Carbon formation and removal in the steam reforming process

J. Carrasquero, OQ8, Duqm, Oman

Syngas produced by the steam reforming process is widely used across the refining and petrochemical industries for ammonia (NH₃), methanol (CH₃OH) and hydrogen (H₂) production. While processing different feedstocks—from almost pure methane (CH₄) to naphtha—the conversion of the steam reforming reaction is influenced by temperature, pressure, catalyst activity and steam-to-carbon ratio.

If not properly addressed during unit design and operating phase, carbon formation is inevitable during the reforming process, decreasing catalyst activity and increasing pressure drop across the reforming tubes. Carbon removal reactions must be utilized until the net accumulation of carbon is eliminated. This article documents research by various recognized authors and sources on the basis of reforming carbon formation and removal, how catalyst potassium promotion can help to remove the carbon laydown and reduce the steam-to-carbon ratio, as well as the calculation basis of the steam-to-carbon ratio over the steam reforming process.

Steam reforming reactions. The reaction between hydrocarbons and steam takes place over a catalyst to produce H₂ and carbon monoxide (CO); it is then followed by the water-gas-shift reaction.

The most important methane steam reforming equilibrium reactions are described by Eq. 1 (methane steam reforming) and Eq. 2 (water-gas shift):

CH₄ + H₂O ⇔ CO + 3H₂           (1)

CO + H₂O ⇔ CO₂ + H₂             (2)

The steam reforming of gases with a higher molecular weight than methane is described in Eq. 3 (steam reforming), Eq. 4 (methanation) and Eq. 5 (water-gas shift):

C_nH_m + nH₂O → n CO + (n + m/2) H₂           (3)

CO + 3H₂ ⇔ CH₄ + H₂O             (4)

CO + H₂O ⇔ CO₂ + H₂               (5)

Side reactions leading to carbon formation¹ are described in Eq. 6 (methane cracking), Eq. 7 (CO reduction) and Eq. 8 [CO disproportionation (Boudouard reaction)]:

CH₄ ⇔ 2H2 + C                          (6)

CO + H₂ ⇔ C + H₂O                   (7)

2CO ⇔ C + CO₂                          (8)

Due to the temperatures at which steam reformers operate, carbon is constantly being formed from the hydrocarbon feedstock, through cracking (Eq. 6).² However, carbon removal reactions (Eqs. 7 and 8) can also simultaneously occur that remove the carbon laid down, meaning there is no net accumulation of carbon in a well-run plant. With a given catalyst loading in the reformer, the rate of carbon removal (Eqs. 7 and 8) is fixed by the catalyst type and the process conditions.

However, the rate of carbon laydown is a function of numerous conditions, such as the catalyst activity, the degree of sulfur poisoning and the heat input to the tubes. The rate of laydown is more likely to vary compared to the rate of carbon removal; therefore, the selected catalyst should have appropriate activity or alkali promoters to ensure that the carbon removal rate is faster than the carbon formation rate, which would result in net-zero carbon laydown.

Carbon formation. From Eq. 1, the steam stoichiometric requirement per carbon atom is 1. Nevertheless, it has been demonstrated that this is not practical, since carbon forming reactions are promoted under steam reforming conditions.³

The carbon deposition severity generally depends on the following parameters:

1. Feedstock composition [natural gas, naphtha, refinery offgas, liquefied petroleum gas (LPG)]
2. Operating conditions (steam-carbon ratio, temperature, pressure)
3. The reformer design (top-fired, side fired, bottom fired)
4. The catalyst loading quality.

In some instances, carbon will be much more likely to form than in others. Carbon forming reactions are suppressed by using an excess of steam with the result that the practical limit for the minimum steam-to-carbon ratio with methane feed is approximately 1.7. The tendency towards carbon formation on catalysts when stoichiometric ratios of carbon and steam are used is greater with naphtha than with methane, and the minimum practical ratio lies around 2.2 with a typical naphtha feed. Additionally, when considering the steam-to-carbon ratio for the reformer, it is also important to understand the effect on the high-temperature shift unit downstream.

At low catalyst temperatures [600°C (1,110°F)], low-density filaments of carbon are formed in large numbers. These can generate sufficient force inside the catalyst pores support to shatter the pellet. At higher temperatures [650°C–700°C (1,200°F–1,290°F)], a high-density platelet form of carbon is produced, which would have a significant adverse effect on the steam reforming activity by encapsulating the active catalyst surface. Any further increase in temperature [> 700°C (> 1,290°F)] causes a dramatic decrease in the rate and quantity of carbon deposition.

An equilibrium line, relating gas composition and temperature, can be drawn for each of the reactions to show the carbon formation zone. FIG. 1 shows the methane cracking equilibrium (Eq. 6) line for the range of temperatures typically seen in a reformer.

The effects of the carbon removal reactions (Eqs. 7 and 8) are illustrated in FIG. 2, which shows the reduction in size of the carbon formation zone.

The effect of alkali on the minimization of carbon deposition is well-known, both in theory and practice. Alkali metals are known to be active carbon gasification/removal catalysts at the temperatures found in steam reformer tubes; therefore the presence of alkali in steam reforming catalyst enhances the rate at which carbon is removed, as illustrated by FIG. 3.

In a naphtha steam reformer, cracking of the higher hydrocarbons to form carbon also occurs. At any point in the steam reformer tube, the process gas composition determines the direction in which each reaction will proceed. The process gas temperature and catalyst activity will determine the relative rates of reaction. The reactions in Eqs. 7 and 8 are carbon-removing at typical steam reforming conditions, while the reaction in Eq. 6 is carbon-forming in the upper part of the tube—therefore, there is a dynamic equilibrium between carbon formation and removal. Overall, steam reformers must operate in the carbon-removing region.

Potassium promotion. It is well known that carbon formation on a surface, whether the support or catalyst, is affected by the acidity of that surface.² Positively charged acidic sites on a surface will increase the rate of carbon formation, which is partly due to acidic sites catalyzing the cracking reaction. Alpha alumina (a common catalytic support) contains acidic sites, and adding Group 2 metals such as magnesium or calcium neutralizes these, making the surface less acidic.

For a supported nickel catalyst, the steam-to-carbon ratio at which a catalyst would run without forming carbon can be decreased by approximately 16% through the addition of dopants (e.g., calcium, magnesium), compared to an undoped alumina. A way to further increase the surface basicity is to add a potassium-containing compound (such as potash) as a dopant, which will lead to an increased prevention of carbon formation. For alkalized calcium aluminate catalyst, the steam-to-carbon ratio can be reduced by approximately 65% without forming carbon compared to an undoped alumina. This is due to the acceleration of the carbon gasification reactions in Eqs. 7 and 8 and the suppression of carbon formation reactions.

In addition to increasing the surface basicity, the potassium will form hydroxide species in the presence of steam and these will aid in any removal of carbon that is formed on the surface. Depending on the conditions, carbon will form on hot surfaces within the reformer (e.g., the inner tube wall). This is especially likely if heavier species slip further down the tube where the wall is hotter. That carbon must be removed at a faster rate than it is formed to prevent any buildup.

The history of potassium-promoted catalysts goes back to 1975 when a trial was carried out on the No. 1 low-pressure ammonia plant in Billingham, UK.⁴ During the trial, it was shown that the promoted catalyst, where the potassium was incorporated in the support, was successful in the suppression of hot bands that had been seen for the previous charge of unpromoted catalyst. These hot bands associated with carbon formation appeared after only a few months of operation, and it was thought at the time that they were due to a plant uprate. Alkali metals were known to inhibit the steam reforming reaction, but during the plant trial no such inhibition was seen due to the way in which the potassium was incorporated into the support. The effect was confirmed by laboratory experimental testing.

After 9 mos of operation, the reformer was inspected and the tubes containing potassium-promoted catalyst were running cooler with a more uniform temperature than adjacent tubes, which contained unpromoted catalyst. The material was discharged and when examined, only a very limited potassium loss was detected. As the feed is becoming heavier (e.g., from natural gas to heavy naphtha), the percentage of alkali promoters tends to be higher, typically 0% for natural gas and between 6% and 7% for heavy naphtha. However, the amount of carbon protection required also depends on both the steam-to-carbon ratio and the overall heat flux.

Potassium is incorporated into the catalyst in ceramic phase reservoirs with a precise stability to regulate the rate of release onto the surface. This leads to the right level of potassium and hydroxide species on the surface to ensure carbon removal (Eqs. 7 and 8) from all nickel sites throughout the catalyst’s lifetime. The potassium-containing phases present in the catalysts depend on the specific application—typically, they are either a potassium-aluminosilicate or potassium-aluminate, which is incorporated in the support.

The use of a range of phases allows for the release of potassium at an appropriate rate under a range of process conditions and maintains high activity in terms of carbon removal. This also ensures that any adverse effect on the steam reforming activity is minimized. FIG. 4 shows an electron probe microanalysis (EPMA) of a potassium-promoted catalyst that clearly shows areas rich in aluminum (FIG. 4, left) and potassium (FIG. 4, right). It can be seen that where there is a high abundance of potassium, there is also high aluminum content. This clearly indicates that areas of potassium-aluminates act as potassium reservoirs for the catalyst.

Steam-to-carbon ratio. Steam-to-carbon is the molar ratio between steam and carbon flow at the reformer tube inlet, where the carbon molar flow is coming only from hydrocarbons inside the feed. Any feed molecule that does not have both H₂ and carbon in its structure (e.g., N₂, CO, CO₂) is not accounted for in the carbon molar flow calculation during the reformer normal run. For instance, CO₂ in the feed promotes the carbon removal reaction in Eq. 8 under normal reforming operating conditions, so it helps to remove the carbon generated by methane cracking. Nonetheless, under transient conditions, startup and shutdown, or emergencies, the equilibrium can be shifted to the opposite side if the proper measures are not applied—carbon laydown can then happen.

TABLE 1 shows steam-to-carbon ratios that can be applied depending on the feedstock used and the type of catalyst selected.⁵ The catalysts are divided in three groups:

1. No alkali doping, typical for natural gas processing
2. Lightly alkalized, typical for feedstocks such as refinery gases containing H₂ but also a heavy hydrocarbon tail 
3. Heavily alkalized, for naphtha processing.

In traditional H₂ plants with CO₂ absorption and methanation, the steam-to-carbon ratio has always been high—with natural gas as feedstock, the steam-to-carbon ratio is typically 5:5.5.

The reason for such a high ratio was partially due to the low steam reformer outlet temperature, which is often 800°C–820°C (1,472°F–1,508°C). The main reason is the need to achieve a low methane slip exiting the reformer, since this methane was not removed and ended up in the product H₂. So, to produce H₂ of 97 mol% purity, the methane slip ex-reformer had to be ≤ 2.5 mol%.

Steam-to-carbon ratios are narrowed down to their operational limits for the pressure swing adsorption (PSA)-type of H₂ production unit (HPU) and the syngas-producing steam reformers. In the PSA-type of HPU, low steam-to-carbon ratios improve the thermal efficiency, whereas, in syngas-producing steam reformers, the steam addition is kept low to enhance CO formation in favor of H₂ production.

A steam-to-carbon ratio of 3:0 has become a general standard for H₂ plants around the world. Note: This level is also preferred in the proper operation of high-temperature shift (HTS) reactors, since a minimum steam-to-gas ratio is required for the water-gas shift reaction. Lower ratios are applied only in steam reformers that produce syngas for methanol plants.

Steam-to-carbon ratio calculation. The steam-to-carbon molar ratio is dependent on the steam molar flow and carbon molar flow from the hydrocarbon feed (TABLE 2).⁵ To calculate the steam-to-carbon ratio, the steam and feed mass compensated flow and composition analysis are needed.

For each feed mixture, the steam and gas mass flow are the same; therefore, the steam:gas mass ratio is 4 for all cases. Likewise, two conclusions can be highlighted:

1. Steam:gas (mole:mole ratio) is proportional to the feed gas molecular weight.
2. Steam:carbon (mole:mole ratio) is inversely proportional to the carbon molar flow. HP

LITERATURE CITED

1. Snoek, J.-W., G. Fremont and M. Fowles, “Kinetic evaluation of carbon formation in steam/CO₂-natural gas reformers. Influence of the catalyst activity and alkalinity,” International Journal of Chemical Reactor Engineering, 2003.
2. Carlsson, M., “Carbon formation in steam reforming and effect of potassium promotion,” Johnson Matthey Technology Review, 2015.
3. Foster, A. I., et al., “Katalco—The theory and practice of steam reforming,” Johnson Matthey, 2005.
4. Lord, L. “Internal Report RD/CC430” ICI, 1976.
5. Grootveld, G., “Reformer and CO shift operation,” Shell Global Solutions, Process Guide on Hydrogen Manufacturing Units, Chapter 6, April 2011.

JOHAN CARRASQUERO has more than 10 yr of experience in the refining industry as a Process Engineer for hydrocracker, hydroprocesssing and H₂ production units, and has worked with major national oil companies like PDVSA and OQ8. Carrasquero holds a BS degree in chemical engineering, a graduate diploma in Petroleum Studies (Major in refining) and an MS degree in company management.