Appendix 2 GTL through CO2/steam reforming
A review of the CO2/steam reforming process from basic chemical equations is worthwhile as an initial exercise to understand better the published data.
The process is summarised in the following equilibrium reactions that result in the production of(a mixture of and carbon monoxide).
Equation (3) results in an H2:CO molar ratio of 1:1 and Equation (4) gives a ratio of 3:1.
There is another reaction involved that affects the H2:CO molar ratio. This is the water gas shift reaction -
An H2:CO molar ratio of 2:1 is required to convert syngas into paraffinic syncrude using theprocess. This is described by the equation below.
represents the Fischer-Tropsch product called syncrude. This product is a mixture of paraffins of carbon chain lengths n, ranging from 5 to 100. This product is upgraded by separation into saleable GTL products, namely Naphtha (n = 5 to 10), Kerosene (n = 10 to 14) and Gas Oil (Diesel) for which n = 14 to 20.
It is evident that if the proportion of CO2 in the raw feed gas is in the right range, then a syngas could be produced satisfying the requirement that the H2:CO molar ratio is equal to 2:1.
We estimate the amount of syngas that could theoretically be created from the same raw feed gas rate (781 MMscf/d) and CO2 content (40 mol%) that we assumed for a benchmark 3 Mt/yr LNG train6. We start with reacting the GTL feed CO2 volume (279 MMscf/d, Table 25) with feed CH4 in accordance with Equation (3). As shown in Table 26, this creates 1,115 MMscf/d of syngas by consuming the stoichiometric amount of methane (279 MMscf/d).
Table 25 – Published pilot plant feed data and scaled-up data
Table 26 – Reforming of methane with CO2
The remaining amount of CH4 in the feed available for steam reforming is 139 MMscf/d. This is the GTL inlet CH4 (418 MMscf/d, Table 25) less the amount already consumed in the reaction described by Equation (3) (279 MMscf/d). This gives an additional theoretical syngas quantity of 557 MMscf/d, as shown in Table 27.
Table 27 – Steam reforming of methane
Adding the theoretical amount of syngas from Equation (3)/Table 26 and Equation (4)/
Table 27 gives us 1,672 MMscf/d with a H2:CO molar ratio of 1.4:1. To increase this ratio to the 2:1 needed for the Fischer-Tropsch reaction, we find that 140 MMscf/d of CO needs to consumed by the water gas shift reaction as shown in
Table 28. The resulting syngas quantity stays the same (1,672 MMscf/d) but it now has the needed H2:CO ratio of 2:1. The theoretical amount of syngas is 9% more than the scaled-up published data  as shown in Table 29.
Table 28 – Water gas shift reaction
Table 29 – Published and scaled-up syngas yield by CO2/steam reforming
|Published data for 40% CO2 in feed gas ||Worked example for 3 Mt/yr LNG plant feed (40% CO2)|
|Scale up of pilot plant data by factor of 3.9 per Table 25||Theoretical' syngas (from Table 26, 21 and 22)|
|H2||293 333||262||26 158||1 022||101 913||1,115||111,244|
|CO||146 667||131||183 108||511||713 391||557||777,323|
|Total||440 000||393||209 267||1 533||815 304||1,672||888,566|
In Table 30 the indicative yield of the GTL process is given both for the published case and the scaled case. Table 31 details the sources and flow-rates of carbon for the GTL process. Finally, Table 32 estimates CO2 emissions from the GTL process.
Table 30 – Yield of GTL products and associated CO2 emissions from combustion
|Published data for 40% CO2 in feed ||Scaled data for comparison with 3.0 Mt/yr LNG train feed|
|Scaled GTL production (factor = 3.9)||CO2 emissions factor7||Estimated CO2 emissions from combustion by user|
|GTL products||bbl/d||bbl/d||lb CO2/bbl||lb/d||Mt/yr|
|Naphtha (C5 – C10)||4,374||17,041||886.0||15,098,087||2.50|
|Kerosene (C10 – C14)||6,069||23,645||904.6||21,389,122||3.54|
|Gas oil (C14 – C20)||4,557||17,754||940.1||16,690,614||2.76|
Table 31 – Estimated mass flow of carbon entering GTL production plant
Table 32 – Inferred CO2 emissions from GTL process by carbon balance
We compare the overall system performance of the hypothetical CO2/steam reforming GTL plant taking the same feed as the example 3.0 Mt/yr LNG train. This comparison shows that the estimated total system CO2 emissions are the same (15 Mt/yr). This is expected, since all the carbon atoms produced from thein both cases must end up as CO2 in the atmosphere, unless the CO2 is captured and stored.
Figure 21 – Hypothetical Japan-GTL® CO2 emissions
From the discussion above, we conclude that the new GTL process (Figure 21) is expected to produce slightly less plant CO2 emissions (6.2 Mt/yr) than an equivalent LNG process (7.3 Mt/yr). This is because the new GTL process transfers 66% of the carbon content of the raw feed gas into the product sold to customer. In contrast, the LNG process only passes on 51% of the carbon content in the feed to the customer.
The LNG process requires a capture plant, followed molecular sieves to remove all feed CO2. This requires significant energy for reboiler heating. The same need to remove all CO2 applies to conventional GTL processes.
This process produces more energy at higher product value from the same amount of raw high-CO2 feed gas than does LNG. In addition, compared to LNG, it is easier to transport the GTL products to market.
6 For a 3.0 Mt/yr LNG train, the product is close to 400 MMscf/d of methane. For 40% CO2 in raw feed, we assume that the methane required to power CO2 removal and refrigeration is equivalent to 17% of production. It can be shown by a material balance calculation that the required raw gas feed rate is 781 MMscf. CO2 emissions arising from the production process are 7.3 Mt/yr, consisting of 6.0 Mt/yr from CO2 extraction and 1.3 Mt/yr from fuel combustion. Theoretical CO2 emissions from LNG combustion are 7.7 Mt/yr, resulting in approximately 15 Mt/yr CO2 emissions over the production and consumption cycle.
7 CO2 emission factors for the combustion of petroleum fuels are derived from EIA published data, see http://www.eia.doe.gov/oiaf/1605/coefficients.html