Some Observations relating Kinetics, Chemistry, and Product - - PowerPoint PPT Presentation

Some Observations relating Kinetics, Chemistry, and Product Structure of Hydrating Cement Paste – Reaction Mechanisms

Hamlin M. Jennings

MSE and CEE Northwestern University

July 09 Kinetics Summit

Mechanism of reaction

• Must explain kinetics

– Divided into several periods

• Must explain thermodynamics

– Equilibrium

• Must explain location of reaction
• Must explain morphology

Required Steps in Reaction

• Dissolution
• Diffusion
• Precipitation
• Growth
• Location, morphology, and rate of

growth of product(s) must be considered

Rate of hydration (kinetics)

Traditional view

Kinetic description – 5 periods

Acceleratory Period Deceleratory Period

Rate of hydration (kinetics)

Time (hours) Rate of reaction Induction (dormant) 2 Nucleation And Growth Diffusion 10 24 From a mechanistic view there are 3 periods:

There is an induction period under some conditions

10 20 30 40 2 4 6 8 Time (days) 1% sugar 0% sugar Rate of heat evolution (J/(h*g initial cement))

Garci Juenger and Jennings CCR 2002

Hydration of C3S two products form

C3S +(3 x + y)H Cx S Hy + (3 x)CH

X = 1.7 Y = 4

• Volume of solids increases
• Total volume decreases (Chemical shrinkage)

C3S + 72.5 5.3H 95.4 167.9 166.9 C1.7SH4 + 124 1.3CH 42.9 Phase volume: Total:

Why is there an induction period?

• Protective layer forms and is later disrupted

– For

• Physical evidence
• Equilibrium and if so with what?

– Against

• Other Physical evidence
• No obvious reason for disruption

Hypothesis #1

Why is there an induction period?

Induction period X C3S Equilibrium in C3S Both under-saturation and over-saturation return to phase line Implies equilibrium with a protective layer

0.1 1 10 100 1000 5 10 15 20 25 30

[SiO

2

µ M [CaO] (mM) Aqueous Phase C-S-H + Aqueous Phase

S M

After Jennings et al ICCC Sweden 1997

The concentration on M is stable for weeks

H.M. Jennings, “queous Solubility Relationships for Two Types

• f Calcium Silicate Hydrate,” Journal of the American Ceramic

Society, 69 [8] 614‑618 (1986).

But is layer protective? SEM dry (two products form)

2 hrs 4 hrs

0.2 µm 0.2 µm

Layer

• Layer prevents high concentrations in

aqueous phase

– Equilibrium is established quickly

• Inconsistent physical evidence

Why is there an induction period?

• Delayed nucleation (Le Chatelier)

– For

• Ca++ concentration
• Kinetics – nucleation and growth

– Against

• What is the seed (CH or C-S-H)? nothing works

Hypothesis #2

Why is there an induction period?

[Ca] Time

Supersaturation in calcium at early times

End of the Induction Period

CH saturation Retarders poison the precipitation

CCaO (mM)

5 10 15 20 25 SiO2 1 10 100 1000

A C' C'' C Ca(OH)2

Phase diagram: equilibrium but some supersaturation

Curves C, C', C" and A represent a spectrum of C-S-H structures Tobermorite-like

• short chain length -

no Ca-OH

5 A

Jennite-like

• long chain length
• more Ca-OH

5 A

A Silicate polymerization is key to variations

Chen, J.J., J.J. Thomas, H.F.W. Taylor, and H.M. Jennings,

• Cem. Concr. Res., 2004. 34(9): p. 1499-1519.

C-S-H nor CH accelerate much

Gartner and Gaidis Materials Science of Concrete I (1989)

Acceleration period

• Nucleation and growth has dominated

modeling

– Avrami: transformation throughout volume – Boundary: transformation starts at boundary

• Diffusion control -- very little argument

Late period

Location and Morphology

• One valiant attempt in the 1979’s

– Reverse silicate garden -- membrane forms and breaks from osmotic pressure resulting in the formation of needles

• Double et al.
• Birchall et al.
• Otherwise not much, with confusion
• ver morphologies such as Hadley

grains

What is new?

• C-S-H is colloid

– Detailed model of density and pore structure – Packing and morphology change with time which explains many morphological variations

• Kinetics described exactly by boundary N+G
• Seed -- C-S-H can work well -- nucleates in

volume

What is new: C-S-H is gel

1 um Wet TEM taken at Imperial College London (1980) Small particles after Powers

Micrographs of shrinkage Huge deformation on drying

Wet Dry 3 day .5 w/c

Fig 4

Surface area, density -- colloid model

HYPOTHESIS: So what is the mechanism controlling early rate?

• Layer, thermodynamically separating

particle from aqueous exists

• Normal hydration starts when nuclei

form in the layer -- N+G controls rate

• Autocatalytic growth of product into

pore space -- both CH and C-S-H

Must explain kinetics Picture of self stimulation = kinetics (middle)

Must explain retarders: Sugar

• Thomas and Birchall showed that sugar

poisons C-S-H

• Under normal conditions nucleation
• ccurs in layer where some degree of

supersaturation exists

Delayed addition of sugar greatly reduced effectiveness = nuclei formed within layer --

H.M. Jennings, H. Taleb, G. Frohnsdorff, and J.R. Clifton), Proceedings of the 8th International Congress on the Chemistry of Cement, Rio de Janeiro, Brazil, III 239‑243, (1986)

Must explain accelerators

Seed from soluble salts - dispersed with active surface

(Jeffrey J. Thomas, Hamlin Jennings, and Jeffrey J. Chen), Journal

• f Physical Chemistry C, 113, 4327-4334 (2009).

Must explain location Schematic of N + G

(Jeffrey J. Thomas, Hamlin Jennings, and Jeffrey J. Chen), Journal of Physical Chemistry C, 113, 4327-4334 (2009).

Figure 3: SEM micrographs of hydrated paste made without C

• S-H seed (left) and with 2% C -S-H seed by mass
• f C3S (right), after [ Error! Bookmark not defined. ]. Both pastes are 28 d old and were made at w/c = 0.5. Black

is capillary porosity, grey is hydration product, and white is unreacted C

• 3S. Note the much lower amount of

capillary porosity in th e seeded paste at right.

Surface area m easured by small angle neutron scattering (m2/cm 3), one year old paste of white Portland cement paste , w/c=0.5 Compressive strength (kN/m 2) 16 day old paste of

• rdinary

Portland cement paste , w/c=0.5 (from [ 1])

No additives 123 35 Additive 143.5 40

[1] Millea, J. The Effects of Calcium Silicate Hydrate Seed on the Compressive Strength of Portland Cement Past e, Senior Thesis, Northwestern University, Evanston, 2006.

Thomas, Jennings, Chen, Physical Chemistry C, 2009

First principles control of microstructure

Must explain morphology

Slow and fast drying = very open packing at early time

3 Day Old - rapid dry 3 Day Old - 18 day dry

Fonsica and Jennings submitted

Seed activates Slag when soluabilized

Late Reaction

• Possibly not diffusion control
• Rate controlling step is difficulty in

finding active sites or, equitantly the nucleation process just continues on slowly

Consumption of active sites

• n particles

Not diffusion: D2O “Effects of Deuterium Oxide and Mixing on the

Early Hydration Kinetics of Tricalcium Silicate,” (J.J. Thomas and H.M. Jennings), Chemistry of Materials, 11, 1907-1914 (1999).

Summary: Reaction kinetics /

Mechnaism and control of microstructure

• Nuclei must form from some degree of

supersaturation -- normally within layer

• Nuclei once formed stimulate new

product when surface is available -- supersaturation not required

• Formation of nuclei can be poisoned but

active surface can not

Particle dissolves If seed with active surface Stimulates new CH and C-S-H product growth Growth continues until active sites exhaust If no seed Nuclei form on surface of cement in layer Supersaturation – likely in layer

Reaction Steps

Overview

• Dispersed seed accelerates - diffusion into

pores does not seem to be rate limiting

• Sugar retards -- poison formation of nuclei

– Sugar does not prevent autocatalytic growth – Seed, if formed and dispersed, trumps sugar

• Seed accelerates activated slag
• But seed must have active surface --

prehydration does not work well because much surface is not available

Monolayer of water IGP Interlayer water C-S-H dry 2.85 g/cm3 A B C

Fig 1

D Globule

Model of particles and pore

CM: Colloid Deformation mapping

Dry to 50% rh Dry to 5% rh

Total shrinkage is sum of shrinking and restraining phases

C.M. Neubauer and H.M. Jennings, J. Mater. Sci. 35, 5741 (2000)

ESEM of Wet Samples

0.5 Hours 8 Hours

40 80 120 160 100 200 300 10 20 30 40 50 60 70

Surface Area (m2/cc) Heat Evolved SANS Surface Area Heat Evolved (Joules) Hydration Time (hours)

Surface area development and heat evolution*

OPC Paste, 20ºC

*J.J. Thomas, H.M. Jennings and A.J. Allen, Cem. Concr. Res. 28, pp. 897-905 (1998).

SANS, LOI, and N2

LD early

Mikhail and Abo-El-Enein (1972)

Aging 125 days

Two densities

1um

N+G in two areas

• On surface of particles

– See Jeff Thomas

• In volume between particles