Concrete deterioration modelling by using …multi-scale/Yamada.pdfConcrete performance/ deterioration modeling by using thermodynamic equilibrium codes Kazuo YAMADA, National Institute

Concrete performance/ deterioration modeling by using thermodynamic

equilibrium codes

Kazuo YAMADA, National Institute for Environmental Studies

Yoshifumi HOSOKAWA, Taiheiyo Cement

2015/07/13 1 JCI-MultiSc_KYamada

Contents 1. Impact of phase composition.

Performances and durability of concrete is determined by hydrates and pore solution chemistry.

2. Basics of reaction transfer. Phase equilibrium, Effective diffusion coefficient, Modeling of

C-S-H.

3. Examples. Behaviors of alkali chlorides Sulfate expansion

4. Required subjects to be clarified. More database on hydrates. Formation factor/ effective diffusion coefficient.


Recent trends of materials for sustainability and durability

SCMs/ addition, blends Limestone Mineral powder Pozzolans Slags Etc. multi-blends

Recycling of wastes as raw meal

Strength and durability of concrete

Pore structure Volume, tortuosity,

constrictivity

Pore solution

chemistry

Hydrates composition

Environmental effects

+

Modification of the nature of cement clinkers and hydrates


New technology requirements

• If pore solution is in an equilibrium with solid phase, geochemical code such as PRHREEQC or GEMS will be useful.

Clinker with

wastes

Blended cement

XRD/Rietveld Image Analysis of BEI/EBSD

Cement hydrates QXRD, IA, Selective dissolution, NMR, TG/DSC, TEM, AFM…

Pore structure pore solution chemistry

Diffusion tests Steady or non-steady Electrical migration…

Squeezing Phase equilibrium

calculation

Solids

Pores

MIP, BET, IA, H-NMR…


Characterization & Modeling

What was the purpose of phase equilibrium calculation by geochemical code?

• Durability is the center of interests for various concrete structures. – An example of nuclear power stations . – OECD/NEA/CSNI CAPS ASCET (Assessment of Structures subject to ConcretE

PaThologies) made a questionnaire for 16 member countries. – According to Neb Orbovic 2014 (OECD/NEA/ASCET – RILEM/ISR Meeting),

followings are the major concerns. – 10/16 ASR (pH) – 5/16 Sulfate attack ([SO4

2-]), Rebar corrosion ([Cl-]/[OH-]) – 4/16 Irradiated concrete, Freeze-thaw, Carbonation (CO2) – 3/16 Chloride interaction ([Cl-])

• Majority is determined by the pore solution and solid chemistry of concrete and they are assumed in some equiliblium.

• Not only one mechanism, but combined degradation is happened. • Then, phase equilibrium calculation is required.


1. Impact of phase composition Ex. 1: Impact of minor limestone powder addition

2015/07/13 JCI-MultiSc_KYamada 6

Hoshino et al. ACT, 2006

5% of limestone addition increased strength - 10 % for OPC, - 30 % for slag blended cement by the formation of carbonate hydrates.


Hoshino et al. ACT, 2006


D. Helfort, Anna Maria Symp. 2009

OPC Constant W/C

A case of more reactive alumina.

It is possible to reduce 20% of CO2 by optimizing clinker composition and SCMs.

Ex. 2: C/S and pH and ASR expansion

• SCMs suppress ASR expansion. The mechanism has been thought as the decrease in pH by the formation of low C/S C-S-H.

• It will be possible to estimate the effects of SCMs by calculating C/S of C-S-H from the data of QXRD and selective dissolution.

0

0.2

0.4

0.6

0.8

1

1.2

0 20 40 60 80

expa

nsio

n ra

tio

SCM replacement level (vol%)

OPCFA(A)FA(B)FA(C)FA(D)BFSBFS

FA(A)

0% 20% 40% 60% 80% 100%

BFS-60%

BFS-50%

BFS-40%

FA(D)-20%

FA(C)-20%

FA(B)-20%

FA(A)-30%

FA(A)-20%

FA(A)-10%

OPC

volume fraction (%)

pore

C-S-H

AFt

Mc

Hc

CH

cement

FA

BFS

Kawabata & Yamada, JSCE 2013

QXRD & selective dissolution. Paste cured at 40 ºC for 28 days.

Mortar bar expansion at 40ºC for 52 weeks. Na2Oeq = 1.2% of cement.


Direct correlation between estimated [OH-] and ASR expansion

• Different SCMs affect C/S in different ways because of the different natures of them.

• However, it will be possible to estimate [OH-] based on C/S of C-S-H based on many experiments.

• There is a simple positive correlation between [OH-] and ASR expansion.

• Therefore, the calculation of phase equilibrium is fundamental procedure for ASR estimation also.

1

1.2

1.4

1.6

1.8

2

2.2

2.4

0 20 40 60

Ca/

Si m

olar

ratio

SCM replacement level (vol%)

OPC FA(A)FA(B) FA(C)FA(D) BFS

BFS

FA(A)

0

0.5

1

1.5

0.2 0.4 0.6 0.8 1 1.2

expa

nsio

n ra

tio

ca lculated [OH] (mol/l)

OPC FA(A)-10%FA(A)-20% FA(A)-30%FA(B)-20% FA(C)-20%FA(D)-20% BFS-40%BFS-50% BFS-60%

0.2

0.4

0.6

0.8

1

1.2

1 1.5 2 2.5

calc

ulat

ed [O

H] (

mol

/l)

Ca/Si molar ratio of C-S-H gel

OPCFA(A)-10%FA(A)-20%FA(A)-30%FA(B)-20%FA(C)-20%FA(D)-20%BFS-40%BFS-50%BFS-60%

Kawabata & Yamada, JSCE 2013 & unpublished. 2015/07/13 10 JCI-MultiSc_KYamada


Ex. 3: Application of material transport model for life estimation of chloride attack

Many chlorides from sea.

Rebar corrosion

How is it possible to design the service life? How long period survive this structure more?

Harsh environments of chloride attack


Traditional way for chloride attack

Cl-

Cl-

Cl-

Cl-

Cl- Cl-

Cl-

Cl-

Cl-

Cl-

Inside of concrete Rebar

Assuming Fick’s diffusion law for Cl penetration, Cl concentration at the position of rebar is calculated by using analytical solution of diffusion equation. -> Service life is judged by Cl threshold for corrosion.

2

2

xCD

tC

∂∂

=∂∂

=DtxCtxC s 2

erfc),(

Basic approach in many guidelines


Problems in traditional material transfer model

Sea water

Environments

Cl－

Cl－

Cl－

Cl－

Cl－

Cl－ Cl－

Cl－

Cl－ Cl－

Cl－

• Assuming homogeneous cement hardened body • Only Cl penetration is considered. • Steel corrosion is basically governed by [Cl-]/[OH-].

[Hardened cement]

Rebar

Degradation = chemical interaction between hydrated cement and ions. However, chemical action = phase equilibrium has not been considered.


Model considering chemical reaction (phase equilibrium) and multi-species transfer 2. Basics of REACTION TRANSFER

Friedel’s salt AFm Ca(OH)2 C-S-H Caobo-

aluminate

[cement paste] [pore solution]

Sea water Environ.

Material transfer

相平衡

Rebar

SO42－ H4SiO4

K+

Al(OH)4－ CO3

2－

OH－

Ca2+ Na+

Na+

Mg+

CO2 CO2 moisture H2O

Cl－

Cl－

Cl－ Cl－

[gas] Concentration gradient Electrostatic interaction

Phase equilibrium

Transfer

Phase equilibrium

FEM

PhreeqC

Poisson – Nernst – Planck equation

Thermodynamics Combined

This model has been developed by Dr Bjorn Johannesson (now DTU) and Dr Yoshihumi Hosokawa (Taiheiyo Cement Corp.) in middle 2000’.

Hosokawa et al. 2011


Calculation method of material transfer

iiiiii m

xzcB

xxcD

xtc

+

∂∂

∂∂

+

∂∂

∂∂

=∂

∂ φττ

∑=∂∂

iiiw czF

x2

2φε

Nernst – Plank equation

Poisson equation

• i = Ca2+, Cl-, SO42-, Na+, K+, CO3

2-,・・・ • All simultaneous equations are solved by FEM. • Unknown parameters：ci , φ (number of ion species + 1)

Ionic concentration Electrostatic potential

Ionic concentration Electrostatic potential


Calculation method of phase equilibrium Example: Dissolution equilibrium of Ca carbonate CaCO3 Ca2+ + CO3

2- H2CO3 HCO3

- + H+ CO32- + 2H+

]CaCO[]CO][Ca[

3

23

2

1

−+=K

Solutions of simultaneous equations → Concentrations of each ion，pH，CaCO3 amount

However, in reality, there are many unknown parameters.

• Ion conc.：Ca2+, Al3+, Na+, OH-, SO42-, ・・・

• Hydrates amounts：CSH, Ca(OH)2, AFt, AFm, ・・・

Non-linear multiple simultaneous equations

Numerical calc. software：PHREEQC

→ ←

• Mass action law • Mass balance • Electrical nuetrality

→ ← → ←

]COH[]HCO][H[

32

32

−+=K =3K


Application of phase equilibrium model to cement system

Na+ K+ Cl− SO42− Ca2+

C-S-H Fr Ms Ett

Al(OH)4−

SiOH

CH

Solution

Solid

Dissolution/ precipitation equilibrium

Speciation of ions

Enable to calculate pore solution composition and hydrates volume in cement paste

• C-S-H → Dissolution equilibrium and C/S variation is reproduced by Nonat’s model (This is the most unique point).

• MS – Fr is modellized by anion exchange reaction.

Ion exchange reaction


Ca2+

Al3+

SO42－

Ca(OH)2

ettringite AFm

Cl－

Neutralization Sulfate attack DEF

gypsum

Si4+

thaumasite

Acid resist.

Thaumasite Sulfate attack

Chloride attack

pH ASR Na+, K+

CSH Dissolution Heavy metals, nuclides, Organic substances

Environments evaluation

Redox reaction Various species, hydrates, reaction are quantified.

Every degradation can be considered.

Combining material transport and reaction equilibrium

Feature and effectiveness of this model


Analysis procedure by this combined model

Material cond. Boundary cond. • Chem comp of cement • Mix (W/C，C content) • Porosity, water content • Reaction ratio of minerals

• Ion conc in solution • Partial gas pressure • Analysis period

Input parameters

Initial cond. • Hydrates amounts • Solution amount • Pore amount

Reaction transfer Combined analysis

Ion concentration, hydrates amounts, gas pressure

Analytical results

• Any cement type is acceptable.

• Any environments can be considered.

Phase equilibrium calculation


Application of phase equilibrium calculation for materials transfer

• Two mechanisms determining transfer should be considered. – Fixation: Ion fixation may delay transfer. – Pore structure: Delayed transfer by complicated pore structure.

Cl Cl

Less permeable More permeable

Pore structure

Cl

Cl Cl

Cl

Fixation

AFm, C-S-H Paste

Pore

Evaluation: Effective diffusion coefficient De from self diffusion coefficient D0 by considering pore structure in dilute solution under high ionic strength calibration.

Small Large Evaluation: Distribution coefficient, binding isotherm

Can be generalized by phase equilibrium model. Problems?


How describe ion fixations of C-S-H with different C/S by phase equilibrium mode?

• What is the mechanism of ion fixation? • Which dissolution equilibrium model should be used?

1. Review of dissolution equilibrium of C-S-H and determination of model should be used.

2. Review of ion fixation by C-S-H and modeling it based on its mechanism.


CaO [mM]

SiO

2 [m

M]

Jennings (1986)

C-S-H dissolution equilibrium model Characteristics of C-S-H dissolution equilibrium

Ca2+

H4SiO4

OH－

C-S-H particle

Pore solution


C-S-H(I)

Invariant point

Invariant point

Jennings (1986)

C-S-H dissolution equilibrium model Characteristics of C-S-H dissolution equilibrium

Ca2+

H4SiO4

OH－

C-S-H particle

Pore solution

0,00

0,50

1,00

1,50

2,00

2,50

0,00 10,00 20,00 30,00 40,00

CaO mmol/l

C/S

Flint and Wells at30°CTaylor at 17-20°C

Lecoq 20°C

Thordvaldson 25°C(+expé sursat)

Modelling of phase equilibrium of C-S-H

Reproduction

Continuous change of Ca/Si

Invariant point

Invariant point

Reproduction of continuous change of Ca/Si

Important point of modelling


C-S-H dissolution equilibrium model Traditional models and their difficult points

• Mainly solid solution model of two end members. –Kulik: Congruent SS between. tobermorite-jennite –Berner: Ingongruent variable end member, Log K depending on C/S


C-S-H dissolution equilibrium model Traditional models and their difficult points

• Load for equilibrium calculation is heavy. Especially for variable Log K. • Poor compatibility with surface potential calculation.

Adopting Nonat’s model not using solid solution concept. [Features of Nonat’s model] •Modelling based on the chemical composition and structure change of C-S-H.

•Surface potential can be calculated.

2015/07/13 JCI-MultiSc_KYamada

C-S-H dissolution equilibrium model Nonat’s model

• Based on 1.1nm tobermolite structure

CaO2 plane →

CaO2 plane →

Si tetrahedra

26


Nonat’s C-S-H model

Si[Q1]

Si[Q2] Non bridging tetrahedron

Si[Q2p] Bridging tetrahedron

Low Ca/Si High Ca/Si

A model based on measurements

29Si solid NMR spectrum of synthesized C-S-H

2015/07/13 JCI-MultiSc_KYamada

28


• Based on 1.1nm tobermolite structure

CaO2 plane

Si tetrahedra

CaO2 plane Ca/Si = 0.66 Imaginary composition. C-S-H is not stable less than 0.8 of C/S.



• Bridging tetrahedron

Decrease in SiO2 tetrahedra → Increase in Ca/Si

SiO2 tetrahedra →

SiO2 tetrahedra →

CaO2 plane →

CaO2 plane →



• Existing ratio of bridging tetrahedron and variation of Ca/Si

Ca Ca Ca Ca Ca Ca Ca

Ca Ca Ca Ca Ca Ca Ca Ca


Ca/Si = 0.66

Ca/Si = 0.8

Ca/Si = 1.0 Si OH

Silanol

Negative charge

Negative charge

Negative charge

Si OH

→ + H+ Si O−

Si O−

+ Ca2+ → Si OCa+

Negative charge

Positive charge



• Coordinate of Ca2+ on silanol and ζ potential

Ca Ca Ca Ca Ca Ca Ca


Ca/Si = 1.0

Ca/Si = 1.5 Zero charge

Ca Ca Ca

Zero chargeﾞ

Ca＋ Ca＋

Ca Ca Ca＋ Ca＋

Increase in Ca/Si Increase in surface charge

-30

-20

-10

0

10

20

30

0.001 0.01 0.1 1 10 100

[Ca2+] (mmol/l)

potentiel zéta (mV)


-SiOH + Ca2+ -SiOCa+ + H+

-SiOH SiO- + H+

2 -SiOH + Ca2+ -SiOCaOSi- + 2 H+

2 -SiOH + H4SiO4 -SiOSi(OH)2OSi- + H2O

-SiOH + CaOH+ -SiOCaOH + H+

Ca2H2Si2O7 + 4 H+ + H2O 2 Ca++ + 2 H4SiO4


K1

K2

K3

K4

K5

K6

Mechanisms mentioned can be summarized in the following six equations.

Installation in phase equilibrium calculation tool ‘PHREEQC’ Ca Ca

Reaction of silanol

Reaction of dimer


Reproduction of phase equilibrium of C-S-H by Nonat’s model

0,000001

0,00001

0,0001

0,001

0,01

0,1

0 0,01 0,02

[CaO] mol/kg

[SiO

2] m

ol/k

g

0,6

0,8

1

1,2

1,4

1,6

1,8

0 0,01 0,02

[CaO] mol/kg

Ca/

Si

[solution composition] [Ca/Si variation]

0,00

1,00

2,00

3,00

4,00

5,00

0,6 1,1 1,6

Ca/Si

Q1/

Q2

[silicate anion chain length]

short

long -10-505

101520253035

0,001 0,01 0,1

[CaO] mol/kg

zeta

(mV

)

Equivalent electric point


Alkali ions fixation by C-S-H

SO4--

Ca++

K+

Na+

Cl-

Cl-

K+

Cl-

C-S-H H+

OH-

Electrical double layer O Ca+

O Ca+

O- O-

Change of composition in EDL

Surface

Zeta-potential measurement of C-S-H suspension

NaCl

Change in surface potential

Phase equilibrium model of C-S-H

Estimation

Verification

Alkali fixation was calibrated by sorption test


0123456789

10

0.5 1 1.5 2

Rd(N

a)

C/S(CSH)

Na100mmol/L

THC_NaCl

Glasser_NaOH

0.1

1

10

100

1000

0.5 1 1.5 2

Rd(N

a)

C/S(CSH)

Na1mmol/LTHC_NaCl

Glasser_NaOH

0 1 2 3 4 5 6 7 8 9

10

0.5 1 1.5 2

Rd(K

)

C/S(CSH)

K 100mmol/L

THC_KCl

Glasser_KOH

0.1

1.0

10.0

100.0

1000.0

0.5 1 1.5 2

Rd(K

)

C/S(CSH)

K 1mmol/LTHC_KCl

Glasser_KOH

Assuming reactions between silanol and alkali, solubility constants are obtained.

C-S-HのCa/Si

Log Kna Log Kk

0.8 −10.8 −9.9 1.0 −11.4 −10.9 1.2 −12.1 −11.9


SiOSi0.5OH + Na+ = SiOSi0.5ONa + H+ Kna SiOH + Na+ = SiONa + H+ Kna SiOSi0.5OH + K+ = SiOSi0.5OK + H+ Kk SiOH + K+ = SiOK + H+ Kk 0.1

1

10

100

0.1 10 1000

Naの

Rd(m

ol/k

g)

初期Na濃度(mM)

0.1

1

10

100

1000

0.1 10 1000

KのRd

(mol

/kg)

初期K濃度(mM)

C/S=1.2

Na (mM)

K (mM)

Rd-

K (m

M)

Rd-

Na

(mM

)

C/S=1.0

C/S=0.8

3.1 Examples - Behaviors of alkali chlorides.


0.0

0.1

0.2

0.3

0.4

0.5

012345678910

0 5 10 15 20

Cs, C

l, N

a, K

（m

ol/k

g）

Ca （

mol

/kg）

Depth （mm）

OPC mortar (CsCl 500mM)

CaCsClNaK

0.0

0.1

0.2

0.3

0.4

0.5

012345678910

0 5 10 15 20

Cs, C

l, N

a, K

（m

ol/k

g）

Ca （

mol

/kg）

Depth （mm）

FAC mortar (CsCl 500mM)

CaCsClNa

0.00

0.10

0.20

0.30

0.40

0.50

0 5 10 15 20

Conc

entr

atio

n (m

ol/k

g)

Depth (mm)

OPC, CsCl=500mM Tot-ClTot-NaTot-KTot-Cs

0.0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20

Conc

entr

atio

n (m

ol/k

g)

Depth (mm)

FAC, CsCl=500mM Tot-ClTot-NaTot-KTot-Cs

• Cs and Cl ingress to and Na & K leached out in the same depth. • Ingress amount of Cs seems the same with leached amounts of alkalis in mole. • Strong Cs fixation at the surface was caused by the dissolution of Ca. • Cl concentration is higher and this may be caused by the formation of Friedel’s salt.

Effects of concentration and C/S on alkali chloride ingress


05101520253035404550

0.000.010.020.030.040.050.060.070.08

0 5 10 15 20Cl

（m

mol

/kg）

Cs (m

mol

/kg)

Depth (mm)

FAC (3mM CsCl) Cs Cl

0

5

10

15

20

0 5 10 15 20

Conc

. (m

mol

/kg)

Depth (mm)

FAC (3mM CsCl)

Cs Cl

0

5

10

15

20

25

30

0.000.010.020.030.040.050.060.070.08

0 5 10 15 20

Cl （

mm

ol/k

g）

Cs (m

mol

/kg)

Depth (mm)

OPC(3mMCsCl)

Cs Cl

0

5

10

15

20

0 5 10 15 20

Conc

. (m

mol

/kg)

Depth (mm)

OPC(3mMCsCl)

Cs Cl

0.0

0.1

0.2

0.3

0.4

0.5

0.00

0.02

0.04

0.06

0.08

0 5 10 15 20

Cl （

mol

/kg）

Cs （

mol

/kg）

Depth （mm）

OPC (CsCl 500mM)

Cs

Cl

0.0

0.1

0.2

0.3

0.4

0.5

0.00

0.02

0.04

0.06

0.08

0 5 10 15 20

Cl （

mol

/kg）

Cs （

mol

/kg）

Depth （mm）

FAC (CsCl 500mM)

Cs

Cl

• [CsCl] affects fixed amounts but not penetration depth. Nothing of fixation to diffusion. • FA addition reduce penetration depth same for Cs & Cl. Nothing of surface charge on diffusion.

Reproduction by reaction transfer model


0.000

0.005

0.010

0.015

0.020

0.00

0.10

0.20

0.30

0.40

0.50

0 5 10 15 20

Conc

entr

atio

n (M

, Cs,

Cl)

Conc

entr

atio

n (M

, Na,

K)

Depth (mm)

OPC, CsCl=3mM

Tot-Na

Tot-K

0.000

0.005

0.010

0.015

0.020

0.00

0.10

0.20

0.30

0.40

0.50

0 5 10 15 20Co

ncen

trat

ion

(M, C

s, C

l)

Conc

entr

atio

n (M

, Na,

K)

Depth (mm)

FAC, CsCl=3mM

Tot-NaTot-KTot-ClTot-Cs

0.0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20

Conc

entr

atio

n (m

ol/k

g)

Depth (mm)

OPC, CsCl=500mM

Tot-ClTot-CsTot-NaTot-K

0.0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20Co

ncen

trat

ion

(mol

/kg)

Depth (mm)

FAC, CsCl=500mM

Tot-ClTot-CsTot-NaTot-K

0

5

10

15

20

0 5 10 15 20

Conc

entr

atio

n (m

mol

/kg)

Depth (mm)

FAC (3mM CsCl)

Cs Cl

0

5

10

15

20

0 5 10 15 20

Conc

entr

atio

n (m

mol

/kg)

Depth (mm)

OPC(3mMCsCl)

Cs Cl

JCI-MultiSc_KYamada 40

3.2 Examples - Sulfate expansion

• Expansion by ASTM C1012

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0 20 40 60 80 100 120

Immersion age (week)

Exp

ansi

on(%

)

OPC OPC+SL (SO3 2.6%) (4.2%)

(5.8%)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0 20 40 60 80 100 120

Immersion age (week)

Exp

ansi

on(%

)

OPC OPC+SL (LS 0%)

(4%)

(8%)

<Effect of Limestone filler> <Effect of SO3 content>

Blast furnace slag suppresses sulfate expansion but limited. Limestone suppresses it much more efficiently. Gypsum addition is also effective. Ogawa et al. 2012

2015/07/13


OPC after 38 weeks

OPC+SL after 38 weeks, (expansion:0.301%)

OPC+SL/L4 (SO3=4.2) after 6 years, expansion:0.149%

Expansion by sulfate attack

Ogawa et al. 2012 2015/07/13


Calculation results

• OPC

C4FH13

C-S-H

Ca(OH)2

C4AH13MsEttringite

0

50

100

150

200

250

300

350

400

0 7.5 15 22.5 30

Depth from the surface (mm)

Vol

ume

of h

ydra

tes

(cm

3/1L

)

Na+

Ca OH-

: :

SO42-

C4FH13

C-S-H

Ca(OH)2

C4AH13MsEttringite

Gyp

0 7.5 15 22.5 30


1095 days

SO4

Volume

Ogawa et al. 2012 2015/07/13


Differences of volume changes and sulfate ingress (1095 days)

C4FH13

C-S-H

Ca(OH)2C4AH13

Monocarbonate

Ettringite

Gyp

0

50

100

150

200

250

300

350

400

0 7.5 15 22.5 30


Volu

me o

f hyd

rate

s (c

m3/1L)

C4FH13

C-S-H

Ca(OH)2

C4AH13

MsEttringite

Gyp

0

50

100

150

200

250

300

350

400

0 7.5 15 22.5 30


Volu

me o

f hyd

rate

s (c

m3/1L)

C4FH13

C-S-H

Ca(OH)2

C4AH13

Ms

Ett

Gyp

0

50

100

150

200

250

300

350

400

0 7.5 15 22.5 30


Volu

me o

f hyd

rate

s (c

m3/1L)

C4FH13

C-S-H

Ca(OH)2

C4AH13

MsEttringite

Gyp

0

50

100

150

200

250

300

350

400

0 7.5 15 22.5 30

Depth from the surface (mm)V

olu

me o

f hyd

rate

s (c

m3/1L)

<OPC>

<OPC+SL(SO3 4.2%)>

<OPC+SL>

<OPC+SL+Lsp>

SO4SO4

VolumeVolumeVolumeVolumeVolumeVolumeVolumeVolume

SO4SO4

SO4SO4

VolumeVolumeVolumeVolume

SO4SO4

VolumeVolumeVolumeVolume

Slag blended Ingress → suppressed (fixing effect on diffusivity) Volume → no effect

Slag blended + SO3 Ingress → suppressed Volume → suppressed (initial Ett formation)

Slag blended + Lsp Ingress → suppressed Volume → suppressed (initial Mc formation)

Ogawa et al. 2012 2015/07/13


• Comparison between expansion by ASTM C1012 and estimated volume change of hydrates due to sulfate ingress by calculations

0

5

10

15

20

25

30

35

120% 130% 140% 150% 160%

Volume changes due to sulfate ingressby calculations (%)

Est

imat

ed

tim

e t

o fai

lure

whic

h w

asas

sum

ed

to o

ccur

at 0

.1% e

xpan

sion

by A

STM

C1012 (m

onth

s)

Ogawa et al. 2012 2015/07/13

4. Required subjects to be clarified

• One difficult point of geochemical code is the database of solubility constant for various cement minerals. – Mg-S-H and its atomic scale model. – C-A-S-H and its atomic scale model (partially done by Haas & Nonat, CCR

2014. Amount of Al in C-A-S-H is determined by [Al(OH)4-])

– Description of reactions for above phases and their dissolution coefficients.

• Formation factor/ effective diffusion coefficient. • Reaction speed.


Al in C-S-H

• According to Haas & Nonat, CCR 2014, Al content in C-A-S-H is determined by Al concentration in solution.

• Al concentration is determined by the equilibrium of the system including all phases and ions.

• Once, Al position in C-A-S-H structure and its chemical reaction with solubility constant are described, uptake amount of Al by C-A-S-H will be possible to calculate.

• Then, automatically, the amounts of other Al relating phases and porosity are calculated.

• Finally, strength development, durability/ degradation will be estimated and it will become possible to design sustainable materials and service life of concrete structure quantitatively.


Formation factor/ effective diffusion coefficient

• In many reaction transfer model, pore structure is given as a formation factor measured by some method such as diffusion or migration test, or immersion test. – These tests are basically aimed to evaluate formation factor including the

effects of tortuosity and surface charge (if exists) and others. – No problem in steady state diffusion test. – Problems in electrical migration test.

• One is primary misunderstanding for the consideration of porosity (direct results is not effective diffusion coefficient but required to divide by porosity).

• For the measurement of electric resistance R (inverse of current), the ratio of R(Liquid) to R(Solid), RL/RS is the reducing factor of flux by pore structures.

• In order to obtain effective diffusion coefficient, pore volume must be considered. • De = (RL/RS)D0 (Wrong misunderstanding!) • Second one is the effect of electric migration current under DC and AC method

should be adopted. – Immersion test affected by ion fixation can be used with reverse analysis

assuming diffusion equation considering some non-linear binding. 2015/07/13 JCI-MultiSc_KYamada 47

0( / )e L SR R=J J 0( / ) /e L SD R R D φ=

Problems of electrical migration method

1. Generation of electric immersion current by applying direct voltage. When walls have charges, solution is reversely charged. In this condition, applying direct voltage results in movement of solution = Electroosmotic flow (more current in finer pores).

Solution

Measurement under alternative current

2. Difficult to eliminate the effect of electroosmosis.

+ + + + + + + + + + + - - - - - - - - - - - + + + + + + + + + + + - - - - - - - - - - -

Even in ion exchanging membrane having too small pores for solute diffusion, applying voltage allows movement of ions by ion exchanges. It is difficult to ignore electroosmosis effect when small pores have charges. The effect is more evident under smaller pores and lower electrolyte concentration.

Measurement under infinite concentration of electrolyte solution

+ + + + + + + + + + + - - - - - - - - - - -

+ + + + + + + + + + + - - - - - - - - - - - - + - + - + - + - +- - - - - -

+ - + - + - - - - -

It is required to have electrical neutralization in solution. If not, what happen?

Solution Ichikawa et al. 2013


Sample = sintered plate without surface charge Electrolyte solution = KI (no binding with cement paste)

Correlation with resistance of solution having the same shape: De = 0.0827 D / 0.323 (porosity)

No electroosmosis

Resistance change of a sample containing 0.1M KI kept in 1M KI solution. Dot line is a theoretical calculation based on reference data D ≈1.95x10-9m2/s.

0 1 2 3 40

50

100

150

Time/104s

Res

ista

nce /

Ω

: Observed

: Theoretical (De = 4.99x10−10m2/s)

0 5 10 150

0.05

0.1

0.15

Electric resistance of solution (inverse of KI concentration)

Rat

io o

r R (L

iqui

d/ S

olid

) (in

vers

e of

cur

rent

=

Diff

usio

n in

Sol

id/ L

iqui

d)

Ichikawa et al. 2013 2015/07/13 49 JCI-MultiSc_KYamada

Mortar plate (W/C = 0.55, S/C=3）

Higher resistance ratio of solid (less current) in less resistance of KI solution (higher concentration) Correlation with resistance of solution having the same shape:

De = 0.00224 D / 0.189 (porosity) Significant electroosmotic flow!!

Resistance change of a sample containing saturated Ca(OH)2 solution kept in 1M KI solution. Dot line is a theoretical calculation based on reference data D ≈1.95x10-9m2/s (electoosmotic flow was ignored)

0 10 20 30 400

200

400

600

Time/104s

Res

ista

nce

/ Ω

: Observed

: Theoretical (De = 0.231x10−10m2/s)

0 5 100

0.01

0.02


Ichikawa et al. 2013

Interaction delays diffusion? • Cl penetration seems free from Cl concentration. • Fixation ability only changes total Cl concentration. • When the rate of fixation reaction such as ion exchange of

AFm to Friedel’s salt is not fast enough although that is the basic assumption of reaction transfer, fixation may not affect diffusion so much.


0

0.2

0.4

0.6

0.8

1

0 10 20 30

Cl conc(p

aste

%)

Distance from surface(mm)

12month

24month

36month

0

0.2

0.4

0.6

0.8

1

0 10 20 30

Cl conc(p

aste

%)


0

0.2

0.4

0.6

0.8

1

0 10 20 30C

l conc(p

aste

%)


Concrete Block, W/C=0.50, Sea shore exposure at 22ºC (annual average) OPC FA15% FA25%

Conclusions

• Impacts of phase composition on the performance and durability and importance of phase equilibrium calculation were introduced.

• Basics of reaction transfer model composed of phase equilibrium calculation based on PHREEQC and multi-species transfer based on Nernst-Plank equation were summarized.

• Unique description of C-S-H for continuous change in C/S based on Nonat’s crystal model was explained.

• Examples of application of reaction transfer calculation for alkali chloride behaviors and sulfate expansion were introduced.

• Required subject to be clarified such as further modeling of M-S-H, C-A-S-H, formation factor/ effective diffusion coefficient determination, fixation rate.


Documents

Concrete deterioration modelling by using …multi-scale/Yamada.pdfConcrete performance/ deterioration modeling by using thermodynamic equilibrium codes Kazuo YAMADA, National Institute