PhD Defense – Chirayu Kothari

Advancing Characterization of Cementitious Systems via Raman Imaging

Advisor: Professor Nishant Garg

Abstract

Rising urbanization and population growth are driving the demand for concrete infrastructure, with

the global building stock expected to double by 2060. Meeting this demand will require producing

large quantities of cement, a manufacturing process that is highly energy-intensive and a major

source of anthropogenic CO2 emissions. These challenges can be partially addressed by developing

durable, alternative cementitious binders. Achieving this goal requires a fundamental

understanding of their mineralogy, hydration behavior, and microstructural development, as these

factors collectively govern the binder performance. However, many of the mineral phases within

cementitious systems responsible for these properties are polymorphic, nanocrystalline,

amorphous, or spatially localized, making them difficult to fully identify and quantify with

conventional bulk characterization techniques alone. This dissertation addresses these challenges

by developing Raman imaging as a complementary, spatially resolved characterization platform

for investigating cementitious systems in anhydrous, hydrating, and deteriorating states.

A Raman imaging workflow was first developed to identify and quantify phases in anhydrous

ordinary Portland cements (Type I/II) using large-area scans (5 mm x 5 mm, 250,000 spectra). A

spectrum-based binarization algorithm was introduced to reduce operator subjectivity and improve

pixel labeling by incorporating signal-to-noise (SNT) and epoxy-to-signal (EST) thresholds. It was

observed that the choice of analysis parameters can greatly influence the final quantitative data

obtained from a given technique. Quantitative comparison with X-ray diffraction (XRD)-based

Rietveld refinement showed excellent agreement for typical Type I/II cement phases (R2 > 0.99,

Δwt.% < 2.5%) at SNT of 5 and EST values ranging from 5 to 15. Representative area analysis

further showed that a scan area of 9 mm2 can provide reliable quantification while reducing scan

time from approximately 8 h to 3 h.

Building upon this quantitative framework, the methodology was extended and applied to

anhydrous calcium sulfoaluminate (CSA) cements, chemically complex binders containing a

variety of polymorphs, solid solutions, and phases with overlapping diffraction patterns. Raman

imaging and XRD provided complementary information for initial phase identification, while

independent quantification showed high correlation (R2 > 0.99) and agreement across techniques

(Δwt.% < 5 wt.%). Although the individual orthorhombic and cubic polymorphs of ye’elimite

differed between the two techniques, the total ye’elimite content differed by only ~1.5 wt.%.

Beyond phase quantification in anhydrous systems, Raman imaging was also applied to investigate

the hydration mechanism in magnesium oxide (MgO)/nesquehonite (Nq) binders, which are

emerging low-carbon cementitious materials whose cementation is associated with the formation

of a carbonate-rich microstructure. Using in situ underwater Raman imaging, a previously

unreported dissolution-precipitation pathway was revealed in which nesquehonite addition

promotes the formation of amorphous magnesium hydroxy carbonate (AMHC) on MgO particle

surfaces, followed by the growth of hydrous carbonate-containing brucite (HCB). Specifically,

AMHC serves as a carbonate-rich intermediate, supplying carbonate to form HCB, the structurally

disordered binding phase. These findings provide a novel mechanistic insight into cementation and

carbonate speciation within MgO/Nq systems.

Finally, confocal Raman imaging was used to investigate alkali-silica reaction (ASR), one of the

most pervasive durability-related degradation mechanisms in concrete. Spatially resolved analysis

revealed variations in the polymerization state of amorphous ASR gels, with Q3/Q4-rich gels in

aggregate cracks and calcium-modified Q2-rich gels near the aggregate-paste interface. These

observations provide direct molecular-scale evidence of gel calcification and spatial heterogeneity

of deleterious amorphous reaction products in mortar samples. Characterizing the spatially

resolved speciation of these gels is a stepping stone towards understanding the ASR-related

damage mechanism.

Collectively, this thesis establishes Raman imaging as a versatile and complementary

characterization approach for quantifying clinker mineralogy, resolving hydration mechanisms,

and investigating durability-related degradation processes. By linking micron-scale spatially

resolved chemical information with phase evolution across multiple stages of a cementitious

material's life cycle, the methodologies developed in this dissertation provide new quantitative

insights into cement chemistry and offer a framework to accelerate the design, optimization, and

implementation of next-generation cementitious materials.