A Study of Portland Cement Hydration
by Paramagnetic Iron Suppression of Proton Magnetic Resonance
J. C. MacTavish. L. M i l j k o v i ć * . L. J. Schreiner, and M. M. P i n t a r
Guelph-Waterloo Program for Graduate Work in Physics, Waterloo Campus, University of
Waterloo, Waterloo. Ontario, Canada N2L 3G1
R. Blinc and G. L a h a j n a r
Institute J. Stefan. University E. Kardelj, Ljubljana, Yugoslavia
Z. Naturforsch. 40a, 3 2 - 3 6 (1985); received September 3, 1984
The dynamics of the hydration of a white cement with negligible iron content and of Portland
cement with a considerable amount of iron has been studied by proton magnetic resonance. The
presence of iron in Portland cement results in the protons close to the free electron spins of iron
being out of resonance and their signal being suppressed. The evolution in time of the percentage of protons out of resonance in a Portland cement and the growth of the solid component have
been determined as a function of hydration time.
I. Introduction
Nuclear magnetic resonance ( N M R ) has o f t e n
been used to study the hydration of v a r i o u s systems
[ 1 - 3 ] , particularly of cements [ 4 - 7 ] . T h e effects of
paramagnetic iron suppression of the p r o t o n N M R
in cements, however, have not yet b e e n c o n s i d e r e d .
A comparative N M R study of a w h i t e c e m e n t
(negligible iron content) and a P o r t l a n d c e m e n t
(containing F e 2 0 3 ) was therefore u n d e r t a k e n .
The total a m o u n t of water in a c e m e n t paste at a
given hydration time can be m o n i t o r e d by w e i g h i n g
the sample. N M R determinations of the total p r o t o n
magnetization through m e a s u r e m e n t of the F r e e Induction Decay ( F I D ) should yield the a m o u n t of
"liquid like" water present at a given h y d r a t i o n
time. However, p a r a m a g n e t i c centres in the s a m p l e
may make some of the protons u n o b s e r v a b l e . N M R
techniques such as F I D m e a s u r e m e n t s a n d spin
echo measurements using variations of the C a r r Purcell (CP) pulse sequences, can also give i n f o r m a tion on the spin-spin relaxation times, T2, of the
protons in the cements. These time constants can b e
used to characterize the environment of the p r o t o n s
(see Table 1) for some examples). In fact, the C P
technique can also be used to m e a s u r e directly the
* On leave from the University of Nis, Nis. Yugoslavia.
Reprint requests to Prof. G. Lahajnar, Institute J. Stefan.
University E. Kardelj. Jamova 39. POB 53. 61111 Ljubljana. Jugoslawien.
Table 1. NMR relaxation times for protons in various
materials [4],
State of water
T 2 (US)
Bulk water
Physically absorbed water
Zeolitic water
Interlayer-clay water
Ice
Hydrogen in calcium hydroxide
2 500 000
1000- 3000
100- 1000
100- 1000
7
4- 7
fraction of water which exists in a " l i q u i d - l i k e "
environment (i.e., with T2 > 1 ms) at a given h y d r a tion time.
II. Experiment
The dry type 10 normal Portland and white
cements with compositions given in T a b l e 2, were
obtained from C a n a d a C e m e n t Lafarge Ltd. (Belleville. Ontario). Doubly distilled water was mixed
with the dry cement p o w d e r in the ratio 0.42 g
water/g dry cement. T h e cement paste was then put
into 8 m m O D glass N M R tubes. To s i m u l a t e hydration under normal conditions the s a m p l e tubes
were left open for the d u r a t i o n of the experiments.
The N M R measurements were p e r f o r m e d with a
Bruker SXP pulse spectrometer (Bruker Instruments
Inc.. Billerica Mass.) operating at a f r e q u e n c y of
40 MHz. All experiments were p e r f o r m e d at 20 ° C .
The spin-spin relaxation m e a s u r e m e n t s were
m a d e using two techniques. T h e F I D was m e a s u r e d
0340-4811 / 85 / 0100-0032 $ 01.30/0. - Please order a reprint rather than making your own copy.
Unauthenticated
Download Date | 6/15/17 9:12 PM
J. C. MacTavish et al. • A Study of Portland Cement Hydration
Table 2. Oxide composition of Lafarge Canada Portland
No. 10 and white cement powders [13].
Oxide
CaO
Si0 2
AI 2 O 3
S03
MgO
Fe,03
K20
L.O.I.
Free CaO
Na,0
TiO,
SrO"
p2o5
well as the evolution in t i m e of the p e r c e n t a g e of
protons out of resonance in the Portland c e m e n t d u e
to paramagnetic iron suppression is d e p i c t e d . In
Fig. 4 shows the h y d r a t i o n t i m e - d e p e n d e n c e of the
T2 for the two samples.
Weight percent
Portland No. 10
White cement
63.38
20.94
4.60
3.89
2.33
2.05
1.06
1.05
0.60
0.52
0.17
0.14
0.12
65.65
23.12
4.74
3.37
0.88
0.03
0.08
1.69
0.07
0.04
0.21
0.04
0.08
directly after a 9 0 ° r.f. pulse; this gave T2 of t h e
solid. T h e F I D was recorded at 110 positions along
the decay. Carr-Purcell M e i b o o m - G i l l ( C P M G )
modified spin-echo experiments [8] were also performed to determine T2 and the m a g n e t i z a t i o n
fraction of protons in the liquid-like e n v i r o n m e n t ,
with T2 > 1 ms. Typically 100 signals were a v e r a g e d
using a Biomation 805 transient w a v e f o r m r e c o r d e r
(Biomation; Cupertina, CA). T h e true e q u i l i b r i u m
magnetization for the samples was d e t e r m i n e d by
Fitting the F I D and extrapolating back to zero t i m e
immediately following the r.f. pulse. T h e m a g n e t i zation fraction for the liquid-like spins was obtained from the t = 0 extrapolation of the spin e c h o
decay normalized to the total m a g n e t i z a t i o n as
determined by the total F I D . recorded with the first
9 0 ° pulse of the C P M G sequence. T h e F I D a n d
C P M G experiments were p e r f o r m e d at v a r i o u s
hydration times. T h e weight loss of the s a m p l e s was
also recorded. T h e observed total m a g n e t i z a t i o n s
and weights of the samples d u r i n g h y d r a t i o n w e r e
normalized to those measured at the start of h y d r a tion; the magnetization of the liquid-like p r o t o n s
was normalized to the total m a g n e t i z a t i o n determined from the F I D experiments.
33
Hydration t i m e [ h o u r s ]
Fig. 1. The weight loss, and the magnetization loss in the
FID and CPMG measurements for Portland # 10 cement.
c
B
2
III. Results
T h e weight loss, and the m a g n e t i z a t i o n loss f r o m
the F I D and C P M G measurements, for the P o r t l a n d
# 10 cement and the white cement as they h a r d e n e d ,
are shown in Figs. 1 and 2, respectively. In Fig. 3 t h e
growth of the solid c o m p o n e n t of these c e m e n t s as
Hydration
time [ hours]
Fig. 2. The weight loss, and the magnetization loss in the
FID and CPMG measurements for white cement.
Unauthenticated
Download Date | 6/15/17 9:12 PM
J. C. MacTavish et al. • A Study of Portland C e m e n t H y d r a t i o n 34
34
IV. Discussion
a -
Portland * 1 0 cement - Paramagnetic
y -
W h i t e cement -
ß
P o r t l a n d *10 cement - v i s i b l e s o l i d
-
solid
6
•out of
EXP. ERROR
solid
resonance
t37„
B-C_PortJand._
./JTI
/
r >HV_
a
"ß
v
/
/
/
/
/
/
/
/
X
/
/
5
I I ] N«^ I I I I I I I I I . I I ! ! I I I I I I ! I
0
5
10
15
20
25
30
Hydration time [ h o u r s ]
Fig. 3. Fractions of the proton magnetization of the solid
environments of the cements and fraction of protons out of
resonance in the Portland cement sample vs. hydration
time. The curves A. B, and C are shown in Figs. 1 and 2.
White
J
i
[
I
I
10
1
L
i
i
i
10 c e m e n t
i
i
i
i
i i
30
Hydration t i m e
As hydration progresses, m o r e and m o r e w a t e r
becomes associated with rigid e n v i r o n m e n t s as the
crystallization of gel c o m p o u n d s increases (Fig. 3,
curves a and y). N o t e that this growth a p p e a r s to b e
negligible until after ~ 1 h in Portland and ~ 3 h in
white cement. In Portland cement a d o r m a n t p e r i o d
is apparent between 5 and 9 h h y d r a t i o n time. A p a r t
from these divergences the solid growth a p p e a r s to
be smooth in both cases, showing no structure. In
white cement at 30 h hydration t i m e the solid c o m ponent makes up ~ 35% of the total m a g n e t i z a t i o n
while the "visible solid" c o m p o n e n t of Portland
cement makes u p ~ 40%.
The samples were kept in open tubes with a f r e e
contact area of ~ I cm 2 and, therefore, w a t e r was
able to evaporate d u r i n g the h a r d e n i n g process.
This is evident f r o m the weight-loss curves of
Figs. 1 and 2. T h e rate of water loss f r o m the w h i t e
cement sample was s o m e w h a t greater t h a n that of
the Portland during the first 30 h, p a r t i c u l a r l y
during the first 10 h. This implies that the a m o u n t
of water available for evaporation in Portland
cement decreased m o r e rapidly with t i m e t h a n t h a t
in white cement, as supported by the greater rate of
structural development in Portland c e m e n t indicated in Figure 3. T h e a m o u n t of water loss in
both cements a p p e a r s to a p p r o a c h ~ 30% at large
hydration times.
cement
Portland *
Prior to the addition of water, both the P o r t l a n d
No. 10 and white cement clinkers were f o u n d to
contain (4 ± 3)% water by mass in a solid-like environment, and no water in a liquid-like environment. A few minutes after the beginning of t h e
hydration (time t = 0) the percentage of p r o t o n s in
the solid milieu seemed to be (20% ± 3%) for the
Portland cement and (10% ± 3%) for the w h i t e
cement (cf. Figure 3). This difference, however, is to
be explained by the presence of iron in the Portland cement since any protons in close p r o x i m i t y
with iron free electron spins in a solid or liquid
matrix will not be observed. Evidently u p to ~ 10%
of the protons in the Portland cement e x p e r i e n c e
such an effect at time t ~ 0.
[hours]
Fig. 4. The evolution in time of T2 as measured by the
CP method. The solid T2, which is on the order of 10 (as, is
essentially constant and is not shown.
T h e loss of signal f r o m protons in the liquid-like
environment in Portland cement (Fig. 1, curve C) is
extremely different f r o m the c o r r e s p o n d i n g loss in
white cement (Fig. 2, curve C). F o r e x a m p l e , at 30 h
hydration time the liquid-like proton signal in Portland cement is only a b o u t 7% of the initial p r o t o n
Unauthenticated
Download Date | 6/15/17 9:12 PM
J. C. MacTavish et al. • A Study of Portland Cement H y d r a t i o n
N M R signal. On the other hand the liquid w a t e r
signal in white cement amounts to still 40% at 30 h.
T h e difference between the liquid water signals
f r o m these two samples is the result of the strong
magnetic interaction [9] between an Fe a t o m ' s electron spin and the water in its vicinity. As a result of
this interaction, the water proton L a r m o r f r e q u e n c y
becomes vastly different from the Proton L a r m o r
frequency in the external field H0 (a>0 = yH0 w h e r e
y is the proton gyromagnetic ratio); it is s h i f t e d
considerably into the wings of the N M R a b s o r p t i o n
spectrum. Therefore some of this water is not even
excited by the r.f. irradiation at w 0 . H o w e v e r ,
because a spread in this frequency occurs as a result
of the short (high power) 90° pulses used, s o m e of
this paramagnetically suppressed water m a y still b e
excited. The FID of such water t h o u g h would occur
in a fraction of a microsecond, and would thus be lost
in the dead time
8 (is) of the receiver of t h e
N M R spectrometer. Therefore, the paramagnetically
suppressed water is simply not observed.
As a result of the extreme loss of this liquid w a t e r
proton N M R signal (Fig. 1, curve C), the P o r t l a n d
cement paste exhibits a dramatically d i f f e r e n t total
proton signal intensity (Fig. 1, curve B). T h e d i f f e r ence between these two curves (Fig. 3, curve ß)
represents the a p p a r e n t relative intensity of t h e
N M R signal of protons in solid-like e n v i r o n m e n t s .
This function shows structure with a peak at 7 h,
and an apparent acceleration of the rate of g r o w t h
of the solid-like phase between 14 and 18 h h y d r a tion time. The unusual character of this f u n c t i o n
reflects the transfer of F e between the solid-like a n d
liquid-like environments. The percentage of p r o t o n s
that are out of resonance and hence u n o b s e r v a b l e is
contingent primarily u p o n the a m o u n t of iron in t h e
liquid-like milieu, a n d is depicted in Fig. 3, c u r v e <5.
T h e peak in this curve at 4 h h y d r a t i o n t i m e indicates that between approximately 4 and 7 h, m o r e
water was finding its way into a solid m a t r i x of
relatively low iron content than iron was dissolving
into solution. D u r i n g this period t h e r e f o r e , relatively rapid hydration of e.g. the calcium silicate
phases must occur in an environment isolated f r o m
iron.
This may correspond to the so-called stage of
inter-granular fibrillar growth [10] d u r i n g w h i c h
finger-like protuberances of both i r o n - f r e e and ironcontaining hydrates [11] crystallize on the s u r f a c e of
the cement grains.
35
As there are no p a r a m a g n e t i c i m p u r i t i e s in white
cement, its solid proton c o m p o n e n t (Fig. 3, curve y)
reflects directly the a m o u n t of solid present at a
given time. T h e growth of this c o m p o n e n t is a
monotonous function of time, depicting the fastest
rate between 5 and 10 h, and a p p a r e n t cessation of
growth at about 26 h h y d r a t i o n time.
The time-evolutions of T2 of the two samples
differ markedly (Figure 4). T h e T2 of w h i t e c e m e n t
decreases gradually f r o m a b o u t 5 ms at t % 0 to
about 0.5 ms at 30 h h y d r a t i o n time. Its s h a p e is
correlative of that of the solid proton growth curve of
Fig. 3 (curve y) during this time. T h a t T2 is only
5 ms at t ~ 0 indicates that there is no bulk water
present in the white cement paste, since bulk water
has a T2 on the order of a second. T h e water is
however still in a liquid-like state but with the
correlation times for translational and rotational
diffusion longer by a b o u t three orders of m a g n i tude, as the relaxation rate for water is p r o p o r t i o n a l
to the correlation time.
V. Conclusions
Paramagnetic iron suppression of p r o t o n magnetic resonance a u g m e n t s the capacity of N M R to
non-invasively monitor the evolution in t i m e of the
constituent solid and liquid c o m p o n e n t s of hydrating cement pastes. T h e effects of p a r a m a g n e t i c
impurities on the results of N M R studies of c e m e n t
are profound and have hitherto not b e e n observed.
The dynamics of h y d r a t i o n of the two cements
studied as seen by p a r a m a g n e t i c a l l y suppressed
proton N M R differed markedly. S o m e i m p l i c a t i o n s
of these differences have been discussed in the text.
Paramagnetic iron suppression of p r o t o n m a g n e t i c
resonane can be a useful tool in e n h a n c i n g the N M R
resolution of heterogeneous systems a n d in N M R
imaging. In order to better investigate the subtleties
of cement hydration a c o m p a r a t i v e study of white
and Portland cements by N M R m e t h o d s [12] is
currently underway.
Acknowledgements
The authors wish to t h a n k Dr. H u n g C h e n f r o m
Canada Cement L a f a r g e Ltd. for s u p p l y i n g the
cement samples and for h e l p f u l discussions.
Unauthenticated
Download Date | 6/15/17 9:12 PM
36
[1] H. Peemoeller and M. M. Pintar. Biophys. J. 28, 339
(1979).
[2] R. Mathur-De Vre et al.. Rad. Res. 90,441 (1982).
[3] C. J. G. Bakker and J. Vriend. Phys. Med. Biol. 29,
509 (1984).
[4] P. Seliaman, J. Res. Dev. Lab. Portland Cem. Assoc.
10,52 (1968).
[5] J. R. Zimmerman and W. E. Brittin, J. Phys. Chem.
61,1328 (1957).
[6] G. Lahajnar et al., Cem. Concr. Res. 7, 385 (1977).
[7] R. Blinc et al.. Amer. Ceram. Soc. 61, 35 (1978).
[8] E. Fukushima and S. B. W. Roeder, Experimental
Pulse NMR. Chapter 1, Addison-Wesley Publishing
Company, Reading, Mass. 1981.
J. C. MacTavish et al. • A Study of Portland Cement Hydration 36
[9] A. Abragam. The Principles of Nuclear Magnetism,
Chapter 6, Clarendon Press, Oxford 1961.
[10] J. D. Birchall. A. J. Howard, and J. E. Bailey. Proc.
Roy Soc. London A360,445 (1978).
[11] P. Barnes, A. Ghose, and A. L. Mackay, Cem. Concr.
Res. 10,639 (1980).
[12] L. J. Schreiner et al., N M R Line Shape-Spin Lattice
Relaxation Correlation Study of Portland Cement
Hydration, in press.
[13] Canada Cement Lafarge Ltd., Belleville, Ontario,
Canada.
Unauthenticated
Download Date | 6/15/17 9:12 PM