Professional Documents
Culture Documents
Aci sp-239-2006
Aci sp-239-2006
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Chemical Admixtures
in Concrete
Proceedings
Eighth International Conference
Sorrento, Italy, 2006
V. M. Malhotra
SP-239
Editor
V.M. Malhotra
American Concrete Institute®
Advancing concrete knowledge
SP-239
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The Institute is not responsible for the statements or opinions expressed in its
publications. Institute publications are not able to, nor intended to, supplant individual
training, responsibility, or judgment of the user, or the supplier, of the information
presented.
The papers in this volume have been reviewed under Institute publication procedures
bv individuals expert in the subject areas of the papers.
Copyright© 2006
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
All rights reserved, including rights of reproduction and use in any form or by any means,
including the making of copies by any photo process, or by any electronic or mechanical
device, printed or written or oral, or recording for sound or visual reproduction or for use
in any knowledge or retrieval system or device, unless permission in writing is obtained
from the copyright proprietors.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
During the conference, a special dinner was held in honor ofMichele
Valente to recognize his outstanding and sustained contributions in the area
of chemical admixtures, and his continued support of the CANMET/ACI
symposia and conferences.
October 2006
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
j
Copyright American Concrete Institute
vi
Provided by IHS under license with ACI Licensee=UNI OF NEW SOUTH WALES/9996758001
No reproduction or networking permitted without license from IHS Not for Resale, 08/10/2015 01:25:47 MDT j
Daneshlink.com
TABLE OF CONTENTS
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
SP-239-5: Artificial Superplasticized Aggregates for Decorative Applications .......... 67
by G. Ferrari, T. Cerulli, A. Lolli, G. Novella, and R. Pellay
SP-239-10: Novel Slurry Rheology Modifier for Materials of Civil Engineering and
Construction ..........................................................................................................137
by H. Yamamuro, K. Koyanagi, and T. Kojima
SP-239-26: Changes in Cement Paste and Mortar Fluidity after Mixing Induced
by PCP: A Parametric Study ................................................................................... 389
by L. Regnaud, A. Nonat, S. Pourchet, B. Pellerin, P. Maitrasse, J.P. Perez,
and S. Georges
SP-239-31: Effect of Advanced Hybrid Admixture on High Strength Concrete ......... 471
by K. Saito, M. Kinoshita, K. Okada, and T. Nawa
by j. Zhor
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
During the last decade a new family of polymers based on polycarboxylate as the
main polymer chain and polyether as side graft chains have been developed. This
new family of products appear to be in general more effective in terms of higher water
reduction, lower slump loss and lower retarding effect at very early ages.
More recently these performances have been enhanced in view of some specific
applications: a) in precast concrete structures, the very early strength (such as at 12-
16 hours) can be increased even in cold climates and in the absence of steam curing;
this effect depends on the number and length of the polyether graft chain which are
responsible for the dispersion effect related to the steric hindrance; b) in ready-mixed
concrete mixtures, the slump maintenance behavior can be designed as a function of
the transport time and placing temperatures, so that fresh superplasticized concrete
mixtures can be transported from the batching plant to the job site without any
slump loss even in summer times; this effect depends on the gradua !liberation of
special molecular groups which are responsible for the adsoption of these polymers
on the surface cement grains; c) a new series of polycarboxylate polymers has been
synthetized in which functional chemical groups, acting as shrinkage-reducing
admixtures, again are gradually liberated by the superplasticizer polymer as a
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
During the last three decades [ 1-2] the main ingredients of the superplasticizers were
synthetic water-soluble polymers such as sulfonated melamine formaldehyde (SMF)
condensate and sulfonated naphthalene formaldehyde (SNF) condensate. These
admixtures are polymers based on a single repeating unit (sulfonated naphthalene or
melamine molecule) which do not allow for much variation of the molecular structure.
Advances in superplasticizers, containing alternative water soluble synthetic products,
have been proposed in the last decade to reduce the slump-loss drawback [3-17]. More
recently, these new superplasticizers - all based on the family of polycarboxylate
polymers (PC) - have been deeply investigated, and numerous papers on these admixtures
have been presented at the CANMET-ACI International Conferences on
"Superplasticizers and Other Chemical Admixtures" [18-24]. Polycarboxylate
superplasticizers are based on two or more structural groups that provide for more
diversity in possible molecular structures with respect to those based on SNF or MSF
[25]: for example, by changing the number of carboxyilic groups per polyether unit, the
adsorption of polymers on the cement surface can be significantly changed; moreover,
the molecular weight of polyether molecule can be adequately adjusted to form polymers
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
particles. In other words, the pendant graft chains of the polymer molecules adsorbed on
the surface of cement would hinder by themselves from flocculating into large and
irregular agglomerates of cement particles (Fig. 3).
SLUMP RETENTION
Slump loss appears to be the most serious problem at least for ready mixed concrete
industries. When a concrete mixture must be transported for a long time, particularly in
hot weather, it should keep as far as possible at its initial slump level to avoid the practice
of redosing the concrete with water above and beyond that required in the mixture
proportion. Results of investigations of retempered concrete indicate that many of the
properties of the hardened concrete (strength, durability, abrasion resistance, etc.) are
significantly affected, since retempered concrete does not perform as well as concrete
which has not been retempered [28]. However, slump loss is un-avoidable because ofthe
intrinsic requirement for concrete mixtures which should set and harden in a relatively
short time. Therefore, a right and proper compromise would be a zero-slump-loss
concrete mixture for about 1 hour. By using superplasticizers based on SNF or SMF
polymers it is not easy to achieve this target, because in general slump loss is higher in
superplasticized concrete with respect to the corresponding plain mix at given initial
and alkali, as well as the form of calcium sulfate used as set regulator, can affect the rate
of slump-loss.
Also redosing the superplasticizer at different intervals of time has been suggested
[30] to reduce slump-loss, but this method appears to be not always easy to be adopted in
practice. Moreover, the total dosage of superplasticizer, as well as the relative cost,
cannot be kept under control according to a given plan.
Therefore a superplasticizer is required which by itself is capable to maintain the
slump for a long period of time independently of the temperature or the type and content
of cement. Collepardi et al. [13] studied the effects of a PC-based superplasticizer on the
properties of ready-mixed concrete mixtures. This admixture acts as both an immediate
superplasticizer and a slump loss reducing agent, and appears to be more effective than
that based on SNF. Although the PC-based superplasticizer has been used at a lower
content of active polymer (0.30% versus 0.40%), it was more effective than that based on
SNF for the water reducing capability (w/c = 0.43 versus 0.47) as well as for the
maintainance of the initial slump level (Fig. 6).
Tanaka et al. [17] studied the effect of a CLAP-based superplasticizer (Cross Linked
Acrylic Polymer) on the slump-loss of concrete mixtures. This superplasticizer is a
partially cross-linked copolymer of acrylic acid and polyethylene glycol mono-alkyl ether
(Fig. 7). According to Tanaka and co-workers the cross-linked polymer is hydrolyzed by
the alkaline water phase of the cement paste and then converted into a PC-based polymer
(Fig. 8). The negative carboxylic groups due to the alkaline hydrolysis would be adsorbed
on cement surface of cement particles and then would be responsible for the dispersion of
cement particles and the fluidizing action of the admixture. The low slump-loss effect of
this superplasticizer should be related with increasing number of the protruding side
CONCLUSIONS
REFERENCES
[17] Tanaka M., Matsuo S., Ohta A. and Veda M., A New Admixture for High
Performance Concrete, Proceedings of the "Concrete in The Service of Mankind",
Editors: R.K. Dhir and M.J. McCarthy, pp. 291-300 (1996).
[18] Jeknavorian A.A., Roberts L.R., Jardine L., Koyata H. and Darwin D.C.,
Condensed Polyacrylic Acid-Aminated Polyether Polymers as Superplasticizers
for Concrete, Proceedings of the Fifth International Conference on
"Superplasticizers and Other Chemical Admixtures" Editor V.M. Malhotra, ACI
SP-173, Rome, Italy, pp. 55-82 (1997).
[19] Sakai E. and Daimon M., Dispersion Mechanisms of Alite Stabilized by
Superplasticizers Containing Polyethylene Oxide Graft Chains, Proceedings ofthe
Fifth International Conference on "Superplasticizers and Other Chemical
Admixtures" Editor V.M. Malhotra, ACI SP-173, Rome, Italy, pp. 187-202
(1997).
[20] Ohta A., Sugiyama T. and Tanaka Y., Fluidizing Mechanism and Application oj
Polycarboxlate-Based Superplasticizers, Proceedings of the Fifth International
Conference on "Superplasticizers and Other Chemical Admixtures" Editor V.M.
Malhotra, ACI SP-173, Rome, Italy, pp. 359-378 (1997).
[21] Shonaka M., Kitagawa K., Satoh H., Izumi T. and Mizunuma T., Chemical
Structures and Performance of New High-Range Water-Reducing and Air-
Entraining Agents, Proceedings of the Fifth International Conference on
"Superplasticizers and Other Chemical Admixtures" Editor V.M. Malhotra, ACT
SP-173, Rome, Italy, pp. 599-614 (1997).
[22] Hamada D., Sato H., Yamamuro H., Izumi T., and Mizunuma T., Development oj
Slump-Loss Controlling Agent with Minimal Setting Retardation, Proceedings of
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Superplasticizer
-+
Superplasticlzer
-+
Fig. 1- Relationship between cement grains dispersion and fluidifying
effect on the cement.
SNF
Skle sulphonic group (negatiVely Chai\Jed)
L,
Main potymerchalnrj. . . . . . . . . . . . . .. . .
/
/
Sulphoni<o groupe
responslbt& for Cement grain
actsorptJon
Fig. 2 - Schematic picture of sulfonated polymer (SNF) and its electrostatic repulsion
effect on the dispersion of cement particles.
pol~ll*
0-·"-"''"""'• .....
c;hlln
\
Std• lorJV graft ch..n
l'"'utral)rn.pontlbleiOf'
•tarkhlfl.dr•nc•
Fig. 3- Schematic picture of the polycarboxylate (PC) superplasticizer and its steric
hindrance effect on the dispersion of cement particles.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
MIXING . . . . JUST I
WATER~C?o· .
. U·\:)·. /
. / AFTER......._Q00
Q....
d~~/ 1 HOUR'-.._
AFTER @ .•
~-"·
HYDRATED CEMENT
E 200
E
2~ r-------------~
--------•,--•-~/C=0.43
, _ .
PC (0.30%) - -.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
a: 150
:iii
:J
...I 100
(I)
50
0 30 60 90 120 150
TIME, min
IBefore Mixing J
low
[AhrMixlng ~7:;:;:;:;:;:;:;:;~b:::::
Fig. 8- CLAP-based superplasticizer hydrolyzed by the alkaline water of cement
paste: the number of coo-
progressively increases with time and this is responsible
for the low slump loss.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
...,.Looo_...,.AgontCOLCAI
................u,
" • : • • !
+ -4
-~ ___l/... _____1_¥- b-&--1L
~ ~-~~l
... , . i..••.••••••••.••.•
Fig. 9- The chemical structure of polycarboxylate type (PC), polyether type (PE),
and slump-loss-controlling agent superplasticizer (SLCA); q and pare the number of
moles of EO in PC and PE or SLCA polymers [22].
Polycarboxytate based (PC) Poly.- based (PE) Stump Loss Controlling Agent
(SLCA)
::s;iW Jl±J
Fig. 10- The schematic molecular structure of polycarboxylate type (PC), polyether
Carboryl unit
(COO") (COOR)
L;J : :~ I
~~ @)
Fig. 11- The mechanism and the hydrolysis reaction ofSLCA in alkaline condition of
cement paste (adapted from Reference 22).
e
§.
c..
E
:I
Ci5
OL-----------------------------~
0 30 60 90 120
Time (mm)
Fig. 12- Schematic trend in the slump loss behavior with different acrylic
superplasticizers (PC, PE, SLCA, PE+SLCA) in concrete mixtures with CEM II A/L 32.5R
(340 kg/m 3), w/c = 0-45.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~~----------------------~
PE
6
t.-25
1-
~L-----------------------~
0 5 10 15 20 25
Time (hours)
40r---------------------------~
oL---------------------------~
0 12 24
Time (hours)
HiCH.-r-~CH•
r= ~Hs
_'f=O H
.
OM 9 PC
...
"'0)
.,
Si'·z
H CHr¢~n¢~~¢~q H
1 9
(EO)n
H a
OM ?
(EPBE)
b
?;S.3
14
i 5
"' 6
~
g7
8
9 L-------~--------~--------~
0 10 20 30
Age (week)
Manufactured Aggregates
Synopsis: Aggregate shape, texture, and grading have been known to have a
significant effect on the rheological performance of fresh concrete. Moreover, while
the optimization of aggregate selection can provide both technical and economical
benefits, the availability of materials and construction operations can often dictate
the use and proportioning of certain aggregate sources, such as manufactured sands,
which can adversely impact the rheology of cementitious mixtures. The use of certain
chemical admixtures has been found to often minimize the need to increase cement
and water contents in orde~ to overcome the loss of workability that can accompany
aggregate sources which feature flat, elongated, angular, and rough particles.
In this study, a wide range of natural and manufactured sands, characterized for
gradation, mineralogy, shape, texture, and cleanliness, are evaluated for their effect
on mortar rheology, with and without a viscosifying-type chemical admixture. While
associations between aggregate characteristics and their impact of mortar rheology
may not be readily evident, the ability of this class of admixture can be shown
to mitigate the rheological effect of certain sands, and in some cases allow for
optimizing the mixture to lower paste contents.
Dr. Kati Hazrati, Product Manager, W.R. Grace & Co.- Conn., Cambridge, MA 02140.
Hideo Koyata, Senior Research Associate, W.R. Grace & Co.- Conn., Cambridge, MA
02140.
Dr. Paul Sandberg, Senior Research Associate, W.R. Grace & Co. -Conn., Cambridge,
MA02140.
Derek McGuire, Technical Specialist, W.R. Grace & Co.- Conn., Cambridge, MA
02140.
INTRODUCTION
Diminishing local availability of naturally occurring sands for use as fine aggregates in
the production of concrete, coupled with regulatory pressures to utilize by-products from
quarry operations, has brought about a growing dependency on manufactured sands [1,2].
Moreover, an economic advantage is often possible when replacing fine aggregates with
manufactured sands. The production of manufactured sands results from by crushing rock
deposits to produce a fine aggregate, which is generally more angular and has a rougher
surface texture than naturally weathered sand particles. Furthermore, rock dust
(microfines), a by-product of the crushing operation consisting of particles less than 75
1-\ID, can amount to as much as 20-30% of the manufactured sand. While some countries
such as India, Spain, and Australia permit from 15-25% microfines in the fine aggregate,
concrete fine aggregate gradation limits such as found in the ASTM C 33 Standard
Specification for Concrete Aggregates [3) permit up to 7% finer than 75 !-LID for some
applications assuming the absence of clay or shale. To comply with various particle
grading standards, the rock dust is removed by washing with water creating a waste
slurry whose accumulation imposes considerable detrimental impact on the aggregate
industry.
When manufactured sands are substituted for naturally sourced fine aggregate, even with
the fine than 75 ~--tm fraction controlled, the workability of concrete mixtures are
generally reduced, and in some cases, the concrete mixture can be more prone to
segregation. Furthermore, the increased friction that results causes higher pumping
pressures, increase the effort associated with concrete placement and finishing operations
[4). These performance attributes associated with manufactured sands result from a
combination of the angular particle shape and rough surface texture [5], both contributing
to increased void space in the aggregate packing. Both factors contribute to increased
surface area, which will require increased water content to maintain consistent
workability.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
This study is concerned with evaluating the interaction of both natural and manufacturing
sands with a bio-polymer type viscosity-modifying agent (VMA), a class of admixture
finding increased use with angular-shaped fine aggregates. From various particle shape
and mineralogical analyses made on sands from five sources, along with yield and
viscosity measurements on mortar mixtures, prepared with and without VMA, predictions
are attempted to associate the key particle attributes responsible for changes in mortar
rheology and the effectiveness of the VMA to maintain mortar workability without
altering mixture components.
TEST PROGRAM
Seven (7) manufactured and eight (8) natural sands, identified in Table I, were tested for
gradation (ASTM C l36 [7]), particle shape and texture (ASTM C 295 petrographic
analysis [8]),% finer than 75 j..lm, and Uncompacted Void Content (ASTM C 1252 [9]).
Mortar mixtures were prepared using the mixture proportions reported in Table l with
each manufactured and natural sand. Yield and viscosity measurements were performed
on the mortar mixtures using a prototype Contec coaxial cylinder rheometer shown in
Figure 1. The outer cylinder rotates at a programmable velocity while the inner cylinder
is stationary and registers the applied torque from the mortar. By assuming the mortar to
flow like a Bingham fluid, the yield value and plastic viscosity are calculated from the
slope (viscosity) and yield value (intercept) in a torque-velocity diagram. In this study,
focus was mainly on the yield value as it has been sown to correlate inversely with slump.
RESULTS
Mortar Rheology
The mortar yield stress results for all the mortar mixtures, prepared at constant w/c and a
constant weight of sand, either from manufactured or natural sources, are reported in
Figure 1. No consistent trend is apparent between the yield stress versus sand type
indicating other factors or combination of factors must be influencing water demand
changes created by the various sand properties. Mortar yield stress results, obtained with
the Contec rheometer, were found to correlate well with mortar flow, and thus help to
validate measurements with the rheometer (Figure 2). Attempts to further characterize
the rheological impact of natural versus manufactured sands by plotting yield stress
versus viscosity as shown in Figure 3 could not discern any specific trends, indicating
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
All the sand samples were subjected to a petrographic analysis performed according to
ASTM C 295 and C 136 Sieve Analysis. Key parameters noted during the examination
were mineralogy, shape, texture, and cleanliness, and are reported in Figures 4-7 for two
pairs of natural and manufactured sands, which some concrete producers are known to
either blend or use interchangeably. Numerical values, ranging from 1 to 10, were
assigned for shape, texture, and cleanliness, where lower the number implies high quality
and suitability for concrete manufacture. As expected, the manufactured sands designated
with the "HS" code have higher values versus natural sand available from a nearby
location.
In an attempt to identity the sand properties responsible for effecting the workability of
mortar and concrete mixtures, the yield stress of mortar mixtures prepared with both
natural and manufactured sands were correlated with fineness modulus, percent un-
compacted voids, and a combined factor representing the sum ofthe values assigned for
shape, texture, and cleanliness (Figures 8 -1 0). No clear correlations were readily
evident suggesting that either certain critical sand parameters have yet to be measured, or
that a higher level regression-type analysis is required to uncover more complex
relationships.
Probing Interaction of Sand-type using VMA Modifying Agent and Cement Paste
With the measured sand properties providing limited predictive information concerning
mortar workability, changes in mortar rheology as a function of VMA addition and
increased paste content were next evaluated. Starting with the control mixture described
in Table 2 and prepared with each of the natural and manufactured sands identified in
Table 1, the paste content was increased from 10 to 40% by mass. In addition, the control
mixture with the manufactured sand was dosed with 0.003% VMA actives by weight of
cement(% s/s). Yield stress was measured for the mixtures. A sample plot of the yield
stress as a function of increasing cement paste is shown in Figure 11. Interestingly, the
pair of selected manufactured and natural sands, exhibit a similar drop in yield stress with
increasing paste content. In the case of the VMA -treated mortar containing the
manufactured Mari-HS sand, a 75 Pa drop in yield stress was observed. Extrapolating the
lower yield stress to the yield stress-cement paste plot, the VMA appears to have the
potential of lowering the paste content by about 12% without altering the yield stress.
In Figure 12, the reduction in paste content observed with the addition of 0.003% s/s
VMA to II mortar mixtures prepared with the various manufactured sands. Interestingly,
the sands appear to have quite a diverse response to the VMA additive, exhibiting a six
(6)-fold change in yield stress.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
To understand if the yield stress reductions observed by both the addition ofVMA and
increased paste content have a similar mechanism in reducing the increased frictional
particle-to-particle forces associated with manufactured sands, the change in yield stress
produced by the VMA was correlated to the ratio of percent change in yield stress to
percent added paste (Figure 15). A somewhat unexpected qualitative trend suggests that
sands with a low to moderate yield stress response to the VMA have a high response to
increasing cement paste. This could possibly be explained by the proposed mechanisms
illustrated in Figure 16, where the increased paste content enables more significant
particle-to-particle separation under various degrees of shear, whereas the effect ofVMA
on yield stress is probably more shear history dependent.
CONCLUSIONS
Acknowledgement
The authors wish to thank Derek Brown, senior petrographer at the Grace Research
Laboratories, for the characterization of the sand samples.
References
4. Hudson, B., "Manufactured Sand- How it Effects Concrete," Quarry 4 (5), 30-39, 1996.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
100
50 -t-BI~III::--M-.--111111---f' !---li':J------~1'..ll--r.A-'I~.t---vl--1"...r-tlll-iB:-,..,.h
Q ~~~~~--~~~~~~~~~U4~~~~
R ~ ~ ~ 0 ~ W • ~ ~ ~ ~ G • ~ ~ ~
~ ~ ~ ~ ~ H ~ w oo ~ ~ d c ~ w n ~
~ $ M ~ ~ ~ ~ ~ ~ ~ ~
HS
WithoutVMA
+MNS
rn MGCRB
HM
~;CGC
&. FL
oO
130
E 120
E
l 110
~100
~
90
90
70
100 ..
, 200 200
Yield Stress, PA
300 350 •oo
10 20 30 40 50 60 70 80
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
HSSLO HS
100
.
= 8070
90
I. 60
1E
50
40
30
""
.... 20
10
100
90
."'
.5
60
..
!l
~
6()
50
l 40
E"30
..."
u
20
10
Shape: l Equant
Texture: 4 Sub angular
Cleanligess: I Clean
STC Factor: 6
100
90
{ 80
70
:. 60
~ 50
"'
.!!
~ 40
E 30
,.
~
(,)
20
10
Central Phoenix NS
100·~---~
90~:
so FM: 2.93
10 %Passing #100: 1.8
eo %UCV: 43.7
50 G'; 312.5 Pa
40 W:52.5J>a
30
20
10
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
•
<> Hanson Califomia NS
<> Florida Umestone NS
350 Central Phoenix HS
0 Central Phoenix NS
OCP2575
y 300
0 QCP4060
t>CP6040
I A+ <>CP8020
0 Hanson Durbin NS
E Hanson SLO HS
250 0 Hanson SLO NS
L ([]
•<>o.
d
G' 200
(Pa)
150
100
<>
• 0
50
2 2.5 3 3.5 4 4.5
FM
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
400.
• Granite Canada~
• Mtricapa River Bottom HS
• Hanson M:Jxican HS (Barge)
6. Mlricopa HS
•
A Hanson Vulcan NS
350
• Henson Califom8 NS
o Florida Unestone N5
CentraiftloenixHS
o Cerrtral Ftloenix NS
D CF'2515
•
0 Cf'ro60
h 0'0040
0
••
300 oCI'8020
o Hanson OJrbin NS
Hanson SLO HS
e HansonSlONS
(, 250
:!!
0 0
.!!!
>-
200- • 4,
150
••
<>
100- •
50
35 37 39 41 43
% Uncompacted Voids
450
400
DolomiteHS
350
o Grarille Canada HS
300 • Maricopa River Bottom HS
• Hanson M!xlcan HS (Barge)
'i' tJ. M3ricopa NS
e.,. 250
... Hanson SLO HS
b
:!! 200 0 Hanson SLO NS
~ • Hanson Dur!Jin NS
-' Central Phoenix HS
150
.t. Central Phoenix NS
100
50
0
3 4 5 6 7 6 9 10 11 12 13
STC Factor (Shape+Texture+Cieanliness)
Figure 10. Correlation of yield stress with combined aggregate properties of shape,
texture, and cleanliness.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Ui 200
't:l
1150
>-
100
50 1 - - - - - - - - - - , - . - - - - - - - - - - - ' - - - l
0 20 40
%Additional Cement
Figure 11. Change in mortar yield as a function ofVMA and cement paste
addition for mortar mixtures prepared with natural and manufactured sands.
18
c 16
0
=~ 14
!.$ 12
10
Ill
..
Ctl
CL
c
Cll
8
6
4
E
Cll
0 2
.~ 0
~
:;; 0 0
.. ~
:q" 1l
(.)
&~
8_ E
~
0
0
0:
0
1£ ~ 0 il
::; ~.!!
Figure 12. Reduction in cement paste content with 0.003% VMA s/s at
constant yield stress.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
20
~16-
18 •
Stress 14
• •
w/\IMA
12
••
10
8 •
6 ••
4
2
•
o-
0 2 4 6 8
Change in Yield Stress(% added paste
Figure 14. Correlation ofVMA and cement paste additions on yield stress of mortar
with various harsh sands. Sands with low response to VMA appear to respond well to
added cement paste.
....
grain surface polymer alignment creates sliding surface
•
Addition of -~'1
-..m
MortarmlX / .•
~'4f
Under low energy Under high energy
(Gravity, finishing) (Pumping, Vibration,
Cement paste ~
Paste volume Grain separation allows Enough separation
separates grains adequate flow creates sliding surface
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: Recently, the concrete workability has become regarded as much more
important due to large demands for high durable concrete with low water content
and unstable quality of concrete ingredients. The new superplasticizer was found to
provide better workability controlling the rheological properties of the fresh concrete
than the previous polycarboxylate type superplasticizers regardless of the quality
of concrete ingredients. The framework structure of the new superplasticizer was a
hyper-branched polymer in which specific function monomers were co-polymerized.
The new polymer of the hyper-branched structure was more condensed than that of
polycarboxylate type polymers and was highly adsorptive to cement particles, which
made a higher density of the polymer adsorption onto cement surfaces possible.
This dense and strong adsorption of the superplasticizer polymer molecules provided
the unique rheological properties of the concrete or the mortar such as the lower
viscosity of the concrete giving much easier handling. This new superplasticizer also
provided the robustness against varieties of the concrete ingredients qualities. This
superplasticizer for new generations has already been made practicable successfully
for several constructions, and this technology will bring various possibilities and
merits in concrete construction fields such as shortening the construction period
thanks to better workability.
INTRODUCTION
The rheological property of fresh concrete or mortar is one of the most important
parameters to detennine the workability in pumping, placing, and filling in a concrete framework.
The fresh concrete has been generally treated as a Bingham fluid [I] according to the following
equation (1) and Fig.l
r = (r - ry) I ?Jpi (1)
where r is the shear rates ofthe Bingham fluid; r is the shear stress; rv is the yield value;
is the plastic viscosity.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
7J p1
The shear rate r which is detennined by the plastic viscosity 7J pi and the yield value
r y represents the mobility of the fluid when the shear stress r is imparted. Therefore, r can
be considered as the representative of a fresh concrete workability. r 1 and 7J p1 can be
controlled by the mixture proportions of the concrete and the superplasticizer, while r depends on
the working condition at the construction site. Thus, the dispersion state of the cement particles by
the superplasticizer in the concrete would affect on the meological parameter r1 and 7J pi •
The dispersion state is controlled by the adsorptive states that are determined by the chemical
features of the superplasticizers.
In recent years, there is an increasing trend toward higher durability of concrete and, for
The chemical structure and the molecular structure of the new superplasticizer polymer
(NHBP) are shown in Fig. 2 and Fig. 3 in comparison with the conventional polycarboxylate based
superplasticizer (PC). Polycarboxylate-based superplasticizers (PC) or polyether-based
superplasticizers (PE) have carboxyl units and EO polymer units as the side chains, which are
believed to impart electrostatic repulsion and steric repulsion, respectively [2, 3, 4, 5, 6, 7].
In the new superplasticizer molecule, the carboxyl units are chemically modified by the
hydrophilic anionic groups (Z) to create a stronger adsorptive function. Moreover, the
difunctional monomers are introduced in the structure, so that the molecular structure of the new
superplasticizer is a hyper-branched polymer not like a linear polymer of a polycarboxylate based
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Cement and water mixed with 3000 ppm of the superplasticizer were placed in a
cylindrical vessel at W/C = 1.0. Mixing was carried out for 30 seconds. The supernatant liquid
was sep<mtted centrifugally at 3000 rpm from the paste. The total organic carbon content of this
liquid was then measured using Shimadzu TOC-5050A, and the concentration of the
superplasticizer remaining in the supernatant liquid was obtained from a calibration curve prepared
in advance. This yielded the amount of the superplasticizer adsorbed into the cement as the
difference between the amount of superplasticizer added and that remaining in the liquid.
In an other set of tests, 2000 mol/L of Na2S04 wa~ added together with the
superplasticizer solution. The adsorption amount of the superplasticizer with Na2S04 was also
measured to evaluate the effect of SOl ions, on the adsorption amount of the superplasticizers.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Cement and sand and water mixed with the superplasticizer were placed in the stainless
vessel of a capacity of 1000 ml (and of0 120 mm inside diameter). The mixture proportions and
the ingredients of the mortar are listed in Table 1. Mixing was carried out for 3 minutes at the
rotational speed of 200 rpm by the stirrer Z-2310 (of 0 110 mm with six choppers of 30 mm
height) manufactured by EYELA, measuring the electrical torque that corresponded to the mortar
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
viscosity. The mortar preparation apparatus is shown in Fig. 5. The mortar fluidity was
evaluated by the spread measured when a truncated cone (of 0 I00 mm and 0 70 mm diameters,
and 60 mm height) filled with the mortar was pulled away vertically.
Rheological Properties
The :rheological properties of the mortar samples as referred to the above were
measured by KMS Ball measuring system (where the ball diameter was 12 mm; the arm radius
was 37 mm; the thickness of mortar sample was 50 mm; the vessel diameter was 0 115 mm; the
ball center position was 18 mm higher from the bottom) of MCR500 manufactured by Paar
Physica (Fig. 6). The shear stress dependence of the shear rate was measured (from I0 Pa to 130
Pa for 40 seconds).
Cement and sand and water mixed with the superplasticizer were placed in a mortar
mixer at W/C = 0.5. Four kinds of European cements were used for the evaluations. The
mixture proportions and the used ingredients of the mortar are listed in Table 2. Mixing was
carried out for 2 minutes at the low rotational speed of 62 rpm. The mortar fluidity was evaluated
by the spread measured when a truncated cone (of 0 100 mm and 0 70 mm diameters, and 60 mm
height) filled with the mortar was pulled away vertically.
Concrete ~nufacturing
A biaxial-forced mixer with a capacity of 50 liters was used to mix 30 liters of concrete.
The cement and both sand and gravel were placed in the mixer and mixed for 10 seconds, then
The concrete flowing speed that was the characteristic index of the workability was
measured when the concrete ran down the tilting lane (23oC) by Tilting Flow Tester manufactured
by Newtech shown in Fig. 7.
Adsomtive properties
Figure 8 shows the adsorption amount of the superplasticizers (PC, NHBP) at W/C
= 1.0 with or without Na2S04• The polymer samples for the evaluation were selected to fix the
length ofEO chains (pin Fig.2, p = 9, 23 moles) and 1he molar mtios of the EO chain units (min
Fig.2) in their polymer structures. The new hyper-bmnched polymer (NHBP) showed much
larger adsorption ammmt and was less influenced by sot ions than the polycarboxylate-based
polymer (PC).
The higher adsorption amount ofNHBP could be explained by the compactly arranged
adsorption of the condensed hyper-branched structure on the cement particle surfaces.
The adsorptive force of PC was weaker than SOl ions so that the adsorption of PC
was inhibited by sot ions, as previously reported [8]. On 1he other hand, the adsorptive force of
NHBP was thought to be so strong that the adsorption was not inhibited by SOl ions.
Mortar viscosity
Figure 9 shows the relationship between the mortar fluidity and the viscosity that was
converted from the electrical torque during the mixing. NHBP showed much lower viscosity than
PC at the same mortar fluidity. The lower viscosity meant that the amount of the committed water
by the gathering cement particles was smaller, which suggested that NHBP dispersed the cement
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
To investigate the difference of the rheological properties ofPC and NHBP in details,
their dynamic rheological properties were studied. Figure I0 shows the shear stress dependences
of the shear rate of PC and NHBP at the same mortar fluidity of200 mm. Both PC and NHBP
exlubited the rheological behavior like Bingham fluids. Table 4 shows the plastic viscosity and
the yield value for each mortar sample. As the paste volume in the mortar mixture for PC type
was increased while W/C ratio was constant, both of the plastic viscosity and the yield value were
decreased dramatically, which could be supported by the excess paste theory [I 0].
As can be seen from the figure, NHBP showed much lower plastic viscosity and lower
yield value than PC though the mixture proportions and the mortar fluidity of them were same.
The lower plastic viscosity ofNHBP than that of PC could be explained by the higher
adsorption amount of the superplasticizer molecules, that lead to the denser adsorption state of EO
chains on cement surfaces [11].
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The lower yield value ofNHBP might also be explained by the denser adsorption state
of EO chains on cement surfaces. A yield value of a slurry is supposed to be reflected in the
coagulation state of the particles. The dense adsorption state of EO chain on cement surfaces (m
I) was considered to enhance the repulsive force (Pp) according to the following formula (3)
[12, 13], so that the dispersion state at the dynamic condition became stable and the coagulation state
at the static condition became soft:
PR(h)==kT7ramF [(D0 !hf3 + 1] (3)
where a is the particle radius (m); h is the distance between particles (m); D 0 is the thickness of
adsorbed layer (m); K is Boltzmann constant; Tis the absolute temperature (K); m is the number of
EO chains in one molecule; r is the number of adsorbed molecule per unit area.
The lower plastic viscosity and the lower yield value ofNHBP based superplasticizer
would conduce to better workability in the concrete than PC based one. Furthermore, as can be
seen in Table 4, the required cement paste volume in a concrete mixture design might be able to be
reduced to get the predefined workability for NHBP compared to PC.
Figure II shows the compatibility test results ofNHBP compared with the commercial
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
PC based superpiasticizers, PC-EI and PC-E2. The dosage of the each superplasticizer was
determined to get the fluidity of 2I 0 ~ 220 nun when testing CEM I . Although the significant
differences of the mortar fluidity were obseiVed between the cements for the PC based
superplasticizers, NHBP based superplasticizer showed stable mortar fluidity independent of
cement varieties. The robustness against cement varieties of NHBP might be explained by the
stronger adsorptive force.
Concrete tests
Table 5 shows the flowing time, the slump retention and the hardening properties of the
concrete with NHBP based superplasticizers, (NHBP-1, NHBP-2, NHBP-3), the EO chain length
and molar ratio of the components of which were varied each other, and with the commercial PC
based superplasticizer (PC-A). The concrete were manufactured according to the mixture
proportions (WIC ~ 0.35, 0.55) shown in Table 3.
The excellent workability of NHBP based superplasticizers was observed in the
concrete tests. The flowing speed of NHBP-A was much faster than that of PC-A, so that the
better workability ofNHBP was confirmed also through the concrete tests. Figure 12 shows the
comparison of the fresh concrete prepared with NHBP-A and PC-A just after the initial slump test
The fresh concrete prepared with NHBP looked much softer and much more workable as the
gravels in the concrete were homogeneously covered by the mortar layer.
The slump retention and the required dosage ofNHBP-A were similar to those of PC-A,
and NHBP-B and NHBP-C showed better slump retention with higher dosage than PC-A The
setting time ofNHBP-A was almost same as that of PC-A, and the setting time ofNHBP-B and C
were a bit longer than that ofPC-A. Therefore, the slump retention and the setting time ofNHBP
could be controlled freely by adjusting the polymer structure.
The strength developments ofNHBP were slightly better than that of PC. The finer
dispersion state by NHBP based superplasticizer that would lead to the improvement effect of the
cement hydration might explain the superiority in the compressive strength by NHBP.
NHBP based superplasticizers have already been used in many applications both for
ready-mixed concrete and precast concrete, and are currently being used on a daily basis. The
CONCLUSIONS
I) The hyper-branched polymer with the hydrophilic anionic group was developed as the new
superplasticizer (NHBP).
2) NHBP based polymer showed much higher adsorption amount and stronger adsorptive
force to cement particles than the conventional polycarboxylate type polymer (PC).
3) The plastic viscosity and the yield value of the mortar prepared with NHBP were much
lower than those with PC, which could be explained by the denser adsorption state of EO
chains on cement surfuces.
4) NHBP based superplasticizer appears to be much more robust in terms of stable
performance of the initial fluidity against the cement variations.
5) The better workability induced by NHBP based superplasticizer was confirmed through the
concrete tests.
6) The basic performances ofNHBP based polymer required as a superplasticizer such as the
fluidity retention and the strength development were confirmed through the concrete tests
including the job-site evaluations.
REFERENCES
[I] Tanigawa, Y., "The Fluid Properties ofFresh Concrete", Cement Journal, 2004.
[2] Shonaka, M., Kitagawa, K., Satoh, H., Izumi, T. and Miztmurna, T., "Chemical Structure
and Performance of New High-Range Water-Reducing and Air-entraining Agents",
Superplasticizers and Other Chemical Admixtures in Concrete, Proceedings; Ed. Malhotra,
V., M., 5th CANMET/ACI International Conference, Rome, Italy, 1997, pp.599-614.
[3] Hamada, D., Sato, T., Yamato, F., and Mizunuma, T., "Development of New
Superplasticizer and Its Application to Self-Compacting Concrete", Superplasticizers and
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Other Chemical Admixtures in Concrete, Proceedings; Ed. Malhotra, V., M., 6th
CANMET/ACI International Conference, Nice, France, 2000, pp.269-290.
[4] Izumi, T., Satoh H., Yamamuro, H., Hamada, D. and Mizunurna, T., "A New Hybrid Type
Superplasticizer", 7th CANMET/ACI International Conference, Berlin, Germany, 2003,
Supplementary Papers, pp.67-81, available from CANMET, Canada
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Admixtures in Concrete, Proceedings; Ed. Malhotra, V, M., 5th CANMET/ACI
International Conference, Rome, Italy, 1997, pp.187-202.
[8] Yamada, K., Ogawa, S. and Hanahara, S., "Working Mechanism of Poly-Beta-Naphthalene
Sulfonate and Polycarboxylate Supe~plasticizer Types from Point of Cement Paste
Characteristics", Superplasticiz.ers and Other Chemical Admixtures in Concrete,
Proceedings; Ed. Malhotra, V, M., 6th CANMET/ACI International Conference, Nice,
France, 2000, pp.367-382.
[9] Aoi, K., Kakimoto, M., "Dendritic Polymer'', N.T.N., 2005.
[10] Kennedy, C. T., "The Design of Concrete Mixtures", Proc. ACI, 28, 1932, pp.419448.
[11] Shirota, K., et al., "Effects of Structures of Polycarboxylate Dispersants on Viscosity in
Concentrated Slurries", J. Am. Ceram. Soc., to be published.
[12] de Gennes, P. G, "Polymers at an Interface; A Simplified View" Review Article, Advances
in Colloid& Interface Science, 27, 1987, pp.189-209.
[13] de Gennes, P. G, "Scaling Concepts in Polymer Physics", 1979.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
SP W/C
J'C.A
NHBP
-A
0.35 0275 149" 4:45 6:50 62.8 74.7
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
·>-
-en
CD
I!!
....
ro
I
I
CD
:J::.
(/)
l:'y
.
Shear stress -r
Fig. 1 -- The rheological behavior of Bingham model. The shear stress dependence of
the shear rates. (-r v : yield value, 11 pt : plastic viscosity)
PC type
·""'""'";,;""''"~. . . / f,--r )
difunctional monomer cH,
1!
!:...................................!
:
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
PC type NHBPtype
EO chain
1111111111111~ ~\............,........,..
Main chain
0 r-----------------------------~
-0.2
-0.4
~
8 -0.6
Ul
5 -0.8
()
·u;
:S -1
c:
;; -1.2
.3
-1.4 ----~- NHBP ___ _
(E0=9mol)
-1.6
a=0.22
-1.8 L.-______.________...._______...._______.
Fig. 4-- The relationship between the molecular weight [M] and the intrinsic viscosity
[TJ] oft he superplasticizer polymer with the side chains of EO 9mols.
(Tri-SEC RALLS System)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
1000ml
stainless
vessel
110mm
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
c 1000
"E
0
"'c::
0 500
e.
0
"'
"0 0
"' PC-1 PC-2 NHBP-1 NHBP-2
(E0=9) (E0=23) (E0=9) (E0=23)
Fig. 8 -· The adsorption amount of the superplasticizer at W/C = 1.0 with or without
Na 2 50 • (3000 ppm of superplasticizer, 2ooo mol/L of Na2 50)
4
Q)
:::>
~ 3200
Q)
:5 3000
E
,g "'2aoo
al ~
"Iii ~2600
>~
c:
8 2400
~
§ 2200 i-- -------:---
Fig. 9 -- The relationship between the mortar fluidity (flow) and the viscosity
converted from the electrical torque when mixing.
- - - - - - - - - - t - - - - - - - - - - _;_
: I •
•
-~~~1!1-~l_ ~----------
m ' '
iii 10.0 ----------~-----------~-----------L----------
:pc
5.0
0.0
0 50 100 150 200
Shear stress E'aJ
Fig. 10 -- The shear stress dependence of the shear rates of the mortar prepared with
PC and NHBP (at W/C = 0-40 and the mortar flow of 200 mm).
280 r:::====~-------,
260 I!ICEM I 42.5R
ClCEMllAM42.5N
240 EICEMmA32.5
~ 220 Ji---=-:=:.:.:..:..:c:.:.::=-:___.J_
-; 200
0
': 180
~
~ 160
140
120
100
PC-E1 PC-E2 NHBP
Fig. 11-- The test results of the cement compatibility in the mortar
(W /C = o.so, PC-E1, PC-E2; European commercial products)
Fig. 12 -- The comparison of the fresh concrete prepared with PC-A and NHBP-A after
the slump test. (W/C = o.ss)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In this study, the authors have developed a new shrinkage-reducing SP by mixing a
shrinkage-reducing component into a SP for ultra-high-strength concrete, and have
studied the effects of this new admixture on both shrinkage reduction and strength
development of concrete. In particular, the effect on strength development was studied
under conditions at high temperature history and of standard curing.
SUMMARY OF TESTING
various measurements and prepared specimens to test the hardened properties. The test
temperature was set at 20°C.
The measured properties of fresh were slump flow, time taken for the flow to reach
the 50-cm point, air content and setting time. The hardened properties were compressive
strength up to the age of 91 days (by standard curing and simple adiabatic curing), and
autogeneous and drying shrinkage strains. The method of measurement for each item is
shown in Table 4.
For simple adiabatic curing, the authors prepared 10 cylindrical specimens (diameter
of 10 em x 20 em in height) sealed-cured and placed in an insulated box (outer
dimensions of 80 x 80 x 80 em, inner dimensions of 40 x 40 x 40 em) as shown in
Figure 4. The internal temperature of one specimen was measured. After sealed curing in
Table 5 shows the results of the tests of fresh properties. The dosage of SP required
to obtain a slump flow of 70 em increased as the amount of shrinkage reducing
component increased. It increased by 0 to 15% compared with SPH. The dosage of
SPSR(l) containing the standard amount of shrinkage-reducing component increased by
approximately 8%.
The setting time was delayed as the dosage of shrinkage reducing agent increased.
The setting time of SPSR(l) containing the standard dosage of shrinkage-reducing
component was retarded by 15 minutes in concrete with a W/C = 0.22, 20 minutes in
W/C = 0.18 concrete, and 1 hour in W/C = 0.15 concrete compared with SPH.
Compressive strength
In general, when concrete with a specified strength of 80 MPa or less is placed into a
concrete column having a relatively large sectional area, it is subjected to high
temperature history at an early age due to a build-up in internal temperature. It is well
known that the compressive strength at the age of 7 days of concrete in an actual structure
is higher than that of concrete cured by the standard method. On the other hand, after the
age of 28 days, the compressive strength of the actual structure is lower than that of the
standard-cured concrete, as seen in Figure 6. However, in Japan, there are some examples
(including the authors' results) of concrete having the specified strength range of 100
MPa or more and exposed in high temperature history at early age, exhibiting a much
higher strength development than that of standard-cured concrete even at the age of 91
days [9,10], as seen in Figure 7. The reason of this is because the W/B ratio is extremely
small in this strength range. As a result, hydration cannot proceed easily under standard
curing conditions, and even by the age of 91 days it is still less than the hydration level of
concrete that has been subjected to a high temperature history at an early age. In other
words, if concrete in an actual structure is subjected to a higher temperature history,
strength development up to the age of 91 days will increase. Therefore, if a shrinkage
reducing admixture that retards setting time of concrete is used in ultra-high-strength
concrete, the retardation will affect the internal temperature history. Therefore, one has to
be aware of the decrease in both the maximum temperature and accumulated temperature.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Autogeneous shrinkage
Figure 9 shows the autogeneous shrinkage up to the age of 91 days. The autogeneous
shrinkage of ultra-high-strength concrete containing SPSR decreased by 15 to 30%
compared with that of SPH concrete. The reduction in autogeneous shrinkage increased
as the dosage of the shrinkage-reducing component increased. The absolute value of
reduction increased as the W/C ratio decreased. The autogeneous shrinkage strains of
concrete with W/C = 0.22, 0.18 and 0.15, containing SPSR(l) at the age of 91 days
decreased by lOOJ..Lm, 150J..Lm, and 220J..Lm, respectively, compared with SPH concrete.
The autogencous shrinkage strain of W/C=0.15 concrete containing SPSR( 1) can be
reduced to a level equivalent to that ofW/C=0.22 plain concrete (without the shrinkage-
reducing component).
Figure 10 shows the drying shrinkage strain up to the age of 91 days. The drying
shrinkage of ultra-high-strength SPSR concrete decreased by 3% to 25%, compared with
SPH concrete. In the same manner as the autogeneous shrinkage strain, the shrinkage
reduction increased as W/C decreased. However, there was a difference in the reduction
rate between drying shrinkage strain and autogeneous shrinkage strain in terms of the mix
proportion or amount of shrinkage-reducing component. In W/C = 0.22 concrete,
SPSR(l) at the standard dosage lowered drying shrinkage strain by 15% or so, almost
equal to the 17% achieved by SPSR(3); however, SPSR(2) containing less component
lowered the drying shrinkage strain by as much as 5% or so. In addition, SPSR(l) at the
standard dosage decreased the drying shrinkage strain by l 5% and SPSR(2) containing
less component decreased it by approximately 8% in W /C = 0.18 concrete.
CONCLUSIONS
REFERENCES
3
s/a Unit content (l<g/m )
W/C SP
Water Cement Sand Gravel
(%) Kind Dosflge
SPH Cx1.3%
SPSR(1) Cx1.4%
0.22 48.1 150 682 768 828
SPSR(2) Cx1.35%
SPSR(3) Cx1.5%
SPH Cx1.55%
0.18 43.4 150 833 634 828 SPSR(1) Cx1.65%
SPSR(2) Cx1.55%
SPH Cx1.7%
0.15 37.1 150 1000 489 828
SPSR(1) Cx1.85%
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
SPH 1.7 123 150 - 158 156 162 170 Strain (•10"") -770 -790 -802 -110 -210 -250 -290
0.15 Strain ( •1 O"") -552 -560 -569 -90 -160 -190 -225
SPSR(1) 1.85 122 152 - 165 158 169 169
Reduction (%) 28.3 29.1 29.1 - 23.8 24.0 22.4
Daneshlink.com
6o Sugamata et al.
1200
e Metllod 111iag· tlae -bedded
-=tooo
..&
.)II.
1: 800
electric strain pu• traaadacer
•t:
.l!
Ill
Ill 600
~
0
Cl>
1:
Cl>
400
1 200
~
0
1200
-¢-50-0 -Q-50-7.5 -l:r-5&-15 *-30..0
! Cl>
1000 +3o-7.s -o-3o-ts +2o-7.s
"'
~ 800
-e
.c
Ill
Ill 600
=
2
1:
Ill
400
Cl
.s 200
.¥
0
t to 100 1000
Time (Ages)
Explanatory note represents:
0 --~~ ~---~----·----~---------~---··----~---···---:----·-····:·---------:-···
. ' . ' . . '
c:
:g
Ill
-200 ____ .. J.;~--~--~.-~..J.;.-:-:..~..-:-:--~__j_~~..~---:-:..-:..-:7..:.-:-:..:-:-:..:-:-:..:-:-:..;..:.:-;:,..:-:-..,.,........~.:L:Pc~o·~~----
: : LPC+SF 0.18 : :
G)
til
I o
'I .:
0
0.02 ~----------------f.~::0::jLUPiCC~+:SS~F00~.1iil2
~LPC+SF 0.15
0.00 +--ill-
""-------------1~LPC+SF 0.18
~ ~LPC+SF 0.22
+-.....,,'LT:l.....~:::::-::---)..-----__,
.!o
;
j
~
c -0.04
Ill
'fi
-0.02
t--~,~~~L~~~~~~~~~~~~~~~~~~
~
+----"'·"'-..---------------1
....... OPC 0.55
~ ~ -~
~
g' -0.08 +-----------=~~-
-0.06
-.----... . . . .-.===~----~
....•
-0.10 + - - - - - - - - - - - - - - - - - - - - t
-0.12 +---""'T"--r-.------..........,--..,-._-r....,~,....-.4
4 12 20 28 36 44 52
Age(weeks)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
80cm
r l :.~ ; 'l}om
L.ll<ll~:~~~-:_-:_-_-_-_-_-_-_-40-c_m__-_-_-_-:_-_-_-:~'~'..,~-.J"'~cm
Figure 5 Specimen of autogeneous shrinkage
1 28 58 91
Age(days)
7 28 56 91
Age (days)
E
.! 4!!
i1.)
c
0
t)
30
20
10
80
0
t lw/C=0.1al
2 3
?::l
4 5
--SPH
6 7
0 70
--SPSR(1)
I!! 60
~Ill
· ·- ·SPSR 2)
a. 50
E
s 40
sI!!
1.)
30
c
0 20
t)
10
0 2 3 4 5 fl 7
80
,...
0 70
e
a!:! 60
a. 50
ID
E
.s 40
.se 30
1.)
1:
0 20
t)
10
0 2 3 4 5 6 7
Age Oays)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
4( -1000 .-
0 20 40 60 80 100
Age(Days)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
l
'lii -200 --<>-SPSR(3)
41
i ~00
-~ 4oo ro;===~---------~--J
i!' I W/C=0.221
0
~oo+---,----r---r--~--~
0 20 40 60 80 100
AAe(Days)
0~------------~====~
i; -<r-SPH
:'E.-100 +-'lr-----------l~SPSR(1)
.5 -o - SPSR 2)
i~oof-~~~~~~---~::::::::~
Cll
:f
~~00+--------~~--~~~~-4
"C
~
"'a40oro===~-----------i
~ IW/C=0.18I
c ~0 +----.-----.------.-----.----1
0 20 40 60 80 100
AaeiDavsl
i;
E. -100 1-"H"-....::----------j
c ~~~~"
~ -200 +------==:::::.;:~=====:::il-..__=--,------l
41
"'
~-300+-------------~~~
"C
~
"'
~4oor.;===~----------~
·~ lwtc=o.1sl
0
~00+---,------r--~---,--~
0 20 40 60
Age(Days)
80 100
Figure 10 Drying shrinkage strain
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: Natural stones have always represented major aesthetic factors in building
construction. One ofthe more ancient and important applications, is the production
of decorative floorings. The most famous example in the flooring field is the "Terrazzo
alla Veneziana". ln recent years, the extensive exploiting of quarries and consequent
environmental problems have reduced the availability of good low cost quality stone.
In the present work, artificial superplasticized aggregates have been produced and
used as decorative elements in this new "Terrazzo" flooring. The substitution of
natural stones with these artificial materials is a first application of new practice.
These aggregates are produced by combining inorganic binders, finely divided
carbonaceous and/or siliceous minerals, superplasticizers and water. The use of
nanostructural polycarboxylate superplasticizers allows to obtain very low water to
cement ratios and, consequently, the final products are characterised by physical
and mechanical properties, similar to those of natural stones. Additional ingredients,
such as pigments and other materials have been added in order to obtain artificial
aggregates with the desired properties and outstanding aesthetic characteristics.
Tiziano Cerulli graduated in organic chemistry at the State University of Milan, Italy.
Since 1989 he has been working as chief in charge of the analytic department of Mapei
Research Laboratory in Milan. He is author of several papers and patents, most of them
in the cement field.
Roberto Pellay is managing director of IN. T.EC., Venice, Italy, a company specialised in
the research and application of new technologies in the environmental field. He is author
of several patents in this field.
INTRODUCTION
This paper describes the production and the characteristics of artificial cement
aggregates and their application for the production of artificial stone ("Terrazzo")
floorings. This technique, produces artificial stones of selected colour and size, which are
incorporated in a fresh binder matrix, according to predetermined mixture proportions.
After the hardening of the binder, the subsequent smoothing and polishing operations
emphasize the design and the colour stone effect. The traditional and most ancient
technique, for the casting of the "Terrazzo", employs calcium hydroxide as binder but,
more recently, portland cement and epoxy resins have also been used. In other
applications, selected coloured stones can be combined with different types of binders,
based on both cement and organic resins, in precast elements of various shapes. These
precast elements are used for the creation of decorative objects, such as linings, tiles,
shelves and bases for furniture. Furthermore, natural stones, without a binder, can be used
as important elements in interior and exterior decorative gardens and flower-beds.
Process of production
the grains and the production of a roundish granular material, with the desired size, is
accomplished. The curing of the fresh grains may be conducted in bulk, at ambient
temperature or by thermal treatment. The artificial aggregates, obtained by the
aforementioned process and used for the "Terrazzo" flooring, are shown in Fig. l.
Materials
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
produces artificial aggregates of suitable gradation and excellent yield. These aspects, in
conjunction with the strict control of colour by the dosage of the selected pigment or
mixtures of pigments, make these artificial aggregates much more uniform, in term of
aesthetic characteristics, as compared to natural stone. Table 3 shows the physical and
mechanical characteristics of the different artificial aggregates. These results confirm the
combined effect of the reduction of W /C and the addition of calcium stearate in reducing
water absorption of the aggregates. Particularly, the addition of calcium stearate is
necessary to reduce water absorption below 4 per cent. Furthermore, the addition of
calcium stearate does not produce any adverse effect on the resistance to abrasion and
wear, which resulted comparable to those of natural stones (3).
Water to cement ratio- The mixing mass, during the production process appears
quite different from that of conventional cement mixtures. Although the concept of
flowability cannot be applied in this case, the results confirmed a surprising and
remarkable effect of the superplasticizer on the amount of the water necessary to form the
artificial aggregates. As already pointed out by other authors (4), the addition of the
superplasticizer to a cement mixture results in the liberation of the water that normally is
entrapped inside cement particle agglomerates. In this case, the higher amount of
available water makes the mixture more wet and increases the plasticity and the
stickiness of the "grain seeds". The rotating plate generates larger agglomerate growth to
the final aggregate size. The test results of Table 3 confirmed that the reduction of the
mixing water greatly improves all the physical and mechanical characteristics of the final
product.
Colour stability - The pigments used for the production of artificial aggregates
are based on metal oxides or organic compounds stable to the light and in the alkaline
environment of cement mixtures and no change in the colour of the aggregates was
observed with ageing. Furthermore, the addition of calcium stearate can effectively
prevent the efflorescences from calcium hydroxide produced by cement hydration, when
the aggregates are used for outdoor applications.
Resistance to abrasion and wear -- Abrasion and wear tests were conducted only
on samples from 4 to 6 according to methods EN 1097-2 and 1097-1, respectively. In all
the cases, the results indicated that the aggregates showed an excellent resistance both to
abrasion and wear, comparable to or higher than the natural aggregates. In fact, as
reported by other authors, typical values of the Los Angeles abrasion test for basalt,
limestone, quartz and granite are 14, 26,28 and 38, respectively (5).
Preparation and Casting of the Mixture - Artificial aggregates, water and mortar were
added to the cement mixer in the following proportions.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The cement mixtures were poured and levelled over a concrete screed,
previously treated with a synthetic primer, at a thickness of about 20 mm by means of a
levelling bar. The paving was air cured for two days, before starting smoothing and
polishing operations.
Smoothing and Polishing the Hardened Mixture - The hardened raw surface was
worked with a dry-smoothing machine to remove the first 5 mm of the surface, which
exposes the aggregate colour and shape. A phase of the smoothing process in shown in
Fig. 5. By changing the fineness of the abrasive heads, the same machine has been used
for polishing operations, giving a final surface with large colour attraction (Fig. 6). In Fig
7, the appearance of the final beautiful "Terrazzo" flooring is shown. Finally, the floor
was treated with a synthetic sealant and waxed.
CONCLUSIONS
The results of this work describes a new application of superplasticizers, for the
production of artificial aggregates, which can be used for decorative applications with
excellent aesthetic characteristics. The function of the superplasticizer is to disperse
cement particles, and release the water entrapped by the cement and the filler
agglomerates. The superplasticizer also makes "grain seeds" previously formed and
favours their agglomeration in a rotating plate. The use of nanostructural superplasticizer
combined with hydrophobic agents, such as calcium stearate, allows high yield
production of artificial aggregates, with physical and mechanical characteristics, suited to
be used as concrete aggregates. On of the most fascinating applications of the artificial
aggregates is the production of"Terrazzo" flooring.
REFERENCES
2. Clemente, P., Dal Negro, E., Ferrari, G., Gamba, M., Surico, F., "Nanostructural
Superplasticizers for In-Lining Concrete in Underground Projects", in
Applications of Nanotechnology in Concrete Design, Editors R.K. Dhir, M.D.
Newlands, L.J. Csetenyj, Thomas Telford Pub!, 2005, 97-104
4. Dodson, V., "Concrete Admixtures", Van Nostrand Reinhold, New York, 1990,
pp. 39-40
Sample Calcium
Cement Water CaC03 Pigment Superplasticizer
Stearate W/C
N. (%) (%) (%) (%) (%)
(%)
I 24.9 12.5 61.2 1.8 0.0 0.0 0.50
CrzOJ % O.o7
Mn02 % 0.05
c,s % 39.9
c,s % 45.1
C 3A % 9.3
C.AF % 0.5
Mean particle size J.liD 29.70
Volume of particles below 40 ~m % 80.2
Resistance to Resistance to
Sample Density Water Absorption
Abrasion Wear
N. (kglm3) (%)
(LA units) (MnE units)
I 1990 10.0 - -
2 2060 6.9 - -
3 2290 5.5 -
4 2350 4.5 22 13
5 2340 2.0 21 13
6 2340 1.0 22 14
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
High Mw PEO I
I (4,000-6,500) ~
E
c
....
lowMwPEO
(750-1,500) ...
0
4.5 nm *l!!ll!ll!l!!!!l!l!l!!ll!l!lll!
29.5 nm 19.3 nm
Conventional PC Nanostructural PC
Superplastlclzar Superplasticizar
MW50.000 MW61.000
.u
""
1-
26
24
22
20
0 5 10 15 20 25
TIME (HOURS)
50
Q40
w
;!;
~30
w
a:
Z20
s~ 10
11.
12 8 passing 1
SIEVE MESH (mm}
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
The control of shrinkage is needed when external or internal restraint induces stresses
higher than the tensile strength of concrete. The reduction of shrinkage may also be
necessary in long bridge structures due to an increase of second order effects. In order to
decrease shrinkage deformation of concrete, a simple solution is the use of shrinkage
reducing admixtures (SRA) 1-6 . The accepted mechanism of action of these non-expansive
products is surface tension reduction of capillary pore solution 1•6 . Knowing that the
surface tension of a solution of water and SRA is proportional to the content of SRA, and
considering surface tension decrease the only effect of these admixtures on concrete, we
can expect proportionality between shrinkage reduction and admixture dosage, assuming
valid the liquid depression given by Young-Laplace equation (1/:
2cr
p -p 1 =-case (I)
g r
where Pg is the gas pressure, PI the liquid pressure, cr de surface tension of liquid, e the
contact angle and r the capillary radius. In fact, this is not the case and the use of a SRA
leads to changes on concrete other than surface tension reduction of pore liquid8- 11 •
liquid contains SRA molecules. The change in surface tension not only has a direct effect
of decreasing the liquid depression, but also indirectly changes the mobility of the
adsorbed water in the pore structure 18 •
This study does not investigate the existence of these different mechanisms in mortars
with SRAs, but rather evaluates the overall influence of increasing dosages of two
different SRAs on total (unsealed specimens) and autogenous shrinkage, flexural and
compressive strength, and mass loss of mortar specimens. The results help to identify
changes in these properties that can be used to better understand secondary effects when
using this kind of products.
Nine mortar mixtures were prepared using portland cement, CEM I 42.5 R, (Table l ), in
accordance with EN 197-1; siliceous sand, in accordance with CEN standard sand
described in EN 196-1, two liquid SRAs, SRA-0 (high molecular weight polyglycol) and
SRA-K (alkyl-ether), Table 2, and distilled water. The mixing of mortars was performed
following EN 196-1. The dosages of each batch are presented in Table 3. The SRAs were
diluted in distilled water.
TEST METHODS
The following tests were performed on the nine mixtures: flexural strength, compressive
strength, weight loss, total shrinkage and autogenous shrinkage. The compressive
strength was measured according EN 196-l standard. Mass loss and shrinkage were
measured on 40 mm x 40 mm x 160 mm prisms, prepared according to EN 196-1. The
prisms were removed from the moulds at 18 hours. The prisms for autogenous shrinkage
measurements were then sealed with several plastic sheets and placed in a closed plastic
box, to avoid water drying, in a laboratory room at 20±2 °C. The prisms for mass loss and
total shrinkage measurements were put in a laboratory room at 20±2 oc and 65±5 % RH.
In addition to tests on mortars, surface tension measurements were performed in solutions
of distilled water and SRA. The measurements were made with a Kl2-Tensiometer, from
KROSS GmbH, using the plate method, at 20 oc.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
on the prisms for shrinkage tests. Each result presented is the average value of three
measurements.
Tables 4 and 5 present flexural and compressive strength results, respectively. The
strength decrease, in percentage, with the SRA dosage is presented in Figs. 1-2. Fig. 1
shows a significant decrease of the flexural strength with increasing SRA dosage at 18
hours. However, at later ages the difference between reference and SRA mixtures is
reduced, being very small at 90 days. Fig 2 shows a similar trend on compressive
strength, but the difference remains clear at later ages, being about 20 % at 90 days with
the highest dosage of SRA. The different behavior on flexural and compressive strength
al later ages may be related with the conservation of the specimens. The prisms for these
two tests were kept under water, as specified in EN 196-1. The flexural strength mainly
depends on the tension resistance of surface layers of the specimens. For long periods of
contact between mortar and water, the leaching of interstitial pore liquid on the surface
leads to a similar superficial pore structure of the prisms with or without SRA. This
phenomena has minimal influence on compressive strength which mainly depends on
Fig. 3 shows surface tension of solutions made with distilled water and the SRAs. The
graphic on the right, Fig. 3b, is an enlargement to show better surface tension changes for
usual contents on interstitial pore structure. The SRA 0 is shown to be more efficient in
reducing surface tension. The product alone, 100 % SRA, has a surface tension of about
30 mN/m, at 20 °C.
Fig. 4 shows the mass loss ofthe specimens used for total shrinkage measurements. Mass
loss of the specimens used for autogenous shrinkage was also measured, and the higher
result obtained was 0.18% at 181 days (an evaporation rate lower than 0,001 %/day),
which was considered low enough to despite the mass exchange.
In general, the mass loss of non-sealed specimens of SRA mortars was higher than the
mass loss of reference specimens. The evaporation increased for higher SRA dosages.
The performance of the two SRAs tested was similar for low dosages, but diverge for
high dosages. Particularly, the mixture with 3% of polyglycol-based SRA 0 evidenced a
high initial drying rate. The main difference between the mortars 04 and K3 was detected
at two days. This worse behavior of04 can be attributed to the low surface tension of the
interstitial pore liquid and to the low initial rate of hydration (see also strength increase
between 18 hours and 2 days, Figs. 1-2 and Table 4-5). The results show that the mass
loss does not depend only on the surface tension, but also on the pore structure. Figs. 3-4
show that the surface tension variation of the pore liquid between 03 and 04 is small, but
the increase of mass loss is very high. Changes in pore structure, which leads to different
mass losses, may modify not only RH profiles but also hydration conditions of surface
and inner parts of the specimens. These phenomena may have major effects on total
shrinkage results.
When extrapolating these results for concrete specimens it is necessary to account for
different surface/volume changes. As reported by others4•6•8•9 •10, the mass loss of SRA and
reference mixtures is not much different. In fact, with the exception of mixture 04, which
had a significant increase of water evaporation in the first days, despite the high value of
surface/volume of the specimens and the short period of cure, the difference between
reference and SRA mortars is less than 1%. For concrete specimens with much lower
surface/volume values and longer periods of cure, the negative effect of the possible
increase of water evaporation rate due to SRA is not expected to be significant in
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
due to changes in surface tension 18 , and possible curing effect promoted by increase of
SRA concentration at the surface9 , is expected to have beneficial effects on reducing
shrinkage.
Fig. 5 shows autogenous shrinkage of the nine mortar mixtures (contraction represented
by positive values). The results indicate that the two products perform differently but
both are effective in reducing autogenous shrinkage. The differences between SRA 0
and SRA K mortars are not only on the amplitude of shrinkage reduction, but also on the
relative behavior, which depends on SRA content and age. At later ages SRA 0 seems to
decrease continuously the autogenous shrinkage with the dosage, which is not the case
with SRA K, which shows approximately the same shrinkage at 181 days for all dosages
of SRA. At early ages, and for 2-4 % of admixture, the results indicate a higher potential
of the alkyl-ether SRA (K) to decrease autogenous shrinkage.
The influence of SRA on autogenous shrinkage may be due to: changes in chemical
shrinkage rate, modification of the solid part of the pore structure due to differences in
hydration rate, changes in internal pore filling and mobility of the pore liquid due to
surface tension reduction; modification of surface free energy and disjoining pressure due
to the presence of SRA molecules in the pore structure. The different chemical nature of
the SRA products, polyglycol (0) and alkyl-ether (K), may have influence on all these
mechanisms. The chemical structure of SRA molecules was not provided by the
manufacturers. However, if the SRA molecules are adsorbed on the C-S-H particles, a
high molecular weight of the polyglycol SRA may increase the gap between two
neighboring C-S-H surfaces, which increases the drying shrinkage 15 , but not the chemical
shrinkage. On the contrary, the hydration delaying effect of SRA should decrease the
chemical shrinkage. Analyzing Figure 2, it is clear a higher influence of SRA on
compressive strength at early ages. Comparing this indirect measure of the hydration
level with the autogenous shrinkage shown in Figure 5, which shows a higher effect of
the SRA at later ages, it is suggested that the effect of SRA on the hydration reactions is
not the dominant mechanism of shrinkage reducing provided by SRA. The autogenous
shrinkage depends on the hydration level as well as the viscoelastic behavior of the
mortars. Using the compressive strength as an indicator of elastic modulus and relaxation
capacity, mortars with higher content of SRA are expected to exhibit, for the same
chemical shrinkage, higher autogenous shrinkage at early ages and lower autogenous
shrinkage at later ages. The higher shrinkage at early age is due to the decrease of elastic
modulus. At later ages, the presumed higher relaxation capacity of the mortars with more
SRA (in fact, it is well known the deep increase of creep coefficients when the structures
are in charge at early ages 19) is responsible for lower autogenous shrinkage values. To
better understand the early age results of autogenous shrinkage it should be taken into
account that shrinkage deformations were measured only after 18 hours. Beginning the
measurements at this time, part of the deformations due to chemical shrinkage was not
recorded. Considering the chemical shrinkage between 18 hours and 2 days as charge on
the specimens skeleton, due to the internal restrained decrease of volume, and using the
Fig. 6 shows total shrinkage. Total shrinkage may be considered as a sum of autogenous
and drying shrinkage. However, when measuring autogenous and total shrinkage in
different specimens, drying shrinkage is not the difference between total and autogenous
shrinkage, since the hydration conditions are not the same in free and sealed specimens.
In spite of this, the autogenous part of total shrinkage in unsealed specimens and the
autogenous shrinkage of sealed specimens should exhibit similar tendencies of variation
with increased SRA dosages.
As observed for the sealed specimens, the influence of the SRAs on total shrinkage is
higher at later ages, but with different effects. Additional dosages of SRA 0 seem to be
effective only until about 1.5 %. Comparing with Fig. 5 after 28 days, it can be observed
that for dosages greater than 1.5 % there is an inverse tendency of autogenous shrinkage
in relation to total shrinkage. This indicates a significant increase in drying shrinkage
with higher dosages of SRA 0, the polyglycol product. Fig. 4 indicates only a small
increase of mass loss after 28 days, but in Fig. 6 a significant increase of total shrinkage
is observed, which suggests a worse behavior in drying shrinkage of mortars with high
content of high molecular weight polyglycol (SRA 0), which is consistent with the
hypothesis of an increased gap between C-S-H surfaces.
The increase of the dosage of alkyl-ether product (SRA K) was efficient until 3 %. A
slight increase of the total shrinkage was observed between 3 % and 4 % of this SRA, at
all ages. Fig. 4 indicates a continuous increase of water loss with increasing dosage of
SRA. The autogenous shrinkage of SRA K mortars indicates no systematic variations
with increase of SRA dosage, showing similar values at 181 days. At this age, the
different behavior on total shrinkage can be attributed to drying shrinkage. The
simultaneous effect of decreasing the surface tension and increasing the rate of
evaporation may have opposite global effects on drying shrinkage. Between 1% and 3 %
the first effect seems to be dominant, assuming capillary pressure the mechanism of
shrinkage reducing, while, for higher dosages the increase of the water evaporated
becomes as relevant as the additional decrease of surface tension. However, this simple
analysis does not take into account other physical phenomena that may occur in the pore
structure due to changes in solid and liquid phases. In fact, surface free energy and
disjoining pressure depend on surface area of hydration products, al well as, the
dimension of the pores, and they are not only dependent of surface tension or rate of
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
CONCLUSIONS
Based on the results obtained, with two commercial SRAs, the following conclusions
may be drawn:
• SRA tends to delay hydration reactions and decrease the long term strength;
• SRA is efficient in reducing both autogenous and total shrinkage;
• shrinkage reduction is not proportional to decrease of surface tension;
• an optimum dosage of SRA to decrease total shrinkage that depends on the type of
SRA exists;
• shrinkage reduction seems to depend not only on surface tension reduction but also on
changes of disjoining pressure and surface free energy;
• early and long-term shrinkage reduction due to the use of SRAs seems to depend on
changes of viscoelastic properties at early ages and on molecular weight ofSRA.
REFERENCES
1. Shah, S.P.; Karaguler, M.E.; and Sarigaphuti, M., 1992, "Effects of Shrinkage-
Reducing Admixtures on Restrained Shrinkage Cracking of Concrete", ACJ
Materials Journal, V.89, N° 3, May-June 1992, pp. 289-295.
2. Berke, N.S.; Dallaire, M.P.; Hicks, M.C.; and Kerkar, A., 1997, "New
developments in Shrinkage-Reducing Admixtures"; Superplasticizers and Other
Chemical Admixtures in Concrete: Proceedings of the Fifih International
Conference CANMETIACJ, Rome, Italy, 1997 (SP-173), ed. V.M. Malhotra,
American Concrete Institute, pp 971-978.
4. Ribeiro, A.B.; Goncalves, A.; and Carrajola, A., 2002, "Shrinkage Reducing by
Chemical Admixtures", Structural Concrete 2002: National Meeting, Lisbon,
November 27-29, ed. Grupo Portugues de Betao Estrutural (Portuguese National
5. Gettu, R.; Roncero, J.; Martin, M.; and Agull6, A., 2002 "Experimental Study of
Concretes Incorporating Shrinkage Reducing Admixtures", Report C-409811,
Universitat Politecnica de Catalunya, Barcelona, Spain, March, 15 pp.
7. Adamson, A.W., Gast, A.P., 1997, Physical Chemistry of Surfaces, 6th Edition,
Wiley-Interscience, John Wiley & Sons, Inc., New York, pp 784.
8. Roncero, J., Gettu, R. and Martin M., 2003, "Evaluation of the influence of a
shrinkage reducing admixture on the microstructure and long-term behavior of
concrete", Superplasticizers and Other Chemical Admixtures in Concrete
Supplementary papers of the Seventh CANMETIACI International Conference,
Berlin, Germany, October 2003, pp 207-226.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
9. Bentz, D.; Geiker, M.; and Hansen, K., 2001, 'Shrinkage-reducing admixtures
and early age desiccation in cement pastes and mortars", Cement and Concrete
Research, V. 31, N° 7, July, pp 1075-1085.
11. Weiss, J. and Berke, N., 2003, "Admixtures for reduction of shrinkage and
cracking", RILEM Report 25: Early Age Cracking in Cementitious Systems,
RILEM Publications, pp 323-335.
12. Powers, T.C., 1965, "Mechanism of shrinkage and reversible creep of hardened
cement paste", The Structure of Concrete and Its Behaviour Under Load:
Proceedings of an International Symposium, London, Septemberi965, Cement
and Concrete Association, pp. 319-344.
14. Hansen, W., 1987, "Drying shrinkage mechanisms in portland cement paste",
Journal ofthe American Ceramic Society, V. 70, N° 5, pp. 323-328
17. Bisschop, J., 2003, "Drying shrinkage cracking at early ages", RJLEM Report
25: Early Age Cracking in Cementitious Systems, RILEM Publications, pp. 47-
55
18. Srinath, M., 2002, Wetting on flexible substrates: formulation, PhD thesis,
Drexel University, Philadelphia, 318 pp.
19. Lange, D.A. and Altoubat, S.A., 2003, "Early Creep," RJLEM Report 25: Early
Age Cracking in Cementitious Systems, RILEM Publications, pp. 57-62.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
7d. 41.1 41.2 36.4 36.0 30.6 36.8 33.0 34.9 32.2
28d. 50.3 49.0 49.6 46.8 39.5 47.6 47.7 45.1 42.5
60 d. 56.6 56.0 50.3 46.8 44.1 51.2 51.9 48.4 45.7
90d. 58.4 56.8 51.8 51.0 47.4 51.5 53.8 49.6 47.2
i'Ett:s§i2
50 .
0.0% 1.0% 2.0%
.
3.0%
.
4.0%
~
i
90
80
gc 70
11
t=-";~~~:::::::t:=-----1
-t-------~~~
60-t------------~
ro+---,----.--.--~
7days 28days ~
e- 1oo ·~--.::::=~----------, e _1oo
;: ~ 90 r==_-;;:.~.~~~=i il ~ 90
.....= 8070 -+-------·~-==""='--j
~ ~ i 80 +---- ----1
~
~!
c t------------j
60 - t - - - - - - - - - - - - j
ti..
~
70
60+----
~ ro+------,--------l ro+---.---.--.---~
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
60 days I• K • o I 90days ~
_1oo l='=1t::!S~:::;~~~
e~~90 e _1oo
~ il:;. 90
:. =
> c
eo +---------------_,
"' 70 +----------------1
~ ~ 80 - ------·-
i! 60
~
---------
50 +--.,----.,----.-----1
>i :~ -----·--
0.0% 1.0% 2.0% 3.0% 4.0% 0.0% 1.0% 2.0% 3.0% 4.0%
SRAdosage _ _ _ _ _SRA dosag_e_ _ _
& 40+---~---.---.--~
0,0% 1,0% 2,0"/b 3,0%
SRAdosage
4,0%
(."' "' '·"" '·"' '·"" '·"' SRAdosage
~Esc+---------------~ =: j BO +-------------------1
.; .. 50 +---------------~ i .. 50 +---------------1
a: 40 +---~---.---.--~ a: ~+----~---.--~------1
0,0% 1,0% 2,0% 3,0% 4,0% 0,0% 1,0% 2,0% 3,0% 4,0%
SRAdosage SRAdosage
60days ~ 90 days I• K • o I
,; 100 .. ,...__ . • .; 100 .,_,.._=,..,..._~----.---------,
~""""'.,._,~~------j
.. - 90 +------II eieo+---~---~~~~=-----1
~ ~ BO +-------~--~ ..,·_··..,•.--·_::_··:._,·~-.,.. "-- BO +----·--___-.:..:...:..:_·~·,--,;.:!l~·c.:;.·-~·="·•._.
~,70+------------------~~ Ei 70 +---------------.;---1
..8 !c so +-------------------1
i~so++==============j
&"'50
40+---~---.---.---4
Jl! .. 50 +-----------------~
~ 40+--~--~--.--~
0,0% 1,0% 2,0% 3,0% 4,0% 0,0% 1,0% 2,0% 3,0% 4,0%
SRAdosage SRAdosage
80 -,-------------------, so~--------------·
~70 fro
sao r:;:Kl S.so 1-iH<l
~
c
~
~~
140 t-- --- --------
------------------ !so -
Jl 40
La
"
Ul
~~
0
3
-----------------
" 20+-~--.--~--.-~
~+--~--~-----.--~
0% 20% 40% 80% 80% 100% 0% 2% 4% 6% 8% 10%
SRA content ('lQ SRA content('IQ
(a) (b)
Figure 3-- Surface tension of solutions with distilled water and SRAs.
6,0 - 6,0
~5,0 ~
., 5,0
7,0
6,0
li=.
5,0
4,0
3,0
:;; 2,0
1,0
0,0 0,0 -1---~--~-~--J
0,0% 1,0% 2,0% 3,00,(, 4,0% 0,0% 1,0% 2,0% 3,0% 4,0%
SRA dosage (%) SRA dosage(%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~ 5,0 ~ 5,0 -- --- ------ --- --
o.o
0,0% 1,0% 2,0%
·'
3,0% 4,0%
I
II! 2,0
1,0
- - - -- - - - - - - -- - - - - - - - -
---------------------
0,0 -+--~---.--~--1
0,0% 1,0% 2,0% 3,0% 4,0%
SRA dosage (%) SRA dosage(%)
Figure 4 --Mass loss of the prisms used for total shrinkage measurements.
~ ::: ~ - - - - - - - - - - - - - - - - - - -I o,5o 1 j
!
~ i :·::s-~~~~~~~~~~~~~~~~~~
ji" E. o,2o
0" 030+ ---------------
1i 0:201~~~~--------------- '
----- -----------
~~ 0,10t::=s?:::----t ~~ 0,10t---~
0,00 0,00
0,0% 1,0% 2,0% 3,0% 4,0% 0,0% 1,0% 2,0% 3,0% 4,0%
SRA dosage (%) SRA dosage (%)
....._K...,._O
181 days
0,50 , - - - - - - - - - - - - - - - - ,
.;- 0,40
!~
g .e 0,30
j} 0,20
~.e
i 0,10
0,00 .J-----~--~--- 0,00 +----~---.------,----1
0,0% 1,0% 2,0% 3,0% 4,0% 0,0% 1,0% 2,0% 3,0% 4,0%
SRA dosage (%) SRA doaage (%)
.E 0.60 '~
! 0.60 --------------------
t
·E .ro.4o ~ ~ 0.40 --- -
28days GO days
0.80...-------------,
0.00 -l--~---,----.----i
0.00 +--~--~--...----1
0.0% 1.0% 2.0% 3.0% 4.0% 0.0% 1.0% 2.0% 3.0% 4.0
SRA doaaga (%) SRA dosage (%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Irene Schober received her Ph.D. in Chemistry from the Technical University of Vienna.
She is head of the Polymer Synthesis Department of Corporate Research at Sika
Technology AG in Zi.irich, Switzerland.
INTRODUCTION
The rheology of cementious materials is profoundly affected by dispersants as in
any agglomerating system. In the case of cementitious, materials, one must worry not
only about interparticle forces but also about the impact that dispersants have on cement
hydration, which may not only affect the strength development but also rheological
properties in the fresh state.
In cementitious materials, dispersants are referred to as superplasticizers. The
latest type of these are comb-type copolymers with an poly-carboxylic backbone and side
chains of polyalkylene oxide. The ionic groups on the polymer backbone allow the
polymer to adsorb on the surface of fine particles that include cement, fly-ash, filler and
silica fume (if enough positive ions can assist the adsorption), while the polyethylene
oxide side chains induce steric repulsion that causes dispersion of these particles.
This type of polymer offers the possibility to perform a wide range of changes in
the molecular structure of the polymer. Also, producers of admixtures for cementitious
materials seek to understand the impact of varying molecular structure on the rheology of
cementitious materials. In doing so, they are confronted to systems in which not only the
degree of dispersion but also hydration can impact a given macroscopic property.
While trying to master these effects, industrial researchers largely opt for simple
and quick tests that offer an indication of the flow properties of the materials. In the case
of admixtures, a mini slump test consisting of a cylindrical mold of 50mm in diameter
and 50mm in height can be used to assess flow (see for example Kinshita et al2000). The
mold is filled with the suspension and then lifted, and the diameter of the spread sample
is measured. Surprising relations are reported in a separate paper (Schober and Flatt
2006) showing a linear relation between this diameter and polymer adsorption (Figure 2).
In this paper the possible origin of this unexpected result is examined in the perspective
of mechanistic characterization of the performance of superplasticizers.
T =4 pVg (1)
0 .J3 nD2
where pis the density of the sample, Vis its volume, g is the gravity constant and Dis the
spread diameter.
The result in the above equation differs only slightly from what was initially
proposed by Murata (1984), but Pashias and coauthors probably came up with the most
eloquent description of this test "the 5c rheometer"(Pashias et al. 1996).
At large spreads, the same authors indicate that the link between spread diameter
and yield stress is given by:
2
(2)
T
0
= 55.904pV 2 D- 5 - AD
4V
where Jc is a constant linked to the liquid vapor interfacial energy and the wetting angle
on the plate.
Thus, the spread diameter can be related to the yield stress with an exponential
function:
T0 = a£exp(-bD) (3)
Po
where a and b are fitting parameters given in Figure I b and which depend neither on the
polymer nature nor on the volume fraction of the suspension. It must be emphasized that
those coefficients are for a cylinder of 50mm diameter and 50mm high. The reference
density Po is 2.25 g/cm3 for the value of a given in the Figure 1.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
DISCUSSION
The relationships presented between yield stress and spread are rather
unexpected and are now examined in further detail. It should be emphasized that the
dominant effect of superplasticizers on fluidity of cementitious systems is attributed to
changes of interparticle forces (Yoshika et al. 1997, Lewis et al. 2000, Flatt 2001, Flatt et
al 2004). Thus the existing knowledge about yield stress from other suspensions than
cement can be used.
An important reference in this respect is found in a review of work of Boger's
group (Zhou et al. 2001 ). Indeed, their results show that yield stress can be written as a
product between interparticle force and a function of the volume fraction of solids <j>. This
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
function can be written to account for the number of interparticle contacts as well as a
divergence as maximum packing is approached (Flatt and Bowen 2005):
(4)
where Fmax denotes the maximum attractive force among particles, a is a numerical
constant and./(¢) is a function that accounts for the volume fraction dependence of yield
stress.
Such expressions were developed for homogeneous systems. However, in the
case of cement suspensions, not only are the particles composed of various phases, but in
most cases of practical interest, their surfaces are not fully covered by the added
dispersant. Nevertheless, it is assumed that equation (4) remains valid provided an
adequate average force is used.
Combining Eqs. (3) and (4), the spread diameter can be written as:
(5)
which, for a given volume fraction, can be written in a more convenient way by
introducing a constant, {J, that only depends on the volume fraction of solids <j>:
1 (6)
D = P--Ln(F
b max
)
where {}is the fraction of the total surface covered by the polymer indices, B and P refer
respectively to portions of the surface that are either bear or contain adsorbed polymer
(ex: Fmax,B-B is the maximum attractive force among two particle that are in contact with
protions of their surfaces that do not contain adsorbed polymer).
If the dispersant is powerful enough to cancel attractive forces between any two
surfaces of which at least one is covered with the said polymer, then:
~ (1- e)
2
Fmax FmaxB-B (8)
The authors emphasize that the surface coverage should be expressed in terms of
the saturation plateau of an adsorption isotherm, but with respect to real complete
coverage. Assuming that under the later condition all these polymers would occupy the
same surface area per unit mass of polymer, then Bcould be expressed as:
(9)
2 {10)
D = {3 2 - y;Ln(moo- m)
An alternative approach is to assume that the average force between the surface
is estimated from a separation distance, 8, that is assumed to be proportional to surface
coverage. In this case Fmax can be expressed as (Flatt and Bowen 2005):
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
F
max
::::::5-2 :::::: g-2 (11)
2 (12)
D= /33 +-Ln(m)
b
The above equation does not require any assumption with respect to the value of
the saturation dosage. It turns out that a logarithmic fit to the data gives a fit which is just
as good as the linear relation previously proposed, as indicated in Figure 3.
(13)
2 (14)
D= /] 3 + -Ln(m- m 0 )
b
when the value ofm 0 is set to 0.84 which is estimated from the fitted curves in Figure 4, a
much better correlation could be found as indicated in Figure 5.
In this case, the agreement is much better between the experimental data and the
predicted behavior. However, the assumption of excluding the irreversibly adsorbed
amount as well as its estimation would need further validation. A number of alternative
solutions could also be considered to account for the same initial discrepancy between the
measured and predicted values. This is, however, beyond the scope, of this paper, but will
be reported elsewhere at a later date. Thus, at this stage this result should only be taken as
an indication that the general approach to analyzing simple flow experiments in terms of
interparticle forces may provide much more insightful information than expected up to
now.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
REFERENCES
Flatt R.J., 2001, "Dispersants in concrete", Chapter 9 in Polymers in Particulate Systems:
Properties and Applications, Ed. Hackley V.A., Somasundran P. and Lewis J.A.,
Surfactant Science Series. Marcel Dekker Inc., pp. 247-294.
Flattt R.J. and Houst Y.F., 2001, "A simplified view on chemical effects perturbing the
action ofsuperplasticizers", Cern. Caner. Res., V. 31, pp. 1169-1176.
Flatt R.J., Martys M. and Bergstrom L. , 2004, "The rheology of cementitous materials",
Mat Res Bull. V. 29, No.5, pp. 314-318.
Flatt R.J., Larosa D. and Roussel N., 2006, "Linking yield stress measurements: spread
test versus Viskomat", Cern Caner Res, V. 36, pp. 99-109.
Flatt R.J. and Bowen P., 2006, "YODEL: a Yield Stress Model for Suspension", J. Am.
Ceram. Soc, V. 89, No.4, pp. 1244--1256.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Kinoshita M., Nawa T., Iida M. and Ichiboji H., "Effect of Chemical Structure on
Fluidizing Mechanism of Concrete Superplasticizer Containing Polyethylene Oxide Graft
Chains", 6th CANMET/ACI International Conference on Superplasticizers and Other
Chemical Admixtures in Concrete, Nice, France, 2000, ACI SP-195, V. M. Malhotra
(Ed.),pp 163-179.
Lewis J.A., Matsuyama H., Kirby G., Morissette S. and Young J.F., 2000,
"Polyelectrolyte effects on the rheological properties of concentrated cement
suspensions", JAm Ceram Soc 83, pp. 1905-13.
Pashias N., Boger D.V., Summers J., and Glenister D.J., 1996, "A fifty cent rheometer
for yield stress measurement". J. Rheol., V. 40, pp. 1179-1189.
Roussel N. and Coussot P., 2005, "Fifty-cent rheometer" for yield stress measurements:
from slump to spreading flow" J. Rhea/., V. 49, No. 3, pp. 705-718.
Roussel N., Stefani C. and Leroy R., 2005, "From mini cone test to Abrams cone test :
measurement of cement based materials yield stress using slump tests", Cern. Caner.
Res., V. 35, pp. 817-822.
Schober I. and Flatt R.J., 2006, "Optimizing Polycarboxylate Polymers", in Proc. 8th
CANMETIACI Conf on Superplasticizers and Other Chemical Admixtures, Sorrento, Ed.
Malhotra M., in press.
Yoshioka K., Sakai E. and Daimon M., 1997, "Role of steric hindrance in the
performance ofsuperplasticizers in concrete",J. Am. Ceram. Soc., pp. 2667-2671.
Zhou Z., Scales P.J. and Boger D.V., 2001, "Chemical and physical control of the
rhoelogy of concentrated metal oxide suspensions", Chern Eng Sci, 56, pp. 2901-2920.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
b) o Combined data
7 • Data for regression 25
" relafive e1TOr
6 ....
15 2':.
'iii' 5 ""
!!,. " ~
" ~
"'"' 4
" 5
e-
~ 3 ..!l
!I " .5:
;E_2 """1:. " -5 E
_g
z 1
...... 1:.
A
6 -15 g
0- w
-1 -25
5 10 15 20 25
Spread diameter [em]
Figure 1. Interpolation between models in Eqs. (1) and (2), a) predictions of both
models, b) minimum yield predictions of both models as well as data selected for
the interpolation are shown as open and filled squares respectively. Open triangles
indicate the error from the interpolation.
210+---------------------
E 190 +-----------------,~07""~
.§.. 170 +-------------~~
i :E FE
~ 150 +------------:7"""'-----
----------1
230~-------------------------.
.§.. 170
~ 150
ii: 130
j
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
110
.... 90
70
50 -+-----..----r-----r-----1
1.0 1.2 1.4 1.6 1.8
Adsorption (mg polymer/g cement)
Figure 3- Data from Figure 2 fitted with a logarithmic function (continuous line) and a
linear regression (discontinuous line).
2.5
"".Ec
..
u
2
c."'
~
-.
.§. 1.5
E
0
1
-e0..
0.
"0
"0
. 0.5
< 0
0 2 4 6 8
Polymer in solution [mg/g cement]
230~------------------------.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
210
'E
.§. 170
19o
--- • ---------#
~ 150
Lt 130
-! 110
:. 90
70
50 +-----~~-----r------~----~
1.0 1.2 1.4 1.6 1.8
Adsorption (mg polymer/g cement)
Figure 5· Data fitted the same amount of irreversibly adsorbed polymers for which
surface coverage is corrected.
Stefan Dikty is a research chemist of Kao Chemicals GmbH, Germany. He has worked
in the field of focussing on the development of new superplasticizer of polycarboxylate-
based polymer.
INTRODUCTION
Two different types of portland cement (CEM I 52.5R and CEM I 42.5R), that
had different strength properties, were used to establish the performance of
The chemical structure of NPD is shown in Fig. 1. This polymer has a bone
structure of cellulose principal chains with two functional groups of opposing natures,
ionic functional groups (X) and hydrophobic groups (R).
Supernatant liquid separated centrifugally from cement paste (w/c = 0.4) was
used to establish the behaviours of the thickeners. Thickener aqueous solutions were
prepared to dissolve with the required dilution of the above supernatant. The viscosity
was measured with a Type B viscometer (at 30 rpm).
Mortar test
Cement, fine aggregate and water containing PCE and NPD were placed in a
mortar mixer at a w/c of0.5. Mixing was carried out for 2 min at 62 rpm. The fluidity of
mortar was evaluated by the spread measured when a truncate cone ( cf> 100 mm of lower,
cf> 70 mm of upper, 60 mm of height) filled with mortar was pulled up vertically. The
bleeding test was performed according to JIS A-1123 that defined the measurement of
bleeding amount. The exothermic curve of mortar at 25 oc was obtained using TAM Air
calorimeter manufactured by Thermometric. The hydration time was calculated from the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
second exothermic peak of cement hydration, which resulted mainly from C 3S hydration.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
RESULTS AND DISCUSSION
The liquid phase in concrete contains a large number of inorganic metallic salts
dissolved from cement, including bivalent ions such as Ca2+ and a small amount of
trivalent such as Al 3+, additionally such monovalent as Na+ and K+. As pH is strongly
alkaline, the conditions are quite severe for thickeners. Fig. 2 shows the results of
viscosity of aqueous solution. Various thickeners were dissolved under three different
types of liquids; one was a cement supernatant liquid (pH 12.9), another was supernatant
diluted 7 times (pH 12.2), and the third was ion-exchanged water (pH 7 .0). The dosage of
each thickener were determined as the viscosity of solution dissolved with ion-exchanged
water adjusted to around l 000 mPa ·sec. The dosage of each thickener was described
below the name in Fig. 2.
The viscosity of both HEC and MC were gradually reduced with increase of ion
concentration. In the case of HEC, the molecular structure expanded with hydration in
The rheology curves of both ascending and descending slopes are depicted in
Fig.3. Each sample was dissolved with cement supernatant liquid. From a rheology curve
point of view, the difference between ascending and descending slopes suggests
thixotropy of solution. With respect to the rheology curve of NPD, a bigger difference
was observed between both slopes. The hysteresis area enclosed with both slopes
of NPD is 174.5 Palsec and that of HEC is 28.0 Palsec. This indicates that NPD forms
twisted structures as networks under high ionic strength.
Fig. 4 shows the influence of metallic ion concentration on the hysteresis area of
NPD. As cement supernatant liquid is diluted, the hysteresis area ofNPD tends to reduce
and show a minimum at ion-exchanged water. Since the value of hysteresis area achieves
to zero, the twist ofNPD remains partly even if under ion-exchanged water.
From the experimental results, the behaviour of NPD in aqueous solution is
established as follows:
Under ion-exchanged water, NPD is supposed to be present with a small amount
of polymer twist due to the strong electrostatic repulsion of ionic function group in it as
shown in Fig. 5. As metallic ion concentration increased, electrostatic repulsion caused
by ionic function group would be weakened gradually. Therefore, each NPD polymer
shrank due to loss of internal stretch. Furthermore, the hydrophobic function groups
approached one another and accelerated to form twists and to cohere in aqueous solution.
This cohesion occurred not only in a polymer, but also among ones like cross linkages.
Finally, large network structures were formed in aqueous solution as shown in Fig. 6.
On the other hand, ordinary cellulose derivatives were also affected by metallic
ion and shrank in solution. Since they had no strong hydrophobic groups in themselves,
their molecules were held shrunk. This is why they showed lower viscosity under high
ionic strength condition.
To summarize the above results, conventional thickeners cannot efficiently
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
realize their inherent viscous properties in the presence of salts, so these are added in
excessive amounts for concrete application. However, since NPD displays greater
viscosity under cement supernatant liquid, it can be used at a lower dosage. Moreover, it
may be possible to develop a so-called two-in-one product that possesses both
dispersability and viscous ability, because NPD can be mixed with a superplasticizer
easily without increasing solution viscosity.
The first application test was to study the possibility ofNPD as an anti-bleeding
agent. The mortar tests were performed under two different cement samples at a w/c of
0.5. The results are shown in Table 1 and 2. As expected, NPD indicated good anti-
bleeding performance at a low dosage. Further, low retardation of cement hydration is
also observed in both conditions. This effect appears to depend on the high viscosity of
liquid phase in mortar.
Concrete tests for SCC
The second application was to investigate the use of NPD in SCC. The mixture
proportions and test result are shown in Table 3 and 4, respectively. No segregation was
found at 600 ± 30 mm over 90 min. In comparison with HEC, similar viscosity was
observed in the concrete with NPD at half the dosage of HEC. Furthermore, the setting
time was 3 hours earlier than that of HEC. This suggests that NPD minimizes the
retardation of cement hydration due to a lower dosage.
Finally, the application of SCC in the wall of a LNG storage tank was undertaken
[2]. The mixture proportions and test result are shown in Table 5 and 6, respectively.
Belite-rich cement was used to reduce the heat amount of cement hydration. A liquid-
type admixture of NPD blended with PCE was prepared for this project. The self-
compactability required in this test was Rank 1 [7]. This R number indicates the level
of self-compactability. Required value and test condition are dependent on the R number.
The filling height has to be more than 300 mm when Rl (obstacle) is used.
Satisfactory flowability of concrete was obtained for 90 minutes. Good
segregation resistance was also observed due to the effect of NPD. Therefore, the
quantity of powder in the proportions could be reduced. Limestone powder having a
Blaine of 750 m2/kg was used for this project. If the quantity of limestone powder were
reduced without sacrificing the performance of fresh concrete, total concrete cost would
be reduced.
Even when the limestone powder content was reduced to 81 kg/m 3 , as shown in
mixture proportions No.2, good SCC was achieved. Furthermore, good self-
compactability was achieved, when the limestone powder was replaced by cement. The
PCE mixed with NPD indicates that it is not necessary to weigh the thickener separately.
CONCLUSIONS
1) NPD forms large network structures due to the cohesion among hydrophobic groups in
the presence of inorganic metallic ions, and as a result, exhibits high viscosity in
concrete.
2) NPD provides good anti-bleeding performance at low dosage. Furthermore, NPD can
be used to produce good self-compactability and to minimize retardation of cement
hydration due to its relatively low dosage.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
ACKNOWLEDGEMENT
REFERENCES
2. Yamamuro H., Hamada D., and Shounaka M., "Development of a liquid type
admixture for self-compacting concrete", 2"d International Symposium on Self-
Compacting Concrete, 2001, pp. 179-186.
3. Yamamuro H., Izumi T., and Mizunuma T., "Study of non-adsorptive viscosity
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
agents applied to self-compacting concrete", 5th CANMET/ACI International
Conference on Superplasticizer and Other chemical Admixtures in Concrete, 1997,
pp. 427-444.
5. Hamada D., Sato T., Yamato F., Mizunuma T., " Development of new
superplasticizer and its application to self-compacting concrete", 6th CANMET/ACI
International Conference on Superplasticizer and Other chemical Admixtures in
Concrete, 2000, pp. 269-290.
7. Guidelines for High Plowable Concrete, Concrete Library, JSCE, 1998, No.93, July,
pp. 10-11.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Compressive
Dosage Time after preparation (min.)
No. Test Strength (MPa)
(%)*
5 30 60 90 cr3 cr7
SF 665 680 690 660
LTA** FS 7.5 7.6 8.7 10.3
1 16.2 29.2
1.4 Air 3.3 --- --- ---
U-test 310 --- 310 ---
SF 670 690 680 670
LTA** FS 6.4 7.4 7.5 9.7
2 14.2 24.2
1.4 Air 4.3 --- --- ---
U-test 312 --- 309 ---
SF 670 675 685 670
LTA** FS 7.2 7.6 9.5 10
3
1.4 --- --- --- 29.8
Air 3.6 ---
U-test 310 308--- ---
*: app. dosage (CX %) **: LTA (hqUid type admixture)
SF: slump flow (mm), FS: flowing time to SF 500 mm (sec)
Air: concrete air(%), U-test: filling height (mm)
0
y X Ionic functional group
~ OH OH
0 ; '
\
OH ?
0
o~-R
OH Hydrophobic group
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
NPD MC HEC XG
(1.3%) (1.4%) (1.2%) (1.0%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
,-.. 30
~
~ 20
j
<Zl 10
50 100
Shear rate (~ec· 1 )
Cement supernatant
diluted 2 times
Cement sup,emata:nt
diluted 7 times
IY Li
Fig. 5- Status of NPD under low metallic ion concentration
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: In this paper, the influence of high volumes of ultra-fine fly ash, raw fly
ash, silica fume and natural zeolites on the properties of self-compacting concrete
is evaluated. Three different samples of ultra-fine fly ash obtained after high energy
milling of raw material were employed. Concretes were prepared employing various
kinds of mineral additions as partial replacement (40%) of cement or in addition
to it. Fresh concretes prepared were submitted to technological characterization
through the determination of normal slump and J-ring slump flow values. The values
of the normal slump flow were found to range between 604 and 785 mm, while the
differences with the ]-ring slump flow were less than 30 mm.
Specimens were tested to evaluate the mechanical properties of the hardened
concrete after 7- and 28-day curing. Specifically, the modulus of elasticity and
compressive strength were determined. Significant strength increases were observed
for the concrete containing ultra-fine fly ash. Finally, no segregation phenomena
were observed in the case of cylindrical column specimens (30 x 150 em). All the
results obtained show that environmentally sustainable, high workability concrete
could be successfully prepared using large volumes of mineral additions.
industrial solid wastes and sediments and the preparation concrete made with coal fly
ashes, MSWI ashes, etc. All the above researches are proved by more than one hundred-
twenty papers.
industrial waste in concrete and in geo-environmental applications. He has been and still
is member of several research projects with public and private societies. All the above
Engineering Faculty of the Federico II University of Naples, Italy. Research interests are
in the broad field of microporous and mesoporous materials with particular emphasis in
applications connected with their extensive surface. Research activity is proved by more
Materials and Production Engineering, Federico II University of Naples, Italy. Her main
research interest is zeolite synthesis and natural zeolite characterization and application in
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
to try to select them, in order to relate their properties to concrete characteristics, such as
workability and segregation. It is usually reported that, if the volume of solids is held
performance, such as durability or strength, but increases the water demand to attain
Furthermore, in some cases it is also reported that the use of fine mineral
addition can reduce water demand or increase the slump . Indeed, it has been found that,
when a very fine blast furnace slag is added, an optimum amount of mineral admixture
reduces the water demand for a specific flow (3). Furthermore, enhanced workability is
achieved when spherical shaped fine mineral admixtures are used, especially fly ash (4-
6).
granulated slag, the workability was improved by replacing part of the binder with ultra-
fine materials (7). In other cases, some similar materials were not effective in improving
workability (8,9).
fluidity on the composition, shape and average particle size of mineral additions must be
(SCCs) not only flow under their own weight but also fill the entire form and achieve
uniform consolidation without segregation. sec is used in structures with closely spaced
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
It is important to outline that the properties of freshly mixed SCC are highly
dependent on the mixture proportioning. This is an important factor as the grain size
admixtures. Furthermore, since chemical admixtures are expensive, the use of mineral
additions is also important as they could increase the. flow behaviour of concrete but not
its cost.
powder, fly ash, silica fume, blast furnace slag, etc. Moreover, the additions can lower the
price of concrete when they are industrial wastes and have a considerable influence on
In this paper, SCCs have been prepared in order to evaluate the influence of high
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
volumes of ultra-fine fly ash, raw fly ash and natural zeolites on the normal slump and I-
ring slump flows of fresh concrete as well as the mechanical (compressive strength and
elastic modulus) properties of hardened concrete. Investigations have been carried out
employing amounts of each addition calculated as 40% by weight of cement content. This
amount was used as a partial replacement of either cement or natural aggregate in order
Furthermore, a combined blend of silica fume and raw fly ash was also used at a
Materials
additions, such as fly ash (FA), silica fume (SF) and natural zeolites (NZ), were also used
in order to improve and maintain the cohesion and segregation resistance. The fly ash
employed came from a power plant located in Brindisi (Italy). Furthermore, in order to
investigate the influence of shape and particle size of ash on concrete properties, this raw
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
material was subjected to wet high energy milling, carried out by means of Netzsch
Labstar equipment This is a grinding unit that allows exact scale-up to comparative
production mills.
During the experiments, the rotating speed, milling time, water-to-solid ratio,
and size of milling media were optimized in order to obtain powders with different
fineness and to reduce treatment costs. At the end of this phase, the operative parameters
were fixed as follows: rotating speed: 3500 rpm; milling time: 30 min; water to solid
ratio: 1.2; and size of milling media: 0.8 to 1.0 mm. The final product of the wet milling
treatment was a slurry with a 55% water content, which was taken into account during the
Three different size fractions of milled fly ash were produced. The samples
obtained as described above and marked FAI, FAil and FAIII presented 90% or more of
particles finer than 10, 20 and 30 ~m, respectively. The particle size distributions of raw
and milled fly ash are shown in Fig l.Two natural zeolites (NZI and NZII) were
employed. These were commercial products. The two materials selected were a
The silica fume used was a commercially available material. In particular, this
special addition was employed in order to enable a comparison with the ultra-fine fly ash
terms of main oxides, is reported in Table 2. The superplasticizer (SP) used was an
acrylic base with 40% solid content and a specific gravity of 1.2 kg/dm 3 . It was used in
all SCC mixtures, and its amount was kept constant for each type of concrete. Even the
water content of the superplasticizer was considered during the mixture proportioning.
An inorganic viscosity modifying agent (VMA) was also used in mixtures to enhance
stability (I 1).
Mixture proportions
Two different types of SCC mixtures were proportioned and cast. Type A
concrete was prepared using each mineral addition as a partial replacement of the fine
additions were employed as a partial substitution of cement. The amount of each addition
was calculated as 40% of cement content. The silica fume was blended with raw fly ash
in amounts of 10% and 30% of cement content, respectively. Considering the well known
characteristics of silica fume, such as the large specific surface, the high content of free
silica and its high cost, its content generally did not exceed 10-15% of the binders. A
control concrete not containing mineral addition was also made and marked as C
concrete.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
the first part of the mixture code gives information on the type of concrete while the
second part indicates the mineral additions contained. The control concrete was prepared
with a cement content of 365 kg/m 3 . For each type of mixture a similar aggregate-to-
cement ratio was employed and the filler content of aggregate was taken in to account
during the mixture proportioning phase. The water-to-cement ratio was kept constant at
0.50 for all the concretes. The mix proportions were chosen to satisfY all performance
criteria for the concrete in both the fresh and hardened state.
Methodology used
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In order to manufacture concrete, a 90 dm3 capacity conventional concrete
mixer was used. After mixing, fresh concrete was used to determine normal and J-ring
slump flow according UNI 11041 and UNI 11045 Italian Standards, respectively. The
slump flow test measures the mean diameter of concrete spread after the removal of the
slump cone. A spread of at least 600 mm is required for SCC. Slump flow gives
information about the flowability of a fresh mixture in unconfined conditions. All the
mixtures prepared did not reveal any segregation phenomena, as shown in Fig. 2 (left).
The J-Ring slump flow equipment consists of a ring of reinforcing bars that fit
around the base of a standard slump cone. After the slump cone is lifted, the concrete
flows under its dead weight through the steel bars which simulate the flow through
reinforcement in a real formwork. J-Ring flow values give information about the
capacity of the fresh mixture to flow through confined spaces and narrow openings
(passing ability), as shown in Fig. 2 (right). The differences between the slump flow
and 28 days of water curing, according to UNI 6132 Italian Standard. Three specimens
(1 00-mm cube) were cast for each data point. The dynamic modulus of elasticity was
This method, according to Rilem NDT 1 Standard, links the velocity with which an
ultrasonic pulse passes through a material, with its elastic modulus. The specimens were
The ultrasonic pulse velocity measurement was also carried out to verify the
shown in Fig. 3, were cylindrical columns of 30-cm diameter x 150-cm height. Sixteen
measurements of velocity were taken for each specimen at different heights and the
Table 4 shows the results of normal and J-ring slump flow measurements
carried out on type A concrete. The values reported were separately obtained
immediately after, and 30 minutes after the mixing of the concrete. The differences
between the two above values recorded, expressed in terms of percentage of slump loss,
are also shown. It is seen that the systems containing raw fly ash and silica fume show
slump flow values higher than those of the other mixtures. These results are clearly
affected by the particle shape and fineness of the additions and, in actual fact, the
spherical shape of samples is one of the main factors that improves the workability (4).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
a lower fineness.
A relevant consideration can be drawn once the results for the systems FAI,
FAIT and FAIII are taken in to account. It is seen that the irregular (sharp) particle shape
of ground material and their high fineness do not significantly affect the flow behaviour
of concrete. Indeed, the slump flow values are fairly high, and equal to 722, 735 and 753
mm, respectively for the systems under consideration. These high values suggest that, in
concrete (about 600 mm of slump flow) reducing the amount of superplasticizer added.
Moreover, the role played by the two kinds of zeolites on the slump flow of mixtures is
clearly affected by their different fineness and the behaviour shown by the coarser sample
is worse than that of the finer one. This is due to the presence in sample NZII of a
Finally, analysing the results of the C mixtures makes it possible to draw the
conclusion that when the concrete is prepared without mineral addition the slump flow
value is lower than those of zeolites containing ones. As regards the slump flow
measurements carried out 30 minutes after mixing, it is clearly seen that the values of
slump loss found for all the mixtures tested in this study lay within the 5.4- 8.7% range.
This technological aspect is of great importance in the use of self compacting concrete at
a considerable distance from the production plant. The values observed for SF mixture
are higher, due to the chemical and physical characteristics of silica fume ( 16). Regarding
the J-ring slump flow values measured for the same type of mixtures described above, it
is seen that the differences between normal slump and J -ring slump values are between
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
the self-compacting concrete prepared and studied in this work have a good passing
ability in terms of capacity to flow through confined spaces. Moreover, the data also give
information on the fact that no blocking phenomena are shown when the tests are carried
Considering the important technological and economic factors such as the VMA
content, it is possible to say that, as reported in Table 3, all the mineral additions tested
are able to favour a reduction in comparison to the amounts required in the control
mixture. Table 5 shows the results of normal slump and J-ring slump flow test carried out
on type R concrete. It is possible to see that all the values measured are lower than those
found for the type A concrete, the reason being that the mixture proportioning is different
for the type R concrete, as shown in Table 3. The data show that the higher slump flow
values are found for the mixes containing silica fume while the lowest values are
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
measured for the zeolite-containing mixtures. Furthermore, in the cases of the concrete
employing the raw and ultra fine fly ash, the slump flow values range between 671 and
635 mm. The latter value (F AI system) is lower than that found for the C mixture.
Finally, the values found carrying out the J-ring slump flow test confirm that all
the mixtures tested show a good behaviour in respect to their capacity to flow through
narrow openings (steel bars). As in the case of the type A systems, the differences
between normal and J-ring flow values are less than 30 mm and the workability loss after
30 minutes is low.
measurements carried out on concrete cylindrical columns show that almost all the
specimens manufactured with mineral additions are homogenous along their height (see
Fig. 3). The only exceptions were found for the C and NZII mixtures, in which ultrasonic
pulse velocity variation (5% and 8%, respectively) were measured. Therefore, the clearly
observable effect due to the use of large amounts of mineral additions is the absence of
The results of these real scale tests are of some interest when the amount of
development for type A and R concrete are shown in Figs. 4 and 5. Figure 4 indicates
that, after 28-day curing, all the mixtures give a strength higher than that of the control
concrete. In particular, the mixture containing silica fume shows a much greater
mechanical strength than the others (16). Moreover, for the mixtures containing fly ash,
the strength increases as the grain size become finer and strengths of 59.0, 59.4, 62.5 and
64.7 MPa have been found for the systems FA, FAIII, FAil and FAI, respectively. The
mixture NZI gives a strength (58.5 MPa) which is very close to that found in the case of
raw fly ash (59.0 MPa), while the mechanical behaviour of the NZII mixture is similar to
As far as the type R systems are concerned, Fig. 5 shows that, after 28-day
curing, all the values are, as expected, lower than those for type A systems and, even for
this type of concrete, the higher strengths were found when the fineness of the additions
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
the only one that shows a higher value than the control mixture. In all the other cases, the
lowest value of strength loss is 5% for the F AI mixture, while the highest is 25% for the
mix containing coarser zeolite (NZII). Analysing all the above results, it can be said that
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
the development of strength, up to 28-day curing, is not governed by chemistry but by the
The dynamic modulus of elasticity values are reported in Fig.6. After 28-day
curing type A concretes show a higher modulus value than concrete C, whereas the
values recorded for the type R mixtures are lower. These experiments confirm that, with
the exception of SF systems, the higher values of modulus are found for the finest
CONCLUSIONS
All the experiments carried out m this investigation allow the following
conclusions to be drawn.
prepared employing considerable amounts of mineral additions. The first type, containing
the additions as a partial replacement of the finer fraction of aggregate, showed slump
flow values higher than those for the other type in which the cement was replaced. For all
the mixtures investigated the behaviour was better or comparable to "standard" self
compacting concrete. When compared with concrete containing raw fly ash, the mixtures
prepared with high energy milled fly ash have shown a decrease in slump flow of less
than 8%. The flow data recorded for mixtures made with zeolite were found to be lower
than those for mixtures containing fly ash. In all cases, the particle-size distribution of the
fine fly ash tested are able to favour a reduction of the chemical admixtures content in
The compressive strength determined for the mixtures made with the finest
milled fly ash was found to be not much lower than those of the silica fume-concrete
specimens. This is true both for the two curing times investigated and for the two types of
concrete prepared. Significant differences were seen for the mixtures containing zeolite.
The modulus of elasticity values confirmed the trend observed for the
REFERENCES
1
V.S. Ramachandran, 1995, Concrete Admixtures Handbook; Properties, Science and
2
C.F. Ferraris et a!., 2001 , "Fresh Concrete Rheology - Recent Developments",
3
F. Lange, eta!., 1997, "Dense Packing of Cement Pastes and Resulting Consequences
on Mortar Properties", Cement and Concrete Research, Vol. 27, No.IO, pp. 1481-1488.
4
V.M. Malhotra, P.K. Mehta, 2002, "High-performance, high-volume fly ash concrete:
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
6
V. M Malhotra, 2002, "High-performance high-volume fly ash concrete. An
7
F. Collins, J.G. Sanjayan, 1999, "Effects of ultra-fine materials on workability and
strength of concrete containing alkali-activated slag as the binder", Cement and Concrete
8
PK Mehta, 1999, "Advancements in Concrete Technology", Concrete International,
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Vol. 21, No.6, pp. 69-76.
9
Bakharev T. et a!., 2000, "Effect of admixtures on properties of alkali-activated slag
concrete", Cement and Concrete Research, Vol. 30, No.9, pp. 1367-1374.
10
Collepardi, 2003, June, "Innovative concrete for Civil Engineering Structures: SCC,
HPC & RPC", Proc. of the Workshop on New Technologies and Material in Civil
11
J. Ambroise et al., 1999, Self-Levelling Concrete-Design and Properties, Concrete
13
F. Colangelo et al., 2004, "Influence of admixtures on the properties of Self-
14
S.D Fossey et al., 2003, "Superclassified PFA for self-compacting concrete", Proc.
lith Int. Cong. on the Chern. of Cement, Durban, South Africa, pp.769- 779.
15
N.Bouzoubaa, M. Lachemi, , 2001 "Self-compacting concrete incorporating high
volumes of class F fly ash. Preliminary results", Cement and Concrete Research, Vol.Jl,
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
16
ACI - SP233 CD, 2006, Workability of SCC: Roles ofIts Constituents and
17
B. Liguori, D. Caputo, M. Marroccoli, C. Colella, 2004, Evaluation of zeolite-bearing
International Conference on Fly Ash, Silica Fume, Slag and Natural Pozzolans in
Concrete, American Concrete Institute, Farmington Hills, MI, U.S.A., Editor: V.M.
Malhotra, p. 319-334.
Composition, kg/m,
System CEM FA FAI FAll FAIII SF NZI NZII SP VMA Aggr.
c 365 5.20 1.92 2335
A-FA 365 146.0 7.26 1.20 2303
A-FAI 365 146 7.26 0.65 2309
A-FAIT 365 146 7.26 0.75 2306
A-FAIII 365 146 7.26 1.00 2305
A-SF 365 109.5 36.5 7.26 0.60 2285
A-NZI 365 146 7.26 1.40 2327
A-NZII 365 146 7.26 1.70 2322
R-FA 219 146.0 5.20 1.90 2279
R-FAI 219 146 5.20 1.80 2285
R-FAII 219 146 5.20 1.80 2282
R-FAIII 219 146 5.20 1.90 2280
R-SF 219 109.5 36.5 5.20 1.80 2266
R-NZI 219 146 5.20 1.90 2295
R-NZII 219 146 5.20 1.90 2291
100
.• ,
90 /
/ _l .... FA!
/, I
[!()
/;
I
i: I
--FAIT
I:
: I I
I / I
- .. FAIII
/
.,.
-
20
10 ~,..
0 -FA
0.1 10 100 1000
Sew,mm
Fig. 2- Details of normal slump flow (left) and )-ring slump flow (right) tests.
150
140
130
120
110
100
90
e
:=" eo
·r:c 70
80
50
40
30
20
10
55 60 65
Vt'loeUy.J.Lt
Fig. 3- Real scale specimens for resistance to segregation test (left) and velocity of
ultrasonic pulse as function of height (right).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
i /UO i!l28~
I ::
t::
8 :no
~ 10.0
M~~~~~~~~~~~~~
1 !,'
; I II I 7 r.<-
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Ii i .Iii I
c RooFA RooFAI RooFAn RooFAni RSF R<-NZI &NZII
--------------------------------~------~--~
,----
({)
~ r-;.
~
i IIII I~ I
f 50
f F
1 40
ll:l
~
1 III III
"a
I I
s :l) :\\:
1 3)
10
0
II: ~)~ ::j:J ~~j~ ~ m
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: A novel slurry rheology modifier (Vl) has been developed for slurry,
mortar and concrete using an inorganic powder such as cement. This chemical
admixture mainly consists of cationic and anionic surfactants. VT can efficiently
exhibit unique viscoelasticity in spite of low molecular weight when the cationic and
anionic surfactants coexist in slurry at the same time. This technology gives very good
segregation resistance to cement slurry and concrete under several conditions, e.g.,
water, vibration, and high water-cement ratio and without setting retardation. The
viscoelasticity mechanism ofVT is caused by the formation of higher-order structure
of surfactants in slurry. Controlling the viscoelasticity ofVT is an effective method
to prevent segregation for cement slurry, various grouts for repair and ground
improvement. VT has various potential abilities in civil engineering and construction
fields because of unique viscoelastic behavior.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
Surfactants have been extensively made available for our daily life and industrial
activity such as detergents, for cosmetics, medicinal products, agricultural chemicals,
food products, construction and civil engineering. The main functions of surfactants are
detergency, emulsification, dispersibility, wettability and so on. In recent years, the
higher-order structure formed by molecule association of surfactants has been
investigated by academic chemists and industrial scientists. The technology development
utilizing the functions derived from higher-order structure has been expected in the
various fields. The authors applied the higher-order structure of surfactants to the
viscosity agents of concrete and developed a novel slurry rheology modifier (VT) for
cement slurry, mortar and concrete.
There have been many studies on the properties of viscosity agents for concrete
[1,2]. The most typical viscosity agents are water-soluble polymers, e.g., cellulose
derivatives [3], fermentation polysaccharide [4,5,6] and various types of synthetic
polymer [7], and are added to prevent segregation of concrete. However these viscosity
agents have several problems; for example, the dissolution rate of viscosity agent's
powder is slow due to the very high molecular weight, and handling and workability
become worse in the case of dissolving viscosity agents in water and setting retardation.
In contrast, VT gives excellent segregation resistance to concrete in spite of very low
molecular weight. Furthermore, since VT dose not inhibit the hydration of cement;
setting retardation is very small compared to conventional viscosity agents.
This paper reports the general properties ofVT in cement slurry and concrete.
The mixture proportions adopted for laboratory testing purposes in this study and
the specifications of the materials used are listed in Table 1.
Slurry test: Cement slurry was prepared by mixing cement together with water containing
viscosity agent and/or superplasticizer at a water-to-cement ratio of 0.5-6.0. In this
connection, the water-to-cement ratio of 0.49-0.55 in Table 1 was selected to get the
required compressive strength (more than 24.5MPa at 28 days). The slurry was mixed
with a hand mixer for 1.5 minutes. In the case of VT, mixing was carried out for 1.0
minute with the addition of compound A and then for 0.5 minutes with.the addition of
compound B.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Hydration property: Using the slurry prepared as a sample, exothermic curve of hydration
was obtained with TCC2-6 calorimeter. Hydration property was evaluated by measuring
the second exothermic peak time of C3S hydration.
Concrete tests: A biaxial forced-mixing mixer with a capacity of 50 liters was used to mix
40 liters of concrete sample. To prepare the sample, the cement and both coarse and fine
aggregates were placed in the mixer and mixed for 10 seconds; then water containing the
superplasticizer and the viscosity agent was added, and mixing continued for a further 90
seconds. The temperature of the concrete during testing, and also the laboratory ambient
temperature, were held at 20.0 ±0.5°C. After the completion of mixing, the slump flow
value, underwater segregation resistance, air volume were measured in accordance with
JIS AllOl and 1128. Self-compactability of the concrete was determined by using the
Box-type test apparatus (2]. The compressive strength test was determined according to
JIS A-1132 and ns A-1108.
Figure 2 shows the viscosity behavior of viscosity agents in water. lt is clear that
the viscosity of methylcellulose (MC) solution increases strikingly with increasing
concentration. On the other hand, the viscosity of compound A and B solution do not
almost change with increasing concentration, because both compound A and compound B
are not polymer but monomer of a few hundred of molecular weight as shown in Fig.!.
The viscosities of solutions are less than 20mPa · s at 20°C. This indicates that compound
A and compound B solution are very easy to handle.
The results for the viscosity of cement slurry (water-to-cement ratio of 0.7)
containing VT are shown in Fig.3. Compound A and compound B did not exhibit
viscosity individually. However, they could exhibit excellent viscosity in cement slurry
The viscosities of cement slurry vs. the VT dosage of plots for cement slurry with
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
different water-to-cement ratio are shown in Fig.5. A desired viscosity value can be
obtained for slurries of various water-to-cement ratios by adjusting the dosage ofVT.
Figure 6 shows the fluidity aspect of cement slurry having ditierent concentration
of VT at same water-to-cement ratio of 0. 7. The flowable viscous slurry and gumlike
viscoelastic slurry were manufactured by changing the concentration of VT.
Figure 7 shows the angular frequency dependence of G' (G' is called as storage
modulus and defined as G'( w )= G ( w )I y 0 (t=O) w :angular frequency, G :strain,
y :stress, t: time) of VT solution and MC solution. The MC solution has the frequency
dependence of typical flexible polymer solution. On the other hand, the VT solution has
what is called a Maxwell-type frequency dependence. The G' ofVT solution was higher
than that of MC solution at a broad range of angular frequency. This indicates that VT is
effective to avoid the segregation of slurry and thus to obtain the homogeneous slurry.
The rheological property of the CaC03 slurry ofVT was compared to that ofMC in
Fig. 8. The CaC03 powder was used to prevent the influence of hydration reaction on
rheological property during the measurement of lhr. The results of slurry were reflected
in the aqueous solution experiments, that is, the G' of VT slurry was extremely high
compared to that of MC slurry at the slurry viscosity 800mPa · s. Additionally, the fresh
behavior of CaC0 3 slurry was very similar to that of cement slurry.
Figure 9 shows the picture of the VT in coexistence with the blast-furnace slag
powder by using Cryo-SEM, VT of0.05 mol/1. The higher-order structure of30-50 nm in
diameter spreads homogeneously around the powder particle. This structure could give
the slurry the unique viscoelasticity, so that the cement slurry containing VT had
excellent segregation resistance in water, as shown in Figure 10.
The second exothermic peak time of C3S hydration was measured to investigate the
influence of viscosity agents on the hydration of cement (Fig.!!). In the case· of MC, the
hydration was strikingly inhibited compared to the control. In contrast, though VT slurry
was about the same viscosity as MC slurry, retardation of VT was within 1 hour.
Moreover, the VT slurry showed the retardation of mere 2 hours regardless of triple
viscosity 15,000mPa · s for MC. This suggests that water molecule used in hydration can
get through the higher-order structure and then get around to the surface of cement
particles.
The cement slurry at water-to-cement ratio of > 1.0 was prepared without
The interaction between the VT and superplasticizers was shown in Fig.l4. In the
case of Polyether-based superplasticizer (PE)[8)NT, the viscosity of slurry did not
significantly change with increasing dosage of PE. In contrast, the viscosity of
naphthalene-based superplasticizer (NS)NT slurry decreased with increasing dosage of
NF. It was thought that the phenomenon of NSNT was caused by inhibiting the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
formation of high-order structure due to the pi electron of naphthalene ring. It is clear that
PE was more suitable for VT slurry.
The underwater concrete was examined for self-compactability and good hardening
property by utilizing the feature of VT. The results of concrete tests are shown in Table 4
and Table 5. In the case of MC/Melamine-based superplasticizer (ML), the
self-compactability was not observed because MC/ML concrete could not get the
sufficient fluidity in spite of increasing dosage of ML. In contrast, VT/PE concrete had
good self-compactability with segregation resistance in water at some mixture
proportions. Additionally, the strength ofVT/PE concrete was more than 5N/mm2 in 24 hr
However, MC/ML concrete did not harden in 24 hr. It was confirmed that good hardening
property was compatible with self-compactability and underwater segregation resistance
by usingVT.
CONCLUSIONS
2. The viscoelasticity ofVT is exhibited when the cationic and anionic surfactants coexist
in slurry at the same time, and the viscosity of each surfactant solutions is very low.
5. A cement slurry was produced at water-to-cement ratio of> 1.0 without segregation and
retardation by adding VT and the test pieces had extraordinary low-strength and a high
water-absorbing property.
REFERENCES
[1] Izumi, T., Tadokoro, T., Mizunuma, T. and Kawai, T., 'Characterization of
thickener's adsorption in view of mortar flow', Proceedings of the
Japan Concrete Institute, Vol.l7, No.I, 1995, pp. 63-68
[2] Hibino, M., 'Effect of Viscosity Enhancing Agent on Self-Compactability of
Fresh Concrete', Sixth CANMET/ACI, 2000, pp. 305-320
[3] Yamamuro, H., Hamada, D. and Shonaka, M., 'Development of a liquid type
admixture for self-compacting concrete' 2nd Symposium on Self-Compacting
Concrete, Oct. 2001, pp. 179-186
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
[4] Kang, K.S. and Pettitt, D.J., 'Industrial gums third edition'
(R.L. Whistler, J .N .Bemller edition), Academic Press, pp354-, 1993
[5] Shindoh, T., Matsuoka, Y. and Sornnuk T.,' Fundamental study on properties of
super wakable concrete ' Proceedings of the Japan Concrete Institute , Vol.l3,
No.1, 1991, pp. 179-184
[6] Khayat, K.H. and Saric-Coric, M., 'Effect of Welan Gum-Superplasticizer
Combinations on Properties of Cement Grouts', Sixth CANMET/ACI, 2000,
pp. 249-268
[7] Yamamuro, H., Izumi, T., Mizunuma, T. and Kawai, T., 'A research on high
flowable concrete with non-adsorptive thickeners', Proceedings of the
JapanConcreteinstitute, Vol.17, No.I, 1995, pp.99-104
[8] Mizunuma, T., Izumi, T., Kanzaki, S. and Iizuka., 'Properties of high range
water reducing concrete with new high range water reducing agent(AE type)'
Proceedings of the Japan Concrete Institute, Vol.l4, No.1, 1992, pp. 33 7-342
Compound A CompoundB
R4
R 0 S0 3 Na
I
R - N -R
1 I 2
oX
R3
Fig. 1-Chemical structure ofVT
~- -o- I
i-;;- 40000
~
30000
Compound A
.....__ Compound B
_._ Methyj cellulose
~ 20000 II
't
.t'
8"'
.~
_
10000
i---'1'-------======>i
..' r:ote;:~ ;;1
Viscosity of Compound AlB
are less than 20mPa ·s ~\)~
> o .::.- L_______o--e--'~-o-•-
0 5 10 15 20 25
Concentration of viscosity agent (%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
>"'
too~~~--~----~--~~~~
additive-free A B AB AB AB
Fig. 3-Relationship between viscosity of slurry and concentration of VT
100000 , - - - - - - - - - - - - - - - - - - - ,
~ 10000- -
f
"""0
.£'
"'
0
~
> 100~---.~--~--~----~--~
0.0 0.2 0.4 0.6 0.8 1.0
Mole fraction of Compound A (NA+B)
Fig. 4-lnfluence of mole fraction of compound A on viscosity of slurry
9000
W/C
r-. 0.70
;;"'
} 6000
""'"'0
0 3000
·:;;
0
()
.;!l
>
0
0 1.0 2.0 3.0 4.0 5.0 6.0
Dosage of Compound A+B (WX%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
I 0,000 mPa · s
0.1
0.01 0.1 10 100
Angular rate w (rad/s)
Fig. 7-Frequency dependence of G' ofVT and MC solution
i
0..
G' 1000
~
1100
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
rrJ 800 mPa·s
1 ~--~----~----~--~
without
VT
Fig. 13-Test of grout with and without VT for bored precast pile method at
sand ground
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Vi
'0 2000 '\
.€
8 1000 ~NS
~ 0 .~
0 0.2 0.4 0.6 0.8 1 1.2
Dosage of superplasticizer (CX%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Presence of Gypsum
~\fnopsi~ Different types of superplasticizers have been widely used over the past
few decades in order to produce a more fluid or very high strength and durable
concrete. These chemical admixtures interfere with the various physico-chemical
processes occurring in early cement paste.
In this paper we present results from a study on the influence of superplasticizers
on pure tricalciumaluminate hydration in presence of gypsum. The suspensions
hydration has been investigated by conductimetry, isothermal calorimetry and total
organic carbon analysis of the liquid phase.
The time taken for ettringite formation has been determined without superplasticizer
and in presence of three different types of superplasticizers: polynaphtalene
sultanates (PNS), polycarboxylate-polyox (PCP) and diphosphonate terminated
polyoxyethylene.
Whereas diphosphonate terminated polyoxyethylene does not seem to modify
tricalcium aluminate hydration carried out in presence of gypsum, PCP and even
more PNS slow down ettringite formation. This effect seems to be largely due to a
decrease of the C3A dissolution rate and might be connected to an adsorption of
PCP. or PNS observed from the early C3A hydration. Such an adsorption does not
happen with diphosphonate terminated polyoxyethylene superplasticizer. Moreover
the presence of PCP superplasticizer causes a decrease in the size of the ettringite
crystals formed.
INTRODUCTION
Moreover it is reported that the type of superplasticizer (SP) and the mixing
method as preparations conditions or the surrounding environment of concrete could
significantly affect the fluidity of the paste [3-14].
The main purpose of our study was to investigate the influence of the chemical
structure of various SPs on the interaction with early cement hydration products in order
to have a better understanding of the paste fluidity evolution.
Portland cement mainly consists of four mineral phases: the most abundant
phases are tricalcium silicate (C 3S, alite) and bicalcium silicate (C 2 S, belite) and the other
phases are tricalcium aluminate, Ca3A]z0 6 (C 3A) and calcium aluminoferrite
Ca4A[zFez0 10 (C 4AF, celite). Moreover gypsum CaS0 4 • 2H20 is added to clinker to
regulate the reactivity of the aluminate phases. When cement is mixed with water, it
undergoes a dissolution reaction generating calcium ions, hydroxyde, silicate, aluminate
Materials
Three types of superplasticizers were investigated for this study. Their chemical
structures and molecular weights are described in Table 1. The pH of these additives was
adjusted to 9.60 by adding 1M NaOH in order to neutralise all the acidic functions
present in the additive. Since adsorption mechanisms are governed by the concentration
of the additive in the solution, the superplasticizer dosages were calculated in order to
have the same concentration in solution as in the pore solution of a typical concrete. The
three superplasticizers were:
A polycarboxylate-type superplasticizer with a polyoxyethylene graft chain
(PCP) produced by Chryso. The main carboxylic and grafted polyoxyethylene chain
lengths are reported in Table 1. --`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The C3 A-gypsum mixture was hydrated at 25°C in dilute lime saturated suspensions with
a liquid/solid ratio equal to 25 under inert atmosphere in order to avoid carbonation. The
suspension was continuously subjected to mechanical stirring. The chemical evolution
was followed by isothermal calorimetry (Tian-Calvet Setaram) and by electric
conductivity measured with an XE 150 electrode (Tacussel).
To determine concentrations of species in solution ([Ca 2+], [Ae+], [SO{]) or the polymer
adsorption during the C 3A hydration reactions, small portions of the suspensions were
removed , filtered through 0.3 !Jill millipore filters and acidified to avoid carbonation. A
carbon analyser, model TOC 5050 (Shimadzu) was used to determine the evolution of
the adsorption of the polymer during the C 3A hydration. Atomic Absorption
Spectrometry (Perkin Elmer) was used for the calcium and aluminium concentrations
and ionic chromatography (Dionex) equipped with a conductivity cell was used to
determine sulfate concentration.
The reference experiment was carried out by adding 37mmol of C3A and
7 .25mmol of gypsum to 250mL of lime saturated solution. The evolution of the sulfate,
calcium and aluminium ions concentrations and the heat flow deduced from the
isothermal calorimetry are reported in Fig. I. The Fig. 1 shows two parts.
During the first part, sulfate ions are still present in the solution and they are consumed to
precipitate a hydrate. That is the situation during the workability period. The second part
begins when the sulfate concentration becomes equal to zero.
First part-- When C3 A-gypsum mixture is introduced in a lime saturated solution without
additive, it undergoes dissolution reactions generating calcium ions, hydroxyde,
aluminate and sulfate anions in the interstitial solution according to the dissolution
reactions:
The suspension very quickly becomes supersaturated with respect to ettringite and AFm.
Previous studies have clearly showvn that although ettringite is the more stable hydrate,
AFm phase precipitates at the very beginning of the C3A hydration and the early AFm
precipitation depends on different factors as for example the sulfate type used, C 3A
granularity or superplasticizer presence (18, 19]. Consequently the exothermic peak
obtained at the beginning of the first part expressed the very high reactivity of the system
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Ettringite precipitation consumes calcium, aluminium and sulfate ions which causes
further C3A and gypsum dissolution. A stationary state is then established resulting
from dissolution reactions and from ettringite precipitation reaction. When gypsum is
totally dissolved, ettringite precipitation rate can be deduced from the decrease of
sulfate ions concentration. As usually reported and observed in these systems, the
average rate of ettringite precipitation from C3A and gypsum, is decreasing according
to time. For instance Fig.1 shows that 70 minutes (duration of the "sulfate plateau") are
needed to precipitate 1.4 mmo1 of ettringite from 4.2 mmol of gypsum and 1.4 mmol
of C3A (according to reactions I, II and III), while 140 minutes are needed to "convert"
the last 3 mmol of sulfate into I mmol of ettringite. Moreover, because sulfate
concentration decrease is linear, we assume ettringite precipitation is quite constant
during this period. Such an evolution on the rate of ettringite precipitation from C3A
and gypsum has already been reported [18]. The average rate of ettringite precipitation
clearly depends on the C3A granularity but also on other parameters as sulfate
concentration.
Second Part.--This part begins when sulfate concentration becomes equal to zero. C3A
hydration without sulfate leads to a metastable calcium hydroaluminate precipitation
according to the reaction:
The exothermic peak appearing at the beginning of this period is attributed to C3A
dissolution which is going on faster and C4AH 13 precipitation. Due to C3A dissolution (I)
and C4AH 13 precipitation (IV), calcium ions concentration decreases and consequently
conductivity too, whereas aluminium ions concentration increases.
In the same time monosulfoaluminate(3Ca0 AbOJ.CaS04 ,12H2 0) becomes more stable
than ettringite and consequently precipitates by using sulfate ions liberated by ettringite
dissolution. This last reaction is very slow and ettringite is still present two days later.
Finally, all these reactions give rise to a stationnary state.
reported in Fig.ll. An initial adsorption is only observed for both PCP and PNS. In the
case of the PNS, a large part of the superplasticizer, about 150mg ( 30%), is adsorbed
during the first minutes. The initial adsorption could happen on the anhydrous phases but
also on the AFm formed at the very begining of the C3 A hydration, because of its
important specific surface area.
In a second time, a quite linear increase of adsorption is observed during the ettringite
precipitation with all superplasticizers. Nevertheless, whereas about lOOmg of PCP or
PNS are adsorbed on ettringite, only 1Omg of diphosphonate seems to be adsorbed on
ettringite. These amounts are larger than the values previously calculated for the PCP and
PNS (~30mg for PNS and PCP). These values are reported infFig.12. These differences
can result from the adsorption of other phases which could also be formed (AFm for
CONCLUSIONS
Diphosphonate does not modify the average rate of ettringite precipitation. No adsorption
happens during the first minutes and the adsorption level of the polymer on ettringite is
low.
The PNS slows down the ettringite precipitation. The decrease of the average rate of
ettringite precipitation is higher with higher PNS dosages. Moreover, the time needed to
precipitate ettringite varies as a function of the initial PNS adsorption level. PNS may act
on C 3A hydration by decreasing its dissolution rate, in consequence of its strong initial
adsorption. By using PNS delayed addition, the strong initial adsorption on the aluminate
phases can be avoided and consequently more PNS can act efficiently on silicate phases
in slowing down its hydration and delaying setting.
PCP also slows down ettringite precipitation, but the decrease of ettringite precipitation is
less strong than with PNS. Whereas PCP may decrease the C3A dissolution rate, it mainly
acts on ettringite growth giving rise to smaller size of ettringite crystals.
REFERENCES
/ CH2-PO/, 2Na+
Diphosphonate CH3 -[0-C~-CH2Jn-N+-H
PCP
t CH2-t cj~ c~)
too t=o
CH2-t -
6
Na+ n p q
( tH2CH20)m
~~_.'
PNS
S03Na
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Hydrogarnet Monosulfo
Cornpmmd c3~ ettringite aluminate
Part1 Part2
- ( d ) Conductivity mS/cm
20
o~~~~~~~~~~--~---=~==~~==~o
0 100 200 300 400 500 600 700
llme(min)
Fig. 1-- Evolution of heat flow, electric conductivity and ions concentration during C A
3
hydration (10g) in saturated lime solution in presence of 1.25g of gypsum. The liquid
to solid ratio was adjusted to 25
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Fig. 2·- Evolution of electric conductivity during C3A hydration (10g) in saturated lime
solution in presence of 1.25g of gypsum and in presence of o.5g of superplasticizer.
The liquid to solid ratio was adjusted to 25.
12
Fig. 3 -- Evolution of the electric conductivity during C3A hydration (tog) in a lime
saturated solution with or without o.5g of diphosphonate and in presence of different
quantities of gypsum. liquid/solid= 25.
-{a)conductlvlly relaranca
- -(b) conductlvttyPNS 2%
······ (c)oonductiYity PNS 5%
--[sulfate] mmoVL reference
"""'!'l •[sulfate]mmolll PNS 2%
- .. -[sulfate] mmolll.. PNS 5%
'
\
Fig. 4-- Evolution of the electric conductivity during C3A hydration (tog) in a lime
saturated solution in presence of t.25g of gypsum and in presence of different
amounts of PNS. The liquid to solid ratio was adjusted to 25.
- 600
~
;sao
.
.
~
=
'!i
400
300
.a.
.a.
j 200
i!,
.. 100
4 10 12 14 16
PNS lniUally adsorbed rnglg OfC3A
Fig. s-- Evolution of the duration of ettringite precipitation versus initial adsorption
level of the PNS. tog of C A were hydrated in zsoml of a lime saturated solution and
3
in presence of 1.25g of gypsum.
14 ---------------------------------------,100
••• 90
• r~-------------.--:~::::~=o···- 80
70 §
•
! ...._(aluminium] mmoi/L
aoJ
so'lil
o o o o
•
•
t
I
o [PCP3S%] adsorbed %
o (calcium] mmoL'L
401B
o oo o o o • o f
~0 --------------------f~-
o.!.·'o 20
0 0 <t....-"' ..
~~,! __ 1>.,_!.---~---~--~.--------.. • 0 0 10
~-.~~~~~~~. .~~~-------,~----~~----+0
0 100 200 300 400 500 600
Time(min)
Fig. 6-- Evolution of the electric conductivity and ions concentration during C A
3
hydration (wg) in a lime saturated solution in presence of 1.25g of gypsum and o.sg
of PCP. The liquid to solid ratio was adjusted to 25.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Fig. 7 ·-:Evolution of the electric conductivity and the sulphate concentration during
C A hydration (10g) in a lime saturated solution in presence of 1.25g of gypsum and in
3
presence of different amounts of PCP.
- • - [calduml mmolll
---[suffate]mrnol/1..
--[ahJminiuml mmo\f\.
turraymbola,..fllter01J!m
Fig. 8-- Evolution of the aluminium, calcium and sulfate concentrations of the
solution filtered through 0.3 11m or o.liJm millipore filters during the C3A-Gypsum
hydration in presence of s%PCP.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
3 JliD
Fig. 9-- Typical SEM images of ettringite obtained 1 hour after the beginning
of the Cl-gypsum mixture hydration in presence of s% of PCP or without
superplasticizer (down).
20 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - .
18
16
1,,,""
--------------------------
Fig.b
~ 14 I
,
I
I
12 -C3AH6
10 - •· ettringite
......,...... m on oau lfoafum in ate
!I 8
if"' 6
20 .--------------------------------------------,
18 Fig.c
-----------·
.
;,.' .·•· · · ~---=--
•· · · · · · · · · · • .
16
"'14 ------- -•-C3AH6
E /,- ··+·· ettring ite
-g 12
.,..
€ 10
n.
0
n.
0
6
8 ,. /
"'
.......-monosulfoaluminate
~-J
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
4
2 .......__._...,!/
---------------_
0
0 0.5 1 .5 2 2.5 3
[PCP] giL
Fig. 10 a,b,c.-- Adsorption of the diphosphonate (a), PNS (b) and PCP (c) on aluminate
phases: ClH 6 , ettringite and monosulfoaluminate in a lime saturated solution and
with a liquid-to-solid ratio adjusted to 25.
~100
€
0
~ 80
.r-··-·-
~
:il 60 ( . ·············•······-··········•··················· ........................... .
0
.!!!
~40
a.
r.-----~;
' i
.
""'
~ 20
__.. .•..-!: ......-""' · ·- · ·% diphosphonate adsorbed
-%PNS adsorbed
0
.............. -
..'
- •% PCP adsorbed
120
a; II experimental adsorption Ill adsorption calculated from isotherm curves
!:?
gtoo
~c
.
~
<=
0
80
~
60
.0
II)
'0
c.. 40
en
'0
E 20
~"
«J
0
PCP PNS Diphosphonate
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
BACKGROUND
means, that the portion of flocculated cement particles in the system is low. The cement
particles flocculation is due to attractive interparticle forces which can he of van der
Waals or electrostatic origin. The magnitude of these forces increases as the interparticle
distance decreases. This may happen at rest and over the time and lead to the system
showing a slump loss. The stronger the attractive interparticle forces are, the larger is the
yield stress related to the slump loss. However, not only the interparticle forces but also
their number influence yield stress. High volume fractions increase the number of
interparticle contacts and thereby the interparticle forces.
Thus, in order to increase the fluidity of the system one can attempt either reduce
the number of possible interparticle contacts (by increasing the water content) or by
decreasing the magnitude of the attractive interparticle forces (adding a superplasticizer).
We also note that mixing will have temporarily reduce the number of particles in close
contact and have an effect on both the number of particles in close contact as well as on
the magnitude of the interparticle forces.
Increasing the W/C or reducing the cement content in the concrete at a given
amount of water unfortunately leads to reduced strength of the hardened material and to
reduced durability. Therefore superplasticizers or high performance water reducers
should be preferred. These substances adsorb on the cement particles. If particles come in
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
performance as plasticizer with respect to flow, flow retention, setting time and
sensitivity to changes in the cement solution 4·6 . The adsorption of a polymer also depends
on the substrate, in this case on the type of cement and the solution chemistr/· 8 .
Therefore in some cement-superplasticizer combinations unsatisfying fluidizing
properties or fast stiffening occur which are described as cement-superplasticizer
incompatibility.
EXPERIMENTAL
Materials
Polycarboxylate-ethers--In the first series (PEI-PE4) only the ratio of
methacrylic acid to polyalkylene glycol methacrylate was varied keeping the length of the
side chains the same. In the second series (PE5-PE10) the length of the side chains was
varied too (Table 1). For the use in mortar or concrete the polymer solutions were
defoamed with a commercial defoaming agent.
Methods
Adsorption tests--200 g cement and 100 g water containing the polymer are
blended with a mechanical stirrer. After the defined time part of the slurry is taken and
the liquid phase is separated from the cement by centrifugation with 8000 rpm for 8
minutes. The amount of polymer in the supernatant phase is determined by HPLC with a
light scattering detector. The difference in the polymer concentration in the liquid before
and after adsorption is calculated as adsorbed polymer.
Mortar test--Mortar with aggregates from 0-8 mm is used (Table 3). The cement
is mixed with the sand in a Hobart blender for 1 minute. Then the water containing the
superplasticizer is added and the mortar is blended for an additional 3 minutes. A flow
cone of 100 mm lower and 70 mm upper diameter and a height of 60 mm is filled with
the mortar and lifted. The flow table spread is the diameter of the mortar cake produced
when the flow table is dropped 15 times from a height of 20 mm. The measurement of the
flow is repeated after 30 and 60 minutes with a re-mix of 30 seconds.
SCC--The mixture proportion of SCC is given in Table 4. The fly ash is added to
avoid bleeding and segregation. The mixing is done according to EN 480-1. Aggregate,
sand and filler are placed first in the mixer. Half of the mixing water is added and all is
mixed for 2 minutes. The mixer is stopped and after 2 minutes of rest the cement and the
fly ash are added and all is mixed for 30 seconds. Next the remaining water containing
the superplasticizer is added within 30 seconds and the concrete is mixed for another 2
minutes. A flow cone of 200 mm lower and 100 mm upper diameter and a height of 300
mm is filled with the concrete and lifted. The spread of the cement cake is measured
when the concrete stops flowing and this value is called "slump flow". The measurement
is repeated after 30, 60, 90 and 120 minutes with 15 seconds re-mixing.
The adsorption after 30 minutes of PE5 to PEIO was measured on NPC3, NPC4
and NPC5 (Fig. 4) . With increasing dosage, the amount of adsorbed polymer increases
but stops at a certain level. This dosage where no more adsorption takes place when
polymer is added is the so called dosage of surface saturation which is dependent on the
structure of the polymer and the type of cement. Again NPC4, the cement with the
highest content ofC 3A, adsorbs/consumes the highest amount of polymer. The adsorption
on NPC3 and NPC5 is comparable. When we compare the polymers with the same
(short) length of the side chains (PE5, PE6 and PE7) we see, that the polymer with the
high amount of methacrylic acid and with the longer backbone (PE6) adsorbs most,
whereas a low amount of methacrylic acid, as seen before, results in poor adsorption
(PE7). This is also confirmed by the high adsorption ofPE8 with it's high content of
methacrylic acid in the backbone and medium length of the side chains. Increase of the
side chain length by keeping the amount of methacrylic acid in the backbone, as done for
PE9 in comparison to PE8, reduces the adsorption. A high degree of adsorption with low
amount of methacrylic acid in the backbone can be reached with a combination of very
short and medium length of the side chains (PEIO).
Mortar tests with PE 1 to PE4 with NPC 1 and NPC2 (Fig. 7 ) and the mortar tests
ofPE5 to PEIO with NPC3, 4 and 5 (Fig. 8) show, that not only each cement has it's
optimum superplasticizer but that also different W/C show different optimal polymers. In
the first mortar series (Fig. 7) the best superplasticizer for NPC 1 at higher W/C and lower
polymer dosage is PEl, the polymer with the lowest amount ofmethacrylic acid in the
backbone. The flow table spread decreases from PEl to PE4 in which the content of
methacrylic acid in the backbone of the polymer increases (Fig. 7 a), but when we reduce
Another mortar series with PES, PE6, PE8 and PE9 was tested. PE7 and PEl 0
could not be measured at the same time therefore PES was tested again in order to be able
to compare the performance ofall6 polymers (Fig. 8). Again no correlation of the mortar
flow with the amount of adsorbed polymer is found. The polymer dosage and Wle as
well as the type of cement influence which polymer is the best performing regarding the
flow of the mortar. Polymers with high adsorption like P£8, for example, show poorer
relative flow than PES at low dosage and higher W/e whereas better relative flow at
lowerW/e and higher dosage with NPCl. Whereas PE7, that has low adsorption rate, is
better at lower dosage and higher W/C than at higher dosage and lower W/C.
There is no trivial link between polymer adsorption and flow table spread of
mortars, which we attribute to the nature of this test. We infer that a similar conclusion
would be drawn for vibrated concrete.
The effect of the surface saturation on the slump retention can also be seen in
standard concrete where we measure the flow with 15 shocks on the flow table. Concrete
with a target flow of 45 +/- 2 em was produced at 3 fixed W/C ratios and NPC5. The
dosage of polymers PE 10 and PE7 was adapted to reach the desired initial flow value.
The change of the flow within 2 hours was determined. Figure 12 shows again the
increase of flow retention with polymer dosage. At a W/C of0.39 the dosage of PElO
with 0.18% (=1.8 mg /g cement) is far below the saturation level of about 3.0 mg/g,
therefore we find fast stiffening of the concrete. The 0.21% (2.1 mg/g cement) dosage of
PE7 is close to the saturation level (about 2.4 mg/g) and therefore only moderate
stiffening with time is recorded. Reducing the W/C to 0.36 and increasing the polymer
dosages to 0.28% and 0.53% for PEIO and PE7 respectively (Fig. 12 b) the saturation
level of the polymers is nearly reached or overreached leading to almost constant flow
over 2 hours. These tests again indicate, that the flow retention is correlated with the
degree of surface saturation.
CONCLUSIONS
In the work presented here and before 6 we were able to demonstrate the
importance of the adsorption of polycarboxylate ether polymers on the cement in order to
behave as high range water reducer or superplasticizer. The higher the adsorbed amount
of polymer the higher was the flow in the cement paste. Since the paste flow and the yield
12
stress are directly correlated"' , increasing the amount of adsorbed polymer reduces the
yield stress of a cement paste. The degree of adsorption is dependent on the polymer
structure and the cement. Higher ionic content in the polymer backbone and increased
backbone length increase the adsorption. The length of the side chains influences the
adsorption in as much as it influences the overall ionic content of the polymer and is
15
more important when very low W/C ratios are used 13 • •
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
f1ow retention, whereas dosages far below this surface saturation show fast stiffening. In
general, polymers with a high adsorption level reach this value only at comparably high
dosages compared to polymers with low adsorption rate. Polymers that show high
adsorption usually are already very good plasticizers at dosages far beyond this level of
surface saturation and are mainly used for precast conrete. They are in most applications
under dosed with respect to the surface saturation and therefore can give problems with
faster stiffening. Polymers with lower adsorption are generally dosed higher to give the
desired flow and they reach the surface saturation already at lower dosages compared to
polymers with high adsorption. These polymers are therefore usually very good in
applications where long slump retention is desired like in ready-mixed concretes.
ACKNOWLEDGEMENTS
This work received funding from the Swiss Federal Office for Education and Science
(contract No 00.0273-2) and was part of a project funded by the 5th European Framework
Programme (Contract G5RD-CT -200 1-00435).
We want to thank our collaborators, Domenico Larosa and Jabbar AIShemari for their
help in producing the polymers and performing most of the application tests and our
colleagues from the Materials Testing Department for performing all concrete tests.
REFERENCES
1. Uchikawa, H., Hanehara S. and Sawaki, D.:"The Role ofSteric Repulsive Force
in the Dispersion of Cement Particles in Fresh Paste Prepared with Organic
Admixture", Cern. Concr. Res. 27 (1997), pp 37-50.
2. Yoshioka, K., Sakai, E., Daimon M. and Kitahara, A.: "Role ofSteric Hindrance
in the Performance ofSuperplasticizers for Concrete", J. Am. Ceram. Soc., 80,
(1997), pp 2667-1671.
3. Flatt, R.J., Houst, Y.F., Bowen, P. and Hofmann, H.: "Eiectrosteric Repulsion
Induced by Superplasticizers between Cement Particles - an Overlooked
Mechanism?", Proc 6 1h CANMET/ACI Int Conf on Superplasticizers and Other
5. Ohta, A., Sugiyama, T. and Tanaka, Y.: "Fluidizing Mechanism and Application
ofPolycarboxylate-Based Superplasticizers", Proc 5th CANMET/ACI Int Conf
on Superplasticizers and Other Chemical Admixtures, Ed. V.M. Malhotra,
Rome, Italy, 1997, ACI SP-173, pp 359-378.
8. Yamada, K., Ogawa, S. and Hanehara, S.: "Controlling of the Adsorption and
Dispersion Force of Polycarboxylate-Type Superplasticizer by Sulfate Ion
Concentration in Aqueous Phase", Cern. Concr. Res. 31 (2001), pp 375-383.
9. Comparet, C., Nonat, A., Pourchet, S., Mosquet, M. and Maitrasse, P.: "The
Molecular Parameters and the Effect of Comb-Type Superplasticizers on Self-
Compacting Concrete: A Comparison of Comb-Type Superplasticizer
Adsorption onto a Basic Calcium Carbonate Medium in the Presence of Sodium
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
10. Flatt, R.J. and Houst, Y.F.: "A Simplified View on Chemical Effects Perturbing
the Action ofSuperplasticizers", Cern. Concr. Res. 31 (2001) pp 1169-1176.
11. Roussel, N., Stefani, C., Leroy, R.: "From Mini-Cone Test to Abrams Cone
Test: Measurement of Cement-Based Materials Yield Stress Using Slump
Tests", Cern. Concr. Res. 35 (2005), pp 817-822.
12. Flatt, R.J., Larosa, D. and Roussel, N.: "Linking Yield Stress Measurement:
Spead Test Versus Viskomat", Cern. Concr. Res. 36 (2006), pp 99-109.
14. Vickers, Th.M., Jr., Farrington, S.A., Bury, J.R and Brower, L.E.: "Influence of
Dispersant Structure and Mixing Speed on Concrete Slump Retention", Cern.
Concr. Res. 35, (2005), pp 1882-1890.
15. Yamada, K., Takahashi, T., Hanehara S. and Matsuhisa M.: "Effects ofthe
Chemical Structure on the Properties ofPolycarboxylate-Type
Superplasticizers", Cern. Concr. Res. 30 (2000), pp 197-207.
-
T a ble 3 M'IXture proportions of mortar
Series 1 Series 2
Cement 750 g 880 g
Limestone filler ISO g 320 g
Quartz sand 0.08-3.2 mrn 2700 g
Limestone 0-1.2 mm 950 g
Limestone 1.2-4 mm 800 g
Limestone 4-8 mm 1250 g
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
non-ionic polyether side
chains (non-adsorbing)
anionic backbone
(adsorbing)
It ot-----~~--------~
0
6 8 10 12 14
Separation [run]
Figure 3- Adsorption of polymer PE1 to PE4 on NPC1 and NPC2 at dosages of 2 and
4 mg solid polymer per g of cement after 30 minutes adsorption time
t)
A. 2.0 +----1---+---i
~
"'
.§. 1.5 +---±:;.,.,.."':,;;.....~;:......o
1:
e.
0
1.0 -t----u.......~~-t.,J--"-'"-=--'f
t)
2.5
NPC4 ........
~ II.
~ 0.5 -t--BJ---t-----+-----1 z 2.0
~
0.0 r l f - - - - 1 - - - - + - - - - - l g 1.5
0.0 1.5 3.0 4.5
2.5 ,..---..,.----,---....,
"0
= 1.0
e-
t) NPCS
~ 2.0 +---+---+---:-.......~
..
0
'0
4:
0.5
~
0.0
_!1.5-l---\--;-?;.::::l!:;:::;;:::::i~
0.0 1.5 3.0 4.5
c:
~ 1.0 -t---b~::Y-=---n--~~""1.,1 Dosage (mg/g cement)
e0 -+-PES --PE6 • D ·PET
~ 0.5 • 0. ·PES • 0 • PE9 _.,_PE10
Figure 4- Adsorption of PEs to PElo on cements NPC3, NPC4 and NPCs at different
dosages and after 30 minutes adsorption time
250 -+-PES
--PE6
• D • PE7 -----j~;:;:o"~~
• 0.• PES
• O• PE9
Figure 5- Flow of the cement paste with cement NPC3 at a W/C of 0.23 after
30 minutes
e
230
210
190
p.:>r
?-:::. PE&-
-5. 170 _........ PE 10 -'-=------
~ 150
~ PE9
U::1l0
~ 110 ~E5
PE7
a.. 90
70
50
~0 L2 1A 1~ 1~
Adsorption (mg polymerlg cement)
a) b)
220 -,---.,..--....-----,----. 220 ,---...,----..---.---,
NPC1 i
i
-
200 e NPC1
2oo -t-- ~-~~+-~~~·~---+-~···---~-!----- -1
;: 180 s 180 -1---+••---+---+---1
~ 160 1 160
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
1&.
~140
0
~ 140 +--+111: 1-+11111--+11
2 120 +a;:-~ :i 120 +=J,:.+BI:'• t-a-•1··
100 +-"--..._....,..c..;...-o:....-....,_...,.L..f 100 -1-"""--+"'"'-"""4-"..,.4-'llll:..dl~
PE1 PE2 PE3 PE4 PE1 PE2 PE3 PE4
c) d)
220 ,...--...,.----,---,---, 220 .---.---.-~--..,
Figure 7- Flow of mortar with polymers PE1 to PE4; a) W/C 0-47. polymer dosage
0.18% solids, b) W/C 0.38, polymer dosage 0.35%, c) W/C 0-40, polymer dosage
0.18% and d) W/C 0.39. polymer dosage 0.35%
220
200
e.E. 180
~ 160
140
120
100
PE5 PE6 PE8 PE9 PES PE10 PE7 PE6 PES PES PE9 PES PE10 PE7
240
220
200
e
,E_180
~ 160
li:
140
120
100
PES PES PES PE9 PE5 PE10 PE7 PES PES PES PES PES PE10 PE7
240
220
200
1180
£160
140
120
100
PES PE6 PE8 PE9 PES PE1D PE7 PE5 PES PES PE9 PE5 PE10 PE7
Rgure 8- Flow of mortar with polymer dosages of 0.15% (a, c, e) and 0.30% (b, d, f). --`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
TheW /C was 0-40 and 0.42 for a, 0.35 and 0.38 forb, 0-43 and 0.44 for c, 0.38 and
0.39 ford, 0.40 and 0-42 fore and 0.34 and 0.35 for f.
Figure 9- Adsorption of PE10 on NPCs and on a blend of NPCs with fly ash in the ratio
of6o/4o
70
e
.2.60
;:
£50
ll.
§ 40
u;
30
20
PE10 (0.21%) PE7 (0.26%) PE7 (0.35%)
Figure 10- Slump-flow of sec with PE10 and PE7 at W/B 0.34 with a blend of NPCs
and fly ash in the ratio of 78/22
.t.PE10 : NPCS
[
z I .5
~
•PE7
····:·' , -+-,-~-.r,· •
~r···
E
:i"t.o ! )···~
e- '.· v:
T
l 0.5 . . ; '
4 /:
0.0 '
0.0 1.0 z.o 3.0 4.0 s.o
Dosage (mglg cementl
Figure 11- Adsorption and estimated dosage for surface saturation of PE10 and PE7
on NPCs after 30 minutes adsorption time
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
-
55
50 +----t-t.-----1
!iio;- ·
l4s '•30'
1111160'
~ 40
!o90'
u:: 35
~120'
30
25
PE10 PE7 PE10 PE7
Figure 12- Concrete flow with PEto and PE7. a: W/C 0.39 and 0.18% PEto, 0.21% PE7;
b: W/C 0.36 and 0.25% PE10, 0.53% PE7
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
architectural quality surface. Similarly, the City of Montreal has successfully used SCC
for the rehabilitation of bridge abutments and parapet walls since 2000 (2).
sec used for the repair of concrete infrastructure should have high deformability,
passing ability, and resistance to segregation. The workability characteristics of SCC and
the performance of the hardened concrete, as well as the technical and economical
advantages of using sec can vary with a number of parameters, including mixture
proportioning, chemical admixture combinations, binder type, and fiber reinforcement. In
particular, the high-range water reducing admixture (HRWRA) type is shown to have
significant effect on workability and hardened properties. The use of polycarboxylate
(PCP)-based HRWRA can lead to high early compressive strength. Concrete subjected to
prolonged mixing or agitation after the introduction of PCP-based HRWRA could lead to
increase in air volume, without necessary refinement in the size of the air bubbles.
Polynaphthalene sulfonate (PNS)-based HRWRA is also used in SCC, especially in the
case of ready-mix concrete and in applications, including repair applications.
In order to better comprehend the influence ofHRWRA type on the performance ofSCC
used in cast-in place applications, an experimental work was undertaken to evaluate the
effect of PNS-based and PCP-based HRWRAs on key properties of the plastic and
hardened concrete. The study included the evaluation of HRWRA demand, mixture
robustness, passing ability, stability, mechanical properties, restrained shrinkage,
transport properties, frost durability, and scaling resistance.
EXPERIMENTAL PROGRAM
The SCC mixtures were prepared with wlcm of 0.42, two different types of PCP-based
HRWRAs (PCP! and PCP2), and a PNS-based HRWRA. For each type of HRWRA,
compatible viscosity-enhancing admixture (YEA) and air-entraining admixture (AEA)
supplied by the same admixture manufacturer were incorporated, as presented in Table I.
A lignosulfonate-based water-reducing admixture (WRA) and a set-retarding agent
(SRA) were used in mixtures proportioned with the PNS and PCP2 HRWRAs to enhance
fluidity retention. This was not required for the mixtures made with the PCPI HRWRA,
since the sec exhibited adequate fluidity retention.
Two commercially available blended cements that are often used in repair applications in
Quebec were employed. The Bl cement (CSA Type GUb-F/SF) contains approximately
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
25% Class F fly ash and 5% silica fume, by mass of cementitious materials, and the B2
cement (CSA Type GUb-S/SF) contains approximately 25% slag and 5% silica fume.
Continuously-graded crushed limestone aggregate with a maximum size of 10 mm and
well-graded siliceous sand were employed. The particle-size distribution of the aggregate
lies well within CSA A23.1 recommendations. The coarse aggregate and sand have
fineness moduli of 6.4 and 2.5, bulk specific gravities of 2.73 and 2.64, and absorption
values of0.5% and 1.2%, respectively.
All of the SCC mixtures were prepared with 4 75 kg/m3 of cementitious materials, which
is typically used in SCC for repair applications. The sand-to-coarse aggregate ratio (by
volume) was fixed at 1.0. The HRWRA and AEA concentrations were adjusted to secure
initial target slump flow consistency of 660 ± 20 mm and air volume of 6.5 ± 1.5%.
Three types ofHRWRAs (PNS, PCP!, and PCP2) were used with the Bl blended cement
The PNS and PCP! HRWRAs were incorporated to mixtures prepared with the B2
blended cement. The mixture proportioning is presented in Table 2. The targeted
minimum 56-d compressive strength of the evaluated mixtures is 35 MPa. This could be
easily achieved in air-entrained concrete made with w!cm of 0.42, regardless of the
HRWRA and binder types.
The mortar flow test was used to evaluate the flow characteristics of CEM. The test
consists of determining the variations of fluidity of a given mortar with changes in water-
to-powder ratio (VwNp), by volume. As indicated in Fig. 1, the intercept of the curve
with the ordinates axe (VwNp) and the slope of the curve represent the minimum water
content that is needed to initiate flow and the relative water demand required to increase a
given fluidity, respectively. The reduction of minimum water content corresponds to an
increase in packing density and, for the same cementitious materials, can lead to decrease
in HRWRA demand. As indicated in Fig. 1, for a given increase in water content (ilW), a
CEM that has higher relative water demand would exhibit smaller increase in fluidity
compared to that of similar mixture with lower relative water demand. This indicates that
the former mixture has greater tolerance to undergo same addition of water with limited
increase in fluidity; this would lead to greater robustness.
Workability ofSCC was evaluated using the slump flow, V-funnel, L-box, caisson filling
capacity, and surface settlement tests. Detail descriptions of these tests are elaborated in
reference 4. The apparent yield stress (g) and torque plastic viscosity (h) were evaluated
using a modified two-point workability rheometer. The workability testing was carried
out within 20 min after the end of mixing.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In general, the SCC made with PNS-based HRWRA exhibited higher HRWRA demand
compared to that prepared with PCP-based HRWRA. Similar findings were obtained for
the flow characteristics of CEM, as shown in Fig. 3. For a given HRWRA concentration,
CEM made with PNS-based HRWRA exhibited higher MWC to initiate flow compared
to similar mixture prepared with PCP-based HRWRA (1.23 versus 0.86 and 0.81). This
can be due to lower packing density of the CEM made with PNS-based HRWRA --`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Fresh properties of SCC mixtures are summarized in Table 3. Regardless of the HRWRA
and binder types, the targeted air content of 5% to 8% was achieved because AEA dosage
was adjusted to secure the targeted value. Regardless of the binder type, concrete made
The effect of HRWRA type on surface settlement varied with the binder type. In the case
of SCC made with the B 1 blended cement, the PNS mixture exhibited higher settlement
value of 0.44% compared to 0.16% and 0.21% obtained for SCC prepared with PCP-
based HRWRA, as presented in Fig. 4. This is attributed to the retarding effect of the
PNS incorporated at relatively high concentration. On the other hand, SCC mixtures
made with the B2 blended cement and two types ofHRWRAs exhibited the same surface
settlement of 0.49%. It is important to note that, regardless of the binder and HRWRA
types, maximum settlement values of the evaluated mixtures are lower than 0.5%, which
is recommended by Hwang eta!. (7).
For a given HRWRA type (PNS), SCC made with the B2 binder developed higher 1-d
compressive strength of I0.6 MPa compared to 3.7 MPa for the concrete prepared with
the B I binder. Similarly, the former concrete had higher 28- and 56-d compressive
strength than the SCC made with the B l binder. This can be in part due to the relatively
high reactivity of blast-furnace slag in the B2 binder compared to fly ash in the BI binder
and lower air content of 5.8% for the B2 mixture compared to 7.0% for the SCC
containing the B I binder.
In general, SCC proportioned with PNS-based HRWRA and the Bl binder exhibited
higher drying shrinkage after 56 days of drying compared to similar concrete made with
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
For a given HRWRA type, SCC made with the Bl binder exhibited longer time to
cracking than similar SCC prepared with the B2 binder. For example, the Bl-PNS
mixture developed longer time to cracking of 17.2 days compared to 4.5 days for the B2-
PNS mixture. As presented in Fig. 5, this spread is mainly due to relatively greater tensile
creep coefficient and stress relaxation in tension for the Bl-PNS mixture, thus reducing
tensile stress by restrained shrinkage. On the average, at a given time, the B 1-PNS
mixture had approximately 60% lower tensile stress compared to the B2-PNS mixture.
Similarly, the 81-PCPI mixture exhibited relatively long time to cracking of 6.3 days
compared to 2.8 days for the SCC made with the B2 binder and the same HRWRA
(PCP!).
The results of the transport properties, air-void system, frost durability, and resistance to
de-icing salt scaling are summarized in Table 5. The mean 56-d RCP and chloride-
diffusion coefficient values are compared in Fig. 6. For a given binder type, the SCC
mixtures made with PNS-based HRWRA exhibited slightly lower diffusion coefficients
of2.2 and 2.6 X 10- 12 m2/s compared to the sec made with PCP-based HRWRA that had
values of 2.6 to 3.5 x 10- 12 m2/s. On the other hand, no significant spread was found in
the RCP results of mixtures made with PNS- and PCP-based HRWRA. The B2-PCP1
concrete had the highest diffusion coefficient of 3.5 x 10- 12 m2/s and RCP value of 905
coulomb. The remaining mixtures had 56-d ·ReP values ranging between 505 to 620
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
coulomb.
The results of frost durability and de-icing scaling resistance of the mixtures made with
various HRWRAs and different binder types are compared in Fig. 7. Regardless of the
HRWRA, the optimized SCC exhibited excellent frost durability with durability
coefficient greater than 94%. All mixtures had spacing factor lower than 200 !liD, except
for the B 1-PCP2 concrete that had a spacing factor of 260 !!ill. The optimized SCC
mixtures had excellent resistance to de-icing salt scaling with cumulative mass loss after
50 freeze-thaw cycles less than 80 g/m 2, regardless of the HRWRA type. This is in
exception of the B l-PCP2 concrete that had a scaling value of 375 g/m 2, given its higher
spacing factor value of 260 !liD and lower hardened air content of 3. 7% (Table 5).
scaling residue values are expressed in the opposite direction. This is done to maintain a
representation whereby a broader plot would correspond to better performance.
For a given binder type, SCC made with PNS-based HRWRA exhibited relatively higher
"closed-loop area" values of 15,660 and 17,600 compared to 8,720 to 14,600 for similar
concrete prepared with PCP-based HRWRA, indicating superior performance of the
former set of mixtures. The main difference between the performance of the hardened
SCC made with PNS- and PCP-based HRWRAs lies in the time to cracking by restrained
shrinkage that was longer for the forrner set of mixtures compared to similar mixtures
made with PCP-based HRWRA. As presented in Fig. 8, the SCC mixtures made with
PNS-based HRWRA had relatively longer time to cracking of 4.2 and 17.2 days
compared to 2.8 to 8.8 days in the case of similar SCC mixtures prepared with PCP-based
HRWRA. This is despite the relatively lower compressive strength and higher drying
shrinkage of the SCC made with PNS-based HRWRA.
For a given HRWRA type, SCC proportioned with the B2 binder exhibited higher
compressive strength, lower drying shrinkage, higher frost durability, and higher
resistance to de-icing salt scaling. As presented in Fig. 8, mixtures proportioned with the
B2 binder and the PNS- and PCP-based HRWRAs exhibited greater "closed-loop area"
values of 17,600 and 14,600, respectively, compared to 15,660 and 8,720 for similar
mixtures made with the Bl binder and the same HRWRA. SCC proportioned with Bl
binder exhibited higher resistance to restrained shrinkage cracking than similar concrete
made with B2 binder.
CONCLUSIONS
Given the testing program presented in this paper to evaluate the performance of SCC
mixtures for repair of concrete infrastructures that were proportioned with 0.42 wlcm and
two different types of ternary cements, the following conclusions can be made:
I. CEM and SCC mixtures proportioned with PNS-based HRWRA necessitated higher
values of minimum water content to initiate flow and higher HRWRA demand,
respectively, compared to mixtures prepared with PCP-based HRWRA. The CEM
proportioned with PNS-based HR WRA exhibited higher relative water demand
2. For the selected binder type (B 1 binder), the SCC proportioned with PNS-based
HRWRA had greater maximum settlement values of 0.44% compared to 0.16% to
0.44% obtained in the case of similar SCC made with PCP-based HRWRA.
slightly higher drying shrinkage compared to similar concrete made with PCP-based
HRWRA and the same binder (660 J..!Strains versus 555 to 615 J..!Strains). However,
the PNS mixture developed longer time before cracking due to restrained shrinkage
of 17.2 days compared to 6.3 to 8.8 days for similar SCC prepared with PCP-based
HRWRA.
4. The HRWRA type does not seem to have any significant effect on transport
properties, frost durability, and de-icing salt scaling resistance for the optimized sec
made with the same cement and wlcm. Optimized SCC mixtures exhibited 56-d RCP
values lower than 900 coulomb, frost durability coefficients higher than 90%, and
cumulative mass of scaling residues lower than 400 g/m2, regardless of the HRWRA
and binder type.
ACKNOWLEDGEMENTS
The authors wish to thank the Natural Sciences and Engineering Research Council of
Canada (NSERC) as well as the industrial partners participating in the NSERC-Industry
Coop project involving the development of high-performance SCC for repair
applications: Axim, Chryso, Ciment Quebec, City of Montreal, De gus sa, Euclid Canada,
Handy Chemicals, Lafarge Canada, Ministry of Transport of Quebec, St-Laurence
Cement, and W.R. Grace. The assistance of Olivier Bonneau and Boualem Abdi in
conducting the experimental work is especially acknowledged.
REFERENCES
1. Hovington, A., "The Experience of the Quebec Ministry of Transports with Self-
Consolidating Concrete (1997-2000)," (in French), Proceedings, Annual ACI
Quebec and Eastern Ontario Annual Meeting, Quebec, Canada, 2000, 8 p.
2. Morin, R., and Khayat, K.H., "Performance of Self-Consolidating Concrete Used to
Repair Parapet Wall in Montreal," Proceedings, 1st North American Conference on
the Design and Use of Self-Consolidating Concrete, Chicago, USA, 2002, pp. 475-
481.
3. Khayat, K.H., Petrov, N., Morin, R., and Thibault, M., "Rehabilitation Strategies
and Material Performance of SCC Used for the Repair of the Jarry/Querbes
Underpass in Montreal," Proceeding, International Conference on Concrete Repair,
Rehabilitation, and Retrofitting, Ed. by Alexander, Capetown, South Africa, 2006,
pp. 1009-1013.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Table 1 -Characteristics of chemical admixtures supplied by three different
suppliers
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
475 5.8 3.3 350
B2-PNS' 200 811 785
(B2) (PNS) (VEA3) (AEA3)
475 1.7 2.0 45
B2-PCPI 200 811 785
(B2) (PCP!) (VEAl) (AEAI)
475 0.2%···
CEM-PNS
(Bl)
200 - 828
(PNS)
- -
475 0.2%
CEM-PCPl
(Bl)
200 - 828
(PCP!) - -
475 0.2%
CEM-PCP2
(Bl)
200 - 828
(PCP2)
- -
' Bl-PNS and B2-PNS were incorporated with a water-reducing admixture
" B t-PCP2 was proportioned with a set-retarding admixture
'" 0.2% solid content by mass of cemeotitious materials
Table 3 -Fresh properties of SCC made witb various HRWRA and binder types
wlcm=0.42
Bl-PNS Bl-PCPl B1-PCP2 B2-PNS B2-PCP1
HRWRA demand, Lfm3 7.0 2.9 3.8 5.8 1.7
!Omin 650 670 665 675 670
Slump flow, mm
70min 600 640 650 635 630
!Omin 94 88 91 94 93
Filling capacity,%
70min 85 94 96 82 83
!Omin 4.0 2.7 3.0 4.4 3.1
V-funnel flow time, sec
70min 4.3 2.8 3.2 4.4 4.0
L-box blocking ratio, !Omin 0.83 0.78 0.80 0.82 0.75
hv'b, 70min 0.72 0.87 0.74 0.73 0.67
10 min 0.6 0.4 0.2 0.4 0.7
1i g,N.m
70min 0.6 0.4 0.3 0.6 0.9
i!
~""
h,N.m.s
!Omin
70min
6.0
7.3
4.0
4.6
5.2
4.9
8.1
10.6
4.6
7.0
10min 6.2 7.2 6.4 6.8 7.1
Air volume, %
70min 5.0 7.0 5.7 5.0 7.4
Max. surface settlement, % 0.44 0.21 0.16 0.49 0.49
Rate of settlement at 30 min, o/ofh 0.14 0.08 0.07 0.15 0.16
Initial setting time, hour 13.5 9.2 8.7 9.7 8.0
Final setting time, hour 15.0 11.0 10.0 11.0 10.3
Table 5- Transport properties, air-void system, frost durability, and de-icing salt
scaling of SCC made with various HRWRA and binder types (at 56 days)
Relative flow
3DD m m
blood base
coven=d w:ilh
325 am tafbn sheet
~DD mm
~
~ 1.2
l8
t 0.8
~
§
.§ 0.4 0.05 0.04 1
~ ~
0,6
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
D-3
0 Maximum settlement ;?
• Rate of settle meat ~
~
~
0.5
0.49 0.49 e..
e.. 0.44 .se;
5 OA O.i' 0
""'';!
f~
0.3
0.21
5
§
.§
1.1
0.2 Dol
i
....
;:;: 0.1 0
~
,.:
Fig. 4- Comparison of static stability of sec made with various HRWRA and
binder types
.soil 12
llJiLI I D
B
4.5
,..."
.§ 4 2.8
0 D
B1-PNS B1·PCP1 B1·PCP2 B2·PNS B2-PCP1
:2 1000 =RCP
e ~Diffusion coefficient 3.5
..!2
0=
-;;-
~ 800 N-s
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
3 N
$
.g
g
600
5" 2
5
·a
!E...
"'" 400
8
.,·.:-~
l ·~"
..!2 200 @
"fl iS
:s!
§- 620 505 575 615 905
~
B1-PNS B1-PCP1 B1- PCP2 B2·PNS B2-PCP1
Fig. 6- Diffusion coefficient and RCP values of sec made with various HRWRA and
binder types determined at s6 days
98
~...
~
400
~
·ail
96
=
"0
..
300
-~
!50 94
"
-~
:;::
-e 92
200
1
.....
0
0= lDD ~
90 ::E
88
B1-PNS B1-PCP1 81- PCP2 B2· PNS B2-PCP1
Fig. 7- Frost durability coefficient and scaling resistance of sec made with various
HRWRA and binder types
80 660 35 560
29.5
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
2.6
30.0
2.8
905
3.4 3.5
Mus of Shrinkage
Rcr n .....
@56d
575 cracking, let
DiiTWJioD
2.6 coefficient@ 56 d
Synopsis: Pure C A hydrates under formation of layered phases (e.g. C AH, and
3 4 9
C2AH 8) which quickly convert to the cubic phase C3AH 6 • In presence of polycarboxylate
(PC) superplasticizers during C A hydration, intercalation of the PC molecule into
the layered calcium aluminum hydrate occurs. The resulting organa mineral phases
Ca-Al-PC-LDHs are very stable. They were characterized by powder XRD, elemental
analysis, IR spectroscopy and TEM microscopy. XRD patterns show that with
increasing side chain length (nw) of PC, the distance (basal spacing) between the
[Ca 2AI(OH) 6]+ layers increases. PCs with side chain lengths up to 45 ethylene oxide
(EO) units intercalate well whereas a PC with 111 EO units no longer intercalates.
Calculations indicate that the PC side chains mostly are compressed or tilted to fit in
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Dr. Zhimin Dai is a post-doc researcher at the Chair for Construction Chemicals in
Munich, Germany. Previously, she worked at the State Key Laboratory for Inorganic
Chemistry at Jilin University in China. Her research focused on the synthesis of hybrid
organic-inorganic materials.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Dr. Dorina Vlad studied chemistry and physics at the University in Bucharest, Romania,
before joining the Chair for Construction Chemicals in Munich, Germany. In her Ph. D.
thesis, she studied the sulfate effect on polycarboxylate superplasticizers and their
interaction with binder surfaces.
INTRODUCTION
Hydration of the pure aluminate phase (C3A) first produces minerals with layered
structure, e.g. C4AH 19 and C2AH 8 [8]. At room temperature, these metastable phases
quickly transform into the more stable cubic phase C3AH 6 • Under heat of hydration from
C3A, the conversion is almost instantaneous. The structure of the layered phases derives
from brucite Mg(OH 2), in which some Mg 2+ cations are replaced by a trivalent cation.
with Mil= Ca 2+ Mg2+ Zn 2+ Co 2+ Ni2+ Cu2+ Mn 2+. M m = Ae+ Cr3+ Fe3+ Co3+ N?+
Mn3+; X = inte~l~yer 'anion: e.g. 'oR", 'cr, N0 3-, Clo:-, cot, 'so/ and ~any 'other
anions.
Several authors have shown that not only inorganic, but also organic guest anions can
intercalate into the layered calcium aluminate structure. Meyn et al. [9], for example,
confirmed the intercalation of citrate, diglycolate, phthalate and aryl/alkyl sulfonates such
as dodecylbenzenesulfonate. Others showed that also polymers such as puly(ethylene
oxide) derivatives [10], poly(u,~-aspartate) (11], poly(acrylic acid), poly(vinylsulfonate)
and poly(styrene sulfonate) [12] intercalate into LDH phases. With respect to concrete
applications, Fernon et al. (7] demonstrated the successful intercalation of BNS polymer
into the Ca-Al-LDH structure. These articles generally show that the steric size of the
intercalated anion determines the interlayer distance as expressed by the gallery height or
the basal spacing (see Fig. 1). Sterically very large anions are more difficult to
intercalate.
between the polycarboxylate and the sot anion which supposedly has a high tendency
to intercalate, pure C 3A clinker without sulfate was used. Several synthetic routes for the
incorporation of PC were studied, among them direct anion-exchange with Ca-AI-OH-
LDH as precursor, coprecipitation of calcium and aluminum nitrate in aqueous PC
solution at pH value of 11.7, and rehydration of calcined C3A in aqueous PC solution.
The rehydration method produced particularly good results and was selected for all
syntheses. Polycarboxylates with increasing poly( ethylene oxide) side chain length (nEO =
4.5 to 45) were used to study the steric influence of molecule size.
Tricalcium aluminate was prepared via a sol-gel process as precursor before calcination.
Aqueous colloidal alumina (VP disp. W630, Degussa, Hanau, Germany) containing 29.7
wt.-% Alz03 with a medium particle diameter of about 200 nm was mixed with
stoichiometric amounts of neutral Ca(N0 3)z solution, which was freshly prepared from
CaC0 3 and HN0 3 (p.a. Merck, Darmstadt, Germany). The mixture was heated at 70°C
and stirred at pH 8 for 10 to 12 h until gel formation. The gel was transferred into
platinum crucibles and slowly heated to 900°C. Water evaporated and Ca(N03) 2
decomposed to the metal oxide. Finally, the sample was heated at l350°C for 9 h with
two intermediate grindings. Investigations by X-Ray powder diffraction confirmed the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
resulting product to be pure tricalcium aluminate.
Preparation of Ca-AI-PC-LDH
A typical synthesis of a PCn-LDH (n denotes the number of EO units in the side chain) is
described as follows: polymer PCn=1.s (Mn=14, 000 Da; Mw=67,460 Da) was dissolved in
100 ml of deionized water at a concentration of 2.5 wt. %. 1.5 g of C3A was added to the
PC solution (pH ~ 12). The system was stirred under argon flow at 75°C for 48 h. The
precipitated intercalate was separated by centrifugation and washed with deionized water
in order to remove residual polymer. The resulting solids were dried at 65°C under high
vacuum (10-2 ~ 10-3 mbar) for 12 hours. Similar routes were used to prepare the other
three PC-LDHs, the number of EO units being 8.5, 17 and 45. Ca-Al-OH-LDH was
prepared by reacting monocalcium aluminate (CA) in deionized water [13].
Compositions of the organomineral compounds were determined by X-ray powder
diffraction, elemental analysis, IR spectroscopy and transmission electron microscopy
(TEM).
Properties of PC superplasticizers
From the data of Table 2, a schematic of the molecular size and conformation of the PC
copolymers was developed (see Fig. 3). It shows that 4.5PC and 8.5PC are worm
molecules while 17PC and 45PC are polymer brushes. It also gives one the idea that PC
molecules with longer side chain will not easily intercalate into Ca-Al-LDH.
Tn spite of the large size of the PC molecules, all synthesized copolymers intercalated
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
easily into the lamellar calcium aluminum hydroxide structure. Incorporation was evident
from X-ray powder diffractograms developed on Bruker axs D9 instrument. The basal
spacing for each intercalate was taken from the (001) reflections and is shown in Fig. 4.
As can be seen, the increasing side chain length of PC as expressed by increasing nw
causes a considerable widening of the inter layer distance. The basal spacing d increases
from 1.15 nm for 4.5PC-Ca-Al-LDH to 4.27 nm for 45PC-Ca-Al-LDH (see Table 3). In
comparison, Femon et al. [7) report a gallery height of approx. 1.5 nm for the BNS-Ca-
Al-LDH.
Infrared Spectroscopy
IR spectra of all synthesized intercalates were taken between 4000 and 400 cm· 1 (see
Fig. 5).
At 3600 cm- 1, all spectra show valence vibrations for the -OH groups in the cationic
inorganic layers. The absorption of the interlayer water appears around 3400 em·I. The
asymmetric and symmetric valence vibrations of the carboxylate group appear around
1560 and 1455 cm· 1• They confirm the presence ofpolycarboxylates in the mineral phase.
Elemental Analysis
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
organa-mineral phases as welL The results are shown in Table 4.
As can be seen, the carbon and hydrogen content increases with increasing side chain
length of the polycarboxylate. For PC with nEO = 17 and 45, the organic part in the
organo mineral phase exceeds 50%.
Although some PC adsorption on the cationic surface of the top and bottom layer of an
Ca-Al-LDH particle can be expected, an organic content of between 30 and 50 % is too
high for only adsorption. Thus, it confirms that intercalation has occurred. Based on the
analytical data, the approximate composition for each Ca-Al-PC-LDH was calculated and
is shown in Table 4.
The layered structure of the PC intercalates also was con finned by TEM. As an example,
the layer stackings for Ca-Al-45PC-LDH is shown in Fig. 6. As expected, it shows a
parallel stacking of layers with a basal spacing of approx. 4 nm corresponding to the
basal spacing of 4.27 nm obtained from XRD analysis.
Surprisingly, intercalation of large PC molecules into the layered C3A hydrate structure
occurred easily and completely. When repeating the intercalation experiment in air, i.e.
without argon protection and in presence of C0 2, the same intercalates were obtained.
This is remarkable because it is well known that C0 32- has an extremely strong tendency
to intercalate. It prevents most other anions from being intercalated [14]. This experiment
confirms the high tendency of PCs to intercalate into C3A hydrate phases. In our
experiments, the only limiting factor we found to prevent intercalation was the length of
the PC side chain: Whereas 45PC with a side chain length of 45 EO units intercalates
well, lllPC shows very little intercalation. It can be concluded that there is an upper
limit for the gallery height between the [Ca2Al(OH) 6t main layers. It is determined by
the steric size of the ethylene oxide side chain.
To investigate on the steric arrangement of intercalated PC, for each intercalate the
distance (gallery height) between two neighboring [Ca 2Al(OH)6t main layers was
determined. The gallery heights are obtained by subtracting the thickness of the cationic
layer which is 0.48 nm (see Fig. I) from the basal spacing d obtained by XRD (see Fig.
Conclusion
Most remarkable is the stability of the organa mineral phases containing PC. Whereas
layered Ca-Al-OH-LDH quickly converts to the more stable cubic phase C3AH6 , the PC
intercalated phases maintain their lamellar structure once they are formed. Based upon
these results, the possibility of PC intercalation into C3A hydrate phases also in the
presence of sulfate which occurs in cement pore solution, should be considered.
According to a model developed by Flatt et al. [15], such intercalation might explain the
differences in concrete flowability observed for some PCs upon early and delayed
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
addition to the concrete.
ACKNOWLEDGEMENT
The authors wish to express their appreciation to following institutions for generous
support of this work: "Jtirgen Manchot-Stiftung", "Deutsche Forschungsgemeinschaft
(DFG)" and "Friedrich Schiedel-Stiftung". Dr. Marianne Hanzlik from TU Munich is
thanked for taking the TEM photos.
References
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Double Hydroxides", Inorg. Chern. 29, 1990, 5201-5207
10 F. Leroux, P. Aranda, J.-P. Besse, E. Ruiz-Hitzky, "Intercalation of
Poly(Ethylene Oxide) Derivatives into Layered Double Hydroxides", Eur. J.
Inorg. Chern. 6, 2003, 1242-1251
11 N.T. Whilton, P.J.Vickers, S. J. Mann, "Bioinorganic clays: synthesis and
characterization of amino- and polyamino acid intercalated layered double
hydroxides", J. Mater. Chern. 7,1997,1623-1629
12 C.O. Oriakhi, I.V. Farr, M.M. Lerner, "Incorporation of poly(acrylic acid),
poly(vinylsuljonate) and poly(styrenesuljonate) within layered double
hydroxides"
J. Mater. Chern. 6, 1996, 103-107.
13 P. Faucon, T. Charpentier, D. Bertrandie, A. Nonat, J. Virlet, J. C. Petit,
"Characterization of Calcium Aluminate Hydrates and Related Hydrates of
Cement Pastes by 27Al MQ-MAS NMR", Inorg. Chern. 37, 1998,3726-3733
14 C. Delmas, Y. Borthomieu, "Chimie Douce Reactions: A New Route to Obtain
Well Crystallized Layer Double Hydroxides", J. Solid State Chern. 1993, 104,
345-352.
15 R.J. Flatt, Y.F. Houst, "A simplified view on chemical effects perturbing the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
...0N
4.5 PC 6.5 PC
Main chain: 3:1.:3 nm Main chain: 84.1 nm
Side Chain: 1.2 nm Side chain: 2.4 nm
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
17 PC 45 PC
Main chain: 34.4 nm Mam chain· 27.5nm
Side chain: 4.6 nm Side chain: 12.6 nm
doo1.03rm
0 10 Ill 30 10 10
2 Theta-scale
a) 8.5 PC
11111111111 II 1111
or
b)4.5 PC, 17 PC or 45 PC
JJIIIIIIIIIIIIIIIIII
or
or
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
In recent years advanced superplasticizers have been developed in order to meet the
increasing performances required by the application. The new technology is essentially
based on polycarboxylate-type admixtures, which allow to reach outstanding
performances, both for fresh and hardened concrete, in terms of high fluidity, high
workability retention, very low water to cement ratio, high strength and high durability.
But if on one hand this new generation of superplasticizers is able to give such properties
to the concrete, on the other hand it shows a very high sensitivity towards cement
composition (mainly C3A and alkali content), cement fineness, amount and type of added
calcium sulfates, SP dosage and addition method, mix design, mixing procedure for
concrete preparation, etc [ l-4].
Among these parameters, many studies have been carried out in the last years about the
influence of sulfates (amount and type) on the performances of superplasticizers,
recognizing them as key factors for understanding the reasons of possible problems of
incompatibility between cement and admixtures [5-15]. Especially soluble sulfates (alkali
sulfates) are believed to play the main role in the interaction with superplasticizers.
According to earlier studies, the behaviors of polynaphtalenesulphonate-type (PNS)
superplasticizers and polycarboxylate-type (PCE) superplasticizers are significantly
different. Regarding PNS, Jiang et al. showed that an optimum soluble alkali content
exists with respect to the fluidity and fluidity loss, which was found to be 0.4-0.5% Na20
equivalent and which is independent of the superplasticizer dosage and C 3A content: a
lack in the soluble sulfates content (in the form alkali sulfates of the clinker) causes
problems of incompatibility between cement and NS-admixture, that can be overcome by
adding the right amount of soluble sulfates. This addition leads to significant increases in
fluidity and fluidity retention[5]. The same conclusions were drawn by Li eta!. [6).
Nakajima and Yamada examined the influence of the kind of calcium sulfates on the
dispersing ability of PNS, concluding that the kind of calcium sulfates influences flow
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
loss of cement paste, which is lower when the cement used contains a higher hemihydrate
to gypsum ratio [7].
The first studies regarding the influence of sulfates on the performances of
polycarboxylate-type superplasticizers were carried out by many researchers in the 1998
- 2003 period [8-14]. Results of these works show that the sulfate ion concentration in
cement paste affects the adsorption behavior of PC on cement particles: the adsorption of
EXPERIMENTAL DETAILS
Materials
Four Italian commercial portland cements type I 52,5 (according to EN 197/l) and one
portland cement type I 42,5 were investigated, which will be named in the following as
cement A, B, C, D and E respectively. Their complete chemical analysis was carried out
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
according to the methods reported in EN 196/2. Total and soluble sulfates (within 2
minutes at w/c of 0.5) were determined by means of ionic chromatography. Three
different superplasticizers were investigated, one based on ~-naphtalensulphonate (PNS)
and two based on ether polycarboxylates (PCE-1 and PCE-2). All the superplasticizers
are water solutions with 38% solid content, pH 7.5 and density 1.195±0.05 g/cm3 for
BNS, 35% solid content, pH 7 and density 1.095±0.05 g/cm 3 for PCE-1 and 30% solid
content, pH 7 and 1.060±0.05 g/cm3 for PCE-2. PCE-1 and PCE-2 are characterized by
the same chemistry but different side chain length (PCE-1 has shorter side chains than
PCE-2) and were chosen to study the possible influence of the side chain length on the
interaction with sufates.
Minislump tests
A cement paste prepared with cement, water and superplasticizer (dosage 0.40% active
bwc for PNS, 0.20% active bwc for PCE-1 and PCE-2) was poured, after mixing, into a
metallic truncated cone mould (internal top diameter: 21 mm, external top diameter: 31
Adsorption Analysis
The degree of adsorption onto cement of the ~-naphtalensulphonate-based and
polycarboxylate-based admixture was determined in cement paste by T.O.C. analysis
performed with a Shimadzu 5000-A. BNS was dosed at 0.40% active bwc, while PCE-1
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
and PCE-2 were dosed at 0.20% active bwc. The w/c was kept constant at 0.5. The
samples were prepared according to the following procedure: powder and water,
containing the admixture, were stirred for the defined time and then filtered by a vacuum
pump. The amount of polymer adsorbed was determined by difference between the
organic carbon originally contained in the polymer solution and the amount left in the
solution filtered from the paste.
Mortars
Standard mortars were prepared according to EN 196/1 with all the five cements studied,
in the presence either of PNS and PCE superplasticizers. PNS was dosed at 0.40% active
bwc, PCE-1 and PCE-2 at 0.20% active bwc. The flow was evaluated for each mortar
according to the EN 413-2 after 15 shocks.
Concrete tests
Concretes were prepared with cements A and B at 20°C, adopting the mixture
composition shown in Table 1.
The dosages of the superplasticizers were 0.40% active by weight cement for PNS,
0.20% active bwc for PCE-1 and PCE-2. A second series of concrete mixes was carried
out, only with cement B, adding the amount of S04=necessary to bring the soluble sulfate
to the level of sample A (0,77% So 4= added as K2S0 4). The concrete composition was
the same as reported in Table I.
Chemical Composition
Table 2 shows the chemical composition of the five cement samples, obtained according
to the methods reported in EN 196/2. The most relevant differences between the five
samples is seen in the soluble sulfates contents: sample A contains the highest amount of
soluble sulfates, while sample B the smallest ones. Sulfate contents of sample C, D and E
are between those of samples A and B. It is interesting to point out that the total sulfate
contents of sample A and B are not so different. As the soluble sulfates are taken after
two minutes of hydration, it is reasonable to suppose that most part of these sulfates are
in the form of alkali sulfates, which are the most soluble.
Table 3 shows the phase composition of the five samples, obtained by means of the
Bogue equation: cement D is characterized by the highest C3A content (~ 11%).
Adsorption tests
The results of the adsorption tests are shown in Fig. 2, 3 and 4.
There is a quite good correlation between the adsorption values and the soluble sulfates
content in the case of PCE: the admixtures are in fact adsorbed in higher amount on
cements that present the lower soluble sulfates content. This seems to confirm the fact
that there is a competition between sulfates and polymer for the adsorption onto cement
particles. In the case of PNS, the adsorption is very high but very similar for all the
cements, meaning that the sulfates content does not seem to be a key parameter in the
interaction between cement and PNS-type admixtures.
The adsorption generally remains constant over time for all the cements and all the
admixtures.
Regarding polycarboxylates, PCE-2 shows the widest range of adsorption, from 30 to
80%, but it is adsorbed on all the cements to a lower extent with respect to PCE-1, which
had shown a better water reduction capability in cement paste. Cement E always shows
the lowest adsorption values, due to its lower specific surface area.
Mortar tests
The results of mortar tests are presented in Fig. 5-8. Table 5 shows the water reduction
capability of every admixture, and Table 6 the corresponding spreading flows. Regarding
the water reduction of polycarboxylates, in this case also, a quite good relation between
water reduction and soluble sulfate content can be found: cements B and C, which
contain the lowest amounts of soluble sulfates show the highest percentage of water
reduction. This behavior is also in accordance with the results of the adsorption tests.
Looking at the workability retention, some considerations can be drawn from the analysis
of the results:
regarding PCE-1, the workability retention seems to improve for the cement
with the highest content of soluble sulfates. Mortars made with cement B and E
contain the same amount of water, but the workability retention is very different,
as cement B, which contains the lowest amount of soluble sulfates, shows a
.workability reduction within the first 30 minutes. Cement A, which contains the
highest amount of soluble sulfates, retains workability better than cement D,
although the water content of the mortar is lower;
also in the case of PCE-2 a high content of soluble sulfates seems to improve
the workability retention. Mortar made with cement B looses workability very
quickly with respect to the one made with cement C, which contains the same
amount of water but a higher amount of soluble sulfates;
regarding PNS, no clear correlation was found between the soluble sulfates
content and the workability retention, although mortar made with cement B,
which contains the lowest amount of soluble sulfates and the highest amount of
water, shows a reduction in the workability within the first minutes.
A possible explanation for the better workability retention could be the fact that a higher
amount of superplasticizer remains in the pore solution, thus acting as a dispersant for a
longer time.
Concrete tests
The results of concrete tests are presented in Tables 7, 8 and 9. Figure 9 shows the
compressive strengths of all the trials. A very good correlation was found between the
water reduction and the soluble sulfates content, as the water reduction obtained with
cement B, which contains the lowest amount of soluble sulfates is considerably higher
than the one obtained with cement A.
According to the mortar tests, PCE-1 seems to perform better than PCE-2 with cement A,
while the performances of the two PCE admixtures are the same with cement B. No clear
correlation was found between PNS performances in mortar and in concrete tests, as the
water reduction capability of PNS is higher with cement A than with cement B in mortar
and lower with cement A than with cement B in concrete. Since the water content of the
concrete mixtures made with cement A and B is so different, the workability retention
was not recorded. The compressive strengths of concrete made with cement B, although
the water content of concretes made with PCE-1 and PCE-2 is the same, is higher at all
ages.
Further concrete tests were made with cement B adding the amount of soluble sulfates
(0, 77% of S04 ~ as K 2S0 4} necessary to bring the soluble sulfate to the level of sample A.
Results are shown in Table I 0.
CONCLUDING REMARKS
Among the chemical parameters that characterize the analysis of cement, the amount and
type of sulfates seem to be the most important to understand the interaction between
cement and admixtures. In particular, the soluble sulfates content (alkali sulfates in the
clinker phase) plays a crucial role in determining the performances of superplasticizers.
Polycarboxylate-type superplasticizers are proved to be very sensitive to the presence of
soluble sulfates, which compete with polymers for the adsorption on cement particles. It
was demonstrated that high amounts of soluble sulfates lead to low water reduction
capability and generally to a better workability retention, at least in mortar, as the
superplasticizer that remains in the pore solution can act as a dispersant for a longer time.
No clear trend was detected regarding the influence of the chemical structure (side chain
length) on the interaction with sulfates. PNS-type superplasticizers seem to be much less
sensitive to the soluble sulfates content, as no clear correlation between water reduction,
workability retention and sulfates content was found.
Of course, there are several other chemical parameters, such as C 3A content and type,
that have to be taken into consideration for a deeper understanding of the mechanism of
action of superplasticizers.
REFERENCES
CEMA CEMB
Table 10- Concrete tests with CEMB and addition of 0.77% S04-
Admil:ture dosage Water Initial slump
Admixture W/C
(active%) (Vm) (mm)
244 0.61 200
PCE-1 0.20 192 0.48 210
PCE-2 0.20 192 0.48 200
PNS 0.40 !96 0.49 Selflevelling
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
0,9
0,79 0,80
0,8
0,7
0,6
~ 05
3:: '
0,4
0,3
0,2
0,1
0
CEMA CEMB CEMC CEMD CEME
100
90
80
CEMB ---:::::::: -
~
70 ~
~
II:
w
::1! 60 ~
~
0
a. 50
~
cw
Ill
II: 40
0
fll
~ 30
20
10 --
0
0 10 20 30 40 50 60
TIME(MIN)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
90
~
~
I1J
...~
...0
0
I1J
80
70
60
50
---
CEMB
CEMD
----
Ill .,~EMA
~ 40
0
rJ)
-
Q CEME
c 30
20
10
10 20 30 40 50 60
TIME(MINI
100
90
"'"ceMo
80
CEME
70
~
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~
I1J
IE 60
...>-
fl 50
Q
I1J
Ill
~ 40
0
rJ)
Q
c 30 --------
20
10
0 10 20 30 40 50 60
TIME(MINI
0,9
0,8
0,7
0,64
0,6
(J
3: 0,5
0,4
0,3
0,2
0,1
0
CEMA CEMB CEMC CEMD CEME
i•Piain.PCE1 DPCE2DPNS\
140r-------------------------------------------------------~
20
10 20 30 40 50 60
TIME(MIN)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
80
~
~
....... 80
40
CEMB
20
0
0 10 20 30 40 50 80
TIMECMIN)
140
120
100
80
~
~....
... 60
40
20
10 20 30 40 50 60
TIME(MIN)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Dr. R. Magarotto, graduated in Industrial Chemistry, Venice University, Italy, 1993. She
has been working in the field of concrete admixtures, focusing on the development of
new polymeric superplasticisers. At present she is Development Head for the Business
Line Admixture Systems at Degussa Construction Chemicals in Treviso (Italy).
INTRODUCTION
Cementitious materials have been in use for construction for about 9000 years. They have
undergone numerous changes in composition, performance and applications. Depletion of
natural resources has led to an increased concern for sustainability in the construction
industry. It has been estimated that some 8% to I 0% of carbon dioxide emitted into the
atmosphere is traceable to the construction industry (cement and steel production,
aggregate processing, mixing, transporting and placing the concrete, etc.). Major steps
have been taken by cement producers to utilise secondary or supplementary materials in
cement production. Some of these steps include:
• Use of secondary fuels such as used tyres, plastic bottles, oil industry residues,
(petroleum coke), paints and solvents, oily rags; just to cite a few.
• Use of supplementary cementitious materials such as fly ash, natural pozzolan, silica
fume and blast furnace slag.
• Limestone blended cements such as type L or LL with limestone content between
6% and 35%; EN 197-1[1].
In order to reduce or eliminate the discharge of the cement kiln dust (CKD) in to the
atmosphere or in landfills, CKD is recycled back into the cement, thus increasing the
presence of unknown or unfamiliar substances in to the concrete mixture. The Portland
Cement Association of the USA sums this up very nicely with the question:
"What has the cement industry done to improve its manufacturing process?"
The answer is "New technology and equipment, coupled with increased use of alternative
fuels and raw materials, have reduced energy consumption by one-third since 1975.
Today, cement and concrete can contain recycled materials that utilize industrial by
products that would otherwise clog landfills. The industry has reduced its own waste by
recycling more than 75 percent of cement kiln dust (nearly eight million tons each year)
directly back into the cement kiln as raw material. By doing so, manufacturers conserve
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The situation in Europe is more or less similar. The need and the importance of
sustainable construction has been recognized the building industry, especially regarding
energy consumption and the use of natural resources. A European set of guidelines has
been set up with more than 200 recommendations, subdivided in to 16 categories. These
recommendations not only cover environmentally friendly constructions but, also cover
social and economic aspects. The most important benefit will be of help, in the very near
future, to include sustainable construction practices [2]. Therefore, for environmental
considerations the concrete industry will continue to increase the use of recycled and by-
product materials, originating within the industry and from other sources, in concrete
production. It is foreseen that by 2030, the concrete industry will achieve zero net waste
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
in its production process. Fig. 1 shows the trend towards the increase of secondary fuels
in cement production, and Fig. 2 indicates the amount of supplementary cementitious
materials that will have to be used in concrete, perhaps as cement replacement, in order to
reach zero emission of carbon dioxide from cement production.
The continuous demand for extreme performances of admixtures has led to an excellence
in tailor made solutions which lead to an economy for the state of the art applications in
concrete. This requires a:
• New understanding of chemical-cement interaction of the polymer
• Better knowledge of relationship between chemical structure of the polymer and its
performances in cement base systems
• Synthesis of tailor made polymers
• Extended competence from organic chemistry to inorganic chemistry of cement and
concrete
• Solid connection between the market and R&D
Recent trends in admixtures science indicate that the ready-mixed concrete industry
demands that the admixtures provide the following characteristics:
• Stronger water reduction for improved durability
• Faster hydration rate for rapid strength development and construction
• Longer workability retention time for reliable transportation and placing
• Rheology control of concrete for easy form filling
The challenge for the admixtures science and technology is to propose a valid solution for
the above stated requirements. To meet all these properties simultaneously is an arduous
task. Strong water reduction leads cement particles to be very close to one another and
workability is lost very quickly. Long workability retention requires extra water to be
added to the concrete mixture.
Nanotechnology has brought new light in to the admixture science. It is now possible that
chemicals or polymers can be engineered to bring together functional groups aimed at
targeted performance. Polymers are built for strong or weak adsorption onto the cement
particles for dispersion effectiveness and for hydration control p J
The Total Performance Control (TPC) concept ensures that the concrete producers,
contractors and the owner get a concrete that is of the same high quality as originally
specified by the design engineer; starting from the production at the central hatching
plant to the delivery and placing into the forms and followed by its hardening. Through
TPC I it is possible to regulate the:
• Rheology of the fresh concrete
• Placing
• Reliable form filling
• Optimal slump retention
and through the water cement ratio reduction it is possible to regulate the:
• Strengths development
• Durability
• Engineering properties
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The positioning of such a superplasticiser (TPC) in comparison with a state of the art
high-range water reducing superplasticiser (HRWR SP) and long slump retention
superplasticiser (LSR SP) are synthesised in Fig. 3.
In actual practice it has been observed that polycarboxylate ether (PCE) superplasticisers
show different performances with different cements and materials comprising the
concrete. The differences may be in the water reduction or workability retention. The
incompatibilities that interfere with the performance are traceable to alternative fuels
used in cement production, secondary components in cement and concrete production,
mineralogical and chemical composition of aggregate and the chemical nature of the PCE
superplasticisers.
The consequences of secondary fuels usage in clinker production affecting fluctuation of
performances are the degree of sulphatization which affects the kinetics of ettringite
Cement is very often blended with secondary components such as slag, pozzolan,
limestone, fly ash, silica fume, etc. In some cases these materials are added to cement
during the grinding and in other cases in to the concrete during the mixing. Very often,
secondary materials are not inert materials. In many cases they interact vigorously with
admixtures. Admixtures are designed to act with cement, but in many instances they
interact with secondary materials too and are also adsorbed on their surface. This
behaviour subtracts active ingredients from being effective in water reduction or
workability retention. Some crushed limestone aggregate, especially sand, slags and
pozzolans show absorption up to 80% of the active polymer of a superplasticiser. The
adsorption of a PCE polymer on eight different limestone and quartz sands, two types of
fly ashes and five types of limestone fillers are shown in Fig. 4 to 6. The chemical
analysis of the fly ashes and the limestone fillers are reported in Tables 2 and 3 and their
porosity parameters are reported in Table 4.
The adsorption of the polymer on to the sands, limestone fillers and the fly ashes may be
summarised as follows:
• Sands show a different adsorption behaviour with a particular polymer
• Fly ashes show very high adsorption values. The reason could be related to their
physical and chemical nature (shape of the pores, high amount of un-bumt carbon).
See figure 6.
• Limestone fillers show a large range of adsorption values. Limestone Ll and L2
used, show very high adsorption. They have low purity and high surface area.
L4 and L5 show very low adsorption. They have high purity and low surface area.
L3 show higher adsorption than L4 and L5. It has a higher surface area and lower pore
size.
There seems to be fairly good correlation between physical and chemical properties such
as purity, size and shape of the pores, total porosity, surface area, and the adsorption.
Preventive characterization of supplementary cementitious materials represents a fast and
easy way to highlight possible problems of incompatibility.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In our experience if the adsorption value is high (over 30%), these materials could take
the admixture away so that the remaining admixture cannot act as an efficient dispersant
of the cement. It is therefore important to check the adsorption of these concrete
mixtures component.
Under these conditions, only by abnormally increasing the dosage of superplasticiser, the
required performances in concrete can be achieved. However, a too high dosage may lead
to retardation and low early strength.
Most common cases where tailor made solutions may be required are:
• very absorptive cement (low workability retention)
• ground slag showing re-fluidification of concrete (or delayed fluidification)
• very absorptive limestone material such as fillers and sand (no fluidification)
"Tailor" made TPC compositions have to be engineered in the above cases to overcome
inadequate performances. Some cases of this approach are now presented.
Four cements corresponding to type CEM II 42.5 AIL, conforming to EN 197, from four
different plants and four PCE polymers (PCE l is a high-rage water reducing
superplasticiser, PCE 2 is a long slump retention superplasticiser, PCE 3 is a high-range
water reducing and long slump retention superplasticiser compatible with most of the
cements and PCE 4 was a "tailor made" superplastiser that performs well also with
problematic cements, listed as cement type C and cement type D. Adsorption tests were
carried out with a total organic carbon (TOC) analyzer in order to compare the amount
of polymer adsorbed on to cement particles. The results are given in Fig. 7. It seems to
be important that at the beginning of the hydration reaction only a reduced amount of
polymer is adsorbed on to the cement particles. The free polymer in water seems to play
an important role for workability retention [4].
Sixteen concrete mixtures were made using aggregates from the same source, the four
cements and the four PCE polymers. The basic mixture composition was:
The results of these tests are presented in Table 5 and clearly indicate that PCE 1 has
good water reducing capacity, but does not maintain workability for more than 30
minutes, while PCE 2 has lower water reducing effect but good workability retention
even up to 60 minutes. PCE 3 performs well for water reducing and workability retention
PRACTICAL APPLICATIONS
One of the most significant example of the "tailor made" solutions was developed for the
construction of the Italian high speed rail network which is a part ofthe European Project
for the reorganisation of the rail transport system to link the countries of the European
Union, by 2010. The section between Milan and Bologna has a length of 182 km and
crosses Lombardia and Emilia regions. In the province of Modena there is a complex
system of bridges that cross the existing motorways, urban roads, the Panaro and Secchia
rivers. The main contractor for the 24,760 m bridges is the Modena Searl Consortium.
The bridge network consist of 755 omega shaped precast beams of 24 to 31.5 m span and
9 cast in situ beams of 136 m span. This complex system required innovative
construction techniques and a special precast factory was set up by the contractor
Pizzarotti. The prefabrication process starts with the assembly of the reinforcement cages
in 8 templates that are able to produce 2 cages per day. The assembly is completed with
the anchor heads and ducts with the
pre-stressing cables. It is then positioned on the casting bed. The particular shape of the
beam and the density of the reinforcement required a concrete with some specific
properties. Durability was another requirement. The concrete for the beams had some
special requirements. The characteristic compressive strength (Rck) was 45 MPa and the
minimum strength at 24 hours was 30 MPa for the application of the initial post
tensioning (35% of the design value). The cement used was type CEM I 52.5. A
limestone filler, two types of sand and three classes of coarse aggregates with a
maximum size of 32 mm were employed. The Consistence class of the fresh concrete was
S5 (slump> 220 mm). The mixture composition, per cubic meter, is indicated in Table 6.
Preliminary tests showed that the required performance of the concrete could not be met
with the usual superplasticisers based on beta naphthalene sulfonates. Therefore, a
superplasticiser had to be specially formulated for this project and is based on the Total
Performance Control concept. The concrete pour for the standard 31.5 m beam requires
276 m3 of concrete and was completely placed in four hours. To ensure trouble-free
casting the owner insisted that the concrete had to retain workability for 4 hours! After
about 24 hours, when the concrete had achieved a compressive strength of 30 MPa, the
internal formwork was extracted and the forms at the two extremities of the beam were
removed. At this stage, the beams were partially prestressed to 35% of the design value
and transported to the stockyard. Two beams are produced daily for a total of ten beams
per week. Figure 8 shows the set-up of the formworks in the casting shed. In spite ofthe
complexity of the works the project was completed ahead of schedule.
The concrete industry is heading towards a more sustainable development which aims to
make concrete an even more cost-effective building material. In this effort, the cement
industry uses increasing amounts of secondary fuels to improve efficiency. Also, the use
of supplementary cementitious materials, marginal and recycled aggregate in concrete
production is increasing. These practices introduce materials, not foreseen, in the past, in
to the concrete composition.
They are not always inert and often affect the performance of the superplasticisers
negatively. "Tailor made" TPC superplasticisers are required to overcome these
incompatibilities.
The new engineered polymer superplasticisers developed shows superior performance in
water reduction and workability retention. The innovative aspect of this new polymer
type is that it acts essentially on the kinetics of adsorption on to the cement. The
laboratory tests and field applications validate these aspects. This system can be tailored
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
to meet the individual needs of the producer and consumer of concrete in order to obtain
specific benefits.
REFERENCES
Parameters Cement without secondary fuels Cement with secondary fuels and
without adjustment
c3s (M %) 60.1 59.7
c2s (M %) 13.6 13.2
C4AF (M %) 5.1 4.9
o-C3A (M %) 6.0 10.0
c-C3A (M%) 5.3 2.2
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Table 2 -- Chemical analysis of the fly ashes
CaC03 CONDUCTIVITY
(% by weight) (p.S/cm)
Limestone 1 52 400
Limestone 2 73 2600
Limestone3 99 47
Limestone4 99 40
Limestone 5 99 33
Table 7 -- Application of "tailor made" superplasticizers in ready mix concrete in several European countries
Cement type CEM CEM CEM CEM CEM CEM CEM CEM
I 52.5 I 52.5 I42.5 I 42.5 I42.5 I 42.5 I 42.5 I 42.5
Cement content 222 + 114 222 + 114 375 375 330 330 300 300
FLY ASH FLY ASH
Admixture type LSRSP TAILOR LSRSP TAILOR HRWR TAILOR HRWR TAILOR
+AEA MADESP MADESP SP MADESP SP MADESP
+AEA
Admixture 1.2 I I I I 1.2 I 0.86
content (%)
W/C 0.56 0.51 0.43 0.39 0.47 0.48 0.49 0.50
Air content (%) 7.1 6.4 1.8 1.6 1.5 0.8 1.6 1.8
Slump (mm) 250 250 200 220 180 210 130 100
illt-5min.
(tiJ t- 30min. 240 250 190 195 130 210 60 80
(< t-60min. 230 250 180 190 7 21 0 8
(G t-90 min. 220 230 170 180 3.5 18 0 7
Temperature of
fresh concrete 23.2 27.1 27.0 29.5 25.8 28.6 30.0 32.0
Not for Resale, 08/10/2015 01:25:47 MDT
Licensee=UNI OF NEW SOUTH WALES/9996758001
COCJ
Compressive
strength (MPa)
@!day 7.4 11.5 20.8 27.4 29.1 28.6 25.8 24.9
Daneshlink.com
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
242 Corradi et al.
Table 8 -- Application of "tailor made" superplasticizers in ready mix concrete in Italy
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
30 30 30
fresh concrete ( 0 C}
Compressive
strength (MPa)
@!.day 25.0 20.5 24.1 20.3 25.0
Compressive
strength (MPa)
@28days 49.9 45.4 52.0 27.5 36.7
M 20
a
t
e 15
r
i 10
a
I
s 5
(%)
0
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
lliiY.......,u"IDa
Olhar I!Ur'Dp<l
llluodi.... Food..... ~
• 11'110
..J
l
Nerlll AniiOftm ln.
-·1-
«l&&INS.Amori..., sw
AM""' 24V.
Mioi. . ECIGO
IBiootAila
I
2ft
SOI!Iheii'A.,
o.-• s:n 41
Figure 2 --Supplementary cementitious material use for zero carbon dioxide increase
Extended
Workability
Adsorption of PCE
70
~50
BO
~ ----== --+-S1
i
I;
~ -40
11 30
-----=== - _,.
-+-S2
__._S3'
--><--84
--oE-S5
~
i 20
---+---87
-sa
10 r--- --
10 15 20 25 ao
timo(mln)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
p 50
0 40
L
y 30
M
E 20
R 10
(%) 0
0 5 10 15 20 25 30
TIME (MIN)
A 100
D
5 90
0
R so ~
B 70
E -+--Limestone 1
D 60 ~Limestone 2
p 50 -+-Limestone 3
0 40 --- Limestone 4
L
y 30 --Limestone 5
M
E 20
R
10
(%)
0 10 15 20 25 30
TIME(MIN)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
F~~~-2? o
~
iJ---..-~-~---~~~~~-~ EJI
~r==i ~-~:~------! 1==:1
~::r····_
--
,D ~:f~=~--~~~-~~-=~~-=u-1
--
01
.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: This paper presents five years of data on the effectiveness of six
commercial corrosion inhibitors: four admixed and two migrating. Both chloride
and carbonation induced corrosion were considered, and the performance of the
inhibitors was evaluated by long-term rebar corrosion monitoring, chloride profiles
and carbonation depth. ln chloride containing concrete, inhibitors seems to increase
time-to-corrosion: this is mainly related to a reduction of chloride penetration rather
than a significant increase in critical chloride threshold for corrosion initiation. The
inhibitors had negligible effect on carbonation rate; once corrosion occurred, they
were not effective in reducing corrosion rate. Migrating inhibitors delayed initiation
of chloride-induce corrosion if applied before corrosion initiation. No effect was
observed on corrosion rate.
S. Goidanich graduated in Chemistry at the Universita degli Studi di Milano in 200 I; got
PhD in Materials Engineering at the CMIC Department of Politecnico di Milano in 2005.
INTRODUCTION
Corrosion inhibitors are chemical compounds that delay initiation and/or propagation
of corrosion of reinforcement in concrete structures. They may be used both as preventive
measures for new concrete structures, through addition to the fresh concrete, as well as a
repair techniques for existing structures. Obviously, the prevention of corrosion is
primarily achieved in the design phase by using high quality concrete, adequate concrete
cover and suitable casting and curing. Additional prevention methods are mandatory
when severe environmental conditions occur or on structures requiring very long service
life (Bertolini eta!. 2004 ) . Among available methods, corrosion inhibitors can offer a simple and
cost effective prevention technique to delay chloride induced corrosion and as a remedial
job for carbonation (Eisener 20011 .
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Organic corrosion inhibitors act by adsorption on the metal surface, forming a layer,
that may inhibit both anodic and cathodic processes (mixed inhibition mechanism (Elsener
2001
; Welle eta!. 1997 ; Nmai and Paci 2003 l. Laboratory tests, both in solution and in concrete, showed
conflicting results about the efficiency and the minimum effective inhibitor concentration
(Pazini eta!. 1998; Trepanier eta!. 2001; Saricimen et al. 2002\ moreover, their relatively recent introduction
and fewer field applications provide little data on their long-term efficiency.
contradictory. There are doubt on migration ability (Trirthart 2003 l, or inhibitors may penetrate
only in conditions (low concrete cover, high permeability concrete) that favours
penetration of aggressive agents too (Andrade et • 1· 1996l. Probably not surprisingly, only tests
carried out by the manufactures show the good penetration of MCls (Bjegovic and Miksic 1999;
BavarianandRainerZOOJ). Some authors affirmed that MCis do not influence the corrosion rate in
mortar specimens exposed to chlorides. According to the literature, the low inhibiting
properties can be attributed to the fact that one of the inhibitor constituents does not reach
the embedded steel, reacting instead with cement paste (Eisener 2001 >. Contrasting to this
opinion, other authors claim that MCis are very effective in reducing reinforcement
corrosion (Bavarian and Rainer 2003; Mader 2005).
EXPERIMENTAL
Concrete specimens
Concrete specimens (20 x 25 x 5 em) were cast using 367 kg/m3 of cement CEM II
AIL 42.5R (as specified in EN 197-1), w/c = 0.6 and 1,770 kg/m 3 of limestone aggregate
with 1.2 em maximum diameter. Concrete mixture proportion and curing conditions are
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Exposure conditions
Three different corrosion conditions were studied: Series 1: concrete specimens cast
with admixed chlorides in dosages 0.8% and 1.2% by cement weight; specimens were
exposed to the atmospheric condition of Milan. Series 2: concrete subjected to ponding
cycles to accelerate chloride penetration, each cycle being three weeks long: first week of
wetting the concrete surface with I L of 3.5% NaCI solution, and two weeks of drying in
the laboratory (T = 20-25°C; RH = 50-60%). Series 3: concrete exposed to carbonation.
Carbonation was achieved by fluxing in a chamber pure C0 2 one hour per day until
complete carbonation of concrete samples. After carbonation, concrete specimens were
exposed outside in Milan.
Inhibitors dosages
Three organic commercial ACis (org A, B and C) and a nitrite-based product (N)
were added to the fresh concrete following the manufacturers' prescriptions (Table l).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Two commercial MCis were tested: they were amines and alkanolamines based with
phosphorous inorganic compounds in their mixture. Applied dosages were 400 g/m2 of
concrete surface for inhibitor MIG-1, and 250 g/m 2 for inhibitor MIG-2. The number of
applications depended on the test series (Table 2). MCis were applied once on specimens
with 0.8% admixed chlorides; three times on specimens with 1.2% admixed chlorides,
and twice on carbonated concrete. In the case of specimens of series 2, two conditions
were analysed: applications of MCis before corrosion initiation (series 2a) and after
corrosion initiation (series 2b). In this latter case, at the end of experimentation, since no
corrosion rate reduction was observed, 1 L of inhibitor was put on the upper surface of
the concrete samples for more then 8 months to verifY if in this case the inhibitors would
be able to reduce corrosion rate.
Corrosion tests
Corrosion potential was measured with an external saturated calomel reference
electrode (SCE) placed on the concrete, following the procedure in ASTM C 876.
1978
Corrosion rate was evaluated by the linear polarisation technique (Andrade and Gonzales >,
applying a potential scan rate of 10 mY/min in the range of ±10 mV with respect to the
free corrosion potential. Mean corrosion rate was calculated by means of the Stem-Geary
relationship: icorr (rnNm2) = B/Rp, where B is equal to 26 mV and Rp (0 m2) is the
measured polarisation resistance. In the case of carbon steel, 1 rnNm2 corrosion rate
correspond to a penetration rate of 1.1 7 J.Lmlyear. Since corrosion penetration lower than
l-2 J.Hnlyear can be considered negligible (Andrade 1998>, a polarisation resistance value of
Chlorides profile
Concrete cores, extracted from concrete specimens, were sliced. Each slice, after
milling, was then dissolved in nitric acid and the total chloride ions content was
determined by potentiometric titration with 0.1 moi!L AgN0 3, as specified in UNI 9944.
Carbonation penetration
Penetration of carbonation was detected by means of phenolphthalein tests on
100 mm cubic concrete specimens (as specified in UNI 9944), and was also confirmed on
concrete cores extracted from the reinforced concrete specimens.
RESULTS
were applied three times, after 8 months, 1 and 2 years of atmospheric exposure. Also in
this case, polarisation resistance, and so corrosion rate, seem to be not affected by the
presence of the two inhibitors (Fig. 3). Comparing Fig. 2 and Fig. 3, no difference in
behaviour of rebars in concrete without inhibitors or applied MCis was observed. The
potential increase may be attributed to seasonal variation.
Five chloride profiles were estimated during the five years test, respectively at the
14'h, 30'\ 47th, 65th and 86th ponding cycle, at the end of the test. Results are described in
Bolzoni et al. 2006. After I 4 ponding cycles, chloride content was higher than 0.1% by
concrete weight only for concrete depths lower than 20 mm, so lower than the cover. No
difference was observed among the concrete specimens cast with or without inhibitors.
As expected, chloride contents increased with time: the highest values were always
measured in concrete without inhibitor, while the lowest in concrete with organic
inhibitor A. Concrete with inhibitor N behaved as the reference concrete. At the end of
the tests, at the rebar level (2-3 em), the chloride content in the reference concrete is in
the range of0.6-0.7% by concrete weight; similar values were measured in concrete with
inhibitor N, while in the presence of organic inhibitors chloride concentration is in the
range ofO.Z-0.4%.
Series 2: Ponding tests with migrating inhibitors applied before corrosion occurrence
Fig. 5 reports potential trend for rebars embedded in concrete samples subjected to
accelerated chloride penetration. Before initiation of corrosion, migrating inhibitors MIG-
I and MIG-2 were applied at the 8th and 14th ponding cycle. Up to now 80 ponding cycles
have been performed. In the case of concrete treated with migrating inhibitor MIG-1 a
potential decrease was detected on all the rebars at the !51\ 24t\ 27th and 71 st cycle, while
in the case of migrating inhibitor MIG-2 only three of the four rebars showed a potential
reduction, at the 23'd, 28'h and 57th ponding cycle, respectively. One rebar is still passive
after 80 ponding cycle.
Series 2: Ponding tests with migrating inhibitors applied after corrosion occurrence
Four concrete specimens were cast with w/c = 0.65 in order to accelerate chloride and
inhibitor penetration. After 4 ponding cycles, that is approximately 80 days, both
potential and polarisation resistance lowered, indicating corrosion occurrence. Then
application of migrating inhibitors was carried out after 100 and 210 days. Measured
rebars polarisation resistance are reported in Fig. 6. No variation was detected after each
corrosion inhibitor application. Since ponding cycles continued after MCI application,
chloride concentration at the rebars level increased reaching values close to 2% by
was verified on a concrete core extracted from the samples). Both corrosion potential and
polarisation resistance decreased to values close to -400 mY SCE and lower than
lO Q m2 (Fig. 7), irrespective of the presence of any inhibitors.
DISCUSSION
An effective corrosion inhibitor should influence service life: it could delay time-to-
corrosion (i.e. initiation), by increasing the critical chlorides threshold or reducing
chloride ingress or carbonation penetration rate, or, once corrosion occurred, reduce
corrosion rate (i.e. propagation).
Chloride-induced corrosion
Influence of inhibitor on time-to-corrosion - The presence of admixed inhibitors
increase initiation time in the case of chloride-induced corrosion. Fig. 9 reports the
number of the ponding cycle in which the occurrence of corrosion was measured by the
reduction of both potential (Fig. 4) and polarisation resistance. Admixed organic
inhibitors A and C provided the strongest increment, while no effect was noticed in the
presence of inhibitor B. Nitrite-based inhibitor (N) effect is scattered: time-to-corrosion
ranged from values similar to those measured in the absence of inhibitors to values
slightly lower than those obtained in the presence of inhibitor C.
In the absence of inhibitor, critical chloride content is in the range 0.4-1.3% with
respect to cement weight: those values are in accordance with the literature, at least the
minimum ones (Bertolini et al. 2004 J. Organic inhibitors A and C provided significant increase
in critical chloride content (both minimum and maximum are doubled with respect to
reference condition), while inhibitor B did not show any significant influence. The effect
of inhibitor A was also confirmed by results of specimens of series I: in the presence of
1.2% admixed chlorides no corrosion has been measured even after 5 years. In the case of
nitrite-based inhibitor (N), considering inhibitor dosage and the obtained critical chloride
content, the estimated molar ratio [N0 2-]/[Cr] is in the range 0.21-0.44, that is lower
than 0.5, considered the threshold value above which the inhibitor is effective (Berke et al.
1990; Berke and Wei! 1992; Berke and Hicks 2004)
It is worth noticing that in any case values are scattered (Table 4): so the effect of
inhibitors is not so clear, above all in experimental conditions different from those used
in this research. Nevertheless, five years of testing show that organic inhibitor A and C
provide an increase in the critical chloride threshold.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
significant.
The nitrite dosage suggested by manufacturer, in this case 1% by cement weight, was
not adequate to reduce corrosion rate in carbonated concrete. This is in agreement with
literature data that report a dosage close to 2-3% in order to reduce corrosion rate in
carbonated COncrete (Alonso and Andrade 1995).
The application of MCis slightly reduced corrosion rate: in the presence of MIG-2,
polarisation resistance increases from 2-5 n m2 to 20 n m2 (Fig. 8), so the corrosion rate
decreased from 10-20 1-lm/y to values close to 2 1-lm/y.
Taking into account results obtained in five-years research, the tested organic
inhibitors are able to delay the onset of chloride-induced corrosion, while, on the other
hand, no significant effect was detected on both the penetration rate of C0 2 and the
reduction of corrosion rate, once corrosion occurred. The delay of chloride-induced
corrosion is mainly related to a reduction of the penetration rate of chlorides (as
suggested by the lower estimated chlorides diffusion coefficient) rather than an increase
in the critical chlorides threshold for the occurrence of localised corrosion. It is possible
to state that the commercial organic inhibitors act reducing the ingress of chlorides by
filling concrete pores and blocking the porosity of concrete with the formation of
complex compounds. This effect is also reported in literature (Nmai and Paci 2003 l. The slight
increase in the critical chlorides threshold is not so important, and it seems not always
reproducible and also influenced by experimental test condition. Summarising, the
inhibitors interact mainly with the cement paste, whereas no significant actions has been
measured at the interface with the reinforcement.
Nevertheless, it must be underlined that the effect on chloride diffusion is not a pure
electrodic inhibitive action, but it is only related to interaction of organic compounds with
cement paste. A similar or stronger reduction of chlorides diffusion may be obtained
simply with lower w/c, or by using blended cement, such as pozzolanic, fly ash or ground
granulated blast furnace slag addition (Bertolini et al. 2004).
chlorides content (Table 4), a molar ratio in the range 0.21-0.44 has been estimated, thus
confirming the fact that the inhibitor is effective only if the molar ratio [N02-]/[Cr] is
higher than 0.5. In the case of carbonation-induced corrosion the considered dosage is too
low to guarantee corrosion prevention.
REFERENCES
Alonso C. and Andrade C., 1995, "Effect of nitrite as corrosion inhibitor in contaminated
and chloride-free carbonated mortars", ACI Material Journal, 3-4, p. 130.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Andrade C., 1998, "An initial effort to use the corrosion rate measurements for estimating
rebar durability", Symposium on corrosion rate of reinforcement in concrete, ASTM,
Baltimore, USA.
Andrade C., Alonso C., Acha M. and Malric B., 1996, "Na2P04F as Inhibitor of
Corroding Reinforcement in Carbonated Concrete", Cement and Concrete Research, 26,
pp. 405-415.
ASTM C876-91, 1991, Standard Test Method for Half-cell Potential of Reinforcing Steel
in Concrete, American Society for Testing and Materials.
Bavarian B. and Rainer L., 2003, "Improving durability of reinforced concrete structures
using migrating corrosion inhibitors", Int. Conference EUROCORR, Budapest, Paper 19.
Berke N.S. and Hicks M.C., 2004, "Predicting long-term durability of steel reinforced
concrete with calcium nitrite corrosion inhibitor", Cement and Concrete Composites,
Vol. 26, pp. 191-198.
Berke N.S. and Wei! T.G., 1992, "World-Wide Review of Corrosion Inhibitors in
Concrete", Advances in Concrete Technology, Ed. V.M. Malhotra, Athens, Greece,
pp. 899-924.
Bertolini L. Elsener B., Pedeferri P. and Polder R., 2004, Corrosion of steel in concrete,
1'1 edn, Wiley-VCH, 392 pp.
Bolzoni F., Lazzari L., Ormellese M. and Goidanich S., 2006, "Prevention of corrosion in
concrete, the use of admixed inhibitors", International Corrosion Conference NACE, San
Diego, California, Paper 344.
Cerulli T., Coppola L., Maltese C., Lazzari L. and Ormellese M., 2003, "Laboratory tests
111
for evaluation of corrosion inhibitor efficiency", 6 CANMET Int. Conference on
Durability of Concrete, Ed. V.M. Malhotra, ACI International SP 212, Paper 68,
pp. 1109-1125.
Collepardi M., Fratesi R., Moriconi G., Corradetti C. and Coppola L., 1990, "Use of
nitrite salt as corrosion inhibitor admixtures in reinforced concrete structures immersed in
sea-water", Proceeding of the International RJLEM Symposium on Admixtures for
Concrete - Improvement of Properties, Barcelona, Ed. Vazquez, Chapman and Hall,
London, UK, pp. 279-288.
Elsener B., 2001, Corrosion Inhibitors for Steel in Concrete - an EFC State of the Art
Report, EFC, Number 35, IOM Communications, Cambridge, 68 pp.
Fratesi R., Moriconi G. and Simoncini S., 1994, "Corrosione delle armature di acciaio nel
calcestruzzo fessurato immerso in acqua di mare", Giornate Nazionali sulla Corrosione e
Protezione, AIM, Milano, 1994, pp. 139-148 (in italian).
Mader U., 1999, "A new class of corrosion inhibitors for reinforced concrete", Concrete,
p. 215-232.
Mader U., 2005, "Surface applied corrosion inhibitors for concrete repair strategies", Int.
Conf. 16'" ICC, Bejing, China.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Nmai, C.K. and Paci P.E., 2003, "A critical review of corrosion inhibitors and their use in
improving concrete durability in the middle east", th Int. Conf. on Concrete in Hot and
Aggressive Environment, Bahrain, Paper 14.
Pazini E., Leao S. and Estefani G., 1998, "Corrosion inhibitors. Behaviour ofNaN02 and
Amine-based Products in the Prevention and Control of Corrosion in Reinforced
Concrete"; NACE Latin-American Congress, Cancun, pp. 134-143.
Saricimen H., Mohammad M., Quddus A., Shameem M. and Barry M.S., 2002,
"Effectiveness of Concrete Inhibitors in Retarding Rebar Corrosion", Cement and
Concrete Composites, 24, pp. 89-100.
Trepanier S.M., Hope B.B. and Hansson C.M., 2001, "Corrosion inhibitors in concrete-
part III: effect on time to chloride-induced corrosion initiation and subsequent corrosion
rates of steel in mortar", Cement and Concrete Research, 31, pp. 713-718.
Tritthart J., 2003, "Transport of a surface-applied corrosion inhibitor in cement paste and
concrete", Cement and Concrete Research, 33, pp. 829-834.
Welle A., Liao J.D., Kaiser K., Grunze M., Mader U. and Blank N., 1997, "Interactions
of N,N-dimethylaminoethanol with rebar surfaces in alkaline and chlorine solutions",
Applied Surface Science, n. 119, pp. 185-190.
2a Chlorides penetration 2 - after the 8th and the 14th ponding cycle
2b Chlorides penetration 2- after the 5th and the 8th ponding cycle
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Chloride-induced corrosion
Carbonated concrete
Inhibitor C/ = 0.8% vs cern C/ = 1.2% vs cern
mean max min mean max min mean max min Tf*
No inhib 1.8 3.9 0.2 9.0 20.0 2.5 6.8 15.5 2.0 ---
A I.l 3.5 0.3 2.5 11.8 0.5 6.7 14.6 2.5 1%
Org B 3.1 5.4 1.4 9.1 20.0 4.0 5.1 11.2 1.5 25%
c 2.6 6.8 0.5 6.7 12.1 2.4 5.2 9.3 2.1 23%
NIT 0.9 4.6 0.2 2.7 9.8 0.5 5.7 12.5 1.9 15%
o I o
100 !00
<;;
!,
-100 ~ -100
t! -200 t! -200
-400 -400
chlorides "" 00~ chlorides ""'0.8%
-500 -500
500 1000 1500 2000 500 1000 1500 2000
Time(days) Time(days)
100
~ -100
~ -200
! -300
-400
chloride$ = 1.2%
-500
0 500 1000 !500 2000
Tlme(days)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~ -200 g
...
! -300 0:
-400
-500 +-----,---'-~--~-~
0 500 1000 1500 2000 500 1000 1500 2000
Time(days) Time(days)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
------MIG-I ------MIG-I
100 --6-----MIG-2 1000
------6--MIG-2
--MClapplication
--MCiapp!ication
~ -100 100
"8
~ -200 g
...
~ -300
0:
10
-400
-500 +--'-.1....,-....L.~--~-~
0 500 1000 1500 2000 500 1000 1500 2000
Time(days) Time (days)
111
30111 cycle
111
15 cycle 45 cycle 601hcycle 75"'cycle -.-no inhib
100 -IJ.~orgA
-o-orgB
0
-o--org C
~ -100 . , ---N
t:l
"'~
~
-200
-300
vf1 I l
-400
-500
~
0 500 1000 1500 2000
Time(days)
-400 -400
MCI MC!
1000 appllcatkm 1000 application
! illbibiturMIG·I \
100 100
M-e
'<g&' 10 ,j--'ll!aiH- ---,- -
' - - g lO .,.--,-go,.--,.- ---,- - - -
: ~
-700 +--L----.---....----,--.,
500 1000 1500 2000 500 1000 1500 2000
Time(days) Time(days)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
-100 b. inlnbitorMIG-2
;;
.!
fol
-200 100
1.)
"S
"'!l -300 g
i
a-400
: 10
£.
-500
-600
0 500 1000 1500 2000 500 1000 1500 2000
Time(days) Tlme(days)
100
admixed inhibitor migrating inhibitor
80 ~
.!I
.;;"
t' 60
l.= 40
20
0
nolnhlb orgA orgB orgC N MIG-I MIG-2
Figure 9-Ponding cycle in which a potential reduction was measured for rebar
embedded in concrete with admixed inhibitors
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Cement-Superplasticizers Interaction:
The Influence of Superplasticizers on
the Formation ofSyngenite
The aim of this study was to reveal the influence of potassium and sodium sulphates
on the dispersing effect of superplasticizers. Two types of superplasticizers have
been used for investigations: polycarboxylate (PC) based and a naphthalene
sulphonate formaldehyde condensate (PNS). By using the combination of rheological
testing, quantitative X-ray diffraction analysis, environmental scanning electron
microscopy and measuring the superplasticizer adsorption, It was possible to
investigate the different influence of K2 50 and Na 2 50 4 on the fluidity of cement
4
pastes. Thereby it can be proven that syngenite formation is an important factor
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
that changes the rheological behaviour of cement pastes and that it is specifically
influenced by superplasticizers.
Christiane RoBier, Dipl. earth sc., Ph.D. student F.A. Finger Institute for Building
Materials Science, Bauhaus-University Weimar, Germany.
INTRODUCTION
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
monosulphate, secondary gypsum and syngenite). 1• 2
Soluble alkalis affect both the mechanism of hydration in the cement and the
adsorption of a superplasticizer on cement particles. 3• 4• 5• 6 Alkalis in the cement are
either bound in the structure of clinker minerals (Na 20, K2 0) or they are present in a
form of alkali sulphates. Often calculated as alkali oxide in composition of cements K20
(0.1...1.5 wt.%) outweighs Na20 in European cements (0.1. .. 0.8 wt.%; in US cements:
LO wt.%). 7 After water addition to the cement powder alkali sulphates dissolve
immediately. Thereby the level of sulphate-, potassium- and sodium ions in the pore
solution of the cement paste increases. The amount of soluble alkali sulphates present in
the cement is known to be very important for the rheological behaviour and setting of
cement.
The first hydration product during Portland cement hydration is ettringite. Calcium,
sulphate and aluminate ions are consumed by ettringite crystallisation. At a certain level
of calcium-, potassium- and sulphate ions in the pore solution the supersaturation for
syngenite is reached, leading to crystallisation of syngenite. The syngenite forming
reactions can be represented as (J.Bensted) 8 :
Crystals of syngenite are long prismatic and thereby they are able "to bridge" cement
particles and cause a false setting. Similar effect can be observed if secondary gypsum
precipitates. No sodium analogue of syngenite is known. 8 Kurdowski and Grzeszczyk
showed that yield stress and plastic viscosity of cement pastes increase when potassium
sulphate concentrations exceed a certain value (2 %). 9
Materials
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The low alkali content was the main reason for selecting this cement type. Mixtures
of cement and solutions of potassium sulphate, or sodium sulphate, at different
concentrations were prepared in order to simulate cements with higher alkali content.
This method assured invariable cement properties, except for the content of alkali
sulphates.
Assuming that all the water soluble potassium is bound to sulphate the starting
cement contains 0.74 % K 2 S04 • The potassium sulphate content of the cement was
increased in three steps as shown in Table 3. Added quantity of 1.94 % (1.05 % K20,01 )
simulates high alkali cement. In practice higher alkali quantities are not found. For the
present investigation, one higher alkali level was chosen to enhance effects and to
compare with findings for regular alkali contents. For comparison the content of sodium
sulphate was also increased to the same molarities (0.12; 0.18; 0.29 M) in parallel batches
of cement pastes (Table 3).
The water to cement ratio of 0.39 was constant for all samples. The mixing of
cement with water or potassium/sodium sulphate solution was carried out with a standard
mixer according to a defined protocol (Table 4). For investigation of hydration by
Environmental scanning electron microscope (ESEM) and TOC-analysis, samples were
mixed by hand.
The added amount of SPs was determined by preliminary rheological testing (see
below). The superplasticizer was first mixed with water or with the solution of alkali
Methods
Fluidity of cement pastes - The rheology of the cement pastes was investigated using a
true speed controlled viscometer Viskomat NT, manufactured by Schleibinger Gerate
GmbH, Germany. One of the possible applications of this device is to monitor changes in
the fluidity of cement pastes due to hydration reactions. The torque (T) with time (2
hours) was measured under the constant rotation speed (60 rpm) and temperature (19.5 ±
0.5 °C). An increase in torque is equivalent with a decrease in fluidity of cement paste.
Viscometer measurements at constant rotation speed are affected by both factors that
define fluidity: (apparent) viscosity and (apparent) yield strength.
X-ray diffraction - Syngenite content in the cement pastes was quantified by Rietveld
Analysis of XRD data. Therefore cement hydration was ,stopped" after 10 minutes by
adding 2-propanol to the paste. The samples were dried at 35 oc on a metal plate and then
homogenised. Due to the fact, that syngenite is a minority phase in the hydrating cement,
it was necessary (to increase precision of the results) to extract the cement phases in the
solution of salicylic acid and methanol. By applying this procedure calcium silicate
phases were removed from the cement paste. The remaining powder was analysed by
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
XRD (05000, Siemens, Germany). The CuKa (A-= 1.5418 A) radiation was generated at
40 rnA and 40 kV. Data have been collected over a Bragg angle range of 8-60° (28),
using an angular step size of0.05° and 5 s counting time.
Calculation of the syngenite content was done by Rietveld analysis with software
Autoquan (BGMN, Germany). Because only short hydration periods were investigated it
has been assumed that the samples contain a small amount of amorphous phases,
therefore no standard was added.
ESEM-FEG Analyses- The morphological analyses of cement pastes with various alkali
sulphate levels and containing the different superplasticizers were performed by using an
Environmental Scanning Electron Microscope (ESEM, XL30, FEI Netherlands) equipped
with a Field Emission Gun (FEG). The operational parameters for the "Environmental"
mode were the following: water vapour in the chamber was set to 8.32- 9.36 mbar and
sample temperature to 12 °C (humidity between 75 % and 90 %). For ESEM
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
increasing progressively the added amount of superplasticizer during the measurement of
the fluidity (i.e. torque at constant rotation speed in the Viskomat NT). Results are shown
in Fig. I. A gain in fluidity (minimum in torque) curves was reached in the samples
without addition of SP after 20-30 minutes. According to the flocculation model 10 , the
floc structure is destroyed by mixing and rebuilt at rest of the paste. The minimal torque
is achieved as soon as the flocculation - deflocculation equilibrium of cement particles
occurs. The torque value which was reached after 30 minutes of measurement was
determined (Fig. 1). Subsequently SP dosage was increased at defined steps until a
maximum fluidity was gained. On this way saturation point can be measured.
It is obvious (Fig. I) that the saturation point of both PC-type superplasticizers lays
around 0.8 %. The addition of 1.8 % PNS superplasticizer to the studied cement is
necessary for reaching the saturation point.
The sodium sulphate content in the m1xmg water was increased to the same
molarities as potassium sulphate. Compared to potassium sulphate, increasing amounts of
sodium sulphate cause higher fluidities (lower measured torque) of the pastes. The torque
maximum immediately after starting of measurement is significantly lower if sodium
sulphate instead of potassium sulphate was added (compare Figs. 2a and 2b). By
The content of syngenite that was formed after 10 minutes of hydration in the
presence of superplasticizers was determined by XRD-Rietveld analysis. It can be seen
that the type of superplasticizer influences the syngenite content in cement pastes
specifically (Fig. 7). Especially PCI (superplasticizer for ready-mix application)
decreases the syngenite content significantly. At the same time, compared to PC2 and
reference pastes, PCl achieves relatively high fluidification at all levels of potassium
sulphate addition (Figs. 2a-5a). These data show that in the presence of syngenite the
fluidity of PC-containing cement pastes is decreased.
It is known that due to the increasing concentration of sulphate ions in the cement
pore solution, the adsorption ofsuperplasticizers is decreased6• 12 • According to Yamada6
the fluidity of cement paste with PC-type superplasticizers is diminished by addition of
sodium sulphate. Nawa 12 has reported that sodium sulphate increases the fluidity ofPNS
containing cement pastes due to inhibited adsorption (competitive adsorption). Kim et al.5
determined an optimum content of soluble alkalis in cement that results in a minimal
fluidity loss. These results are confirmed by the data presented in Figs. 6, 2b-5b and
extended by results in Figs. 2a-5a: there it is obvious that the decrease in fluidity due to
increased sodium sulphate addition is much less than for potassium sulphate addition.
Due to the low alkali content, the investigated cement is incompatible with the PNS-
type superplasticizer; this results in a high fluidity loss of the paste (Fig. 3b: OM Na2S0 4).
After the addition of 0.12 M Na2 S04 and 0.18 M Na2 S0 4, the fluidity of cement paste
with PNS was increased. The results shown in Fig. 3b are in accordance with previous
studies (5, 6 , 12). In the presence of syngenite the gain in fluidity of PNS-containing
paste is generally lower (Fig. 3a). During the first 60 minutes of hydration, at highest
molarity of mixing water (Fig. 3a: 0.29 M K2 S0 4), the fluidity of the paste is even
decreased as compared to the reference sample without superplasticizer. This loss in
fluidity can definitely be attributed to high rates of syngenite formation.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
%). Addition of potassium sulphate causes a decrease of superplasticizer adsorption in all
investigated samples. Comparison of PC! and PC2 adsorption shows that at higher
sulphate content, adsorption ofPCl is more influenced by potassium sulphate than PC2.
XRD-Rietveld analysis was furthermore carried out to determine, whether the C3A
consumption is affected by alkali sulphate addition. Therefore after 10 minutes of
hydration the C3A content in cement pastes was determined (Fig. 8). The graph in Fig. 8
Fig. 10 illustrates the model for the rheological behaviour of cement pastes in the
presence of long prismatic hydrate phases (like syngenite or secondary gypsum). It is
known 16• 17 that the long prismatic crystals rotate into the shear direction in simple shear
conditions. Depending on the crystal asymmetry the particles reorient to the shear
direction in preference (high crystal length/diameter ratio) or rotate infinitely
(axisymmetric particles) during continuous stirring. The alignment of long prismatic
crystals into the shear flow is accompanied by gain in fluidity. Considering the model in
Fig. 10 it is possible to explain why the high decrease in fluidity of cement pastes with
high potassium content, which occurs immediately after addition of mixing water, can be
overcome by intensive stirring.
CONCLUSIONS
The results show that for a complete understanding of the action of superplasticizers
the amount and type of alkali sulphate, and the respective hydration products and their
morphology, have to be considered. It was shown, that syngenite formation is a factor
that has a marked influence on fluidity of cement pastes with higher content of potassium
sulphate. The amount of syngenite formed is specifically influenced by the presence of
superplasticizers. The increase in syngenite content results in a decrease of the cement
paste fluidity in the first minutes of hydration. The gain in fluidity during continuous
stirring is related to the subsequent alignment of elongated crystals into the shear
direction.
ACKNOWLEDGMENTS
3. Gartner, E.M.; Young, J.P., Damidot; D.A., Jawed I., 2000, "Hydration of
Portland Cement" in Bensted, J.; Barnes, P., Structure and Performance oj
Cements. 2nd Edition, E+FN Spon, pp.57-139. ISBN 0-419-23330-X.
5. Jiang, S.; Kim, B.-G.; Aitcin, P.-C., 1999, "Importance of adequate soluble
alkali content to ensure cement/superplasticizer compatibility", Cern. Caner.
Res, V. 29, pp. 71-78.
6. Yamada, K.; Ogawa, S.; Hanehara S., 200 I, "Controling of the adsorption and
dispersing force of polycarboxylate-type superplasticizer by sulphate ion
concentration in aqueous phase", Cern. Caner. Res, V. 31, pp. 375-383.
7. Stark, J.; Wicht, B., 2000, Zement und Kalk. Der Bausta.ff als Werkstoff, I st
Edition, Basel; Boston; Berlin: Birkhauser, ISBN 3-7643-6216-2.
10. Tattersall, G. H, 1955, "The rheology of Portland cement pastes", Brit. J Appl.
Phys., V. 6, pp. 165-167.
11. Yang, M.; Neubauer, C.M.; Jennings, H.M., 1997, "Interparticle potential and
sedimentation behaviour of cement suspension - Review and results from
paste", Advn Cern Bas Mat, V. 5, pp 1-7.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
13. Uchikawa, H.; Hanehara, S.; Shirasaka, T.; Sawaki, D., 1992, "Effect of
admixture on hydration of cement, adsorptive behavior of admixture and fluidity
and setting offresh cement paste", Cem Caner Res, V. 22, pp. 1115-1129.
15. Flatt, R. J.; Houst, Y.F., 2001, "A simplified view on chemical effects
pertrubing the action of superplasticizers", Cem Caner Res, V. 31, pp. 1169-
1176.
16. Jeffery, G.B, 1922, "The motion of ellipsoidal particles immersed in a viscous
fluid". Proe. R. Soc. London Ser. A, V.l02, pp. 161-179.
18. Ailcin, P.-C., 1998, High Performance Concrete.!" Edition, E&FN SPON,
London.
19. Flatt R. J., 2004, "Dispersion forces in cement suspensions", Cem. Caner. Res,
v. 34, pp. 399-408.
Table 1 • Chemical composition of cement CEM I 52.5 R ·given in wt. o/<>l
CaO Si02 AhOJ FC!03 K20 Na20
66.0 22.9 4.0 1.3 0.66 0.23
K20•ol Na20•ol S03 CaO&ee MgO Blaine
0.4 0.06 3.1 0.0 0.6 5170cm2/g
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
s~ _]lrocedure duration
1. cement added to water or potassium/sodium sulphate solution <10 s
containing superplasticizer
2. mixing in laboratory mixer according EN 196-1 by 140 rpm 60s
3. stop for cleaning 30 s
4. mixing by 140 rpm 120 s
TblSCharac
a e . . of superpJasllctzers
tenzatton 1·· -
Description PC1 PC2 PNS
Naphthalene
Type Polycarboxylate Polycarboxylate sulphonate
Formaldehyde
Backbone chain Polymetbacrylic acid --
Side chain Metboxypolyetbylenglycol --
Connection between
Ether --
backbone and side chain
COO'/etber 3.5 8.0 --
Molecular weight of side
chains
Low High --
Solid content (wt.- %) 30
Application Ready-mix Precast Ready-mix
~o.--,---,--f-:~P~C~1_f0~M~~~S~0~4--~-~~-~PC~1_no1,1RBMMK~;S~0~4
~··~••,!---+----i ........ pc2_o M ~so4 · ... PC2_0,1BM ~S04
200
e
E
1so ··.. ·· ..
~ •• ---NSF OM~S04 . .,. NSF_0,1BM~S04
~ 160 ~~·• .....,~ ••
gE 140
120
\\.'·
.
~.
~ 100 H .'··;:-r-- - .._.~s;_ ! ---
'·f~
"'
~ :BO
20
c___ \ \ . ·....
\-.._ ·..
\..._-.......;~:i ...
•
··.:-:-.._
.. ..~ ....
........... ........
0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8
superplasticizer dosage %
270
180
150
•""-
~
-0.12M
K~SO,
-~MK,S~
tlme(mln
20 40 00 80 100 120
Fig. 2a- Increasing K,SO 4 content, without addition of SPs. The torque in the cement
paste with 0.29 M K,S0 4 solution is out of measurement range (>)oo Nmm).
300
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
270
1240 -0.12M
i 210
-0.29 N
Nazsf. - OM
N~
Na,SO,
180 ~ Na,s\o,
~
160 Ume(ml~)
20 40 60 100 120
150
120
My
--- ---
-0.2811
i 00
K2 S0 4
1
i """""""'
00
I i~:o~
•• -o.u_M"
K2so ..
0 lme (ml~)
• •• •• •• •• 100
•••
Fig. 3a- Increasing K2504 content and addition of PNS.
:'-o::N.
120
-o.2eM
K,$0~
30 - K.SO
20 40
-0.12_1#"
K~o.
80
-o~A Kti04
80
time(mln)
100 120
150
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
120
Ii!:. eo
j eo
-o.29M
! Na~o.
-o.12M
/Na2so. N~;8o~
-
30
-N~% tlme(mi':l)
20 40 60 80 100 120
120-H.----------------j
}•o~~~~--~~~~~~~~~
"'" -OM K,SO
}eo ~::::===::::::;5~~
0~-~-~--~-~-~~~
0 20 40 60 80 100 120
180,---------------------.
I •o
i 60~:::::::;~~~
tlme(mln)
0~--~--~----~--~--~~~
0 20 40 60 80 100 120
90
80
70
0~
I: 60
0
i.
... 50 --·- f-·· --
~ 40 '-·· -- 1-
~ 30
20
Pc-'11 0.4%
Pczl.j
1-
I- - 1- -- 1- :--
10
0.4~ PC 1 1.5% 1[ PC 2 1.5% l[ NSF 1.5%
0
0 0.18 0 0.18 0 0.12 0.18 0.29 0 0.12 0.18 0.29 0 0.12 0.18 0.29
K2S04 concentration of mixing water (M=molesflitre]
~ 2,0
~
E
fii
1.5
~ 1,0 .J-----------·--
:0
0
~ 0.5
0,0
0 0.12 0.16
K2S04 concentration of mixing water [molesll~re]
0 0~2 Q18
K2S04 concentration of mixing water [moles/litre]
Fig. 8- C3A content in cement pastes after 10 minutes hydration. C A content in the
3
unhydrated cement was 7-5 %.
Fig. 9a- Reference cement paste without addition of alkali sulphate: short prismatic
ettringite (approx. 500 x 200 nm) covers cement particles.
Fig. 9b- Addition of 0.18 M K,S0 : long prismatic syngenite and short prismatic
4
ettringite on cement particle.
Fig. 9d- Cement paste with PNS and without addition of alkali sulphates: short
prismatic ettringite, microstructure similar to reference (Fig. 9a).
Fig. 10- Rotation of long prismatic crystals like syngenite into the shear direction
(indicated by arrows).
Adsorption of Lignosulphonates
on Cement and the Hydration
Products of Cements
by B.O. Myrvold
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
BACKGROUND
EXPERIMENTAL
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Substrates
Adsorption isotherms
r = (C,,-C,)*V/m,ub" (I)
(2)
With m,ub in grams, C,, and C, in ppm and a fixed volume of 20 ml the factor 50
gives r in mg lignosulphonate pr. gram substrate.
r was plotted as a function of C,. In a reactive system like cement C, is not a true
equilibrium value.
There were a few slight variations to the procedure above. With the hydrated
clinker phases, I 0 ml of a lignosulphonate solution of twice the final strength was used to
keep the total volume at 20 ml.
Lignin modification
Most of the studies presented here were performed with the same batch of high
molecular weight softwood sodium lignosulphonate. (Mw = 47.8 kDa, M, = 5.8 kDa,
reducing sugar 0.4%, Ca2+ 0.01%)
phenolic groups. The amount of phenolic hydroxy Is left was determined by standard UV
spectroscopic methods [23], while the carboxylic acids were determined by conductivity
titration [24].
Isotherms
In all the isotherms presented here we use the term "consumed" about the
reduction of lignosulphonate amount in the solution. There might be several mechanisms
operating. There might be true adsorption where the lignosulphonate interacts strongly
with the surface. There are several possible isotherms, but the most common types will
all have a plateau value where the surface is covered by one monolayer of the solute.
There is also the intermediate situation of the cation assisted adsorption [25].
Here the substrate provides ions that, partially or fully cancels the charges in the
lignosulphonate molecule. These new ion pairs adsorb to the surface. There will be no
adsorption in the absence of the ions. This is different from the precipitation, and the
lignin will not be consumed without a substrate to adsorb to. There will be a plateau
value where the surface is fully covered by the lignosulphonate.
For cement we also have the possibility that the sulphonate groups of the
lignosulphonate takes part in reactions (probably replacing sulphate). This will bury the
lignosulphonate inside the newly formed reaction products.
(3)
where 0 is the fractional coverage and k, is a rate constant. Thus the rate of desorption is
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
proportional to the amount adsorbed. The rate of adsorption (R,) is given by:
R, = k, (l-0)c, (4)
where c is the concentration in solution and k, is a rate constant. At equilibrium the two
rates are the same.
R, = Rd or k, (l-0)c = kd 0, (5)
rearranging we obtain
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
some again at the highest concentrations where the system is well dispersed [2, 15]. The
reaction will proceed faster in the samples with low concentrations of lignosulphonate
than in samples with much lignosulphonate. This could lead to abnormally shaped
isotherms. The effect will depend on the relative affinity of the starting material and
reaction products towards the lignosulphonates.
Cement isotherms
Two sets of isotherms are shown in Figure I and 2. In the first (Ia) the cement
has been hydrated for the given time before the lignosulphonate solution is added. The
solution and prehydrated cement has been in contact for 25 minutes. We could not
prehydrate cement for more than 8 hours due to the hardening. We clearly see how the
adsorption rises strongly with time, except for the product hydrated for 1 hour, where
there is a small decrease. It has been shown that hydrated C,A and C4AF [1] consumes
less lignosulphonates than the unhydrated material. C,A and C4AF also are reported to
adsorb less calcium lignosulphonates with longer contact times [27], but then contact
times of 12-24 hours were used. On the other hand the hydration of C3S/C 2S proceeds so
slowly that it has hardly started at 1 hour [27]. It takes about 4 hours before these
minerals reacts to any significant degree [9). This can explain the small decrease when
the aluminate phases are converted to their hydration products, while strong rise is due to
the new surfaces created by the hydration of the silicate phases. A new and larger surface
is formed and thus the plateau value increases, strongly. In Figure I b the data is plotted as
fractional coverage. The initial slope of the adsorption isotherms are the same in all cases
with prehydrated cement. Thus it seems reasonable to assume that a lot of new surface is
created, but the nature of the surface is the same. The material formed is the C-S-H phase
A value of 4.0 10-2 ppm- 1 was found for the constant bin eq. 6.
The differences between the two sets of isotherms clearly show that the surface
grows in contact with water, but that this growth is stopped very efficiently with
lignosulphonates present. The reactions between water and the cement minerals release
ions into the solution. These ions might modify the properties of the lignosulphonates.
However, for the pre-hydrated cement (Figure 1) these ions will be washed away when
the water is sucked off, and the hydrated cement isolated. There will thus be less ions in
this system. The larger adsorption can thus not be due to more precipitation due to the
ions. In the system where the cement is in contact with both the water and the
lignosulphonate at the same time, all the ions leached out will also be available to interact
with the lignosulphonate in solution. Still, this does not give a higher consumption.
In Figure 3 the r values for the adsorption plateaus as a function of time are
compared. The growth of new surface is much smaller when lignosulphonate is in contact
with the cement.
Modified lignosulphonates
When we compare the slope of the curve with the setting times in Table 1 it can be seen
that, the lignosulphonates that adsorbs strongest gives the shortest setting times, e.g. the
least retardation. When all the lignosulphonate is adsorbed to the aluminate phases it will
be Jess left to interfere with the hydration of the silicate phases, and thus a lower
retardation.
Table 1 summarises the properties of the different modifications.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Cement consists of several minerals and the system is thus highly inhomogenous
When suspended in water the cement minerals will partially dissolve and the surfaces
gain a charge (as measured through their zeta-potential). The charge is different for the
different clinker minerals. The C,A!C,AF phases are positively charged, while the
C,SIC,S are negatively charged [20].
The lignosulphonates adsorb to the positively charged areas of the surface. This
cancels the positive charges on the C,AIC,AF parts of the surface or even reverses them
to negative charge. This explains why the dosages of lignosulphonate used in concrete are
much smaller than that needed for full surface coverage. Only the positively charged
areas need to be covered. It also explains why cement that has a zero or negative zeta-
potential will gain a more negative surface potential when lignosulphonates are adsorbed
[16, 21]. The measured zeta-potential for cement is an average over the positive and
negative areas. When lignosulphonates are adsorbed the negative areas stay unchanged
while the positive areas also becomes negatively charged, the average will thus become
lower. It is agreed in the literature that lignosulphonates interact strongly with the pure
aluminate phases [3, 4, 10,11].
Not all of the lignosulphonate adsorbs to the C,AIC,AF part of the surface. Some
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
remain in solution. The lignosulphonate left in solution will interact with the ions
dissolved into the pore water. These ions will partially cancel the negative charges in the
lignosulphonate molecule due to ion pairing. The lignosulphonate with cancelled charges
can adsorb or precipitate onto the C,SIC,S parts of the cement surface. This will interfere
with the normal hydration of the surface and causes the retardation of the setting.
CONCLUSIONS
During hydration the surface area of the cement grains grow through the formation of the
C-S-H phase. The growing surface areas are seen by an increased plateau value for the
adsorption isotherms. However, the slopes of the adsorption isotherms remain constant,
suggesting that the surfaces formed have the same characteristics. When lignosulphonates
are added, ions that are released by the rapid hydration of the C,AIC,AF phases interact
with and modify the lignosulphonate. The modified lignosulphonate covers the silicate
phases and there is a very slow growth of new surface area when the lignosulphonate is in
contact with the cement.
For the best dispersing effect we should have a lignosulphonate that adsorbs strongly to
the aluminate phases. As much lignosulphonate as possible should be consumed by the
aluminate phases, as the excess lignosulphonate will interact with the ions released and
then interfere with the hydration of the silicate phases.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
17. M. Georgescu and I. Carazeanu, Influence de quelques additiefs fluidifiants sur le
processus d'hydratation et durcissement de 4Ca0Al,O,·Fe,0 3, Rev. Roumanine Chemie
40(10) (1995) pp. 983-992
19. M. Michaux and E. B. Nelson, Flash-set behaviour of oil well cement slurries
containing lignosulphonates. 9th International Congress on the Chemistry of Cement,
New Dehli (1992) pg. 584-590
21. H. Uchikawa, S. Hanehara and D. Sawaki, Effect of electrostatic and steric repulsive
force of organic admixture on the dispersion of cement particles in fresh cement paste,
Proceedings 101h ICCC, Gothenburg (1997) paper 3iii001
22. F. M. Ernsberger and W. G France, Some physical and chemical properties ofweight-
fractioned lignosulfonic acid, including the dissociation of lignosulfonates, J. Phys.
Colloid Chern 52 (1948) pp. 267-276
23. Y.-Z. Lai, Determination of phenolic hydroxyl groups, pp 423-434 inS. Y. Lin and C
W. Dence (eds.) Methods in Lignin Chemistry, Springer Verlag (1992)
27. I Odler and Th. Becker, Effects of some liquefying agents on the properties and
hydration of Portland cement and tricalcium silicate pastes, Cement Concrete Res. 10
(1980) pp. 321-331
# m1lheqmvalents pr gram.
~
tJ)
20
.. .. lo1hour
E 15 .. 6 • "2 hours
"CC • • 4 hours
..''
CD c
E 10 D
•
:I
Ill
"• 0
• 0 0 16 6 hours
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
c 5- .. 8 hours
0
u I • None
0
0 200 400 600 BOO 1000
Concentration I ppm
• 1 6 0 rJie o o •
$' 0.8 ~ rf'O
!; 11.•
~ 0.6 ..
c "
~ 0.4 "'
£o.d
200 400 600 800 1000
Concentration I ppm
'---------------- -- -
12 1o 8 hours
~ 10 • ~ • . :~
""
"..
E •4 hours
'0
8-. ~ 1c 2 hours
a> '"D.
E 6
I
1.o.1 hour
:J
Ill 4
le.30 min
'
c
0
u 2 • 16 hours
0
0 100 200 300 400 500 600 700
Concentration I ppm
~
: 15
-g 10
i!0
.a 5- --with LS
-a- Without LS
<C
0 4 8 12 16
Time/hours
Figure 3-- The plateau values forthe different isotherms, as a function oftime.
Filled symbols are with lignosulphonate added initially. Open symbols are for
lignosulphonates added after the given delay. --`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
.
§ 0.8
¢
..
8 0.6 ..
"ii ~ o Mod-1
6 0.4 0 ..
.. Mod-2
'iS oMod-3
I! 0.2
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
u.. • Mod-4
0
0 500 1000 1500
Concentration I ppm
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
NSR-1, having hydrophobic groups, adsorbs on cement particles and is able to give
hydrophobicity on pore surfaces. As a result, NSR-1 reduces the interfacial tension
between pore surfaces and water in these pores, and can reduce drying shrinkage
effectively.
Drying shrinkage behavior was measured by mortar with W/C of 44%. At 28
days, the shrinkage strain of NSR-1 was 390 x 10·6 at dosage of 0.3% by mass of
cement, and those of conventional DSRA were 380 x 10·6 to 405 x 10·6 at dosage of
2.0% by mass of cement. The NSR-1 showed the same drying shrinkage-reducing
effect at a dosage of about one-seventh of the add it ion amount for the conventional
DSRA.
Tsuyoshi Hirata is Nippon Shokubai Fellow and General Manager of Research Project,
Nippon Shokubai Co. Ltd., Japan
Dr. Toyoharu Nawa is a Professor at the Division of Solid Waste, Resources and
Geoenvironmental Engineering Laboratory of Eco-Materials and Resources, Hokkaido
University, Japan
INTRODUCTION
I) The surface tension of DSRA in the cement filtrate (A) (Method was shown in
experimental 2)-a))
The relationship between the concentration of DSRA in the cement filtrate (A) and
the surface tension is shown in Figure l. The surface tension of the cement filtrate (A),
containing SRA-3 could not be measured because SRA-3 did not dissolve in the cement
filtrate (A). The surface tensions of SRA-1 and SRA-2 in the cement filtrate (A)
decreased with increasing SRA concentration, and achieved constant values of about 40
mN/m at dosages of more than 1.0 mass%. On the other hand, that of NSR-1 achieved
constant value at 55 mN/m at dosages of more than 0.5 mass%. Comparing the surface
tensions of SRA-1 and SRA-2 with that ofNSR-1 at 1.0 mass%, the former were 37.0
mN/m (SRA-1) and 43.2 mN/m (SRA-2), and the later (NSR-1) was 55.7 mN/m. Hence,
the surface tension ofNSR-1 was higher than the others.
4) Concrete properties
Table 5 shows properties of fresh concrete containing DSRA. Figure 5 shows air
contents and spacing factors at 5 and 30 minutes after adding water.
Concrete containing NSR-l(C-NSl) achieved the target of slump of 20 em with a
dosage of 0.7%, by mass of cement, and thus C-NS 1 was evaluated without the addition
of a water reducing agent. On the other hand, concretes with 2.0% by mass of cement, of
SRA-1, SRA-2 and SRA-3 (referred as to C-SRl, C-SR2 and C-SR3 respectively) did not
achieve the target of slump of 20 em, and therefore a polycarboxylate-type
superplasticizer (PC-1) had to be used. Further, the air content was controlled by using
air-entraining admixtures or defoamers, if necessary. However, it was difficult to control
the air content in C-SR3, containing SRA-3. Thus C-SR3 was evaluated as non-AE
concrete.
The NSR-1 delayed initial setting by 40 minutes and also final setting by 60 minutes
compared to plain concrete. On the other hand, all the mortars containing SRA-1, SRA-2
and SRA-3 showed no retardation of setting.
The spacing factor of the concrete containing NSR-1 was 230 ~-tm. This value was
much smaller than the values of 300 to 1,200 ~-tm for SRA-1, SRA-2 and SRA-3. This
result means that the air void structure for NSR-1 whould be the most excellent.
Furthermore, Figure 5 shows that the air content and the spacing factor of NSR-1 were
maintained at the superior values of 3.3% and 220 ~-tm between 5 and 30 minutes after
adding water. It is well known that when the spacing factor of hardened concrete is below
200 to 250 ~-tm, hardened concrete shows higher resistance to freezing and thawing [5].
From the results of the spacing factor measurement on fresh concrete, NSR-1 has the
potential for better resistance to freezing and thawing in hardened concrete. In the next
investigations, we would like to evaluate durability of hardened concrete, using NSR-1.
CONCLUSION
REFERENCES
1. Tada S., "Water Transport and Drying Shrinkage Mechanism", Concrete Journal,
Vol.43, No.5, 2005, pp. 43-50
2. Tanimura M. and Tomita R., "Physics and Chemicals of Mass Change Control"
Concrete admixture Handbook, 2004, pp. 78-90
3. Tanimura M., "Countermeasures to Reduce Cracks", Concrete Techno, Vol.23, No.1,
2004, pp. 38-47
4. Fujita Y., "The Mechanism and the Effect of a Drying Shrinkage Reducing Agent",
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
;J>
if ~
s"'s·
0
-
~
§.
"
....
...., ....
v. "'
;;g
u,
t
.
Oc
;." Table 3 - Mixture proportions of concrete
g
Water reducing agent Air control admixture
Unit content(kg/m3) DSRA
W/C
"'
Plain 60.9 50 195 320 899 929 - - - - -
~ C-SR1 PC-I O.ll5 DEF-1 0.012 SRA-1 2.00
....,
...
v. z §j" C-SR2
50.0 48 160 320 869 973
PC-I 0.140 AE-1 0.005 SRA-2 2.00
N
Oo b "'~ 3 C-SR3 PC-I 0.150 - - SRA-3 2.00
z
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~
C-NS1 - DEF-1 0.014 NSR-1 0.70
Not for Resale, 08/10/2015 01:25:47 MDT
Licensee=UNI OF NEW SOUTH WALES/9996758001
"'"'3
....,
N ~
Oo ::tl
Cf
~
"'
Daneshlink.com
Superplasticizers and Chemical Admixtures in Concrete 305
Table 5 - Properties of fresh concrete, using DSRA
Plain C-SRl C-SR2 C-SR3 C-NS1
Formulation
Water reducing agent PC-I PC-I PC-I
dosage (wt"/o/C) 0.115 0.140 0.150
DSRA SRA-1 SRA-2 SRA-3 NSR-1
dosage (wt"/o/C)
· · · - ·io- -~--+--z-.o--1--o-.7--1
80
'§' 70
i§ 50
60
'§ 40
.S:!
0) 30
~ 10
0
0.0 2.0 4.0 6.0
Concentration (wt%)
700 ·X····control
~~ 600 -~ -SRA1(0.3)
3 -•·-SRAl(l.Oli
~ 500 • · SRA1(2.0) i
= -D SRA2(0.3) j
·~ 400 - SRA2(1.0)\
· • SRA2(2.0) ·
i
-'<)
300 · ll SRA3(0J) i
=
] 200
-6- ·SRA3(2.0)f
···0· '
"' 100
0
0 10 20 30
Figure 5-- Comparison of air content and spacing factor determined of 5 and 30 min
after adding water.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
An Amphoteric Copolymer as a
Concrete Admixture
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
EXPERIMENTAL WORK
Materials
The materials used include Type I Portland cement, fine and coarse aggregate, fly
ash and blast-furnace slag, and two chemical admixtures. Cement is from the Taiwan
Cement and complies with ASTM Cl50. Fine aggregate is from the bank of Pearl River,
China. The specific gravity of the sand is 2.59 and its fineness modulus is 2.8. Coarse
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
PDA was first dissolved in water to form an aqueous solution with 20% by
weight in solid content. Cement pastes were made by mixing water and cement, with or
without addition of polymer solution. The water/cement ratio (W/C) was fixed at 0.3; the
polymer/cement ratio (SP/C) ranged from 0 to 1 % by mass of cement, expressed in
terms of dry material. The fluidity of cement pastes was indicated by the measured mini-
slump value, which was determined from a mini-slump test.[9] The apparatus for this test
consists primarily of a hollow mold in the form of a frustrum of a mini cone with the base
40 mm in diameter, the top 20 mm in diameter, and the height 60 mm. The minicone was
sat on a plexiglass plate and filled with cement pastes. Ten strokes were given on the top
of the mini cone before it was raised rapidly. The mini-slump value is indicated by the
measured spread diameter of pastes on the plate. Higher spread diameter value represents
greater fluidity of materials.
Concrete was made by mixing water, binder (cement+ fly ash+ slag), sand, and
coarse aggregate, with or without the presence of polymer solution. The water/binder
ratio (W/B) was 0.37; the polymer/binder ratio (SP/B) ranged from 0.54 to 0.8% by
weight. Table 3 lists the mixture proportions of concrete made with various PDA and
NSF. The workability of concrete was indicated by the slump and slump flow values
determined shortly after the end of mixing (0 min) and 60 min after using the slump test
according to ASTM C143.
Adsorption Measurements
Cement pastes (W/C = 0.6) with polymer (SP/C = 0.5 wt %) were prepared. After
the samples were mixed and centrifuged, the clear solutions on the top of the paste
samples were taken. The residual dispersant concentration in the clear solutions was
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
effect is in the order: P22 > P12 > P32. The optimal AAM/DAE ratio appears to be about
5/1, which results in mixtures with the highest fluidity on cement pastes. Besides, the 60-
min slump value is higher than the initial slump value by 13.5%.
Figure 5 shows the 60-min spread diameter of cement pastes made with different
admixtures used at various dosages. It is clear that the two PDA samples are more
effective in maintaining the fluidity than NSF, as the saturation points occur at lower
polymer dosage and the achieved slump values at the plateau are higher for the former.
Again, the difference in slump-retention property between PDA and NSF could also be
attributed to different adsorption behaviors onto cement particles. Furthermore, the
spread diameter of cement pastes with either Pl2 or P22 at 60 min was observed to be
even greater than that initially, indicating that the polymer with proper reactant ratio and
molecular weight would behave like CAE, has slump-releasing property along with
elapsed time.
Although CAE has slump-releasing property, the addition of CAE into concrete
incorporating NSF could not improve the slump retention of the cementitious material
due to the incompatibility of these two admixtures.[12] As PDA shows the slump-
releasing effect, it will be interesting to know if blending this polymer with NSF could
improve the slump-retention property of the commercial admixture. Figure 6 shows the
0- and 60-min spread diameter of cement pastes incorporating NSF/P22 at different
blended ratios. The admixture dosage is 0.8%. Along with increasing P22/NSF ratio, the
0-min slump increases initially, reaches a maximum value at about a ratio of 4/6, and
then decreases afterwards. In contrast, the 60-min slump increases monotonically with
P22/NSF ratio. Apparently, P22 is compatible with NSF. More importantly, blended
admixture replaced by PDA show improved fluidity retention over those containing pure
NSF.
Workability of Concrete
Concretes proportioned with different admixtures were produced and their slump
and slump flow values were measured, as shown in Table 3 and Table 4, respectively. In
this study, concrete did not flow if it did not incorporate any chemical admixture. As the
polymer was added, both the slump and slump flow values of the concrete increased due
to the dispersing effect of the admixture. For concrete made with P22 polymer ( CP 1) at
0.54% by mass of binder, the slump and slump flow value were about 230 mm and 370 x
370 mm 2 at 0 min, and 275 mm and 640 x 630 mm2 at 60 minutes, respectively. It is
noted that the latter slump flow value is higher than the former one. This agrees with the
test results of cement pastes and confirms that PDA indeed could provide slump-
releasing effect. Besides, concrete made with higher admixture dosage showed greater
initial workability. When the polymer dosage was increased to 0.6%, both the initial and
60-min slump values of the concrete (CP2) were about 265 mm, and the initial and 60-
min slump flow values of the concrete (CP2) were about 650x650 mm 2• In contrast,
concrete made with NSF required higher superplasticizer dosage to achieve similar
slump and slump flow. Both CNI and CN2 had worse workability-retention properties
than the CPl and CP2, because concrete containing NSF shows some slump loss along
with elapsed time.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
CONCLUSIONS
ACKNOWLEDGMENT
The authors are grateful for the support of this work by the National Science
Council of The Republic of China (Contract# NSC-94-2211-E-003-001).
REFERENCES
1. Rixom, M. R., Mailvaganam, N. P., Chemical Admixtures for Concrete, 2"d Edition,
E & F. N. Spon, London, 1986, pp. 12-19.
2. Collepardi, M., "Superplasticizers and Air Entraining Agents: State of the Art and
Future Needs", Concrete Technology Past, Present, and Future: Proceedings ofV.
Mohan Malhotra Symposium, San Francisco, 1994 (SP-144 ), ed. P. K. Mehta,
American Concrete Institute, pp. 399-416.
4. Peng, J. H., Qu, J. D., Zhang, J. X., Chen, M. F., Wan, T. Z., "Adsorption Character-
istics of Water-Reducing Agents on Gypsum Surface and its Effect on the Rheology
of Gypsum Plaster", Cement and Concrete Research, 35 (3), 2005, pp. 527-531.
5. Li, C.Z., Feng, N. Q., Li, Y. D., Chen, R. J., "Effects ofPolyethlene Oxide Chains on
the Perfonrmance of Polycarboxylate-Type Water-Reducers", Cement and Concrete
Research, 35 (5), 2005, pp. 867-873.
6. Jeknavorian, A. A., Roberts, L. R., Jardine, L., Koyata, H., Darwin, D. C., "Condensed
Polyacrylic Acid-Aminated Polyether Polymers as Superplasticizers for Concrete",
Superplasticizers and Other Chemical Admixtures in Concrete: Proceedings of the
Fifth CANMET/ACI International Conference, Rome, Italy,1997 (SP-173), ed. V.M.
7. Hanehara, S., Yamada, K.,"Interaction between Cement and Chemical Admixture from
the Point of Cement Hydration, Adsorption Behavior of Admixture and Paste
Rheology", Cement and Concrete Research, 29 (8), 1999, PP. 1159-1165.
8. Chen, L. P., Hsu, K. C., "Amphoteric Water-Soluble Copolymer for Barium Titanate
Slurries. I. Synthesis and Dispersing Ability", Journal of Applied Polymer Science, 96
(4),2005,pp.l443-1450.
9. Aitcin, P. -C., High-Performance Concrete, E & FN SPON, New York, 1998, pp. 170-
190.
10. Ferrari, G., Cerulli, T., Clemente, P., Dragoni, M., Gamba M., Surico, F., "Influence
of Carboxylic Acid-Carboxylic Ester Ratio of Carboxylic Acid Ester Superplasticizer
on Characterics of Cement Mixtures", Superplasticizers and Other Chemical
Admixtures in Concrete: Proceedings of the Sixth CANMET/ACI International
Conference, Nice, France,2000 (SP-195), ed. V.M. Malholtra, American Concrete
Institute, pp. 505-520.
11. Ye, Y. S., Huang, H. L., Hsu, K. C., "A Water-Soluble Acrylate/Sulfonate Copoly-
mer. I. Its Synthesis and Dispersing ability on Cement", Journal of Applied Polymer
Science, 100 (3), 2006, pp. 2490-2496.
12. Coppola, L., Erali, E., Troli, R., Collepardi, M., "Blending of Acrylic
Superplasticizer with Napthalene, Melamine or Lignosulfonate- Based Polymers",
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Figure 1. Chemical structure of PDA.
c=:J 0 min
200
-60min
'E
.5. 160
I
.!2 120
0
lQ. so
en
40
SP/C(%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Figure 3· Mini-slump of cement paste (o min).
0---L~--L-~-L~~~_L~~
0 20 40 60 80 100
Time(min)
200
'E
.s 160 ,..-
~E
a 120 /
-c
"'
I!! / --o-- P22
c% 80 ~ P12
/ _ . NSF
*E
~ 120
lQ.
!/) 80
40 -
NSF/P22 10/0 8/2 6/4 4/6 218 0/1 0
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
overall influence of superplasticizers on fresh cementitious systems.
The search for alternate approaches to investigate the mode of action of concrete
chemical admixtures has occasionally led to studies on inert mineral systems (l· 6l.
Aqueous slurries of various minerals, typically quartz, titanium dioxide, magnesium
hydroxide and calcium carbonate have been used as reference (or model) systems for
cement pastes. If the reference system adequately represents 'non-hydrating' cement
particles, various physical processes involving the admixture, such as binding to the
mineral surface, calcium ion complexation, solution and paste rheology, etc, can be
characterized and better understood. Through a comparison of observations on the
reference system and on cement pastes, a more accurate description of the mode of action
of concrete admixture should then emerge.
In recent work (7, 8l, the authors have established that calcium carbonate exhibits colloidal
properties very similar to those of Portland cement. Indeed, CaC0 3 and cement slurries,
examined over a wide range of the water/solid ratio, showed remarkably similar
flocculation, stability and rheological behavior shortly after mixing. At longer times (> I
hour), the properties (stability, rheology) of the cementitious slurry depart from those of
the limestone slurry, reflecting the consequences of hydration reactions. The work
described below was undertaken to further explore the potential of limestone pastes as a
model system to elucidate the mode of action of chemical admixtures. In its initial phase,
the investigation focused on the influence of surface active admixtures and their role on
Materials
A common GU (ASTM Type I) Portland cement was used, the composition of which is
reported in the Table 1. The specific surface area and mean particle diameter (weight
average) of the calcium carbonate and cement used are presented in the Table 2. The
surfaces of both solid materials are comparable, although the particles of the limestone
mineral are somewhat finer.
The nature and origin of the different surfactants used in this study are presented in Table
3, together with some of their key physicochemical properties, namely, critical micelle
concentration (CMC), hydrophilic-lipophilic balance (HLB) and their foaming ability.
The surfactants selected cover different groups of surface-active molecules having widely
different properties. The choice of surfactants was also guided by their availability as
pure substances, or as commercial products of known composition.
Methods
Typical conductivity vs. time curves obtained in calcium carbonate pastes, at different
depths in the sample (different electrode pairs), are illustrated in Figure 1 (a); the curves
obtained for cement pastes (9- 10) are somewhat more complex, since the electrolytes which
dissolve upon cement hydration change the conductivity of the solution phase with time.
Nevertheless, the variance in the conductivity values at different depths at a given time is
a direct measure of the heterogeneity of the sample at that instant; such data have been
used to define various indices describing the stability of a cementitious system.
The conductivity curves were analyzed through a protocol described elsewhere (?, 11 ) to
extract density profiles, from which two fronts can be identified:
the bleeding front, which represents an interface between accumulated bleeding
water and the rest of the sedimenting column, and
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The evolution of these fronts with time defines the bleeding and sedimentation curves
shown in Figure 1(b). These curves provide quantitative information on the rate of
surface bleeding and the rate of formation of a layer of maximum compaction at the
bottom of the conductivity cell. The bleeding curve derived from conductivity
measurements was shown consistent with results of a standard protocol (ASTM C232) as
illustrated in Figure l(b); the ASTM method involves collecting the surface bleed water
as function of time. Details on the instrumentation, optimization of the experimental
parameters and data analysis procedures may be found elsewhere (Ill.
Rheology. -- All rheological measurements reported in this study were performed with a
Viscotech rheometer (ATS RheoSystems) using cup and bob geometry at constant
temperature (25 'C). The yield stress values of the pastes were determined via a stress-
growth experiment using a vane tool to avoid "wall slippage" (llJ. In these experiments
the pastes were sheared at a low and constant shear rate(- 0.01 rad/s), and the resulting
shear stress was measured as a function of time; the yield stress value was determined as
the maximum stress in the stress-time curve.
In order to obtain pastes with reproducible viscoelastic behavior, all samples were pre-
sheared 30 s at 200 Pa and then allowed to rest for 5 minutes before the beginning of the
dynamic viscosity measurements. In this way, each paste was reduced to a reproducible
reference state prior to rheological characterization. In oscillation stress sweep
experiments, the shear sweeps, usually between 0.1 and 10 Pa were run at a constant
frequency of I Hz; the elastic and loss modulus and the phase angle were determined in
the linear viscoelastic region. Frequency sweeps were also performed at frequencies
ranging from 0.01-IO Hz; the applied stress during the frequency sweeps was well below
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
the critical stress of the particular sample. From these experiments, other rheological
parameters were determined, namely the relaxation time (inverse value of the frequency
at which the phase angle reaches 45') and the zero-shear viscosity (low frequency
extrapolation of the measured complex viscosity).
Adsorption. -- The adsorption of admixtures onto mineral particles in the pastes was
determined by measuring the residual concentration of the admixtures in the liquid phase
of the concentrated suspensions of limestone or cement particles. The pastes for
adsorption measurements were prepared in the same way as pastes for stability and
rheological measurements (see below). A one hour period was considered to be sufficient
for adsorption equilibrium at ambient temperature. After centrifugation of the paste
sample, the supernatant was diluted and the residual concentration of the admixture was
determined by the suitable analytical technique. The admixtures which adsorb in UV and
visible region of the spectrum (PNS and Triton X-100 surfactant) were quantified
spectrophotometrically using a Hewlett Packard 8452A Diode Array spectrophotometer.
The residual concentration of the other organic admixtures in the solution was obtained
by chemical oxygen demand (COD), which determines the total organic carbon in the
solution (IJJ.
Mixing protocols and measurement sequence. The samples used for stability and
rheological measurements were prepared with a Hobart mixer according to ASTM
procedure C305. The required amount of water or electrolyte solution (10 g/1 NaCl in
calcium carbonate pastes) and admixtures were introduced into a mixing bowl before
mixing, except as otherwise indicated. The solids were then slowly introduced in the
mixer and the time recorded as time zero for bleeding and sedimentation measurements;
the total mixing time was 4 minutes. The sample was then transferred into the
conductivity cell and conductivity readings started within 5-6 minutes of initial
solid/liquid contact.
The remainder of the sample was used for rheological characterization as described
above, and for evaluation of air entrainment in the pastes when necessary. The latter
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
measurements were made 6 minutes after mixing through the following paste protocol. A
portion of the sample was transferred to overfill a Plexiglas cylinder (9 .5 em height and
5.2 em interior diameter), and the excess paste was skimmed-off with a straight edge. The
air content of the sample was determined by measuring the mass of the sample in the
cylinder, and by comparing this mass with the mass calculated for an air-free paste. Paste
fluidity measurements were also performed 8 minutes after mixing on another portion of
the sample, using the Cantro mini-slump method (l 4l; the paste spread diameters reported
are obtained as an average of 4 readings.
The various data illustrating the behavior of the reference limestone and cement pastes, as
well as the influence of admixtures on the properties of these pastes, are collected in
Tables 4 and 5 (4: bleeding and sedimentation, 5: rheology). For discussion purposes, the
following quantities have been extracted from bleeding and sedimentation curves similar
to those in Figure 1 (b):
All these rheological parameters have been obtained in measurements carried out at very
low shear stresses, so they should be relevant to bleeding/sedimentation phenomena
which take place at low stresses, involving small deformations in the sample.
In a previous study <7J the authors concluded that calcium carbonate (CaC0 3) appeared as
the most relevant model mineral to elucidate the factors contributing to the stability and
the rheological properties of cement-based materials at short times (i.e., <I hour). CaC0 3
has surface properties and flocculation behavior comparable to slightly hydrated cement
particles, namely, an irregular spheroid shape, a low surface charge (zeta potential) in
water and a very low solubility, even at high pH (- 12-13) relevant to cementitious
systems. To further illustrate the similarity in the behavior of CaC0 3 and cement pastes,
the rheological properties and the bleeding-sedimentation behavior of the following
systems were examined: a CaC0 3 paste (LIS=O.S) and a cement paste (LIS=O.SS). These
LIS ratios were selected to yield stable pastes with similar fluidities.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The paste rheology reflected in the minislump test results (Table 5) shows minor
differences in the paste fluidities, minislump spread values decreasing in the order :
cement> CaC0 3 > CaC03 pH 12.5.
Observations on paste bleeding and sedimentation -- Figure 3 illustrates the bleeding and
sedimentation behavior of the CaC03 and cement pastes obtained from the analysis of
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
conductivity measurements. As indicated by the limiting slopes of the curves, the initial
bleeding and sedimentation rates (respectively Q0 and V,) of the two pastes are identical
within experimental error. However, after approximately 90-120 minutes, the limestone
pastes exhibit a rapid increase in the bleeding and sedimentation rates, which are believed
to be associated with the formation of channels (Qch) (see discussion below). At the same
time, in the cement paste, the hydration reactions begin to play a dominant role in the
flocculation-sedimentation of the particles. The hydrating cement particles form a three-
dimensional network which inhibits the separation of the liquid and solid phases, as is
readily apparent from the data in Figure 3. Hence, the accelerated bleeding and
consolidation processes are not observed in the cement paste. In accordance with
observations on the rheological properties, the presence ofCa(OH) 2 stabilizes the CaC03
paste, reducing the initial bleeding rate (Q 0).
The remarkable similarity in the early bleeding and sedimentation behaviors of the
CaC0 3 and cement pastes confirms that the key physical effects responsible for bleeding-
sedimentation behavior in cement paste are adequately represented by those which occur
in the CaC0 3 paste. Consequently, the factors influencing the early migration of liquids
and solids in cementitious systems, and the influence of many types of chemical
admixtures on these processes, can be usefully investigated in model limestone pastes.
Likewise, observations on paste rheology confirm that investigations on the mode of
action of admixtures in limestone should provide information highly relevant to
cementitious systems.
The surface active compounds selected to elucidate the influence of air-entraining agents
(AEA) on the properties of cementitious systems comprised various types of molecules,
namely:
a foaming anionic surfactant typically used as an AEA, Sodium Decylsulfate
(DSSi' 5l
a foaming non-ionic surfactant of size comparable to DSS, Triton X-1 00
a low molecular weight ethylene glycol polymer (PEG) selected as a reference
organic polymer with little surface activity
a non-foaming low molecular weight propylene glycol polymer (PPG) representing
a surface active polymer
a block EG-PG-EG surfactant co-polymer with low foaming properties labeled as
L-61
To further assist the interpretation of paste properties in the presence of surface active
compounds, the surface tension of saturated CaC0 3 solutions were determined as
function of increasing surfactant concentrations; the results are illustrated in Figure 4. As
expected from their relative hydrophobic character, PEG induces only a minor decrease
in the surface tension of water, while the other surfactants lead to a significant lowering
of the surface tension. In some cases, DSS, Triton and L-61, a significant slope change is
observed in the surface tension vs concentration curve; this signals the occurrence of
micellization, a surfactant aggregation process; the sharper the break in the surface
tension curve, the more cooperative the aggregation process (l 6l.
As with any concrete chemical admixture, the interaction of surfactant molecules with the
surface of the particles (CaC0 3 or cement) may play a role in their overall effect on these
pastes. To asses the significance of such effects, the binding of several surface active
molecules was investigated, and the adsorption isotherms obtained are presented in
Figure 5. The results indicate that CaC0 3 binds the L-61 quite strongly in a Langmuir-
type adsorption, while Triton X-100 shows a Freundlich type adsorption isotherm
(commonly known asS-type); DSS is only weakly adsorbed. On the other hand, the same
surfactants were only faintly adsorbed by cement particles.
Effect of surfactants on paste rheology -- The addition of (non or low foaming) polymeric
surfactants to either CaC03 or cement pastes has only marginal effects on the rheological
parameters of these pastes (Table 5). The values of <o, G' and G" all remain relatively
constant through the range of surfactant concentration investigated in CaC0 3 pastes at
equilibrium pH(- 10), or at pH 12.5, as well as in cement pastes. In CaC0 3 pastes at pH
12.5, a low concentration ofL-61 reduces the G' and G" values by approximately 10%,
but little effect is observed at higher concentrations, or in other pastes; although this
effect is marginal, it may be real, considering the important influence of L-61 observed in
other properties described below.
In the presence of the foaming surfactants DSS and Triton, but without entrained air, the
yield stress values of all the pastes is decreased significantly; the elastic and loss moduli
also decrease moderately with increasing surfactant concentration (the paste pre-shearing
operation may entrain some air, which will also contribute to decrease G' and G"
values). These results point to a light fluidifying influence of the surfactants, which is
also manifested by a modest increase ofthe minis lump spread areas. On the other hand, if
air entrainment is allowed, the DSS surfactant induces a low increase in the yield stress of
The results examined above support the conclusion that, at short times, the rheological
properties of cement and limestone pastes are not significantly altered by the surfactants
in the absence of entrained air. Any influence of the surfactants on interfacial tension,
drag reduction, or particle-particle interactions does not appear to modify the dynamic
response of the pastes under shear stress. In the presence of entrained air, the fluidity of
pastes is decreased; a modest increase in the yield stress values of pastes having high air
content suggests that the surface tension forces merge boundary films of neighboring air
voids into a network, which contributes to stabilize the whole sample volume. This is
further supported by a decrease of the phase angle value in air-containing pastes,
indicating an increased solid-like behavior.
The influence of surfactants on the migration of water and solid particles may be quite
complex, since in contrast to rheological properties measured at very short times after
sample homogenization, the bleeding-sedimentation experiments extend over several
hours, during which the paste properties evolve continuously. The time evolution of
particle-particle interactions and of flocculation-consolidation processes will affect the
mechanism through which the solids can sediment or consolidate in the sample column,
as well as the water transport through and out of the consolidating materials.
In studies on settling processes in cement-based systems, Loh et. al (I?J concluded that
channeling can occur when the governing process is sedimentation in the upper part of
the sample, while consolidation is taking place in the sedimented zone at the bottom;
however, if the settling is due to consolidation only, no channeling should occur. The
bleeding and sedimentation curves derived from paste conductivity measurements show
an initial period of constant bleeding and constant sedimentation rates (Figure I). Beyond
the initial period, sharp increases in the bleeding and sedimentation rates are observed;
this observation points to changes in the mechanism of the phase separation process,
which can be attributed to the onset of channeling.
The data collected in Table 4 provides some indication on the influence of surfactant
admixtures on bleeding sedimentation processes. These will be examined in the same
order as that followed for examination of the rheolgical properties.
A cursory examination of the cumulative bleeding volumes collected over CaC03 and
cement pastes shows that, in the absence of entrained air, surfactants tend to slightly
increase the total bleeding volume, except perhaps for the CaC0 3 pastes with lime. In
pastes with entrained air, the total bleeding is substantially reduced, as expected from
observations in mortars and concrete (ISJ_ The bleeding volume increase is likely due to
the reduction of the interfacial tension of water, which provides increased spontaneous
access through small capillaries in the porous mass of the solids during sedimentation and
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Accelerated bleeding and sedimentation at later stages from Ooh values -- As noted
earlier, in calcium carbonate pastes, the bleeding rate increases substantially after - 90-
120 min, as shown by the Qch values; the latter represents the steep downward slope of
the bleeding curve shown in Figure l. The important change occurring in the bleeding
process, and the concomitant sedimentation· consolidation processes, are significantly
affected by the presence of surfactants. In all limestone pastes investigated in the absence
of entrained air, all surfactants increase the Qch values (except PEG which has low
surface activity), reflecting a destabilization of the paste. The magnitude of the effect
depends on the nature and concentration of the surfactant; the destabilizing effect
manifested by the Qch values exhibit sharp variations as function of surfactant
concentration.
The introduction of PPG in limestone pastes induce a slight increase in the Qch rate
parameters, the other surfactants strongly enhance the Qch values. The most important
effect is observed with L-61 for which the bleeding-sedimentation (consolidation) curves
in limestone pastes are illustrated in Figure 6. As can bee seen from the bleeding curves,
the L-61 surfactant enhances bleeding at low concentrations, but stabilizes the pastes at
the higher concentration examined. A similar but less noticeable effect is observed in
systems containing Triton or DSS. The bleeding and sedimentation curves presented in
Figure 7 shows the influence of L-61 on limestone pastes in lime-saturated water; in the
latter, L-61 stabilizes the limestone paste at all concentrations tested.
When sufficient air entrainment is allowed in the paste, as with DSS added at the
beginning of the experiment, the Qch value decreases markedly. This is consistent with
observations on paste rheology, i.e., the presence of entrained air improves the stability of
the paste.
The influence of L-61 on the bleeding and sedimentation curves obtained with cement
pastes is shown in Figure 8. As noted earlier, channeling effects were not observed in the
Mechanistic aspects-- The bleeding rate enhancement observed in the 1-2 hour period is
believed to be related to the onset of channeling through the consolidating sediment
column. In this framework, the ability of surfactant molecules to promote channeling in
limestone pastes (pH~ 10) apparently increases in the following order: DSS <Triton<
L-61. This order may be dictated by the intrinsic molecular features and solution
properties of the surfactants, namely, HLB, surface tension reduction, water binding, type
of micellar aggregate, solution viscosities, etc. However, because the water migration in
bleeding processes and the movement of solids in the sedimentation-consolidation
processes are inherently coupled, the impact of surfactants may also be determined by
their interaction with the solid particles, and the consequences of these interactions on
interparticle forces, static (interaction potentials) or dynamic (drag).
With respect to surface tension, it was found that, although PPG induces surface tension
reduction comparable to L-61 (Figure 4), its effect on bleeding and sedimentation
phenomena is negligible. Also from the surface tension data in Figure 4, the order of
surface tension lowering at the CMC of the various surfactants is Triton > DSS > L-61;
since L-61 is most potent for accelerating the bleeding and sedimentation, effects other
than surface tension reduction must be involved.
The surfactant concentration, at which the rate of the solid/liquid phase separation is the
most atiected, appears related to the CMC values of these surfactants in CaC0 3 -
saturated solutions (Figure 4). The concentrations of maximum effect for L-61 and Triton
X-100 are respectively 0.01 and 0.025 wt% with respect to the solids; this corresponds to
solution concentrations of ~5x 10-3 and ~1.3 x 10-2% (neglecting adsorption), which is
close to the CMC values of these surfactants. Hence the role of a surfactant in the
acceleration of phase separation processes is related to its ability to form micelles,
although, once micelle formation occurs, the surfactant effect levels-off (in spite of the
fact that much more surfactant is present in the system).
The propensity of surfactants to form micelles should be related to their HLB values
(within the same family of compounds). Attempts to correlate the de-stabilizing effects of
the surfactants on limestone pastes to their HLB values, however, failed to show a
significant relation.
On the basis of the above observations, the accelerating influence of surfactants on the
bleeding-sedimentation processes does not appear dominated by their solution properties.
On the other hand, the surfactant-particle interaction, as evidenced by the adsorption
isotherms presented in Figure 5, can be seen to play an important role. The surfactant
binding affinity, illustrated by the shape of the adsorption curve (extent of adsorption at
low concentration), increases in the order DSS <Triton< L-61, as does the surfactant
effect on paste destabilization.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
CONCLUSION
The results derived from the present investigation indicate that surfactants, at
concentrations typical of concrete admixture dosages, can exert complex influences on
some of the properties of cementitious or model limestone pastes, even in the absence of
entrained air.
For a group of surfactants having widely different properties, detailed paste rheological
measurements demonstrated only marginal surfactant effects in the absence of entrained
air, up to 0.1 wt% dosages. Similarly, surfactants have little influence on the initial
bleeding and sedimentation of stable cement and limestone pastes. However, at longer
times, typically > 1 hour, surfactants can enhance the bleeding and sedimentation
processes associated to channel formation in the consolidating sediments; the magnitude
ofthe rate enhancement depends on the properties and concentration of the surfactants.
The surfactant effects were observed mostly in limestone pastes; in the cement pastes
investigated here, the hydration reactions and products hinder the bleeding-consolidation
process after 1-2 hours, i.e., before channel bleeding becomes significant. Clearly,
however, in less stable cementitious systems (higher w/c, presence of retarders), channel
bleeding will likely set-in before the paste microstructure develops to stabilize the
systems. Surface active admixtures could then significantly alter the properties and
behavior of these materials.
REFERENCES
2. Pierre, A., Mercier, R., Foissy, A. and Lamarche, J.M. "The Adsorption of Cement
Superplasticizers onto Mineral Dispersions", Adsorpt. Sci. Technol., Q, 219-231
(1989).
4. Pierre, A., Lamarche, J.M., Mercier, R. and Foissy, A. "Adsorption d'un tluidifiant
du ciment sur le carbonate de calcium", Cern. Concr. Res., 19, 692-702 (1989).
5. Comparet, C., Nonat, A., Pourchet, S., Mosquet, M. and Maitrasse, P. "The
Molecular Parameters and the Effect of Comb-Type Superplasticizers on Self-
Compacting Concrete: A Comparison of Comb-type Superplasticizer Adsorption
onto a Basic CaC03 Medium in the Presence of Sodium Sulphate", Proceedings of
the 71h CANMET/ACI International Conference on Superplasticizers and Other
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Chemical Admixtures in Concrete, Berlin, 2003, Malhotra, V.M. (Ed.), SP-217-13,
pp. 195-209.
7. Mikanovic, N., Jolicoeur, C., Khayat, K. and Page, M. "Model Systems for
Investigation of the Stability and Rheological Properties of Cement-Based
Materials", Proceedings of the 8'h CANMET/ACI International Conference on
Recent Advances in Concrete Technology, Montreal, Canada, 2006, Malhotra, V.M.
(Ed.), SP-235-22, pp. 323-356.
10. Pavate, T.V., Khayat, K.H. and Jolicoeur, C. "In-Situ Conductivity Method for
Monitoring Segregation, Bleeding, and Strength Development in Cement-Based
Materials", Proceedings of the 6th CANMET/ACI International Conference on
Superplasticizers and Other Chemical Admixtures in Concrete, Nice, 2000,
Malhotra, V.M. (Ed.), SP-195-33,pp. 535-559
12. Saak, A.W., Jenning, H.M. and Shah, S.P. "Influence of Wall Slip on Yield Stress
and Viscoelastic Measurements of Cement Paste", Cern. Caner. Res., 11. 205-212
(2001).
13. APHA Standard Methods, 19th ed., p. 5-15, method 5220 D (1995).
15. Jolicoeur, C., To, T.C., Nguyen, T.S., Hill, R., and Page, M. "Investigation of
Physico-Chemical Aspect of Air Entrainment in Cementitious Systems",
Proceedings of the 7th CANMET/ACI International Conference on Superplasticizers
and Other Chemical Admixtures in Concrete, Berlin, 2003, Malhotra, V.M. (Ed.),
SP-217-39, pp. 595-619.
16. Shaw, D.J. "Introduction to Colloid and Surface Chemistry", 2"d Edition,
Butterworth & Co. Publishers Ltd., England, 1970.
17. Loh, C.-K., Tan, T.-S., Yang, K.-Y. and Wee, T.-H. "An Experimental Study on
Bleeding and Channelling of Cement Paste and Mortar", Adv. Cern. Res., .ill( l ), 1-
16 (1998).
18. Powers, T.C. "The Properties of Fresh Concrete", John Wiley & Sons, Inc., New
York, 1968.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Table 1 --Properties of Portland ~ement
Composition(%)
Chemical analysis
CaO 62.9
SiO, 21.0
AhO, 4.2
Fe,O, 3.1
MgO 2.3
so, 2.7
Na,Oeq. 0.76
Loss of ignition
Mineralogy
c,s(%) 52
c,s(%) 21.5
c,A(%) 5.7
C.AF(%) 9.5
Physicall!r!ll!erties
% passing45 ~ 96
Specific surface BET (m2/kg) 1200
Specific surface Blaine (m2/kg) 345
CMCat Ross-Miles
Surfactant Fonnula
Mw 25°C HLB foam height
(g/mol)
(o/om/v) (!nmj'
n-decyl
n-C 1oH21 0S0 3Na 260.3 0.8 40 150
sulfate Na
CsH 1,CJ{;O
Triton 625 1.5xl0"2 13.5 110
(CH,CH,O)..,ofl
HO(CH2CH20),·
Pluronics
(CH2CH(CH3)0)y- 2000 1.4xl0-3 3 10
L-61
(CH2CH20).H
PPG HO(CH2CH(CH3)0).H 1025 . 7.3 .
PEG HO(CH,CH,Ol,H 2000 . 25.3 .
* 0.1% solution
System stability
(~)
System Admixture Air Bleeding Qo Vs
(%) Jml)_ _(cm!h} Jcm!h}
<1.1 42. w us 4..0
0.01% DSS 2.9 35.5 0.15 0.4 3.0
0.025%DSS' 9.3 29.5 0.1 0.15 2.6
0.05%DSS' 16.2 17 0.04 - 2.45
0.01%DSS 0.4 44 0.35 0.95 4.0
0.05%DSS" 0.4 43.5 0.3 0.75 4.5
o.t% oss" 0.5 45 0.25-0.3 0.8 4.5
a
u..
0.005% Triton
0.01% Triton
< 0.1
<0.1
43.5
43.5
0.25-0.3
0.35
0.9
0.85
3.7
4.4
0.025% Triton <0.1 43 0.3 1.15 3.8
~ 0.05% Triton < 0.1 46 0.35 1.05 4.3
a
e 0.005%L-61 <0.1 44 0.4 1.15 . 4.35
e. 0.0!%L-6l < 0.1 44 0.4 1.5 4.6
.,....
0
0.025%L-61
0.05% L-61
< 0.1
< 0.1
43.5
43.5
0.5
0.4
1.2
1.15
4.0
4.3
::J 0.1%L-61 <0.1 45 0.25 0.8 4.3
O.Ol%PPG <0.1 44 0.35-0.4 0.6 3.9
0.05% PPG <0.1 45 0.4-0.45 0.7 4.2
O.l%PPG <0.1 44 0.35 0.65-0.7 4.15
0.15% PPG < 0.1 44 0.3 0.5-0.55 4.0
O.OI%PEG <0.1 41 0.35-0.4 0.65 4.35
0.05%PEG <0.1 40 0.4 0.55 4.0
O.l%PEG <0.1 44 0.4 0.65 4.4
<:l.t 30 tJ.lS -~ :M'
e 0.0 I% Triton < 0.1 26 0.2 - 5.1
!"l 0.025% Triton <0.1 26.5 0.2-0.25 - 4.7
i-=~i
0.05% Triton <0.1 25 0.2-0.25 - 5.45
O.Ol%L-61 <0.1 28 0.3 - 5.5
::J 0.025%L-61 <0.1 28 0.25 -- 5.7
..
0.05% L-61 <0.1
'~:f.}'
27
u
0.25
... 5.2
tt;.;
~.!
. ..
Ei ..
0.01% Triton
0.1% Triton
<0.1
<0.1
8.5
9
''"
0.3
0.25-0.3
-
--
6.2
5.8
uc. 0.01% L-61 < 0.1 10 0.25 8.2
0.1% L-61 <0.1 II 0.25 - 8,4
' Admixture added at tbe beginning of the mixing procedure directly in liquid phase
" Admixture added at the end of mixing procedure manually. All other surfactants were
added at the end of mixing procedure.
Rheological properties
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
[><
"' 0.025% Triton 80 6.8 51.5 33.6 33.2
!ii 0.05% Triton 80 6.6 36.0 22.1 31.5
'"=
"8=-
~
0.01%L-61
0.025% L-61
77
77
7.35
7.3
56.1
57.7
39.3
43.9
35.0
37.3
0.05%L-61 77 7.3 55.3 42.4 37.5
..."' ... 0.01% Triton 96 5.05
,5}
18.4 16.9 42.6
'
20
s ~ 0.1% Triton 98,5 4.8 15.6 14.2 42.3 21
d Q, 0.01% L-61 96,5 5.2 18.4 17.6 43.7 22
0.1% L-61 94 5.15 18.9 19.1 45.3 17.5
' Admixture added at the beginning of the mixing procedure directly in liquid phase
" Admixture added at the end of mixing procedure manually. All other surfactants were
added at the end of mixing procedure.
4
Time (h)
JOW-------------------------------~30
>t~ + •• (b)
f:
X
=s 15
~.
Position of •
sedimentation front
-.
+X
+X
.x
Position of
bleeding front
)
29
28
g +------· +~">+ XX
.::: 10
...
= ••
• Sedimentationl 27
~5 ; +Bleeding
-~ASTM C23~-.
0-=·~----.----.----....-----+26
0 4 8
Time (b)
Figure 1 -- (a) Typical conductivity curves obtained in experiments with CaC0 3 pastes
at L/S= o.s (pH- to); and (b) Bleeding and sedimentation curves derived from
conductivity measurements.
<><>~.,
•
Gj'
e:,
t?
1000 •
10
0,1
(a)
I
<>. 10
Shear stress (Pa)
100000
~
'i'
e:,
G
1000 •
(b)
g.
10
0,1 I 10
Shear stress (Pa)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
80
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
e:z; 0 x ..
......... ....
.- ..
v. 0 0 0
! 60
.
~ ~
·!= ill.:" • .. •)(
.
=
~
~
!l 40 .
i! XDSS
.L-61 -:t·~·.1"-x~-:.x ·><
"'"' .l TritonX-100
+PPG(M.W.J025)
0 PEG (M.W.2000)
t "
I ..
DSS
Triton
20
-5 -4 -3 -2 -I
LogC(%)
i1il .L-61
..
~ 0,3
04
' A Triton X-100
~
.'"
!
~ 0,2
0,1
z
Figure 5 --Adsorption isotherms for surfactant molecules on CaC0 3 (solid lines) and
cement particles (dashed lines).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
•Reference
! ... 0.005% !Ail
..
."
-== 29
.0.01% L-61
X 0.025% !Ai 1
X0.05%L-61
1. •>¢Ill 0 •
00.1%!Ai1
·~ 0 • ----
'l; 28
•:u. • ::K • o 1M o
,g
~ (a)
27+------.----.----------~
0 6
Time (b)
30
Ez5
~
(b)
0
• • • ::1( 0 0
....
,g" 20 .X A 0 •Rc:ferc:ncc
... 0.005% L-61
e
115
.uo .0.01%L-61
lal(j.() X0.025% L-61
.....
:10
.. ~
X0.05%L-61
OO.l%L-61
~,. ~
(II:
0
a
0 4
Time (b)
!g
30
.... ~.
+Reference
.0.01%L-61
.
ol: 29,5
= ~~ :.;
& O.o25% L-61
X0.05%L-61
i
:;;; •
co
-~
29
• ·~•
l • I •
'
(a)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
28,5
4
Time (h)
!25
30
(b) • • •
X
X
;;
£20 •x •Reference
•o.oJ%L-61
:ae !5 ~
•x &0.025% L-6!
XO.OS~L-61
...
~
.,.
: 10
c
'!:
-~
=-
0
0
••- 4
Time (h)
e
~ •• i •.0.01% L-61
Reference
ij
..
0
-= 29,8 • .O.I%L--61
•!!!
....
..."
~ '·• .~
• t
'E 29,6
"
~
's:l • JJ ...
0
~ (a)
29,4
0 0,5 I 1,5 2,5
Time (h)
30
'""'
]Y
=
~ 20
(b)
ffi~ -I
eference
Ol%L--61
l%L-61
a
.... .......
115
...:w
0
<= •
• '
-~ 5
•
••
~
0
0 0,5 1,5 2,5
Time (h)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: In the present work, the influence of two organic admixtures (a vinyl
copolymer and one based on polypropyleneglycol in the alkaline activation process
of blast furnace slag has been studied. The effect of these admixtures on the
rheological and mechanical properties, volume stability and durability ofwaterglass-
activated slag mortars has been studied.
The incorporation of these organic admixtures delays the activation
process of waterglass-activated slag pastes and, consequently, the initial and final
setting time. The presence of the admixtures also increases the total heat released
during the activation. Vinyl copolymer and shrinkage-reducing admixtures both
decrease the yield stress of waterglass-activated mortars during the first minutes
after mixing, and consequently, mortars with a lower liquid/solid ratio and higher
mechanical strengths are prepared. On the other hand, the incorporation of the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Francisca Puertas holds a Dr. Sc. in Chemistry and is a Scientist Researcher at the
Eduardo Torroja Institute and she has been working on building materials for 20 years.
She is author or coauthor of more than I 00 papers and editor in chief of "Materiales de
Construcci6n".
Phil Banfill holds a Ph.D. in Construction Materials and has been Professor of
Construction Materials at Heriot-Watt University, Edinburgh since 1995. He is author or
coauthor of 2 books and over 50 papers on the rheology of cement based materials for
construction.
INTRODUCTION
The use of organic admixtures in the preparation of Portland concretes has become a
common practice during the last decades. Superplasticizers are the admixtures most
frequently used, and the resulting concrete has excellent rheological properties and a low
water/cement ratio, which ensures high strength and durability. Several authors (Mollah,
M. Y.A., et al, 2000; Berke, N. S. et al, 1997; Palacios, M. and Puertas, F., 2005) have
studied the mechanism governing the effect of these admixtures on Portland cement
systems, which consist of their adsorption on the cement particles, where the electrostatic
or steric repulsive forces generated prevent flocculation. Shrinkage-reducing admixtures
based on polypropyleneglycol, in turn, decrease the surface tension of the water in the
pores, reducing the capillary stress in the pore structure and, as a result, shrinkage when
the water evaporates (Berke, N. S. et al, 1997).
Nonetheless, very few studies have been conducted on the behaviour and properties
acquired by alkali-activated slag cements and mortars in the presence of these admixtures
(Palacios, M. and Puertas, F., 2005). Moreover, the results obtained vary widely due to
differences in the vitreous composition of the slag, the nature and concentration of the
alkali activator, and the type and dosage of the admixture used. Palacios and Puertas
(Palacios, M and Puertas, F., 2004; Palacios, M and Puertas, F., 2004-b) reported
substantially reduced fluidifying action by some superplasticizer admixtures, specifically
those based on polycarboxylates or vinyl copolymers, in alkali-activated slag systems as
a result of the modification of their chemical structure in highly basic media (pH> 13 ). By
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
However, nothing is known about the effect that admixtures have on other properties of
alkali-activated slag (AAS) mortars. Consequently, the objective of the present study is to
analyze the influence of a vinyl copolymer superplasticizer admixture and a shrinkage-
reducing admixture based on polypropyleneglycol on the mechanical strength, shrinkage
and durability of AAS mortars.
Spanish granulated blast-furnace slag and CEM I 42.5N Portland cement were used in
this study. The chemical composition of the two materials is shown in Table 1. The
fineness of the slag and the Portland cement were 325 and 360 m2/kg, respectively and
the vitreous content of the slag was 99%.
The blast furnace slag was activated with waterglass (Na20·nSi02 ·mH20 + NaOH)
solution with 4% Na20, by mass of slag, and a Na20/Si0 2 ratio in the range 1.0-1.2. I%
of a vinyl co-polymer (V) and I% of a shrinkage-reducing admixtures (SRA), by mass of
slag, were added to the alkaline solution.
Alkali-activated slag (AAS) and Portland cement mortars were prepared. The binder I
aggregate ratio used was 1:2 by mass. Spanish and European standard UNE-EN I96-I
sand with a Si02 (quartz) content of over 98% was used. The liquid/solid (Us) ratio used
was determined by the slump test described in Spanish standard UNE-80-1I6-86. The
composition of the different mortars prepared is shown in Table 2.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Tests conducted
- Mechanical strength
AAS and cement mortars (4 x 4 x 16 em) were prepared according to Spanish standard
UNE-I97-l. These waterglass-activated and cement specimens were cured for 48 hours
and 24 hours, respectively, at 20± 2 oc and 99% relative humidity (RH). They were
subsequently removed from the moulds and cured until the day of the strength test at 99%
relative humidity (RH).
The flexural and compressive strengths of the specimens were measured after 2, 7 and 28
days as specified in European standard UNE-EN 196-1.
- Mortar shrinkage
AAS and Portland cement mortars were prepared with a binder: sand ratio of 1:2. The
effect of 1% of the vinyl copolymer and the shrinkage-reducing admixure on the
rheological properties of the AAS mortars was studied.
- Durability
Cubic specimens (3-cm) of the mortars shown in Table 2 were prepared. The specimens
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
were kept in a moist chamber for 28 days, whereupon the respose to carbonation and
exposure to high temperature were studied.
Carbonation test
Mortars were placed in the carbonation chamber where the relative humidity was kept at
43.2% using a saturated K 2 C0 3 solution according to E 104-02 ASTM "Maintaining
Constant Relative Humidity by Means of Aqueous Solutions". The mortar specimens
were removed after four and eight months of exposure. Some of the specimens were
protected with insulating tape on all but one face. The chamber was COrsaturated by
filling it with the gas at atmospheric pressure twice a day.
The phenolphthalein test was used to measure the carbonation front on the side exposed
to C0 2 • Compressive strength, total porosity and average pore diameter, before and after
the carbonation, were measured.
Mortars were exposed for 2 hours to temperatures in the range 100°C-1 000°C.
Subsequently, the effect of the temperature exposure on the mechanical strength of the
mortars was determined.
Mechanical strength
The development of flexural and compressive strength of Portland cement and AAS slag
mortars, with and without admixtures, are given in Figure 1.
At 2 days of curing the mechanical strength of AAS mortars without admixtures are
lower than the mechanical strength of cement mortars. However, at 7 and 28 days of
curing the mortars present similar strength behaviour. The incorporation of organic
admixtures to the AAS mortars slightly increases the flexural and compressive strength
due to the decrease of the liquid/solid ratio that the admixtures induce.
Shrinkage of Portland cement and waterglass-activated slag mortars, with and without
admixture, under both curing conditions (99% and 50% RH), are given in Figure 2.
In both curing conditions, AAS mortars experience a shrinkage four times higher than the
cement mortars.
This difference in shrinkage is essentially due to two factors: on the one hand,
waterglass-activated slag mortars have a higher percentage of small pores (diameter<
0.01 ~-tm) (Shi, C. 1996; Shi, C., 1992), generating higher capillary stress in the pore
structure, and on the other, their C-S-H gel has a higher water content, which causes
greater shrinkage as water is lost during drying.
At RH=99%, the inclusion of 1% and 2% SRA reduced shrinkage by 50% and 85%,
respectively, with shrinkage values, in the latter case, similar to those recorded for
Portland cement. Shrinkage in waterglass-activated slag mortars containing the V
admixture was slightly smaller than in mortars with no admixture, although after 365
days the results were practically the same.
The flow curves (hysteresis loops) for waterglass-activated slag and Portland cement
mortars are given in Figure 3, and presented as torque on the impeller (proportional to
shear stress) and speed of rotation (proportional to shear rate).
There is a visible difference in the rheological behaviour of AAS and Portland cement
mortars. The curves for AAS mortars fit the Herschel-Bulkley curve (equation 1), while
where 1 is the shear stress, 'to is the yield stress, J..l is the plastic viscosity , r
is the shear
rate and A and B are constants. The width of the hysteresis loop can be taken as an
indication of the extent of shear-induced structural breakdown during the test. This
appears to be more significant in the AAS mortars which suggests differences in the
water-binder interactions in the mortars.
A comparison of the yield stress values for the two types of mortar shows that this
rheological parameter is up to two times greater in Portland cement than AAS mortars.
This is an indication that the interactions between the floes formed are much stronger in
Portland cement mortars than in waterglass-activated slag mortars (Tattersall, G. H. and
Banjill. P. F. G., 1983). Taken with the comments on loop shape above this suggests that
further research on the rheology and structure of fresh AAS mortars is justified.
Fig. 4 shows that in waterglass-activated slag mortars, the admixtures initially reduced
the yield stress by up to 50%, although after 9 minutes, the presence of the admixture
induced no significant change in the value of this rheological parameter with respect to
AAS mortar without admixture. The loss of the fluidizying properties of the vinyl
copolymer in waterglass-activated slag systems is due to its instability in the highly
alkaline waterglass solution, which hydrolyses the admixture and changes its chemical
structure.
Durability
Carbonation
The phenolphthalein test showed that the carbonation front was no deeper than 1 mm in
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Portland cement. Carbonation was much more intense and deeper in the activated slag
mortars, however. The inclusion of admixtures V and SRA did not protect the interior of
the mortar specimens from carbonation.
Table 4 gives the variation in total porosity and mean pore diameter for mortars A, B-1, C
and D at the three carbonation times. Total porosity declined slightly with carbonation
time in all cases, but most markedly in cement mortars. Whereas the mean pore diameter
increased significantly with carbonation in waterglass-activated slag mortars, it declined
substantially in cement mortars.
According to the phenolphthalein test, carbonation was deeper and more intense in
waterglass-activated slag than in Portland cement mortars, and this concurs with results
reported by other authors (Bakharev, T. et al, 2001). In Portland cement mortars, the
portlandite is carbonated first, quickly reacting with the C0 2 to form CaC0 3 •
Subsequently, as the amount ofportlandite begins to decline, the C-S-H gel is carbonated,
generating a mix of C-S-H gel with a low Ca content and silica gel. The precipitation of
CaC0 3 crystals in the pores reduces total porosity and mean pore diameter, obstructing
C0 2 diffusion through the mortar while at the same time increasing mechanical strength.
Conversely, as Bakharev (Bakharev, T. et al, 2001) concluded, in waterglass-activated
slag mortars the C-S-H gel is carbonated directly. The decalcification of this gel as a
result of carbonation causes a loss of cohesion in the matrix, owing to the increase in pore
diameter and causing a drop in mechanical strength. The presence of admixtures does not
appear to affect the carbonation resistance of waterglass-activated slag mortar.
Figure 5 shows the residual strength of the mortars after the high-temperature test.
Mechanical strength was significantly higher in Portland cement than waterglass-
activated slag mortar at all test temperatures except 1000 oc, where it was lower. The
inclusion of admixtures SRA and V increases the residual mechanical strength of AAS at
all the test temperatures.
The loss of mechanical strength in the mortars is due firstly to the elimination of free
water between I 00 and 200 oc (Georgali, B. et al, 2005). The loss of chemically bonded
water begins at around 250 °C, prompting the disintegration of the C-S-H gel. The a->P
transformation in the quartz comprising the aggregate takes place at around 550 °C, in
conjunction with the expansion of this mineral that causes deterioration of the mortar. In
Portland cement mortars, in tum, the portlandite dehydroxylation that takes place
between 400 and 500 oc entails an increase in porosity and therefore a greater decline in
mechanical strength than in waterglass-activated slag mortars, which have no portlandite.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
CONCLUSIONS
The main conclusions to be drawn from the present study are as follows:
1.- The inclusion of admixtures SRA and V in waterglass-activated slag mortar enhances
its strength slightly as a result of the minor decrease in the liquid/solid ratio induced.
ACKNOWLEDGEMENTS
The authors wish to thank the Ministry of Science and Technology for the funding for
project MA T200 1-1490, without which this study would not have been possible. The
authors wish to thank A. Gil and J. L. Garcia for their collaboration in the durability test.
M. Palacios also wishes to thank the Comunidad de Madrid for the fellowship awarded.
REFERENCES
Banfill, P.F.G., 1994; "Rheological methods for assessing the flow properties of mortar
and related materials", Constr. Build. Mater. 8, 43-50.
Mollah, M.Y.A. eta!, 2000; "A review of cement-superplasticizer interactions and their
models". Adv. Cern. Res., 12 (4) 153-161
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Shi, C. 1996; "Strength, pore structure and permeability of alkali-activated slag mortars"
Cem.Concr. Res. 26, 1789-1799
Tattersall, G. H. and Banfill. P. F. G., 1983; "The Rheology of Fresh Concrete". Pitman,
London
RHaSO%
--A
--8
- ... C·1
....,.._C2
- ... D
.,.
~ w
~z
.. iii! o.
:z:
. .... • • •
0
_....., ___ .......__.__.---·-·-
100
DAYS
200 300
Ill
0,
Fig. 2.- Shrinkage percentage in AAS and cement mortars under different curing
conditions: a) RH = 99%; and b) RH =so%.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
e CYCLE1
~ 0.1
!
! CYCL£2
CYCL£3
CYCLE4
cYCLE!
2 4
0,06,--------.====;----,
0,05
---..._l~gl
£0,04
j., 0,03
·---.___ •
.... ...__
~ 0,02
0,01 ~l;;.;;~-::.:r:::"·""-""b:::----.!
10 15 20 25
Tlma (minutes)
~ ~ ~ ~ ~ ~ ~ ~ ~ m
1'emperatln r'C)
Fig. s.- Residual strength of cement and AAS mortars after high
temperature treatment.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
1 INTRODUCTION
Superplasticizers, and especially those made from PEO side chains grafted on a
PolyCarboxylate backbone (named PC-PEO afterwards), have greatly contributed to the
development of high performance concretes (HPC) and/or high fluidity concretes as Self
Compacting Concretes (SCC). These polymers adsorb onto cement grains surface. Thus,
they induce particles deflocculation, releasing the water trapped in cement floccs. Hence,
superplasticizers enable a reduction of the water required to reach a given fluidity level,
so that strength and durability are enhanced. Alternatively, superplasticizers can increase
the fluidity at constant water content, which confers a better flowability for the placing
and laying.
Superplasticizers should usually fulfill two requirements: their addition at reasonable
dosages (about 0.05 to I weight percent of dry matter with respect to cement) should
induce, in the cementitious formulations, both a good initial fluidity and a low fluidity
loss during the processing (fluidity retention for 1 or 2 hours). ln addition, it is required
that a given superplasticizer be active with the largest number of cements. Now, the
interaction between cements and PC-PEO depends on both PC-PEO molecular
architecture (backbone molar mass, side chains molar mass and grafting degree), and
cement composition. In fact, PC-PEO action is the result of a steric repulsion between the
side chains of the polymers adsorbed on the surface of cement grains. Superplasticizers
with high affinity for the surface of cement grains are bound to give good initial fluidity
(superplasticizer is rapidly adsorbed on cement grains) but important fluidity loss (little
superplasticizer remains in the aqueous phase to cover newly hydrated surfaces). On the
other hand, PC-PEO with low affinity for cement particles show a rather limited initial
fluidity (the coverage of grains by adsorbed layers of polymer is insufficient) but a rather
low fluidity loss since the aqueous phase behaves as a polymer reservoir for later
adsorption on newly hydrated surfaces.
Incompatibility issues are a well-known concern; they include poor initial fluidity for
usual superplasticizers amounts, as well as excessive fluidity loss during concrete
processing. The literature on this subject has mentioned various hypotheses to rationalize
the behaviors observed. Among these, the idea that part of the superplasticizer would be
lost by incorporation into organo-mineral phases should not be the explanation for PC-
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
2 EXPERIMENTAL STUDY
2.1 Materials
Two Axim-Italcementi Group PC-PEO superplasticizers (named SPl and SP2, main
characteristics in Table 1 ) designed to show low fluidity loss (generally used for ready
mix concrete) were tested respectively with six laboratory CEM I type Portland cements,
and six Ciments Calcia-Italcementi Group industrial CEM I type cements. The laboratory
cements were prepared so as to control cement composition (that it be clinker's chemistry
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
or set regulator nature and amount) and Blaine specific surface area. In this respect, it
was chosen to analyze, at constant Blaine specific surface area, the influence of alkali
sulphate level, while the nature of set regulator was fixed but its amount was varied.
However, these laboratory cements only present two alkali levels. The six industrial
cements were then selected so as to increase the number of different alkali levels
considered. However, the use of industrial cements allowed to control neither Blaine
specific surface area, nor set regulator's nature nor amount.
The 6 laboratory cements were obtained by grinding two Ciments Calcia-Italcementi
Group clinkers (named K1 and K2) with various amounts of a quite pure industrial
gypsum (more than 93 weight%) to reach a 4000 cm2 /g Blaine specific surface area. The
two clinkers were selected for their quite similar contents in silicate and aluminate phases
(respectively C3S+C2S and C3A+C4AF molar %), but varying alkali amount and
C3A/C 4AF molar ratio (Table 2). In these clinkers, the values of the molar ratio (Na20 +
K20)/S0 3 suggest that most of the sulphates are under the form of alkali sulphates in Kl
(the ratio is about I) [7], whereas in K2 the sulphates exceed the stoichiometric quantity
of alkalis available. According to H. F. W. Taylor [7], calcium langbeinite
(2CaS04.K2S0 4) should then form. Gypsum content was varied to obtain, for each
clinker, three cements of S03/C weight ratio respectively optimal in terms of strength, or
inferior or superior of about 0.5 weight % to this optimal value (Table 3). These ratios
were respectively named Opt, Opt- and Opt+.
The 6 CEM I type industrial cements studied with SP2 are referenced from Cl to C6.
Their main characteristics are listed in Table 4. The molar percentages of silicate and
aluminate phases determined with Rietveld analysis from XRD diagrams are calculated
say the plot of CAS versus CAs which, for a BET law, leads to a line with
Qcons·(l-CAs)
slope (K-l)I(Q 0 .K) and a y-intercept li(Q0 .K). While Q 0 determination, essentially linked
to inverse slope's value, is quite precise, K determination is much more uncertain because
small y-intercept variations lead to huge K-variations. These uncertainties are integrated
into error bars on the values obtained. These parameters are plotted versus SOiC ratio for
K I and K2 cements in Fig. 5.
We observe that Kl cements, whose aqueous solutions are more concentrated in
sulphates than those of K2 cements, show a higher adsorption (higher Q0 than that of K2
cement for equivalent K value for Opt cements; a reverse situation is observed for Opt-
and Opt+ cements).
In addition, adsorption by Kl cements increases more between 15 and 90 minutes than
the adsorption by K2 cements.
Finally, gypsum content variations through SOiC ratio influence, even though weakly,
the adsorption (Q0 is maximum and K minimum for Kl Opt cements, the K-situation is
reversed and Q 0 more or less constant for K2 cements), whereas sulphate concentration in
the aqueous solution is not affected. On the whole, it seems that the higher the sulphate
concentration in the aqueous solution, the lower the SP 1 adsorption because of a
competition with sulphates, since the S031C ratio is a second order parameter as
compared to clinker's chemistry. Nevertheless, a complete balance of the sulphates (in
aqueous phase, or integrated in mineral phases as AF 1 for instance, or adsorbed) is
required for a better correlation between the sulphate concentration in aqueous solution
The main conclusions are the same than for laboratory cements (3. I. I):
./ The composition of the aqueous solution is rapidly fixed by the cement
chemistry. Neither duration (15 or 90 min), nor superplasticizer dosage (SP/C = 0 or 0.40
%) modify (beyond experimental error) the values of theCa, K, Na or S concentrations
measured for one given cement.
./ Consequently, one of the experimental conditions, namely SP/C = 0 % and 15
min, is chosen to discuss the variations of calcium, alkalis and sulphate concentrations
(respectively noted [Ca], [Na]+[K] and [S]) versus Na20eq (Fig. 8)). Alkalis and sulphate
concentrations in the aqueous solution increase with Na2 0eq whereas calcium ions
concentration decreases. Calcium and sulphate concentrations in the aqueous solution
were plotted versus potassium concentration in the aqueous solution (Fig. 9): in
agreement with the results ofL. Nachbaur [10], we observe that [Ca] decreases while [S]
increases with [K]. Our interpretation is that the solubility ofportlandite is reduced in the
presence of increasing concentrations of potassium, and that most of the soluble sulphates
originate from alkali sulphates from the clinker phases.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
of adsorption sites for SP2.
4 CONCLUSION
The results on laboratory and industrial CEM I type cements illustrated the influence of
cement composition and especially clinker chemistry. As a matter of fact, the quantity of
sulphates present in the aqueous phase is governed by the alkali level: the rapidly soluble
sulphates are the alkali sulphates of the clinker. In addition, the higher the sulphate
concentration in the aqueous solution, the lower is the superplasticizer's adsorption. For
laboratory cements, the distinction, between the respective influence of the number of
adsorption sites and their affinity, in the variations of superplasticizer's adsorption is not
yet explained. Nevertheless, the yield stress level is governed by the superplasticizer
adsorption (for a given time, we obtain a common curve for the 6 laboratory cements
considered). For industrial cements, the number of adsorption sites of the superplasticizer
decreases while the sulphate concentration in the aqueous solution increases. This
reinforces the idea that this species is in competition with superplasticizer from an
adsorption point of view. Correspondingly, for a given superplasticizer dosage, the yield
stress of the grout is all the higher (and so the fluidity all the lower) as the number of
adsorption sites for superplasticizer is lower. Finally, we evidenced that set regulator
composition and its amount were, in the cements considered, at most second order
parameters. Indeed, concentrations in the aqueous solution (especially sulphate
concentration), adsorption level and thus yield stress value of the grouts were mostly
governed by clinker's chemistry.
5 REFERENCES
[1} Fernon, V., Vichot, A. et al., Interaction between Portland cement hydrates and
polynaphtalene sulfonates, 51h CANMET, SP173, 225-248 (1997).
[2} Flatt, R., Houst, Y, A simplified view on chemical effects perturbing the action oj
superplasticizers, Cement and Concrete Research, 31, 1169-1176 (2001).
[4] Yamada, K., Hanehara, S., Interaction mechanism of cement and superplasticizers -
The roles of polymer adsorption and ionic conditions on aqueous phase, Concrete
Science and Engineering, 3, 135-145 (2001).
[5] Yamada, K., Ogawa, S., Hanehara, S., Controlling of the adsorption and dispersing
force of polycarboxylate-type superplasticizer by sulfate ion concentration in aqueous
phase, Cement and Concrete Research, 31, 375-383 (2001).
[6] Jiang, S., Kim, B., Artcin, P. C., Importance of adequate soluble alkali content to
ensure cement I superplasticizer compatibility, Cement and Concrete Research, 29, 71-78
(1999).
[8] Prince, W., Espagne, M, Ai"fcin, P.C., Ettringiteformation: A crucial step in cement
- superplasticizer compatibility, Cement and Concrete Research, 33, 635-641 (2003).
[9] Tagnit-Hamou, A., Artcin, P.-C., Cement and superplasticizer compatibility, World
Cement, 38-42 (2003).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
silicate tricalcique, composant principal du ciment Portland. Caracterisation des
interactions al'origine de Ia prise, PhD thesis, Universite de Bourgogne (1 997).
[11] Plummer, L. N., Parkhurst, D. L., Flemming, G. W., Dunkle, S. A., A computer
program incorporating Pitzer's equation for calculation of geochemical reactions in
brines, U.S. Geological Survey, Water Ressources Investigations Report 88-4153 (1988).
[12] Sun, L.-M, Meunier, F., Techniques de l'Ingenieur, Adsorption, Aspects theoriques,
12-730, 1-16 (2003).
Technique Quantity Kl K2
C3S+C2S 85.4 88.9
XRD, Rietveld analysis C3S/C2S 2.2 2.5
(n/n %) C3A+C.0,F 11.2 7.1
C3A/C.0,F 0.7 1.3
Na20-= 0.24 0.61
XRF (w/w%)
sol 0.32 1.35
XRF(n/n) Oia20+K20_2/ SOl 0.97 0.59
Table 3: values of SOJC weight ratio for the six laboratory cements measured with a
sulphur analyzer.
SO)IC Opt- Opt Opt+
Clinker
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Kl 2.3 2.9 3.3
K2 2.8 3.6 4.0
~~;o:===~=~==;~;::••=f
100
K1 K2
--- -B- SP/C = 0%, 15 min
----- - e- SP/C = 0 %, 90 min
-4--- -A- SP/C = 0.40 %, 15 min
--+-- - +- SP/C = 0.40 %, 90 min
25
U U U U ~ U M U U U
SOsfC (wiw%)
.--..
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~E
E
....... 100
·~
~
cl.
0
II
if
</)
g
10
0.25 0.30 0.35 0.~0 0.45 0.50 0.55
Na 20eqc (wlw %)
Figure 2: Concentrations in the aqueous solutions of plain grouts (SP/C = o%) after
15 min versus the weight ratio Na 2 0eq of the clinker.
l
~
~
1.5
1.0
f----W-+1---i
K1 K2
~
• 0 SP/C = 0%,15 min
0.5 •• 0 SP/C = 0 %, 90 min
+
'iii' .. I>
<)
SP/C = 0.40 %, 15 min
SPIC = 0.40 %, 90 min
~ 0.0 I I I
2.4 2.8 3.2 3.6 4.0
S01C (w/w%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
2.0
cCD I
E
CD 1.5
0
0
~
c.. 1.0
(/)
"'
.5 0.5
§
0
0.0
0.000 0.005 0.010 0.015 0.020 0.025
CAS (gSP/gAS)
3.0 600
2.5 500
~
2.0 400
.5 1.5 I 300
0
0 1.0 200
0.5 100
0.0 0~~--~~~--.-~
2.25 2.75 3.25 3.75 4.25 2.25 2.75 3.25 3.75 4.25
SOjC(wiw%} SOiC (wiw%)
0 15 30 45 60 75 90
t (min)
Figure 6: Evolution of yield stress with time for K1 Opt grouts and various
SP /C ratios of SP1.
15min 90 min
10
K1
•
K2
A .
K1
A
K2
Opt-
•• 0
0
Opt
Opt+
5 & 7 8 9
0.1
Qcon;mm (mg SP/g of cement)
Figure 7: Yield stress versus superplasticizer's consumption after 15 or 90 min for SP1
containing grouts.
-::1 50 +
~
400
'5 40
E
Q
!. 300
ro 30 $
2. 200
20 33
0
100
10
E
0 0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Na20eq (wtw %)
Figure 8: Concentrations in calcium, alkalis and sulphates versus Na 2 0,q for plain
grouts of industrial CEM I type cements (SP /C = o%) after 15 min.
eo
60
f • ~I
[Cal
c
0
A
<>
SP/C • 0'%, 15 min
=
SP/C O"k, 90 min
=
SP/C 0.40%, 15 min
SP/C = 0.40%, 90 min
it::
, "'
200
150
~
0
E
_#- ]l
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
..§. 40 100 33
-;;; g,
2. .!:::
20 50
·--
01-----.-----.----.----------~
100 200 300 400 500
[K) (mmol/1.)
-l.
~
~
:::::..
2.0 -
Figure 10: ([Na]+[K])/(2.[5]) molar ratio measured by lCP-AES in the aqueous solutions
versus the weight ratio Na 2 0eq of the cements.
1.2 1.2
.,
"% "%
E
8
1.0 "'E
8
1.0
0 0.8 0 0.8
::?
"-
rn
0.6 -"'
a.
UJ
0>
0.6
.5."'., 0.4
.5. 0.4
c
8
0.2 "'c
8
0.2
X-~~-Z
0 0.0 0 0.0
0.000 0.005 0.010 0.015 0.020 0.000 0.005 0.010 0.015 0.020
CAs(9 SP I gAS) CAS(g SP/g AS)
Figure 11: Consumption isotherms ofSP2 by industrial CEM l type cements after
15 and 90 min.
1.2 4000
"%
"E 1.0
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
3000
8 0.8
0
~
a.
0.6 1 2000
~
Ch
0.4
! 0.2
1000
o· 0.0 0
0 40 80 120 160 200 40 ao 120 160 200
[S]AS (mmoVL) [SJAs (mmoi/L)
Figure 12: Parameters of the BET fit of the SP2 consumption isotherms(~ on the left,
K on the right) versus sulphate concentration measured in the aqueous solution after
the same time.
Figure 13: Yield stress versus superplasticizer (SP2) content of the grouts, SP/C.
Figure 14: Relative yield stress versus adimensional SP2 consumption, after 15 min.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
or CSA as expansive agents combined with a Shrinkage-Reducing Admixture (SRA)
in the presence and in the absence of wet curing. Due to the synergistic action of
the SRA with the expansive agent, the restrained expansion can occur even in the
absence of any wet curing. However, this action in much more effective when SRA is
combined with CaO rather than with CSA.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Engineering Concrete, Ponzano Veneto, Italy. He is author of numerous papers in the
field of concrete technology and in particular of chemical and mineral admixtures.
INTRODUCTION
Recently it has been found that when a CaO-based expansive agent is combined with
a Shrinkage-Reducing Admixture (SRA), the expansion can occur even in the absence of
a wet curing [2].This effect would improve the robustness of a shrinkage-compensating
Two commercial expansive agents were studied : one based on dead burnt CaO and
the other on CSA. The expansion produced by CaO or CSA are related to the following
chemical processes:
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In general the recommended time of wet curing is 1-2 days for cementitious material
containing CaO and 5-7 days for cementitious materials containing CSA. This time is
related to the different hydration rates of the chemical processes.
The other ingredients used in the study are:
to open air. Curing procedure iii) is intended to simulate the behavior of reinforced
structures exposed immediately to open air as sometimes occur with sprayed mortars
without any protection and wet curing.
The restrained length change of the reinforced specimens was measured as a function
oftime up to 2 months. The results are shown in Fig.6-10.
Figure 6 shows the behavior of the control mixture and that of the "SRA mortar"
following curing procedure ii): there was no length change in both mortars during the
first two days when the specimens were wrapped with plastic sheet; then, when the
specimens were exposed to the air at R.H. of 60%, the "SRA mortar" shrank at a lower
rate with respect to the control mortar: about 50% less at early ages and about 30% less at
later ages. So, the effect of the chemical admixture SRA used in this work is similar to
that recorded by others in the technical literature [5].
Figure 7 shows the behavior of the "CaD" mortar when tested according to the three
curing procedures: the restrained expansion reached in about two days the level of 800
microstrains under permanent wet curing with little change at later ages; when the
specimens are protected from drying by the plastic sheet the initial expansion of 600
microstrains is gradually relaxed by the subsequent shrinkage and it is completely lost in
one month; in the absence of the early protection, the initial expansion in the reinforced
specimens is only 300 microstrains and it is lost in about two weeks.
The combination of SRA with the CaO-based expansive agent (Fig. 8) gives a
synergistic effect with respect to the mortar with only SRA (Fig. 6) or only CaO (Fig. 7).
First, the expansion on the "CaOISRA" mortar according to curing procedure ii) is the
same as that of the same mortar permanently kept under water: the restrained expansion
is 800 microstrains at about 2 days regardless of the curing. Second, the shrinkage of the
"CaO/SRA" mortar after removing the plastic sheet is much less with respect to that of
the corresponding mortar with CaO without SRA (Fig. 7); therefore, the initial expansion
after the demolding is relaxed by the subsequent shrinkage to a lower extent so that there
is still a restrained expansion even after the exposure of the specimen to air with R.H. of
60% for two months. Third, even when the restrained specimens are exposed
immediately to the air with R.H. of 60%, the reduction of the expansion due to the
subsequent shrinkage is lost only at two months; this means that cracking in the
"CaO/SRA" mortar would be unlikely: only when the tensile stress, after two months, is
higher than the tensile strength of the mortar.
Figure 9 illustrates the restrained length change of the "CSA" mortar according to the
- the ~ effect which resultsg in a lower shrinkage due to the expected influence of the
SRA on the shrinkage behavior;
- the a effect due to the un expected higher expansion of CaO related to the presence
ofSRA.
X-ray diffraction analysis (Fig. 12) did not show any influence of SRA on the CaO
hydration in terms of the Ca(OH) 2 peaks. The termogravimetric analysis (Table 3), again,
does not show any influence of SRA on the amount of Ca(OHh .Therefore, presently the
increase in the restrained expansion of CaO with SRA cannot be explained. Also
unexplainable is the influence of SRA on the restrained expansion is much more effective
with the CaO expansive agent (Fig. 8) than with the CSA agent (Fig. 10).
CONCLUSIONS
The combined action of SRA with an expansive agent such as dead burnt CaO or
calcium sulfa-aluminate (CSA) increases the restrained expansion in the absence of wet
curing. The effect is much more enhanced when SRA is combined with CaO than with
CSA.
In particular, the combined action of SRA with CaO will permit the manufacture of
shrinkage-free mortars as repair cementitious materials for damaged concrete structures,
when they are placed into formworks and exposed to air after two days: even after two
REFERENCES
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Malhotra, Barcelona (Spain), pp. 1209-1220,2000.
[2] M. Collepardi, A. Borsoi, S. Collepardi, J.J. Ogoumah Olagot, R. Troli, "Effects
of Shrinkage-Reducing Admixture in Shrinkage Compensating Concrete Under Non-Wet
Curing Conditions", Cement and Concrete Composities, 6, pp. 704-708, 2005.
[3] C. Maltese, C. Pistolesi, A. Lolli, A. Bravo, T. Cerulli and D. Salvioni,
"Combined Effect of Expansive AND Shrinkage Reducing Admixtures to Obtain Stable
and Durable Mortars", 12, pp 22444-2251,2005
[4] D. Bischoff and A. Toepel, "Laboratory Testing of Portland Cement Concrete
Patch Material, Modified to Reduce or Eliminate Shrinkage", Wisconsin Department of
Transportation Final Report, pp. 1-81.
[5] N.S. Berke, L. Li, M.C. Hicks, J. Bal, "Improving Concrete Performance with
Shrinkage-Reducing Admixtures", Proceedings of the Seventh CANMET/ACI
International Conference on Superplasticizers and Other Chemical Admixtures in
Concrete, Editor V.M. Malhotra, Berlin, pp. 37-50, 2003.
DE BONDING
Fig. 1- Cracks in the repair material and debonding from the concrete substrate
adapted from[t].
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
.
,. "cracking
,;
...
.. ... .. 'Ot
,;
,,
,;
Time
I"'LENGHT
INCREASE
FROM
EXPANSIVE
AGENT
(1-7 days)
~
Fig. 3 -Concrete substrate and reinforcements restrain the length increase of
the expanding repair mortar. The repaired portion of the structure accumulates
compressive stress (o), adapted from [1).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
, "'
............................................
'', ""
...... - ,"
"Gc
wetcurfna
expansion air d
Time
Fig. 4- Compressive stress (8J due to the restrained expansion compensates the
subsequent relaxation and tensile stress induced by drying shrinkage. The resulting
net stress is never higher than the tensile strength (f,).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
<Yt /
/
·············-~·-··························
Time
Fig. 5- Due to inadequate wet curing, the compressive stress (8<) is very low and
then it cannot compensate the subsequent shrinkage: tensile stress 81 can reach the
tensile strength f, and cracking can occur.
Fig. 6- Restrained length change of the control mortar and SRA-based mortar both
protected by plastic sheet for 2 days and then exposed to air (R.H. = 6o%).
1000~-----------------------------,
l ::?t; -
... .oKlO p!l 1!1,
wet curing - -
fl. IJ '- protected 2 days
~ zoo p '--.. lEI .... b.! plastic sheet
j
o ------------o,.:.:.-:--!'::.fk-;.;-.:.·::·----------------------
~ -zoo
E
· c-. . ~-
"It - - - - - -21
" -roo a1r cunng · -o- . -[
J-e ·600
R.H. = 60%
6j -BOO
-1000 ..___. . . ____.___ ___..___......_ . . . .____,
......~
0 10 20 30 40 50 60 70
Time (days)
Fig. 7 -Influence of the curing on restrained length change of the CaO based mortar.
·!
g.
kl -400 D..,
t!l. , ~ protected 2 days
fl.~ !I~ by plastic sheet
.S zoo . ""'0. ~ - ~ - - -!1 - - 11- -!1-
i 0 •.....•••.•••.•.••••. :.r::!:-::: ..:..':7.:-::-•• .C::-:=::.IJ..:-e-.
·~ -200 air curing
1!1. -400 R.H. =60%
~ ·600
~ ·800
-1ooo..__ _.__ _.__ _.__ ___._ _,___ _,____.J
0 10 20 30 40 50 60 70
Time (days)
1000r-------------------------,
c 800
l
wet curin!l
.E 200
i... 0 _g_ .:,s.. ......... ---. ---·-.. -----·----· .. -------·.----·. ---·
'-' '!1.. protected 2 days
~ -200 ·[]-....::_... ... by plastic sheet
i -400 • ~ ~ - - - ~- - - - - - 1!:1
~c -600 LJ-. . - - IJ - - · --C
air curing
~ _
800
R.H. = 60%
-1000.___ _,___ _.__-J.._.....,I._ _ .___....___._~
0 10 20 30 40 50 60 70
Time (days)
Fig. 9- Influence of the curing on restrained length change ofthe CSA based mortar.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
1:
wet curing
61 ·800
·10001......-_.__........_--'--.1.....-....&...-......1.---'
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
0 10 20 30 40 50 60 70
Time (days)
Fig. 10- Influence of the curing on restrained length change of the CSA/SRA
based mortar.
1000
,.cc0 100
600
:1
~ 400
cao/SRA
.5 200
I!
'il 0 -----------------------
~-200
cao
j -400
~ -600
~ .-oo
·1000'----'--......L.---i--.1.....-....&...-......1.-__...l
0 10 20 30 40 50 60 70
Time (days)
Fig. 11- Restrained length change in "CaO mortar" and "CaO/SRA mortar" exposed
to air (R.H. = 6o%) after 2 days of protection by plastic sheet.
1000
~800
S600
u
400
200
0.0
13.0 14.0 15.0 16.0 17.0 18.0 19.0
29
1200
1000
~800
8600
400
200
0.0
13.0 14.0 15.0 16.0 17.0 18.0 19.0
29
Fig. 12- X·ray diffraction analysis of the "CaO Mortar" (A) and "CaO/SRA mortar" (B).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Sebastien GEORGES and Jean Philippe PEREZ - Jean philippe Perez and Sebastien
Georges work for Lafarge (LCR) as research engineers on organa-mineral interactions.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
INTRODUCTION
Polycarboxylate -type superplasticizer has become popular because of its superior water-
reducing ability and fluidity retention ability. The interaction mechanism between
polycarboxylate polymers (PCP) and cement is not yet fully understood and the
dispersibility of PCP on cement particules can be affected by various conditions (1).
Indeed, it is assumed that PCP causes dispersion of cement particles by the steric
repulsive forces caused by the PCP adsorbed to the surface of solid particles (2).
It has been reported that the cement characters affecting the fluidizing ability of PCP
were the amount of alkaline sulphate, the kind and amount of calcium sulphate, and the
amount and activity of interstitial phases (3).
Some studies pointed out that the surface area on which PCP was adsorbed and the
adsorption amount were the most important factor based on the assumption that PCP
works after the adsorption on cement hydrates (4, 5). As for the PCP adsorption, the
sulphate ion concentration in solution phase was thought to affect it. The adsorption of
PCP decreases in the case of high sulphate ion concentration in aqueous phase because
the sulphate ion and PCP are adsorbed on cement particules competitively (6).
In actual placing of concrete, in which PCP was used, it is experienced in some cases that
the fluidity of concrete tends to gain few minutes after its mixing. We called "over-
fluidification" this phenomenon. This increase in fluidity makes it difficult to pour the
Our study was thus designed to better understand this phenomenon of "over-
fluidification". Our study first consisted in highlighting the phenomenon and quantifying
it on mortar and cement paste thanks to a rheometry tool.
A parametric study of the phenomenon was then undertaken. We studied the parameters
related to the manufacturing process, the parameters related to the PCP structure, and the
chemical composition of cement.
In order to understand the origin of the phenomenon, the evolution of the rheological
properties of cement paste is drawn in parallel with measurements of PCP adsorption on
cement particles.
Our study should allow the understanding of the origin of the phenomenon of super-
fluidification and the parameters which control the action of the PCP at the beginning of
the cement hydration in order to overcome the incompatibility problems.
EXPERIMENTAL
In the following sections, the characteristics of the material (polymers and cement) and
the details of the method used in the experiments of the present study are described.
1-Materials:
Polycarboxylate superplasticizers:
Industrial reference:
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The industrial HRWRA selected as reference is a standard PCP (called PCP,er) used for
ready mix applications, characterized by a charge amount ccoo· concentration/g of
polymer) of 0.94 m.eq/g. Some slump increases with time were reported when using this
HRWRA.
Model PCP:
In order to analyse the influence of the PCP structure, we worked with several PCP
model with different charge amount.
The PCP were synthesised by copolymerizing methacrylic acid (CA) and
methoxy(polyoxyethylene) methacrylate of Mw = 1100 g/mol (CE) at different molar
ratio.
We obtained a range of five PCPs with charge amount between 0.53 m.eq/g and 2.6
m.eq/g. The average molecular weights (Mw) of these polymers were about 45 000 g/mol
with a polydispersity index about 2,5. (Table 1)
All the polymers were neutralized with NaOH to a pH value of 7.
Cement:
Lafarge Portland cement named Cement H was used to study the influence of process
parameters and of PCP structure. It is characterized by a low tricalcium aluminate phase
and low alkaline sulphate quantity. (Composition in Table 2, Calcium sulphate : 2/3
gypsum, 1/3 hemi-hydrate)
Clinker:
In order to control the type of calcium sulphate added and the proportion of alkaline
sulphate, we used an industrial clinker containing low alkaline sulphate, namely the
clinker L. (Composition in Table 2)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In order to investigate the influence of the type of calcium sulphate introduced, we added
a certain amount of calcium sulphate to the clinker in order to have 3.7% of total S03 •
Different kinds of calcium sulphate were used: gypsum only or a mixture of gypsum and
hemi-hydrate at different ratios. The calcium sulphate and the clinker were mixed in an
asymmetrical mixer during three days.
To investigate the influence of the proportion of alkali sulphate, we introduced Na2S0 4 or
K2S0 4 to the solution before mixing with the clinker powder.
11-Mixture proportion :
Most of the experiments were carried out on cement paste but we also worked with a
mortar in order to evaluate the results on the two different systems.
All manipulations were performed with a constant PCP dosage of 0.16% and at a
temperature of 20°C except the case of the temperature influence. The water content
varied with the nature of the mineral composition and admixture in order to retain the
same initial viscosity of about 2 Pa.s at 50 s- 1 for the cement paste and a spread of slump
flow of 280 mm for the mortar except in the case of the study of the process parameters
where the W/C remains constant. This adjustment guarantees to have the same initial
state for all the different conditions thus assuring reliable comparisons.
HI-Experimental procedures:
The fluidity of mortar was approximately evaluated by flow value. Cement and sand were
mixed for 30s in a mortar mixer. We added water mixed with superplasticizer and we
mixed everything for 1 min 30s at a constant rotational speed. The fluidity of the paste
was evaluated by the spread measured when a truncated cone (Abrams (7) of 0 100
mm and 0 50 mm and 150 mm of height) filled with the mortar was pulled away
vertically. In order to examine the change in the flow value over time, we measured the
flow value every 0, 15, 30, 45, and 60 minutes after initial mixing with a mixing of 30s
before each measurement. The slump flow on mortar highlighted the phenomenon of
"over-fluidification".
In order to work with a simple system and to more precisely examine the phenomenon of
"over-fluidification", we developed a rheometry protocol.
This protocol permits to measure the apparent viscosity of the cement paste at 0, 15 and
30 minutes after mixing in the same conditions as the slump flow test.
As a cement paste is close to a Bingham fluid (8), its apparent viscosity T]app depends on
the shear rate applied y and on the yield stress cy, considering the law:
We used an ARES rheometer with a rheo reactor geometry illustrated in Fig. 1. In order
to ensure good reproducibility (8), the cement paste was mixed in the rheo reactor.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Thanks to a vibrating spatula, the cement was progressively mixed with the solution
(water+ PCP) previously introduced in the tank. When all the powder is introduced the
cement paste was mixed for one minute at a shear rate of 80 s· 1• After that, the plastic
viscosity is measured at a shear rate ranging from 50 to 0.2 s· 1• To do so, the tank turns at
different rotation rates and the sensor measures the torque that the cement paste creates
on the agitator. After a rest of 15 minutes, the paste was again mixed for 30 sat a rate of
50 s· 1 and we again measure the viscosity. The same operation was repeated 30 minutes
later after initial mixing.
In this way we measured the variation in apparent viscosity versus shear rate at different
times after initial mixing.
The results show that this % of over-fluidification corresponds well to slump flow
measurements for rates between 10 s· 1 and 50s· 1. Indeed, the scale of shear rates applied
on a fluid during gravitating flow is ls- 1 to 10s" 1 (9). During the slump flow on mortar,
the cement paste undergoes both the shear of the flow and the shear of the aggregates so
the scale of shear rates which correspond with rheometry protocol seems coherent.
Thus in order to make a comparison, we calculated the average of the % of over-
fluidification for rate between 10 and 50 s· 1:
withj being the number of measurements made with shear rate between 10 and 50s· 1
and i being the shear rate between 10 and 50 s· 1
For each condition, the experiments were repeated 5 times and we calculated an average
of the results.
This protocol therefore results in a quantification of the phenomenon and in comparison
of this phenomenon in different operational conditions.
Measurement of adsorption
For each measurement, we worked with a cement paste at W/C = 0_5 mixed in a small
mortar mixer at a temperature of 20°C. After mixing, we took a sample of the paste. For
each sample, supernatant liquid was separated centrifugally at 3000 rpm from the paste.
Then the total organic carbon content of this liquid was measured using Shimadzu TOC-
5000 A, and the concentration of the superplasticizer remaining in the supernatant liquid
was obtained from a calibration curve prepared in advance. This yielded the amount of
superplasticizer adsorbed on the cement as the difference between the amount of
superplasticizer added and that remaining into the liquid.
For each condition, the measurements were repeated 3 times and we calculated an
average of the results.
So we obtained the evolution of the PCP adsorption during the thirty minutes after initial
mixing of cement paste. In order to interpret results, two parameters are calculated, the %
of PCP initial adsorption and the relative increase of the adsorption between t = 0 min
and t = 30 min.
Figure 2 shows the results of the slump flow test with the use of PCP ref which leads to a
"over-fluidification". When PCP ref is used we can observe an increase of the spread of the
slump flow during the 30 minutes after initial mixing which is representative of the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Figure 3 shows the measurement of viscosity of cement paste according to shear rates
over time using PCP,er when a phenomenon of "over-fluidification" is noted. For each
shear rate, we note that viscosity decreases between 0 and 30 min. The % of super-
fluidification calculated is therefore positive contrary to cement paste without PCP which
gives a constant viscosity over time and therefore no over-fluidification. (Fig. 4)
11-Parametric study
Rheology results:
Process parameters :
The influences of temperature and mixing were studied. We can note in Fig. 5 that the
higher the temperature, the lower the % of super-fluidification. This result is in
accordance with the observation that during winter more problems of super-fluidification
arise with use ofPCP(ll).
To measure the influence of mixing, two parameters were studied, the rate of the mixing
and the duration of the mixing. We can see in Fig. 6 and 7 that the longer the the mixing
duration and the stronger the mixing rate, the lower the over-fluidification. Since duration
and shear rate of mixing are proportional to the mixing energy, we can say that the higher
the mixing energy expended on the cement paste, the lower the over- fluidification.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In Fig. 10, we can observe that with a constant dosage in PCP, the higher the quantity of
Na2 S0 4 in solution, the lower the over-fluidification. We can also observe that the water
needed to obtain the same initial fluidity increases with the Na2 S04 concentration. This is
in accordance with the fact that there is a competition in adsorption between PCP and
sulphates ions (6). We observe the same effect with the addition ofK 2S04.
Adsorption results:
Process parameters:
Figure 11 shows that the initial adsorption of the PCP varies with the temperature: The
higher the temperature, the higher the initial adsorption. We also observe that adsorption
increases with time and that this increase depends on temperature.
As regards mixing, the observation is similar (Fig. 12): The longer the mixing duration,
the higher the initial adsorption of PCP. We also observe that the evolution of PCP
adsorption over time depends on the mixing duration.
Structural parameters:
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In Fig. 13, we can observe the adsorption results for the PCP model with different charge
amount. We observe that the higher the charge amount, the higher the initial adsorption.
This finding is in accordance with the fact that the higher the charge amount, the more
carboxylic groups contained in the PCP and the higher the adsorption on the cement
particles (13). Moreover, for each charge amount, adsorption with time increases and
seems dependent on the PCP charge content.
In order to draw a parallel, we reported the % of initial adsorption and the % of the
relative increase in adsorption during the 30 minutes after the initial mixing as a function
First, we observe a link between the initial adsorption, the % of the relative increase in
adsorption and the % of over-fluidification. Indeed, for each parameter, if the % of over-
fluidification increases, the % of initial adsorption decreases and the % of relative
increase in adsorption increases. There is a single exception at a point which corresponds
to the PCP which has a charge amount of 2.6 m.eq/g (Fig. 17). With this PCP, the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
phenomenon of over-fluidification is very slight, its initial adsorption is high, but its
increase in adsorption is also high. We think that the evolution of adsorption for this PCP
has no effect on the fluidity of the paste because its initial adsorption is sufficient to
assure the maximal dispersion.
We can therefore conclude that there is a link between the phenomenon of over-
fluidification, and the evolution of adsorption of the PCP. The lower the initial adsorption
and the higher its relative increase with time, the more marked the phenomenon of over-
fluidification.
Ifwe compare the ratio A= ~% of over- fluidification and
~% of initial adsorption
the ratio B =
ll% of over- fluidification , we observe that A> B
ll% of relative increase of adsorption
for each parameter. Then, the initial adsorption seems to be the parameter the most
critical for the phenomenon of over-fluidification.
With regards to the alkaline sulphate, we have observed that the decrease in the
phenomenon of over-fluidification related to the alkaline sulphate concentration is linked
to a strong decrease in the initial adsorption of the PCP. This is caused by the competition
of adsorption between sulphate and PCP ( 5). So, assuming a constant dosage of PCP, the
increase in the amount of alkaline sulphate creates a major decrease in the adsorption of
PCP. In these conditions, the PCP becomes inoperative and loses all dispersion effect.
Consequently, the phenomenon of over-fluidification disappears because this
phenomenon seems to be the consequence of the PCP adsorption.
CONCLUSION
This study first highlights the phenomenon of over-fluidification by slump flow tests on
mortar. The rheometry protocol used here precisely quantifies this phenomenon on
cement paste. This protocol provides information in agreement with slump flow tests on
mortar.
Thanks to this rheometry protocol, we conducted a parametric study of the phenomenon
of over-fluidification. We have observed that the temperature, the mixing, the charge
amount of the PCP and the kind of calcium sulphate added have a significant effects on
the phenomenon of over-fluidification. A study of PCP adsorption in the same conditions
showed that when the phenomenon takes place, the adsorption of the PCP increases with
REFERENCES
2. Uchikawa, H., S. Hanehara, and D. Sawaki, The role ofsteric repulsive force in
the dispersion of cement paricles in fresh paste prepared with organic
admixture. Cement and Concrete Research, 1997. 27(1): p. 37-50.
3. Matsuhisa, M., eta!., Effects of cement character on the fluidity ofcement paste
added with beta-naphtalene sulfonate type or polycarboxylate type
superplasticizers. Proceedings of the Japan Concrete Institute, 1998. 20(2): p.
67-72.
4. Negishi K., N.H., Sugaya H., Goto T., Fluidity ofpaste added with
polycarboxylate-type superplasticizers. Cement science and Concrete
Technology, 1998.52: p. 138-139.
7. Saak, A.W., H.M. Jennings, and S.P. Shah, A generalized approach for the
determination ofyield stress by slump and slump flow. Cement and Concrete
Research, 2004. 34(3): p. 363-371.
8. William D.A., S.A.W., Jennings H.M., The influence of mixing on the rheology
offresh cement paste. Cement and Concrete Research, 1999. 29: p. 1491-1496.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
12. Comparet, C., Etude des interactions entre les phasesmodeles representatives
d'un Ciment Portland et des superplastijiants du beton. 2004, universite de
Bourgogne.
13. Yamada, K., et al., Effects of the chemical structure on the properties of
polycarboxylate-type superplasticizer. Cement and Concrete Research, 2000.
30: p. 197-207.
14. Nakajima Y., Y.K., The effect of the kind of calcium sulphate in cements on the
dispersing ability ofpoly -naphatalene sulphonate condensate superplasticizer.
Cement and Concrete Research, 2004. 34: p. 839-844.
Chorge omount
PCP
(m.eq/g)
Mw(glmol) lp
1 2.6 I I
2 1.43 45 777 2.59
3 1.23 47462 2.6
•
SOip~~:jl~ c c ,:he: ............
M
Cement paste
c:
Angular displacement
Over -lluidification
340
320
300
I 280
a!
I! 260
".
240
220
200
0 10 20 30 40
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
time (min)
Fig. 2: Evolution ofthe spread of Mortar over time (slump flow tests),
PCP ref= 0.16%, W/C = 0.4, Cement H.
0.1 10 100
shear rate (s-1)
Fig. 3: Evolution of Apparent Viscosity versus Shear rates over time after initial mixing
for a Cement paste with PCPref = 0.16%, W /C = 0.32, Cement H:
A decrease of viscosity is observed with time after mixing.
' ----
1:
I - ---- -without additiw
40
~
c: 30
'6
·s 20
- % super-tluidificationm representative of the
~CD slump flow test on mortar
10 -
~
;fl.
0
~.... 10 20 30
-
_40_____ @~
-10 --e., ___ ..
---
-20
shear rate (s ., )
E
60
c::
0 50
~
c: 40
:2:I 30
~Gl 20
~ 10
?fl.
0
10 20 30
Temperature (°C)
80s-1 100s-1
shear rate applied for 3 min rnxlng
35,--------------------------,
W/C=0.32
I
E 30 +----------~-----!
25
, 20+----
=
~ 15+-----
5
'$.
10+----
W/C=0.3
60,-----------------------------.
e 50+-------------
C
~ 40+------~-----------
~ 30+-~~WL-------------
'6
·s 20
'Z
5 10
"$. 0
- 10 ..L---'c"Cal"-:ci""um"----""'-.>i
sulphate
kind of calcium sulfate added
Fig. 9: Influence of the kind of calcium sulphate added to Clinker Lon the% of over-
fluidificationm (see Equation 2): Cement paste with PCP ref= 0.16%, W/C adjusted to
·get the same initial viscosity, initial mixing: 1 min, So s-1, T= 20°C.
90 9
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
c
.2 80 8 c
e. i..
,..,.
0 70 7
0
60 6
0
. 50 5
i
~
.13
!
.5
40
.E
4 II.
30 3 ~
1 0
..,.
i0 20
10
2
1
..,.
0
10 20 30
Temperature ("C)
Fig. 11: Parameters representative of PCP adsorption evolution over time according to
temperature: Cement H paste, PCP ref= 0.16%, W/C = o.s, T = 20°C,
initial mixing: 1 min.
Fig. 12: Parameters relative to PCP adsorption evolution over time according to
duration of mixing: Cement paste, PCP ref= 0.16%, W/C = o.s. T = 20°C,
Initial mixing: 1 min.
50
40
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
:.< 30
20
10+-----
null
0·1-----
0.53 1.06 1.23 1.43 2.6
Charge amount (m.eq/g)
Fig. 13: Parameters representative of PCP adsorption evolution over time according to
the PCP charge amount: Cement paste, Additive= 0.16%, W/C = o.s,
Initial mixing: 1 min, T = 20°C.
Fig. 14: Parameters representative of PCP adsorption evolution over time according to
the kind of the calcium sulphate added to Clinker L, Cement paste, PCP ref= 0.16%,
W/C = o.s, T = 20°(, Initial mixing: 1 min, 8os·'.
8
~
80 0 "
"'0
16
1i 5
7 '
',~
....... '
'
----- 70
60
50
""e-
.,...0
0
..
Ill
'1'1 '' \ !
.5 4 40 u
D.
!i
0 2
3 ------ --. "' .... '
''
- - 30
20
.:
Ill
.i!:
1!I!!
~ ''
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
' ... 10 0
~
0 0
0 10 20 30 40 50 60
% over-fluidificatior~m
c 5 -- -- -- / 90 "&.
~
-/
0
~4..
'
--- 80 .....
70 '0
i '
--~
60 ..
~ 3
'
''
'' :: J
-
~
2
''
' ' ... 30 if
'51
~
l ~ -% of PCP initial adsorption
- % of relattw increase of adsorption
~
20
10 '0
~
0 0
15 20 25 30 35 40 45
% over-fluidificationm
40
.. ,, 60
.~
35
30
-- --,_
'
~._
I
1 ----..,II
-, ,,
25
~ 20
CA_
.i
~ 15
~ }'\
~ \
''
~ 10 \
% over-fluidificatio!lm
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~-
. ,
,--
' ,,
..................... /
I ~-~-%ofPCPinitialedsorption
~' - - - % ~frelativeincreaseo~d~orption
' -
p '
'
...<'
\
' - - - ------ ..
-10 10 20 30 40 50 60
% over-fluidification..
Fig. 18: Parallel between %of over-fluidificationm, the% of PCP initial adsorption
and the relative increase in this adsorption over time with the results concerning the
effect of the kind of the calcium sulphate added to clinker L.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Rheological Behavior of
Fresh Cement Pastes: Interaction
between Superplasticizer and
Viscosity-Enhancing Admixture
Synopsis: Self-compacting concrete (SCC) has high flowability and can be placed
without vibration. It is defined as concrete that exhibits high deformability and
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
good resistance to segregation. This kind of concrete is of great interest and has
gained wide use especially in the case of structures that presents difficult casting
conditions, such as heavily reinforced sections. From a rheological point of view, the
use of a viscosity-enhancing admixture (VEA) along with an adequate superplasticizer
(SP) content can ensure high deformability and stability. However, little is known
about the interactions between SP and VEA. Hence, a study on several cement pastes
formulated from the original paste of a typical SCC is proposed. One of the major
aims of this paper is to show that empirical tests, such as spread and flow time, are
suitable to characterize the rheological behavior of cement paste instead of more
complex ones. Rheological properties, i.e. viscosity and shear yield stress can be well
correlated with empirical test results in the range offlowable mixtures. Moreover, the
experimental program leads to emphasize the effects of the mixing procedure and
especially the introduction of the SP on the rhelogical properties of cement paste.
Finally, test results enable to underline the interactions between SP and VEA used in
designing SCC. VEA affects both viscosity and shear yield stress. However, its effect
depends on the SP dosage.
Julie Cordin is a civil engineer from the French national state school of public works
(ENTPE) in Lyon. She was temporary affected at the LCPC for her degree.
INTRODUCTION
Self-compacting concrete (SCC) is a highly fluid concrete that does not require
any vibration during the placement process. Thus, this kind of concrete is of great interest
from economical, technical and environmental points of view 1' 2' 3 • Fresh SCC must be
stable to ensure structure strength homogeneity and durability. However, several
problems like bleeding, settlement or segregation can occur, sometimes simultaneously,
on construction sites. Segregation can appear during placing (referred to as "dynamic
segregation") or afterwards, during the dormant stage (referred to as "static
segregation"t "Static segregation" consists in the sedimentation of the coarsest
aggregates of the suspension under gravity forces. A voiding segregation is a matter of
both cement paste rheology and granular skeleton 4 . The cement paste has to be
sufficiently fluid to ensure the high fluidity of the concrete and sufficiently "viscous" to
suspend coarse aggregates. In fact, whereas a non-zero shear yield stress enables to avoid
the initiation of segregation, the viscosity and the thixotropy enable to limit its effects 4 .
According to this point of view, a better characterization and understanding of the
rheological behavior of fresh pastes for sec is the first step to control the stability and
robustness of sec.
In this paper, the rheological characterization of several fresh cement pastes
obtained by varying the superplasticizer (SP) and the viscosity-enhancing admixture
(VEA) contents of a typical SCC paste, is reported. This enables to underline the possible
interactions between SP and VEA usually used in designing SCC. Moreover, possible
correlation between empirical tests like spread or flow time and rheological ones (i.e.
shear yield stress or apparent viscosity) has been investigated to simplify the rheological
characterization of flowable mixes. Finally, the influence of the mixing procedure on
rheometric measurements has been underlined.
EXPERIMENTAL PROGRAM
Cement pastes are designed from one of the SCC mixtures studied in the
framework of the French national project BAP. The 28-day strength of this SCC is about
40 MPa. Constituents include Portland cement (CEM I 52.5 N CE CP2 NF according to
the EN 197-1 5), limestone filler and aggregates, SP and VEA. The SP is a new generation
one with a modified polycarboxylate base and the VEA is a polysaccharide suspension in
vegetal oil. The cement paste referred to as "mixture 0" is the same paste as in the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
0.45 and 0.27). The SP dosage has been limited to the «saturation value» whereas four
increasing YEA contents have been investigated. However, little is known about the
influence of the VEA dosage on "the saturation value".
Beyond "the saturation value", an increase in the SP dosage does not
significantly improve the flowability, i.e. spread or flow time (Fig. 1). "The saturation
value" is usually determined from flow time results even if spread measurements are in
good agreement with flow time ones 6 :
- flow time is measured for various SP dosages,
- the logarithm of flow time is plotted against the SP dosage, expressed in terms of dry
content of SP to binder ratio,
-the saturation dosage is the point of tangency with a straight line whose slope is 2/5.
By using this method, "the saturation value" is independent of the paste volume
used 6. Preliminary tests were performed and led to the conclusion that the "saturation
value" of 0.2%, is not affected by changes in the VEA dosage, whereas the rheological
behavior is clearly modified (Fig. 2). The greater the VEA dosage, the lower the fluidity.
Empirical tests, namely spread and flow time, are performed on the twelve
cement pastes to characterize their rheological behavior just after mixing. Spread is
measured by using a mini cone (diameters: 10 and 7 em, height: 5 em). The truncated
cone mould is placed on a glass plate, filled with paste and lifted. The resulting final
diameter of the fresh paste sample is the mean value of two measurements made in two
perpendicular directions (Fig. 1). Flow time is measured by using a Marsh cone. Its
geometry and dimensions are specified in the European standard EN 445 7• The test
consists in measuring the time required for a given volume of paste (400 ml) to flow
trough the nozzle (diameter: 8mm). The initial paste volume poured into the cone is
800 ml (Fig. I).
Apparent viscosity and shear yield stress are also determined by using a rate-
controlled viscometer Haake VT 550. A rotational speed is imposed to the tool and a
torque is measured. Both rotational speed and torque are then converted to shear rate and
shear stress according to the tool geometry used and the calibration of the device.
Concentric cylinder geometry is first used. Both, the inner and outer cylinders
are covered with a rough paper to prevent any slip-surface to occur (roughness 100 11m).
A gap of 1.5 mm between the concentric cylinders is sufficient when the maximum
particle size is 100 jlm (Fig. 3). The whole test procedure is shown in Fig. 3. A pre-
shearing phase of 120 s at a shear rate of l 00 s·' is first applied to the cement paste (phase
1 and phase 2 of the test procedure). After this pre-shearing phase, each measurement is
8
carried out under steady state flow to avoid any bias • An increasing shear rate phase
followed by a decreasing one are then successively applied (phase 3 and phase 4 of the
test procedure).
v(r)= ~ (1)
r
Where r and f are shear stress and shear rate, respectively. For this
investigation, apparent viscosity V was defined at the end of the pre-shearing phase.
Two methods are used to determine the shear yield stress of each paste. The first
one consists in fitting the Herschel Bulkley model (Eq. 2) on flow curves obtained with
the concentric cylinder geometry: shear stress vs. shear rate during phases 4 of the test
procedure. Hence, the same test enables to determine both apparent viscosity and shear
yield stress. The Herschel Bulkley model is a three-parameter model typically used to
describe the behavior of cementitious materials 9 • This viscoplastic material exhibits a
yield response with a power law relationship between shear stress and shear rate above
the yield stress r0 .
(2)
Where r 0 , K and n are the parameters of the model. A second method is used
to determine the shear yield stress. The test consists of a mould filled with paste and a
Vane that rotates at a constant and extremely small speed. The generated torque evolution
is measured and is related to the shear stress. As shown in Fig. 4, the maximum value of
shear stress corresponds to the shear yield stress. Before the peak has been reached, the
sample deforms elastically. Then, as soon as the structure of the fresh material breaks
down, a maximum stress is obtained. The relevance of the testing procedure and the
definition of the experimental constraints (dimensions of the Vane compared to those of
the sample) have been already studied by Dzuy and Boger 10' 11 • This procedure has also
been applied to cement paste 12 • Measurements have been carried out for various times of
rest (ranging from 0 to 30 minutes) and hence, enable to characterize the evolution of the
yield stress r 0 • The thixotropy of the material can be quantified through the evolution of
shear yield stress with the time of rest, and the possible interactions between SP and VEA
can be evaluated accordingly. Moreover, both methods of shear yield stress evaluation
can be compared.
Repeatability has been carried out on a given cement paste in order to determine
the accuracy of each rheological parameter determination. At least five tests have been
performed in each case. The calculated accuracies are summarized in Table 2. These
values enable to give a better confidence to the analysis of results. For example, one can
notice that the accuracy of the shear yield stress determined by the Vane is better than
that extrapolated by the Herschel Bulkley model on flow curves. Both kinds of results are
in good agreement. However, the slight difference between Vane and coaxial results can
be attributed to the shearing conditions. Yield stress is measured at rest with the Vane
geometry and under steady shear flow in the case of the concentric cylinders. Yield stress
measured by the Vane will be referred to as "at-rest yield stress" in what follows.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The empirical spread and flow time test results are summarized in Table 4 and
presented in Fig. 5. Tests have been performed just after mixing. As expected, the lower
the YEA dosage and the higher the superplasticizer dosage, the greater the fluidity, i.e.
the shorter the flow time and the greater the spread. However, the fluidity is all the more
reduced by an increase in YEA, since the superplasticizer dosage is low. This is more
apparent in observing the flow time results in Fig. 5.
Rheological parameters
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The correlation curve between spread and at-rest yield stress just after mixing is
an exponential one. This can be verified by plotting the logarithm of spread against yield
stress. The correlation does not depend on the VEA dosage (Fig. 9). The relationship
between spread and shear yield stress can be written as follows:
Where v and Flt are apparent vi~cosity measured at f = 1OOs- (mPa s) and
1
flow time (s), respectively. Linear correlations can also be obtained between flow time
and viscosity defined at other shear rates ranging from 1S to 100 s- 1• The parameters are
shear rate-dependent, and a power law function enables to describe their evolution in
relation to shear rate (see reference 13 for more details).
Finally, these correlations lead to the conclusion that easier empirical tests can
be performed instead of more complex rheological ones, when determining the
rheological behavior of flowable cement pastes. They support the points of view of
several authors about the modeling of empirical tests 15 · 16 . However, according to other
published papers, weak correlations are obtained between empirical tests and rheological
parameters 17•18 . The only way to explain that good correlations have been established in
this study is to assume that the parameters of the models probably depend on the studied
binder, SP and VEA.
In this study, twelve cement pastes have been designed from the paste of a
typical SCC mixture by modifying the SP and the YEA dosages. Their rheological
behavior has been characterized to evaluate the effect of variations of SP and YEA
dosages. Test results enable the following conclusions:
1 - The use of YEA affects both viscosity and shear yield stress. An increase of YEA
content leads to an increase of rheological parameters. However, the saturation dosage of
SP is not modified by the YEA dosage.
2 -For SP dosage close to the saturation point, the YEA does not seem to modify the
rheological behavior of the paste.
3 - The shear yield stress values obtained from the Vane Geometry are in good
agreement with those obtained from flow curves (extrapolation by the Herschel Bulkley
model).
4 -From a statistical point of view, significant correlations are obtained between spread
and yield stress, and between flow time and apparent viscosity. Hence, easier empirical
tests can be performed to characterize the rheological behavior of cement pastes for a
given set of constituents.
ACKNOWLEDGMENTS
This study was carried out in the framework of the French national project BAP,
supported by the RGCU (Reseau genie civil et urbain) and the DRAST (Direction des
affaires scientifiques et techniques), France. It also received the financial support of the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
French Ministry of Research as part of a special study dealing with the stability of
cementitious materials.
REFERENCES
1. Okamura, H.; and Ouchi, M., "Self Compacting Concrete. Development, present use
and future", in: A. Skarendahl and 0. Petersson (Eds), Self Compacting Concrete,
Proceedings of the first International RILEM Symposium, RILEM Publications, La Ferte-
Mace (France), 1999, pp. 3-14.
2. Skarendahl, A., "The present- The future", in: 0. Wallevik and I. Nielson (Eds), Self
Compacting Concrete, Proceedings of the third International RILEM Symposium,
RILEM Publications, 2003, pp. 6-14.
3. Walraven, J., "Structural Aspects of Self Compacting Concrete", in: 0. Wallevik and I.
Nielson (Eds ), Self Compacting Concrete, Proceedings of the third International RILEM
Symposium, RILEM Publications, 2003, pp. 15-22.
6. De Larrard, F.; Bose, F.; Catherine, C.; and Deflorenne, F., "The AFREM method for
the mix-design of high-performance concrete", Materials and Structures, V. 30, No 201,
1997, pp. 439-446.
8. Geiker, M.R.; Brandl, M.; Thrane, L.N.; Bager, D.H.; and Wallevik, 0., "The effect of
measuring procedure on the apparent rheological properties of Self Compacting
Concrete", Cement and Concrete Research, V. 32, No 11, 2002, pp. 1791-1795.
9. De Larrard, F.; Ferraris C.F.; and Sedran, T., "Fresh concrete: A Herschel Bulkley
material", Materials and Structures, V. 31, 1998, pp. 494-498.
10. Dzuy, Nguyen Q.; and Boger D.V., "Yield stress measurement for concentrated
suspensions",Journal of Rheology, V. 27, No 4, 1983, pp. 321-349.
11. Dzuy, Nguyen Q.; and Boger D.V., "Direct yield stress measurement with the Vane
method", Journal ofRheology, V. 29, No 3,1985, pp.335-347.
12. Saak, A.W.; Jennings, H.M.; and Shah, S.P., "Characterization of the rheological
properties of cement paste for use in self compacting concrete", in: A. Skarendahl and 0.
Petersson (Eds), Self Compacting Concrete, Proceedings ofthe first International RILEM
Symposium, RILEM Publications, La Ferte-Mace (France), 1999, pp. 83-94.
13. D' Aloia Schwartzentruber, L.; Le Roy, R.; and Cordin J., "Rheological behavior of
fresh cement pastes formulated from a Self Compacting Concrete (SCC)". Accepted for
publication in Cement and Concrete Research, to be published in 2006, V. 36, No 7,
pp. 1203-1213.
14. Agullo L.; Toralles-Carbonari, B.; Gettu, R.; Aguado, A., "Fluidity of cement pastes
with mineral admixtures and superplasticizer - A study based on the Marsh cone test",
Materials and Structures, V. 32, 1999, pp. 479-485.
15. LeRoy, R.; and Roussel, N., "The Marsh cone as a viscometer: Theoretical analysis
and practical limits", Materials and Structures, No 38, January-February 2005, pp. 25-30.
17. Ferraris C.F.; Obla, K.H.; and Hill, R., "The influence of mineral admixtures on the
rheology of cement paste and concrete", Cement and Concrete Research, V. 31, No 2,
2001, pp. 245-255.
18. Lachemi, M.; Hossain, K.M.A.; Lambros, V.; Nkinamubanzi P.C.; and Bouzoubaa,
N., "Performance of new viscosity modifying admixtures in enhancing the rheological
properties of cement paste", Cement and Concrete Research, V. 34, No 2, 2004, pp. 185-
193.
*: A continuous flow was difficult to obtained for mixture n°8. The measured value is not
significant in this case.
i :--- __' \ 7
\\ I
\
\1'
.
' '
' '
nozzle
ll~$-.snnn
,--c::-,ro;;;
- 7 \____/
= 400ml
t2 ;
i
~
.s
'Y(lls)
r1 ~I9mm
h~58mm
r, ~ 20.5mm
1e=r, -r1 =l.Smm
Coaxial Geometry Test Procedure:
Phases 1 and 2: pre-shearing at a
constant rate of !Oos· during 120 s,
Phases 3 and 4 : steady state flow:
decreasing and increasing shear rate.
1<----~
Dv=4cm
D=9.9an
Fresh cen:tent paste sample: 850ml
Figure 5 --Flow time and spread of pastes formulated from SCC mix.
'""
100
Figure 6- At-rest yield stress evolution according to time of rest (Vane Geometry).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
0150-300
11!!450-600
.750-900
VEAO X4
VEA
0.25
~·
e,!!:.
i.,
rn
~
:
.c
rn
40
20
v v •
0
20 40 60 80 100
At-rest Yield Stress (Vane) (Pa)
Figure 8 --Comparison between shear yield stress values determined on flow cuNes
(Herschel Bulkley model) and those measured by Vane.
100
80
"'
0"-,~
y = -0.23x + 8.0518 _
2
R = 0.9927
I 60 \
~
~
~ 40 "'-.
~
20 \ u·v~o
+vEAx2
""'- ""-
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~
OVEAx4
-1 •vEAx6
10 20 30 40 10 20 30 40
Mini cone diameter
Spread(cm) Spread(cm)
1000
H. J-- • VEAO X2
o VEAOX4
_y = 49.354x- 302.68
R2 =0925
I
. /"
~v
1/0
10 15 20 25
Flow time (s)
Figure 10 -- Correlation between flow time and apparent viscosity ( y=1oo 1/s).
Properties of Superplasticized
Cement Mortars
by J. Golaszewski
Synopsis: The methodology and the results of the investigation into influence of
the temperature on the rheological properties of fresh mortars containing different
superplasticizers and cements are presented, and discussed in this paper. Influence
of temperature was studied using Two Point Workability Test (TPwn and modified
standard mortars ace. PN EN 196-1 which can be also considered as a model of
concrete. The following factors were taken into consideration: mix temperature (10,
20, 30oC), specific surface of cement (320, 370, 420 m2/kg), C3A content in cement
(2, 7, 12%), Na20e content in cement (0.3, 0.7, 1.1%), and superplasticizer type
(polyeter and polycarboxylate acid based superplasticizers with different molecular
weight and polymer structure).
The obtained results show that temperature strongly influences rheology of mortars
and that the nature and range of this influence changes in wide range depending
on the properties of cement and superplasticizer used. Basic effects of temperature
influence on rheological parameters of mortars in relation to mentioned above
factors are presented and discussed. These relationships can be applied to control
workability of fresh mortar and concrete. Discussion about the results covers also
mechanism of temperature influence on rheological properties of mortars. It is
concluded that compatibility of cement- SP system must be tested taking into
account temperature. TPWT made on mortars makes possible to select a cement
- superplasticizer system that is optimal in given temperature, and to collect data
necessary for control of fresh mortar and concrete workability in various temperature
conditions.
INTRODUCTION
temperature of the fresh concrete (1-3, 9-11). Knowledge of how the temperature
influences the rheology of fresh concrete with different in properties of cement -
superplasticizer systems is crucial for concrete mixture proportioning and placing. This
knowledge makes it possible to select an optimal cement - superplasticizer system, and to
control fresh mortar or fresh concrete workability in various temperatures. Unfortunately,
until now, the systematic experimental data on this topic has been limited.
The main objective of this paper is to explore the influence of temperature on the
rheological properties of fresh mortars containing different superplasticizers and different
cements. The effects of temperature on the rheological properties of mortars were
investigated using Two Point Workability Test (TPWT) (2, 7, 8).
EXPERIMENTAL
It is well documented that fresh mortar, like fresh concrete, behaves as Bingham
material, whose properties can be expressed by two fundamental rheological parameters,
the yield stress and the viscosity according the formula:
(1)
Testing program
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In this research, the following factors were taken into consideration:
• mixture temperature 10, 20, 30°C;
• cement specific surface 320, 370, 420 m2/k:g;
• C3A content in cement 2, 7, 12%
• Na20eq content in cement 0.3, 0.7, 1.1%
• superplasticizer type PEl, PE2 and PC according to Table 2;
Superplasticizer dosage (2%) and w/c (w/c = 0,45) were kept constant in the tests.
The value of g of mortars with cement of specific surface Sse = 320 m2/kg
insignificantly decreases with rising temperature; this effect is most distinct for SP PE2
mortars. At the temperature of 10°C, the g value of Sse = 320 m2 /kg cement mortars,
independently on SP type, is clearly higher then that of mortars with cements of specific
surface Sse= 370 and 420m2/kg.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The value of g of SP PE I, PE2 and PC mortars with cements of specific surface Sse
= 370 and 420m2/kg increases with rising temperature. The effect of rising temperature
on g of mortars is clearly proportional to cement specific surface, the higher the cement
specific surface, the more g increases due to rising temperature. The effect of rising
temperature on g of mortars is greatest when SP PEl is used.
Test results and its statistical analysis show that the nature of the influence of
temperature on rheological parameters of mortars depend both on the superplasticizer
type and on properties of the cement (that is to say: specific surface of cement and
contents of C3A and Na20eq in cement). The influence of temperature on rheological
parameters of SP PE and PC mortars decreases when superplasticizers with polymers of
greater molecular weight and of longer main and side chains are used, as well as with
decreasing specific surface of cement, decreasing the C3A contents in cement, and with
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Generally, in the case of cements of low specific surface and low contents of C3A
(that is to say cements of low initial chemical activity), the influence of temperature on
rheological properties of the mix is less intensive, because the increase in temperature
does not cause an increase in the hydration process speed of such importance that would
outbalance increased adsorption of the superplasticizer. As a result, g of mortars with
such cements, is independent of temperature and in some cases it even decreases with
rising temperature. Relatively small variations of h that accompany it, are probably a
consequence of changes in amount of adsorbed superplasticizer. In the case of cements of
high C3A content and/or of high specific surface (that is to say: cements of high initial
chemical activity), at higher temperature the speed of cement hydration process prevails
over the effect of increased adsorption of the superplasticizer. In such cases, especially
when the amount of free water in the mix is small and/or the amount of unabsorbed
superplasticizer in the solution is small, and/or if SP PC and PE polymers has short side
chains, changes in g and h with time may take place very fast.
CONCLUSIONS
The results presented in the paper show that temperature strongly influences
rheological properties of superplasticized mortars. Changes in rheological properties of
these mortars can vary in a wide range, showing ambiguous trends depending on
temperature, SP type, C3A and Na20eq content in cement, specific surface of cement and
interaction of these factors.
The relationships presented in the current paper can be used for control and
adjusting workability of not only mortars but also of fresh concrete. However, due to the
variety of cements and superplasticizers, the influence of temperature on workability
should be always experimentally verified.
[1] Aitcin, P-C. High Performance Concrete, EF&N SPON, London, 1998.
[2] Neville, A.M., Properties of Concrete (in polish), Polski Cement, Krakow, 2000.
[9] Grzeszczyk, S., Sudol, M. The effect of temperature upon the efficiency of new
generation superplasticizers (in polish), Cement-Wapno-Beton, Vol. VIII/LXX,
2003, No 6, pp. 325 - 331.
[10] Jolicoeur, C., Sharmann, J., Otis, N., Lebel, A., Simard, M-A., Page, M., The
Influence of Temperature on the Rheological properties of Superplasticized
Cement Pastes, 5th CANMET/ACI International Conference "Superplasticizers
and Other Chemical Admixtures in Concrete", ACI SP 173, Ed. Malhotra V.M.,
Rome, Italy 1997, pp. 379-391.
[11] Nawa, T., Ichiboji, H., Kinoshita, M., Influence of Temperature on Fluidity of
Cement Paste Containing Superplasticizer with Polyethylene Oxide Graft Chains,
6th CANMET/ACI International Conference "Superplasticizers and Other
Constituents f% Specific
Cement surface
Si02 CaO Ah03 F~03 Na,o.. so3 c3s c,s C3A C4AF
fm2 /kg
#I 22.4 66.6 4.3 5.5 0.7 3.0 60 19 2 17 370
#2 23.3 66.6 4.7 3.5 0.3 3.0 52 28 7 9 370
#3 23.3 66.6 4.7 3.5 0.7 3.0 52 28 7 9 370
#4 23.3 66.6 4.7 3.5 1.1 3.0 52 28 7 9 370
#5 23.3 66.6 4.7 3.5 0.7 3.0 52 28 7 9 320
#6 23.3 66.6 4.7 3.5 0.7 3.0 52 28 7 9 420
#7 20.3 65.8 6.6 3.1 0.7 3.0 58 15 12 10 370
Density Concentration
SP Base component
fg/cm3l f%1
PEl Ipolyeter (low molecular weight, short side chains) 1.09 18
PE2 Ipolyeter (high molecular weight, long side chains) 1.05 36
polycarboxylate acid (highest molecular weight,
PC . 1.06 40
longest main chain, shortest side chains)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Source of _g_ after 10 min h after 10 min~ after 60 min h after 60 min
variation Sig. Sig. Sig. Sig.
F-ratio F-ratio F-ratio F-ratio
level level level level
A: Temperature 40.974 0.007 0.620 0.595 119.769 0.001 22.600 0.016
B: Cement
155.093 0.001 35.892 0.008 213.552 0.001 166.787 0.001
specific surface
C: SPtype 156.911 0.001 971.961 0.000 121.349 0.001 1387.864 0.000
AB 96.373 0.002 18.353 0.019 128.083 0.001 4.637 0.119
AC 18.296 0.019 25.041 0.012 22.645 0.014 14.496 0.027
BC 8.634 0.054 12.601 0.032 13.229 0.030 24.175 0.013
ABC 5.938 0.085 22.600 0.013 9.743 0.044 8.695 0.051
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~ Sw c- 420m2/kg
E 30
E
z g after 60 rrin
ti 20
-+- Sw c - 320 m2/kg
10 ~ Sw c- 370m2/kg
- - - Sw c - 420 m2/kg
0
10 20 30
TBIM"ERA TlJRE, "C
50 g after 10 rrin
SP PE2 ~ Swc- 320m2/kg
40
-lr- Sw c - 370 m2/kg
~~
z g after 60 rrin
ti 20
-+- Sw c - 320 m2/kg
10 ~ Swc- 370m2/kg
- - - Sw c - 420 m2/kg
0 +------r------r-----~
10 20 30
TBIM"ERAlURE. •c
50 g after 10 rrin
SP PC ~ Sw c- 320m2/kg
40
-lr- Sw c - 370 m2/kg
~ Sw c- 420m2/kg
~
E 30
E
z g after 60 rrin
ti 20
-+- Sw c - 320 m2/kg
10 ~ Sw c- 370m2/kg
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
- - - Sw c - 420 m2lkg
0
10 20 30
Te.IIPERA lURE. •c
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
10 --+-C3A-7%
-C3A-12%
0
10 20 30
TBIIIPffiA lURE, °C
50 g after 10 rrin
SP PE2 ~C3A-2%
40
--tr- C3A - 7%
E 30 -o--C3A -12%
E g after 60 ITin
z
o; 20 -+-C3A-2%
10 --+-C3A-7%
-C3A-12%
0
10 20 30
TEMPERA lURE, oc
50 g after 10 rrin
SP PC ~C3A-2%
40
--tr- C3A - 7%
~
E 30 -a- C3A - 12%
E g after 60 rrin
z
o; 20 -+-C3A-2%
10 --+-C3A-7%
-C3A-12%
0
10 20 30
TBIIIPffiAlURE, oc
Figure 4 --The effect of temperature on g of mortars with CEM I cements different in
C A content and with SP PC and PE.
3
~
~C3A-7%
20 -o-C3A-12%
"'
E
h after SO rrin
E
z
~ 10 --C3A-2%
......... C3A- 7%
---C3A-12%
0
10 20 30
T8\11PffiATURE, OC
30 h after 10 rrin
SP PC --¢-C3A-2%
~
~C3A-7%
20
-o-C3A-12%
"'
E
E h after SO rrin
z
.: 10 --C3A-2%
......... C3A-7%
---C3A -12%
0
10 20 30
T8I.FERA TURE, OC
~
E 30 -o-Na20e-1,1%
E
z g after 60 nin
.,;; 20
-+- Na20e - 0,3%
10 __.__ Na20e- 0,7%
-11-Na20e -1,1%
0
10 20 30
TEI.1PERA ~ •c
50 g after 10 nin
SP PE2 --Q- Na20e - 0,3%
40
---lr- Na20e- 0,7%
E 30 -o-Na20e -1,1%
E
z g after 60 nin
.,;; 20
-+- Na20e - 0,3%
10 __.__ Na20e- 0,7%
- I I - Na20e- 1'1%
0
10 20 30
TEMPmATURE, •c
50 g after 10 nin
SP PC --Q- Na20e - 0,3%
40
-t.- Na20e- 0,7%
~
E 30 -o-Na20e-1,1%
E
z g after 60 nin
.,;; 20
- - Na20e- 0,3%
10 __.__ Na20e- 0,7%
-11-Na20e-1,1%
0
10 20 30
TEI.1PERA TURE. •c
10 20 30
1Bv1PERAlURE, •c
Figure 7-- The effect of temperature on h of mortars with CEM I cements different in
Na 2 0eq content and with SP PC and PE.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Hydrophobic Admixtures to
Control Corrosion of Galvanized Steel
Reinforcement in Concrete
INTRODUCTION
Static, dynamic and cyclic loading as well as shrinkage, creep and thermal stress or even,
in precast concrete, mechanical shock and flexural stress induced during transportation,
lifting and mounting, can cause concrete cover cracking. Cracks greatly increase the
concrete surface permeability since they represent preferential paths for penetration of
aggressive agents such as chloride ions, which promote corrosion of steel ( 1-6). After the
initial induction period, the degradation process grows very rapidly, since the
deterioration mechanisms show a destructive expansive nature (7).
In this respect, it is evident that steel corrosion would not occur if water, which acts as
the main carrier for aggressive substances and the medium where corrosion reactions
develop, is permanently prevented from wetting the concrete porous structure. The water
in contact with a porous material such as concrete penetrates the material by means of
capillary forces following the Washburn equation, but the molecular attraction between
water and the concrete pore walls can be lowered by surface treatment based on
impregnation with hydrophobic agents, such as those currently named silanes and
siloxanes (9-21). Their alkoxylic groups are chemically bound to the concrete hydrated
silicates by a condensation reaction while their hydrophobic alkylic groups come out on
the pores surface (22).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
However, if galvanized steel reinforcement is used, in which case it has been already
demonstrated that passivation is mainly promoted by oxygen (43), it is easy to predict
that the hydrophobic admixture could effectively counteract the corrosion process.
Therefore, the present paper is aimed to verify the validity of this intuition by monitoring
the corrosion behavior of galvanized steel reinforcement embedded in sound or cracked
hydrophobic concrete exposed to chloride environment.
EXPERIMENTAL
Concrete mixtures with w/c of 0.45 and 0.75 were manufactured by using commercial
Portland cement type CEM IIIA-L 42.5 R with and without silane as a hydrophobic
admixture. The silane admixture was used in form of 45% aqueous emulsion of butyl-
ethoxy-silane at a dosage of 1% by cement weight. Natural sand and gravel with 6-mm
and 11-mm maximum size, respectively, were used as aggregate. The proportions of the
concrete mixtures are given in Table 1, where compressive strengths at 2, 7, and 28 days
are also reported. In the presence of silane admixture, a reduction in the concrete
compressive strength of about 18 and 23% was recorded with respect to that of non-
hydrophobic concrete when w/c ratio of0.45 and 0.75 was used, respectively.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
For each concrete type, nine prismatic specimens (70x70x280 mm in size) were
produced. These prismatic specimens were reinforced with a hot dip galvanized steel
plate (2IOx40xJ mm) embedded at mid depth from a specimen side. The zinc coating,
obtained by immersion in molten zinc, was 100-f..lm thick with an outer pure zinc layer
about 20-f..lm thick. An electric cable, isolated through a PVC sheath, was connected to
each steel plate by means of spot welding. The weld was protected by epoxy resin before
casting the concrete mixture. Steel plates instead of usual rebars were used in order to
facilitate the cracking of specimens and to modulate the crack width.
After casting, all the specimens were wet cured for 2 days and air dried for 1 month at
room temperature. Then, six specimens for each concrete type were deliberately cracked
by a flexural stress in order to open a crack 0.5 or 1 mm wide in the middle zone, with the
crack apex reaching the steel reinforcement. In order to localize the crack formation at a
predetermined point, a preformed V-notch, 10 mm deep and 14 mm wide, was produced
on these specimens by placing a plastic rod on the bottom surface of the mould. Once the
crack of required width, measured by an extensometer, was obtained its opening was
Subsequently, the specimens were exposed to wet-dry cycles (2 days wet followed by 5
days dry) in a 10% NaCl aqueous solution for four months.
The corrosion risk of the reinforced concrete specimens exposed to the chloride
environment W!iS evaluated by free corrosion potential measurements .with respect to a
saturated calomel electrode (SCE) as reference, while the kinetics of the corrosion
process was followed by polarization measurements. The polarization resistance, in
inverse relation with· the corrosion rate, was measured through the galvanodynamic
method, using an external graphite bar as counter-electrode, by calculating its average
value between the anodic and cathodic zones. The electrochemical values reported in the
graphs are averaged among the ~easurements carried out· on three specimens of each
concrete type during the immersion period.
In order to validate and to complete the evaluation of the electrochemical behavior, the
galvanized steel plates were extracted by splitting the concrete specimens after 16 wet-
dry cycles in the chloride solution and the corrosion extent was assessed by visual
observation. Metallographic analysis was carried out on a cross section of the galvanized
steel plates in order to evaluate the decrease in the zinc coating thickness due to the
corrosive attack. A cross section 1 em from the crack apex was chosen in order to reflect
the effects of a real corrosive process, avoiding the extreme situation produced at the
crack apex by a strongly accelerated corrosion process in order to reduce experimental
times. Zinc corrosion products were identified by X-ray diffraction. Additionally, the free
chloride concentration derived by water extraction from the cement paste in contact with
the steel reinforcement in the same cross section was measured at the end of the exposure
time to the chloride environment.
Electrochemical measurements
Figure 1 shows the free corrosion potential values of galvanized steel plates embedded in
concrete specimens with w/c = 0.75 as a function of wet-dry cycles in the aggressive
environment. In non-hydrophobic concrete they assume active values of about -1000
mV/SCE reflecting a generally great corrosion risk and they keep constant for almost all
the test time. On the other hand, in bulk hydrophobic concrete, just after the exposure to
the aggressive environment the galvanized steel plates assume values of -800 mV/SCE in
the presence of concrete cracks regardless of the crack width, and -600 mV/SCE in sound
specimens. Moreover, the free corrosion potentials move towards more passive values,
indicating a very low corrosion risk, after a few wet-dry cycles. Cracks in the concrete
cover do not significantly affect the reinforcement potential.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The two different crack widths considered do not affect the corrosion results in non-
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
hydrophobic concrete specimens, due to the high porosity of the cement matrix, while, in
the hydrophobic ones, just a little bit higher corrosion rate is detected in the presence of
the wider crack.
A good quality concrete matrix with a w/c as low as 0.45 seems to partially hide the
beneficial effect of the hydrophobic admixture. The corrosion risk described by the free
corrosion potential measurements (Figure 3) is about the same regardless of the cement
matrix. However, the polarization resistance of steel reinforcement (Figure 4) is always
higher in the hydrophobic concrete specimens with respect to the corresponding non-
hydrophobic ones, even if the difference is not as evident as observed in the more porous
cement matrix. In particular, in this case, the hydrophobic effect was not able to cancel,
from a corrosion point of view, the detrimental effect of the concrete cover cracking as
previously observed in concretes with w/c = 0.75. Moreover, due to the low porosity of
the cement matrix, the crack width influenced the test results both in the presence and in
the absence of silane with corrosion risk and corrosion rates which increase with the
crack width.
Finally, by comparing the results obtained in concrete specimens with w/c = 0.45 and
0.75, it is evident that the hydrophobic admixture, even in the presence of concrete
cracks, decreases the corrosion rate monitored in very porous concrete, as that
manufactured with w/c = 0.75, to values comparable with those obtained in good quality
concrete, as that manufactured with w/c = 0.45. In other words, silane cancels the
detrimental effect, at least from the corrosion point of view, of a large porosity of the
cement matrix.
Visual evaluation
Visual and metallographic observations carried out on the galvanized steel plates
removed from the specimens after the aggressive exposure confirmed the electrochemical
measurements.
On the other hand, for the galvanized steel plates embedded in hydrophobic concrete
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
(Figure 5b), zinc grains were still well visible on the galvanized plate surface and close to
the crack apex a white surface deposit of adherent and compact zinc corrosion products,
later identified by X-ray diffraction as zinc oxides, hydroxides and calcium
hydroxyzincate, appeared. In particular calcium hydroxyzincate is a well known
passivating zinc corrosion product that, once formed, protects the underlying pure zinc
layer from further corrosion as metallographic analysis has well demonstrated.
As a matter of fact, the galvanized steel plates extracted from non-hydrophobic concrete
(Figure 6a) showed a very deep general corrosive attack with red rust appearing on the
surface indicating that total consumption of the protective zinc layer had occurred.
On the other hand, the zinc layer on steel reinforcement embedded in hydrophobic
concrete (Figure 6b) appeared thicker, as shown by the metallographic section, in spite of
an apparently stronger corrosive attack with respect to that observed in hydrophobic
concrete with lower w/c ratio (Figure 5b), indicating that the corrosive attack is weaker.
It must be pointed out that this behavior was observed regardless of the chloride
concentration. This was measured close to the plates embedded in hydrophobic concrete
with w/c = 0.75 and it resulted in a concentration four times higher with respect to that
measured with w/c = 0.45 (Table 2). Such an observation confirms greater oxygen
diffusion, as that assured by more porous hydrophobic cementitious matrix manufactured
with higher w/c, can really favor galvanized reinforcement passivation.
CONCLUSIONS
The use of hydrophobic admixture can improve the corrosion resistance of galvanized
steel reinforcement in concrete specimens exposed to wet-dry cycles in a chloride
aqueous solution, even in the presence of concrete cracks, especially when a high w/c is
used.
Bulk hydrophobization reduces water absorption approximately 50% and, at least from
the corrosion point of view, cancels the detrimental effect of a large porosity of the
cement matrix as well as, in high porous concretes, of concrete cover cracking.
Galvanized steel in hydrophobic concrete not only assures the concurrence of the benefits
derived from the use of either steel galvanization or concrete hydrophobization, but it
could further provide a useful synergistic effect. As a matter of fact, hydrophobic
REFERENCES
1) Allan, M.L., and Cherry, B.W., 1992, "Factors Controlling the Amount of Corrosion
for Cracking in Reinforced Concrete", Corrosion, V. 48, No.5, pp. 426-430.
4) Sakai, K., and Sasaki, S., 1994, "Ten Years Exposure Test ofPrecracked Reinforced
Concrete in a Marine Environment", Durability of Concrete: Proceedings of the
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
5) Tuutti, K., 1982, "Corrosion of Steel in Concrete", Report 4/82, Swedish Cement
and Concrete Research Institute, Stockolm, Sweden, p. 468.
6) Ueda, T., et al., 1998, "Analytical Study on Concrete Cover Cracking Due to
Reinforcement Corrosion", Concrete under Severe Conditions: Proceedings of the
Second International Conference, Tromso, Norway, 1998 (CONSEC '98), eds. O.E.
Gj0rv et al., E & FN Spon, London, U.K., pp. 678-687.
7) Page, C.L., and Treadaway, K.W.J., 1982, "Aspects of the Electrochemistry of Steel
in Concrete", Nature, 291, pp. 109-115.
8) de Sitter Jr., W.R., 1983, "Costs for Service Life Optimization, the Law of Fives",
Durability of Concrete Structures: Proceedings of CEB-RILEM International
Workshop, Copenhagen, Denmark, 1983, CEB Bulletin d'Information, No.152, pp.
131-134.
10) Basheer, P.A.M., 1995, "Surface Treatments for Concrete", Corrosion and
Protection of Metals in Contact with Concrete: Proceedings of COST 509 Workshop,
Sevilla, Spain, 1995, eds. C. Andrade, R. Cox and B. Pukl, Key-Note Lectures, 50
pp.
12) Basheer, P.A.M., et al., 1997, "Surface Treatments for Concrete: Assessment
Methods and Reported Performance", Construction and Buildings Materials, V. 11,
No. 7-8, pp. 413-429.
13) Moriconi, G., et al., 2002, "Review of silicone-based hydrophobic treatment and
admixtures for concrete", The Indian Concrete Journal, V. 76, No. 10, pp. 637-642.
14) Chowdhury, A., et al., 2001, "Vapour phase application of hydrophobic coatings to
cement-based materials", Magazine of Concrete Research, V. 53, No. 5, pp. 347-
352.
16) ACI Committee 515, 1985, A Guide to the Use of Waterproofing, Damp-roofing,
Protective and Decorative Barrier Systems for Concrete (ACI 515.1R-79, Revised
I985), American Concrete Institute, Farmington Hills, MI, 44 pp.
17) Syropolou, S., et al., 2003, "Development of a Rapid, Non-destructive In-situ Test
for Assessing Surface Treatments on Concrete", Durability of Concrete:
Proceedings of the Sixth CANMETIACI International Conference, Thessaloniki,
Greece, 2003, Supplementary Papers, pp. 557-571.
19) Rodrigues, M.P., and Salta, M.M., 2003, "Surface Treatments for Protection of
Reinforced Concrete against Corrosion", Durability of Concrete: Proceedings of the
Sixth CANMETIACI International Conference, Thessaloniki, Greece, 2003,
Supplementary Papers, pp. 297-311.
22) Wong, K.H., et a!., 1983, «The Retardation of Reinforcing Steel Corrosion by
Alkyl-Alkoxy Silane", Cement & Concrete Research, V. 13, No.6, pp. 778-788.
23) Rodrigues, M.P., 1998, "Evaluation of the durability of surface coatings for concrete
protection", Corrosion of Steel in Reinforced Concrete Structures: Proceedings oj
· COST 521 Workshop, Utrecht, The Netherlands, 1998, ed. R.B. Polder, Annual
Progress Report 1997/1998 Project P4. .
24) Raupach, M., and Wolff, L., 2003a, "Long-Term Durability of Surface Protection
Systems", Durability of Concrete: Proceedings of the Sixth CANMETIAC]
International Conference, Thessaloniki, Greece, 2003 (SP-212), ed. V.M. Malhotra,
American Concrete Institute, pp. 45-58.
25) Raupach, M., and Wolff, L., 2003b, "Investigations on the Long-Term Durability of
Hydrophobic Treatments on Concrete", Durability of Concrete: Proceedings of the
Sixth CANMETIACllnternational Conference, Thessaloniki, Greece, 2003 (SP-212),
ed. V.M. Malhotra, American Concrete Institute, pp. 409-426.
26) Raupach, M., et a!., 2002, "Monitoring the durability of hydrophobic treatment at
existing concrete structure'', lnternationale-Zeitschhrift-fuer-Bauinstandsetzen-und-
Baudenkmalpflege, V. 7, No. 2/3, pp. 255-269.
28) Sandin, K., 1999, "Influence of cracks on moisture conditions in facades with water-
repellent treatments", Internationale-Zeitschhrift-fuer-Bauinstandsetzen-und-Bau-
denkmalpjlege, Vol. 5, No.5, pp. 499-521.
29) Gaislon, R.S., 1999, "Water repellents: protection of facades of concrete in a wet
climate", lnternationale-Zeitschhrift-fuer-Bauinstandsetzen, V. 5, No. 3, pp. 251-
272.
30) Calder, A.J.J., and Chowdhury, Z.S., 1996, "A Performance Specification for
Hydrophobic Materials for Use on Concrete Bridges", Corrosion of Reinforcement in
Concrete Construction, eds. C.L. Page, P.B. Bamforth and J.W. Figg, SCI Special
Publication No.183, pp. 556-566.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
33) Jones, M.R., et al., 1995, «Concrete Surface Treatment: Effect of Exposure
Temperature on Chloride Diffusion Resistance", Cement & Concrete Research, V.
25, No. 1, pp. 197-208.
34) Sauder, M., 1999, "Damage caused by water repellent agents- reasons and counter-
measures", Internationale-Zeitschhrift-fuer-Bauinstandsetzen, V. 5, No. 3, pp. 311-
321.
35) Vivian, H.E., 1960, "Some Chemical Additions and Admixtures in Cement Paste and
Concrete", Proceedings of the Fourth International Symposium on the Chemistry oj
Cement, Washington, U.S.A., 1960, pp. 909-926.
36) Tittarelli, F., et al., 2000a, "Comparative Evaluation of Traditional and Innovative
Corrosion Protection Methods in Cracked Reinforced Concrete Exposed to Chloride
Environment", Durability of Concrete: Proceedings of the Fifth CANMETIACJ
International Conference, Barcelona, Spain, 2000 (SP-192), ed. V.M. Malhotra,
American Concrete Institute, pp. 613-627.
39) Fratesi, R., et al., 1997, "The Influence of Hydrophobic Concrete on the Corrosion of
Rebars", Superplasticizers and Other Chemical Admixtures in Concrete:
Proceedings of the Fifth CANMETIACI International Conference, Rome, Italy, 1997
(SP-173), ed. V.M. Malhotra, American Concrete Institute, pp. 105-122.
40) Alred, J.M., et al., 2001, "The effect of initial moisture content on water transport in
concrete containing a hydrophobic admixture", Magazine of Concrete Research, V.
53, No.2, pp. 127-134.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
43) Tittarelli, F., et al., 1999, "The Role of Oxygen on the Passivation Mechanism of
Galvanized Steel Rebars", Infrastructure Regeneration and Rehabilitation -
Improving the Quality of Life Through Better Construction - A Vision for the Next
Millennium: Proceedings of International Conference, Sheffield, U.K., 1999, ed.
R.N. Swamy, Sheffield Academic Press, pp. 517-526.
Table 1 - Concrete mixture proportions (kg/m3) and related compressive strengths (MPa)
w/c = 0.45 w/c=0.75
Mixtures w/c= 0.45 w/c = 0.75
hydrophobic hydrophobic
w/c= 0.45
0.2 0.4 0.5
hydrophobic
w/c= 0.75
1.4 1.5 1.9
hydrophobic
8 10 12 14 16
Number of cycles
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Figure 1- Potential of the galvanized steel plate embedded in concrete specimens
with w/c = 0.75 as a function of the number of wet-dry cycles.
1600.-------------------------,
1000
16
Nu.f.ber of c~cles
12 14
-!ZOO,------------------------,
~no crack hydrophobic .... no crack non-hydrophobic
-1100 ~0.5 mm crack hydrophobic . . . 0.5 mm crack oon-bydrnphobic
-i:P 1 mm crack hydrophobic _._ 1 m.m enclt .non-hydropbobk
-1000 t-----------------------1
~-~+---------------------~
~
~ ~00+---------------------~
!.
lilil -700
400L--~--~--~-----~-~--~-~
B 10 12 14 16
Number of cycles
1000 ....,
§:..., ~
r:l" ..,._/
~ .....- ~
/:..
400
... ......
~
zoo
-- ~ 'iF""'
0
0 8 10 1Z 14 16
Number of cycles
~no crack n
I mm crack ,
~
, 'v
.•.
.
'!', 4
'
'
< '
1 111m crack ~
:
t
' '
(a) (b)
Figure 5- Visual observation and metallographic section of the galvanized steel plate
embedded in cracked concrete specimens with w/c = 0.45 in the absence
(a) and in the presence (b) ofthe hydrophobic admixture.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
(a) (b)
Figure 6- Visual observation and metallographic section of the galvanized steel plate
embedded in cracked concrete specimens with w/c = 0.75 in the absence
(a) and in the presence (b) of the hydrophobic admixture
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: The interactions between six commercial cements and three types
of plasticizers, namely naphthalene sulfonate formaldehyde condensate (SNF),
lignosulfonate (LS) and polyacrylate grafted with polyether (PA) are investigated
using cement paste. PA was found to be the most efficient plasticizer of the three
tested even though it was found to present lower adsorption onto cement than SN F
and LS. SNF and LS brought about comparable results. PA was observed to induce
flow gain of pastes within 2 hours of rheological measurements, while pastes made
with SNF and LS exhibited flow loss as a function of time.
The adsorption characteristics were found to depend on the plasticizer type. The
plasticizer saturation dosages were, moreover, found to depend on cement surface
area (Blaine fineness), amount of cubic C3A and soluble sulfates.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Dr. Harald Justnes is Chief Scientist at SINTEF Concrete, Trondheim, Norway. His field
of interest covers the chemistry of cement, concrete, admixtures and additives from
production, through reactivity, to durability. Dr. Justnes is Adjunct Professor in "Cement
and Concrete Chemistry" at NTNU, Trondheim.
Renato Figi is leader of the analytical chemistry group at EMP A Swiss Federal
Laboratories of Materials Testing and Research, Diibendorf, Switzerland.
INTRODUCTION
EXPERIMENTAL
The cementitious pastes were prepared using cements 1-5 described in Table 1 and
all three admixtures. Mixtures made with SNF and LS were prepared with w/c of 0.40.
The blending of the pastes was performed in a high-shear mixer from Braun
(MR5550CA) by adding solids to the water and mixing for 'li minute, resting for 5
minutes and blending again for 1 minute. The superplasticizers were either added with
the water first (immediate addition) or 5 minutes after water addition (delayed addition).
The pastes were filtered through 0.45 11m-filters 20 minutes after water addition. The
pore solutions were diluted 50-200 times with a solution of"artificial pore water" (NaOH
and KOH with a KINa molar ratio equal to 2 and pH = 13.2). The amount of
superplasticizer in the water phase was read from the standard curve which had been
made with diluted series of SNF (290 nm) and LS (283 nm). The standard curve is thus
the relationship between the measured absorbance and the plasticizer concentration in the
pore water. The difference between the added and measured contents of superplasticizer
gave the consumed portion, i.e. plasticizer adsorbed onto cement grains and hydration
products and intercalated portion of the superplasticizer. The consumed amounts of
plasticizer are distinguished as adsorbed and intercalated since previous reports have
indicated that some of the plasticizer molecules will be embedded by the hydration
products<8'9>.
The concept of yield stress is sometimes a very good approximation for practical
purposes. The Bingham model applies, however, often only for limited parts of the flow
curve if the tested material has shear thinning or shear thickening flow behaviour. Shear
thickening flow behaviour may furthermore result in negative yield stress at high shear
rates. Such a negative yield stress has no physical meaning. There has been raised a
question whether or not a yield stress exists or whether all non-Newtonian materials will
exhibit a finite zero-shear viscosity. Barnes and WaltersC 13l states that the yield stress
concept is an idealization and that given accurate measurements, no yield stress exists.
They claim that even concentrated systems flow in the limit of very low stresses and that
these materials appear not to flow merely because the zero shear viscosity is so high.
Yahia and Khayat114> support this statement by finding that the yield stress correlated well
to the apparent viscosity at low shear rate (5 s· 1).
It was in this work decided to use the area under the flow curve (Pals) as a
measurement of "flow resistance" rather than describing the flow curve with the usual
Bingham model. The flow resistance was calculated as the area under the flow curve
using the trapezoidal rule:
(2)
The flow resistance could, of course, have been obtained at constant shear rate, but a
shear rate range of 43-8.8 s- 1 consisting of six measuring points relevant for mixing and
pouring concrete05 l was used in order to compare with parameters from the Bingham
model. The question is whether the area under the flow curve represents something more
"physical" than an "apparent" yield stress from Bingham modeling. In a parallel plate set-
up with shear area, A [m 2], and gap h [m] between the plates;
~y = vlh (4)
(shear rate difference over an interval)
where F [N] is the force used to rotate the upper plate, and v [m/s] is the angular velocity.
where V [m3] is the volume of the sample between the parallel plates.
The unit of the area under the curve can be expressed as N·mlm3·s, J/m3·s or W/m3 • It
is in other words, the power required to make a unit volume of the paste flow with the
prescribed rate in the selected range. The power, P [W], required to mix concrete for a
certain time interval is actually sometimes measured by simply monitoring voltage (U
[V]) and current (I [A]) driving the electrical motor of the mixer, since P = Uxl.
amount of polyelectrolyte adsorbed is always less than the polyelectrolyte dosage, with
the isotherm gradually tapering off with increasing polyelectrolyte concentration rather
than attaining a definite adsorption plateau07l. Such a linear isotherm is consistent with
conditions in which the number of adsorption sites remains constant throughout the
whole range of solute (plasticizer) concentrations until the pore water is saturated. This
means that the surface available for adsorption expands (due to dispersion and hydration)
proportionally with the amount of plasticizer adsorbed< 16l.
In some cases as seen for the delayed addition of SNF to Cement 1 (Figure 1) and
immediate addition of PA to Cement 2 (Figure 2) the consumption decreases after
reaching a maximum. This might indicate that surplus molecules in the water phase are
compressing the ionic double layer or that adsorbed molecules expand and hinder
molecules in the water phase to attach at the surface. A similar phenomenon was
discovered in the case of cement by Andersen et al.< 5l and Yoshioka et al. 08l. The
occurrence of a maximum is especially pronounced for the PA-isotherms. This is
probably caused by the high molecular weight of this superplasticizer which renders the
Experiments with delayed addition of polyacrylate were not performed since Vikan
and Justnes< 20 ) established that delayed addition of PA only causes marginal decrease of
flow resistance. Delayed addition of SNF plasticizer to Cement 1 (Figure 1) resulted,
however, in strong reduction of flow resistance and reduced amount of consumed
plasticizer. The main reasons for the decrease in flow resistance and amount of consumed
superplasticizer in the case of delayed addition may be reduced intercalation of the
plasticizer in the newly formed ettringite dominating a possible increased surface
adsorption. Furthermore, "free" or available excess superplasticizer in the interstitial
solution which do not intercalate or adsorb onto cement particles have also been found to
contribute to increased flow of the paste< 21 • 22 ).
It is not straight forward to analyze the adsorption data as a function of time since
the concentration of superplasticizer in the pore water might decrease by intercalation in
the hydration products or increase as the pore water is consumed by the hydration
process. The concentration of PA, LS and SNF in the pore water did, however, not
change noticeable with time. This observation is backed by Chiocchio and Paolini(lS) who
claimed that the adsorption of SNF and SMF is completed after only 1-2 minutes.
It was found that flow change with time is dependent on the cement type and type
and dosage of superplasticizer. Flow loss was typical for pastes with SNF. Pastes with LS
generally developed lower flow loss than pastes with SNF and flow gain was observed
for several pastes. Similar results have been found by Uchikawa et al.<26) who found that
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
LS forms complex salts with Ca2+ more easily than SNF. Thus, the complexation of ci+,
delays the saturation of Ca2+ in the pore water, which influences the morphology of
hydrate produced (e.g. larger [Ca3(Al,Fe)(OH) 6]X3 ·xH20 (AFt) crystals are produced
with LS than with SNF) and delays the setting time of cement.
Another noticeable feature for SNF is that the flow loss has a turning point at the
saturation dosage whereafter the flow loss rates increase noticeably. Nawa et al.< 27 ) found,
Flow gain was the general trend for most pastes with PA. These results which are
backed up by Nawa et al. (Z?l and Golaszewski and Szwabowskic23 l might be due to
delayed adsorption or preferred adsorption on hydration products.
No turning point for flow loss as a function of PA dosage has been detected which
might indicate that the grafted chains of the molecules are long enough to provide steric
stabilization even if the stem is adsorbed and covered by hydration products. Table 2
shows flow loss data for Cement 2 with different dosages of the three plasticizers studied.
Ranking of the adsorption capacity of the cements seems to depend upon the
superplasticizer dosage. SNF molecules adsorb somewhat stronger on cement than LS
below saturation while the opposite trend is found at saturation dosage. All in all, it can
be concluded that SNF and LS are adsorbed roughly to the same extent to the cements.
Cement 3 (cement with fly ash) consumes less superplasticizer than expected
regarding its relatively high surface (Blaine) as shown in Table 3. Experiments to study
the influence of fly ash on the superplasticizer adsorption were therefore performed.
Adsorbance of LS and SNF were measured on pastes of pure Cement 1, pure ground fly
ash (FA), pure FA with 2% by weight Ca(OH) 2 and pastes with 50% FA and 50%
Cement 1 measured by weight. The results given in Table 4 ·show that pure fly ash
adsorbs less superplasticizer than cement, while the amount of consumed plasticizer for
the cement-FA mix was found to be close to the results of pure cement. These results
seem to coincide with the work of Nagataki et al.C 28 l. Pure FA pastes with 2% Ca(OH)z
adsorbed approximately the same or slightly higher amount of plasticizer as pure cement
paste. This result seems to indicate that Ca2+ adsorbs on the FA surface and that the
anionic superplasticizer molecules adsorb on the positive charged sites thus created.
Jolicoeur and Simardc29l pointed out a similar effect after the observation of SNF being
adsorbed to a minor extent by silica fume slurry, but being equally adsorbed by cement
paste and cementitious paste with silica fume replacement.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Good correlations were found for the saturation points of SNF with the product of
Na.,qv and Blaine fineness as well as the product of Naeqv and cubic C 3A. The best
correlation was found for the product of cubic C3A and Blaine as illustrated by Figure 4a.
This is logical since high surface and aluminate content in the cement means high C3A
surface area enabling more intercalation.
CONCLUSIONS
PA was found to be a more effective plasticizer than SNF and LS, even though it
was consumed to a lesser extent by the cements. SNF and LS gave comparable results.
P A resulted moreover in flow gain as a function of time while SNF and LS resulted in
cement pastes with flow loss.
REFERENCES
1. Bonen, D., Sarkar, S. L., 1995, 'The Superplasticizer Adsorption Capacity of Cement
Pastes, Pore Solution Composition, and Parameters Affecting Flow Loss", Cem. Caner.
Res. V. 25, No.7, pp.l423-1434
2. Ai"tcin, P.-C., Jolicoeur, C. and MacGregor, J. G., 1994, "Superplasticizers: How They
Work and Why They Occasionally Don't", Concrete International, V. 16, No.5, pp. 45-
52
4. Ohta, A., Sugiyama, T., Tanaka, Y., 1997, "Fluidizing Mechanism and Application of
Polycarboxylate-Based Superplasticizers", SP 173-19, Proceedings of 5th CANMET/ACI
International Conference on Superplasticizers and Other Chemical Admixtures in
Concrete, Rome, Italy, October 7-10, pp. 359-378.
5. Andersen, P. J., Roy, D. M. and Gaidis, J. M., 1988, "The Effect of Superplasticizer
Molecular Weight on its Adsorption on, and Dispersion of, Cement", Cem. Caner. Res.,
V. 18, No.6, pp. 980-986
6. Basile, F., Biagini, S., Ferrari, G., Collepardi, M., 1989, in third CANMET/ACI
International Conference, Superplasticizers and Other Chemical Admixtures in Concrete,
Ed .. Malhotra, Ottawa, pp. 171-188
7. Spiratos, N., Page, M., Mailvaganam, N. P., Malhotra, V. M., Jolicoeur, C., 2003,
"Superplasticizers for Concrete: Fundamentals, Technology, and Practice",
Supplementary Cementing Materials for Sustainable Development Inc., Ottawa, Canada
9. Prince, W., Edwards-Lajnef, M., AYtcin, P. -C., 2002, "Interaction Between Ettringite
and a Polynaphtalene Sulfonate Superplasticizer in a Cementitious Paste", Cem. Caner.
Res. V. 32, pp. 79-85
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
11. Atzeni, C., Massida, C. and Sanna, U., 1985, "Comparison Between Rheological
Models for Portland Cement Pastes", Cern. Concr. Res. V. 15, pp. 511-519
12. Yahia, A. and Khayat, K. H., 2003, "Applicability of Rheological Models to High-
Performance Grouts Containing Supplementary Cementitious Materials and Viscosity
Enhancing Admixture", Materials and Structures, V.36, pp. 402-412
13. Barnes, H. A., and Walters, K., 1985, "The Yield Stress Myth?", Rheol Acta V. 24,
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
pp. 323-326
14. Yahia, A. and Khayat, K. H., 2001, "Analytical Models for Estimating Yield Stress of
High-Performance Pseudoplastic Grout", Cern. Concr. Res. V. 31 pp. 731-738
15. Wallevik, 0. H, 2003, "Rheology of Cement Based Particle Suspensions such as Self-
Compacting Concrete", IBRI Iceland, Chapter 3 p.13.
16. Giles, C. H., D'Silva, A. P., Easton, I. A., 1974, "A General Treatment and
Classification of the Solute Adsorption Isotherm. Part I and II", J. Colloid Interface Sci.
V. 47, No. 3, pp. 755-778
17. Ratinac, K. R., Standard, 0. C., Bryant, P. J., 2004, "Lignosulfonate Adsorption on
and Stabilization of Lead Zirconate Titanate in Aqueous Suspension", J. Colloid and
Interface Science, V. 273, pp. 442-454
18. Yoshioka, K., Tazawa, E., Kawai, K., Enohata, T., 2002, "Adsorption Characteristics
ofSuperplasticizers on Cement Component Minerals", Cern. Concr. Res. V. 32, pp.l507-
1513
19. Myers, D. 1992, "Surfactant Science and Technology", 2nd Edition, VHC, pp.267-290
20. Vikan, H. and Justnes, H., 2004, "Influence of Silica Fume on the Rheology of
Cementitious Paste", Proceedings of the 81h CANMET/ACI International Conference on
Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete (Ed. V. Mohan Malhotra),
Las Vegas, USA, May 23-29, CANMET/ACI SP-221-25, pp. 427-442.
22. Kim, B. -G., Jiang, S., Jolicoeur, C., Ai"tcin, P. -C., 2000, "The Adsorption Behaviour
ofPNS Superplasticizer and its Relation to Fluidity of Cement Paste", Cern. Caner. Res.
v. 30, pp. 887-893
23. Golaszewski, J. and Szwabowski, J., 2004, "The Influence of Superplasticizers on the
Rheological Behaviour of Fresh Cement Mortars", Cern. Caner. Res. V. 34, pp. 235-248
24. Hanehara, S. and Yamada, K., 1999, "Interaction Between Cement and Chemical
Admixture From the Point of Cement Hydration, Absorption Behaviour of Admixture,
and Paste Rheology", Cern. Caner. Res. V. 29, pp. 1159-1165
25. Chiocchio, G., Paolini, A. E., 1985, "Optimum Time for Adding Superplasticizer to
Portland Cement Paste", Cern. Caner. Res. V. 15, No.5, pp. 901-908.
26. Uchikawa, H., Hanehara, S., Shirasaka, T., Sawaki, D., 1992, "Effect of Admixture
on Hydration of Cement, Adsorptive Behavior of Admixture and Fluidity and Setting of
Fresh Cement Paste", Cern. Caner. Res. V. 22, No.6, pp. 1115-1129
27. Nawa, T., Ichiboji, H., Kinoshita, M., 2000, "Influence of Temperature on Fluidity of
Cement Paste Containing Superplasticizer with Polyethylene Oxide Graft Chains",
Proceedings ofthe 6th CANMET/ACI International Conference on Superplasticizers and
other Chemical Admixtures in Concrete, Ed. V.M. Malhotra, Nice, pp. 195-210.
28. Nagataki, S., Sakai, E., Takeuchi, T., 1984, "The Fluidity of Fly Ash-Cement Paste
with Superplasticizers", Cern. Caner. Res. V. 14, No.5, pp.631-638
30. Bilanda N., Fierens, P., Tirlocq, J., Tenoutasse, N., 1980, "Hydration of Tricalcium
Aluminate Doped by Sodium Oxide", 7th International Congress on the Chemistry of
Cement, V. IV, Paris
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
32. Nawa, T., Eguchi, H., Fukaya, Y. ,1989, "Effect of Alkali Sulphate on the
Rheological Behaviour of Cement Paste Containing a Superplasticizer", Proceedings of
the 3rd CANMET/ACI International Conference, Superplasticizers and Other Chemical
Admixtures in Concrete, Ed. V. M. Malhotra, Ottawa, pp. 405-424.
33. Taylor, H .F. W., 1990, Cement Chemistry, Academic Press, pp. 236-237
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Orthorhombic 1.1 1.7 3.4 3.0 1.7
Aluminate
Lime 1.0 1.2 0.2 0.6 0.7
Periclase 1.6 1.2 1.7 0.3 0.4
Gypsum 0.0 0.2 0.0 1.4 1.3
Hemihydrate 1.8 2.0 2.1 1.5 0.4
Anhydrite 0.6 0.5 0.4 0.4 0.4
Calcite 0.5 4.6 0.6 4.0 0.7
Portlandite 0.3 0.4 0.5 0.3 0.2
Quartz 0.0 0.4 1.5 0.4 0.2
Arcanite 0.3 0.4 0.1 0.0 0.0
Mullite - - 2.6 - -
Amorphous - - 9.3 - -
Blaine 546 360 467 364 447
Fineness
(m2/kg)
K(mg/1) 920 880 580 320 360
Na (mg/1) 220 170 116 74 84
N!!eqv (mg/1) 760 690 458 262 296
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Cement 1 • SNF
--FR
- • -FRO
--cons.
SNF
- • -Cons.
01---....,..=_ _ _ _ _ _-l SNF D
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
0 0.5 1 1.5 2 2.5 3 3.5
Dosage of plasticizer
(% of cement)
- ---·---·-- -······-·--·-·-· --· --·-·--·---···--·-·-·----- ---·-··-····-····-·-----
i
2l
6000
<='>
f
2l
6000
.§ 4000 ~ 4000
.,
·~ ·~
~ 2000 ~ 2000-
0
0
u: u:
0 0.1 0.2 0.3 0.4 0.5
Consumed Plasticizer Consumed Plasticizer (%
cement)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
SNF lS
1.6 1.6
0 R' =0.97 0 R' =0.99
/' V'
~~1.2 ~~ 1.2
·-"' ·-"'
03:
~~0.8 ~~ 0.8
""
oa>
:o::e <=e
~"'
~ ~0.4 ~ u 0.4
en <n
0 0
0 1000 2000 3000 4000 0.0 2.0 4.0 6.0
cC3A'Biaine Na(eqv)'cC3A
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Toyoharu Nawa, Dr. Eng., Professor, Division of Solid Waste, Resources and
Geoenvironmental Engineering, Graduate School of Engineering, Hokkaido
University,Kita 13, Nishi 8, Kita-ku, Sapporo, 060-8628 Japan. His current research
interests are fluidity and shrinkage of concrete incorporated with mineral and organic
admixture and the utilization of recycle aggregate.
INTRODUCTION
High strength concrete has been increasingly used for high-rise reinforced concrete
collective housing, mostly in urban areas. Higher strength generally requires higher
binder content, leading to greater autogenous shrinkage of concrete. For this reason,
simple, effective, and low-cost measures for reducing the shrinkage of high strength
concrete have been anticipated in order to prevent early cracking. It is known that
shrinkage reducing admixture reduce autogenous shrinkage as well as drying
shrinkage[l],[2]. With this as background, the authors have developed one-pack hybrid
admixture (HSP) for concrete in the high and ultra high strength range with a target
strength measured on cylinders of 60 to 150 MPa that is having a water-cement ratio
(W/C) of 0.35 to 0.15, based on the investigation into items (1) to (3) given below with
the aim of developing a new-type admixture combining high fluidity, viscosity-reducing
property, and shrinkage-reducing property.
(1) Development of a water-reducing component having high fluidity,
viscosity-reducing property, and high compatibility with the shrinkage reducing
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
admixture to be adopted
(2) Selection of an effective shrinkage reducing admixture compatible with the
water-reducing component described in (1) above
(3) Investigation into optimum proportions of(2) and (3) above
This paper reports on the outline of the development of HSP and its properties on
concrete presenting a strength in the range of 60 to 150 MPa.
Materials
Cement
Normal portland cement (N), low-heat portland cement (L), and silica fume cement
(SF C) were used. SFC is a premixed cement containing silica fume (SF) in place of 10
wt% of L cement. Their chemical and physical properties are given in Table 1 and Table
2.
Aggregate
Fine aggregates were land sand with a saturated surface dry (SSD) density of 2.58
g/cm3 and fineness modulus of 2.67 (S1) and crushed sand with a SSD density of 2.64
g/cm3 and fineness modulus of 2.67 (S2). Coarse aggregates were crushed stone with a
SSD density of2.68 g/cm 3 and solid volume percentage of60.0% (Gl) and crushed stone
with a SSD density of 2.66 g/cm3 and solid volume percentage of 60.0% (G2).
Procedures
Concrete proportioning
Table 3 gives the mixture proportions of concrete. The type of cement was changed
depending on the target strength while fixing the unit water content constant. N cement, L
cement, and SFC were used for compressive strength of 60, 80, and I 00 MPa or more,
respectively. Aggregate with high strength was used for the target strength of 150 MPa.
Concretes with W/C's of 0.35 and 0.25 were entrained with target air as 3%, whereas
those with W/C's of 0.20 and 0.15 were not air-entrained because previous study showed
that high strength concrete with lower W/C than 0.20 have sufficient freezing and
thawing resistance without entraining air[3].
Mixing concrete
Concretes were mixed using a forced-mixing-type mixer with a capacity of 55 liters
as follows: Dry-mix fine aggregate and cement for 15 sec. Add water and admixture and
mix for 60, 90, and 300 sec for W/C's of 0.35, 0.25 and 0.20, and 0.15, respectively. Add
coarse aggregate and mix for 120 sec. The test temperature was 20 °C. The batch size
was 30 liters. The air content was adjusted by the use of an air-entraining admixture or
defoaming admixture.
Test methods
Slump flow-- Slump flow was measured in accordance with JIS A 1150. The slump flow
was determined as the average diameter of the fresh concrete after spreading by
measuring in two directions at right angles.
L-flow velocity -- L-flow velocity (Lv) was measured by using L-flow meter [3] shown
in Fig. 1. The flow speed of concrete between 5 and I 0 em from the flow start line of the
L-flow meter was defined as L-flow velocity.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Setting time -- Setting time was measured in accordance with JIS A 1147 by means of
penetration resistance measurements on mortar sieved from concrete.
Drying shrinkage test -- Drying shrinkage test was conducted in accordance with JIS A
1129. The specimens for length change molded in a size 100xJOOx400 mm were
demolded after 24 hours, cured in water until the age of 7 days and kept in a room with
temperature of 20 oc and 60%RH for 6 months. The base length of the specimen was
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
measured at the age of 7 days.
Autogenous shrinkage test -- Autogenous shrinkage test was conducted following the
method established by JCI Research Committee for Autogenous Shrinkage[4]. the strain
was measured by the data logger using embedded strain gauges in the middle section of
the form shown in Fig.2, and the changes in the strain after initial setting were regarded
as autogenous shrinkage strain. The specimens molded in a size lOOx100x400 mm were
sealed and kept in a room with temperature of 20 oc until the age of 28 days.
Compressive strength test-- Compressive strengths were tested with 200x100 mrn
cylinder specimens in accordance with JIS A 1108. The specimens were water-cured
immediately after being demolded until compression testing at 7, 28 and 91 days.
Freezing and thawing resistance test -- Freezing and thawing resistance test was
conducted in accordance with ASTM C 666. the specimens molded in a size
1OOx 1OOx400 mm were taken out of the molds at the age of 24 hours and subjected to the
test after 28 days curing in 20 °C water. They were frozen in water at temperature of -18
± 2 oc and thawed at 5 ± 2 oc until 300 cycles.
Accelerated carbonation test -- Accelerated carbonation test were conducted in
accordance with JIS A 1153. The specimens molded in a size 100x100x400 mm were
demolded after 24 hours, cured in water with 20 oc until the age of 28 days and cured in
air with 20 oc and 60%RH until the age of 56 days before they are subjected to the test.
The specimens were exposed to an atmosphere containing 5% C0 2 with 20 °C and
60%RH. The carbonation depth was measured by applying phenolphthalein color
indicator on the fracture surface of freshly broken off fragment of the specimens at 1, 4, 8
13, 26 weeks after starting to be exposed.
Water-reducing component
A polycarboxylic water-reducing component (SSP) configured by polymeric
molecular design having improved water-reducing performance in the high and ultra high
strength range, viscosity-reducing performance, and compatibility with shrinkage
reducing admixtures was used along with a conventional polycarboxylic high-range
Shrinkage-reducing component
A shrinkage reducing admixture presenting high compatibility with high-range
water-reducing admixtures was selected from five types, SRA1 to SRAS, having different
chemistry. Their chemical compositions are given in Table 4. The shrinkage-reducing
performance and compatibility of SRAl to SRA5 are given in Table 6.
The shrinkage-reducing effect was evaluated by conducting tests on concrete made
with W/C of 0.35 as given in Table 3 and calculating the percentage of shrinkage
reduction based on the drying shrinkage strain after 26 weeks with respect to that of
reference concrete containing no shrinkage reducing admixture. The dosage of each
shrinkage reducing admixture was 2 kg/m3 . SSP was used as the water-reducing
component in combination with the shrinkage reducing admixtures in all cases.
The compatibility of these shrinkage reducing admixtures was examined by
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
blending each type in SP or SSP at a ratio of 15, 33, or 50% and leaving to stand for one
month for visual judgment of the state of solution. In this test, SRA 1 to SRA5 were used
with their solid concentrations adjusted to 30 %. Uniform and clear solution was marked
"0," whereas solution with segregation and/or cloud was marked "x" in Table 6.
By overall judgment, SRA3 with a highest shrinkage-reducing effect and high
compatibility with SSP was selected as the shrinkage-reducing component for this study.
Autogenous shrinkage
The autogenous shrinkage strain of concrete containing SP with W/C's of 0.35 (N),
0.25 (L), and 0.20 (SFC) was relatively small, in the range of215 to 315 x 10·6 at an age
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Drving shrinkage
HSP produced a shrinkage-reducing effect of 7.6 to 8.8 % at an age of 26 weeks.
However, its effect of drying shrinkage reduction is weaker than its effect of autogenous
shrinkage reduction. Differences in the shrinkage-reducing effect due to proportioning
differences were also small.
Compressive strength
Fig.8 shows the compression test results. No marked difference was observed
between the strength-development in presence of SP or HSP. Both achieved strengths
equivalent to the target strength by the age of 28 or 91 days, proving that the addition of a
shrinkage reducing admixture produces no adverse effect on concrete strength.
concretes containing SP and HSP were 90 or more after 300 cycles of freezing and
thawing. Both were found to provide sufficient resistance to frost damage with an air
content of around 3% for W/C's of 0.35 and 0.25 and 2% or less for W/C's of 0.20 and
0.15.
Carbonation resistance
Table 11 gives the results of accelerated carbonation test. No carbonation was
observed in concretes containing SP and HSP at the end of an accelerated carbonation test
period of 26 weeks. These were therefore proven to be sufficiently resistant to
carbonation.
CONCLUSIONS
REFERENCES
Referenced standards
ASTM
C666 Test Method for Resistance of Concrete to Rapid Freezing and Thawing
Cited references
2. Berke, N.S. and Li, L., 2003, "Effectiveness of shrinkage reducing admixtures
inreducing total shrinkage," Advances in Cement and Concrete, Proceedings of a
conference held at Copper Mountain, Colorado, USA, pp.l01-109.
3. Mitsui, K.; Yonezawa, T.; Kinoshita, M. and Shimono, T., 1994, "Application of a
New Superplasticizer for Ultra high-Strength Concrete," Superplasticizers and Other
chemical admixtures in concrete: Proceedings of the Fourth CANMETIACI
International Conference, Montreal, Canada, 1994 (SP-148), ed. V.M. Malhotra,
American Concrete Institute, pp.27-46.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
5. Kinoshita, M.; Suzuki, T.; Yonezawa, T. and Mitsui, K., 1994, "Properties of an
Acrylic Graft Copolymer-Based New Superplasticizer for Ultra high-Strength
Concrete," Superplasticizers and Other chemical admixtures in concrete: Proceedings
of the Fourth CANMETIACI International Conference, Montreal, Canada, I994
(SP-148), ed. V.M. Malhotra, American Concrete Institute, pp. 281-299.
6. Kinoshita, M.; Suzuki, T.; Soeda, K. and Nawa, T., 1997, "Properties ofMethacrylic
Water-Soluble Polymer as a Superplasticizer for Ultra high-Strength Concrete,"
Superplasticizers and Other chemical admixtures in concrete: Proceedings of the Fifth
CANMETIACI International Conference, Rome, Italy, 1997 (SP-173), ed. V.M.
Malhotra, American Concrete Institute, pp.143-162.
2.20 14,100 92.2 0.60 0.70 0.27 1.00 1.50 0.49 2.20
70
0.35 550±50 3±0.5 160 457 - - 880 - 868 -
70~
100
0.25 650±50 3±0.5 160 - 640 - 756 - 852 -
too~
120
0.20 650±50 2> 160 - - 800 624 - 852
120~
150
0.15 700±50 2> 160 - - 1067 - 412 - 846
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Cone.
Symbol Type Admixture Composition and Properties
(%)
Methacrylic type polycarboxylate-based
SP Conventional 30
copolymer
New Maleic type polycarboxylate-based copolymer 30
SSP
copolymer
SRAl Low alcohol series 100
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
SRA2 Shrinkage Polyether series 100
SRA3 reducing Polyether derivative 100
SRA4 admixture Low alcohol alkylene oxide adduct 100
SRA5 Glycol ether derivative 100
Mixture with Maleic type
HSP Hybrid polycarboxylate-based copolymer 30
and Shrinkage reducing agent (SSP+SRA3)
Target Target
Kind of Dosage Slump flow L-flow velocity
W/C slump flow air (Cx%)
admixture (mm) (cm/s)
(mm) (%)
SRA1 15.6 X 0 0 0
SRA2 18.2 X 0 0 0
SRA3 19.4 X 0 0 0
SRA4 15.8 X 0 X X
SRA5 14.9 X X X X
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
0.35
SP 215 - 618 -
HSP 165 23.3 571 7.6
0.25
SP 240 - 523 -
HSP 144 40.0 481 8.0
0.20
SP 315 - 521
HSP 248 21.3 478 8.3
0.15
SP 730 - 400 -
HSP 412 43.6 365 8.8
SP 95
0.35
HSP 93
SP 93
0.25
HSP 91
SP 99
0.20
HSP 98
SP 99
0.15
HSP 99
Carbonated thickness
Kind of
W/C (26weeks)
admixture
(mm)
SP 0.0
0.35
HSP 0.0
SP 0.0
0.25
HSP 0.0
SP 0.0
0.20
HSP 0.0
SP 0.0
0.15
HSP 0.0
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
g
...
400mrn
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Fig. 2 --Autogenous shrinkage strain tester.
-
800 . . . . - - - - - - - - - - - - - - - - - , 800
W/C0.25,L
~ ~e::::::~-----=o:.:~~ 700
i
700
600 I 600
A soo ~ 500
W/C O.IS,SFC
g 400 ~------------
J 400
"' 300
300 ~-------------
200 '----'----'----~ 200
30 60 90 0 30 60 90
Elapsed time (min) Elapsed time (min)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
14 . . . . - - - - - - - - - - - - - - - - - , -14
W/C0.35,N W/C 0.20,SFC
6 ---------------------
4 '----~-~----~~ 4
30 60 90 0 30 60 90
Elapsed time (min) Elapsed time (min)
12 . . . . - - - - - - - - - - - - - , 10
W/C 0.25,L W/C O.IS,SFC
.:s 6
4 ~
~------------- 2
2 0
30 60 90 0 30 60 90
Elapsed time (min) Elapsed time (min)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
~
1 I
I : :
I
I
I
200 ritiii~~~~~~~~~~~~
~ ----T----~-----,----- ~ ----T----~----~-----
1 I 1 I
I I I I I I
~ ----T----,----~----- 600 ----T----~----~-----
~ L---~~-----Ll----~:----~
1 I 1
I
~ L--~-~--~-~
7 14 21 28 7 14 21 28
Elapsed time (days) Elapsed time (days)
-200 -200
i W/C025,L
l
I I
I I
: : W/C0.15,SFC
----T----~----~----- ----T----~----~-----
:g I I I :g I I I
! •i
g
~ ·i
0 I I I
200 200 ~--i----i----~-----
I I I
I I
~ ...i
0
400 -~--~----i----- 400 T
~: ----~----~-----
~"
600
~ 600
:
800 ~
14 21 28 14 21 28
Elapsed time (days) Elapsed time (days)
200 r-----------------------------------,
180
;?160
6140
~120
~100 ~==~~~--1
·~ 80
~60
8 40
20
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Carlo Pistolesi graduated in Civil Engineering at the University of Ancona. Since 1992 he
has been working in Mapei, Research and Development Centre in Milan, Italy as chief in
charge of the building materials department of the and now he is working in the
development of repair concrete mortars and admixtures. He is author of some papers in
the concrete field.
Anna Bravo graduated in bioorganic chemistry at the University of Milan. In 1997 she
got a P.h.D. in industrial chemistry at the Politecnico of Milan with a internship at the
N.R.C. in Ottawa (Canada). Since 2000 she has been working in the Central Analytical
Department ofMapei Research and Development Centre in Milan and she author of some
papers in the cementitious field.
Fiorenza Cella graduated in geology at the University of Milan. In 2000 she got a P.h.D.
in applied mineralogy at the University of Milan with a internship at the CTG Italcementi
Group in Guerville (France). Since 2000 she has been working in the Central Analytical
Department ofMapei Research and Development Centre in Milan and she author of some
papers in the cementitious field.
Tiziano Cerulli graduated in organic chemistry at the University of Milan. Since 1989 he
has been working as chief in charge of the Central Analytical Department of Mapei
Research and Development Centre in Milan. He is author of several papers, most of them
in the cementitious field.
INTRODUCTION
Alkali-free accelerators are relatively new liquid concrete admixtures used in the
area of underground construction (1, 2, 3, 4, 5). For safety reasons in the European
market they are replacing the classical alkali-rich accelerators, like sodium aluminate or
sodium silicate. Furthermore, the concrete sprayed with these new alkali-free accelerators
can reach higher long term compressive strengths (6). The accelerating admixtures cause
a fast concrete setting allowing to project the cementitious material onto the tunnel wall.
The concrete coat acts as a consolidating shield useful to prevent the tunnel convergence
(7). Several alkali-free accelerators are available on the market. They can be divided into
two main categories both containing aluminium sulphate complexes stabilized either by
inorganic or by organic acids (8). The capability of concrete to stick on the substrate is
related to the efficacy of the reaction between accelerator and hydrating cement. This
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
EXPERIMENTAL DETAILS
Raw materials
Cement pastes (Tab.l) were prepared by mixing a clinker (from Cementerie Barbetti,
Gubbio, Italy - Tabs.2,3), ~-hemihydrate (from Lages, Italy, J.t: 11 J.tm) or anhydrite
(from Yesos Ibericos, Spain, J.t: 8 J.tm), water (water/solid= 0.45). An inorganic acid
based alkali-free accelerator was used (Tab.4).
Mechanical test
The samples were prepared by mixing by a spatula clinker, calcium sulphate (~
hemihydrate or anhydrite) and water for 1 minute. Afterwards the accelerator was added
and mixed for 10 seconds. Final set time (each result is an average of two measurements)
was carried out according to ENV 196/3 (Vicat needle method- Fig.l,2). The procedure
is explained elsewhere (9). Compressive strength development (each result is an average
of four values - Tab.5) was also determined on 25x25mm (diameter x height) cylindrical
specimens cured at 23°C and 50% R.H .. During the first hours of hydration, as the
samples were too soft (mechanical strength lower than 1 MPa), a digital force gauge
(Imada model DPS 100 equipped with a needle -3 mm section) was used to determine the
compressive strength. At later curing time, when the samples became harder
(compressive strength higher than 1 MPa), a machine type Controls L 1400 was utilised.
Paste samples, mixed and cured as described for the mechanical tests, were
studied as follows.
Hydration temperature profile during the first 12 hours of hydration was measured (Fig.3)
by a digital thermometer (Testo model 781) inserted in a 50 g sample. The cell containing
the specimen was kept in a temperature controlled box (20°C).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
RESULTS
Mechanical test
The results summarized in Fig.l point out that calcium sulphate form influences
the final set of non accelerated paste sample (R,1k<Ranhct<Rheta). The accelerator addition
causes a significant setting time reduction on both the samples without calcium sulphate
(ACc1k- Fig.2- in comparison with Rc 1k- Fig.l) or containing anhydrite (ACanhd- Fig.2-
compared to Ranhd- Fig.l). On the contrary, a strong retardation was observed when~
hemihydrate was used as setting regulator (ACheta - Fig.2 - compared to Rheta- Fig.l ). The
compressive strength results (Tab.5) follow a similar trend (R, 1k>Ranhct>Rheta and
ACc~k>ACanhd>ACanhd) even if, after 4 hours, the strength of the references (Rc1k. Ranhd and
Rheta) becomes higher than the accelerated samples.
Chemical-physical analyses
The hydration temperature profile is related to the rate of heat release ( 14). The
exothermic peak observed during the first two hours of curing (Fig.3 - standard
deviation: +/- 0.5°C) is due to early cement hydration and hydrated sulphoaluminate
formation. Figure 3 shows an increase of the peak height on the samples admixed with
the alkali-free accelerator (ACclk, ACbeta and ACanhd in comparison with R, 1k. Rheta and
Ranhct). This effect arises from aluminium and sulphate ions coming from the accelerator
which promote a massive ettringite formation (15). ACanhct shows the highest peak
intensity.
The XRD patterns (Figs.4-6) point out signals of ettringite crystals in all the
samples. The use of the flash setting accelerator seems to cause an increase of the
ettringite peak intensity. ACanhd is the sample with the highest ettringite peak intensity.
The inorganic acid based alkali-free accelerator could retard cement hydration as the
portlandite content after 6 hours of hydration is significantly lower than the reference
(Figs. 4-6). Nevertheless, after 6 hours, also Rheta has a lower portlandite amount
compared to Rc1k and Ranhct, suggesting that this setting regulator could slow down alite
hydration. Both Ranhd and ACanhd show anhidrite signals also after 6 hours of curing. This
Morphological analyses
The results in Tab.7 underline that ACcik has the largest surface area (after 5 minutes of
curing). Furthermore the specific surface area of ACanhd is larger than ACbeta· Therefore
some morphological differences among the accelerated samples should be evident.
The ESEM micrographs of the reference pastes Rc1k. Rheta and Ranhd show significant
differences: the sample without calcium sulphate points out well shaped plates of likely
AFm phase and few crystals of ettringite (Fig. 7); on the contrary the sample containing ~
hemihydrate is characterised by small and rounded masses (Fig.8) on a gel like material
which covers cement grains; whereas few small but well shaped prismatic needles
appears in Ranhd (Fig.9). From the literature (16) it is known that calcium sulphate based
setting regulators promote the formation of an amorphous layer of AFt phase on
hydrating cement grains which slows down further hydration (dormant period). This
protective coat cannot be formed in Rc1k (as it does not contain the setting regulator) and
the low amount of available soluble sulphate ions (for example coming from the fast
dissolution of alkaline sulphate) could give origin to well shaped crystals of ettringite
(Fig.7). Moreover, Fig. 8 supports the hypothesis that a fast nucleation rate occurs in Rheta
(for the higher instantaneous dissolution rate of ~-hemihydrate) and a larger amount of
gel like material could be formed. In Ranhd, for the lower instantaneous dissolution rate of
anhydrite, more ordered and larger ettringite crystals could be obtained. As time
proceeds, the morphology of the samples becomes more similar.
In Figs. 10 and 11 (ESEM micrographs of ACc1k and ACanhd after the first minutes of
hydration) thin and long rods of AFt phase are visible. On the contrary, the
microstructure of ACbeta is deeply different (Fig.l2): short and disordered masses (whose
chemical nature is not clear) are present everywhere. These differences support the
hypothesis about the existence of many nucleation germs in ACbeta, which could hinder a
proper growth of well formed ettringite crystals.
After 6 hours, Figures 13 and 14 (ACc1k and ACanhd) show a significant growth of
ettringite crystals (which appear slightly longer and thinner in ACc~k). At the same curing
time, short prismatic needles of likely ettringite and plates of AFm phase are evident in
ACbeta (Fig.15).
DISCUSSION
According to the literature (17, 18, 19), alkali-free accelerator reacts with
hydrating cement promoting the formation of hydrated sulphoaluminates (particularly
AFt phase) thus determining a rapid viscosity increase of cement pastes which results in a
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
CONCLUSION
This study points out that the initial sulphate ions content could influence the
efficiency of an inorganic acid based alkali-free accelerator. It was found that the lower is
the calcium sulphate instantaneous dissolution rate the more efficient is the accelerator.
This behaviour could be related to the different morphology of the ettringite formed
during the first instant of the reaction between hydrating cement and alkali-free
accelerator. Such a morphology depends on the different nucleation/growth ratio of the
ettringite crystals: anhydrite favours crystal growth whereas J3-hemihydrate promote the
nucleation. For shotcrete applications, it is recommendable to use cement containing
anhydrite as setting regulator or, at least, to avoid cement with ~-hemihydrate.
REFERENCES
hydroxide in calcium silicate and cement pastes", Cern. Con. Res., No. 9, pp. 677-
684·
14) Taylor, H. F. W., 1998, Cement Chemistry, 2"ct edition, p.212 (Publisher: Thomas
Telford);
15) Bravo, A., Cerulli, T., Maltese, C., Pistolesi, C., Salvioni, D., 2003, "Effects of
increasing dosages of an alkali-free accelerator on the physical chemical properties
of a hydrating cement paste", Proceedings of th CANMET/ACI, Berlin, Germany,
2003, ed. V.M. Malhotra, pp. 211-225;
16) Scrivener, K.L., Pratt, P.L., 1984, "Microstructural studies of the hydration of C3A
and C4AF independently and in cement paste", Br. Ceram. Proc., No. 35, pp. 207-
219;
17) Paglia, C., Wombacher, F., Bohni, H., Sommer, M., 2002, "An evaluation of the
sulphate resistance of cementitious material accelerated with alkali-free and alkaline
admixtures: laboratory vs field", Cern. Con. Res., No. 32, pp. 665-671;
18) Burge, T.A., 2001, "Mode of action of alkali-free sprayed shotcrete accelerators"
Shotcrete: Engineering Developments, ed. Bernard, pp. 79-85;
19) Cerulli, T., Pistolesi, C., Maltese, C., Salvioni, D., 2002, "Alkali rich and alkali free
accelerators for shotcrete: physical, chemical and mechanical effects on cement
hydration", Proceedings of ICMA 2002, San Diego, California, pp. 1-17,
www.cemmicro.org;
20) Schwiete, H.E., Ludwig, U., Jager, P., 1966, "Investigation in the system 3CaO-
CaS04-CaO-H20", Proceedings of Highway Research Board, Symposium on
Structure ofPortland Cement Paste and Concrete, Washington, 1966, pp. 353-367;
21) Odler, I., 2000, Special inorganic cements, ed. E&F.N. Spon, p. 205;
22) Locher, F.W., 1981, "Hydration of pure Portland cement" The proceedings of th
!CCC, Paris, 1981, Vol.4, pp. 49-54- (provided by: www.iccc2007.org).
Accelerator 6 6 6
Water 45 45 45 45 45 45
Blaine
Na20
Density specific
CaO(%) Si02 (%) Ah03 (%)Fe203 (%) MgO (%) S03(%) Cl(%) Equivalent (kg/mJ)
surface
(%) 2
area (m /g}
66.45 19.74 6.23 2.52 1.36 1.04 0.1 0.5 3150 331
c~ CaO Ca(OH)2
69.2 8.7 17.1 2.2 0.5 0.2 1.5
(+/- 0.2) (+/- 0.7) (+/- 0.2) (+/- 0.1) (+/- 0.05) (+/- 0.05) (+/- 0.1)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
.1:::-
\0
0\
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
3:
S»
~
(D
"'
Table 5 -Compressive strength development (MPa)
S»
:-
2 hours 0.02 (+/-0.005) <0.01 <0.01 0.092 (+/-0.01) <0.01 0.04 (+/-0.0 I)
3 hours 0.06 (+/-O.oJ) <0.01 <0.01 0.11 (+/-0,02) <0.01 0.04 (+/-0.01)
4 hours 0.08 (+/-O.oJ) <0.01 0.01 (+/-0.005) 0.12 (+/-0.02) <0.01 0.04 ( +/-0.0 I)
6 hours 0.2 (+/-0.02) 0.08 (+/-0.01) 0.14 (+/-0.02) 0.14 (+/-0.02) 0.01 (+/-0.005) 0.05 (+/-0.01)
8 hours 0.4 (+/-0.05) 0.13 (+/-0.02) 0.37 (+/-0.05) 0.16 (+/-0.02) O.Ql (+/-0.005) 0.14 (+/-0.02)
Not for Resale, 08/10/2015 01:25:47 MDT
Licensee=UNI OF NEW SOUTH WALES/9996758001
24 hours 6.3 (+/-1.0) 7.8 (+/-1.0) 7.1 (+/-1.0) 6.4 (+/-1.0) 8.8 (+/-1.0) 6.1 (+/-1.0)
Daneshlink.com
No reproduction or networking permitted without license from IHS
Provided by IHS under license with ACI
Copyright American Concrete Institute
(/)
c
"'C
(D
-"'
'"'I
"'C
-
cu
;::;·
No
(D
'"'I
"'::s
cu
Table 6 - Thennogravimetric analyses
c.
n
::r
R.tk ~ R..bd AC,Ik AC- ACanhd (D
1.5 3
5 min 1.0 0.3 0.9 0.7 0.8
;::;·
Calcium Hydroxide(%)
6 hours
5 min
4.9
1.8
2.2
2.2
2.8
1.3
1.6
7.2
1.0
5.1
1.7
8.8
-
cu
)>
c.
Water Loss(%; 80°- 300°C) 3
~
6 hours 4.1 3.3 5.4 9.0 7.7 9.3
c
ta
Not for Resale, 08/10/2015 01:25:47 MDT
Licensee=UNI OF NEW SOUTH WALES/9996758001
"'::s-·
n
0
::s
~
-
n
ta
(D
Daneshlink.com
\0
.......
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
498 Maltese et at.
Table 7 - Specific surface area
500 +---------
400 +---------
300 +---------
200
100
40~-------
35+--------
30+---------
25+--------
t
ii
20+---------
15+---------
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
10
31 .:_-
29 ------- -
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
27
25
23
19 -~---~---r--~--
0 a o 12 0 12
time (hours)
Countsls
60
40
20
5 mino
100
50
400
portlandite
m
200
100
6 hrs o
100
50
10 15 20 25
Position["2Theta[
"'
150
100
"'
100
100
10 15 20 25
Position ['2Theta)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
ettrin~ite ,, anhydrite
! ------------,
100 ' ',
' -,
: '
100
100
10
Position["2Theta[
Fig. 7- ESEM micrograph of Rcik after 5 min of hydration (unit length: t1Jm).
Fig. 8- ESEM micrograph of Rbeta after 5 min of hydration (unit length: t!Jm).
Fig. 9- ESEM micrograph of Ranhd after 5 min of hydration (unit length: t!Jm).
Fig. 10- ESEM micrograph of ACctk after 5 minutes of hydration (unit length: 21Jm).
Fig. 11- ESEM micrograph of ACanhd after 5 minutes of hydration (unit length: 21Jm).
Fig. 12- ESEM micrograph of ACbeta after 5 minutes of hydration (unit length: 211m).
Fig. 13- ESEM micrograph of ACcik after 6 hours of hydration (unit length: 211m).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Fig. 14- ESEM micrograph of ACanhd after 6 hours of hydration (unit length: 211m).
Fig. 15- ESEM micrograph of ACbeta after 6 hours of hydration (unit length: 211m).
INTRODUCTION
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
MATERIALS
METHODS
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
the first step, a typical model 16 C9 unit segment was designed for each type of
lignosulfonate (SL, KL and OL). It complied with the data on elemental composition,
functional groups content, inter-unit linkages and other molecular characteristics
collected from the literature. In the second step, model molecular structures representing
the individual lignosulfonate fractions analyzed in this study were designed. These were
based on the results of experimental analysis acquired in this work, the model 16 C9 unit
segments from step one and additional experimental data obtained from the literature. As
a result, a complete molecular structure including all functional groups and linkages was
proposed for each lignosulfonate fraction. The detailed procedures and algorithms for
designing model molecular structures are available elsewhere [20].
Correlation Analysis
RESULTS
Three series of lignosulfonate fractions: SLl to SL4, KLl to KL4 and OLl to
OL4 were tested in fresh cement pastes. The results for dispersing (both plasticizing and
fluidifying), set-retarding and air-entraining effects plotted against the weight-average
molecular weight Mw are depicted in Figure 2.
The lignosulfonate fractions were analyzed for molecular weight and chemical
composition. The results are summarized in Table 2. Model molecular structures
representing individual lignosulfonate fractions were designed. Their molecular
parameters matched the analytical data presented in Table 2 as well as the literature data
on functional groups and structural linkages with an accuracy of ±5%. An example of a
proposed model molecular structure for a lignosulfonate fraction (SL series) composed of
fifty-two C9 units is shown in Figure 1. The relationships between selected functional
group counts per 100 C9 units and molecular weight of lignosulfonates (SL, KL and OL)
are depicted in Figure 3 and Figure 4. The complete sets of proposed molecular structures
representing the lignosulfonate fractions and the molecular properties of these structures
are available elsewhere [20].
Correlation Analysis
DISCUSSION
Dispersing Effect
A different scenario than in the MS-test is in place during the T-test. At the time
of lignosulfonate addition, 20 minutes after the initial contact of cement with water,
cement hydration is well under way [20]. As a result, lignosulfonate molecules can
adsorb both on the amorphous C-S-H gel coating covering the cement grains and on
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
ettringite, located both at the surface of cement grains and in the pore solution. These
interactions can not only inhibit and/or alter the ettringite crystal growth, but they also
can affect the dispersability of the whole system. The results are shown in Table 4. The
high positive correlation to methoxyls and C-C bonds indicates that the hydrophobicity of
lignosulfonates molecules is beneficial to the fluidifying effect. In other words, the
"repulsion" between hydrophobic lignosulfonate surfaces and the hydrophilic cement
grains seems, not surprisingly, to be a contributing factor to the dispersion stability.
Similar to the plasticizing effect, the negative correlation to the hydrophilic group counts
indicates that extreme hydrophilicity can be detrimental.
Presumably, the correlation between functional group counts and the overall
dispersing effect presented in Table 5 may provide an improved less test-dependent
insight into their relationship. Here, the methoxyl content is the most positively
correlated, probably due to its contribution to the hydrophobic and steric properties of the
molecule. The high correlation of inter-unit linkages (C-C and C-0-C) indicates the
positive contribution of molecular weight to the overall dispersing effect reported
previously [8,9]. The aliphatic carboxyl content is highly negatively correlated, which is
consistent with its contribution to an excessive negative charge on lignosulfonate
Set-Retarding Effect
increase in carbon-carbon inter-unit bond and the methoxyl group count. This indicates
that retardation may be influenced by the formation of a hydrophobic membrane around
the hydrating cement particles, which slows down the dissolution of cement minerals and
the transfer of hydrated ions in the pore solution. A similar hydrophobic layer on the
surface of cement particles caused by the adsorption of surfactants may reportedly
increase the air stability in cement-water systems [26]. In agreement with the above
concept, polar lignosulfonate functionalities (pKa<12) are instrumental in reducing the set
retardation as they are negatively correlated to the retardation effect.
Air-Entraining Effect
CONCLUDING REMARKS
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
further research is needed in order to explore these interactions in a greater detail.
ACKNOWLEDGEMENTS
This research was funded in part from various resources of the University of
New Brunswick, Canada. Financial support during laboratory experiments granted by Dr.
Bremner is particularly acknowledged. The author also wishes to thank Dr. van
Heiningen, Dr. Ni, Dr. Deslongchamps and Dr. Turner for providing an access to the
analytical equipment and molecular modeling software, and for valuable discussions.
REFERENCES
1. ASTM C 494: Specification for Chemical Admixtures for Concrete, Annual Book of
ASTM Standards, Vol. 04.02: Concrete and Aggregates, ASTM, Philadelphia, PA,
2000.
2. Mark, J. G., Concrete and Hydraulic Cement, US Patent No. 2,141,570, 1938.
10. Zhor, J., Bremner, T. W. and Lora, J. H., Effect of Chemical Characteristics of
Alcell® Lignin-Based Methyl-Sulfonates on Their Performance as Water-Reducing
Admixtures, In Proceedings, 4th International Conference on Superplasticizers and
Other Chemical Admixtures in Concrete, Montreal, Canada, Edited by V. M.
Malhotra, ACI SP-148, ACI, Detroit, MI, pp. 333-351, 1994.
11. Ferrari, G., Basile, F., Dal Bo, A. and Mantoni, A., The Influence of the Molecular
Weight of Beta-Naphthalene Sulfonate Based Polymers on the Rheological
Properties of Cement Mixes, Il Cemento, Vol. 83, No.4, pp. 445-454, 1986.
13. Yamada, K., Takahashi, T., Hanehara, S. and Matsuhisa, M., Effects of the Chemical
Structure on the Properties of Polycarboxylate-Type Superplasticizer, Cement and
Concrete Research, Vol. 30, No.2, pp. 197-207,2000.
16. Perche, F., Houst, Y. F., Bowen, P. and Hofmann, H., Adsorption ofLignosulfonates
and Polycarboxylates - Depletion and Electroacoustic Methods, In Proceedings, 7th
1996.
18. Kauppi, A., Banfill, P. F. G., Bowen, P., Galmiche, L., Houst, Y. F., Lafuma, F.,
Mader, U., Perche, F., Petersen, B.G., Reknes, K., Schober, 1., Siebold, A. and Swift,
D., Improved Superplasticizers for High Performance Concrete, In Proceedings, 11th
International Congress on the Chemistry of Cement, Durban, South Africa, Edited by
G. Grieve and G. Owens, pp. 528-537, 2003.
22. Adler, E., Lignin Chemistry - Past, Present and Future, Wood Science and
Technology, Vol. 11,No. 2,pp.169-218, 1977.
23. Dshalalow, J. H., Real Analysis: An Intoduction to the Theory of Real Functions and
Integration, Chapman&Hall/CRC, Boca Raton, FL, 2001.
24. Young, J. F., Reaction Mechanisms of Organic Admixtures with Hydrating Cement
Compounds, In Concrete Admixtures, Transportation Research Record 564, pp. 1-9,
1976.
25. Chandra, S., Flodin, P., Interactions of Polymers and Organic Admixtures on
Portland Cement Hydration, Cement and Concrete Research, Vol. 17, No.6, pp. 875-
890, 1987.
27. Kobayashi, A., Tadaaki, H., Yamakawa, 1., Sato, K., On the Dispersing and Air
Entraining Effect of Lignosulfonate, Mokuzai Gakkaishi, Vol. 13, No. 3, pp. 118-
122, 1967.
0
II
Sulfonate R-S03H R-S-OH -4.5 - -2.0 10-70
II
0
Aliphatic
R-COOH R-C
1P 1.5 - 5.0 0-50
Carboxyl
'OH
Aromatic
Ar-COOH 2.5 - 4.5 0-50
Carboxyl o-t:H
Phenolic
Ar-OH Q-oH 8.0 - 12.0 10-80
Hydroxyl
H
Primary I
R-CH20H R-C-OH 12.0 - 16.0 20-70
Hydroxyl I
H
R'
Secondary I
R2CHOH R-C-OH 12.0 - 16.0 10-40
Hydroxyl I
H
H
Methoxyl Ar-OCH3 Q-o-b-H >20 60- 110
I
H
e.g.
Ar-0-R R
Inter-Unit
R-0-R' >20 10-80
Ether Bond o-o-t-H
Ar-0-Ar' I
R'
e.g.
Ar-R R
Inter-Unit
R-R' o-b-H N/A 10-80
C-CBond
Ar-Ar' I
R'
Notes: 1. C900 stands for 100 C9 umts. 2. Other carbonyl-contammg functionalities
.. ..
such as ketones, aldehydes and esters, as well as ethylenic structures are also
present in lignosulfonate molecules.
Sample
Code
Mw{kDa) c ("lo) H(%) Na(%) Ca(%) 8(%) 0(%) OCH3(%)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
3 OCH3Group 0.581
Positive 4 OHa~.totGroup 0.576
5 C-C Inter-Unit Bond 0.441
6 OHsecGroup 0.317
I All pKa<12 Group -0.697
2 COOHa~Group -0.619
3 OHphGroup -0.569
Negative 4 COOH101 Group -0.546
5 All pKa<5 Group -0.528
6 S0 3HGroup -0.367
1 OCH3Group 0.533
Positive 2 C-C Inter-Unit Bond 0.467
3 Oet Inter-Unit Bond 0.320
1 All p.K,<5 Group -0.650
2 All p.K,<12 Group -0.451
Negative 3 COOHtotGroup -0.409
4 COOHa1Group -0.369
5 S03HGroup -0.292
Correlation Coefficient
Correlation Rank Functional Group I Bond Count
RpL ~ Ro1s
1 OCH3Group 0.581 0.533 0.557
2 0., Inter-Unit Bond 0.719 0.320 0.520
Positive 3 C-C Inter-Unit Bond 0.441 0.467 0.454
4 OHpnnGroup 0.679 0.154 0.416
5 OHaJ.totGroup 0.576 0.068 0.322
1 All pK0 <5 Group -0.528 -0.650 -0.589
2 All p.K,<ll Group -0.697 - 0.451 -0.574
3 COOHa1Group -0.619 -0.369 -0.494
Negative
4 COOHtot Group -0.546 -0.409 -0.478
5 S03HGroup -0.367 -0.292 -0.329
6 OHphGroup -0.569 -0.068 - 0.318
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
6
5
¢ 4
,......o
f3
2
3
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Figure 1--Basic C9 phenylpropane lignin unit with common notation for carbon atoms
and an example of lignosulfonate model molecular structure (fraction SL4).
rs
<
:52
.!o.28
g
~0.25
1'-
0.22
2.2 4.4 6.6 6.6 11.0 0.0 2.2 4.4 6.6 6.8 11.0
Mw,kDa Mw,kDa
15.0 25
I!!
j!12.6 - 20
.0:: ~
g-10.2 - ~ 15
ii
1 7.8 i
0 10
& ~
i 5.4 -
3.0
0.0 2.2 4.4 6.6 8.6 11.0 0.0 2.2 4.4 6.6 6.6 11.0
Mw. kDa Mw. kDa
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
-~
IIKL J125
~ '
---,I
'
8 40 § 75
e"" 20
~
e"'"" so
" o'--~-'-~--'-~~~-.L-~_j "
0.0 2.2 4.4 6.6 8.8 11.0 2.2 4.4 6.6 8.8 11.0
Mw. kDa Mw.kOa
100 .--~=o---r---r----r==
§ 80
u
z. 60
§
8 40
i
I
I I
20 - - - - r - - - -,- - - - -,- - - - -,-
0.0 2.2 4.4 6.8 8.8 11.0 0.0 2.2 4.4 6.6 8.8 11.0
Mw. kDa Mw. kDa
Figure 3 --Relationship between various fully ionizable (pK0 <12) functional group
counts per 100 C9 units and molecular weight of lignosulfonate samples
(SL, KL and OL).
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
J~ ----r----r----r----~-----
" 40 -{}-
~60 !30
0 0
8 40 8 20
"e"" 20
0
e""
0
10
0.0 2.2 4.4 6.6 8.8 11.0 0.0 2.2 4.4 6.6 8.8 11.0
Mw, kDa M.., kDa
120 r---:r==-,----,----,------,
~~ I I I I
" 90
K8o
§
- - - -:- - -: ~- - 8
~
70
I 1 ~Kl
r 1 - - - -~- - - - -~ 2 60
0
20 L_~~-~~~-L-~~:=o=o~L~
0.0 2.2 4.4 6.6 8.8 11.0 0.0 2.2 4.4 8.8 8.8 11.0
M.., kDa M,,kDa
100 ,---:r:=:--,------,,.------,------,
o.... 100 r-------:~§J=,.:--T,----.,---r:,j=O=SL=I
j
!
~
80
60 f: ~::t : ::::
t=-~~
40
0
e"" 20 (:t Q I : : I
0.0 2.2 4.4 6.6 8.8 11.0 0.0 4.4 6.6 8.8 11.0
Mw. kDa M..,kDa
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Synopsis: A combination of ground granulated blast furnace slag, fly ash and
calcium hydroxide has been used to immobilize acid nuclear wastes in the form of a
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
hardened system in the absence of any aggregate. An aqueous solution of NaOH (19
M) was used for two purposes: first, to neutralize the acid waste and to transform it in
a very basic system (pH>13); second, to activate the hardening ofthe binder.
Due to the liquid-solid ratio lower than 0.70, the fresh mixture is relatively stiff.
Superplasticizers based on sulfonated naphthalene and melamine or acrylic
polycarboxylate do not work to fluidify this system. D-Sorbitol as well as other alditols
are very effective in terms of higher workability and lower workability loss of this
cementitious system. Surprisingly this chemical admixture increases the compressive
strength of this cementitious grout at both early and longer ages at a given water-
binder ratio.
The properties of this cementitious system in the absence and in the presence of
D-sorbitol as chemical admixture (0.2-o.s% by weight of binder) have been studied
in the form of workability of the fresh mixture, early and later compressive strength,
leachability by water and X-ray diffraction analysis of the hydrated products of the
hardened system.
INTRODUCTION
waste should be mixed with a Portland cement-free formulation based on a blend of blast
furnace slag, fly ash and hydrated lime as calcium hydroxide.
The role ofNaOH, used to neutralize the acidic waste, was also that of an activator of
the binding system in order to accelerate the hardening process. On the other hand, the
absence of aggregates in the cementitious system can exclude any form of alkali-
aggregate reaction due to the presence ofNaOH.
The present work is specially devoted to study the influence of D-sorbitol adopted as
chemical admixture to improve the workability of the fresh mixtures. This property,
which is absolutely needed because the liquid waste and the binding system are mixed
Table 5 shows the workability of the fresh grouts as a function of the waste-binder
ratio. In the absence of any drop table, the workability should be as high as possible (at
least 100%). On the other hand, the amount of liquid should be relatively low in order to
reduce the radiological load of the final waste form. The compromise was to choose the
waste-binder ratio of 0.60 (corresponding to a water-binder ratio of 0.39) and to improve
its workability by using a fluidifying chemical admixture. The use of usual
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Table 6 shows the workability of the fresh grouts with addition of D-sorbitol and its
loss during the first hour after mixing. According to the nuclear expertise, the fresh
mixtures should be workable (flow>100%) without any compaction effort at least during
1 h. The results indicate that 0.4-0.5% of a 70% D-sorbitol aqueous solution with respect
to the binder system is good enough for the workability requirements.
Figure I shows the compressive strength of the hardened grouts as a function of the
curing time. Surprisingly the addition of D-sorbitol is not associated with any retarding
effect both at early and at later ages.
Table 7 shows the Leachability Index (LI) values calculated according to the long-
term test procedure (90 days). All the binder systems satisfy the condition that ensures a
good containment of soluble salts (LI>7 for sodium and LI>8 for calcium). Moreover, the
presence of D-sorbitol and the waste-binder ratio do not substantially affect this property
required for wasteform durability.
Results obtained by XRD analysis (Figure 2) show that the binding formulation is
mainly constituted of the amorphous phase (broad hump centred at 28=30°) and some
crystalline phases (quartz, mullite, portlandite). After 28 days of curing new phases are
detected in the hardened grouts; these were recognized as calcite CaC03, hydrotalcite
Mg 6Ah(C0 3)(0H) 16 ·4H20 and calcium silicate hydrate C-S-H.
CONCLUSIONS
In the absence of any mechanical compaction the workability of the fresh grouts
represented the most important parameter to be considered. Due to the presence of a
considerable amount of aluminum in the nuclear waste the workability of the waste-
binder system was very low. Only in the presence of D-sorbitol the fluidity reached an
optimum value, even at water-binder ratio as low as 0.39. Moreover, in the presence of a
70% D-sorbitol aqueous solution (0.4-0.5% by binder weight) the compressive strength
of the hardened system was increased indicating a better durability of the wastefonn.
The results of the research are summarized as follows:
a) The workability increases at very low dosages of a 70% D-sorbitol aqueous solution
(0.2-0.5%)
b) The workability loss satisfy the requirement of a flow> 100 after one hour from the
initial mixing
c) The compressive strength is substantially unchanged at early ages and increases at
later ages
d) The Leachability Index values ofNa+ and ci+ are not affected by the presence ofD-
sorbitol
e) The compounds that form during hydration of grout specimens are not known having
any detrimental effect on wasteform durability.
[1] F.P. Glasser, Fundamental aspects of cement solidification and stabilisation, Journal
ofHazardous Materials 52 (1997) 151-170.
[2] M. Collepardi, "Use of fly ash, silica fume, slag and natural pozzolans in Europe:
future directions", International Workshop on the Use of Fly Ash, Silica Fume and
other Siliceous Materials", Sydney, 4-6 July (1988).
[3] P.G. Smith, H.R. Watling, P. Crew, The effects of model organic compounds on
gibbsite crystallization from alkaline aluminate solutions: polyols, Colloids
Surfaces A: Physicochem. Eng. Aspects, 111 (1996) 119-130.
[4] UNI EN 1015-3:2000 Determination of consistency of fresh mortars (by flow table).
[5] ANSI/ANS-16.1-2003 Measurements of the leachability of solidified low-level
radioactive wastes by short-term test procedure.
Compound mol/L
H"No3 0.79
~hS04 0.02
k-\I(N03 )]"9H20 1.13
Fe(N03)J"9H20 O.Ql
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Waste-binder Water-binder
Admixture Workability(%)
ratio ratio
0.50 0.32 30
no 0.60 0.39 90
0.70 0.46 120
40.0
.. 35.0
a.
:::;;
;; 30.0
"'c:~ 25.0
ti5
.~ 20.0
~c. 15.0
E
8 10.0
5.0
0.0
7 26
Time of curing (days)
""" l PC A 0
p p p
q
FA m q m,.ol m m m I m I
~s
p
-
sllgglasa
p
di)'WndlllgW.IHI
m P m p p
m p
c.s...
-
21da)os"llydmfdlllend
h rnA
10 15 20 2!1 30 35 40 45 50 55 60
"28
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Effectiveness of Superplasticizers
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Incorporating Shrinkage-Reducing
Admixture in Recycled-Aggregate
Concrete
INTRODUCTION
Previous results obtained by the authors (Corinaldesi and Moriconi 2001; Corinaldesi et
a!. 2001; Corinaldesi and Moriconi 2002; Corinaldesi and Moriconi 2003a; Corinaldesi
and Moriconi 2003b) as well as those reported in the literature (Hansen 1992; Kasai
1998; Dhir et a!. 1998) showed how the use of aggregates coming from recycling process
allows the production of structural concrete, at least concerning their mechanical
performance. One ofthe critical points concerning the employment of recycled aggregate
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
is its higher tendency to shrink compared to ordinary aggregate concrete. This is
especially the case when both coarse and fine recycled aggregate are used (Corinaldesi et
a!. 2001; Hansen and B0egh 1985). This behavior could increase cracking vulnerability
of recycled-aggregate concrete even though tensile stresses induced by shrinkage strain
are lower, given the lower elastic modulus of the concrete (Corinaldesi et a!. 2001;
Hansen and B0egh 1985).
In this work an attempt was made to manufacture recycled-aggregate concrete that is less
sensitive to shrinkage which uses only coarse recycled aggregate, water-soaking the
recycled aggregate before its introduction in concrete, thus creating an internal water
reserve, and also introducing water-reducing admixture.
The mechanism through which shrinkage-reducing admixtures are able to reduce plastic,
autogeneous and especially drying shrinkage of concrete seems to be the lowering of
surface tension of the water in capillary pores (Rongbing and Jian 2005; Folliard and
Berke 1997; Bentz 2001). In fact, as water-filled pore begin to loose moisture, curved
menisci are formed, and the surface tension of water pulls the walls of the pores. With the
reduced surface tension of water, the force pulling the walls of the pores is probably
decreased, and the resultant shrinkage strain is reduced in tum.
Natural sand (with a maximum size of 5 mm) was used in each concrete mixture. Its main
physical properties are reported in Table 1, and its grain-size distribution is plotted in Fig.
1.
The kind of coarse aggregate was changed: either an ordinary natural gravel or a
recycled-concrete aggregate from a suitable treatment of precast-concrete scraps or a
recycled aggregate from a crushing plant in which rubble from building demolition is
suitably selected, ground, cleaned and sieved. Their main physical properties of the
aggregates are reported in Table I, and their grain-size distributions are plotted in Fig. I.
As shown in Fig. I, the maximum-particle size of the coarse aggregates was 20 mm; in
particular, it can be noticed that the grain-size distribution curves of 'Natural gravel' and
'Recycled rubble' are quite the same. While the 'Recycled concrete' appears slightly
finer.
In terms of water absorption, strong differences can be detected (Table I). It should be
underlined that both recycled-aggregate fractions were added to the mixture after water-
soaking, in a condition very close to that defined as saturated surface-dried. This was
done to create an internal water supply able to reduce drying-shrinkage as well as to
avoid water absorption of aggregate during mixing and, consequently, to maintain
concrete workability for enough time to be cast.
Seven different concrete mixtures were prepared with the same water-to-cement ratio of
0.45 and the same slump of 170 mm, to guarantee sufficient workability, by varying the
kinds of coarse aggregate and chemical admixture.
In Table 2, the mixtures proportions of the three concretes prepared using natural
aggregate are reported. The first mixture, indicated as 'Nat+SP', was prepared with the
pure water-reducing admixture added at a dosage of 0.8%, in terms of total mass of
admixture by mass of cement, which guaranteed a water reduction of 25%. The same
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
water reduction was achieved using superplasticizer with PC-type SRA dosed at 1.6%, by
mass of cement. A slightly lower water reduction of 20% was obtained by the
superplasticizer with RM-type SRA dosed at 2.0%, by mass of cement. Natural sand and
gravel were combined at a percentage of 40% and 60% by volume respectively in order
to achieve optimum grain size distribution of the whole inert fraction according to the
Bolomey particle size distribution curve (Collepardi 1991).
In Table 3, the mixture proportioning of the four evaluated concrete mixtures prepared
using recycled aggregate are reported. The mixtures indicated as 'RecCon+SP' and
'RecRub+SP' were prepared with the pure water-reducing admixture at the same dosage
of 0.8%, by mass of cement.
The 'RecCon+SP' concrete was prepared using the recycled-concrete aggregate coming
from precasting plant as coarse aggregate; this aggregate fraction was combined at a
volume portion of 70% with 30% natural sand to enhance grain-size distribution
according to the Bolomey particle-size distribution curve. The same kind of mixture was
prepared with superplasticizer and the PC-type SRA at a dosage of 1.6%, by mass of
cement: in this case the concrete mixture is indicated as 'RecCon+PC-SRA'.
The 'RecRub+SP' concrete was prepared by using the recycled-rubble aggregate coming
from rubble recycling plant as coarse aggregate besides to natural gravel. Natural sand,
natural gravel, and recycled-rubble fractions were combined at 40%, 30%, and 30%,
respectively, by volume, for optimizing the grain-size distribution of the whole inert
fraction according to the Bolomey particle-size distribution curve. The same kind of
mixture was prepared using superplasticizer with RM-type SRA at a dosage of 2.0%, by
mass of cement: in this case the concrete mixture is indicated as 'RecRub+RM-SRA'.
Twelve cubic specimens (100-mm side) were cast for each concrete mixture, then wet-
cured at 10°C (natural ambient conditions) in order to evaluate compressive strength
Three other prismatic specimens (1 00 x 100 x 500 mm) were cast for each concrete
mixture, wet cured for 2 days at l5°C (natural ambient conditions), then cured at the
same temperature and relative humidity of 65% (natural ambient conditions) up to 70
days in order to evaluate concrete drying shrinkage. In spite of the different laboratory
temperature and relative humidity, shrinkage testing was carried out according to the
procedure described in the Italian Standard UNI 6555-73 "Hydraulic Shrinkage
Determination".
Compressive strength
The results obtained for compressive strength tests carried out on wet-cured specimens
after 1, 3, 7, and 28 days are reported in Fig. 2. By comparing the 28-day compressive
strength of the three concretes prepared with the pure superplasticizer 'SP' it can be
noticed that the kind of aggregate used does not influence concrete strength. When the
other two chemical admixtures were used, slightly higher compressive strengths were
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
obtained for the PC type and slightly lower values for the RM type, although the
maximum difference detected was within 7 MPa.
The effect of curing condition, wet (100% R.H.) or humid air (80% R.H.), on
compressive strength of concrete containing SRA can be evaluated on the basis of results
reported in Table 4.
Drying-shrinkage
Free drying-shrinkage tests were carried out up to 70 days of curing at l5°C and relative
humidity of 65%. Results obtained are reported in Fig. 3. By comparing drying-shrinkage
of natural-aggregate concretes the effectiveness of the two superplasticizers incorporating
a shrinkage-reducing group can be evaluated: after 70 days concrete shrinkage was
decreased by 24% and 62% in presence ofRM-type SRA and PC-type SRA, respectively.
At the same curing time, Rongbing et al. (Rongbing and Jian 2005) found a 30%-40%
drying shrinkage reduction by means of SRA dosed at 2%, by mass of cement.
Concerning concrete prepared using either the recycled-concrete aggregate fraction or the
recycled-rubble fraction, the drying shrinkage was very low, less than 180 1-1rnlm after 70
days of testing. Actually, for the two concrete mixtures prepared with recycled aggregate
and SRA (i.e. 'RecCon+PC-SRA' and 'RecRub+RM-SRA'), values of drying shrinkage
less than 50 !lmlm were detected: after 70 days concrete shrinkage was decreased by 79%
and 73% in the presence of RM-type SRA and PC-type SRA, respectively, when
compared to the same concrete prepared with pure superplasticizer. The reason probably
lies on the positive effect of the preliminary water-soaking of recycled aggregate in
CONCLUSIONS
The kind of curing, too, does not seem to influence compressive strength of concretes
prepared by using SRA, at least in the range 80%-100% ofre1ative humidity.
For the recycled-aggregate concretes the measured values of drying shrinkage were
always very low, less than 180 f,lmlm after 70 days of testing: the reason probably lies on
the positive effect of the preliminary water-soaking of the recycled aggregate fraction.
Values of drying shrinkage less than 50 f.tmlm were detected for the concrete prepared
with both recycled aggregate and SRA.
ACKNOWLEDGMENTS
The writers acknowledge the General Admixtures Spa ofPonzano Veneto, Treviso (Italy)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
for supplying chemical admixtures.
REFERENCES
Bentz, D.P., Geiker, M.R., and Hansen, K.K., 2001, "Shrinkage-reducing admixtures and
early-age desiccation in cement pastes and mortars", Cement & Concrete Research, V.
31, pp. 1075-1085.
Berke, N.S., Li, L., Hicks, M.C., and Bal, J., 2003, "Improving concrete performance
with shrinkage-reducing admixtures", Superplasticizers and Other Chemical Admixtures
in Concrete: Proceedings of the Seventh CANMETIACI International Conference, Berlin,
Germany, 2003 (SP-217), ed. V.M. Malhotra, American Concrete Institute, pp. 37-50.
Collepardi, M., 1991, Concrete Science and Technology (in Italian), Third Edition,
Hoepli, Milano, Italy, p. 301.
Collepardi, M., Borsoi, A., Collepardi, S., Ogoumah Olagot, J.J., and Troli, R., 2005,
"Effect of shrinkage reducing admixture in shrinkage compensating concrete under non-
wet curing conditions", Cement & Concrete Composites, V. 27, pp. 704-708.
Corinaldesi, V., and Moriconi, G., 2001, "Role of Chemical and Mineral Admixtures on
Performance and Economics of Recycled-Aggregate Concrete", Fly Ash, Silica Fume,
Slag and Natural Pozzolans in Concrete: Proceedings of the Seventh CANMETIAC1
International Conference, Madras, India, 2001 (SP-199), ed. V.M. Malhotra, American
Concrete Institute, pp. 869-884.
Corinaldesi, V., and Moriconi, G., 2002, "Rubble Processing to Manufacture Structural
Recycled-Aggregate Concrete", Cement and Concrete: Proceedings of the 5th
International Symposium, Shanghai, China, 2002, ed. Wang Pei-ming, Tongji University
Press, Volume 2, pp. 796-802.
Corinaldesi, V., and Moriconi, G., 2003a, "The Use of Recycled Aggregates from
Building Demolition in Self-Compacting Concrete", Self-Compacting Concrete:
Proceedings of the 3rd International RILEM Symposium, Reykjavik, Iceland, 2003 (PRO
33), ed. 0. Wallevik, and I. Nielsson, RILEM Publications S.A.R.L., France, pp. 251-
260.
Corinaldesi, V., and Moriconi, G., 2003b, "Recycled Aggregate Concrete under Cyclic
Loading", Role of Concrete in Sustainable Development: Proceedings of the
International Symposium dedicated to Professor Surendra Shah, Dundee, Scotland, UK.,
2003, ed. R.K. Dhir, M.D. Newlands, and K.A. Paine, Thomas Telford, London, U.K.,
pp. 509-518.
Corinaldesi, V., Tittarelli, F., Coppola, L., and Moriconi, G., 2001, "Feasibility and
Performance of Recycled-Aggregate Concrete Containing Fly Ash for Sustainable
Building Development", Sustainable Development of Cement and Concrete: Proceedings
of the Three-Day CANMETIACI International Symposium, San Francisco, US.A., 2001
(SP-202), ed. V. M. Malhotra, American Concrete Institute, pp. 161-180.
Dhir, R.K., Henderson, N.A., and Limbachiya, M.C., 1998, Use of Recycled Concrete
Aggregate, Thomas Telford Publishing, London, U.K., 525 pp.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Hansen, T.C., and B0egh, E., "Elasticity and Drying Shrinkage of Recycled-Aggregate
Concrete", Technical Paper, ACI Journal, Sept-Oct 1985, pp. 648-652.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Kasai, Y., 1998, Reuse of Demolition Waste, Volume Two, Chapman and Hall, London,
U.K., pp. 500-774.
Maltese, C., Pistolesi, C., Lolli, A., Bravo, A., Cerulli, T., and Salvioni, D., 2005,
"Combined effect of espansive and shrinkage reducing admixtures to obtain stable and
durable mortars", Cement & Concrete Research, V. 35, pp. 2244-2251.
Nakanishi, H., Tamaki, S., Yaguchi, M., Yamada, K., Kinoshita, M., Ishimori, M., and
Okazawa, S., 2003, "Performance of a Multifunctional and Multipurpose Superplasticizer
for Concrete", Superplasticizers and Other Chemical Admixtures in Concrete:
Proceedings of the Seventh CANMETIACJJnternational Conference, Berlin, Germany,
2003 (SP-217), ed. V.M. Malhotra, American Concrete Institute, pp. 327-342.
Palacios, M., and Puertas, F., 2005, "Effect of superplasticizer and shrinkage-reducing
admixtures on alkali-activated slag pastes and mortars", Cement & Concrete Research,
v. 35, pp. 1358-1367.
Rongbing, B., and Jian, S., 2005, "Synthesis and evaluation of shrinkage-reducing
admixture for cementitious materials", Cement & Concrete Research, V. 35, pp. 445-448.
Superplasticizin~ admixture
(kg/m)
3 - -
Superplasticizer with PC-type
SRA(kg/m3) - 6 -
Superplasticizer with RM-type
SRA(kg/m3) - - 8
V'l
~
t.)
Q
~
::::1
Q)
_
... ...... _ -
Mixtures
~
.. -~ ·~ ~···~
RecCon+SP
a ~- -~••-• - - -
--- ···-· ·--
RecCon+PC-SRA
-·- - - --
RecRub+SP RecRub+RM-SRA
s::
tD
!!!.
Water/Cement 0.45 0.45 0.45 0.45 Q)
::::s
Q.
Water (kglm3) 165 165 165 175
3:
Cement (kglm3) 365 365 365 390
..,
0
Natural sand (kglm3) 550 545 735 710 -·n
0
::::s
Natural gravel (kglm3) - - 550 530 -·
Recycled concrete (kg/m3) 1195 1190 - -
-
Not for Resale, 08/10/2015 01:25:47 MDT
Licensee=UNI OF NEW SOUTH WALES/9996758001
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Superplasticizers and Chemical Admixtures in Concrete 543
Table 4- Effect of curin~ on 28-day compressive strength of concretes prepared with SRA
Wet curing Humid air curing
Temperature (°C) 10 10
100
[:r-c.- ~... ,..,. .~l!' ~I[ __~!]_,__
1*1:-nr. . . . ;
i
Natural sand
j
-+ R~cled ru~rv_j
---~ :
l .I,,
· I
I /i !,
1(:.: 1 ·+·····'·····
'} i
:
0 .btf
0.1
I I
il Jic
• '
1
Sieve opening (mm)
! !i
10
! I
100
10 15 20 25 30
Curing time (days)
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
400~----~------~------~----------~
0 20 40 60 80 100
Curing time at 1s•c and 65% R.H. (days)
Performance of Self-Consolidating
Concrete made with Diutan Gum
Synopsis: Dynamic and static stability of self-consolidating concrete (SCC) affect the
production, transport, and overall performance of the concrete. Viscosity-modifying
admixtures (VMAs) are often incorporated to enhance the stability of SCC, especially in
the case of cast-in place concrete. The influence of a new type of polysaccharide, diutan
gum, on key characteristics of sec targeted for the construction and repair of concrete
infrastructure is examined. The investigation compares the performance of SCC made
with diutan gum and different types of high-range water reducing admixtures
(HRWRAs), including polynaphtalene sulphonate (PNS)-based and two polycarboxylate
polymer (PCP)-based HRWRAs. The effect of admixture combination on workability,
rheology, stability, and setting time was investigated. The robustness of optimized VMA-
HRWRA system and key durability parameters of the concrete were also determined.
Test results indicate that the use of diutan gum increases the plastic viscosity and yield
stress of SCC. Regardless of the VMA and HRWRA combination, the use of diutan gum
can significantly decrease segregation and bleeding, and lead to greater homogeneity of
the concrete during the dormant period of cement hydration. The performance of SCC
made with diutan gum depends on the type of PCP in use. Out of two PCP-based
HRWRA, one system resulted in lower HRWRA demand, increase in viscosity with the
increase in VMA content, and greater stability. The use of PCP and diutan gum is shown
to lead to adequate robustness where the sec can tolerate small changes in sand moisture
content without significant effect on sec properties. sec made with medium dosage of
diutan gum and PCP is shown to develop adequate air-void system and excellent frost
durability and resistance to de-icing salt scaling.
INTRODUCTION
Polysaccharides are high molecular weight, water-soluble organic polymers that can be
used to enhance the resistance to washout, bleeding, and segregation of cement-based
materials. Such viscosity-modifying admixtures (VMAs) play a key role in the
production of self-consolidating concrete (SCC), especially in the case of ready-mixed
concrete. Incorporating small dosage rates of VMA in SCC can enhance cohesiveness
and stability of the plastic concrete, as well as the robustness of the material. Greater
robustness refers to the ability of freshly mixed concrete to tolerate small changes in
material characteristics without adverse effect on workability. This includes changes in
concrete temperature, sand humidity and fineness modulus, and dosage rate of HRWRA.
Greater robustness can lead to more consistent and reliable daily production of the
concrete.
The increased yield stress of the concrete associated with the incorporation of VMA
necessitates an increase in high-range water-reducing admixture (HRWRA) demand.
Polysaccharides supplied as dry powdered materials are usually converted into liquid-
based suspensions to facilitate dispensing and ensure homogeneous dispersion of the
VMA during concrete mixing. Superplasticizers provide suitable suspension mediums for
polysaccharides. Such slurry medium should prevent the polysaccharide from hydrating
and must be selected to ensure adequate compatibility with the admixtures in use.
The mode of action of VMA depends on the type and concentration of the polymer, and
its performance depends on the type of the HRWRA used in the mixture. In this study,
the performance of a new type of polysaccharide, diutan gum, in enhancing the plastic
properties of SCC is investigated. Diutan gum is produced using a carefully controlled
aerobic microbial fermentation process utilizing a suitable carbon source. The backbone
is comprised of D-glucose, D-glucuronic acid, D-glucose, and L-rhamnose. The diutan
repeat unit contains two L-rhamnose units substituted on C3 of every 1-4-linked glucose
residue. The side chains shield the carboxylate group in the double-helix more completely
by hydrogen-bonding to the oxygen in the carboxylate group (1 ). This renders the
molecule unsusceptible to cation-mediated bridging (2). Van der Waals interactions of
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Diutan rheological properties are defined by the conformation (chain shape) and
molecular weight (chain length), and tendency for the gum molecules to associate. The
chain conformation and molecular weight define the hydrodynamic volume swept out by
the molecule. One could relate conformation to the chain shape assumed in solution and
molecular weight to the chain length. The hydrodynamic volume and the tendency of the
molecule to associate with itself and neighboring molecules define the rheological
performance.
The influence of using diutan gum in SCC at various concentrations along with
polynaphtalene sulphonate (PNS)-based and polycarboxylate polymer (PCP)-based
HRWRAs is investigated in this study. The effect of admixture combination on HRWRA
demand, fluidity loss, stability of air volume, self-consolidating properties, and setting
time is evaluated. Similarly, the robustness of diutan gum used with a compatible
HRWRA is determined, and frost durability and de-icing salt scaling resistance of the
concrete are evaluated.
EXPERIMENTAL PROGRAM
The testing program involved the evaluation of the effect of various VMA-HRWRA
combinations on fresh characteristics of SCC. Diutan gum was combined with different
types of superplasticizers: polynaphtalene sulphonate (PNS) and two polycarboxylate
polymers (PCP), referred to here as PCPl and PCP2. The SCC mixtures were
proportioned with 0.42 w/cm and 475 kg/m3 of blended cement. The diutan gum was
added at three concentrations of0.04%, 0.08%, 0.12%, by mass of water, and the dosage
rate of the HRWRA was adjusted to maintain a fixed slump flow consistency (665±15
mm). Given a w/cm of 0.42, these diutan gum concentrations correspond to 0.017%,
0.034%, 0.05%, by mass of binder. Such values can be considered as very low to medium
concentrations of the polysaccharide for use in sec.
The SCC mixtures were air-entrained to ensure initial air volume of 5% to 8% and were
tested to determine the retention of slump flow, air volume stability, rheological
properties, dynamic and static stability, and setting time. The robustness of SCC made
with diutan gum and PCPl was evaluated for ±10°C changes in concrete temperature and
±1% variations in sand humidity. The mixtures were tested to determine admixture
demand, workability, and compressive strength. The resistance to freezing and thawing
and to de-icing salt scaling ofSCC made with diutan gum and PCPl was also evaluated.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Blended cement containing Type 10 (GU) cement and substitutions of cement with 22%
Class F fly ash and 6% silica fume, by mass, was employed. The chemical and physical
properties of the binder components are given in Table 1. Continuously graded crushed
limestone aggregate with nominal size of 10 mm was used to facilitate the filling of
densely reinforced elements. The coarse aggregate has a specific gravity of 2.74 and
0.3% absorption. Well-graded siliceous sand was used. The sand has a specific gravity of
2.69, 1.22% absorption, and fineness modulus of2.40.
solid content and specific gravity of 1.08. Diutan gum was used in a dispersed solution of
8% concentration with each of the investigated HRWRA in order to facilitate dispending
and ensure better dispersion of the gum in concrete.
The concrete was mixed in an open pan mixer of 100-L capacity. The mixing sequence
consisted of homogenizing the sand and coarse aggregate for one minute in the mixer
before introducing one third of the mixing water. The cement was then added, along with
the HRWRA and the remaining water. After three minutes of mixing, the VMA diluted in
HRWRA was introduced, and the concrete was mixed for two additional minutes. The
temperature of the concrete during sampling was maintained at 20 ± 2°C, unless
otherwise specified. Following concrete mixing, the initial slump flow, concrete
temperature, unit weight, and air volume were determined. This was followed by
evaluating the passing ability, stability, and rheological parameters of the concrete. The
slump flow was determined immediately after mixing, which corresponded to
approximately 5 min of age. The remaining measurements were done between 5 and 15
min following the initial contact of cement with water. The characteristics of the fresh
concrete were again determined at the age of 70 minutes. The fresh concrete remained in
the mixer and was agitated for one minute at 10-min intervals.
The passing ability was evaluated using the J-Ring and L-box test methods. In both cases,
the gaps between the reinforcing bars were set to 35 mm. The filling capacity of the SCC
was evaluated using the caisson filling capacity test (4). The caisson test consists of an
acrylic box with dimensions of 300 x 500 x 300 mm with several closely spaced smooth
horizontal bars with clear spacing of 34 mm. The static stability of the concrete was
evaluated using the surface settlement and electrical conductivity tests. In the former test,
concrete is cast in a PVC column measuring 200 mm in diameter and 600 mm in height.
Settlement is monitored using an LVDT fixed on top of a thin acrylic plate anchored to
the upper surface of the concrete (5). The settlement is monitored until a steady state
condition is reached.
Tables 2 to 4 show the mixture proportioning and properties of the investigated SCC. The
mixtures had w!cm of 0.42, binder content of 475 kg/m3 , and fixed sand/(sand +
aggregate) ratio of 0.50, by volume. The dosage rates of the HRWRA and air-entraining
admixture (AEA) were adjusted to secure initial slump flow and air volume of 665 ± 15
mm and 6.5 ± 1.5%, respectively.
HRWRA demand
The influence of diutan gum on HRWRA demand is illustrated in Fig. 1. SCC made with
PNS necessitated greater HRWRA demand than similar concrete made with PCP,
regardless of the VMA content. The two reference mixtures made without any VMA had
the lowest HRWRA demand; the reference mixtures prepared with PNS and PCP2 had
HRWRA demands of 6.9 and 2.1 L/m3 , respectively (Tables 2 to 4). The combination of
PCP and diutan gum led to considerably lower HRWRA demand than similar concrete
made with PNS and diutan gum. For example, the PNS and PCPI and PCP2 mixtures
made with 0.08% diutan gum had HRWRA demand of 9.5, 4.0, and 5.4 Llm3,
respectively.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Rheological parameters
Table 5 summarizes the rheological parameters of the investigated mixtures. The torque-
angular velocity data are plotted to fit the Bingham model. The correlation coefficients of
the derived linear relationships were greater than 0.99. Despite similar slump flow values
of 665 ± 15 mm, some variations were obtained for the apparent yield stress values (g) of
the various mixtures. The initial values of g varied from 0.3 to 1.0 N.m for all mixtures,
regardless of the VMA-HRWRA combination. The torque plastic viscosity (h) values at
10 min ranged between 3.3 and 12.6 N.m.s. The highest viscosity was obtained in the
case of the SCC made with 0.04% diutan gum. For the mixture made with PNS and
0.04% of diutan gum, the h value was high. The lowest viscosity values were obtained
with the sec proportioned with the mid concentration of diutan gum of 0.08% in
mixtures made with PCP2- and PNS-based HRWRA.
Figure 2 shows the effect of increasing VMA dosage on h of SCC made with different
HRWRA types determined at 10 and 70 minutes. It is important to note that the dosage of
the HRWRA was adjusted to maintain constant slump flow of 665 ± 15 mrn. The
reference non-VMA concrete made with PCP 1 also exhibited slightly greater torque
plastic viscosity than that made with PNS or PCP2. Regardless of the HRWRA type, the
increase in diutan gum from 0 to 0.04% led to higher plastic viscosity compared to non-
VMA mixtures. At low gum concentration of 0.04%, the highest viscosity was obtained
for concrete made with PNS, both at 10 and 70 minutes of age. With the increase in VMA
content to 0.08% and 0.12%, a further increase in viscosity was obtained for the sec
made with PCPl-based HRWRA. The increase ofVMA content to 0.08% and 0.12% in
the case of concrete made with PNS or PCP2 resulted in reduction in viscosity, both at 10
and 70 minutes. This can be due to the high concentration of HRWRA required in the
mixtures compared to the SCC made with PCP!. In other words, the effect of HRWRA
was more dominant compared to that of the diutan gum used at 0.08% and 0.012%
dosage rates. Furthermore, mixtures made with 0.08% and 0.12% diutan gum had for the
most part 1.5% to 2% greater air volume than those made with 0.04% diutan gum and
PNS or PCP2; the increase in air volume contributes to the reduction in viscosity.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
The results of the L-box, J-Ring, and caisson filling capacity tests are presented in Tables
2 to 4. The increase in L-box blocking ratio (h2/hl), J-Ring spread diameter, and filling
capacity reflect an increase in passing ability and resistance to blockage of concrete flow
among closely spaced obstacles. The reference non-VMA mixtures had h2/hl values of
0.74 to 0.90. The incorporation ofVMA at low and medium concentrations led to a slight
increase in passing ability; however, further increase in VMA content to 0.12% led to
some reduction in blocking ratio, especially in the case of the PCPl HRWRA that
exhibited the highest viscosity level.
The spread between slump flow and J-Ring flow values are given in Tables 2 to 4 and
varied from 0 to 80 and 0 to 60 mm after 10 and 70 minutes, respectively. In general,
spread between non-restricted slump flow and restricted J-Ring flow values lower than 50
mm would correspond to adequate passing ability. The increase in VMA concentration
resulted in lower spread between the slump flow and J-Ring flow values at 10 and 70
minutes. The filling capacity ranged between 85% and 99% at 10 min and 61% and 98%
at 70 minutes. In general, filling capacity values greater than 80% would yield adequate
performance for the filling of highly restricted section (5). Compared to reference SCC
made without any VMA, the increase in diutan gum content is shown to enhance the
filling capacity of the SCC, regardless of the HRWRA type, both at the age of 10 and 70
minutes.
Static stability
The stability indices determined from electrical conductivity of SCC made with PNS and
PCP2 are given in Table 6. Based on the conductivity data, it is possible to define
stability indices that determine the evolution of bleeding and segregation of the fresh
concrete (7). The bleeding index is taken here as the cumulative area between
conductivity values of the top electrode pair and the second pair. This area is considered
between the initial and peak values, divided by the effective time required to reach the
peak value. The segregation index is defined as the cumulative area between the bottom
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
In general, SCC mixtures made with either PNS or PCP2 exhibited low stability indices,
which reflect the stable nature of the selected SCC. These mixtures had maximum surface
settlement values lower than 0.5%. For SCC made with PNS, the increase in
concentration of the diutan gum led to sharp decrease in bleeding index nearing zero at
0.12% gum content. Similarly, the increase of diutan gum from 0 to 0.08% resulted in
sharp reduction in segregation and homogeneity indices. However, such added stability
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
was not observed when the gum concentration increased to 0.12%. This is believed to be
due the high content of PNS-based HRWRA that was 11.3 Llm3 . The bleeding and
segregation indices were similarly affected by the type of the HRWRA. In the case of
SCC prepared with PCP2, the increase in diutan gum concentration from nil to 0.08% led
to some increase in the bleeding, segregation, and homogeneity indices. Again, this could
be related to the greater HRWRA demand that increased from 2.1 to 5.3 L!m3 . The
increase in VMA content to 0.12% did not necessitate any additional HRWRA demand
and resulted in about 50% reduction in stability indices compared to the sec made
without any VMA.
Stability of air
As shown in Table 2, the AEA dosage of concrete made with PNS dosage and VMA was
similar to that of the reference non-VMA concrete. On the other hand, the incorporation
of diutan gum necessitated very large increase in AEA in mixtures made with PCP. For
example, the required AEA dosage increased from 50 to 500 ml/m3 with the addition of
0.04% diutan gum in order to maintain the targeted initial air volume of 5% to 8%. This
is thought to be due to the increase in HRWRA demand, and hence increase in
concentration of de-foamer making it more difficult to air-entrain the sec at
conventional dosage rates of the AEA.
As mentioned earlier, the concrete was subjected to occasional agitation in the mixer for
60 min before the second set of testing. The SCC made with PNS exhibited 0.7% to 3.5%
loss in air volume, with the upper values obtained with mixtures made with the higher
VMA contents. Concrete proportioned with PCPl exhibited stable air content between 10
and 70 min, except in the case of the PCPl + 0.12% Diutan where the air content
increased by 2.4% with the occasional agitation. The reference PCP2 concrete had also
stable air content; however, the incorporation ofVMA resulted in 2.0% to 4.5% increase
in air volume between 10 and 70 minutes. Therefore, care must be taken in selecting the
HRWRA type that would not lead to excessive increase in air volume in SCC made with
diutan gum.
The initial setting time of mixtures made with PNS and PCP2 are shown in Fig. 4. The
initial setting time ranged from approximately 7.5 to 14.5 hours, and the final setting time
varied from approximately 8 and 16 hours. For SCC incorporating PCP2, the increase of
diutan gum from nil to 0.08% or 0.12% resulted in an increase in HRWRA demand from
2.1 to 5.3 Llm3 as well as a delay of 6.5 hours in initial setting time. In the case of SCC
made with PNS-based HRWRA, an increase in diutan gum from 0 to 0.08% resulted in a
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
delay of 3.5 hour in initial setting. Further increase in gum content to 0.12% resulted in
further delay of 3.5 hours in initial setting. As discussed earlier, the increase in diutan
gum concentration resulted in an increase in the PNS-based HRWRA demand from
approximately 7 to 11 L!m3 • The higher PNS content can prolong the dormant period.
Based on the above results, liquid-based VMA containing diutan gum and PCP!
HRWRA was prepared for further investigation. The admixture was used at medium
dosage to proportion air-entrained SCC. Five SCC mixtures were prepared with diutan
gum and PCP! to evaluate the robustness to changes in concrete temperature (± l0°C)
and sand moisture content that deviate ± 1% from surface saturated dry condition (SSD).
Achieving robust concrete is quite important since the conditions on the building site vary
frequently, thus affecting water demand, changes of self-consolidating properties,
kinetics of cement hydration, air entrainment, and stability.
Tables 7 and 8 present the mixture proportioning and test results of the SCC evaluated for
robustness investigation. In order to produce concrete of 32°C, the coarse aggregate,
sand, and cement were stored at 38°C for 24 hours before mixing, and the mixing water
was heated to 42°C. For the concrete prepared at l2°C, material constituents were
maintained at 5°C, and small quantities of crushed ice were employed to replace part of
the mixing water. The HRWRA and AEA were diluted with half of the mixing water. The
mixing sequence was the same as that described earlier.
The optimized PCP! + 0.08% Diutan mixture proportioned at SSD showed good
workability and homogeneity at 10 and 60 minutes. The slump flow and filling capacity
values were 675 mm and 99% at 10 min, and 700 mm and 95% at 60 minutes. The torque
plastic viscosity (h) values were 6.8 and 5.2 N.m.s at 10 and 60 min, respectively. The
demand of the PCP! HRWRA in the reference SCC was 4.0 L/m3 •
An increase in sand moisture of+ 1% resulted in additional water (w!cm of 0.44) and an
increase in slump flow of 45 mm without much difference in filling capacity and initial
air content. With the higher slump flow, the SCC with the higher water content exhibited
lower apparent yield stress (g) of 0.2 at 10 and 60 minutes. Regardless of the sand
A reduction in the quantity of water that is equivalent to -1% of sand moisture (wlcm of
0.40) caused an increase of2.6% in air content (total air content of 10%) at 10 min and of
2% (from 9% to 11 %) at 60 minutes for a fixed dosage of AEA. The slump flow value
was identical to the mixture at SSD. The torque plastic viscosity (h), filling capacity, and
24-hour compressive strength values were nearly identical to those of the reference sec.
Small increase in settlement (0.2%) was noted. The 28-d and 56-d compressive strength
results of the SCC made with -1% sand humidity are omitted due to experimental error.
Overall, SCC made with PCP1 and diutan gum exhibited good robustness to small
changes in sand moisture where variations in sand humidity of ±1% did not seem to have
much effect on the self-consolidating properties of the concrete (filling ability, filling
capacity, and stability). With the decrease in sand humidity, an increase in air content of
approximately 2.5% was obtained. Care must be taken to adjust the dosage of the AEA to
maintain the air volume within the targeted range.
Compared to the reference SCC at 22°C, the decrease in concrete temperature to l2°C,
without changing the HRWRA content, had limited effect on slump flow at 10 and 60
minutes. The colder concrete developed lower settlement of 0.1 0%. On the other hand,
the volume of air of 9% at 10 and 60 min was higher than the reference SCC. The filling
capacity was nearly 100% at both 10 and 60 min and was slightly greater than that
obtained for the reference concrete. The torque plastic viscosity (h) values were 5.6 and
4.2 N.m.s at 10 and 60 min, respectively, which were approximately 1 N.m.s lower than
those of the control sec at 22°C.
For the second mixture prepared with the PCP1 + 0.08% diutan at 32°C, the slump flow
at 10 min was 710 mm and underwent a loss of 20 mm after 60 min, which is not the case
for the remaining mixtures made with this HRWRA-VMA combination. This loss caused
an increase in yield stress; the measured value at 10 min is 0.4 N.m, and this value
increased to 0.7 N.m at 60 minutes. Compared to the two other mixtures (cold and SSD
state), the hotter mixture showed slightly higher settlement (0.23%), which is still
considered as stable SCC. Its 24-hour compressive strength was 3.9 MPa. The air volume
was 8.9%, which was slightly greater than that at 22°C (7.4%). This volume was 9.5% at
60 minutes. The 28-d and 56-d compressive strengths of the concrete made with l2°C
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Frost durability and scaling resistance ofSCC made with diutan-PCP 1 combination
Given the good performance of SCC made with diutan gum and the PCPI HRWRA,
further investigation was carried out to determine the frost durability and de-icing salt
scaling for SCC made with PCPI and 0.08% diutan gum. The air-void system of the
hardened concrete was determined using I 00 x 200 mm cylinder in accordance with
ASTM C 457, Modified Point Count method. The air content of samples taken at 10 min
of age and after 60 min (concrete subjected to occasional agitation) were 7.4% and 9.0%,
respectively, in the fresh state; these values were 7.0% and 7.8%, respectively, in the
hardened state. The spacing factor values of the concrete sampled at lO and 60 min were
225 and 170 J..lm, respectively. Frost durability testing (ASTM C666, Procedure A)
revealed excellent resistance to freezing and thawing with 100% durability coefficient
and elongation of 120 ,. un!m after 300 cycles of freezing and thawing (Table 9). Two
square samples measuring 500 cm2 each were prepared for each concrete mixture to
determine the de-icing salt resistance. The samples were water-cured for 14 days,
followed by 14 days of air-drying, and 7 days of immersion in salt-saturated solution, in
accordance to the Quebec Concrete Standards (BNQ). The freezing and thawing cycles
were started at 35 days of age. The samples were subjected to one freeze-thaw cycle per
day, 16 hours of freezing at -18°C and 8 hours of thawing at 23 °C. In total, 50 cycles of
freezing and thawing were applied to determine the resistance of the concrete to de-icing
salt scaling. The cumulative mass loss of the PCPI + 0.08% diutan concrete was limited
to 250 g/m2 (Table 9), which is quite low considering the highly fluid nature of the
concrete.
CONCLUSIONS
The influence of different combinations of diutan gum and HRWRA on key properties of
SCC designed for construction and repair of concrete infrastructure is examined. The
concrete was air-entrained with 0.42 w/cm and contained 475 kg/m3 of ternary cement.
Based on the above results, the following conclusions are warranted:
I. Good compatibility is observed between diutan and PNS-based HRWRA and the two
PCP-based HRWRAs. The HRWRA demand and workability characteristics of the
SCC is shown to depend on the selected HRWRA and dosage of the VMA.
2. Regardless of the VMA concentration, concrete incorporating PCP-based HRWRA
necessitated lower HRWRA demand than that made with PNS. The HRWRA
demand was the lowest in the case of the PCP-I HRWRA.
3. Increase in diutan gum concentration led to greater HRWRA demand, especially in
the case ofSCC made with PNS-based HRWRA.
4. At a given slump flow consistency, the increase in diutan gum is shown to increase
plastic viscosity without significant change in yield stress; the latter is due to the
REFERENCES
1. Cesaro, A., Gamini A., and Navarini, L. "Supramolecular Structure of Microbial
Polysaccharides in Solution: from Chain Conformation to Rheological Properties,"
Polymer, Vol. 33, No.l9, 1992, pp. 4001-4008.
2. Rakitsky, W.G., "Suspending Agent Applications in Cementitious Systems,"
ConChem International Exhibition & Conference, Karlsruhe, Germany, 1993.
3. Talashek, T.A., and Brant, D.A. "The Influence of Side Chains on the Calculated
Dimensions of Three Related Bacterial Polysaccharides," Carbohydrate Research,
1987, Vol. 160, pp. 303-316.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
7. Khayat, K.H., Pavate, T., Assaad, J., and Jolicoeur, C., "Analysis of Variations in
Electrical Conductivity to Assess Stability of Cement-Based Materials," ACJ
Materials Journal, Vol. 100, No.4, 2003, pp. 342-350.
Table 1 - Chemieal and physieal properties of eement, fly ash, and sitiea fume used
in this investigation
Type GU cement Silica fume Fly ash
Si02,% 21.0 93.6 41.9
Ah03,% 4.2 0.3 23.43
Fe203,% 3.1 0.5 18.89
CaO,% 62.0 0.3 7.11
MgO,% 2.9 0.5 0.86
Na20eq.,% 0.74 1.4 1.57
S03,% 2.7 - -
C3S,% 52 - -
C2S,% 21.5 - -
C3A,% 5.7 - -
c~,% 9.5 - -
Blaine specific surface, m2/kg 420 - 410
Surface area B.E.T., m2/kg - 20,250 -
Specific gravity 3.17 2.22 2.53
Percent passing 45 11m, % 95 IOO 90
Bulk unit weight, kg/m3 3150 280 -
L.O.I.,% 2.5 2.8 2.97
~~ t;~
zo tJgga J) ~ .2
S:> ~C!i
0
" ~0~ ~;;;o
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Water kglm' 199.5 199.5 199.5 199.5
PNS-based HRWRA Um' 6.9 7.0 10.0 11.3
AEA mllm' 750 600 700 800
Diutangum g/m' 0 80 160 240
Temperature 10min ·c 21 21.8 21.8 21.9
Unit weight lOmin kglm3 2206 2240 2191 2202
Entrained air lOmin % 6.8 6.2 8.5 8.0
content 70min % 6.1 5.5 5.0 6.2
10min mm 650 680 690 660
Slump flow
70min mm 520 680 665 680
10min nun 620 680 680 610
J-Ring
70min nun 500 620 660 680
Slump flow- 10min mm 30 0 10 50
J-Ring 70min mm 20 60 5 0
lOmin sec 0.53 3.10 1.69 2.69
h2/h1 0.90 0.90 0.88 0.74
L-box
70min sec 1.44 4.01 1.31 2.7
h2/h1 0.54 0.81 0.73 0.8
10min % 90.0 98.5 97.2 93.7
Filling capacity
70min % 61.1 98.1 95.5 97.0
Max. settlement % 0.51 0.4 0.3
Initial hr 7:20 - 10:45 14:15
Setting time
Final hr 8:10 - 12:20 15:45
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Temperature IOmin oc 19.6 20.7 20.0 20.7
Unit weight 10min kglm' 2277 2063 2002 2206
Entrained air 10min % 5.5 10 10 7.6
content 70min % 4.8 10 10 10.0
lOmin mm 655 675 660 650
Slump flow
70min mm 540 650 580 590
lOmin mm 635 600 640
J-Ring
70min mm 620 540 580
Slump flow- 10min mm 40 60 10
J-Ring 70min mm 30 40 10
lOmin sec 0.93 3.22 6.41 4.53
h2/hl 0.74 0.82 0.82 0.75
L-box
70min sec 1.00 3.68 9.00 5.00
h2/hl 0.74 0.85 0.52 0.6
lOmin % 88.3 94.4 91.6 91.0
Filling capacity
70min % 73.4 92.8 69.8 78.3
Max. settlement % 0.70 0.11 0.16 0.40
~~ ~~ s:!ooo
uo
~~do
~·=
~ ~~ Po.
Setting time
Initial hr 08:05 - 14:25 14:15
Final hr 09:20 - 15:55 16:00
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
10 min 70min
Mixtures
g(N.m) h (N.m.s) R2 g(N.m) h(N.m.s) R2
PNSnoVMA 0.3 3.3 0.994 0.6 4.2 0.988
PNS+
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Segregation Homogeneity
Mixtures combinations Bleeding index
index index
noVMA 0.17 0.11 0.16
PNS 0.08%Diu 0.08 0.03 0.05
0.12% Diu 0 0.11 0.12
noVMA 0.06 0.13 0.10
PCP2 0.08%Diu 0.15 0.30 0.45
0.12% Diu 0,03 0.06 0.08
V'l
0\
N
="
:r
~
--
tD
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
D)
Table 7 -Mixture constituents used for robustness evaluation of SCC made with diutan gum and PCPl HRWRA
r
SSD mixture temp. Sand humidity Effect of temperature
Units 22°C +1% -1% l2°C 32°C
w/cm 0.42 0.438 0.404 0.42 0.42
Water Urn' 200 208 192 200 200
3 475 475 475 475 475
Binder kg/m
Coarse aggregate kg!m' 800 800 800 800 800
Sand kg/m 777 769 785 777 777
PCPlHRWRA Lim' 4.0 4.0 4.0 4.0 4.0
Diutan gum - 0.08% glm' 160 160 160 160 160
Set retarder 0.75 0.75 0.75 0.75 0.75
Not for Resale, 08/10/2015 01:25:47 MDT
Licensee=UNI OF NEW SOUTH WALES/9996758001
Llm3
AEA 0.10 0.10 0.10 0.10 0.10
Daneshlink.com
No reproduction or networking permitted without license from IHS
Provided by IHS under license with ACI
Copyright American Concrete Institute en
c
"C
tD
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`--- ""l
"C
Results
Table 8- Results of robustDess investigation of SCC made witll di111Dn gum and PCI'l HRWRA
;::;·
N.
tD
Units
temp. 22°C +1% -1% I2°C 32°C ~
IOmin 21.7 21.6 22.2 13.6 32.1
Temperature oc Q)
60min 21.5 21.1 21.8 16.9 29 ::s
IOmin 675 720 670 680 710 0.
Slump flow mm
60min 700 720 680 720 690 n
Unit weight kg!m'
10min 2163 2149 2086 2115 2134 ::s"
60min 2137 2208 2046 2093 2091 tD
Entrained air IOmin 7.4 7.5 10.0 9.0 8.9 3
content
%
;::;·
-~
60min 9.0 7.2 11.0 9.0 9.5
lOmin 98.8 99.4 97.6 99.6 98.8 Q)
Filling capacity %
60min 95.3 99.6 97.0 99.6 99.8
g,N.m 0.4 0.2 0.6 0.4 0.4
h, N.m.s IOmin 6.8 5.6 6.6 5.6 5.8
Rbeological R' 0.994 0.999 0.998 0.998 0.995 3
parameters g, N.m
h, N.m.s 60min
0.4
5.2
0.2
4.6
0.6
4.9
0.4
4.2
0.7
4.6
8."
c
R' 0.996 0.998 0.998 0.998 0.998 ;
Not for Resale, 08/10/2015 01:25:47 MDT
Licensee=UNI OF NEW SOUTH WALES/9996758001
Compressive
strength
MPa
I day
28days
2.1
36.2
3.3
39.0
2.2
-
2.9
34.7
3.9
38.1
-·
::s
56 days 41.0 44.5 37.8 40.0
~
::s
n
til
;-
Daneshlink.com
U'l
0\
\A)
564 Khayat et al.
Table 9 - Frost durability and de-icing salt scaling of SCC made with 0.08% diu tan gum and PCPI HRWRJ
690
I 680
~ 670
<+::
j 660
U'l
650
640
630
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
14
)1~- -----------------
12
0+-----~----~------~-----,------.-----~
Fig. 2- Effect of diutan gum dosage and HRWRA type on initial rheological parameters
0.7
0.6
~ 0.5
i
I 0.4
rll 0.3
0.2 '\
\-----------------------JI(
0.1
0.0
0 0,02 0.04 0.06 0.08 0.1 0.12
Diutan gwn (%)
900
800
i
'-' 700
-~
gp 600
"fi
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
"' 500
400
300
Length
inch millimeter (mm) 25.4E+
foot meter (m) 0.3048E
yard meter(m) 0.9144E
mile (statute) kilometer (km) 1.609
Area
square inch square centimeter (cm 2) 6.451
square foot square meter (m 2 ) 0.0929
square yard square meter (m 2 ) 0.8361
Volume (capacity)
ounce cubic centimeter (cm 3 ) 29.57
gallon cubic meter (m 3):j: 0.003785
cubic inch cubic centimeter (cm 3 ) 16.4
cubic foot cubic meter (m 3) 0.02832
cubic yard cubic meter (m 3 ):j: 0.7646
Force
kilogram-force newton (N) 9.807
kip-force newton (N) 4448
pound-force newton (N) 4.448
Pressure or stress
(force per area)
kilogram-force/square meter pascal (Pa) 9.807
kip-force/square inch (ksi) megapascal (MPa) 6.895
newton/square meter (N/m 2) pascal (Pa) l.OOOE
pound-force/square foot pascal (Pa) 47.88
pound-force/square inch (psi) kilopascal (kPa) 6.895
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Temperature§
deg Fahrenheit (F) deg Celsius (C) tc = (IF - 32)/1.8
deg Celsius (C) deg Fahrenheit (F) tF = 1.8tc + 32
* This selected list gives practical conversion factors of units found in concrete technology. The reference
source for information on SI units and more exact conversion factors is "Standard for Metric Practice" ASTM E
380. Symbob of metric units are given in parenthe~es.
t E indicates that the factor given is exact.
:j: One liter (cubic decimeter) equals 0.001 m 3 or 1000 cm 3 .
§ These equations convert one temperature reading to another and include the necessary scale corrections. To
convert a difference in temperature from Fahrenheit to Celsius degrees, divide by 1.8 only, i.e., a change from 70
to 88 F represents a change of 18 For 18/1.8 = 10 C.
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Index
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
alkali sulphate, 269 coagulation, 321
alkaline cements, 345 Colangelo, F., 117
amphoteric copolymer, 309 Collepardi, M., 1, 525
artificial aggregates, 67 Collepardi, S., 375
artificial "Terrazzo" flooring, 67 Comparet, C., 151
autogenous shrinkage, 471 compressive strength, 51, 117, 487
autogenous shrinkage strain, 51 concrete, 137, 169, 249, 309
aqueous phase, 357 concrete-equivalent mortar, 185
consolidation, 321
B Corazza, G., 375
Banfill, P.F.G., 345 Cord in,)., 409
Bentur, A., 15 Corinaldesi, V., 533
Bettencourt Ribeiro, A., 77 Corradi, M., 231
blast furnace slag, 525 corrosion inhibitors, 249
bleeding, 321 corrosion rate, 249
Bolzoni, F., 249 critical chloride threshold, 249
Bravo, A., 487
D
c D'Aioia Schwartzentruber, L., 409
calcium carbonate, 321 Dai, Z., 201
Caputo, D., 117 density, 231
carbonation, 249 Dikty, S., 105
Carrajola, A., 77 dispersant, 507
Cella,F.,487 drying shrinkage, 297
cement, 31, 137, 231, 269, 285, 309, 321, durability, 345
357.455
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
fresh concrete, 423 K
frost durability, 545 Khayat, K.H., 185, 545
Khurana, R., 231
G Kinoshita, T., 51, 471
galvanized steel reinforcement, 441 Kojima, T., 137
Georges, 5., 389 Koyanagi, K., 137
Goidanich, 5., 249 Koyata, H., 15
Golaszewski, J., 423 Kucerova, H., 269
Gon~alves, A., 77
Grossi, G., 525 L
Guillot, L., 357 lignin and derivatives, 507
lignosulfonate, 455
H lignosulphonates, 285
Haehnel, C., 357 Liguori, B., 117
Hamada, D., 31 Lolli, A., 67, 249
Hamai, T., 31 Lombois-Burger, H., 357
Harada, K., 51
Hazrati, K., 15 M
high strength concrete, 471 Magarotto, R., 215, 231
high-range water reducing admixture, Maitrasse, P., 151, 389
185 Maltese, C., 487
Hirata, T., 297 Masanaga, M., 297
Hsu, K.C., 309 McGuire, D., 15
Hwang, S.-D., 185 Mechaymech, A., 545
Hwung, 0.5., 309 Mikanovic, N., 321
hybrid, 471 mineral addition, 117
hydration, 151, 507 mix design, 409
hydration products, 285 molecular structure, 507
hydropobic admixture, 441 Monosi, 5., 375
hyper-branched polymer, 31 Moratti, F., 215
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
Palacios, M., 345 169
paste, 321 Schober, 1., 93, 169
paste flow, 169 secondary fuels, 231
PCE superplasticizers, 231 sedimentation, 321
PCP, 185, 389 segregation resistance, 137
Pedeferri, M.P., 249 self-compacting concrete, 105, 117, 409
Pellay, R., 67 self-consolidating concrete, 185, 545
Pellerin, B., 389 setting regulator, 487
Pellizon Birelli, M., 525 setting time, 51, 487, 507
PEO side chains, 357 Shimoda, M., 31
Perez, J.P., 389 Shonaka, M., 31
performance, 231 shrinkage, 1, 77. 345
petrographic analysis of aggregates, 15 shrinkage reducing, 51
Pistolesi, C., 487 shrinkage reducing admixture(s), 77.
Plank,]., 201 297. 375. 533
plastic viscosity, 31 shrinkage reducing agent, 471
plasticity, 507 Skaggs, B., 545
plasticizers, 455 slag, 345
PNS, 185 slump flow, 389
polyacrylate, 455 slump loss, 1
polycarboxylate, 1, 201, 215, 357, 471 slurry rheology modifier, 137
polycarboxylate ether, 169 solidification/stabilization, 525
polyether, 1 sorbitol, 525
polymer structure, 169 stability, 545
polysaccharide derivative, 105 steric hinderance, 1
Pourchet, S., 151, 389 Sugamata, T., 51
Puertas,F.,345 sulfate(s), 215, 357, 385
--`,```,,``,,,`,,,``````,,,,```,-`-`,,`,,`,`,,`---
T
Takahashi, H., 31
temperature, 423
thickener, 105
Tittarelli, F., 441
tricalcium aluminate, 151, 201
Troli, R., 375
u
ultra-fine fly ash, 117
ultra-high·strength concrete, 51
v
Valente, M., 1
Valenti, G., 441
Ventura, G., 525
Vikan, H., 455
viscoelasticity, 137
viscosity, 471
viscosity-enhancing admixture, 409
viscosity-modifying admixture, 545
viscosity-modifying agent, 15
Vlad, D., 201
w
water·reducing admixture, 507
water reduction, 1
workability, 31, 117, 185, 309, 423
y
Yaguchi, M., 51
Yamamoto, T., 297
Yamamuro, H., 105, 137
yield stress, 15
<·miP
American Concrete Institute•
Advancing concrete knowledge
SP-239
Copyright American Concrete Institute
Provided by IHS under license with ACI Licensee=UNI OF NEW SOUTH WALES/9996758001
No reproduction or networking permitted without license from IHS Not for Resale, 08/10/2015 01:25:47 MDT
Daneshlink.com