Carbon negative geothermal: Theoretical efficiency and sequestration potential of geothermal-BECCS energy cycles

The problem

About 70% of the world's geothermal systems are low or medium enthalpy, with reservoirs cooler than 160 °C (Franco and Villani, 2009; Hochstein, 1990). Their low electricity yield often makes them uneconomic to develop. One way to increase the yield is to add a second heat source, such as biomass (DiPippo, 2016). Burning biomass is usually treated as carbon neutral, because the plants took that carbon from the atmosphere as they grew. Capturing and storing that CO2 then removes it from the atmosphere. This is bioenergy with carbon capture and storage (BECCS).

Free-phase CO2 injected in conventional storage is buoyant and could leak unless the caprock is shown to be secure (Kaya and Zarrouk, 2017; Kervévan et al., 2013). Water with dissolved CO2 is denser than fresh water and contains no buoyant free-phase CO2 (Sigfusson et al., 2015). At Hellisheidi in Iceland, CO2 from the geothermal fluid is already dissolved in reinjected water (Sigfússon et al., 2018). Our goal was to combine biomass heating with this form of storage: burn biomass to generate more power, then dissolve the flue CO2 in the reinjected geofluid. We call this geothermal-BECCS.

What we did

We built a systems model of mass and energy balances for each plant. Each design draws 1 kg/s (kilogram per second) of geofluid from a reservoir at 160 °C and 50 bar. The results can therefore be scaled to larger plants, although we did not consider how infrastructure or costs change with size. The fuel is forestry waste from Kaingaroa Forest in New Zealand. Reinjected water leaves the plant at an average of 95 °C and takes up CO2 at 50 bar, 50 times the 1 bar surface pressure in our model. At these conditions, water dissolves up to 0.021 kg of CO2 per kg (2.1 wt%).

We compared four hybrid designs with three benchmarks. Biomass superheats the steam of a flash plant in Configuration 1, and evaporates the working fluid of an organic Rankine cycle (ORC) binary plant in Configuration 2. In Configuration 3, biomass heats the geofluid before it is split between a steam turbine and a binary unit. Configuration 4 is a biomass steam plant that uses geofluid to preheat its water. Benchmarks A and B are conventional flash and binary plants. Benchmark C is a biomass plant with no geofluid.

What we found

In our model, every hybrid design generated more power than the geothermal benchmarks. Per kg/s of geofluid, the flash and binary benchmarks generated 32 and 43 kWe (kilowatts of electricity). Configurations 1, 2 and 3 generated 64, 76 and 62 kWe, at utilisation efficiencies of 9, 9 and 7%, against 6 and 8% for the benchmarks. Configuration 4 generated 589 kWe at 20% efficiency, the same power as Benchmark C from 19% less biomass. However, Configuration 3 almost doubled the power of the flash benchmark but was only 1 percentage point more efficient, and less efficient than the binary benchmark.

Configurations 1, 2 and 3 emit 0.007–0.016 kg/s of flue CO2. The reinjected water can dissolve up to 0.021 kg/s. All of it is therefore dissolved. Configuration 4 emits 0.127 kg/s, about six times that capacity. Only 17% of its CO2 is therefore stored. We count the rest as carbon neutral. Configuration 4 still stores the most CO2 per kg/s of geofluid, 675 tonnes a year, against 217–501 tonnes for the other three.

Per unit of electricity, Configuration 4 stores the least. This is because most of its power comes from biomass, while the CO2 it can store is limited by its 1 kg/s of reinjected water. We estimate emissions intensities of −386, −753, −922 and −131 g of CO2 per kWh for Configurations 1 to 4. Average US natural gas and coal plants emitted 413 and 1012 g per kWh in 2021 (EIA, 2021). These values count operational emissions only.

Scaled to 100 MWe (megawatts of electricity), as in the widget, Configuration 3 would need 1613 kg/s of geofluid and 29 kg/s of biomass, and could sequester 0.8 Mt (million tonnes) of CO2 a year. Configuration 4 would need 170 kg/s of geofluid and could sequester 0.11 Mt a year. Its geofluid demand is about 6% of the take at the 175 MWe Wairakei field (Wymer and Nicholson, 2014). To sequester 1 Mt a year, Configuration 3 would need a rating of 124 MWe. In practice, a site's geothermal and biomass resources would have to be estimated before a plant could be sized.

Limitations

The model treats the geofluid as pure water, with no magmatic CO2. Geofluid at Rotokawa can hold up to 0.35 wt% magmatic CO2 (Addison et al., 2015b). If captured and reinjected, this would use about a sixth of Configuration 4's dissolution capacity. Its sequestration would fall by the same amount. We also neglected parasitic loads, the power a plant uses to run itself, because our modelled plants are small. For Configuration 4, CO2 compression and pumping would use under 1% of output. However, a full-sized plant would need all of its parasitic loads optimised.

A decline in reservoir pressure could form buoyant free-phase CO2 that poses a leakage risk (Kaya and Zarrouk, 2017). Access to a sustainable biomass supply is also likely to limit where these plants can be built. We have not costed the designs. The economic feasibility of a particular site will need further analysis of electricity, feedstock and CO2 sequestration markets.

Glossary

Enthalpy is the heat content of a fluid. Low and medium enthalpy systems have reservoir temperatures below 160 °C.

Geofluid is the hot water and steam produced from a geothermal reservoir.

A flash plant lowers the pressure of geofluid so that part of it boils to steam that drives a turbine.

A binary plant passes heat from the geofluid to a second working fluid in a closed loop. An organic Rankine cycle (ORC) plant uses an organic working fluid, here isopentane.

Superheating heats steam above its boiling temperature. Preheating warms a liquid before it is boiled.

BECCS (bioenergy with carbon capture and storage) burns biomass for energy and stores the resulting CO2 underground.

Flue CO2 is the CO2 in the exhaust gas from a boiler.

Reinjection returns used geofluid to the reservoir through wells.

Free-phase CO2 is CO2 stored as a separate liquid or supercritical fluid, undissolved in water.

A caprock is a low-permeability rock layer above a reservoir that limits upward flow.

kg/s is kilograms per second. kWe and MWe are kilowatts and megawatts of electricity.

Exergy is the part of a fluid's or fuel's energy that could in principle be converted to work, given the surrounding temperature and pressure.

Utilisation efficiency is net electrical power divided by the exergy supplied by the geofluid and biomass.

wt% is the mass of dissolved CO2 as a percentage of the mass of water.

Emissions intensity is the mass of CO2 emitted per kilowatt-hour (kWh) of electricity. A negative value means more CO2 is stored than emitted.

Mt is millions of tonnes.

Magmatic CO2 is CO2 of volcanic origin already dissolved in the geofluid.

A parasitic load is power used by the plant itself, for example to run pumps and compressors.