Cement, Concrete, and Aggregates, Dec. 2004, Vol. 26, No.

2
Paper ID CCA12315
Available online at: www.astm.org

Joerg Rickert1 and Gerd Thielen2

Influence of a Long-Term Retarder

on the Hydration of Clinker and Cement

ABSTRACT: This paper summarizes the results of investigations on the effects of a long-term retarder (LTR) based on phosphonic acid (PBTC)
on the hydration of several clinker phases (C3 S, C3 A, C4 AF), Portland cement clinker, and Portland cement. It could be shown that long-term
retardation is due to a thin sparingly soluble layer of calcium phosphonate on the particle surface. The formation of sparingly soluble calcium
phosphonate requires 2.5 moles Ca2+ per mole PBTC. If enough Ca2+ is available, sparingly soluble calcium phosphonate precipitates and forms
a layer around the cement particles, which retards further hydration immediately. In the contrary a lack of Ca2+ causes a short-term acceleration
in hydration of the reactive clinker phase C3 A. The investigations show that a controlled long-term retardation can only be achieved if the cement
shows an optimal “natural” setting retardation; hence, when the sulphate ingredient is fully compatible with the reactivity of C3 A.

KEYWORDS: C3 A, C4 AF, C3 S, hydration, long-term retarder, phosphonate, Portland cement clinker, Portland cement, retardation

Introduction is still inadequate. Depending on the temperature and the time of

Long-Term Retarders
Long-term retarders (LTR) are organic retarders, which on ac-
count of their composition are capable of inhibiting the hydration
of cement very strongly, i.e., for several days. During their use as
recycling aids for wash water or residual concrete, they facilitate
the cleaning of truck mixers and concrete mixers, and enable the
recycling of the wash water or the residual concrete at sites that
have no recycling water reservoir, or only a small one. Long-term
retarders allowed in Germany as recycling aids consist of phospho-
nates and mainly contain 2-phosphono-butane-1,2,4-tricarboxylic
acid (PBTC) (Bayer AG 1998; Deutsche Bauchemie e.V. 1999;
DIBT 2000). PBTC affects the hydration reactions of the cement
with the mixing water very strongly.
In addition to wash water and residual concrete recycling, long-
term retarders are also used directly to control the hydration of the
cement, e.g., in combination with accelerators for shotcrete in the
wet-spray process for tunnel construction and for deep drilling op-
erations, and in the production of bore piles and roller compacted
concrete (Persinski et al., 1976; Kinney, 1989; Bobrowski et al.,
1990; Fischer, 1991; Bobrowski et al., 1991; Bürge et al., 1992;
MBT, 1993; Ramachandran, 1995; Bobrowski et al., 1996;
Bodamer, 1997; Wenquan et al., 1997; Paolini et al., 1998; Okawa
et al., 2000).
Contrary to the plasticizers, superplasticizers and air entraining
agents, where a wide range of knowledge has been gained concern-
ing identification and mode of action in cement paste, mortar, and
concrete, knowledge concerning long-term retarding admixtures
Manuscript received; accepted for publication July 30, 2004; published
December 2004.
1 Dr.-Ing., Research Engineer, German Cement Works Association (VDZ),
Concrete Technology Department, Duesseldorf, Germany.
2 Prof. Dr.-Ing., Managing Director, German Cement Works Association
(VDZ), Duesseldorf, Germany.
addition, inverse reactions can occur with particular cements.
For example, retarders may suddenly act as setting accelerators
and/or severely impair the strength development of concretes
(Goetz, 1967; Seiler, 1969; Meyer, 1979; Lewandowski, 1983;
Richartz, 1983; Rickert, 2004a). In practice, this can lead to costly
concrete removal and cleaning work, the loss of the mixing truck
drum, or even the destruction of entire bridge spans.
2-Phosphonobutane-1,2,4-tricarboxylic acid (PBTC)
PBTC is a hydroxycarboxylic acid with five acid sites, which can
dissociate in an alkaline medium. Figure 1 shows the structure of
PBTC in the deprotonated state. Above a pH value of approx. 12.7,
five protons (H+ ) fully dissociated.
In this form, PBTC possesses five coordination sites, which are
suitable for the complexation with metal ions. Studies in Bier (1993)
revealed that in the neutral pH region (pH 7), calcium, aluminium,
cadmium, and magnesium ions are only complexed to a maximum
of 3%, and other metal ions, e.g., chromium, copper, iron, man-
ganese, lead, and zinc, to over 85%. In an alkaline medium, the
complexing ability of PBTC with calcium and aluminum ions in-
creases with increasing pH value (Bier 1993).
Influence of PBTC on the Hydration of Clinker and Cement
Skalny et al. (1980) reported that molecules with complex met-
als can be converted from the extended molecular shape into sta-
ble rings through the metal ion. Such chelating agents can retard
the hydration of cement particularly strongly. Here, the retarder
action generally increases with increasing number of free coordi-
nation sites. According to Skalny et al. (1980), compounds with
a β-hydroxycarboxyl group, which can form a stable 6-member
ring with a metal ion, are particularly good chelating agents and
very effective retarders. PBTC can form similar ring structures by
complexation with calcium ions. According to Kinney (1989),

Copyright

C 2004 by ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. 1
2 CEMENT, CONCRETE, AND AGGREGATES
FIG. 1—PBTC—deprotonated.
recycling aids based on PBTC retarded the hydration of C3 S more
strongly than that of C3 A. It is presumed that PBTC slows the for-
mation of C-S-H and calcium hydroxide nuclei, and as a result finer
and more compact hydrates can form.
In Bobrowski (1990, 1991, 1996), it is suggested that the retarder
action of phosphonic acids is due to the binding of calcium ions
in chelates. According to Paolini et al. (1998), Langenfeld et al.
(1998), and Moeser et al. (1999), a PBTC-containing recycling aid
retards the hydration of C3 A and C3 S very strongly. In Langenfeld
et al. (1998) and Moeser et al. (1999), instead of the initial hydration
products, “spherical new formations” about 100 nm in size were
observed on the surfaces of C3 A and C3 S after 20 min. These
are thought to be chelate complexes. In experiments with C3 S, the
formation of these complexes resulted in higher hydration heat rates
in the initial minutes. In experiments with C3 A, the initial evolution
of heat of hydration was decreased by PBTC.
In studies by Lipus et al. (2000) and Rickert (2004a), recycling
aids based on PBTC retarded all further hydration reactions of
Portland cement, except for the formation of small primary ettrin-
gite crystals. The new formations (chelate complexes) observed in
Langenfeld et al. (1998) were not observed on the surface of the
cement particles.
In the literature, there are hardly any other references that ad-
dress mechanism of action of PBTC on the hydration of clinker and
cement. The literature yields only a few and to some extent contra-
dictory findings concerning the influence of PBTC on the hydration
of clinker phases, clinker, and cement. In addition, the often very
different experimental conditions and procedures hinder the ability
to draw conclusions regarding the retarding mechanism of PBTC.
Aim and Extent of the Investigations
The aim of the current investigation is to extend the knowledge
concerning the effects of the PBTC-based long-term retarder on the
hydration of clinker phases, clinker, and cement.
Firstly, the effect of the long-term retarder on the hydration of
individual clinker phases (C3 A, C4 AF, and C3 S) and on the hydra-
tion of mixtures of clinker phases, sulphate carriers, and/or calcium
hydroxide, as stated in Table 1, was studied chemically and miner-
alogically.
The retarding mechanism of long-term retarders associated with
the reaction of C3 A or C4 AF fractions with dissolved sulphate to
give ettringite, was investigated by comparative studies between
two industrial clinkers and on cements produced from them by
the addition of sulphate. In this process, it was also necessary to
examine whether the formation of primary ettringite is a necessary
precondition for the further planned retardation of cement pastes by
long-term retarders, or whether long-term retarders can take over
the regulation of the setting of clinker meal in place of sulphate
carrier optimization. In addition, the sorption behavior of the long-
term retarder on clinker and cement was studied.
For a detailed description of the interactions between cement par-
ticles and long-term retarders, the characterization of the chemical
compounds responsible for the long-term retardation is necessary.
Owing to the relatively small quantities of admixtures used and in
particular owing to their strong incorporation and sorption onto ce-
ment hydration products in the hydrating multicomponent system,
such compounds can be detected with difficulty. Consequently, the
compound that leads to the long-term retardation was synthesized
separately, and chemically and mineralogically analyzed.
On the basis of the findings obtained, a model was derived with
which the mechanisms of action of this admixture can be elucidated.
Test Procedure
An overview concerning the composition and designation of
the samples and concerning the studies performed is given in
Table 1.
Description of the Starting Materials
Clinker Phases, Industrial Clinkers, and Cements—The clinker
phases C3 S, C3 A, and C4 AF in the form of granules were ground
finer than 90 µm in a swing mill with a tungsten carbide grind-
ing vessel. Furthermore, calcium hydroxide (CH) p. a. and β-
hemihydrate (CS̄H0.5 ) and anhydrite (CS̄) produced from gypsum
(CS̄H2 ) p. a. were also used.
An unground Portland cement clinker was obtained from each of
two cement works. By Bogue calculation and by X-ray diffraction
with Rietveld analysis, Clinker K I was C3 A-free and contained
about 17.6 mass % C4 AF and about 72.7 mass % C3 S. Clinker K II
had about 8.3 mass % C3 A, 9.2 mass % C4 AF, and about
66.2 mass % C3 S. Each clinker (K I and K II) was ground in a
ball mill to about Blaine 3200 to 3300 cm2 /g. From one portion
of each clinker meal, Cement Z I and Cement Z II, respectively,
were produced by addition of sulphate carriers normally used in
practice. As sulphate carriers, β-hemihydrate (CS̄H0.5 ) and nat-
ural anhydrite (CS̄) were used. For the production of the hemi-
hydrate, gypsum (CS̄H2 ) was appropriately dehydrated. The sul-
phate content as SO3 was adjusted to about 3.2 mass % of the ce-
ment. This corresponded to sulphate dosages for these clinkers in
practice.
Long-Term Retarder
As long-term retarder (LTR) a commercially available long-
term retarder (recycling aid for wash water), based on
 RICKERT AND THIELEN ON LONG-TERM RETARDER 3

TABLE 1—Overview of experiments with clinker phases, clinkers, and cements.

Composition
Designation (mass of solids) Water/Solid1 LTR in Mass %5 Studies

Clinker phases
C3 A-1 C3 A 1.0 0 ESEM, XRD, DSC
C3 A-2 2 Sorption
C3 A-37 5 Pore solution6
C3 A-Ca(OH)2 -1 C3 A:Ca(OH)2 = 3.3 0
C3 A-Ca(OH)2 -2 2
C3 A-SO3 -1 C3 A:SO3 = 2.62 0
C3 A-SO3 -2 2
C3 A-SO3 -37 5
C3 A-SO3 -Ca(OH)2 -1 C3 A:SO3 :Ca(OH)2 = 2.6:1:0.83 0
C3 A-SO3 -Ca(OH)2 -2 2
C3 A-SO3 -Ca(OH)2 -37) 5
C3 S-1 C3 S 0.5 0
C3 S-2 2
C4 AF-17 C4 AF 1.0 0
C4 AF-27 2
C4 AF-SO3 -17 C4 AF:SO3 = 5.34 0
C4 AF-SO3 -27 2
Clinkers
K I-1 KI 0.50 0
K I-2 2
K II-1 K II 0
K II-2 2
Cements
Z I-1 ZI 0
Z I-2 2
Z II-1 Z II 0
Z II-2 2

1 Admixture was included in the calculation of the water.

2 C3 A:CS̄:CS̄H0.5 = 3:1:1.
0.5 : CH = 3:1:1:0.9.
3 C A:CS̄:CS̄H
3
0.5 = 6:1:1.
4 C AF:CS̄:CS̄H
4
5 Relative to solids.
6 Only clinkers and cements.
7 See (Rickert 2004b).

2-phosphonobutane-1,2,4-tricarboxylic acid (PBTC), was used.
The active substance content (PBTC) was about 15 mass %. In
each case, the quantities of LTR given below relate to the respec-
tive mass of solids.
Experimental Procedures
For the studies, pastes were mixed in teflon dishes with a teflon
spatula under argon using fully demineralized water. The admix-
ture was diluted in the mixing water, with the amount of admix-
ture water included in the calculation of the total water content.
The total mixing time was about 2 min. Suitable portions of the
paste were filled into 50 ml polyethylene flasks, flushed with argon
and stored tightly sealed at 20◦ C over water in the desiccator. At
each test time, a part of the sample was analyzed by Environmen-
tal Scanning Electron Microscope (ESEM) and Energy-Dispersive
X-Ray spectrometry (EDX). The hydration of the rest of the sam-
ple was stopped with acetone and diethyl ether. The hydrate phases
were also studied by Differential Scanning Calorimetry (DSC) and
X-Ray Diffraction (XRD). The DSC studies were carried out over a
temperature range of 25–630◦ C with a heating rate of 30 K/min in a
nitrogen atmosphere. The XRD studies were carried out on powder
samples using a Bragg-Brentano diffractometer with CuKα radia-
tion in the angular range from 5 to 53◦ 2θ and with a step width of
0.04◦ 2θ.
The sorption behavior of the long-term retarder on clinker and
cement was determined on suspensions (w/c = 3.75) over a pe-
riod of 7 days by determination of the organic phosphorus content
in filtered pore water. The admixture content was 2 mass % and
5 mass % relative to cement or clinker. The organic phosphorus
content was calculated from difference of total phosphorus and
inorganic phosphorus content.
The setting behavior was determined on the basis of EN 196-3
(Vicat apparatus). Pore solutions were filtered off from fresh cement
paste with a vacuum pump, and pressed out in the case of hardened
cement paste.
By mixing of a solids-free, calcium hydroxide saturated solu-
tion (pH 12.7) with a diluted long-term retarder solution (50%
admixture, 288 mmoles PBTC/l) with constant stirring, the com-
pound responsible for the long-term retardation was synthesized.
During this, the content of long-term retarder was increased step-
wise. The mixture was filtered with the aid of a vacuum pump.
The filter residue was then dried to constant mass at 40◦ C and the
composition investigated by XRD, X-ray fluorescence analysis, IR
4 CEMENT, CONCRETE, AND AGGREGATES
FIG. 2—C3 S-1, w/C3 S = 0.50, at age 28 days, normal hydration, de-
velopment of C-S-H phases—paste hardened.
FIG. 3—C3 S-2, w/C3 S = 0.50, 2 mass % LTR, at age 180 days, no hydra-
tion products—paste can still be processed.
spectroscopy, and ESEM/EDX. The carbon and hydrogen content
was determined by elemental analysis.
Test Results
Influence of LTR on the Hydration of Synthetic Clinker Phases and
on Mixtures Between Clinker Phases and Calcium Sulphate and/or
Calcium Hydroxide
C3 S Hydration —Figs. 2 and 3 show ESEM micrographs of the
samples C3 S-1 and C3 S-2, respectively, at ×10 000 magnification,
at ages of 28 and 180 days respectively.
Comparison of the micrographs shows clearly that the LTR addi-
tion of 2 mass % (Fig. 3) led to complete blocking of the hydration
of C3 S. While in the case of the unretarded sample C3 S-1 (Fig. 2) a
compact structure of sharp needle-shaped C-S-H phases had formed
as a result of hydration, no hydration products could be observed
on the surfaces of the long-term retarded sample C3 S-2 even after
180 days. After this time, the paste displayed a soft consistency,
similar to the freshly mixed paste. The light areas on the particles
in Fig. 3 display higher contents of oxygen, carbon, and phospho-
rus compared to the remaining particle surface, which indicate a
phosphorus-containing organic compound on the particle surface.
In this region, the carbon/phosphorus ratio was between 2.7 and 2.9,
and thus, approximately corresponded to the carbon/phosphorus ra-
tio of the LTR (about 2.7). It must be presumed that this is a region
where the admixture has become concentrated. After a hydration
period of one year, the sample C3 S-2 had hardened. At this age, the
microstructure, in addition to very compact structures due to the
FIG. 4—C3 S-1, w/C3 S = 0.50, DSC-curves.
FIG. 5—C3 S-2, w/C3 S = 0.50, 2 mass % LTR, DSC curves.
progression of the hydration, sporadically displayed relatively
porous regions. In these regions, poorly hydrated C3 S particles
were present.
In addition to the ESEM micrographs, a comparison of the DSC
measurement curves of both samples, shown in Figs. 4 and 5,
also shows that the unretarded sample C3 S-1 after one year had a
markedly higher calcium hydroxide content than the LTR-retarded
sample C3 S-2. This further confirms that the progress of the C3 S
hydration in this sample still remained lower.
C3 A-, C4 AF Hydration—Figs. 6 and 7 show the course of hydra-
tion of the Samples C3 A-1 and C3 A-2, respectively, up to 28 days on
the basis of DSC curves. Figures 8 and 9 show ESEM micrographs
of the Samples C3 A-1 (without LTR) and C3 A-2 (with 2 mass %
LTR), each at the ages of 3 min (top) and 28 days (bottom).
Whereas in Sample C3 A-1 (Fig. 6 and Fig. 8) cubic C-A-H
(C3 AH6 , hydrogarnets) could mainly be seen, in Sample C3 A-2
(Fig. 7 and Fig. 9) due to the addition of 2 mass % LTR in partic-
ular, hexagonal C-A-H, such as for example CAH10 , C2 AH8 , and
 RICKERT AND THIELEN ON LONG-TERM RETARDER 5

FIG. 6—C3 A-1, w/C3 A = 1.00, DSC curves.

FIG. 8—C3 A-1, w/C3 A = 1.00, at age 3 min: formation of hexagonal

and cubic C-A-H, and at age 28 days: conversion of hex. C-A-H to cubic
C-A-H (C3 AH6 - hydrogarnet).

FIG. 7—C3 A-2, w/C3 A = 1.00, 2 mass % LTR, DSC curves.

C4 AH11-13 , were formed. The formation of cubic C-A-H or the

conversion of the hexagonal CAH into cubic CAH was strongly
impeded by LTR. The same was observed with the hydration of
C4 AF in the presence of LTR. X-ray diffraction studies at the age
of 28 days showed that LTR increased the C3 A hydration (C3 A-2).
On the other hand, the hydration of C4 AF was decreased by LTR
(C4 AF-2).
FIG. 9—C3 A-2, w/C3 A = 1.00, 2 mass % LTR, at age 3 min and 28
C3 A-Ca(OH)2 Hydration—If C3 A was hydrated in the presence days, hex. C-A-H predominantly, no conversion after 28 days.
of Ca(OH)2 and LTR (Sample C3 A-Ca(OH)2 -2), the formation of
hexagonal C-A-H was increased. But compared to the system with-
out Ca(OH)2 , the conversion of C3 A was decreased.
retarded. The retardation commencing after the initially acceler-
C3 A-SO3 , C4 AF -SO 3 Hydration—As can be seen from compar- ated formation of ettringite affected all hydration reactions. Even
ison of the DSC curves in Figs. 10 and 11 after a hydration time of the relatively fast reaction of hemihydrate (CS̄H0.5 ) with water to
3 min and 6 h, in sample C3 A-SO3 -2 (Fig. 11), the formation of pri- give gypsum (CS̄H2 ) was retarded. This was indicated by the pres-
mary ettringite (E) was initially accelerated by the LTR for a short ence of hemihydrate in Sample C3 A-SO3 -2 after 28 days. Whereas
time. The further progress of the hydration of C3 A was strongly in Sample C3 A-SO3 -1 after 28 days hexagonal C-A-H (C2 AH8 ,
6 CEMENT, CONCRETE, AND AGGREGATES
FIG. 10—C3 A-SO3 -1, w/solid = 1.00, DSC curves.
FIG. 11—C3 A-SO3 -2, w/solid = 1.00, 2 mass % LTR, DSC curves.
C4 AH11-13 ) and cubic C-A-H (C3 AH6 ), monosulphate (M), and
ettringite (E) were present. A similar situation was observed in the
Sample C4 AF-SO3 -2.
C3 A-SO3 -Ca(OH)2 Hydration—Comparison of the DSC curves
in Figs. 12 and 13 shows that in Sample C3 A-SO3 -Ca(OH)2 -2 all
hydration reactions, with the exception of the formation of ettrin-
gite, were strongly retarded. Increased formation of ettringite due
to the LTR similar to the case of Sample C3 A-SO3 -2, was not ob-
served. After 28 days, hemihydrate was still present in the retarded
Sample C3 A-SO3 -Ca(OH)2 -2.
For all trials with calcium aluminates, no organic compounds
associated with LTR as found on the surface of C3 S could be
detected by EDX. This is due to the enlarged specific surface area
caused by the hexagonal C-A-H.
FIG. 12—C3 A-SO3 -Ca(OH)2 -1, w/solid = 1.00, DSC curves.
FIG. 13—C3 A-SO3 -Ca(OH)2 -2, w/solid = 1.00, 2 mass % LTR, DSC
curves.
Influence of LTR on the Sorption Behavior of Portland Cement
Clinker and Portland Cement
Figure 14 shows the results of the sorption experiments on clink-
ers (K I and K II) and cements (Z I and Z II) with varying addition
amounts and addition times of LTR. It can be seen from the figure
that the long-term retarder was sorbed independently of the clinker
type, cement type, the amount added, and the time of addition.
Shortly after the addition, in all cases the proportion of LTR de-
tectable in the solution was already only about 7% relative to the
original amount added. As hydration progressed, the proportion
sorbed increased further. After 7 days, the long-term retarder was
completely sorbed on the cement and no longer detectable in the
filtered solution.
 RICKERT AND THIELEN ON LONG-TERM RETARDER 7

FIG. 16—K II-2, w/c = 0.50, 2 mass % LTR, DSC curves.

FIG. 14—Sorption of the long-term retarder onto clinkers (K I and K II)

and cements (Z I and Z II) with varying addition amounts and times; 2s: 2
mass % LTR, added with mixing water; 2n: 2 mass % LTR, added 2 min
after addition of mixing water; 5s: 5 mass % LTR, added with mixing water.

FIG. 17—Z II-1, w/c = 0.50, DSC curves.

FIG. 15—K II-1, w/c = 0.50, DSC curves.

Influence of LTR on the Hydration of Portland Cement Clinker

and Portland Cements
Apart from a few details, the influence of LTR on the hydration
of the two clinkers (K I and K II) and the cements (Z I and Z II) pro-
duced from them was almost identical. Therefore, only the results
for Clinker K II and Cement Z II are shown. Figure 15 contains
the DSC curves for the Sample K II-1 without LTR and Fig. 16 the
DSC curves for the Sample K II-2 with 2 mass % LTR.
FIG. 18—Z II-2, w/c = 0.50, 2 mass % LTR, DSC curves.
Comparison of the DSC measurement curves after 3 min shows
that with the Portland cement clinkers, LTR caused a short-term
increase in the formation of hexagonal calcium aluminate hydrates, The comparative studies on Portland cement clinker meal (K I
such as for example C2 AH8 and C4 AH11-13 . As a result, the stiff- and K II) and on Portland cements (Z I and Z II) produced from
ening or setting of the pastes and mortars was accelerated. On the these clearly showed that the sulphate optimization of the cements
other hand, the hydration of the calcium silicates was very strongly is necessary for a long-term retardation. The formation of primary
retarded. ettringite initiated by calcium sulphate is essential for PBTC to
Pastes and mortars made of Portland cements showed a pro- function as a retarder.
longed workability by addition of LTR. Comparison of the DSC During the retardation period, the pore solutions of the retarded
curves for Sample Z II-1 (Fig. 17) with those for the Sample Z II-2 clinker pastes or cement pastes always displayed higher contents of
(Fig. 18) clearly shows that the formation of primary ettringite was calcium or calcium and sulphate than the corresponding unretarded
not affected by LTR. In contrast, the hydration of the silicate phases sample. This was attributable firstly to the briefly intensified hy-
was particularly strongly retarded by LTR. drolysis of the calcium aluminates and secondly to the stabilization
8 CEMENT, CONCRETE, AND AGGREGATES

of submicroscopic crystal nuclei, e.g., of CaSO4 or ettringite, by the

residual LTR present in solution. After the retardation period, the
calcium and sulphate concentration in the pore solution decreased
as hydration recommenced (Rickert 2004b).
Whereas with pure C3 S (Sample C3 S-2, Fig. 5) the LTR re-
sulted in complete blocking of the hydration up to 180 days, all
pastes and mortars produced with clinker and cement showed after
28 days approximately the same hydration progress as the corre-
sponding unretarded reference sample (see Figs. 15 & 16 and 17
& 18). This means that after the retardation a more rapid hydration
set in.

Synthesis of Calcium Phosphonate

The addition of LTR (containing 15 mass % PBTC) to a solids-
free, saturated calcium hydroxide solution (pH 12.7) immediately
produced a milky turbidity. Up to an added amount of about
6.8 mmoles PBTC /l calcium hydroxide solution (Ca/PBTC molar
ratio ≥ 2.5), a white precipitate was deposited. Analyses showed
that this was a sparingly water-soluble X-ray amorphous calcium
phosphonate, which contained approximately 2.5 moles of calcium.
In order to be able to bind this quantity of calcium, the active sub-
stance PBTC must have been present in fully dissociated form (see
Eq 1). The formation of calcium phosphonate presumably occurs
FIG. 19—Effect of LTR on the hydration of Portland cement clinker
according to Eq 2. without or with non-optimized supply of CaSO4 , (K II-2).

C7 H11 O9 P → [C7 H6 O9 P]5− + 5H+ (1)

(PBTC)
[C7 H6 O9 P]5− + 5H+ + 2.5Ca2+ + 5OH−
←→ Ca2.5 [C7 H6 O9 P] · 3H2 O + 2H2 O (2)
(Calcium phosphonate)

On further addition of LTR, the Ca/PBTC molar ratio fell below

2.5 and soluble calcium phosphonate complexes of lower calcium
content were formed. Hydrating cement makes available an almost
inexhaustible supply of calcium, so that it can be assumed that
on addition of LTR to clinker, cement or calcium aluminates or
calcium silicates, sparingly soluble calcium phosphonate always
forms in accordance with Eq 2. This is most likely responsible for
the long-term retardation on account of its insolubility.

Mechanisms of Action of the PBTC-Based LTR

From the information gained in these studies, the main mecha-
nisms of action of the LTR can be derived. By way of example,
Figs. 19 and 20 respectively show the mechanisms of action of
the LTR on Portland cement clinker (K II-2) and Portland cement
(Z II-2) schematically.
In an alkaline medium, PBTC (C7 H11 O9 P) releases 5 protons FIG. 20—Effect of LTR on the hydration of Portland cement with opti-
and is present in a completely dissociated form ([C7 H6 O9 P]5− ). For mized supply of CaSO4 , (Z II-2).
the immediate formation of sparingly soluble calcium phosphonate
(Ca2.5 [C7 H6 O9 P]), 2.5 moles of calcium are needed per mole of
PBTC. If LTR is added to Portland cement clinker without calcium ions released during this brief acceleration phase are bound by
sulphate carrier (Fig. 19), the calcium content in the pore solution PBTC.
is not sufficient to form sparingly soluble calcium phosphonate. The sparingly soluble calcium phosphonate (Ca2.5 [C7 H6 O9 P])
Because of the calcium demand of PBTC, there is generally a thus formed coats the surfaces of the individual clinker particles
short-term accelerated hydration of reactive clinker phases, such as differently. On the surfaces of relatively unreactive clinker phases,
for example C3 A. This causes an increased formation of hexagonal such as for example C3 S and C2 S, which are still largely free
C-A-H and results in a marked stiffening or setting. The calcium from hydration products when the sparingly soluble compound is

formed, essentially complete coating with calcium phosphonate
can take place. As a result, the further access of water (H) or
further ion exchange are markedly hindered and the hydration is
retarded. Already hydrated surfaces, such as the C3 A, display a
relatively large specific surface area, which cannot be completely
covered with sparingly soluble calcium phosphonate. The access
of water and ion exchange cannot be completely blocked. For this
reason, compared to the less reactive clinker phases, the hydration
of reactive calcium aluminates is only slightly retarded by calcium
phosphonate formed from LTR.
Because of the sulphate carrier (CS̄Hx ) contained in the Portland
cement (Fig. 20), the concentration of calcium and sulphate in
the pore solution is much higher than in system containing only
clinker. A thin layer of primary ettringite can form on the surface
of the calcium aluminate. Moreover, sufficient calcium ions are
available for the immediate formation of sparingly soluble calcium
phosphonate. The calcium phosphonate coats both the surfaces of
the clinker particle and also those of the still undissolved sulphate
carrier. As a result, further hydration is strongly blocked.
The formation of primary ettringite and the coating of the sur-
faces with sparingly soluble calcium phosphonate constitute the
precondition for planned prolongation of the workability by the
long-term retarder. This means that sulphate carrier optimization is
not only the basis for the control of the setting of the cement, but
also a precondition for the desired mode of action of the admixture.
Conclusion
By systematic studies, the state of knowledge concerning the
effects of a long-term retarder (LTR) based on 2-phosphonobutane-
1,2,4-tricarboxylic acid (PBTC) on the hydration of clinker phases
(C3 S, C3 A, and C4 AF), clinker and cement has been substantially
extended.
The following conclusions were drawn from the research results:
r In the hydration of pure C3 A and C4 AF, the retarder results
in increased formation of hexagonal calcium aluminate hy-
drate (C-A-H). Apart from this, the conversion of hexagonal
C-A-H into cubic hydrates is strongly retarded. The conver-
sion of C3 A is increased, and that of C4 AF decreased.
r The hydration of pure C3 S was completely blocked for up to
180 days.
r The retarder action is attributable to the formation of sparingly
soluble calcium phosphonate (Ca2.5 [C7 H6 O9 P] · xH2 O). The
formation of sparingly soluble calcium phosphonate requires
2.5 moles Ca2+ per mole of PBTC. If enough Ca2+ is available
in the pore solution, sparingly soluble calcium phosphonate
precipitates and forms a closed layer around the cement parti-
cles, which retards further hydration immediately. Conversely,
a lack of Ca2+ causes a short-term acceleration in hydration
of the reactive clinker phase C3 A. Due to the formation of
C-A-H the specific surface area increases. Therefore, the for-
mation of a closed layer of calcium phosphonate is impeded
and the hydration of the calcium aluminates progresses.
r With Portland cements, owing to the content of calcium sul-
phate, sufficient quantities of dissolved calcium are present
for the direct formation of calcium phosphonate. The forma-
tion of primary ettringite and the coating of the surfaces with
sparingly soluble calcium phosphonate are the precondition
for planned prolongation of the workability by the long-term
retarder. If calcium sulphate is absent or present in insufficient
amount, for example in the case of Portland cement clinkers,
RICKERT AND THIELEN ON LONG-TERM RETARDER 9
the calcium demand of the LTR active substance PBTC re-
sults in a short-term acceleration of the hydration of reactive
clinker phases, e.g., C3 A.
r The direct formation of an ettringite layer on the cement par-
ticle and the simultaneous supply of calcium ions are not only
important for controlling the setting of the cement but also for
the desired mode of action of the long-term retarder.
References
Bayer, A. G., 1998, BAYHIBIT-Anwendungen, Wirkungen, Eigen-
schaften, Bayer AG, Anorganische Industrieprodukte,
Leverkusen, Germany, 4.
Bier, A. W., 1993, Computer-Einsatz in der Analytischen Chemie
zur Untersuchung von Protolyse- und Komplexbildungs-
gleichgewichten am Beispiel der Phosphonocarbonsäuren.
Heinrich-Heine-Universität Düsseldorf, Mathematisch-
Naturwissenschaftliche Fakultät, Diss.
Bobrowski, G. S., Guecia, G. C. J., Lupyan, D. A., and Kinney, F. D.,
1990, Methods and Composition for Reclaiming Concrete,
United States Patent 4,964,917.
Bobrowski, G. S., Guecia, G. C. J., Lupyan, D. A., and Kinney, F. D.,
1996, Wiederaufbereitung von Betonmischungen, German
Patent DE 3727907.
Bobrowski, G. S. and Ryan, R. J., 1991, Method for Stabi-
lizing Concrete Residues, European Patent 0 324 501
B1.
Bodamer, A., 1997, “Hydratationsgesteuerter Beton im Tunnel-
bau,” Strassen- und Tiefbau, 51, No. 2, pp. 14–17.
Bürge, T. A. and Mäder, U., 1992, Abbindeverzögerer für das
Nassspritzverfahren, European Patent 0 508 158 A2.
Deutsche Bauchemie e.V., 1999, Betonzusatzmittel und Umwelt,
Sachstandsbericht, Mai 1999, 3. Ausgabe, BAUCOM,
Boehl-Iggelheim.
DIBT, 2000, Allgemeine bauaufsichtliche Zulassung, Nr. Z-3.210-
1199, DIBT-Berlin.
Fischer, H., 1991, “Steuerung der Zementhydratation von Spritz-
beton,” Schweizer Baublatt, No. 4.
Goetz, H. W., 1967, “False Set of Cement as Influenced by
Hydroxylated-Carboxylic-Acid-Type Admixture,” Mate-
rial Research & Standards, June, pp. 246–249.
Kinney, F. D., 1998, Reuse of Returned Concrete by Hydration
Control: Characterization of a New Concept, Third CAN-
MET/ACI Conference “Superplasticizers and Other Chem-
ical Admixtures in Concrete,” Ottawa, Canada, SP 119-2,
pp. 19–40.
Langenfeld, M. and Stark, J., 1998, “Der Einfluss von Verzögerern
auf die frühe Hydratation von Portlandzementklinker-
phasen, dargestellt in einem ESEM-FEG”, Thesis, Vol. 44,
No. 1, pp. 82–90.
Lewandowski, R., 1983, “Anforderungen des Transportbeton-
Herstellers an Betonzusatzmittel,” Beton, Vol. 33 , No. 8,
pp. 285–288.
Lipus, K., Rickert, J., and Sylla, H. -M., 2000, ESEM-Unter-
suchungen zur Hydratation von Portlandzement unter Ein-
fluß von verzögernden Zusatzmitteln, 14th ibausil, Weimar,
Bauhaus-Universität Weimar, pp. 301–313.
MBT, 1993, DELVO Betonwaschwasser-Recycling, Master Builder
Technologies (MBT), Brochure.
Meyer, L. M. and Perenchio, W. F., 1979, “Theory of Concrete
Slump Loss as Related to the Use of Chemical Admixtures,”
Concrete International, pp. 36–43.
10 CEMENT, CONCRETE, AND AGGREGATES
Moeser, B., Langenfeld, M., and Stark, J., 1999, Frühe Hydratation
von Portlandzement unter Zusatzmitteleinfluss dargestellt
im ESEM-FEG. 37th DAfStb-Forschungskolloquium, Inst.
f. Konstruktiven Ingenieurbau F. A.-Finger-Institut für
Baustoffkunde, Weimar, pp. 123–134.
Okawa, Y., Yamamiya, H., and Nishibayashi, S., 2000, “Study
on the Reuse of Returned Concrete,” Mag. of Concr. Res.,
Vol. 52, No. 2, pp. 109–115.
Paolini, M. and Khurana, R., 1998, “Admixtures for Recycling of
Waste Concrete,” Cement Concrete Composites, Vol. 20,
No. 2/3, pp. 221–229.
Persinski, L. J., Martin, F. D., and Adams, S. L., 1976, Well
cementing composition having improved flow properties,
containing phosphonobutane tricarboxylic acid, and method
of use, United States Patent 3,964,921.
Ramachandran, V. S., 1995, Concrete Admixtures Handbook: Prop-
erties, Science and Technology, 2nd ed., Noyes Publications,
NJ.
Richartz, W., 1983, “Einfluss von Zusätzen auf das Erstarrungsver-
halten von Zement,” Beton, Vol. 33, No. 11, pp. 425–429
and No. 12, pp. 465–471.
Rickert, J., 2004a, “Influence of Retarders on the Hydration of
Clinker and Cement,” Concrete Technology Reports 2001–
2003, Verlag Bau+Technik, Duesseldorf, 2004.
Rickert, J., 2004b, Zum Einfluss von Langzeitverzögerern auf
der Basis von Phosphonsäure auf die Hydratation einzel-
ner Klinkerphasen, Portlandzementklinker und Portlandze-
mente, Bauhaus-Universität Weimar, Dissertation, (http://
e-pub.uni-weimar.de/volltexte/2004/23/).
Seiler, K., 1969, Die Wirkungsweise von Erstarrungsverzögerern
unter besonderer Berücksichtigung silicofluorhaltiger
Verbindungen. RWTH Aachen, Dissertation.
Skalny, J. and Young, J. F., 1980, Mechanisms of Portland Ce-
ment Hydration; Proceedings, 7th International Sympo-
sium Chemical of Cement, Paris, 1980, S. II 1, pp. 3–
45.
Wenquan, L., Huizhen, L., Zhen, H., and Yajie, L., 1997,
Use of High Range Retarder-Plasticizer for Roller-
Compacted Concrete of the Three Gorges Project in
China, Fifth CANMET/ACI Superplasticizers and Other
Chemical Admixtures in Concrete, Rome, Italy, pp. 835–
845.