The synthesis gas can contain a considerable amount of methane and other light hydrocarbons, representing a significant part of the heating value of the gas. Steam reforming (SMR) converts these compounds to CO and H2 driven by steam addition over a catalyst (usually nickel) at high temperatures (Katofsky 1993). Autothermal reforming (ATR) combines partial oxidation in the first part of the reactor with steam reforming in the second part, thereby optimally integrating the heat flows. It has been suggested that ATR, due to a simpler concept, could become cheaper than SMR (Katofsky 1993), although others suggest much higher prices (Oonk et al. 1997). There is dispute on whether the SMR can deal with the high CO and C+ content of the biomass synthesis gas. While Katofsky writes that no additional steam is needed to prevent coking or carbon deposition in SMR, Tijmensen (2000) poses that this problem does occur in SMR and that ATR is the only technology able to prevent coking.

Steam reforming is the most common method of producing a synthesis gas from natural gas or gasifier gas. The highly endothermic process takes place over a nickel-based catalyst:

Concurently, the water gas shift reaction (see below) takes place and brings the reformer product to chemical equilibrium (Katofsky 1993).

Reforming is favored at lower pressures, but elevated pressures benefit eco­nomically (smaller equipment). Reformers typically operate at 1-3.5 MPa. Typ­ical reformer temperature is between 830°C and 1000°C. High temperatures do not lead to a better product mix for methanol production (Katofsky 1993). The inlet stream is heated by the outlet stream up to near the reformer temperature to match reformer heat demand and supply. In this case less synthesis gas has to be burned compared to a colder gas input, this eventually favors a higher methanol production. Although less steam can be raised from the heat at the reformer outlet, the overall efficiency is higher.

SMR uses steam as the conversion reactant and to prevent carbon formation during operation. Tube damage or even rupture can occur when the steam-to — carbon ratio drops below acceptable limits. The specific type of reforming catalyst used, the operating temperature, and the operating pressure are factors that deter­mine the proper steam-to-carbon ratio for a safe, reliable operation. Typical steam to hydrocarbon-carbon ratios range from 2.1 for natural gas feeds with CO2 recycle, to 3:1 for natural gas feeds without CO2 recycle, propane, naphtha, and butane feeds (King et al. 2000). Usually full conversion of higher hydrocarbons in the feedstock takes place in an adiabatic prereformer. This makes it possible to operate the tubular reformer at a steam-to-carbon ratio of 2.5. When higher hydrocarbons are still present, the steam-to-carbon ratio should be higher: 3:5. In older plants, where there is only one steam reformer, the steam-to-carbon ratio was typically 5.5. A higher steam:carbon ratio favors a higher H2CO ratio and thus higher methanol production. However, more steam must be raised and heated to the reaction temperature, thus decreasing the process efficiency. Neither is additional steam necessary to prevent coking (Katofsky 1993).

Preheating the hydrocarbon feedstock with hot flue gas in the SMR convection section, before steam addition, should be avoided. Dry feed gas must not be heated above its cracking temperature. Otherwise, carbon may be formed, thereby decreasing catalyst activities, increasing pressure drop, and limiting plant throughput. In the absence of steam, cracking of natural gas occurs at temperatures above 450°C, while the flue gas exiting SMRs is typically above 1000°C (King et al. 2000).

Nickel catalysts are affected by sulfur at concentrations as low as 0.25 ppm. An alternative would be to use catalysts that are resistant to sulfur, such as sulphided cobalt/molybdate. However, since other catalysts downstream of the reformer are also sensitive to sulfur, it makes the most sense to remove any sulfur before conditioning the synthesis gas (Katofsky 1993). The lifetime of catalysts ranges from 3 years (van Dijk et al. 1995) to 7 years (King et al. 2000). The reasons for change out are typically catalyst activity loss and increasing pressure drop over the tubes.

Autothermal reforming (ATR) combines steam reforming with partial oxida­tion. In ATR, only part of the feed is oxidized, enough to supply the necessary heat to steam reform the remaining feedstock. The reformer produces a synthesis gas with a lower H2.CO ratio than conventional steam methane reforming (Katof — sky 1993; Pieterman 2001).

An Autothermal Reformer consists of two sections. In the burner section, some of the preheated feed/steam mixture is burned stoichiometrically with oxygen to produce CO2 and H2O. The product and the remaining feed are then fed to the reforming section that contains the nickel-based catalyst (Katofsky 1993).

With ATR, considerably less synthesis gas is produced, but also considerably less steam is required due to the higher temperature. Increasing steam addition hardly influences the H2:CO ratio in the product, while it does dilute the product with H2O (Katofsky 1993). Typical ATR temperature is between 900°C and 1000°C.

Since autothermal reforming does not require expensive reformer tubes or a separate furnace, capital costs are typically 50-60% less than conventional steam reforming, especially at larger scales (Dybkjaer et al. 1997, quoted by Pieterman 2001). This excludes the cost of oxygen separation. ATR could therefore be attractive for facilities that already require oxygen for biomass gasification (Katof­sky 1993).

The major source of H2 in oil refineries, catalytic reforming, is decreasing. The largest quantities of H2 are currently produced from synthesis gas by steam­reforming of methane, but this approach is both energy and capital intensive. Partial oxidation of methane with air as the oxygen source is a potential alternative to the steam-reforming processes. In methanol synthesis starting from C1 to C3, it offers special advantages. The amount of methanol produced per kmol hydro­carbon may be 10% to 20% larger than in a conventional process using a steam reformer (de Lathouder 1982). However, the large dilution of product gases by N2 makes this path uneconomical, and, alternatively, use of pure oxygen requires expensive cryogenic separation (Maiya et al. 2000).

Reforming is still subject to innovation and optimization. Pure oxygen can be introduced in a partial oxidation reactor by means of a ceramic membrane, at 850-900°C, in order to produce a purer synthesis gas. Lower temperature and lower steam to CO ratio in the shift reactor leads to a higher thermodynamic efficiency while maximizing H2 production (Maiya et al. 2000).