Bragg Institute
Vanessa.peterson@ansto.gov.au
Studying the Structure and Hydration Kinetics of Cement 
Systems in Real-time using Neutron Scattering
Vanessa Kate Peterson
boxtargetboxtargetCement Research
head2righthead2right Understand these sensitivities ? more robust material
head2righthead2right Optimizing chemical processes ? better, cheaper product
? Many aspects of cement are unknown 
? Cement is complex - and sensitive.
? Many factors affect the final product
? Controlling these factors is difficult. 
boxtargetboxtarget
What is cement clinker?
? Tricalcium silicate is the main component of cement
Tricalcium silicate
Other major components
> 50 %
? Up to 90 %
Dicalcium
silicate
Tricalcium 
aluminate
Tetracalcium
aluminoferrite
Tricalcium silicate is the fundamental hydraulic 
component of cement
boxtargetboxtarget
Ca3SiO5 + (3 + y ? x)H2O       (CaO)x(SiO2)(H2O)y + (3 - x)Ca(OH)2
Hydration of Ca 3 SiO5 : Following the hydrogen
Real Time 
Hydration: 
30 minute 
time slices
? Applying Quasielastic Neutron 
Scattering to follow the hydration
? Incoherent scattering from H
boxtargetboxtarget
QENS: NIST Center for Neutron Research
Neutron time-of-flight Fermi Chopper Spectrometer (FCS)
? Sample 45 ? to the beam
? Reflection geometry data 
only used 
? Incident beam ? = 4.8 ?
? Beam is pulsed by a 
chopper ? timestamp
? Scattered neutrons arrive at 
a bank of detectors: 2.29 m
? The detectors record the 
neutron arrival time
boxtargetboxtarget
QENS data collection & treatment
Incident neutrons, 4.8 ?
Scattered neutrons:
Energy and momentum
At high Q, QENS spectrum is Q independent. 
Data averaged over a Q range (1.9?2.4 ??1):
- focus on rotational dynamics of water.
- increase the signal-to-noise ratio
boxtargetboxtarget
Ca3SiO5 + (3 + y ? x)H2O       (CaO)x(SiO2)(H2O)y + (3 - x)Ca(OH)2
QENS: Following the hydrogen
Total Hydrogen =
Bound + 
Constrained +
Free + 
boxtargetboxtarget
-0.5 0 0.5
Total H
Bound H
Constrained H
Free H
? Contribution to each QENS spectrum: 
States of hydrogen H2O
H Adsorbed on 
surfaces
States of hydrogen: Profile fitting
Products
meV
head2right Fitted from -2 to 2 meV
boxtargetboxtargetIntegrated areas: Quantitative information for each H state
S(Q,?) =
( )
2
B
0
354.2/W
xx
2
1-
2
B
e
354.2/W2?
B ???????? ?
head2righthead2right Very wide and wide Lorentzians: Free H
head2righthead2right Narrow Lorentzian: Constrained H 
head2righthead2right Gaussian: Bound H
+ ( ) ( )2
C
2
0
C
2/Wx-x
C
?2
W
+
+ ( ) ( )2
F
2
0
1F
2/Wx-x
F
?2
W
1
1
+ + ( ) ( )2F20
2F
2/Wx-x
F
?2
W
2
2
+
Quantitative determination of H: profile fitting
x = energy transfer and x0 is the peak center ? constrained to be the same 
WB: Fixed to instrument resolution (0.147 meV)
WC: Determined for an average of the last 7 spectra
WF1 and WF2: Determined from an average of the first 7 spectra
Fixed
boxtargetboxtarget
QENS profile fitting: Bound Water Index
t = 1 hour t = 40 hours
? Can quantitatively derive the H in each state
Bound + Constrained H
Total H
BWI =
Allows the 
hydration to be 
followed  
boxtargetboxtarget
Experimental: QENS sample preparation
? Water to solid mass ratio of 0.4
? Mixed by hand for 3 minutes
? Sealed in teflon film 
? Placed in can, sealed using In
? Can placed in a closed cycle He 
refrigerator, 30 ?C.
? Data continuously collected for 50 
hours on Fermi Chopper neutron 
Spectrometer (FCS).
boxtargetboxtarget
Tricalcium silicate hydration
1. Induction and dissolution
Ca2+
OH-
H2SiO2-4
H2O
C-S-H
+ 
Ca(OH)2
3. Diffusion limited 
hydration
+ H2O 
?
C-S-H
+ 
Ca(OH)2
2. Nucleation 
and Growth
boxtargetboxtarget
-0.025
0.025
0.075
0.125
0.175
0.225
0.275
0.325
0 10 20 30 40 50 60 70
Nucleation and growth
Following hydration: BWI with time
Q u
a n
t i t y
 o f
 b o
u n
d  H
Diffusion limited
hydrationB W
I
Time (hours)
Time to 
Nucleation 
and Growth
Time to Diffusion 
Limited Hydration
Duration
boxtargetboxtarget
? On heating unit cell expands, then phase transitions occur
? Polymorphs stabilized in clinker at RT by impurity ions
T1 ? T2 ? T3 ? M1 ? M2 ? M3 ? R
620 ?C           920 ?C            980 ?C          990 ?C       1060 ?C            1090 ?C
Tricalcium silicate polymorphism
Different forms exhibit differing strength!
? Compressive strength has been linked to:
Crystal type - symmetry
Type of stabilizing ion
boxtargetboxtarget
QENS: Hydration of tricalcium silicate forms
QENS data and hydration models for Mg stabilized T1 and M3 C3S 
? Significant 
differences between 
hydration behaviors
? Monoclinic sample 
has half the surface 
area
Peterson, V. K. et al. Chem. Phys. 326, 2006.
boxtargetboxtarget
QENS: Hydration of C3S forms
QENS data and hydration models for Mg stabilized T1 and M3 C3S 
Peterson, V. K. et al. Chem. Phys. 326, 2006.
0.20121.6M3
0.188.21.3T1
Rate of 
product 
formation
(dBWI/dt)
Time to 
DL
(hours)
Time to 
NG
(hours)
? Monoclinic has a longer Nucleation and Growth (NG) period.
? Broad tapering into the Diffusion Limited hydration from the NG period is 
typical for a broader particle size distribution (consistent with PSD).
boxtargetboxtarget
-0.025
0.025
0.075
0.125
0.175
0.225
0.275
0.325
0 10 20 30 40 50 60 70
Data
Model
Kinetic parameters of the hydration
BWI(t) = BWI(0) + A[1-exp{-[k(t-ti)]n}]
Modeling of the nucleation and growth period:
Parameter ?A?: Max. product at 
infinite timeB
W
I
Hours
NG Rate constant ? Intrinsic reactivity 
(volume independent)
boxtargetboxtarget
-0.025
0.025
0.075
0.125
0.175
0.225
0.275
0.325
0 10 20 30 40 50 60 70
Data
Model
Kinetic parameters of the hydration
Modeling of the diffusion limited period: 
Parameter ?Di?: Diffusivity of 
the water
BWI(t)= BWI(0) + [1-{[1-BWI(td)]1/3? (R-1)(2Di)1/2(t-td)1/2}3]
B W
I B W I
boxtargetboxtarget
QENS: Hydration of C3S forms
QENS data and hydration models for Mg stabilized T1 and M3 C3S 
Peterson, V. K. et al. Chem. Phys. 326, 2006.
190.160.2M3
0.30.250.1T1
Di (10-15m2h-1)k
(hours-1)
A
? Monoclinic:
? Intrinsically less reactive ? the way in which protons are 
accepted by the structure
? Morphological effect - significant permeability of material.
? Slower reaction and higher permeability produces more 
product (longer NG period) 
boxtargetboxtarget
Effective accelerant
Effective retarder
QENS: Effect of additives on C3S hydration
boxtargetboxtarget
Sucrose: known hydration retarder
head2right Increases duration of induction period 
head2right Some suggestion of ?delayed accelerator?
Atoms: 
C, O, H
Mechanisms of effects are uncertain, particularly details on kinetics
Hypothesis:
? Formation of a half-salt that poisons 
surfaces, disallowing nucleation. 
? Intermediate ability to form this half-salt. 
? Stable: does not undergo ring-opening 
(degradation) in the alkaline paste
head2right Other factors
? Chelates with Ca2+ - depressing solution Ca2+. 
? Solubilizes silicate in hydrating cement.
boxtargetboxtarget
Sucrose - QENS
-0.02
0.03
0.08
0.13
0.18
0.23
0.28
0.33
0 5 10 15 20 25 30 35 40 45
Hydration time (hours)
B W
I
Water
Sucrose 
(High conc.)
Sucrose (Low conc.)
Triclinic tricalcium silicate; H2O:cement = 0.4 at 30 ?C 
0.01 and 0.04 wt. % sucrose
boxtargetboxtarget
head2righthead2right Consistent with the literature, commensurate with conc.  except: 
? No evidence for accelerated reaction.
? What about the ?delayed acceleration? ?
Sucrose ? Kinetic Parameters
12.637.825.20.480.08High %
Sucrose
8.811.02.20.400.18Low %
Sucrose
6.67.91.30.600.19H2O
Nucleation 
and Growth 
duration (h)
td (h)ti (h)Di
( x 10-15m2h-1)
k (h-1)
? Retards induction period
? Lengthens the nucleation and growth period
? Nucleation and growth rate constant reduces
? Slightly denser product, consistent with slower production
boxtargetboxtarget
Sucrose ? Kinetic Parameters
12.637.825.20.480.080.0200.212High %
Sucrose
8.811.02.20.400.180.0300.201Low %
Sucrose
6.67.91.30.600.190.0220.155H2O
Nucleation 
and Growth 
duration (h)
td (h)ti (h)Di
( x 10-15m2h-1)
k (h-1)d(BWI)/dt
(h-1)
A
head2righthead2right More product 
head2righthead2right Higher rate of formation during NG Delayed ?acceleration?
head2righthead2right Higher rate of formation only with low sucrose conc.
head2righthead2right k is the instrinsic reactivity ? volume independent
head2righthead2right d(BWI)/dt is volume dependent (peaks at Akn) 
Hypothesis: Solubilization of the silicate species
? Increased ions in solution (more volume, ? d(BWI)/dt) 
? Competes with the rate constant (k) reduction
? Commensurate with sucrose concentration
boxtargetboxtarget
Comparison of Calorimetry and QENS
-0.02
0.03
0.08
0.13
0.18
0.23
0.28
0.33
0 5 10 15 20 25 30 35 40 45
Hydration time (hours)
B W
I
Water
Sucrose 
(High conc.)
Sucrose (Low conc.) ? Both QENS and calorimetry 
find more product at the end of 
NG after hydration with a low 
sucrose conc.
? The % increase was more with 
QENS than with calorimetry
? Calorimetry observes the 
tricalcium silicate dissolution
boxtargetboxtarget
Calorimetry
? Measures the heat evolved during reaction
? Has been correlated to QENS results for the bound hydrogen 
component - reflects chemically bound H associated with Ca(OH)2
and C-S-H
? Markers = QENS 
(Fractional Bound H)
? Line = Calorimetry 
(Fractional cumulative 
heat evolved)
? QENS 
measures 
constrained and
bound H
? Constrained H 
associated with 
a C-S-H phase 
in which water is 
loosely attached
-0.1
0.1
0.3
0.5
0.7
0.9
1.1
0 5 10 15 20 25
F r
a c
t i o
n  o
f  2
5  h
o u
r  v
a l u
e
0
Time (hours)
F r
a c
t i o
n  o
f  2
5  h
 v a
l u e
Time (hours)
boxtargetboxtarget
Small Angle Neutron Scattering and QENS
? Evidence for two different C-S-H morphologies: 
? Chemically bound H (HS) phase = C1.7SHS(1.6)
? Constrained and chemically bound H (HT) phase = 
C1.7SHT2.7 
J.J. Thomas et al, J Am Ceram Soc 84 (2001). 
Results suggest sucrose induces a high-surface area C-S-H phase in 
which water is loosely bound
head2righthead2right SANS for hydrating tricalcium silicate: Correlates the 
constrained H population (QENS) with a high surface 
area C-S-H phase
Soft Synchrotron X-ray Transmission Microscopy results show a high 
surface area C-S-H with unique morphology 
(Juenger et al, Proc. of the 11th Int. Cong. Chem. Cem., 2003)
boxtargetboxtarget
Combined effects: Hydration of different tricalcium 
silicate forms + additives
boxtargetboxtarget
Triclinic Monoclinic
? Same effects, observed to a lesser extent  with monoclinic ? possibly due 
to the lower reactivity of the monoclinic form
Sucrose added to different C3S forms
Results for the monoclinic form support hypotheses from the triclinic study 
boxtargetboxtarget
The accelerator CaCl2
boxtargetboxtarget
Hydration of mixtures
QENS data and hydration models for Mg stabilized T1 and M3 C3S 
? What happens during 
simultaneous hydration?
? Time to NG and k both 
lower for the mixture!
? May arise from different 
product morphologies and 
solution chemistry
Peterson, V. K. et al. Chem. Phys. 326, 2006.
boxtargetboxtarget
Tricalcium silicate gives 
cement early strength
Interaction of Ca3SiO5 and Ca2SiO4
? Tricalcium silicate is responsible for strength up to 28 days
? Dicalcium silicate is ~ 20 wt. % of cement
? Adds strength later (months, years)
Ca3SiO5 is more reactive than Ca2SiO4.
boxtargetboxtarget
head2righthead2right Independent hydration has been studied checkbldcheckbld
head2righthead2right Interactions are unknown !
QENS: Simultaneous hydration of the calcium silicates
Interaction of tricalcium and dicalcium silicate
head2righthead2right Pure tricalcium silicate
head2righthead2right Pure dicalcium silicate
head2righthead2right Intermediate mixtures of the two
Kinetic parameters 
as a function of 
composition
boxtargetboxtarget
boxtargetboxtarget
Kinetic parameters of the hydration
D i  
( m
2 h
- 1 )
 x  
1 0
- 1 6
A
5.5
1.5
3.5
maximum
A (correlates with early strength) = ?
Di (product permeability) = ? 
% C3S
10080604020
Pure C3SPure C2S
boxtargetboxtarget
QENS: Optimization in the amount of product at 80-85 wt % 
tricalcium silicate.
Application of inelastic neutron scattering
head2righthead2right Confirmation by another method: formation of 
more product 
head2righthead2right Ca(OH)2 or CaO-SiO2-H2O: Use Ca(OH)2
Inelastic Neutron Scattering: Can be done with QENS
Local interactions of atoms and molecules
? Compare sample spectrum to a reference 
? Quantitative!
boxtargetboxtarget
Inelastic Neutron Spectrum of Ca(OH)2: 
Vibrational Density of States
? Filter Analyzer Neutron 
Spectrometer (FANS)
head2righthead2right 41 meV represents the main 
Ca(OH)2 phonon mode
? Sample data taken at 22 hours: 
Diffusion Limited hydration
boxtargetboxtarget
Ca(OH)2 amount with mixture composition
Increase in Ca(OH)2
at the ?critical 
composition? 
identified by QENS
100807060 90
% Tricalcium silicate
boxtargetboxtarget
Complementary analysis:
28 day compressive strength testing
? Mortars of the mixtures were 
prepared at 30 ?C 
? The QENS and INS prediction of 
increased strength was tested?
boxtargetboxtarget
0
100
200
300
400
500
600
700
-0.1 0.1 0.3 0.5 0.7 0.9 1.1
Ratio of tricalcium silicate in the mixture
C o
m p
r e s
s i v
e  S
t r e
n g
t h  
( P
o u
n d
s  p
e r  
S q
u a
r e  
I n c
h )
Strength correlates with QENS results
Maximum strength 
occurs at the 
predicted point, 85 %
20            30    50 70             90
% Tricalcium silicate
boxtargetboxtarget
Cement: Significance of research
? These outcomes represent significant breakthroughs:  
? New insights into the fundamental aspects of the kinetics of 
tricalcium silicate hydration that relate to cement properties
? Effects of: Structural Variation, Mixtures, and Additives
References
? Peterson, V. K.; Neumann, D. A.; Livingston, R. A. J. Mat. Res. 21, 1836-1842, 2006.
? Peterson, V.K.; Neumann, D. A.; Livingston, R. A. NIST Center for Neutron Research Annual Report, 14-
15, 2005.
? Peterson, V.K.; Neumann, D.A.; Livingston R.A., J. Phys. Chem. B, 109, 14449-14453, 2005. 
? Peterson, V. K.; Neumann, D. A.; Livingston, R. A., Mater. Res. Soc. Symp. Proc. 840, Q2.2, 2004
? Peterson, V.K.; Stuzmann, P. Livingston, R.A. J. Mat. Res. In Review
? Peterson, V. K.; Neumann, D. A.; Livingston, R. A., Chem. Phys. Lett., 419, 16-20, 2006.
? Peterson, V. K.; Garci-Juenger, M. C., Physica B, . 385-386, 222-224, 2006.. 
? Peterson, V. K.; Brown, C.; Livingston, R. A. Chem. Phys. 326, 381-389, 2006.
? Peterson, V. K.; Garci-Juenger, M. C., Chem. Mater. 18, 5798-5804, 2006.
? Peterson, V. K.; Livingston, R. A.; Neumann, D. A., Physica B. 385-386, 481-486, 2006. 
boxtargetboxtarget
S(Q,?) =
( )
2
C
0
354.2/W
xx
2
1-
2
C
e
354.2/W2?
C ???????? ?
+ ( ) ( )2
P
2
0
P
2/Wx-x
P
?2
W
+
+ ( ) ( )2F20
1F
2/Wx-x
F
?2
W
1
1
+ + ( ) ( )2
F
2
0
2F
2/Wx-x
F
?2
W
2
2
+
Future directions for QENS and cement:
1. Derive kinetics from models rather than fits
? Method used  in this study: Profile fitting
? BWI = (C +P)/(F1 +F2+C+P)
? Kinetic models fitted to BWI versus time plots
boxtargetboxtarget
1. Derive new parameters from QENS models: 
Self-dynamics of water molecules
Fratini et al Phys. Rev. E, 2001.
head2right Self-dynamic structure factor
head2right Related to S(Q,?) via a time-
Fourier transform
? Contains Intermediate Scattering 
Functions describing the water 
molecule motions as per the 
hydrogen atoms:
1. Ftranslational(Q,t): Of the center of 
mass
2. Frotational(Q,t): Around the center of 
mass
boxtargetboxtarget
? Data from Q = 0.55 to 1.24 ?-1, 5 spectra for each measurement. 
? Transmission geometry at low angles, reflection at high angles
])?/t(exp[C
])?/t(exp[)t(F
p
:)t(F
S
l
v
H
?
?
?
?
+
Q,
Q,
Long term component: ??relaxing cage model?? 
? relaxation + stretching, ? = 0-1. 
? Lorentzian when ? = 1
immobile fraction
1. Derive new parameters from QENS models
Fratini et al Phys. Rev. E, 2001. and Liu et al Phys. Rev. E., 2002
Self-dynamic structure factor
Short term translational vibrations of 
central molecule
Short term lth order rotational correlation 
function
Appreciable 
when Q > 1 ?-1
boxtargetboxtarget
1. Derive new parameters from QENS models
Fratini et al Phys. Rev. E, 2001.
? Time evolution of these parameters not modelled
? Bound water component for the same study shows more interesting
trend with less error  
head2right Better definition of these parameters for cementitious systems may lead 
to application of kinetic models 
boxtargetboxtarget
M3
R
T1
Structural differences
Orientation of SiO4 tetrahedra: Change 
hydration behaviour
2. Couple kinetics with tricalcium silicate structure
boxtargetboxtargetTricalcium silicate: structural modulation 
Electron and Synchrotron Powder Diffraction studies have revealed:
T1 ? T2 ? T3 ? M1 ? M2 ? M3 ? R
head2righthead2right Commensurate + incommensurate
620 ?C           920 ?C            980 ?C          990 ?C       1060 ?C            1090 ?C
head2righthead2right Unmodulated
head2righthead2right Positional modulation of both Ca and Si
head2right Existing crystal structures are averages: 
head2righthead2right Relationship between the forms is unknown
Solve Structures and relate form to hydration behaviour: 
optimum structure?
boxtargetboxtarget
3. Study kinetics of formation of tricalcium 
silicate: Stabilize favourable forms 
? How are these forms are stabilized: in-situ temperature-dependant 
synchrotron X-ray Powder Diffraction studies of the clinkerization (formation) 
processes 
? Use this information to preferentially stabilize the favourable form 
Aranda et al Proc. 12th ICCC, 2007.
Better with neutrons!
Temperature (Deg. C)
boxtargetboxtarget
Australia's OPAL 20MW reactor source
New Neutron Scattering Facility 
boxtargetboxtarget
Neutron Zoo at OPAL
Platypus
(Reflectometry)
Wombat
(Hi-Intensity Powder)
Kowari
(Residual Stress)
Koala
(Single Crystal)
Quokka
(Small Angle)
Pelican
(Polarized Quasielastic)
Echidna
(Hi-Res. Powder)
Taipan
(Thermal Inelastic)
Sika
(Cold Inelastic)
boxtargetboxtarget
Time
(60hrs total)
discharge
charge
discharge
In-situ battery cycling on Wombat 
5 minute acquisition 
time ? as low as 50 ?s
Wombat Kinetics: Looking forward to 
exciting new research on the kinetics of 
reactions and processes
boxtargetboxtarget
Acknowledgements
NIST Center for Neutron Research, USA
? Dan Neumann, Craig Brown, Juscelino Le?o
The University of 
Texas at Austin, USA
? Maria Garci-Juenger
US Department of Transportation, USA
? Richard Livingston
I told you that cement was too wet?
boxtargetboxtarget
-0.025
0.025
0.075
0.125
0.175
0.225
0.275
0.325
0 10 20 30 40 50 60 70
Data
Model
Kinetic parameters of the hydration
BWI(t) = BWI(0) + A[1-exp{-[k(t-ti)]n}]
Modeling of the nucleation and growth period:
B W
I
Hours
Can be determined from QENS - more precisely determined by 
calorimetry 
? n for monoclinic pastes vary from 2.44 - 2.65. 
? M3 form has larger n (2.65) than T1 (2.27) at 30 ?C. 
head2right Model is relatively insensitive to changes in n
? varying n between 2.27 - 2.65 = 0.8 % change in A and 4 % in k. 
n = Dimensionality of growth: 
Product growth occurring in a 3-dimensional pore space