Alkali–silica reaction in concrete Digest 330

Detailed guidance for new construction Part 2 2004 Edition

digest
BRE Centre for Concrete Construction

Concrete can deteriorate as a result of an interaction between alkaline

pore fluids (principally originating from Portland cements) and reactive
minerals in certain types of aggregates.
The mechanism of deterioration is known as alkali–aggregate reaction
(AAR); it can occur in a number of forms, the most common being
alkali–silica reaction (ASR).
This Digest is in four parts
Part 1 gives the background to the detailed and simplified guidance
contained in Parts 2 and 4. Part 2 gives detailed guidance for minimising
the risk of damaging ASR in new construction. Part 3 gives worked
examples. Part 4 gives simplified guidance for new construction using
aggregates of normal reactivity.
Advice on the prevention of ASR caused by opal, glass, calcined flint and
other forms of extremely reactive material is outside the scope of this Digest.

Concrete core surface showing divergent ASR induced expansive

microcracking centred on quartzite coarse aggregate particles

Using this Digest

Table 1 gives recommendations on minimising the risk of
damaging ASR in concrete. It gives advice for cements
and combinations, including ground granulated
blastfurnace slag (ggbs), pulverised fuel ash (pfa) and fly
ash, used with aggregates classified in three reactivity
classes.
Several concepts are embodied in the table: they are
elaborated in the text. The following notes should be
studied before using the guidance in Table 1.
The recommendations are underpinned by a range of
more general considerations; these apply:
a to concrete with contents of cement or
combinations, in general to a maximum of 550
kg/m3 (Table 8 carries the exceptions);
b to cement conforming to British or European
Standards (Table 6) where the ‘declared mean alkali
content’ of the CEM I-type component is declared by
the manufacturer to enable classification as set out
in Table 1 to low, moderate or high alkali.
The manufacturer, on request, will provide the
‘declared mean alkali content’ of the cement in
accordance with BS EN 197-1 NB4;
c to cement, ggbs or pfa where the term ‘declared
mean’ is a declaration by the manufacturer that the
value declared will not be exceeded without prior
notice;
d to CEM I cement or the CEM I-type component of
cement where the term ‘guaranteed ≤ 0.60%
Na2O eq’ describes a guarantee by the manufacturer
that no test result, obtained on any spot sample, will
exceed the stated limit;
e where the alkali content of any cement is derived from
the results of the manufacturer’s latest 25
consecutive determinations;
f to ggbs to BS 6699 where the acid soluble alkali
content is ≤ 1.0% Na2O eq;
g to pfa to BS 3892-1 or EN 450 where the total acid
soluble alkali content is ≤ 5.0% Na2O eq;
h to combinations, defined as a mixture of BS EN 197-1
CEM I cement with ggbs to BS 6699 or to
pfa to BS 3892-1 or EN 450, at recommended target
mean proportions produced at the concrete mixer;
i to proportions of cement, ggbs or pfa, given as target
mean values, derived from the required batch weights
and varying by no more than ±3% from the
recommended values.

A BRE research project funded by

DTI Construction Sector Unit
2

Table 1 Recommended limits for alkali contents of concrete

Aggregate type Alkali content of the CEM I-type component of the cement (Table 6) or the CEM I component of a
or combination combination with ggbs or pfa
Low alkali Moderate alkali High alkali
(guaranteed ≤ 0.60%Na2O eq (declared mean (declared mean
on spot samples) ≤ 0.75% Na2O eq) >0.75% Na2O eq)

Low reactivity Self-limiting: no mix calculation Self-limiting: no mix calculation Limit:

needed † needed † ≤ 5.0 kg Na2O eq/m3 ‡ ◊
Normal reactivity Self-limiting: no mix calculation Limit: ≤ 3.5 kg Na2O eq/m3 ◊ § Limit:
needed #
≤ 3.0 kg Na2O eq/m3 ‡ ◊
3
3

High reactivity Limit: ≤ 2.5 kg Na2O eq/m Limit: ≤ 2.5 kg Na2O eq/m Limit: ≤ 2.5 kg Na2O eq/m3

Notes to Table 1
† This self-limiting condition and associated recommendation applies:
a) where the concrete contains any cement (Table 6) or any combination irrespective of the proportions of ggbs or pfa;
and b) where the alkali contributed to the fresh concrete from other sources (aggregate, admixture etc) is ≤ 0.60 kg Na2O eq/m3. Guidance on situations where this
value is exceeded is outside the scope of this Digest (but see text on the use of RCA).
‡ This recommendation applies:
a) where the concrete contains any cement in Group A, B*, B or C (Table 6), or a combination of CEM I with ≥ 40% ggbs or ≥ 25% pfa;
b) where the alkali contributed to the fresh concrete from other sources (aggregate, admixture etc) is ≤ 0.20 kg Na2O eq/m3. Where this value is exceeded, the
recommended limit should be reduced by the amount of alkali contributed from other sources up to a maximum of 0.60 kg Na2O eq/m3;
and c) where contributions to the initial or a revised limit are derived from the cement or the CEM I component of a combination.
◊ This recommendation applies:
a) where the concrete contains a combination of CEM I with <40% ggbs or <25% pfa. Alkali contributions from either the ggbs or pfa should be calculated in
accordance with the recommendations in Table 7 or 8 and the recommended limit should be reduced by the amount of these alkalis;
b) where the alkali contributed to the fresh concrete from other sources (aggregate, admixture etc) is ≤ 0.20 kg Na2O eq/m3. Where this value is exceeded, the
recommended limit should be reduced by the amount of alkali contributed from other sources up to a maximum of 0.60 kg Na2O eq/m3;
and c) where contributions to the initial or a revised limit are derived from the cement or the CEM I component of a combination.
# This self-limiting condition and associated recommendation applies:
a) where the concrete contains cement in Group A, B or C (Table 6, noting that B* is excluded) or a combination of CEM I with ≥ 40% ggbs or ≥ 25% pfa;
and b) where the alkali contributed to the fresh concrete from other sources (aggregate, admixture etc) is ≤ 0.60 kg Na2O eq/m3. Guidance on situations where this
value is exceeded is outside the scope of this Digest (but see text on the use of RCA).
§ This recommendation applies:
a) where the concrete contains cement in Group A, B or C (Table 6, noting that B* is excluded) or a combination of CEM I with ≥ 40% ggbs or ≥ 25% pfa;
b) where the alkali contributed to the fresh concrete from other sources (aggregate, admixture etc) is ≤ 0.20 kg Na2O eq/m3. Where this value is exceeded:
i) the recommended limit should be reduced by the amount of alkali contributed from other sources up to a maximum of 0.60 kg Na2O eq/m3;
and ii) additionally, for pfa-containing cements or combinations, the recommended maxima in Table 9 may apply;
and c) where contributions to the initial or a revised limit are derived from the cement or the CEM I component of a combination.

This recommendation applies:
a) where the concrete contains either: a cement in Group A or C (Table 6, noting that B and B* are excluded), or a combination of CEM I with ≥ 50% ggbs
or ≥ 40% pfa;
b) where the alkali contributed to the fresh concrete from other sources (aggregate, admixture etc) is ≤ 0.60 kg Na2O eq/m3;
and c) where contributions to the initial or a revised limit are derived from the cement or the CEM I component of a combination.
 3

Using Table 1 Aggregates

Table 1 is a matrix of recommendations, in a variety of forms, Not all rock types and rock combinations behave in the same
which are either inclusive or exclusive of limiting values of way when exposed to alkaline pore solutions in concretes.
alkali content in a concrete mix. One of the most important means of distinguishing between
To locate the appropriate recommendation: rock types and rock combinations is the threshold alkali value
● from the columns, classify the alkali content of the CEM I in concrete which first promotes damaging ASR. In field and
component of the cement, or of a combination; laboratory studies, some rock types and rock combinations
● from the rows, classify the reactivity of the aggregate type (eg greywackes) promote damaging ASR at lower alkali
or aggregate combination; levels and with less associated ASR gel detected than other
● find the intersection of these classifications; more common UK types of aggregates (eg flint sands). The
● from the appropriate note at the foot of the table, establish aggregate’s reactivity can, therefore, be defined in terms of
the major limiting conditions to which the this threshold alkali level. Aggregate types (including
recommendation is subject. (Limiting conditions which aggregate combinations) which produce damaging
apply generally to Table 1 are given on page 1.) expansion at low alkali levels are considered highly reactive;
● if a limit value for alkali content forms the basis for the aggregate types which do not produce expansion even at very
recommendation, calculate the alkali content of the high alkali levels are considered low reactivity.
concrete mix for comparison with the limit value. This Digest uses this concept to classify types and
(Worked examples are given in Part 3 of this Digest.) combinations of aggregates into three broad classes of
● from the information obtained so far, decide whether the reactivity: high, normal and low, and provides modified
recommendations or limiting conditions are appropriate or guidance for each reactivity class.
acceptable, or whether alternative mix constituents should
be considered. Low reactivity aggregates
This class or combination of aggregates includes many of the
rock types that have not, to date, been implicated in damaging
ASR in field concretes in the UK. Table 2 gives a list (not
exhaustive) of these rock types and rock combinations.
Even though an aggregate is of low reactivity when used as
both the sand and coarse fraction in the same mix, it might
cause damaging expansion when used in combination with a
different low-reactivity aggregate type. This feature is found
with flint sands when used at pessimum proportions.
It has been advised previously that aggregate
Table 2 Rocks and minerals considered to be of low combinations of sand and gravels, in which both fractions
reactivity have a high proportion of flint and the total flint content is
This table is based on current experience in the UK. In other countries, greater than 60% by mass, were unlikely to be reactive. BRE
some of the rock types (eg younger basalts) have been found to be more research shows this advice to be still sound in the vast
reactive majority of cases. However, laboratory tests also showed that
andesite gabbro schist abnormal expansion was experienced in very few cases (5 out
basalt gneiss slate of 80 deposits studied) where the total flint content in the
chalk † granite ◊ syenite aggregate combination exceeded 60%.
diorite limestone # trachyte As a precautionary measure, therefore, these combinations
dolerite marble tuff should be considered normally reactive unless other evidence
dolomite microgranite indicates the aggregates as being of low reactivity. Other
feldspar ‡ quartz ‡ § evidence could be a long history of satisfactory use in
air cooled blastfurnace slag concretes of similar or higher alkali content in external
expanded clay/shale/slate exposure. If the aggregate contains less than 10% crushed
sintered pfa greywacke, the concrete prism test in BS 812-123 could be
used as a basis for the assessment.
† Chalk is included in the list since it may occasionally be a minor rock type in Sims and Blackwell[1] have provided some interpretations
concrete aggregates.
of the results (reproduced in Table 3 on page 4) obtained in
‡ Feldspar and quartz are not rock types but are discrete mineral grains occurring
principally in fine aggregates. the BS 812-123 test which have been recommended by BSI
◊ A few cases have been reported in which partially altered feldspars within Working Group B/502/6/10 (who devised the test).
granites have been identified as the source of deleterious ASR. If a low reactivity aggregate is combined with a more
# One UK silicified limestone deposit has been found to induce abnormal
reactive type, and there is no information to suggest
expansion in laboratory concrete expansion tests. However, in the UK, this
deposit appears to be related to particular and localised geological otherwise, reactivity of the combination is classified on the
circumstances. Other UK limestones have been used in concretes over many basis of the higher reactivity aggregate. For example, if a low
years without evidence of ASR. reactivity aggregate is combined with a normally reactive
§ Not quartzite, nor micro-crystalline or crypto-crystalline quartz.
one, the combination is assumed to be normally reactive.
4

Table 3 Concrete prism test from BS 812-123

Suggested criteria for interpretation: applies to low and normal reactivity UK aggregates (BSI Working Group B/502/6/10)
Expansion Classification for Aggregate type Notes Some examples from
for up to the aggregate from Table 1 UK sources
12 months † combination tested ‡
(%)

>0.20 Expansive Normal reactivity Exhibited by combinations known to Chert and flint-bearing sand
have been involved in cases of actual with non-reactive, low porosity
damage to concrete coarse aggregates
0.10 to ≤ 0.20 Possibly expansive Normal reactivity Includes combinations which have Quartzitic gravels and sands
sometimes been involved in cases of from the English Midlands and
actual damage to concrete some crushed rock containing
structures, but also includes some microcrystalline quartz
widely used combinations with no
record of causing damage to
concrete
>0.05 to ≤ 0.10 Probably Low reactivity Combinations in this range have rarely Some of the quartzitic gravel
non-expansive been associated with actual cases of sands from the English
damage to concrete structures. Midlands
However, these might be considered
unsuitable in extreme conditions
≤ 0.05 Non-expansive Low reactivity Combinations which have no record Crushed limestone and igneous
of causing damage to concrete rock aggregates‡; chert and
flint bearing coarse/fine
combinations with > 60% chert
and flint contents

† Occasionally expansion in the test might not have reached its ultimate expansion at 12 months. For example, this is the case with some crushed greywacke aggregates.
‡ Except when contaminated by opaline and/or microcrystalline quartz.

Normal reactivity aggregates High reactivity aggregates

Most aggregates extracted in the UK are expected to be
classified as normally reactive for the purpose of this Digest.
Included in this classification should be all combinations of
sand and gravel, including those which contain greater than
60% flint.
The few exceptions should be classified as high or low
reactivity aggregates as appropriate.
Table 4 shows types and combinations of aggregates known
to be highly reactive. These react at lower alkali levels than
the more common UK concreting aggregates, so more
conservative recommendations are given. In the case of
crushed greywacke or crushed greywacke-type aggregates,
however, a testing protocol for greywacke aggregates[2] has
been developed which can be used to assess, on a quarry-by-
quarry basis, whether a less conservative approach may be
valid. Fortunately, many of the rock types and rock
combinations listed in this reactivity class are rarely used in
the UK as concreting aggregates.

Table 4 Aggregates considered to be of high reactivity

Aggregate type Comments

Crushed greywacke, greywacke-type
sandstones, greywacke-type
siltstone or mudstones, or mixtures
containing more than 10% of these
Applied to aggregate from a primary source and excludes materials found as normal, uncrushed,
constituent of some natural sand and gravel deposits; unless the sand and gravel deposit
contains, or is subsequently blended with, 10% or more greywacke and/or greywacke-type
sandstone and/or siltstone or mudstone material that has been crushed (regardless of whether
this crushed material is from a primary or gravel source)

Recycled and reclaimed aggregates
Recycled aggregates
In the previous edition (1999) of this Digest, it was
considered prudent to classify recycled aggregates as highly
reactive as an interim measure. Recently, however, the topic
has been debated within BSI, leading to the inclusion in
BS 8500 of specific recommendations for the use of certain
recycled aggregates while ensuring that the risk of damaging
ASR is minimised. These recommendations have been
adopted in this Digest.
BS 8500 distinguishes between recycled aggregate and
recycled concrete aggregate, defined as follows:
recycled aggregate (RA) – aggregate resulting from the
reprocessing of inorganic material previously used in
construction
recycled concrete aggregate (RCA) – recycled aggregate
principally comprising crushed concrete
The difference between these materials is exemplified by the
maximum permitted masonry content: 100% for RA and 5%
for RCA. (Note: Annex B of BS 8500-2 gives a test method
for determining the composition of RCA and RA.) While
BS 8500 gives some general requirements for RA, it
considers that the potential composition is so wide that more
specific requirements cannot be given and they should be
assessed on a case-by-case basis. In particular, BS 8500 states
that the project specification for RA should include, among
other requirements, the classification with respect to
alkali–aggregate reactivity and the method for determination
of the alkali content. Because of this absence of detailed
requirements, it is recommended that RA continues to be
classified as being of high reactivity in Table 1. The
recommendations given below are applicable only to RCA
meeting the requirements given in BS 8500.
5
Using RCA
RCA may be classified as having normal reactivity provided
the following conditions apply.
1 The alkali contribution from RCA is assumed to be either:
0.2 kg Na2O eq per 100 kg of RCA, or
where the composition of the RCA is known
the alkali content can be calculated for the original
concrete.
2 The 0.6 kg Na2O eq/m3 limit on alkalis contributed to the
fresh concrete from sources other than the cement or
combination does not apply to alkali contributed by the
RCA.
3 The RCA does not contain highly reactive or extremely
reactive aggregates such as glass or opal (which are
excluded from the scope of this Digest) and the aggregate,
other than the RCA, is not classed as highly reactive or
extremely reactive.
Reclaimed aggregates
Reclaimed aggregates are natural aggregates retrieved from
the concrete production process before the adhering cement
has set. Usually these aggregates are recovered after washing
out the cement from returned ready-mixed concrete; the end
products of the process are essentially unmodified from the
original natural aggregates.
No additional recommendations have been made for
reclaimed aggregates. They should, therefore, be classified in
terms of reactivity of the original aggregate type or
combination.
6
Cementitious materials
Cements
Investigations of UK structures affected by ASR have shown
that damage is nearly always associated with using an
abnormally high alkali cement (Na2O eq >1%). Conversely,
long term experience in both field and laboratory concretes
has confirmed the well established use of a low alkali cement
(Na2O eq < 0.60%) to minimise damage from ASR. For this
reason guidance here differs depending on the alkali content
of the cement under consideration. The underlying principle
of the advice is that more precautions are needed to minimise
damaging ASR as the alkali content of cement increases.
While Table 1 divides cements into three classes of alkali
content – low, moderate and high – Table 5 gives the full
classification system.
Table 5 Classification of cement based on the alkali content of the CEM I-type component
Cement classification Alkali limit
Low alkali cement Cement with a guaranteed alkali content ≤0.60% Na2O eq
Moderate alkali cement Cement with an alkali content ≤0.75% Na2O eq based on manufacturer’s declared mean
High alkali cement Cement with an alkali content >0.75% Na2O eq based on manufacturer’s declared mean
New European cement standards
From May 2002, cement in the UK has been specified to the British
Standard BS EN 197-1 Cement – Composition, specifications and
conformity criteria for common cements. This Standard superseded
the British Standards:
● BS 12:1996 Specification for Portland cement
● BS 4246:1996 Specification for high slag blastfurnace cement
● BS 6588:1996 Specification for Portland pulverized fuel ash
cements
● BS 7583:1996 Specification for Portland limestone cement
Essentially, Portland cement to BS 12 became a BS EN 197-1 CEM I
cement. The BS 4246, BS 6588 and BS 7583 cements were
incorporated into BS EN 197-1 cement types CEM II, CEM III and
CEM IV, but not in a mutually exclusive way.
Not all British Standards for cement were superseded in May
2000, and, of these, the ones still referred to are:
● BS 146:2002 Specification for blastfurnace cements with
strength properties outside the scope of BS EN 197-1
● BS 1370:1979 Specification for low heat Portland cement
● BS 4027:1996 Specification for sulfate-resisting Portland
cement
● BS 6610:1996 Specification for Pozzolanic pulverized fuel ash
cement
Concretes made with cement with a Na2O eq (acid soluble)
less than 0.60% are unlikely to be at risk from damaging ASR
even when using a normally reactive aggregate.
Low alkali cements guaranteed to have a Na2O eq less than
0.60% are available in the UK as an option under BS 4027. In
view of the reduced risk of damaging ASR, more relaxed
recommendations are made here where a guaranteed low
alkali cement is used.
Cements have also been classified into groups (Table 6).
This is to simplify the selection of an appropriate Portland
cement and to meet one of the limiting conditions given in the
notes to Table 1.
For BS EN 197-1, the basis for how the alkali information relevant to
each cement type is made available is set out in the National Annex
NB4. BS 146 refers to BS EN 197-1 as the basis for providing alkali
information where required. The basis for providing the alkali
information for BS 4027 and BS 1370 cements should be taken as
the same as that for BS EN 197-1 CEM I cements. BS 6610 sets out a
basis for declaring alkali content that differs from BS EN 197-1; but,
as the BS EN 197-1 method is the more conservative, it is the
approach adopted for this Digest.
Prior to April 1989, when standard strength classes were
introduced, Portland cement was classified as either ordinary
Portland cement (OPC) or rapid hardening Portland cement (RHPC).
This is mentioned since the terms Ordinary, OP and OPC remain in use
although the correct identification is Portland Cement CEM I.
A list of cements is set out in Table 6 together with the various
proportions of any ggbs or pfa, as the proportion of these materials in
cement may be required to calculate the total alkali contribution from
the cement.
 7

Table 6 Cement groups

Group Cement Cement specification and Specified proportions Additional qualification
designation of ggbs or pfa on proportions
(by mass)

A Portland cement BS EN 197-1 CEM I

Low heat Portland cement BS 1370
Sulfate resisting Portland cement BS 4027
Portland limestone cement BS EN 197-1 CEM II/A–L or –LL
B*† Portland slag cement BS EN 197-1 CEM II/A–S‡ 6–20% ggbs <42% ggbs
BS EN 197-1 CEM II/B–S‡ 21–35% ggbs <42% ggbs
Blastfurnace cement BS 146 BIIIA 36–65% ggbs <42% ggbs
BS EN 197-1 CEM III/A 36–65% ggbs <42% ggbs
Portland fly ash cement BS EN 197-1 CEM II/A–V‡ 6–20% pfa <26% pfa
BS EN 197-1 CEM II/B–V‡ 21–35% pfa <26% pfa
Pozzolanic cement BS EN 197-1 CEM IV/A–V 11–35% pfa <26% pfa
B Blastfurnace cement BS 146 BIIIA 36–65% ggbs ≥ 42% ggbs
BS 146 BIIIB 66–80% ggbs
BS EN 197-1 CEM III/A 36–65% ggbs ≥ 42% ggbs
BS EN 197-1 CEM III/B 66–80% ggbs
BS EN 197-1 CEM III/C‡ 81–95% ggbs
Portland fly ash cement BS EN 197-1 CEM II/B–V 21–35% pfa ≥ 26% pfa
Pozzolanic cement BS EN 197-1 CEM IV/A–V 11–35% pfa ≥ 26% pfa
BS EN 197-1 CEM IV/B–V 36–55% pfa
Pozzolanic pfa cement BS 6610 36–55% pfa
C Blastfurnace cement BS 146 BIIIA 36–65% ggbs ≥ 50% ggbs
BS 146 BIIIB 66–80% ggbs
BS EN 197-1 CEM III/A 36–65% ggbs ≥ 50% ggbs
BS EN 197-1 CEM III/B 66–80% ggbs
BS EN 197-1 CEM III/C‡ 81–95% ggbs
Pozzolanic cement BS EN 197-1 CEM IV/B–V 36–55% pfa ≥ 40% pfa
Pozzolanic pfa cement BS 6610 36–55% pfa ≥ 40% pfa

† B* cements where the proportion of a blastfurnace slag constitutes <42% by mass of the sum of the main and minor additional constituents (formerly called cement
nucleus), or where the proportion of pfa constitutes <26% by mass of the sum of the main and minor additional constituents (formerly called cement nucleus). For these
cements the appropriate alkali contribution from the ggbs or pfa is incorporated into the declared mean alkali content as set out in BS EN 197-1:2000 NB4.
‡ Cements not readily available in the UK.
8
Ggbs and pfa for use in cements and combinations
Requirements for pfa and ggbs, when used as constituents of
cements (Table 6), are given in BS EN 197-1.
Recent research has shown that pfa to BS 3892-1 and fly
ash to BS EN 450 can be regarded as equivalent in their
effectiveness in controlling ASR; so when used in
combinations, ashes conforming to either to BS 3892-1 or
BS EN 450 can be used. Ggbs should conform to BS 6699.
The terms pulverised fuel ash (pfa) and fly ash are usually
synonymous. British Standards such as BS 3892-1 generally use pfa,
whereas European Standards such as BS EN 450 and BS EN 197-1
use fly ash. For simplicity, this Digest uses the term pfa throughout
except where referring to the contents of these European Standards.
Unless stated otherwise, pfa should be taken to include fly ash.
Much of the research work to prove the effectiveness of
these materials in preventing damaging ASR was undertaken
using ggbs and pfa which were commonly available in the
UK at the time. Traditionally, for both ggbs and pfa, the alkali
contents were considered to be low so knowledge of the
effectiveness of high alkali materials in minimising
damaging ASR is limited. It is prudent, therefore, to put
maximum alkali limits on these materials pending further
research work. The acid-soluble alkali content in ggbs
should not exceed 1.0% Na2O eq and in pfa should not exceed
5.0% Na2O eq.
recommended for the appropriate cement and combinations,
depending on the reactivity class of the aggregate. No
account needs be taken of alkali in these materials when they
are used at or above the minimum proportions recommended
for a particular classification.
Minimum proportions are given percentage values, below,
for cases where a classification sequence (described in the
section ‘Using Table 1’ on page 3) is:
● based on a low reactivity aggregate together with a high
alkali cement classification, or
● a normal reactivity aggregate together with a low,
moderate or high alkali cement classification.
The minimum proportion of ggbs should be 40% and the
minimum proportion of pfa should be 25% where either a
cement containing ggbs or pfa (Groups B or C in Table 6) or a
combination has been selected. If proportions less than these
minima are used, a fraction of the alkali contribution from the
ggbs or pfa should be taken into account (Tables 7 and 8).
● based on a low reactivity aggregate together with a low or
moderate alkali cement classification.
No account should be taken of any contribution of alkali from
ggbs or pfa where a cement containing either of these
materials or a combination is used since no minimum
proportions are recommended.

The acid soluble alkali content is determined:
● in ggbs by the method specified in National Annex of BS EN 196-21;
● in pfa using a method specified in BS 3892-1, Clause 14.2b.
Where ggbs and pfa are used to minimise the risk of
damaging ASR, different minimum proportions are
● based on a high reactivity aggregate, irrespective of the
classification of the cement.
Proportions less than 50% by mass for ggbs and 40% by mass
for pfa for either a cement or for combinations are not
recommended.

Table 7 Alkali contributions from ggbs to be taken into account if the target mean proportions are less than those
recommended in Table 6 and in the notes to Table 1
Proportion of ggbs in a cement Fraction of the acid-soluble alkali content of the ggbs to be taken into account
or combination when calculating the contribution from the cement or combination

1
25 – 39% in a cement or combination ⁄2
<25% in a cement or combination All

As a consequence of the more conservative approach to the contribution of alkali from ggbs, as recommended in this table, step changes in the
amounts contributed from the cementitious sources occur around the proportions recommended in Table 6 and in the notes to Table 1.

Table 8 Alkali contributions from pfa to be taken into account if the target mean proportions are less than those
recommended in Table 6 and in the notes to Table 1
Proportion of pfa in a cement Fraction of the total † acid-soluble alkali content of the pfa to be taken into account
or combination when calculating the contribution from the cement or combination

1
20 – 24% in a cement or combination ⁄5
<20% in a cement or combination All

† The total acid-soluble alkali content of a pfa is determined following either a complete acid decomposition or a high temperature fusion.

As a consequence of the more conservative approach to the contribution of alkali from pfa, as recommended in this table, step changes in the
amounts contributed from the cementitious sources occur around the proportions recommended in Table 6 and the notes to Table 1.

Pfa to BS 3892-2 as a Type I addition
Most UK studies into the capacity of pfa to mitigate the
effects of ASR have used it, pfa, conforming to BS 3892-1 or
BS EN 450. Little information is currently available on the
effects on ASR where pfa to BS 3892-2 is used as a Type I
addition. Guidance cannot be given here for using BS 3892-2
ash to modify the risk of ASR because of the lack of
technical data.
However, provision has been made for such pfa if it is
needed for purposes other than to modify the risk of ASR.
It should then be treated in the same way as a BS 3892-1 or
BS EN 450 ash as if it were a component of a combination
for the purpose of its contributions to the recommended
maximum combination content, and to mix alkali contents
(Table 9).
The worked examples in Part 3 of this Digest show how
these recommendations are applied.
9
Alkalis contributed to the fresh concrete from other
sources
Alkalis from other sources (aggregate, admixture etc)
included at mixing have been shown to contribute to
damaging ASR in a few concrete structures. Information is
limited on how these alkalis affect the mechanisms involved,
particularly at higher levels of alkali contribution.
Accordingly, the guidance given adopts a conservative
approach to these alkali contributions by taking all the
alkali into account where the contribution exceeds 0.2 kg
Na2O eq/m3 and by limiting the maximum contribution to
0.60 kg Na2O eq/m3, excluding alkali contributed by
recycled concrete aggregates (notes to Table 1 and the
section on recycled aggregates on page 5).

Table 9 Recommended maximum cement or combination contents where pfa is a constituent

This table does not apply for ggbs and is applicable only where:
● aggregates are classified as of normal reactivity (Table 1)
● cement is classified as of moderate alkali (Table 1)
● target mean proportions for pfa-containing cements or combinations are 25% ≤ pfa <30%
Alkali contributed from other sources † Declared mean‡ alkali content (kg/m3) of the CEM I-type component of a
3
(kg Na20 eq/m ) cement or of the CEM I component of a combination
0.65% 0.70% 0.75%

0.21– 0.30 550 540 530

0.31– 0.40 550 525 510
0.41– 0.50 550 515 490
0.51– 0.60 ◊ 550 505 470

† Other sources are non-cementitious materials included at the time of mixing (aggregate, admixture etc).
‡ Linear interpolation is permitted for intermediate levels of declared mean alkali content.
◊ Guidance where there is >0.60 kg Na2O eq/m3 from other sources is outside the scope of this Digest (section ‘Using RCA’ on page 5).
10

Water De-icing salts

If the water is of potable quality from a UK source, no
account needs be taken of the alkali metal content when
calculating alkali levels in a concrete mix.
Where the water complies with BS EN 1008, an alkali
contribution from the water should be taken as the greater of
1.5 g/l or the declared maximum value.
This applies particularly to aggregates of high reactivity.
Because aggregates of high reactivity react at lower alkali
levels than other common forms of aggregates, alkalis in de-
icing salts may add to alkalis from the binder sufficiently to
promote or aggravate ASR. This has been shown by research
in which concrete specimens have been immersed in
solutions of relatively high salinity for long periods. It is
possible for solutions of high salinity to occur as a result of
ponding of run-off water on poorly detailed or badly
maintained elements and pavements which are frequently
exposed to these salts; this ponding, in extreme
circumstances, may promote localised ASR damage.
If high reactivity aggregates are used in structures exposed
to de-icing salts, the structure or pavement should be detailed
to avoid ponding. In existing structures which contain high
reactivity aggregates, other steps should be taken to avoid
ponding of solutions of high salinity. Care should be taken in
maintenance to avoid ingress and ponding of saline solutions:
for example, along joints.

Table 10 Recommendations for use of metakaolin and silica fume to minimise ASR damage: recommended limits for the
alkali contents of concrete
The alkali contents are calculated in accordance with Table 1 and its notes
Aggregate type Alkali content of cement
Low Moderate High

Low reactivity Self-limiting Self-limiting ≤5.0 kg Na2O eq/m3

with 10% metakaolin
and ≥ 8% silica fume
Normal reactivity Self-limiting ≤5.0 kg Na2O eq/m3 ≤5.0 kg Na2O eq/m3
with 10% metakaolin with 10% metakaolin
≤4.0 kg Na2O eq/m3 ≤3.5 kg Na2O eq/m3
with ≥ 8% silica fume with ≥ 8% silica fume
High reactivity ≤5.0 kg Na2O eq/m3 ≤5.0 kg Na2O eq/m3 ≤5.0 kg Na2O eq/m3
with 10% metakaolin with 10% metakaolin with 10% metakaolin
≤3.0 kg Na2O eq/m 3
≤3.0 kg Na2O eq/m 3
≤3.0 kg Na2O eq/m3
with ≥ 8% silica fume with ≥ 8% silica fume with ≥ 8% silica fume

Table 11 Minimum recommended additions of lithium salts to minimise ASR damage

The recommendations in this table are applicable only if the total alkali content of the mix (in accordance with Table 1 and its notes) does not exceed
5.0 kg/m3
Aggregate type Lithium salts Pfa by mass of Lithium salts or Salt admixture Volume reduction
total binder solution per kg of per kg of Na2O eq in mixing water
Na2O eq per litre addition of
admixture
(%) (Kg) (litres) (litres)

Low reactivity No addition needed

Normal reactivity LiOH.H2O 0–25 0.75 – –
LiNO3, 30% solution 0–25 3.75 3.15 2.6
High reactivity LiOH.H2O 0–14 1.3 – –
15–25 1.0 – –
LiNO3, 30% solution 0–14 5.95 5.0 4.15
15–25 5.2 4.4 3.6

Additional methods of minimising the risk of
damage from ASR
Research completed at BRE in the last couple of years
(IP 1/02) has confirmed the effectiveness of two additional
methods of controlling ASR: by adding metakaolin or
lithium salts to the concrete mix. Recommendations based on
these have been considered and endorsed by a working group
of The Concrete Society ASR Committee.
This working group has also reviewed the information on
using silica fume and air entrainment in controlling ASR.
Whereas it concluded that silica fume was effective, the
evidence of the effectiveness of air entrainment in respect of
ASR was poor. Accordingly, recommendations on using
silica fume are included here.
Metakaolin and silica fume are strongly pozzolanic
materials and the mechanism by which they control ASR is
similar to that of pfa. Lithium salts operate through a different
mechanism. Although it is not completely understood, it
seems that the addition of lithium salts modifies the
properties of the alkali silica gel so that it is not expansive.
These further measures allow greater flexibility in the choice
of methods to minimise the risk of damaging ASR and in
some cases a relaxation of alkali limits. Also they will be
valuable in designing concrete for use in highly aggressive
conditions such as saline environments.
Silica fume
This is a by-product of the manufacture of silicon metal or
ferro-silicon alloys. It is a highly pozzolanic material and is
internationally available from many sources. Specifications
for cements containing silica fume (CEM II/A–D) are given
in BS EN 197-1 and for using as a Type II (ie pozzolanic or
latent hydraulic) addition to concrete in BS EN 206-1. Silica
fume currently available in the UK has amorphous SiO2
contents above 90% and is low in alkali.
To minimise the risk of damaging ASR, the silica fume
should have an amorphous SiO2 content equal to or greater
than 85% by mass. It is recommended that at least 8% by
mass, as percentage of the total binder, is used; either in a
CEM II/A–D cement or equivalent CEM I/silica fume
combination. In this case an increase of 0.5% kg/m3 in the
maximum allowable alkali content of the concrete, with
either a normally or highly reactive aggregate being allowed
(Table 10).
11
Metakaolin
Metakaolin is a strongly pozzolanic material produced by the
calcining of purified kaolinite clay. It can be used as either a
Type II addition or as a component of cement. In order to
minimise the risk of ASR damage in concretes exposed only
to normally aggressive environments, it is recommended that
a 10% addition, by mass of total binder, of a metakaolin
containing at least 45% by mass of SiO2 is used. In this case
the allowable alkali content of the concrete can be relaxed to
5 kg/m3, except where it is self-limiting (Table 10). However,
in highly aggressive conditions, such as concrete containing a
highly reactive aggregate exposed to highly saline
environments, a 15% addition is recommended.
Lithium salts
Adding lithium salts to reduce ASR expansion was first
suggested over 50 years ago. In recent years, however, there
has been systematic evaluation of their use so that solidly
based recommendations can be now made. A number of
lithium salts have been shown to be effective but either
lithium nitrate or lithium hydroxide are preferred because of
their greater solubility. These recommendations are focussed
on their use as admixtures in fresh concrete but trials have
also been undertaken on the addition of lithium-containing
materials to the cement kiln feed and on intergrinding with
cement clinker. Furthermore, field trials have been carried
out on applying lithium salt solutions as remedial treatments
for ASR affected structures.
Addition of lithium salts as an admixture to minimise
risk of ASR damage
The lithium salt should be added at the mixing stage, either as
a solid (lithium hydroxide) or as a 30% solution (lithium
nitrate). Good dispersion into the mix is essential. The
amount needed depends on the reactivity of the aggregate and
the alkali content of the concrete mix (Table 11). There is
evidence that adding lithium salts is effective in mixes
containing lower proportions of pfa than are recommended
on page 8; lower amounts of lithium salts are recommended
therefore in Table 11 for some mixes containing pfa. In the
case of ggbs, however, there is some evidence of increased
expansions in mixes containing lesser proportions of lithium
salts, so using lithium salts in concretes containing lesser
proportions of ggbs than those indicated on page 8 is not
recommended.
12

References and further reading

[1] Sims I and Blackwell B Q. Suggested criteria for the interpretation of the
DD 218: 1995. Concrete Prism Test. Discussion document submitted in 1994 to
BSI Working Group, 502/6/10.

[2] British Cement Association. Testing protocol for greywacke aggregates.

Protocol of the BSI B/517/1/20 Adhoc Group on ASR. BCA Publication 45.044.
Crowthorne, BCA, 1999.

[3] The Concrete Society. Alkali–silica reaction: minimising the risk of damage to
concrete. Concrete Society Technical Report No 30 (Third Edition). Crowthorne,
The Concrete Society, 1999.

BRE
IP 1/02 Minimising the risk of alkali–silica reaction: alternative methods

British Standards Institution

BS 146:2002 Specification for blastfurnace cements with strength properties
BRE is committed to providing
outside the scope of BS EN 197-1
impartial and authoritative information
BS 812-123:1999 Testing aggregates. Method for determination of alkali–silica on all aspects of the built environment
reactivity. Concrete prism method for clients, designers, contractors,
BS 1370:1979 Specification for low heat Portland cement engineers, manufacturers, occupants,
etc. We make every effort to ensure
BS 3892-1:1997 Pulverized fuel ash. Specification for pulverized fuel ash for use
the accuracy and quality of information
with Portland cement and guidance when it is first published.
BS 3892-2:1996 Pulverized fuel ash. Specification for pulverized fuel ash to be However, we can take no responsibility
used as a Type I addition for the subsequent use of this
information, nor for any errors or
BS 4027:1996 Specification for sulfate-resisting Portland cement
omissions it may contain.
BS 6610:1996 Specification for Pozzolanic pulverized fuel ash cement
BS 6699:1992 Specification for ground granulated blastfurnace slag for use with BRE is the UK’s leading centre of
Portland cement expertise on building and construction,
and the prevention and control of fire.
BS 8500:2002 (2 Parts) Concrete. Complementary British Standard to
Contact BRE for information about its
BS EN 206-1 services, or for technical advice, at:
BS EN 196-21:1992 Methods of testing cement. Determination of the chloride, BRE, Garston, Watford WD25 9XX
carbon dioxide and alkali content of cement. (National Annex) Tel: 01923 664000
Fax: 01923 664098
BS EN 197-1:2000 Cement. Composition, specifications and conformity criteria
email: enquiries@bre.co.uk
for common cements Website: www.bre.co.uk
BS EN 450:1995 Fly ash for concrete. Definitions, requirements and quality control
BS EN 1008:2002 Mixing water for concrete. Specification for sampling, testing Details of BRE publications are available
from:
and assessing the suitability of water, including water recovered from processes
www.brebookshop.com
in the concrete industry, as mixing water for concrete or
IHS Rapidoc (BRE Bookshop)
Willoughby Road
Bracknell RG12 8DW
Acknowledgement Tel: 01344 404407
The authors would like to acknowledge the support of the Department of Trade Fax: 01344 714440
and Industry in the preparation of this Digest. email: brebookshop@ihsrapidoc.com

Published by BRE Bookshop

Requests to copy any part of this
publication should be made to:
BRE Bookshop,
Building Research Establishment,
Watford WD25 9XX
Tel: 01923 664761
Fax: 01923 662477
email: brebookshop@emap.com

© Copyright BRE 2004

August 2004
ISBN 1 86081 710 6

www.bre.co.uk