Concrete - Geomaterials: Under the Microscope - page 84

workable enough to compact. The water of hydration
chemically binds in the cement hydrates, while the water
for workability occupies a system of capillary pores
(around 1 μm in size) within the concrete (with some being
lost by bleeding and evaporation). As a general rule, the
higher the W/C, the more microporous and consequently
the less durable, the hardened concrete will be.
The W/C of hardened concrete is usually determined
either as part of routine condition assessment for
concrete structures, or as an aid to determining the cause
of unsatisfactory concrete performance. It is of particular
importance where poor concrete performance is
suspected or alleged to have been caused by incorrect or
improper W/C, for example where too high a W/C has
resulted in inadequate compressive strength. Although
there is no internationally recognized test, original W/C
is most commonly determined using indirect chemical
analysis in accordance with methods such as that
described in BS 1881: Part 124 (British Standards
Institution, 1988). The original W/C is obtained by
combining determined values for the water bound in the
cement hydrates and the capillary porosity. The chemical
method is unsuitable for air-entrained, cracked,
carbonated, or poorly compacted concrete and even in
favourable circumstances (W/C in the range 0.4-0.8) it
only has an accuracy and reproducibility of within 0.1 of
the actual W/C (Concrete Society, 1989).
The relative inaccuracy of the chemical methods has
led to the development of alternative petrographic
methods. These are direct methods based on petrographic
indicators of apparent W/C. In general as W/C increases,
the hardened cement paste exhibits a higher
microporosity, a lower concentration of residual
unhydrated cement grains, and more (and larger)
portlandite crystallites. For normal Portland cement
concrete, an experienced petrographer should be able to
use these microscopical features to estimate visually
apparent W/C to one of the following broard ranges:
Apparent W/C

prepared from concretes with a range of W/C (e.g. 0.30,
0.35, 0.40, 0.45, 0.50, 0.55, 0.60, 0.65, 0.70, 0.75, 0.80),
but with similar ingredients to the sample under
investigation. With reference specimens available there
are two methods used for W/C determination. The most
commonly used involves observing the relative
fluorescence using fluorescent microscopy and the other
method involves observing the relative amounts of
residual unhydrated cement grains.
Apparent W/C determination by fluorescence
microscopy is widely used and a standard method is
available, which is described in Nordisk NT 361-1999.
The sample is first impregnated using fluorescent yellow
dye and a thin section of the sample is prepared to a
thickness of 20-25 μm (rather than the standard 30 μm).
The thin section is subsequently examined micro-
scopically in fluorescent light (either transmitted or
reflected). The intensity of the fluorescence emitted from
the cement paste is proportional to the amount of
intruded resin, which is in turn related to the capillary
porosity and the W/C. Figures 179-181 illustrate the
typical relative fluorescence emitted by hardened
concrete samples with W/C in the low, normal, and high
ranges respectively. The cement paste may appear
homogeneous or show large variation in microporosity.
Large variations can result from improper mixing or
bleeding. Figure 182 shows concrete that exhibits a large
variation in microporosity of the cement paste making
determination of the apparent W/C challenging. In
Denmark, an image analysis software package (W/C-
check) has been developed for the determination of
apparent W/C from camera images of fluorescent
impregnated thin sections when viewed under the
microscope ( 183 ).
The accuracy of the fluorescence microscopy method
of determining W/C is believed to be at least as good as,
and usually better than, the chemical analysis method.
Jakobsen et al ., (2003) claim an accuracy of ±0.02 but
this is likely to be achievable in only the most favourable
circumstances with excellent reference samples. For
unknown concrete from existing structures the accuracy
is more likely to be within 0.1 (Neville, 2003). It is not yet
clear whether the relative fluorescence of the dyed resin
used for impregnation of W/C reference specimens will
degrade with time as the thin sections are stored. Sibbick
et al . (2007) report that they have found no change in
their reference specimens after 15 years.
Another direct method for determining W/C described
by French (1991a) involves the determination of the
number of residual clinker particles per millimetre
traverse of paste. The sample is compared with reference

Range

<0.35 Low

0.35-0.65 Normal

>0.65 High

This approximate determination of W/C is often
sufficient as the petrographer is usually only required to
establish whether or not a particular concrete had a W/C
that significantly exceeded the target value (St John et
al ., 1998).
If more accurate W/C determination is required, for
example in the case of a dispute, it is necessary to
compare the thin section with reference specimens