Effect of Surface Roughness on the Steel Fibre
Bonding in Ultra High Performance Concrete
(UHPC)
T. Stengel1
Abstract. Micro steel fibres are currently added to UHPC to improve ductility. In
UHPC under load, the high fibre strength is only utilised partly when the fibres are
pulled out of the UHPC matrix. Using steel fibres with a nano- and microroughened surface is one measure to enhance the bond between the UHPC matrix
and the fibre. In this contribution the effect of surface roughness on the bond
behavior of steel fibres is shown. Single fibre pull-out tests were performed to investigate the bond behavior. Aim is to improve the bond behavior so that fibre
pull-out takes place under a high load level with respect to the fibre strength.
1 Introduction
Ultra high performance concrete (UHPC) typically has a compressive strength of
200 to 250 MPa. The high compressive strength, the high durability as well as a
reduced maximum grain size diameter of about 2 mm opens up a field of new construction methods and applications for concrete. However, the extreme brittleness
of UHPC, characterized by explosive failure under critical load, restricts the exploitation of the otherwise good properties of this material. Currently, 1.5 to 3.0 %
by volume of micro steel fibres 0.15 -0.20 mm in diameter with tensile strength of
up to 3.500 MPa are used to improve ductility. Short fibre length is necessary for
the typically slender and light UHPC structures. Therefore the micro steel fibres
usually have an aspect ratio (Lf / Øf) of about 50 to 100. High costs as well as a
considerable contribution to the environmental impact of UHPC are caused by the
aforementioned amount of micro steel fibres [1, 2]. It is therefore aim of a research project to reduce the amount of micro steel fibres necessary or to use more
economic and environmental friendly fibres. The amount of micro steel fibres
T. Stengel
Centre for Building Materials (cbm), Technische Universität München (TUM)
e-mail: stengel@cbm.bv.tum.de
www.cbm.bv.tum.de
372
T. Stengel
could be reduced e.g. by increasing the efficiency of every single fibre. If the fibres could be utilized to a higher extent a smaller amount of fibres would lead to
the same composite performance. This can be achieved for example with an optimized bond between the UHPC matrix and the fibres when the fibres are pulled
out under a load near the fibre breaking force. Fibres with a larger diameter of e.g.
1.0 mm, which are much more economic, would lead to unfeasible fibre length of
50 to 100 mm when the same aspect ratio as mentioned before is considered. In
that case an optimized bond behavior may lead to aspect ratios lower than 50 providing for a feasible application of these fibres. This contribution deals with the
effect of surface roughness on the bond behavior of steel fibres in UHPC. Preliminary single fibre pull-out tests using steel fibres 0.15 and 0.30 mm in diameter
with a roughened surface have shown that the bond between the UHPC matrix and
the fibre could be enhanced significantly [3]. Following that in this study three different surface roughnesses and two different UHPC compositions were investigated together with steel fibres 0.98 mm in diameter.
2 Investigations
The fibre surfaces were mechanically treated by abrasive papers with three different roughnesses (grit 180, 240, 400). The surface roughening was done rubbing a
2 cm piece of abrasive paper for ten times along the fibre axis. Surface roughness
parameters of the fibre surfaces were determined according to ISO 4287 with a
confocal LED microscope using a fifty fold objective (vertical accuracy 2 nm).
The compressive and splitting tensile strength of the plain UHPC was tested on
cylinders with a diameter of 50 mm and an aspect ratio of 1 and 2 respectively.
The load controlled tests were done at a rate of 0.5 MPa•ss in the case of comprescompre
sive strength and 0.5 kN/s in the case of splitting tensile strength. The pull-out
tests were performed according to [3] but using a laser sensor (accuracy:
0.001 mm) for the displacement measurement. An embedded length of 7 mm was
chosen for all pull-out specimens. Plain fibres as well as roughened fibres (grit
180 and 400) together with two different UHPC compositions were tested. From
the pull-out curves the bond strength (assuming constant bond stress over the
whole embedded length) as well as the bond energy (integral of load slip curve between a slip of 0 and 0.5 mm) was derived. The steel fibres used in this study had
a diameter of 0.98 mm and a tensile strength of 2,600 MPa. All single fibres were
alcohol cleaned when fixed in the moulds. Two different UHPC compositions –
with and without PVA fibres – were used, Table 1. The PVA fibres (tensile
strength appr. 780 MPa) had a diameter of about 15 μm and a length of 12 mm.
For both mixtures a well-cement strength class 42.5 and a water cement ratio of
0.28 was chosen. The mixing was done according to [3] but applying a vacuum
(50 mbar) during the last 60 s. Within 60 s after the end of mixing, the workability
was tested with a mortar cone (Hcone = 50 mm, Øupside = 70 mm, Øbottom = 100 mm) on
a glas plate. The workability was 32 cm and 26 cm for C4 and C5 respectively.
Specimens were casted without any compaction.
Effect of Surface Roughness on the Steel Fibre Bonding in UHPC
373
Table 1 Composition of the two UHPC mixtures [kg/m³]
Concrete
Cement
Silica
Fume
Quartz
Powder
Quartz
Sand
Water
Superplast.
PVAFibres
C4
680
138
360
990
166
35
0
C5
674
137
352
975
158
54
3
3 Results
Figure 1a to Figure 1d shows parts of the confocal microscope pictures (each appr.
305 × 229 μm²) of the different fibre surfaces. The height scale lasts from 0 (white
colour) to 10 μm (black colour). It can be seen that with increasing grit number
the depth and the distance of the scratches is decreasing while the number of
scratches is increasing.
a
b
c
d
Fig. 1 Confocal microscope pictures of fibre surfaces: a: plain surface, b: grit 180 roughened, c: grit 280 roughened, d: grit 400 roughened
The roughness parameters parallel and perpendicular to the fibre axis (measured length: 250 μm) of the four fibre surfaces are given in Table 2. Only small
differences in surface roughness were found parallel to the fibre axis while the
374
T. Stengel
Table 2 Roughness parameters of the different fibre surfaces parallel (║) and perpendicular
(┴) to the fibre axis
Roughness
parameter
Plain
Surface
Grit 180
Roughened
Grit 280
Roughened
Grit 400
Roughened
Rq ║ [μm]
0.034
0.051
0.053
0.037
Rq ┴ [μm]
0.293
0.879
0.703
0.516
Rt ║ [μm]
0.224
0.379
0.351
0.326
Rt ┴ [μm]
3.241
5.556
4.793
3.254
Fig. 2 Effect of abrasive paper grit number on roughness parameters of steel fibre surfaces
roughness perpendicular to the fibre axis increased with decreasing grit number
linear dependent, Figure 2. The root mean square roughness was between
0.293 μm for plain and 0.879 μm for grit 180 roughened fibre surface.
The mean compressive and splitting tensile strengths of the two concretes are
shown in Table 3. Concrete C4 had a 9% higher compressive strength than concrete C5; the splitting tensile strength was even 58% higher. This may be due to
the diverse workability which causes different deariation behavior. Further investigations will be done on that. The coefficient of variation was in every case less
than 7% (three specimen tested).
Table 3 Mean Compressive and Splitting Tensile Strength of the UHPCs [MPa]
Concrete
Compressive
Strength
Standard
Deviation
Splitting Tensile
Strength
Standard
Deviation
C4
235
3
15.8
1.1
C5
216
9
10.0
0.7
Effect of Surface Roughness on the Steel Fibre Bonding in UHPC
375
Table 4 shows the derived mean bond properties for the two concretes and the
different fibres.
Table 4 Mean bond properties evaluated from single fibre pull-out tests
Concrete /
Fibre
Bond Strength
[MPa]
Bond Energy
-3
[10 J]
C4 / plain
5.83
31.9
C4 / 180
9.10
78.7
C4 / 400
3.69
31.1
C5 / plain
4.27
39.9
C5 / 180
11.5
93.3
C5 / 400
5.94
45.2
For both concretes the grit 180 roughened fibres showed the highest bond
strength as well as the highest bond energy. This is obviously due to the surface
roughness which was the highest for the grit 180 roughened fibre surfaces. Despite
the lower compressive and splitting tensile strength the addition of PVA fibres
seems to affect the bond behavior of steel fibres beneficially when the fibre surface is rough. Using PVA fibres the bond behavior was enhanced significantly for
both surface roughnesses.
4 Discussion
The surfaces of steel fibres were roughened with abrasive paper of different grits
(180, 280, 400) along the fibre axis. The fibres roughened with the lowest grit
showed the best bond properties. This is due to the highest surface roughness
which was obtained with the lowest grit number. The grit 400 roughened fibres
had a similar bond behavior as the plain fibres when using UHPC without PVA
Fig. 3 Reinforcing effect of PVA fibres within the UHPC matrix surrounding a steel fibre
376
T. Stengel
fibres. An addition of PVA fibres affected the bond behavior of all roughened
fibres beneficially. The higher pull-out load in the case of roughened fibres may
result in a higher volume of UHPC matrix involved in the stress transfer, Figure 3.
When using surface roughened fibres the stress transfer volume has an extension beyond the mean roughness and lasts into the bulk matrix where the micro reinforcing PVA fibres can be found, Figure 3. In that case the PVA fibres can
bridge micro cracks and therefore strengthen the matrix volume involved in stress
transfer. This affects the pull-out behavior in a positive way.
5 Conclusions
Surface roughening done with abrasive paper affects the bond behavior of steel fibres in UHPC significantly. The lower the grit number of the abrasive paper the
higher is the roughness of the steel fibre surface. Surface roughness depends on
the grit number in a linear way. Due to the higher surface roughness steel fibres
treated with grit 180 paper showed the best bond behavior in this study. Bond
properties were enhanced by roughly 200% compared to plain fibres. Even better
bond behavior of the roughened steel fibres is achieved when PVA fibres are
added to the UHPC matrix. Bond properties were enhanced by another 20% compared to the plain UHPC without PVA fibres. This may be because of a micro reinforcing effect of the PVA fibres to the UHPC matrix surrounding the steel fibre.
Due to higher pull-out loads when using roughened steel fibres the stress transfer
volume around the steel fibre is significantly larger than in the case of plain steel
fibres. Therefore PVA fibres can bridge microcracks and strengthen the matrix
within the stress transfer volume. To verify this, further investigations on the effect of PVA fibres like SEM analyses of the interfacial zone will be done.
Acknowledgments. The author would like to thank the German government for the financial support within the NanoTecture research programme and the FRT GmbH for the
surface measurements.
References
1. Stengel, T., Schiessl, P.: Sustainable Construction with UHPC – from Life Cycle Inventory Data Collection to Environmental Impact Assessment. In: Fehling, E., Schmidt, M.,
Stuerwald, S. (eds.) Proc. of the 2nd Int. Symposium on UHPC. Kassel University
Press, Kassel (2008)
2. Stengel, T., Schiessl, P.: Life Cycle Assessment of UHPC Bridge Constructions: Sherbrooke Footbridge, Kassel Gärtnerplatz Footbridge and Wapello Road Bridge. In: Kotynia, R., Gawin, D. (eds.) Proc. of the 6th Int. Conference Analytical Models and New
Concepts in Concrete and Masonry Structures AMCM 2008. University of Lodz, Lodz
(2008)
3. Stengel, T.: Optimisation of Steel Fibre Bonding in Ultra High Performance Concrete.
In: Gettu, R. (ed.) Proc. of the 7th Int. RILEM Symposium on Fibre Reinforced Concrete: Design and Application BEFIB 2008. RILEM Publications S.A.R.L., Bagneux
(2008)