Cement and Concrete Research 31 (2001) 707 ± 712

Performance of concrete made with commercially produced coarse

recycled concrete aggregate
K.K. Sagoe-Crentsil*, T. Brown, A.H. Taylor
CSIRO Building, Construction and Engineering, P.O. Box 56, Highett, Victoria 3190, Australia
Received 30 March 1999; accepted 3 November 2000

Abstract

Performance tests have been carried out for fresh and hardened properties of concrete made with commercially produced coarse recycled
concrete aggregate and natural fine sand. Test results indicate that the difference between the characteristics of fresh and hardened recycled
aggregate concrete and natural aggregate concrete is perhaps relatively narrower than reported for laboratory-crushed recycled aggregate
concrete mixtures. For concrete without blast furnace slag having similar volumetric mixture proportions and workability, there was no
difference at the 5% significance level in concrete compressive and tensile strengths of recycled concrete and control normal concrete made
from natural basalt aggregate and fine sand. Water absorption rates and carbonation of recycled concrete and reference concrete were
comparable for most applications. However, the abrasion loss of recycled aggregate concrete made with ordinary portland cement increased
by about 12% compared to normal concrete, while the corresponding drying shrinkage was about 25% higher at 1 year. The ratio of splitting
tensile strength to compressive strength was found to be in good agreement with established values derived for equivalent grade concretes
made with normal-weight natural aggregates. One-year test results indicate that incremental improvements in durability characteristics can
further be achieved with the use of blast furnace slag cement. Enhanced fresh and hardened concrete properties of the investigated recycled
concrete aggregate as compared to aggregate derived from laboratory-crushed concrete arise primarily from improved aggregate grading and
quality achievable in plant crushing operations. D 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Aggregate; Blast furnace slag; Compressive strength; Drying shrinkage; Durability; Performance; Recycled concrete

1. Introduction Recent advances in aggregate ±production technologies

in the areas of rubble screening and aggregate washing,
In recent years, the recycling of concrete to produce and tighter regulation of the recycling industry [4± 6],
aggregates suitable for nonstructural concrete applications have contributed to significant improvements in aggre-
is emerging as a commercially viable and technically gate quality. Accordingly, the performance characteristics
feasible operation. This situation has arisen following of concrete incorporating commercially graded recycled
well over two decades of intensive research, predomi- concrete aggregate, hereafter, referred to as recycled
nantly centered on laboratory-crushed concrete [1 ± 3]. concrete, require reassessment in relation to natural
Hence, the potential now exists for demolished concrete aggregate concrete.
to serve as a source of quality aggregate feedstock in a Literature reports [1,6] confirm that the cement mortar
variety of nonstructural applications, particularly in mod- attached to the aggregate particles primarily determines the
ern urban centers removed from quarry sites. Such recy- performance of concrete made with laboratory-crushed
cling operations have the added benefit of reducing recycled concrete aggregate. This residual mortar alters
landfill disposal, while conserving primary resources and aggregate absorption and density and can have adverse
reducing transport costs. effects on concrete performance. As widely reported [2,7],
typical reductions of the order of 10% in compressive
strength and up to a 70% increase in drying shrinkage are
not uncommon.
* Corresponding author. Tel.: +61-3-9252-6350; fax: +61-3-9252- While studies on the engineering properties of concrete
6244. made with laboratory-crushed recycled concrete aggregate

0008-8846/00/$ ± see front matter D 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 0 8 - 8 8 4 6 ( 0 0 ) 0 0 4 7 6 - 2
708 K.K. Sagoe-Crentsil et al. / Cement and Concrete Research 31 (2001) 707±712

Table 1
Sieve analysis of coarse recycled concrete aggregate and basalt aggregate
Coarse aggregates Percentage of mass passing through sieve, mm
Maximum size 19 mm 13.5 mm 9.5 mm 6.7 mm 4.75 mm 2.36 mm 150 mm
Recycled concrete aggregate 100 91.4 28.7 7.6 5.4 0.2 0.5
Basalt 100 84.0 43.7 5.6 2.1 1.0 0.2

abound [8,9], only limited data are available on commercial- cated in Table 2, recycled concrete aggregate also has a
grade recycled concrete aggregate, including concrete mix- comparatively high water absorption value at 5.6%.
ture proportions, fresh concrete performance, and durability
characteristics. Thus, this paper evaluates aspects of con- 2.2. Binders
crete rheology and performance of N25 (25 MPa)-grade
concrete containing commercially crushed and graded Ordinary portland cement, designated Type GP, and a
recycled concrete aggregate. blast furnace slag cement, Type GB, comprising 65/35
portland cement/blast furnace slag were used in the inves-
tigations. The slag blend was used partly to assess possible
2. Experimental program improvements in fresh concrete cohesion, workability, and
concrete performance.
2.1. Aggregates
2.3. Concrete mixtures
A single source of commercially graded unwashed coarse
recycled concrete aggregate and natural fines was used in all Several preliminary trial mixtures were proportioned to
concrete mixtures. The natural coarse aggregate used was a evaluate water requirements for a nominal 25-MPa con-
nominal 14-mm crushed basalt. The grading of the basalt crete. Mixture proportion data are given in Table 3. The
and the coarse recycled concrete aggregate conformed to the unit water content of the concrete was corrected for free
requirements of Australian Standard AS 2758.1 [10] as moisture in the aggregates. However, the recycled aggre-
shown in Table 1. gate was presaturated for 10 min in the mixer and brought
This grading and a fine-to-coarse aggregate ratio of 46:54 to room temperature prior to mixing the concrete. The
were kept constant throughout the test program for all mixtures were proportioned to have a nominal binder
concrete mixtures. Based on visual inspection, the surface content of 240 kg/m 3 for C0912A, C0912B, and
texture of the plant-crushed recycled concrete aggregate C1212A, while mix C1212B contained an additional 5%
appeared characteristically grainy compared to the basalt. cement to assess the effect of increased cement content.
The recycled concrete aggregate was batched in the as- Slag cement was used in mix C1212A.
received state. Details of the physical properties of both The water± binder ratio of all mixtures was adjusted to
aggregates are shown in Table 2. achieve comparable consistency, and hence equal nominal
Partly as a result of residual cement mortar attached to slump of 80 ‹ 15 mm. A reduction in the water requirement
the commercial recycled concrete aggregate particles, mar- was attained by using a lignosulfonate-based water-redu-
ginally lower density and higher aggregate crushing values cing admixture at nominal doses recommended by the
were observed compared to the basalt aggregate. As indi- manufacturer. The recycled concrete mixtures contained
100% recycled aggregate and natural fine sand, while the
normal concrete mixtures contained all natural coarse and
Table 2 fine aggregates.
Properties of recycled concrete aggregate and basalt aggregates
Recycled 2.4. Specimens and testing
concrete
Property aggregate Basalt Specimens were cast from each mixture to assess com-
Aggregate crushing value, 23.1 15.7 pressive strength, drying shrinkage, expansion, splitting
% (AS 1141.21) tensile strength, and abrasion resistance. The concrete dur-
Bulk density, kg/m3 2394 2890 ability-related tests involved accelerated carbonation and
(AS 1141.6)
Water absorption, 5.6 1.0
water absorption. Unless otherwise specified, all specimens,
% (AS 1141.6) upon their removal from the molds, were stored under
Impurity level, 0.6 < 0.1 standard moist curing conditions at 23°C/ > 95% RH until
% (AS 1141.32) required for testing. Hardened concrete testing was per-
LOI, % 4.9 1.3 formed in accordance with the requirements of AS 1012
 K.K. Sagoe-Crentsil et al. / Cement and Concrete Research 31 (2001) 707±712 709

Table 3
Mix designation and mixture details of concrete specimens
Binder Water ± binder Slump, Wet density, Entrapped Coarse
Mix designation loading, kg/m3 Binder type ratio mm kg/m3 air content, % aggregate
C0912A 242 Type GP 0.76 90 2466 2.4 basalt
C0912B 240 Type GP 0.73 75 2335 2.4 recycled
C1212A 238 Type GB 0.74 95 2321 1.8 recycled
C1212B 254 Type GP 0.70 80 2335 2.3 recycled

[11], with the exception of water absorption and abrasion results plotted in Fig. 1, there is no significant difference
resistance measurements that were respectively tested to between the strength of portland cement concretes, as a
ASTM C 642-1997 [12] and AS/NZS 4456.9:1997 [13]. function of aggregate type, for the grade of concrete
investigated. The observed equivalence in strength between
recycled and normal concretes contrasts the corresponding
3. Results and discussion 10% reduction in strength typical of laboratory-crushed
recycled concrete [6,9].
3.1. Fresh concrete properties As expected, the slag cement concrete with similar
water ± cement ratios achieves much higher later-age
The workability, expressed in terms of slump, varied strength owing to the hydraulic properties of the slag. The
within a 20-mm band. As summarized in Table 3, the increase in strength gain of the slag cement concrete from
entrapped air content of concrete was similar for mixtures 20.2 to 32.6 MPa between 7 and 28 days is significantly
made with portland (Type GP) cement, varying within higher than the equivalent nominal 6 MPa average increase
2.4 ‹ 0.2% compared to the low value of 1.8% obtained for the portland cement concretes. As noted in Fig. 1, the
for mix C1212A made with slag cement. Although the water strength gain for recycled concrete made with portland
absorption of the recycled aggregate is relatively high cement remains virtually unchanged at the 5% significance
compared to the reference basalt aggregate, there was no level beyond 28 days.
difficulty in achieving the desired consistency and subse-
quent compaction of concrete. As suggested by Hansen and 3.3. Splitting tensile strength
Narud [8], the observed marginal difference in measured
wet density between recycled concrete and normal concrete Results of the tests for indirect tensile strength, as
can be attributed to the presence of lower-density residual measured by the splitting tensile test, are shown in Fig. 2
cement mortar attached to aggregate particles. for recycled concrete and the reference basalt mixtures. The
general trend in the tensile strength development appears to
3.2. Compressive strength depend mainly on binder rather than aggregate type. Hence,
the tensile strength of the slag cement concrete improves
Compressive strengths were determined on concrete with curing, while the tensile strengths of portland cement
cylinders continuously stored under moist conditions for concretes remain practically unchanged beyond 28 days for
up to 365 days. As shown by the mean compressive strength the duration of the measurements.

Fig. 2. Relationship between splitting tensile strength of recycled concrete

Fig. 1. Development of concrete compressive strength with age. and reference basalt concrete with time.
710 K.K. Sagoe-Crentsil et al. / Cement and Concrete Research 31 (2001) 707±712

Table 4
Ratios of splitting tensile ( ft) and compressive strength ( fc) of concretes
ft/fc
Concrete mix 7 days 28 days 365 days
C0912A (OPC/basalt) 1.19 1.07 0.99
C0912B (OPC/recycled) 1.19 1.20 0.94
C1212A (slag/recycled) 1.21 0.98 0.96
C1212B (OPC + 5%/recycled) 1.08 0.97 0.89
OPC: Ordinary portland cement.

At one year, the tensile strengths of the slag cement

samples were of the order of 25% above the portland
cement samples compared to an equivalent 15% difference
at 28 days. By contrast, the tensile resistance increases as
Fig. 3. Relationship between drying shrinkage strain of recycled concrete
the cement content of the concrete increases, as indicated and reference concrete with time.
by the difference shown between mixes C0912B and
C1212B, respectively, with nominal and high portland
only a 15% increase in shrinkage strain is obtained for
cement contents. Generally, the pattern of behavior of the
portland cement specimens for the same curing conditions.
indirect tensile strength is similar to that of compressive
However, there appears to be a much less significant effect
strength. Compared with the development of compressive
on drying shrinkage as a result of the 5% difference in
strength, however, the percentage compressive strength
cement content.
gain beyond 91 days is less pronounced except for the slag
It is evident from Fig. 3 that recycled concrete concretes
cement mixture.
display higher drying shrinkage values compared to the
The absence of any detrimental effect on recycled con-
reference normal concrete mixture, possibly partly due to
crete tensile strength is partly indicative of good bond
the lower restraining capacity of recycled aggregate particles
characteristics between aggregate and the mortar matrix.
compared to basalt. Currently, the Australian Standard AS
Thus, the corresponding mechanism of failure of recycled
3600 [16] recommends a basic shrinkage limit of 700
concrete specimens would be expected to be similar to that
microstrain at 56 days.
of the reference concrete.
Overall trends in the expansion profiles of recycled
As shown in Table 4, the splitting tensile-to-compressive
concrete mixtures shown in Fig. 4 are nearly identical to
strength ratio of recycled concretes are comparable to results
those observed in the drying shrinkage plots (Fig. 3). For
obtained for conventional concretes made with natural
portland cement concretes, a marginally lower expansion
aggregates. This ratio provides an indication of the resis-
was obtained for recycled concrete compared with the
tance of a concrete to tensile strain and is dependent on type
control basalt concrete. The low expansion values obtained
and size of coarse aggregate particles, voids in the concrete,
for the higher portland cement content mixture can be
curing conditions, and test conditions. Calculated splitting
attributed to its lower water ± cement ratio and improved
tensile-to-compressive strength ratio values ranged from
paste quality of the matrix. Generally, it would be expected
0.89 to 1.21 for recycled concrete, which is in close
that shrinkage and swelling are similarly affected by the
agreement with published values of between 0.8 and 1.4
same parameters.
for equivalent normal weight aggregate concretes [14].

3.4. Drying shrinkage

The variation of drying shrinkage strain with time for

both recycled and reference concretes is shown in Fig. 3.
Fig. 4 documents the corresponding expansion results of all
concrete mixtures. The drying shrinkage of test specimens
increased with time and stabilized at about 91 days, follow-
ing similar trends reported by several researchers [1] for
laboratory-crushed recycled concrete, although the absolute
shrinkage values are slightly lower in the present case.
While both natural and recycled aggregates display similar
trends with regard to the rate of shrinkage, strains associated
with recycled concrete made with slag cement at 365 days
are over 35% higher than the reference mixture, and closer Fig. 4. Relationship between expansion of recycled concrete and reference
to typical published values of 30 ±70% [6,15]. In contrast, concrete with age.
 K.K. Sagoe-Crentsil et al. / Cement and Concrete Research 31 (2001) 707±712 711

3.5. Abrasion resistance

The resistance to abrasion of the recycled concrete

concretes was compared with the abrasion wear of the
reference normal concrete. White fused aluminum oxide
was used as the abrasive material and dispensed at a
constant rate, to the point of contact between disc and the
specimen surface. Fig. 5 shows the abraded volumes of
concrete specimens tested. The abrasion resistance of
recycled concrete specimens was generally similar. How-
ever, mix C0912B with nominal portland cement content
showed slightly lower abrasion resistance. Reference to Fig.
Fig. 6. Accelerated carbonation rates of concrete in 4% CO2 atmosphere.
5 further shows that the average abraded volume of the
recycled concrete at 212 mm3 was about 12% higher than
that of the reference basalt concrete.
At the lower concrete strengths tested, the effect of By excluding the slag cement concrete, some correspon-
aggregate hardness on abrasion resistance is likely to be dence can be established between concrete strength and the
more pronounced, given that the resistance to abrasion is rate of carbonation of the mixtures investigated for all
influenced by concrete strength, aggregate properties, and remaining concretes. Hansen [1] cites earlier carbonation
quality of near-surface finish. As with normal and light- studies carried out in Japan under different experimental
weight aggregates, the abrasion resistance of recycled con- conditions that showed a 65% increase in carbonation rate
crete may be expected to increase with compressive relative to the control. This trend contrasts with the nominal
strength [17]. 10% rise shown in Fig. 6 for plant-crushed and graded
recycled aggregate.
3.6. Carbonation
3.7. Water absorption
The variation of depth of carbonation with time under
accelerated exposure conditions is shown in Fig. 6 for all Fig. 7 shows percent water absorption for specimens
concretes. In general, a parabolic rate law applies to the subjected to 6 days of moist curing after demolding and 21
recycled concrete as with the reference mixture. The slag days of drying at 50% RH and 23°C. The measured
cement mixture shows a slight deviation from this trend, absorption of water indicates the relative quality of near-
suggesting the possibility of a different mechanism of surface concrete properties.
carbonation. As expected, the rate of carbonation of the Reference to Fig. 7 shows that under the curing regime
slag cement concrete is fractionally higher than the employed, the lower porosity of the basalt aggregates in the
equivalent strength-grade concretes not containing slag. normal concrete mixture restricts the rate of water absorption
However, for portland cement specimens, recycled con- compared to the recycled concrete mixtures that have an
crete carbonation rates appear to correlate well with average 25% higher absorption. The residual mortar attached
cement content. to recycled concrete particles serves as a potential conduit for

Fig. 7. Water absorption rates of recycled concrete and reference

Fig. 5. Abrasion resistance of recycled concrete and reference concrete. normal concrete.
712 K.K. Sagoe-Crentsil et al. / Cement and Concrete Research 31 (2001) 707±712

moisture transport, especially under limited curing condi- References

tions for which the maturity of concrete matrix strength and
aggregate:matrix bond properties remains critical. [1] T.C. Hansen, Recycling of demolished concrete and masonry, RILEM
As exemplified by the carbonation results, some corre- Rep. No. 6 E&FN Spon, London, UK.
[2] Y.A. Frondistou-Yannas, Recycled concrete as new aggregates, Pro-
lation between the quality of the matrix and the transport gress in Concrete Technology, CANMET, Energy Mines and Re-
of fluids or the degree of permeability is to be expected sources Canada, Ottawa, Canada, 1980, pp. 639 ± 684.
for fully cured specimens. Thus, for moist-curing condi- [3] K.K. Sagoe-Crentsil, T. Brown, S.L. Mak, A.H. Taylor, Engineering
tions the contribution of aggregates to concrete perme- properties and performance of concrete made with recycled construc-
ability may be much less pronounced and thus provide a tion aggregates, Proceedings of the National Symposium on the Use
of Recycled Materials in Engineering and Construction, 30 ± 31 May,
better indicator of recycled concrete durability under Institution of Engineers, Sydney, Australia, 1996, pp. 132 ± 135.
service conditions. [4] M.S. Rashwan, S. AbouRizk, The properties of recycled concrete,
ACI Concr. Int. 19 (7) (1977) 56 ± 60.
[5] I.B. Topcu, Physical and mechanical properties of concretes produced
4. Conclusions with waste concrete, Cem. Concr. Res. 27 (12) (1997) 1817 ± 1823.
[6] I.B. Rasheeduzzafar, A. Khan, Recycled concrete Ð a source of new
concrete, ASTM Cem., Concr., Aggregates 6 (1) (1984) 17 ± 27.
1. Plant processing of recycled aggregate produces rela-
[7] R. Sri Ravindrarajah, Y.H. Loo, C.T. Tam, Recycled concrete as fine
tively smoother spherical particles, which leads to improved and coarse aggregate in concrete, Mag. Concr. Res. 39 (141) (1987)
concrete workability in comparison with natural basalt 214 ± 220.
aggregate concrete with equivalent grading and ratio of fine [8] T.C. Hansen, H. Narud, Strength of recycled concrete made from
to coarse aggregate. crushed concrete coarse aggregate, ACI. Concr. Int. 5 (1) (1983)
79 ± 83.
2. For mixtures with similar volumetric mixture propor-
[9] V.M. Malhotra, Use of recycled concrete as new aggregate, Proceed-
tions and workability, there was no significant difference at ings of the Symposium on Energy and Resource Conservation in the
the 5% confidence level in the 28-day compressive Concrete Industry, CANMET Rep. No. 76-8, CANMET, Ottawa, Ca-
strengths of concrete made with commercial recycled aggre- nada, 1978, pp. 4 ± 16.
gate and strength obtained for normal weight natural basalt [10] Australian Standard AS 2758.1, Aggregates and Rock for Engineer-
ing Purposes: Part 1. Concrete Aggregates, Standards Australia,
aggregate concrete. Higher drying shrinkage values are
Sydney, 1998.
however observed. [11] Australian Standard AS 1012, Methods of Testing Concrete, Standards
3. Tensile strength of recycled concretes as measured by Australia, Sydney, 1996.
the splitting tensile method on specimens cured for periods [12] ASTM C 642-97, Standard Test Method for Density, Absorption, and
up to 365 days, showed no statistically significant reduction Voids in Hardened Concrete, American Society for Testing and Ma-
terials, Philadelphia, PA, 1997.
in strength in the period from 91 to 365 days.
[13] AS/NZS 4456.9:1997, Masonry Units and Segmental Pavers Ð Meth-
4. Concrete containing recycled aggregate as the coarse ods of Test Ð Determining Abrasion Resistance, Standards Australia,
aggregate fraction reduced the abrasion resistance by Sydney, 1997.
about 12% compared to the reference concrete made with [14] P.M. Carrasquillo, Concrete strength testing, in: P. Klieger, J.F. La-
basalt aggregates. mond (Eds.), Significance of Tests and Properties of Concrete and
Concrete-Making Materials, ASTM Spec. Tech. Publ. 169C, Ameri-
5. Examination of water absorption and carbonation rates
can Society for Testing and Materials, Philadelphia, PA, 1994, pp.
have shown little difference between commercially pro- 123 ± 139.
duced recycled concrete and the reference concrete having [15] M. Tavakoli, P. Soroushian, Drying shrinkage behaviour of recycled
a cement content of 242 kg/m3. aggregate concrete, ACI Concr. Int. 11 (1996) 58 ± 61.
6. Ratios of splitting tensile strength/compressive [16] Australian Standard AS 3600, Concrete Structures, Standards Austra-
lia, Sydney, 1994.
strength were found to be in good agreement with conven-
[17] ACI Committee 213R-87, Guide for Structural Lightweight Aggregate
tional values derived for concretes made with normal- Concrete: ACI Manual of Concrete Practice: Part 1, American Con-
weight natural aggregates. crete Institute, Detroit, MI, 2000, p. 213R-18.