Feature Article. . .

SHOTCRETE PERFORMANCE OF
REFRACTORY CASTABLES
R. G. Pileggi, Y. A. Marques, D. Vasques Filho and V. C. Pandolfelli, Department of Materials
Engineering, Federal University of São Carlos, São Carlos, SP, Brazil
INTRODUCTION 1. SHOTCRETE APPLICATION 2. NOVEL MULTISTAGE
Sprayed concretes were originally devel- PROCESS CHARACTERIZATION
oped for civil construction in the early 20th Refractory castable shotcrete is a multi- TECHNIQUES
century [1, 2]. The high installation rates, stage application process comprising four Based on the successive steps of the shot-
low cost and good final mechanical proper- consecutive steps, i.e., mixing, pumping, crete application process, a multistage
ties of the applied material are responsible spraying and consolidation. The efficiency experimental setup, comprising mixing,
for today’s widespread use of this placing of this application method is usually quanti- pumping, spray shooting and consolidation
process [1, 2]. fied by the rebound loss, which is defined as characterization, was employed to simulate
The application method consists of pump- the percentage of material that fails to stick and evaluate the performance of three dis-
ing either the dry powder (Gunning) or the to the covered surface [1-5]. tinct commercial shotcrete compositions (A,
fluid concrete (Shotcrete) directly from the It is a generally accepted assumption that B and C).
mixer into the pipeline nozzle, where a spray only self-flowing castables (free-flow Details concerning the castables, such as
flow is formed by the injection of high-pres- between 80 and 110 %) [8] are qualified for their chemical composition, polymeric fiber
sure compressed air (both techniques) and shotcrete applications because of their content, shotcrete additive (sodium silicate)
water (only Gunning). Immediately after it pumpability and lower rebound, which and amount of water required for mixing,
reaches and covers the target surface, the result from the absence of large round were supplied by Saint-Gobain, Brazil
concrete sudden loses its fluidity due to the agglomerates (golf balls) [7] when sprayed. (Table I).
action of cement-setting accelerators and However, this assumption is questionable,
coagulant additives also injected into the since contradictory rheological behaviors 2.1 Mixing
nozzle [1-7]. such as dilatancy and pseudoplasticity can A rheometer [8, 9] developed specifically
Rebound losses and dust released into the also afford high-flow compositions, which, to evaluate the rheological behavior of casta-
air are the major problems associated with however, are not necessarily pumpable or bles was used to mix the compositions. The
gunning [1-2]. These difficulties are greatly sprayable [8]. torque profile and the time elapsed to mix 4
reduced when the concrete is pumped in the Pileggi and Pandolfelli [8, 9] demonstrated kg of each castable were recorded online.
fluid state, which justifies the growth of that the mixing and pumping performance of Castable mixing was carried out at 33 rpm,
shotcrete applications [6]. castables is strongly affected by the rheolog- according to the following experimental
The advantages of this method are well ical behavior of these materials. Dilatant setup: (a) dry-powder homogenization
known, but it has only recently become compositions, for instance, are difficult to (time: 180 s); (b) water addition and mixing
common practice to use refractory castable mix and unsuitable for pumping. On the up to the turning point; and (c) castable
shotcrete for lining large areas or repairing other hand, pseudoplastic castables with homogenization after the turning point
damaged surfaces [6, 7]. excessive water content may segregate dur- (time: 120 s).
Despite the continuous technological ing pumping due to their very low matrix In view of the successive steps involved in
advances in shotcrete machines [1, 2], viscosity. Nevertheless, both behaviors the shotcrete process, it is reasonable to con-
sprayable compositions are still developed might be verified in self-flow castables. sider low mixing energy as a basic require-
based on empirical and semi-empirical In addition to the rheological behavior, the ment for sprayable castables. This character-
methods [6]. This discrepancy reveals the action of cement-setting accelerators and istic offers economical and technical bene-
lack of characterization techniques to simu- coagulant additives is strongly influenced by fits, such as the production of highly homo-
late shearing conditions during the applica- the chemical compositions of castables [1-
tion of shotcrete [6] and the need for further 7].
research on the subject. The capability to shotcrete and the materi-
The main objective of the present study al’s effective hardening on the covered sur-
was to reduce this technological gap. To this face are therefore the outcome of the four
end, a set of experiments based on castable consecutive steps (mixing, pumping, spray-
rheometry was developed to evaluate the ing and consolidation) involved in the shot-
shotcrete performance of refractory casta- creting technique.
bles. In addition, commercial shotcrete com-
positions were evaluated based on this new
testing procedure.

Refractories Applications and News, Volume 8, Number 3 May/June 2003 15

geneous materials without the need for pow-
erful and expensive mixers [9], and without
promoting undesirable castable heating [8,
9] and degradation of the polymeric fibers
usually employed in sprayable compositions
[1, 2, 10].
The results reported on herein (Figure 1)
proved that the fiber-free composition
(comp A) required the lowest mixing energy
(ME) and the lowest torque at the turning
point (TTP), whereas the two fiber-contain-
ing systems (comp B and C) showed higher
values for these two parameters. A similar
trend was reported by Salomão et al. [10],
who demonstrated that the mixing energy
and torque at the turning point scaled with
the size and content of the polymeric fibers
within the castable.
Comparing these results to others
described in the literature [8-10], composi- Figure 1 – Mixing behavior of the compositions A, B and C. Note: the inset in the
tions A and C can be classified as easy mix- graph shows the mixing energy required and the maximum torque at the turning
ing systems. The large amount of highly point.
irregular and coarse abrasive silicon carbide
grains in composition B probably explains
its high mixing energy.
2.2 Pumping
The pumping of refractory castables is a
demanding step in which the material flows
over long distances inside pipelines and is
subjected to high shear rates and volume-
restricted conditions [8].
In the present work, the free-flow value of
each composition was measured (ASTM C-
860 adapted for self-flowing castables)
immediately after mixing. Four kilograms of
each composition was then subjected to two
successive shearing cycles (between 2 and
75 rpm) in a rheometer, in which the castable
was subjected to both unrestricted and vol-
ume-restricted conditions [8]. The rheologi-
Figure 2 – Rheological behavior of the compositions A, B and C measured through
cal results are displayed on Figure 2.
shearing cycles (between 2 and 75 rpm) carried out without and with volume restric-
Self-flowability was achieved only for the tion (RV). Note: gray area highlights the torque range suitable for pumping [8].
fiber-free composition A mixed with 7.0
wt.% of water. Compositions B and C dis- ideal pumping range at high shear. Such rhe- castable layer by spray shooting the material
played a similar low fluidity suitable for ological behavior does not, in effect, inhibit onto the surface to be covered. In practice,
castables applied under vibration [8, 9]. the flow inside pipelines, but high energy the nozzle’s conical geometry and the injec-
would be required to achieve this purpose. tion of high-pressure compressed air and
Based on the traditional cone flow selec-
That being the case, the shotcrete’s perform- additives into the pumped material cause a
tion criteria, only composition A was suit-
ance would be negatively affected by the high-speed spray flow against the surface,
able for pumping and shotcrete applications
undesirable outcome resulting from exces- which adheres to and consolidates on the
[7, 8]. However, both compositions A and C
sive shearing energy, such as excessive heat- surface (Figure 3).
were pumpable due to their pseudoplastic
ing, irregular flow (plug-flow), segregation Ideally, the castable spray flow is com-
behavior, which matched the ideal torque
and structural damage (fiber rupture, particle posed of individual high-speed granules
range for pumping [8] under both shearing
fracturing) [6, 8]. (coarse aggregate covered with a matrix
conditions (unrestricted and restricted vol-
ume). 2.3 Spray Shooting layer). It is reasonable to assume that, in
Regardless of its pseudoplastic character- The essence of the shotcrete application order to ensure the granules’ adhesion, the
istics, composition B greatly surpassed the process consists of generating a compact damping behavior of the matrix layer must

16 May/June 2003 Refractories Applications and News, Volume 8, Number 3

 These flaws have generally been attributed
to the inadequacy of the pumped material to
be sprayed [7] or to some of the multiple
parameters involved in generating the spray
[5]. However, the techniques conventionally
applied to characterize the shotcrete per-
formance of refractory castables do not pro-
vide accurate information about the phe-
nomena involved in the spray shooting
process, since they are based only on post-
application analyses or on rheological evalu-
ations of the matrix [6].
A laboratory test developed specifically to
characterize the shooting performance of
castables was recently proposed by Pileggi
and Pandolfelli [6]. This method involves
Figure 3 – Schematic drawing of castable spray shooting and the material’s consol-
the use of a rheometer to simulate the shear
idation onto the surface.
conditions during the castable’s application.
The method allows for easy quantification of
the impact of any specific factor (additives,
water content, rheological nature, etc.), on
the shotcrete’s performance. However, this
novel technique does not allow for analyses
of the sticking tendency of shot-sprayed
materials or of strength increases in the con-
solidated layer. For this reason, two alterna-
tive tests are proposed herein to improve the
rheometer-based characterization method.
2.4. Sticking Test
After mixing, a conical device was intro-
duced into the mixing bowl in order to pro-
vide a surface target. Additionally, the plan-
etary mixing paddle was replaced by anoth-
er specifically designed for shooting the
castable at a high velocity against the coni-
cal target (Figure 5 (A)).
For the sticking test, the prepared castable
(2 kg) was also homogenized at 33 rpm for
30 s. The shotcrete additive (1 wt. % of an
aqueous solution composed of 0.4 wt. %
sodium silicate and 0.6 wt. % of water) was
injected at the end of the homogenization
step and the revolution speed was suddenly
Figure 4 – Schematic drawing of the different problems that may occur during the escalated to 75 rpm, held there for 10 s, and
shotcreting of refractory castables. Note: light gray represents the ideally consoli- then reduced to zero (Figure 5 (B)). The
dated castable; the dark gray, the distinct consolidation problems. rebound loss was defined as the weight per-
centage of non-adhered material.
be able to dissipate all the kinetic energy In addition to its high application rate, the Large rebound losses were recorded for
associated with the granules when they shotcrete process often involves undesirable the self-flowing pumpable composition A,
strike the surface. material losses (> 15 wt. %) caused by both with and without added fiber, whereas
The dissipation of energy is aided by rebounding or fragmentation of the first compositions B and C actually displayed
reducing the granule size or increasing the deposited layer when successive applica- good sticking features (Figure 6). These
matrix/aggregate ratio. It is worth mention- tions are required [1-7]. Problems associated results confirmed the hypothesis that high
ing that the characteristics of the matrix with the consolidated material, such as creep fluidity is not directly related to shotcrete
layer also control the adhesion of granules to strain, sliding, detachment and lamination / performance. However, the addition of fiber
the surface, as well as its homogenizing cracking of the applied layer, are also com- unexpectedly increased the rebound loss in
capacity and its consolidating nature. mon [1-7] (Figure 4). composition A.

Refractories Applications and News, Volume 8, Number 3 May/June 2003 17

 Considering that the same shotcrete addi-
tive (sodium silicate) was used in each com-
position, it can be concluded that the chemi-
cal composition and physical characteristics
of the tested materials strongly influenced
the material’s adhesion onto the surface.
2.5 Consolidation Test
In order to evaluate the reaction profile
during the shotcrete application process, the
sodium silicate solution was injected into the
castables (2 kg) at the end of the mixing
homogenization step. Immediately follow-
ing this addition, the revolution speed was
abruptly escalated to 75 rpm, held for 10 s,
then reduced to 20 rpm and held there for
120 s to evaluate the torque increase under
low shear conditions (Figure 5 (B)).
In theory, the first 10 s at 75 rpm simulates
the high-speed spray-shooting process. Only Figure 5 – (A) Schematic Drawing for the conical device employed in the sticking
composition A (fiber) reacted instantaneous- test; (B) revolution speed set-up employed in the sticking (only until the 75 rpm / 10
ly (Figure 7), resulting in a positive initial s step) and in the consolidation tests (75 rpm for 10 s + 20 rpm for 120 s). Note: the
reaction rate (estimated through the angular polymeric screen prevents the castable from sliding over the internal surface of the
coefficient calculated from the best linear cylindrical bowl.
fitting between torque and time actually held
at 75 rpm in the high speed step), as demon-
strated on Table II. The other systems
remained almost unchanged in the first 10 s,
probably due to a balance between the
effects promoted by the additional liquid and
the coagulating action of the shotcrete addi-
tive.
The addition of fiber accelerated the coag-
ulating action promoted by the sodium sili-
cate in composition A, possibly due to the
friction promoted by the fibers [10]. This
behavior is the main reason for the highest
rebound loss reached by this system; in fact,
the literature reports that rebound losses are
minimized when the additives react more
slowly in this initial stage [4].
High initial reaction rates harden the mate- Figure 6 – Rebound loss measured in the sticking test for the compositions A, B and
rial inside the nozzle, hindering the forma- C. The shotcrete additive was sodium silicate (0.4 wt %). Polypropylene fibers were
tion of spray and leading to the formation of also added in equal amounts to compositions A, B and C.
large agglomerates [7]. Moreover, hardened
matrixes cannot dampen the collision impact increase at this stage (due to the action of the driving force that increases rebound losses
between the granules and the surface. additive) and after reducing the revolution in refractory castable shotcrete.
speed from 75 to 20 rpm in the rheometer.
In order to produce a dense coating on the Regardless of the rebound, the consolidat-
surface, there must also be a lapse of time According to this parameter, composition ed material must also be strengthened over
before the individual granules begin their A (fiber) showed the shortest resting time, the surface to avoid the creep strain resulting
setting reactions [4]. This condition ensures followed by the fiber-free composition from its own weight, which allows thick lay-
proper binding and homogenization between (comp A), thus further reinforcing their high ers of coating to form. In the experimental
successively applied layers, preventing the rebound loss. Both low-rebound composi- setup proposed herein, the maximum torque
material from sliding and reducing rebound tions B and C displayed a similar resting in the final step (20 rpm / 120 s) describes
losses. time. the material’s consistency after its applica-
This lapse of time was measured indirect- High initial reaction rates associated with tion.
ly through the resting time, defined here as short resting times may be assumed to be the Composition A, with or without added
the time elapsed before restarting the torque fiber, resulted in high maximum torque val-

18 May/June 2003 Refractories Applications and News, Volume 8, Number 3

Figure 7 – Reaction profile measured through the consolidation test for the compo-
sitions A, B and C. Note: sodium silicate (0.4 wt %) was used as the shotcrete addi- Measures Young’s Modulus,
tive. The dark gray arrow indicates the rupture that occurred in composition A Shear Modulus and Poisson’s
(fiber) under shearing. The reaction rate was estimated through the angular coeffi- ratio of various objects by
cient calculated from the best linear fitting between torque and the real time at 75 analyzing their vibrational
rpm (gray area) in the high speed step. behavior following an impulse
excitation.

Available in configurations
adapted for room temperature,
elevated temperature (max.
1100oC), and automated product
testing.

Complies with ASTM C1259,

C666, C215 and E 1876.

Call today
for a free evaluation test.

J.W. LEMMENS
Dynamic Materials Testing Instruments
Website: www.grindosonic.com

ues during this step. Conversely, composi- combination that provided a resting time
USA
tions B and C did not afford significant max- similar to that of compositions B and C, and 3466 Bridgeland Drive
imum torque values, which may have inhib- a final torque such as that of composition A. St. Louis, MO 63044-2606
ited the formation of thick layers of coating. Tel. 314-770-2200
3. FINAL REMARKS Fax 314-770-2262
The fiber-containing composition A devel-
The new testing procedure proposed here 1-800-437-3884
oped excessive rigidity, evidenced by the
demonstrated that high fluidity is not the E-mail: jwlinc@grindosonic .com
rupture of its structure by shearing during
most important feature for good shotcrete
the test (20 rpm). In practice, such a lack of
plasticity may result in lamination or in the
performance. Instead, low mixing energy EUROPE
and low torque values at high shearing rates Geldenaaksebaan 456
rupture of the consolidated layer from the
under volume-restricted conditions were B-3001 Leuven l BELGIUM
impact caused by spraying on successive
found to be better requirements for shotcrete Tel. +32-16 391500
layers of material.
of materials. Fax +32-16 400182
In fact, a suitable behavior would have
been achieved with a composition/additive E-mail: [email protected]

Refractories Applications and News, Volume 8, Number 3 May/June 2003 19

 According to the sticking and consolida- 7. N. Cassens Jr., R.A. Steinke, R.B. commercially available refractories; ii)
tion test results, rebound losses are reduced Videtto, "Shotcreting Self-Flow establish the mechanisms of hydration for
when the shotcrete additive does not cause a Refractory Castables", Proceedings of refractories used in combination with water
high initial reaction rate or a too brief rest- the Unified International Technical cooling jackets and develop solutions for
ing time. High maximum torque after con- Conference on Refractories increasing their life in service; iii) develop
solidation is also required for placing thick (UNITCER’97), 531-544, (1997). new refractories with improved corrosion
layers. 8. R.G Pileggi, V.C. Pandolfelli, and hydration resistance. We are also devel-
In view of all these factors, composition C "Rheology and Particle Size oping new reactive oxide powders used in
displayed good shotcrete performance, Distribution of Pumpable Refractory refractory technologies to allow the forma-
though only for thin coating layers. Castables", Am. Ceram. Soc. Bull., 80 tion of a ceramic matrix in refractory casta-
Composition A resulted in large rebound [10], 52-57, (2001). bles at the operating temperatures of industri-
loss due to its high reaction rate, which was 9. R.G. Pileggi, A.R. Studart, V.C. al furnaces, much lower than in steel tech-
further increased with the addition of fiber. Pandolfelli, "How Mixing Affects the nologies. The new developed refractories are
On the other hand, composition B required a Rheology of Refractory Castables, tested against traditional bricks in the corro-
lot of energy for mixing and proved unsuit- Part 1", Am. Ceram. Soc. Bull., 80 [6], sive gases and metal fumes, similar to those
able for pumping. 27-31, (2001). from the industrial smelting furnaces, before
10. R. Salomão, V.C. Domiciano, R.G. being selected for pilot and industrial scale
The multistage characterization technique
Pileggi, M.D.M Innocentini, V.C. experiments.
developed to evaluate the shotcrete perform-
ance of refractory castables was a further Pandolfelli, L.R.M. Bittencourt, R.P. SERVICES TO INDUSTRY
step used for analyzing the mixing, pumping Rettore, "Rheological Behavior of UBCeram maintains close working rela-
and shooting behavior of these materials. Refractory Castables Containing tionship with a number of BC industries, as
Polymeric Fibers" (In Portuguese). well as research and industrial organizations
ACKNOWLEDGEMENTS Proceedings of the 46o Annual throughout Canada and the US. The majori-
The authors are greatly indebted to the Meeting of the Brazilian Ceramic ty of the research programs are of a collabo-
Brazilian research funding institutions Society, 656-668, (2002). R AN rative nature, including cash and in-kind con-
FAPESP and CNPq, and to Saint Gobain /
tribution from industry. Additionally,
Brazil, for their support of this research
Institute Profile Continued from Page 14 UBCeram is involved in a number contract
work.
services in the areas of expertise, obviously
REFERENCES ries preparation laboratory), two large-vol- related to characterization of ceramics and
1. G.D. Yoggi, "The History of Shotcrete ume air furnaces for processing and charac- refractories but also delivery of trial samples
– Part I", Shotcrete Magazine, Fall, terization of refractories. of the novel materials and techniques devel-
28-29, (2000). The main research project is “Refractories oped through UBCeram research programs.
2. T. Telford, "An Introduction to for Non-Ferrous Metals Smelting” spon-
THE FUTURE
Sprayed Concrete", Internal Report sored by NSERC and five industrial part-
ners. The partnership in this project has an By virtue of changing times and also
From the Sprayed Concrete
equilibrated structure for improving the exis- because of the geographical location of
Association, 33 p., (1999).
tent processes and creating new technologies Vancouver, we are encouraged to “work
3 L.R. Prudêncio Jr., "Accelerating locally and think globally”. The vision of
Admixtures for Shotcrete", Cement for non-ferrous metallurgy. The industrial
partners are TeckCominco Metals Ltd. and UBCeram is that it will become a major
and Concrete Research, 20, 213-219, ceramics laboratory, servicing primarily
(1998). Inco Ltd., two of the largest producers of
Zinc-Lead and respectively Nickel in the Western Canada and Northwest US, but also
4. C. Paglia, F. Wombacher, H. Böhni, active on a world scale, in particular within
world, Hatch Associates Ltd., one of the
"The Influence of Alkali-Free and the Pacific Rim countries. Our research col-
largest metallurgical design companies in the
Alkaline Shotcrete Accelerators laborations with Asian countries increased
world, and two refractory manufacturers,
Within Cement Systems: I. rapidly in the recent 2-3 years. We will con-
RHI, the largest refractory manufacturer in
Characterization of the Setting Time", tinue to build on our expertise in structural
the world and the main supplier of refracto-
Cement and Concrete Research, 31, ceramics, ceramic coatings, ceramic com-
ries for non-ferrous furnaces, and Clayburn
913-918, (2001). posites and refractories. In the short term,
Industries, the only 100% Canadian owned
5. M. Jolin, D. Beaupré, S. Mindess, UBCeram will link together professionals
refractory producer in Canada, also having
"Tests to Characterize Properties of from British Columbia and the other
manufacturing plants in China and USA.
Fresh Dry-Mix Shotcrete", Cement provinces of Canada to form a pool of
and Concrete Research, 29, 753-760, The main objective of this Consortium is to
expertise capable of addressing most of the
(1999). address specific issues regarding manufac-
local and worldwide needs of users and man-
turing and use of refractories for zinc-lead
6. R.G. Pileggi, Y.A. Marques, D. ufacturers of ceramics and refractories. Our
and nickel-copper smelting. Our research is
Vasques Filho, V.C. Pandolfelli, "Wet- web site will serve as a focal communication
concentrated on three directions: i) establish
Shotcrete of Refractory Castables", and information tool for the UBCeram relat-
the mechanisms of failure for refractories
Am. Ceram. Soc. Bull., 81 [10], 51-56, ed research, teaching and service.
used in industrial furnaces and develop solu-
(2002).
tions for increasing the life in service for the
Continued on Page 36

20 May/June 2003 Refractories Applications and News, Volume 8, Number 3