CRACKS AND PORES -THEIR ROLES IN THE TRANSMISSION OF WATER

L.P. Aldridge1 H.N. Bordallo2

K. Fernando1 & W.K. Bertram1

1ANSTO, Private Mail Bag 1, Menai 2234 NSW, Australia2Helmholtz-Zentrum Berlin für Materialien und Energie–

Berlin - D-14109, Germany

Water Transmission in cement paste

• The aim of this work is to differentiate the rate of water transmission through– Cracks– Pores

• Capillary Pores (Diameter > 100 Å)• Gel pores (Diameter < 100 Å)

• Using the techniques of – Permeability – Quasi Elastic Neutron Scattering

Why this study

• Cementitious materials are used as barriers to radioactive wastes– Rate of transmission of radionuclides depends on

rate of water transmission• Durability of concrete related inversely to its

ability to transmit fluids. – Hence an ability to predict future water transmission

gives information on likely service life of concrete structures

• The service life of a low level repository is expected to be greater than 300 years– Tools to demonstrate that this is a likely outcome are

desirable.

Definitions

• Concrete - cementitious materials & aggregate & sand & water

• Mortar - cementitious materials & sand & water• Paste - cementitious materials & water

• OPC Ordinary - Portland Cement manufactured by Blue Circle Southern.

• GGBFS - Ground Granulated Blast Furnace Slag• Marine Cement –OPC with 60% replaced with

inter-ground GGBFS

Experimental ILow & medium water pastes & mortars had similar flow.

• Paste– Low Water

• W/C 0.32– Medium Water

• W/C 0.42– High Water

• W/C 0.6 & 0.8• Mortar

– Low Water• W/C 0.32• Binder/Sand 1.2

– Medium Water• W/C 0.46• Binder/Sand 1.0

Experimental IIShear Mixed and cured 28 days sealed

• OPC & Marine – Mortars Pastes

• Mixed in Wearing Blender• Cured 28 days Sealed• Low Medium similar flow

0.42 0.43 0.44 0.45 0.46

20

40

60

80

100

Flow beforeshear mixing

Flow

(Sec

onds

)

Water to Cement ratio

Ludirinia's apparatus for permeability measurement

• A’(m2) - the cross section of the pipette,

• A(m2) area of specimen• h(m) the water head

– ho is the initial level – hl the final level

• L(m) the thickness• t(s) the time

• Note can only measure when K’ > 1*10-12m/s

KK’’= (A= (A’’ L)/(A t) ln(hL)/(A t) ln(h00/h/hll))

Effect of Crack Width on Water Transmission

ll Length CrackLength Crackw w WidthWidthof crackof crack

dd Depth Cylinder Depth Cylinder

From the From the NavierNavier--Stokes equation Stokes equation

It can be shown that It can be shown that

ww33= 3= 3πμπμdd22 KK’’ / (/ (ρρl)l)

Where w is the width of the crackWhere w is the width of the crackL is the crack lengthL is the crack lengthd is the depth of the cylinderd is the depth of the cylinderμμ is the viscosity of water at 20 degreesis the viscosity of water at 20 degreesρρ is the density of wateris the density of waterKK’’ is the permeability of the sampleis the permeability of the sample

Relationship between crack width and measured Permeability

0 100 200 300

1E-12

1E-11

1E-10

1E-9

1E-8

1E-7

1E-6

1E-5

Per

mea

bilit

y (m

/s)

Crack Width (μm)

Cracks <50Cracks <50μμm m =>=>Permeability >10Permeability >10--88m/sm/s

For pastes (uncracked) it will be the capillary pores that carry the majority of water

• The capillary pores are the space remaining after hydration takes place

• Thus they are highest at the start of hydration• Paste made up

– Un-hydrated cement– Hydrated cement – gel– Gel – pores– Capillary pores– Pores due to chemical shrinkage (capillary pores)

• Volume at a time depends on the extent of the paste hydration (α) which varies between 0 and 1

Powers Brownyard Model – Volume of components depends on α

• Define p the initial porosity of the paste – depends on density of cement water and w/c

• Vol( chemical shrinkage) = 0.20(1-p) α• Vol( capillary pores) = p-1.32(1-p) α• Vol( gel pores) = 0.62 (1-p) α• Vol( gel) = 1.52 (1-p) α• Vol( un-hydrated cement) = (1-p) (1-α)

• Relationship depends on assumptions – e.g that chemically bound water (non-evaporable

water) 0.23 g binds per gram of cement hydrated– Gel water 0.19g binds per gram of cement hydrated

0.0 0.2 0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

0.8

1.0W/C = 0.80

unhydrated cement

gel solids

gel water

capillary pores

chemical shrinkage poresV

olum

e Fr

actio

n

α

So these approximations show that capillary pore volume decreases with hydration

Even low w/c pastes have large capillary pore volumes - when uncured.

0.0 0.2 0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

0.8

1.0W/C = 0.32 chemical shrinkage pores

capillary pores gel water

gel solids

unhydrated cementVol

ume

Frac

tion

α

From work by Powers and his co-workers we also find that capillary pore space is related to the permeability

• Powers plotted permeability for pastes at different w/cratios

• Pastes were almost fully saturated

• Pastes with continuous capillary pores had greater permeability than indicated by line.

• After at low pore volume the capillary pores became discontinuous and results followed the line

• The point of imitation of the discontinuous pores is indicated.

0.0 0.1 0.2 0.3 0.4 0.5

1E-16

1E-14

1E-12

Discontinous Capillary Pores

Continous Capillary Pores

Per

mea

bilit

y (m

/sec

)fractional volume of capillary pores

Further work by Powers and co-workers indicated the relationship between curing time and w/c ratio

• Pastes cured longer were less permeable.

• Pastes made with greater W/C had dramatic differences in permeability

0 10 20 300.01

1E-3

1E-4

1E-5

1E-6

1E-7

1E-8

1E-9

1E-10

1E-11

1E-12

1E-13

Per

mea

bilit

y m

/s

Days Curing (Paste W/C =0.7)

0.2 0.3 0.4 0.5 0.6 0.7

1E-14

1E-13

1E-12

Perm

eabi

lity

m/s

W/C

Powers work indicated that with proper curing at w/c 0.42 the pores should be discontinuous.

• At 7 days the paste with w/cof 0.45 should have an approximate degree of hydration 0.60 and have acquired a discontinuous pore structure.

• However this does assume – Proper Mixing– Proper Compaction – Proper Curing

• Furthermore these are theoretical estimates

• assuming ALL cements hydrate in the same manner.

0.0 0.2 0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

0.8

1.0

unhydrated cement

gel solids

gel watercapillary pores

chemical shrinkage poresW/C = 0.42

Vol

ume

Frac

tion

α

0.0 0.2 0.4 0.6 0.8 1.00.0

0.2

0.4

0.6

0.8

1.0

unhydrated cement

gel solids

gel watercapillary pores

chemical shrinkage poresW/C = 0.42

Vol

ume

Frac

tion

α (degree of hydration)(degree of hydration)

~28day~28day 1year1year

Pore Water Ratios in 1g of Fully Hydrated Pastes (1 year old α~0.95) for w/c=0.42

Mobile water is in both gel and Mobile water is in both gel and capillary porescapillary pores

CC33SS

CHCH

Cap PoreCap Pore> 100> 100ÅÅwaterwater ∼ 2%∼ 2%

Gel pores < 100Gel pores < 100ÅÅAmount of water ~ 43%Amount of water ~ 43%

~52% of initial water ~52% of initial water Bound CH & CBound CH & C--SS--HH

Pore Water Ratios in 1g of Fully Hydrated Pastes (28 day cure α~0.75) for w/c=0.42

Mobile water is in the smaller poresMobile water is in the smaller pores

CC33SS

CHCH

Cap PoreCap Pore> 100> 100ÅÅwaterwater ∼ 25%∼ 25%

Gel pores < 100Gel pores < 100ÅÅAmount of water ~ 34%Amount of water ~ 34%

~41% of initial water ~41% of initial water Bound CH & CBound CH & C--SS--HH

Water Diffusivity can be measured by QENS

• Bulk water 25*10-10 m2 /s • OPC Paste 12*10-10 m2 /s at ΔE = 98 μeV• OPC Paste 6*10-10 m2 /s at ΔE = 30 μeV

QENS Results fromQENS Results from1.1. Bordallo, H.N., Aldridge, L.P., and Bordallo, H.N., Aldridge, L.P., and DesmedtDesmedt, A. (2006) , A. (2006)

Water Dynamics in Hardened Ordinary Portland Cement Water Dynamics in Hardened Ordinary Portland Cement Paste or Concrete: From Paste or Concrete: From QuasielasticQuasielastic Neutron Scattering. Neutron Scattering. J. Phys. Chem. C, 110(36), 17966J. Phys. Chem. C, 110(36), 17966--6.6.

2.2. Bordallo, H.N., Aldridge, L.P., Churchman, G.J., Gates, W.P.,Bordallo, H.N., Aldridge, L.P., Churchman, G.J., Gates, W.P.,Telling, M.T.F., Kiefer, K., Telling, M.T.F., Kiefer, K., FouquetFouquet, P., , P., SeydelSeydel, T., and , T., and KimberKimber, S.A.J. (2008) Quasi, S.A.J. (2008) Quasi--Elastic Neutron Scattering Studies on Clay Elastic Neutron Scattering Studies on Clay InterlayerInterlayer--Space Highlighting the Effect of the Space Highlighting the Effect of the CationCation in Confined in Confined Water Dynamics. J. Phys. Chem. C, 112(36), 13982 Water Dynamics. J. Phys. Chem. C, 112(36), 13982 -- 13991.13991.

3.3. Aldridge, L.P., Bordallo, H.N., and Aldridge, L.P., Bordallo, H.N., and DesmedtDesmedt, A. (2004) , A. (2004) Water dynamics in cement pastes. Water dynamics in cement pastes. PhysiciaPhysicia B, 350, e565B, 350, e565--e568.e568.

QENS – Quasi-Elastic Neutron Scattering

• The signal from H dominates the spectra: We can “see water”, and the signal from the rest is very small

• By using the elastic fixed window approach that is similar to Debye-Waller evolution obtained from X-rays, we can determine the temperature where the water starts to move

Elastic Windows on SPHERES (ns time scale) define when the water motion is unlocked!!!

0 50 100 150 200 250 3000.0

0.2

0.4

0.6

0.8

1.0 ~220K

~210K

240K Paste w/c=0.42 Paste,dried after re-hydration at RH=50% Dried sample at 105C

S Nor

mal

ised

(Q,0

)

T(K)

QENS – Quasi-Elastic Neutron Scattering

• The signal from H dominates the spectra: We can “see water”

• QENS allows measurements in different time and length scales – ΔE (Δt)

• We can differentiate between bound water and “free” water – because they move differently

NEAT - ToF Spectrometer at BENSC

Q: 0.4 Å-1 < Q < 2.1 Å-1

resolution at elastic peak: ΔE ~ 98 or 30μeV

time scale: picosecondmulti-detector

radial collimator

double-trumpet:converging/divergingneutron guide sections

straightneutron guide

multidetector chamber

sample

monitor

chopper housingwith two

choppersdiaphragmsystem 2

diaphragmsystem 1

388 single detectors

Dynamics Model

),(),(e),( 3/22

ωωω QRQTQS Qu⊗=

−

Jump TranslationalIsotropic RotationOscillation

High Frequency: Stretching

<u2>½ Debye-Waller

τrRotational Correlation

Timerg

Radius of Gyration

DtTranslational Diffusion

τtResidence Time

LMean Jump Distance

-1.5 -1 -0.5 0 0.5 1 1.5

Q=0.665Å-1

0.001

0.01

0.1

1

10

S(Q

,ω) (

arbi

trary

uni

ts)

Energy Transfer (meV)

-1.5 -1 -0.5 0 0.5 1 1.5

Q=2.11 Å-1

0.001

0.01

0.1

1

10

S(Q

,ω) (

arbi

trary

uni

ts)

Energy Transfer (meV)

QENS results from NEAT

0 1 2 3 4 5

0.01

0.02

0.03

0.04

0.05

0.06

Γ(Q)=(Dt Q2)/(1+Dt Q

2τ0)& Dt= L2/(tτ0)

Γ(m

eV)

Q2 (Å-2)

QENS results from NEAT

1.5 2.0 2.5 3.0 3.5 4.0 4.51

10

100

Dt *

10-1

0 m2 /s

τ 0 ps

L (Å)

5

10

15

20

25

30OPC0.4GGBFS0.6OPC

RT Cure

OPC40o Cure

H2O

Chemically “bound” water• Heating pastes at 105°C

– removes both glassy and unbound water. Only chemically bound water remains.

– No QE broadening after heating

– Dynamics of chemically “bound” water molecules occurs on a timescale significantly slower than the pico-second.

-0.4 -0.2 0 0.2 0.4

S(Q

,ω) (

Arb

itrar

y un

its)

Energy Transfer (meV)

-0.4 -0.2 0 0.2 0.4

Q = 2.1Å-1

ΔE = 98μeV

S(Q

,ω) (

Arb

itrar

y un

its)

Energy Transfer (meV)

-0.2 -0.1 0 0.1 0.2

S(Q

,ω) (

Arb

itrar

y un

its)

Energy Transfer (meV)

Q = 0.88Å-1

ΔE = 30μeV

-0.2 -0.1 0 0.1 0.2

S(Q

,ω) (

Arb

itrar

y un

its)

Energy Transfer (meV)

QENS of water in paste before and after heating at 1050C

Before Heating After Heating After re-hydration

The red line (Representing translational diffusion) is about 5 times narrower on the left hand spectrum

-0.4 -0.2 0 0.2 0.4

S(Q

,ω) (

Arb

itrar

y un

its)

Energy Transfer (meV)

-0.4 -0.2 0 0.2 0.4

Q = 2.1Å-1

ΔE = 98μeV

S(Q

,ω) (

Arb

itrar

y un

its)

Energy Transfer (meV)

QENS of water in paste before and after heating at 1050C

• After re-hydration

– Similar • To bulk

– Water

0

0,05

0,1

0,15

0,2

0,25

0,3

0 1 2 3 4 5

OPC-40°C; w/c=0.6

OPC-40°C;w/c=0.6 re-hydrated

Γ(m

eV)

Q2 (Å-2)

Conclusions

• For well made cementitious based barriers– Cracking may dominate water transport– Water transport through capillary pores in

cement paste can be estimated

• At low w/c ratios then water transport through gel pores should control water transport

Conclusions – Gel Pores

• Definition of water motion in gel pores is vital to understand (and measure) the durability in concrete.

• We need to understand– Time scale of diffusion through the gel pores – Time scale of diffusion into the gel pores

• We know more than we did when this work was started

• We know less than we would like• We must characterise water motions occurring at

different time scales