Jul-2024

Hydrogen recovery from ROG Part 2 membrane separation and compression

Using next-generation separation membranes to recover unused hydrogen.

Zach Foss
Divigas

Article Summary

Refinery off-gas (ROG) streams will be a crucial source of hydrogen (H2) in refineries as product regulations and crude processing requirements drive increased H2 demands. The need will further increase as energy companies look to limit heavy sources of carbon within their operations to meet net zero goals by 2050.

As refiners shift from steam methane reformer (SMR)-produced grey H2 to green or blue, costs will rise, increasing the necessity for downstream H2 recovery and purification to remain economically viable.

Part 1 in PTQ Gas 2024 previewed sample ROG streams produced from several types of units for their potential separation and purification via Divi-H, a proprietary next-generation H2 separation membrane. It demonstrates separation costs as low at $0.015/kg H₂ separated, with return on investments (ROIs) exceeding 2,400% over the life of the product when compared to grey H₂ production. Part 2 will analyse its effectiveness in hydrotreaters, hydrocrackers, isomerisation units, and catalytic reformers.

Hydrotreaters and hydrocrackers
Hydrotreaters and hydrocrackers are similar use cases with a few key variables. ROG pressure can vary between these units depending on the level of reaction severity needed to meet product specifications and the composition of the feedstock.

The gasoil hydrotreater, for example, can operate at pressures five times higher than the naphtha hydrotreater. This high-pressure operation makes its ROG ideal for membrane separation. There is a significant driving force in the feed, and H2 product pressures are high enough to feed to downstream hydrotreaters without need for any additional compression. It also tends to have higher H2 purities in the ROG, with concentrations up to 70-80 mol%.

For each hydrotreater and hydrocracker case, permeate purities from 95-99.9% were considered (see Table 1). When there is flexibility in the purity of produced H2, membrane separation is incredibly beneficial. On average, a 99.9% purity H2 stream requires five times the membrane Capex investment of a 95% purity stream. These H2 streams leave the membrane systems at a lower pressure than the feed to be sent either directly to any downstream unit operating at lower pressures or to recycle compressors. Ancillary equipment, including valves, piping, and pressure control, would add ~50% to the module cost.

The high-pressure ROG streams produce a purified H2 stream that is still at high pressure. These streams could easily be recycled to lower-pressure systems without any recompression required. For a stream like the naphtha hydrotreater ROG, the H2 permeate would only be between 7-17 bar. It would almost certainly need to be recompressed to either be recycled back to the source unit or utilised elsewhere in the refinery.

It is important to keep in mind that another benefit of membrane separation is that the retentate stream (the H2 poor stream) leaves the membrane system at a near-feed pressure, allowing for a variety of different recovery and recycle opportunities.

Assuming there is available H2 recycle compressor capacity, the costs of recompression to get the H2 stream back up to feed pressure can be considered. For this analysis, we will look at the H2 recovery over the life of a membrane unit (five years) and assume a unit uptime of 95% (to be conservative) with total H2 production. The cost of H2 separation will then be compared back to the cost of producing new H2 via a grey H2 steam methane reformer (SMR) process (with an estimated price per kg of $1.80) to find the breakeven point:
Where
BEPdays = breakeven point (days)
Cmem     = Capex cost of membrane modules ($)
Canc       = Capex cost of ancillary equipment ($)
CH2gen  = Cost per kg fresh H2 generated ($/kg)
RH₂rec    = Rate of H₂ permeate production (kg/hr)
Pcomp    = Compression power required (kW)
Ecost      = Electricity cost ($/kWh)

With the processing units shown in Table 2, all options show a positive ROI but are highly variable depending on stream conditions and required purities. This demonstrates the criticality of designing the membrane system only to the required purity rather than automatically matching the 99.9% purity attainable from a pressure swing adsorption  (PSA) system.

The total H₂ recovery values vs the amount spent in membrane system Capex and compression Opex are significant, with the naphtha hydrotreater 95% purity case recovering more than $121M in H2 value with only a $3.2M Capex investment and $1.8M in compression costs over the membrane life. Figure 1 is a visual comparison of each case vs the cost of new H₂ from a grey SMR H2 plant, including the Opex cost for compression back to feed pressures.

Isomerisation unit, catalytic reformer, FCC
The isomerisation unit typically has much lower ROG flow rates than the hydrotreaters and hydrocracker to upgrade low-quality naphtha into higher-quality gasoline blending components. The catalytic reformer is one of the few units in the refinery that produces H2 as a side product to its main reactions. It frequently supplements the SMR in H2 production for the refinery. H2 from this ROG stream is typically captured via PSA, and this stream will be reviewed for future comparison to PSA separation costs.

H2 recovery from the fluid catalytic cracker (FCC) was not assessed. A combination of it is low operating pressure (2 bar) and low H₂ purity (<10%) make it less ideal for membrane separation processes. The stream would require compression to at least 15 bar before feeding to the membrane module, and the H₂ permeate stream would require additional compression to be useful. While the separation and purification of this stream would not be impossible, it would be expensive via membrane separation in all but a green H2 environment with >$8/kg H2 costs. Refineries interested in H₂ recovery from the FCC are encouraged to look at membrane separation options if they plan to transition to green H₂ soon.

These lower-pressure ROG streams would need H2 permeate recompression after the separation due to their lower feed pressures. However, to compare the pure separation costs with the hydrotreaters and hydrocracker the same analysis was still conducted. For the case of the catalytic reformer purified to 99.9% purity, Table 3 shows that the estimated cost to produce fresh grey H2 from the SMR plant has already been exceeded. For this reason, this case would not be considered unless the refinery was utilising blue or green produced H₂ at higher costs.