Posted in: Chemistry, IBM Research-Brazil, Nanotechnology, Natural Sciences, Physics

Big oil’s next gusher is at the nanoscale

Last year the world consumed almost 97 million barrels of oil per day. What if I told you that many more barrels still remain in those same wells? Deep inside the rock, 60 percent and more of a reservoir’s oil remains trapped in capillaries which are sometimes only tens to hundreds of nanometers wide (For comparison: DNA is 2.5 nanometers wide). It’s because of the porous nature of sandstone and shale that oil can settle into sedimentary rock. But really understanding how to get the oil out of these capillaries has been impossible – until now.

My industrial technology & science team based in Rio de Janeiro published a study in Scientific Reports, Adsorption energy as a metric for wettability at the nanoscale, explaining how the properties of liquid oil molecules behave in completely different and unexpected ways when in contact with a solid, at the nanoscale. Everything the industry knows about how to extract oil, such as calculating the energy it takes for extraction, turns out to be different at the nanoscale.

Simulating and measuring wettability weirdness

At attoliters (10-18), a droplet of liquid ceases to look like what we imagine: spherical or teardrop shapes. Instead, our research found that, ultimately, the nanoscale oil droplet looked much more like a flat film against a solid surface. This increased surface area turned out to represent much more “wetting” than had been accounted for in typical macroscopic measurements. And not only was there more surface coverage in these flat nano-droplets than previously thought, the standard simulation tools and techniques did not take into account the increased energy required to extract these oil molecules.

Figure 3. Droplet adsorption energy: underestimated at the nanoscale. (a) Comparison of the actual surface with an idealized spherical cap fit to the same data. (b) Difference between adsorption energy for the actual surface and that of the spherical cap approximation. The negative differences indicate that the fitted spherical cap underestimates adsorption energy, just as it underestimates contact area (see Fig. 2). For volumes larger than 106 nm3 a spherical cap fit provides a robust estimate of the adsorption energy α.

Figure 3. Droplet adsorption energy: underestimated at the nanoscale. (a) Comparison of the actual surface with an idealized spherical cap fit to the same data. (b) Difference between adsorption energy for the actual surface and that of the spherical cap approximation. The negative differences indicate that the fitted spherical cap underestimates adsorption energy, just as it underestimates contact area. For volumes larger than 106 nm3 a spherical cap fit provides a robust estimate of the adsorption energy α. (Notes: Figure 3b acronyms: AFM-Atomic Force Microscope Measurement, AAMD-All Atom Molecular Dynamics Simulation, CGMD-Course Grain Molecular Dynamics Simulation. Image reprinted from Scientific Reports’ “Adsorption energy as a metric for wettability at the nanoscale”)

Unearthing the nano-level shape change led us to develop oil flow simulations that could better-predict oil extraction from a reservoir.

IBM, though, isn’t an oil and gas company. We don’t have all the data about the materials, core plugs, and specific reservoirs that an oil company would consider its core data. So, to build a computational representation of a reservoir at the nanoscale (video, below), we took rock characterization data from public repositories, such as ETH Zurich’s Rock Physics Network. Then, based on the “reservoir template” made from the geometrical data, we are now able to deploy the nanoscale wetting and flow science that had not been done before.

 

We then showed this new template to oil and gas companies to demonstrate how our nano-flow simulation takes into consideration the properties of the oil trapped in the capillaries of their wells. And while the simulation does not suggest how to extract all of the trapped oil, it does offer different techniques and materials to explore that might help to extract about 1 percent more. In Brazil, which pumps 2.4 million barrels of oil every day, that 1 percent increase in production would add 24,000 more barrels to the daily total – and 8.8 million more barrels every year.

From flow simulations to oil-filtering chips

In our paper, the simulation was calculated using massively parallel processing on Blue Gene – we are now redistributing the simulations to be delivered through the IBM Cloud.

Our wettability discovery is an important step to help oil and gas companies to recover more than the industry average of 40 percent of the oil trapped in their reservoirs. The next step is to study the flow of oil in nano-capillaries. To that end, we have developed an integrated chip platform that enables us to experimentally validate and calibrate nanoscale flow for building better flow simulations (read our paper presented at the 2016 Rio Oil & Gas Expo & Conference: Multiscale Science Enables High-Accuracy Simulations Of Enhanced Oil Recovery).

To do this, we need to scale up: first, we need a physical measurement of a capillary network from a scanning electron microscope, or an x-ray computed tomography. Then, with the pore network’s data, we use an experimentally calibrated flow simulation to determine how much pressure is necessary to pump water, including customized chemicals specifically designed to separate oil from rock, through the nanoscale pore network – and eventually to push out the oil (for which we have a patent: Method and integrated device for analyzing liquid flow and liquid-solid interface interaction).

Today, the industry relies on incomplete physical models to predict oil recovery at their wells. And it could significantly improve its return on investment with higher-accuracy oil recovery predictions. Our research offers a way to improve prediction models to better-account for the oil confined at the nanoscale, which is of particular importance in non-conventional reservoirs. Accounting for the nanoscale, now, could mean another 1 percent yield in oil recovery. And eventually, with better simulation technology and functional materials, perhaps we can get closer to recover the remaining 59 percent, too.

Read more about the work we’re doing at our new NanoLab lab in Rio, here.

Comments

Add Comment

Your email address will not be published. Required fields are marked *

Manager, Industrial Technology & Science IBM Research

Mathias Steiner

Manager, Industrial Technology & Science, IBM Research-Brazil