Report: Field Pilot of Low Carbon Concrete

Overview

The OCP-supported project aims to de-risk the adoption of low embodied carbon concrete in data centre slabs, an application treated as high risk. The July 2026 report describes four full-scale slabs placed on July 31 and August 1, 2024 at Wiss, Janney, Elstner Associates' campus in Northbrook, IL. A contractor, a ready-mix supplier and consultants took part, and the mixes were judged against a collective hyperscaler specification.

The four mixes: Mix 1 (control) is a typical data centre mix with ASTM C595 Type IL cement at 517 lb/yd3 and w/cm 0.50, with embodied carbon of 214 kg CO2e/yd3. Mix 2 keeps w/cm 0.50, cuts total cementitious content to 491 lb/yd3 and replaces 40% of the Type IL cement with ASTM C989 slag cement (149 kg, 30% lower). Mix 3 is about 80% ASTM C1157 performance cement and 20% Type IL at 605 lb/yd3 and w/cm 0.35 (121 kg, 43% lower). Mix 4 is C1157 alone at 750 lb/yd3 and w/cm 0.34 (92 kg, 57% lower).

The laboratory programme covered fresh and hardened properties: slump, air, unit weight, bleed, heat of hydration, setting, compressive and flexural strength, drying shrinkage, cracking resistance, coefficient of thermal expansion and bulk resistivity. All mixes met the typical specification limits of maximum 3.0% air, 5 to 9 in slump and at least 4,000 psi at 28 days. Mixes 3 and 4, however, were extremely thixotropic, lost workability quickly and produced almost no bleed water, which raises early-age plastic shrinkage risk and complicates finishing timing.

In the field all four mixes were pumped, placed, finished and reached design strength. Mixes 3 and 4, the lowest-carbon options, lost workability within 15 to 20 minutes of placement, making surface closing, screeding, floating and troweling difficult. The report concludes that, as formulated, they cannot be used effectively for traditional slab-on-ground construction without changes that improve workability or adjustments to construction practice. The scale is small and rests on a single placement event.

Key points

1. Embodied carbon (kg CO2e/yd3): Mix 1 control 214; Mix 2 149 (30% lower); Mix 3 121 (43%); Mix 4 92 (57%). Against the national 236 kg and regional 231 kg benchmarks Mix 4 is 61% and 60% lower.
2. Mix 2 (40% slag cement) behaved closest to the control and was judged the best performer by the contractor and WJE observation.
3. Mixes 3 and 4 were extremely thixotropic and lost workability quickly; in the field workability dropped 15 to 20 minutes after placement.
4. Mix 3 set very fast (accelerator was reduced) and Mix 4 had a delayed final set; both showed little to no bleed water.
5. 28-day laboratory compressive strength: Mix 1 5,870 psi, Mix 2 5,750, Mix 3 8,100, Mix 4 4,140 psi against the 4,000 psi minimum; Mix 4 has a very thin margin.
6. Drying shrinkage met the specification in all four mixes, with markedly lower shrinkage in Mixes 3 and 4. Restrained-shrinkage cracking potential is moderate to high for Mixes 1, 2 and 3 and low for Mix 4.
7. Floor flatness FF: Mix 1 44.6; Mix 2 44.9; Mix 3 33.5; Mix 4 34.1. Most hyperscalers require overall FF 35 and one requires 25; Mixes 3 and 4 fell slightly short of some requirements.
8. Abrasion testing showed very little abrasion depth for Mixes 2, 3 and 4, so a densified, abrasion-resistant surface is achievable with lower-carbon mixes.
9. Heat of hydration is lower in Mixes 3 and 4, which matters little for this indoor project but helps mass concrete.

Using it in design

• Write the slab embodied-carbon target into the specification together with an EPD requirement and tie it to national or regional benchmarks as a percentage; a 30% reduction against the control is a low-risk starting point.
• When proposing slag cement replacement (for example 40%), require control-like w/cm and strength behaviour as in Mix 2; require a laboratory trial before moving to C1157-heavy mixes.
• Ask for a pre-construction laboratory package beyond standard tests: setting times, bleed, heat of hydration, drying and restrained shrinkage, thermal expansion and workability loss over time.
• Review the floor flatness requirement (FF 35 or 25) alongside lower-carbon mixes, and plan finishing time, crew size and measures such as evaporation retarder for high-binder, thixotropic mixes.
• Leave a strength margin for site variability on borderline mixes such as Mix 4, and for over-strength mixes such as Mix 3 consider cutting binder content to lower carbon further.
• Verify with the structural and construction consultant whether the reduced shrinkage cracking can support wider joint spacing.

Who should read it

Data centre structural and construction engineers, sustainability managers, concrete suppliers and contractors.

Document contents

1. Executive Summary
2. Introduction
3. Compliance with Open Compute Project Tenets
4. Laboratory Evaluations
5. Raw Materials Characterization Results
6. Concrete Physical Testing Results
7. Field Placement
8. Concrete Placement
9. Field Quality Control Testing
10. Field Abrasion Testing
11. Appendix A. Environmental Product Declarations
12. Appendix B. Raw Material Characterization
13. Appendix C. Lab Testing - Plastic Concrete Properties
14. Appendix D. Lab Testing - Physical Testing Results
15. Appendix E. Slab Design Drawings and Locations
16. Appendix F. Field Testing - Physical Testing Results
17. Appendix G. Field Abrasion Testing

Related standards

• ASTM C595
• ASTM C989
• ASTM C1157
• ACI 302.1R
• OCP Embodied Carbon Disclosure Base Specification

Read the document

Publication details

Authors

Chhiba, P. (Google); Kusuma, J. (Meta); Masri, F. (Google); Stanley, B. (Microsoft); key contributor Kim, J. (WJE)

Publisher

Open Compute Project Foundation

Year

2026

Document language

English

Source type

Industry body

Access

Open licence — downloadable

Licence

CC BY-SA 4.0 — share with attribution under the same licence

Licence evidence

www.opencompute.org

Citation