CO2 Sequestration

Peter Newnam points us to an article in The Guardian discussing the implications of a recent paper, Sequestering Carbon Dioxide in a Closed Underground Volume by Economides et al. We do our best to present the content of the paper in the following paragraphs, because we cannot see the content explained anywhere else. For consistency with our other posts, we will convert the paper's imperial units to metric units.

One way to get rid of CO2 is to pump it into the ground and leave it there. Storing CO2 under the ground is called carbon dioxide sequestration. If we pump CO2 into the ground just any old place, it will leak out again. We must find a suitable geological formation that will hold and keep the CO2 indefinitely. An aquifer is an example of such a formation: a layer of permeable rock, perhaps a hundred meters thick, surrounded on all sides with impermeable rock. Aquifers are usually filled with water, oil, or natural gas.

Exhausted oil reservoirs are ideal aquifers in which to store CO2 because they have been depleted of their fluid content. Oil companies even go so far as to buy CO2 and pump it down into oil reservoirs to push the oil out. But exhausted oil reserves are already being used to store natural gas. They are not available for sequestration. Most CO2 sequestration schemes use an aquifer that is saturated with water or brine. We pump the CO2 into the aquifer at sufficient pressure to compress the water and the rock. The rock and water get smaller to make space for the CO2.

The initial pressure of almost all aquifers, before we start pumping in CO2, will be close to Hρg, where H is the aquifer depth, ρ is the density of water, and g is gravity. An aquifer 2 km down will have formation pressure close to 20 MPa = 200 Bar. Aquifers fracture if the pressure gradient from the surface to the aquifer exceeds 20 kPa/m. An aquifer at 2000 m cannot be pressurized to more than 40 MPa = 400 Bar without fracturing. When an aquifer fractures, its contents leak out.

Suppose the aquifer is initially 20% water and 80% rock. The bulk modulus of rock is roughly 70 GPa and of water is roughly 2 GPa. If we raise the pressure in the aquifer from 20 MPa to 40 MPa, the volume of the pre-existing rock and water shrinks by 0.36%. The Prudhoe Bay reservoir is the largest oil field in the United States. It is 1000 km² in area and up to 100 m thick. Suppose our sequestration reservoir is of the same size. When we pressurize it to 40 MPa, the 10¹¹ m³ volume of the aquifer's rock and water contracts by 360×10⁶ m³. Meanwhile, CO2 at this pressure has density 650 kg/m³, so the maximum mass of CO2 we can store in our aquifer without fracturing its impermeable rock walls is 240×10⁹ kg. A 500-MW coal-burning power plant produces 3×10⁹ kg/yr. There are roughly one thousand such coal-burning plants in the US alone. One thousand reservoirs the size of Prudhoe Bay would be required to store the CO2 generated by these one thousand plants over the course of eighty years.

But the paper shows that the capacity of an aquifer is further limited by the rate at which we can inject CO2 into the aquifer. In order to force CO2 through the the porous rock, we must apply a pressure greater than the reservoir pressure, and even with 10 MPa of excess pressure, the CO2 does not rush into the aquifer. It seeps in. An aquifer five times the size of Prudhoe Bay, along with four separate injecting wells, would be required to sequester the CO2 produced by a single 500-MW coal-burning power plant.

Previous work on the subject suggested that we could push CO2 into an aquifer ten to a hundred times faster. Economidese et al. mention two problems with previous work. One problem is that previous work confused the volume of the pores in the aquifer with the volume of the aquifer. The volume of the pores is only 20% of the volume of the aquifer. This confusion led to an over-estimation of the sequestration rate by a factor of five. The other problem is that previous calculations assumed that the pressure at the boundaries of the aquifer was constant. If the pressure at the boundaries is constant, we certainly can push CO2 in faster, perhaps ten times faster. But such an assumption implies that water, and ultimately CO2, will escape the aquifer at the boundaries. In other words: you can fit a lot more CO2 into an aquifer if you allow it to leak out at the edges, but that rather defeats the purpose of sequestration.

In short: the paper concludes that carbon sequestration is impractical. The state of Rhode Island would require an aquifer as large as itself just to sequester 20% of the CO2 generated by its own electricity consumption.

UPDATE: My father is a retired chemical engineer. He was a pioneer in liquified natural gas. He designed the large LNG tank outside Boston, where I live. He also studied CO2 sequestration twenty-five years ago. He read Economides et al. and "did not think much of it". He says there are better places to put CO2 than the aquifers discussed in the paper.