized, modular, ready to install plants at affordable prices by applying assembly-line mass production methodology to chemical plant manufacturing. It is anticipated that China will dominate the industry taking market share away from established Western companies like Casale, Haldor Topsoe, KBR, and thyssenkrupp Industrial Solutions. The competitive advantaged established by these producers is attributable to the level of technical expertise, manufacturing difficulty, and intellectual property they possess in the industry, distributed production overturns this paradigm paving way for a more competitive and consumer-friendly landscape where farmers rather than producers enjoy the upper hand.
Note: most of the price quotes were received from Chinese companies, this was mainly due to the convenience of being able to quickly request quotes on online marketplaces such as Alibaba which have an efficient RFQ “request for quotation” service. There are few Western firms advertising on Alibaba, additionally, many large Western companies due to their prominence are hesitant to spend time giving out quotes to those who aren’t necessarily credible prospective customers, largely due to the reduced market competition they face. Secondly, we believe the prices in China are more reflective of the inherently lower production cost that can be exploited with global manufacturing thanks to modularity and the absence of site specificity.
All components except for the part per trillion gas purifier could be sourced from China, current Chinese manufactures are not yet able to build high purity filters and this technology was sourced from Western firms.
Extant large scale ammonia plants obligate manufacturing and fabrication be to a large extent, on site. Small scale plants can use components sourced in India, China or anywhere in the world where manufacturing costs are the lowest.