Radical shake-up
Rising demand is shaking up the staid business of air separation

Larger and more energy-efficient cryogenic plants are cutting the cost of the oxygen needed for world-scale gasification processes, as well as nitrogen for the natural gas business. Traditional non-cryogenic air separation has moved up to capacities of 300 t/d, and a new generation of high-temperature membrane processes promises a radical shake-up in the years to come.
Industrial gases including oxygen, nitrogen, carbon dioxide, and hydrogen are among the most widely used commodity chemicals. In 2008 the world market is expected to reach $ 52 billion. The top ten industrial gas producers have a market share of over 90 percent, with four companies controlling over two-thirds of the business worldwide (see table on page 22).

Historically, there have been two different approaches to air separation. One is cryogenic distillation in air separation units (ASUs), which are typically reserved for applications requiring tonnage quantities of oxygen or nitrogen. The other operates at ambient temperatures, using either molecular sieve adsorbents or polymer membranes.

More recently, a third category of air separation has emerged based on high-temperature ceramic membranes. While commercialization is still years away, demonstration plants suggest that this approach will yield compact systems producing tonnage oxygen at significantly lower costs than cryogenic ASUs. Meanwhile, both cryogenic and conventional non-cryogenic processes continue to develop.

LNG drives nitrogen demand

Nitrogen is widely used to eliminate the risk of fire and explosion, and in modified atmosphere packaging (MAP) to protect perishable products from atmospheric oxygen. Nitrogen is also pumped into hydrocarbon reservoirs to enhanced oil and gas recovery. Liquid nitrogen finds application in food freezing, process cooling, and cryogenic grinding.

A key driver for nitrogen demand is the production and transport of liquefied natural gas (LNG). Europe’s largest LNG plant in Hammerfest, Norway, was built by gases group Linde for the Norwegian petroleum company StatoilHydro, and came on stream at the end of September 2007. Following the removal of carbon dioxide, nitrogen, and water, the natural gas is compressed, cooled, expanded, and then further cooled using liquid nitrogen down to –162 °C.

Nitrogen is also used further along the LNG supply chain: for inerting during transport and storage, and for blending with the re-vaporized natural gas to maintain a consistent heating value.

Oxygen for gasification…

Industrial oxygen is used to enhance combustion during the manufacture of metals, cement and glass. Oxygen also finds application as an oxidizing agent in the production of chemicals and fuels, and as an environment-friendly bleaching agent in pulp and paper. It is also used to support fermentation processes in biotechnology, and in wastewater treatment.

More recently, oxygen demand has been spurred by growth in gasification processes. These convert coal, petroleum coke, biomass, and waste materials into synthesis gas (syngas) consisting mainly of hydrogen and carbon monoxide. The syngas can then be burned to produce electricity (“integrated gasification combined cycle” or IGCC), or further processed using Fischer-Tropsch catalytic reactions to produce chemicals (“coal-to-liquids”, CTL) and fuels (“gas-to-liquids”, GTL).

Modern gasification plants are pushing the envelope in oxygen production. While the earliest commercial IGCC plants were in the range 250–300 MW, for instance, most plants now under development are sized at 450–630 MW, and will each require 5,000 t/d or more of oxygen.

Commercial GTL and CTL plants now under development include Sasol’s CTL development in South Africa; the Oryx (Sasol and Qatar Petroleum) and Pearl (Shell and Qatar Petroleum) GTL plants in Ras Laffan, Qatar; and Escravos (Chevron and Sasol) in Nigeria. For such plants, the cryogenic ASU needed to supply the oxygen represents 10–15 percent of the total capital cost.

The largest of the “mega-syngas” complexes, Pearl GTL, requires 30,000 t/d of oxygen. For comparison, standard-sized gasification plants for chemicals or fuels require around 5,000 t/d of oxygen, and a typical steel mill consumes 500–3,000 t/d of oxygen.

The gases industry has responded by developing ever-larger ASUs. A decade ago, the largest single-train cryogenic ASUs delivered just over 3,000 t/d of oxygen. Today, single-train ASUs can deliver 4,500 t/d, and larger systems are under development.

Design changes targeted at lowering capital costs and energy use, and shortening project times, include the use of tower packings instead of the traditional trays, higher operating pressures to reduce column diameter, and real-time optimization.

…and oxy-fuel combustion

Another driver for oxygen consumption is oxy-fuel combustion. Burning coal or other fuels in pure oxygen instead of air increases flame temperature, reduces the volume of flue gas to be handled, and makes it easier to recover CO2 from the flue gas. According to gas company Air Products, oxy-fuel combustion has the potential to allow 98 percent of the CO2 produced by coal-fired plants to be captured economically.

Gas company Linde and Swedish power producer Vattenfall have a technology partnership for CO2 separation in coal-fired power plants. A 30 MW pilot plant at Schwarze Pumpe (Lausitz/Germany) is the world’s first oxy-fuel plant of its kind to incorporate carbon capture and sequestration (CCS). Another gas company, Praxair Deutschland, is also working with Vattenfall on oxy-coal technology to reduce CO2 emissions.

The use of oxygen in energy applications is driving a change in the design of cryogenic air separation plants. Historically, cryogenic ASUs were designed to produce high-purity (>99 percent) oxygen. For oxy-fuel combustion, however, oxygen of 95 percent purity works well and is cheaper. Strategies such as multi-tower processes, lower pressures and more-efficient compression are achieving energy savings of around 20 percent, with even greater reductions planned.

High-temperature air separation

Several industrial gas producers are pursuing alternatives to cryogenic and conventional non-cryogenic air separation. The new processes operate at temperatures of 600 °C or higher, using novel ceramic membranes or molecular sieves.

Nearest to commercialization is Air Products’ Ion Transport Membrane (ITM) system, which uses high-temperature ceramic membranes to separate oxygen from air. An ITM plant is built up from a series of modules, each about the size of a loaf of bread and containing a stack of ceramic wafers.

Air Products has operated a 5 t/d ITM pilot plant since 2005, and is working with the US Department of Energy to bring a 150 t/d plant on stream this year. The goal is to commercialize ITM plants to produce high-purity oxygen (>99 percent) at 1,000–2,000 t/d or more. Compared to a conventional cryogenic ASU, an ITM plant will be much smaller, around 35 percent cheaper to build, and require 35–60 percent less energy, Air Products says.

Linde’s Ceramic Autothermal Recovery (CAR) process does not use ceramic membranes; instead, it relies on the ability of the mineral perovskite to adsorb oxygen at temperatures of 600–800 °C. Extruded pellets of perovskite are sandwiched between layers of alumina beads to create multiple fixed beds. To release the oxygen, recycled fluegas or superheated steam is used to reduce the partial pressure of oxygen.

The CAR process operates much like conventional PSA, with the difference that perovskite releases heat as it adsorbs oxygen, while desorption requires heat to be added. According to Linde, by recycling heat between the adsorption and desorption stages, the CAR process requires little or no additional energy input.

Linde too is working with the US Department of Energy, and operates a 0.7 t/d CAR pilot plant with integrated oxy-fuel combustion in partnership with the Western Research Institute (Laramie, Wyoming).

Praxair is developing oxygen transport membranes (OTMs), which use an electric current to separate oxygen from air at low pressures and high temperatures. At the membrane surface, oxygen is adsorbed on a porous, electrically conductive coating. The oxygen dissociates to form oxygen ions, which are transported through the non-porous ceramic electrolyte. Once through the membrane, the oxygen ions lose electrons, forming molecular oxygen, which is then desorbed from the membrane’s surface.

Non-cryogenic processes

Especially for applications requiring quantities of oxygen and nitrogen below the tonnes-per-day scale, air separation methods based on molecular sieve adsorbents or hollow-fiber polymer membranes have been widely used since the early 1990s.

Membrane systems are typically used to produce nitrogen, with the oxygen-rich permeate stream vented to the atmosphere. Systems are built up from tubular modules, each loaded with thousands of hollow-fiber membrane strands thinner than a human hair.

Today’s membrane systems are compact and lightweight, allowing them to be used to create inert atmospheres at the point of use. Applications include protecting perishable or flammable cargo on ships, and even inerting of fuel tanks on commercial aircraft.

Adsorption systems typically use activated carbon to produce nitrogen at 95–99.5 percent purity, or alumina in combination with zeolite silicates to yield oxygen with purities of 90–95 percent. Operating modes cover pressure swing adsorption (PSA), vacuum swing adsorption (VSA), and hybrid vacuum-pressure swing adsorption (VPSA).

Recently, several industrial gas companies have commercialized lithium-based adsorbents, which have greater selectivity and higher mass transfer rates. Combined with other innovations such as radial beds, this has yielded systems with less than 25 percent of the adsorbent volume, and less than 80 percent of the energy consumption, of older adsorption units.

Such developments have made larger adsorbent-based systems economic: where 30 t/d used to be the practical capacity limit, onsite PSA systems producing 200 t/d of oxygen are now widely available. As a result, non-cryogenic systems for onsite oxygen production have moved beyond the traditional 20–100 t/d market to cover applications needing 100–300 t/d and beyond.