About This Project
Concrete, produced at 40 gigatons annually, can naturally soak up CO2 to form a scalable, durable and verifiable carbon sink. However, this process is limited by CO2 diffusion and hydration. We are engineering carbonic anhydrase enzymes to overcome this kinetic bottleneck. Through machine learning, metagenomics, and high-throughput testing, we will optimize enzymes to unlock additional gigaton-scale carbon sequestration within our built environment.
Ask the Scientists
Join The DiscussionWhat is the context of this research?
Concrete, produced at 40 gigatons annually, is the most consumed man-made substance. Concrete production presents both a challenge, accounting for ~8% of global emissions, but also an opportunity. Specifically, its alkaline, calcium-rich composition confers the ability to serve as an annual gigaton-sink for atmospheric carbon removal (ref1, ref2). However, natural carbonation of CO2 is too slow to meaningfully mitigate climate change. By leveraging carbonic anhydrase (CA), a highly efficient enzyme for hydration of CO2, the project aims to dramatically accelerate this process. Through a combination of natural enzyme discovery, machine learning-driven design, and directed evolution, we seek to create enzymes capable of surviving in cement and significantly boosting CO2 uptake.
What is the significance of this project?
We see multiple implications:
1. Increased concrete carbonation rates have the potential to form a gigaton-scale carbon sink (ref).
2. Compressive strength improvements due to carbonation (10-20%, ref1; ref2) can significantly reduce the required amounts of cement that account for 8% of global emissions.
3. With protein production being an established industry, adding carbonation catalysts could be adopted rapidly with minimal changes to existing concrete infrastructure and supply chains.
Finally, alkaline-stable CO2 hydration catalysts are expected to aid in other CO2 sequestration contexts, such as during enhanced rock weathering, liquid direct air capture and end-of-life concrete carbonation, as well as confer self-healing properties to concrete to reduce material replacement needs.
What are the goals of the project?
The project’s overarching goal is to develop biologically enhanced concrete capable of sequestering significant amounts of CO2. To achieve this, the team will engineer carbonic anhydrase enzymes that function optimally in the high-pH environment of cement.
The research plan manages risk with four key aims:
1. Building a comprehensive database, including sampling of natural extreme environments, to identify enzymes adapted to such harsh conditions.
2. Using machine learning to design novel carbonic anhydrases tailored for concrete.
3. Implementing a high-throughput screening platform to identify the most effective enzyme variants.
4. Testing the effect of enzyme addition in concrete to measure the impact on CO2 absorption and material properties.
Budget
We are taking 4 approaches towards engineering carbonation catalysts (carbonic anhydrases) to enable additional, verifiable, durable and scalable carbon removal within concrete:
Gathering deep metagenomic information to discover novel carbonic anhydrases functioning at high pH, and sourcing publicly available metagenomic carbonic anhydrase variants.
Leveraging state-of-the-art protein machine learning models to design high pH carbonic anhydrases.
Systematic, large scale synthesis and experimental testing of 10.000 - 100.000s of carbonic anhydrases for concrete conditions
Material property testing of CO2 uptake and compressive strength/durability testing of carbonated concrete materials.
Project Timeline
We will perform at least 2 rounds of sourcing and screening of ~10,000s semi-rationally diversified, ~1000 metagenomic and ~1000 model-designed carbonic anhydrases. This allows for iterative fine-tuning of protein machine learning models. We will then select the top 10 variants to perform testing within concrete samples. This enables determining the carbon uptake rate of concrete, as well as testing for critical composition and structural properties.
Dec 31, 2024
First round screening of ~1500 carbonic anhydrases, data analysis and iterated model training.
Apr 15, 2025
Second round carbonic anhydrase sequence design, synthesis and testing.
Jul 20, 2025
Carbonation testing for CO2 absorption of catalyst amended concrete samples.
Oct 20, 2025
Structural property testing (compressive strength) and compositional analysis (TGA, XRD) of catalyst amended concrete samples.
Meet the Team
Team Bio
This is a multi-institutional collaborative effort among:
Protein Design: David Baker, Seth Woodbury, Jacob Gershon, Joe Min; Debora Marks, Pascal Notin.
Protein Screening: David Savage, Andrew Hunt, Simon D'Oelsnitz .
Metagenomics: Jill Banfield, Luis Valentin, Andreja Kust, Jordan Hoff.
Civil Engineering: Jiaqi Li, Berkeley Concrete Structures Lab. We will establish further collaborations and have discussed this project with leading concrete experts Paulo Monteiro and Claudia Ostertag.
David Ding
David studied biochemistry and computational genetics at Oxford, and performed his PhD with Debora Marks (Harvard) and Michael Laub (MIT). There he built high-throughput selection assays to measure the function of 100,000s of protein variants, as well as computational methods for predicting protein function. He’s now a postdoctoral scholar at UC Berkeley with Dave Savage, where he works on enzyme engineering towards sustainability applications.
Prior work can be found here.
Jung-Un Park
Jung-Un is an expert in biochemical characterization of enzymes.
Prior work can be found here.
Lab Notes
Nothing posted yet.
Additional Information
Project Backers
- 0Backers
- 0%Funded
- $0Total Donations
- $0Average Donation