Cerâmica 65 (2019) 426-431

http://dx.doi.org/10.1590/0366-69132019653752687 426

Recycling ceramic waste as a raw material in sanitary ware production

(Reciclagem de resíduo cerâmico como matéria-prima na produção de

louças sanitárias)
T. H. Silva1 , A. C. M. Castro1, F. C. Valente Neto1, M. M. N. S. Soares1, D. S. de Resende1, A. C. S. Bezerra1*
1
Centro Federal de Educação Tecnológica de Minas Gerais, Av. Amazonas 5253, 30421-169,
Belo Horizonte, MG, Brazil

Abstract

This study characterized discarded sanitary ware waste (SWW) that could not meet quality requirements. To understand the
composition of such SWW, samples were collected, processed, and characterized. This SWW was used to substitute granite in
ceramic slips, and the final recycled sanitary ware compositions of 5%, 10%, 25%, 50% and 100% were obtained. Then, the mixtures
were evaluated based on density, rheology, linear shrinkage, water absorption, and flexural strength. The results for viscosity and
pyroplastic deformation in specimens containing the ceramic waste were better than the reference slip used in a production line.
Results for density, water absorption and linear shrinkage showed no significant difference from the control samples. Although the
replacement of more than 5% of granite reduced the flexural strength, no tested specimen was below the minimum value required
by the industry.
Keywords: ceramic, slip casting, sanitary ware, waste, sustainability.

Resumo

Este estudo caracterizou o resíduo de louças sanitárias descartadas que não atendiam aos requisitos de qualidade. Para compreender
a composição de tal resíduo, as amostras foram coletadas, processadas e caracterizadas. Este resíduo de cacos de louça cerâmica
moídos foi usado para substituir o pó de granito utilizado nas massas cerâmicas em 5%, 10%, 25%, 50% e 100%. As misturas foram
avaliadas quanto à densidade, reologia, retração linear, absorção de água e resistência à flexão. Os resultados de viscosidade e
deformação piroplástica em barbotinas contendo resíduo foram melhores que a barbotina padrão empregada em um processo de
produção. Resultados de densidade, absorção de água e retração linear não mostraram diferenças significativas em relação às
amostras de controle. Ainda que a substituição de granito superior a 5% tenha reduzido a resistência à flexão, nenhum corpo de
prova testado apresentou resistência abaixo do mínimo exigido pela indústria.
Palavras-chave: cerâmica, colagem barbotina, louça sanitária, resíduo, sustentabilidade.

INTRODUCTION it is classified as fired clay, this ceramic waste is a material

with high strength and resistance to wear [1, 3]. In addition,
A significant type of industrial waste that requires according to the NBR 10004:04 solid waste classification
attention is the ceramic waste, with the majority of this waste [4], this ceramic waste is classified as inert, which facilitates
coming from the disposal of defective and broken ceramic its handling for reuse. On the other hand, the rejected
pieces [1]. Studies of possible applications of the waste ceramic pieces require comminution prior to reuse, and
from the sanitary ware industry can, in addition to reducing although this recycling process involves initial investment
the prevalence and environmental footprint of crushing and energy consumption to grind and mill the waste, there
yards, help reduce the cost of waste disposal. From the are examples of recycling initiatives that are economically
ecological point of view, the recycling of industrial wastes viable [5]. Several papers studied applications for sanitary
is becoming a priority, as a result of an ever-increasing ceramic waste, especially in the field of construction, as an
environmental awareness [2]. In producing sanitary ware, aggregate to concrete [1, 3, 6-8]. However, the recycling of
despite any improvements made to the process, there is this type of waste as a raw material in the process responsible
always a percentage deemed inappropriate for sale. The for generating it has not been thoroughly investigated,
defects in rejected wares include imperfect shapes, cracks, as there are no applicable standards yet on its reuse. The
and visual damage to the glaze, defects which do not alter recycling of this ceramic waste back into its production
the intrinsic properties of the ceramic material. Therefore, as process could have some advantages: it does not require
high transportation costs; it could also decrease expenses in
raw material consumption and waste disposal in landfills.
*[email protected]
https://orcid.org/0000-0003-1670-2376
Independent of the type of ceramic being manufactured,
427 T. H. Silva et al. / Cerâmica 65 (2019) 426-431

the initial slip (or another kind of mix) includes three 3000 g of dry raw materials, 1000 mL of tap water and 6
types of materials, hence the name ‘triaxial ceramics’, a mL of dispersant (sodium silicate). The suspensions were
plastic material (clays) which facilitates casting, a fluxing pulverized in a ceramic ball-mill operating at 70 rpm for
agent (feldspar) that assists melting and helps to vitrify the 1.5 h. Test specimens were made from each slip in a casting
structure, and a non-plastic material (quartz) to provide process using plaster molds. In order to evaluate how the WP
structural strength [2]. Therefore, one of the first steps in substitution degree could influence the ceramic slips, tests of
recycling wastes into sanitary ceramics is determining which density, residue percentage (44 µm sieve), and rheology were
of the three behaviors the waste exhibits. When it comes to performed, including tactile and detachment evaluations.
sanitary ware production, focusing on materials that are used Furthermore, the physicomechanical properties of ceramic
in the ceramic slip, numerous parameters must be verified. specimens were determined through water absorption, three-
These are typically achieved through characterization that point flexural strength for green specimens (before sintering),
determines the particle-size distribution, composition, linear shrinkage and pyroplastic deformation. The density
crystalline phases and material morphology [3]. Furthermore, of the ceramic slips was determined by correlating weight
the ceramic slips must meet standards regarding density and and volume for 100 mL of a ceramic slip in a volumetric
viscosity, and the final product must be in accordance with flask. The residue percentage, which represented the fraction
standards for linear shrinkage, flexural strength, and water of solids from a suspension that was retained after size
absorption [9-11]. This study evaluated the application of exclusion through a 44 µm sieve, evaluated the degree of
sanitary ware waste as a substitute for one of the standard raw slip pulverizing.
materials used in ceramic slips for sanitary ware production.
It is important to note that this study was performed by Table I - Formulation of ceramic slips.
recycling a sanitary ware waste within an industrial process [Tabela I - Formulação das massas cerâmicas.]
responsible for its production. Ceramic Substitution Granite WP
slip degree (%) (%) (%)
EXPERIMENTAL
Control 0 22.50 0.00
Samples of rejected sanitary ware pieces were collected WP-5 5 21.38 1.12
from an industry’s landfill; these were crushed, and then WP-10 10 20.25 2.25
pulverized by a planetary ball-mill with jar and beads of
WP-25 25 16.90 5.60
zirconia. The particle size distribution and humidity of the
waste powder (WP) were determined, the first through laser WP-50 50 11.25 11.25
granulometry (LG), and the second according to the NBR WP-100 100 0.00 22.50
NM 24 standard [10]. The powder was also characterized by Note: in each slip, 13.50% of clay and 64.0% phyllite were used.
X-ray fluorescence spectrometry (XRF), X-ray diffraction
(XRD), and scanning electron microscopy (SEM). The LG Rheology is one of the most important variables for
analysis was performed under ambient conditions using ceramic slips, as it influences the flow of the material,
water as a fluid, without a dispersant, 60 s of ultrasonication which exerts a significant impact on pumping, thus affecting
treatment and obscuration set to 17%. The XRF was production [12]. The analysis on rheology consisted of
performed under a regular air atmosphere and a collimator tests of viscosity, setting rate and drying time. The first was
of 10 mm. The XRD measurements were performed using a measured using a dial-reading Brookfield viscometer, RV
copper X-ray tube with 40 kV accelerating voltage and 30 series, equipped with disc spindle number 2; for measuring
mA current in the continuous scanning mode with a 2θ range the viscosity two speeds were applied, 50 rpm until the
from 10º to 80° at a rate of 2 °/min. The SEM analyses were reading stabilized, let the slip rest for 5 min, then measuring
performed with low vacuum and magnification capacity up it at 2.5 rpm. These values were recorded and converted
to 30000 times and a digital zoom factor of 4x, using an to viscosity using the spindle factors associated with each
accelerating voltage of 15 kV and a backscattered electron rotating speed [13]; also, the ratio between the two measured
(BSE) detector. viscosities was calculated to determine the thixotropic index
Subsequently, to determine which raw material the WP for each slip [14]. The setting rate defines the speed at
should replace in testing, the WP and the raw materials used which a ceramic slip is accumulated on the porous surface
by the industry were separately shaped into cones, which of a plaster cast, while the drying time is the time required
undergone the industrial firing process to compare their for excess water to be drained as the mass on the plaster
after-sintering structure. Once determined which material to cast loses its pearl shine. For this test, a plaster cast with an
be substituted, different formulations of ceramic slips were internal surface area of 509 cm² was filled with the ceramic
prepared and labeled according to the degree of substitution slips, and after 2 min of setting the cast was drained. At
of granite by the WP, therefore the labels were WP-5, WP- this point, the deposited mass still had a pearl shine, which
10, WP-25, WP-50 and WP-100. All formulations applied slowly became matte as the plaster absorbed water; the time
in the tests, including the control, are described in Table I. taken for the ceramic mass to present a completely matte
Each slip was prepared as a water-based suspension, using finish was recorded as the dying time. Simultaneously, the
 T. H. Silva et al. / Cerâmica 65 (2019) 426-431 428

detachment and tactile evaluations were performed to assess 5 100

the difficulty in removing a layer of accumulated material Distribution of particles

Fraction of particles

Cumulative passing
Cumulative passing
from the cast and to subjectively appraise consistency of the 4 80
ceramic slip.

(volume %)

(volume %)
Finally, to analyze the physicomechanical properties of 3 60
the fabricated materials, three types of test specimens were
produced. Cylinders with 200 mm in length and 15 mm 2 40
in diameter were used for the three-point flexural strength
test. Tabular specimens of 290 mm length were measured 1 20
immediately after slip casting and the values were compared
to their lengths after drying and sintering for linear retraction. 0 0
The third specimen, for testing pyroplastic deformation and 0.1 1 10 100
water absorption, had a length of 290 mm, with double the Particle size (µm)
thickness at the initial 100 mm (head) and tabular shaped Figure 1: Particle size distribution curves of waste powder.
for the remaining length (tail). The pyroplastic deformation [Figura 1: Curvas de distribuição granulométrica do pó do resíduo.]
test consisted of placing the head of the specimen between
two layers of refractory material, while the tail was in (Al6Si2O13, COD 9001567) [18] was also detected; this is a
suspension; this allowed the tail to deform as liquid-phase stable phase accountable for mechanical strength in sintered
was produced during sintering; an increase in deformation clay minerals, and is formed starting at 950 ºC up to 1100 ºC
represented an increase in the presence of liquid-phase, [19], which are compatible with the firing curve in use by the
and thus correlating to flux material content in the mixes industry that has a peak temperature of 1200 ºC. All phases
[15]. For water absorption tests, an 80 mm section of the were identified using the software Match! 3.
specimen’s tail was cut, weighted while dried, placed in The particle morphology of the sanitary ware waste was
boiling water for 2 h, rested in that water for 24 h, and then observed through SEM analysis. The micrographs shown
weighted again [11]. in Fig. 3 were in accordance with the LG analysis, as the
material presented heterogeneous particle size distribution,
RESULTS AND DISCUSSION with a significant fraction of lamellar crystals with sharp
edges. Additionally, some particles had brighter shades of
After the comminution process, the powder generated grey, which indicated a denser phase as a result of a higher
from the ceramic shards had white color and a silky touch; atomic number; this was because the displayed brightness is
also, it was very fine and presented a heterogeneous particle- a function of the material interaction with electron beam, and
size distribution. The humidity values obtained were denser materials tend to have a higher signal in backscattered
satisfactory, since none exceeded the 5% limit [10]. The electron (BSE) analysis [20]. The test using firing cones
results from the LG analysis, presented in Fig. 1, showed made from each material demonstrated that the WP acted as
that 90% of the sample particles were less than 34.9 µm a fluxing agent during sintering. In this test, cones made from
in diameter, which had an average of 13.57 µm. These phyllite and clays resulted in a ‘sandy’ texture that was easily
dimensions were in accordance with those for standard raw crumbled apart, which is characteristic of refractory materials.
materials used in the ceramic industry [16]. The result of In contrast, cones made from granite and WP presented flux
oxide composition obtained by energy dispersive X-ray material behavior, with very consistent structure and smooth
(EDX) spectroscopy is listed in Table II. As expected, the texture, vitreous shine on the former and matte finish on the
composition of the WP reflected a mix of the raw materials latter.
used by the industry, with the addition of oxides from dyes Table III shows that density values for all the ceramic
used for glazing. Fig. 2 shows a well-defined presence of slips; they were in accordance with the range commonly
quartz phases (SiO2, Crystallography Open Database - COD used in manufacturing processes, which is around
9013321) [17] in the sample diffractogram. Mullite phase 1.80 g/cm³ [21]. Additionally, when granite was replaced

Table II - Chemical composition of the sanitary ware waste and raw materials (in wt%).
[Tabela II - Composição química do resíduo de louça cerâmica e matérias-primas (% em massa).]
Material SiO2 Al2O3 K2O Fe2O3 CaO TiO2 ZrO2 ZnO BaO Cr2O3 Others
Clay A 50.80 31.40 4.56 10.22 - 1.95 0.33 0.05 0.50 - 0.25
Clay B 63.36 30.45 0.96 2.23 0.15 2.35 0.44 - - - 0.06
Phyllite 56.73 24.34 8.56 5.46 - 3.22 0.23 - - - 1.46
Granite 71.40 13.39 9.32 1.88 1.58 - 0.41 0.01 - - 2.01
WP 64.53 21.47 5.42 3.51 1.50 1.43 1.08 0.58 0.27 0.06 0.15
 429 T. H. Silva et al. / Cerâmica 65 (2019) 426-431

3000 q standard values exhibited by the control samples, as can be

q - Quartz
seen in Table III. The results showed that WP incorporation
2500
decreased the initial viscosity in most formulations, with
Intensity (a.u.)

µ - Mullite
2000 the exception of WP-100; similarly, the thixotropy factor
decreased with WP increase. In other words, as more WP was
1500 added to the mix, it experienced a lower increase in viscosity
1000
over time [22]. This is especially important for the production
q
µ line, where the stabilization of ceramic slips is crucial and
500 µ q usually maintained through constant stirring. In contrast,
µ µµµqq µ µ µqµq µ q when comparing setting rates, there was a direct relation to
0 WP incorporation, with the rate increasing as a higher WP
0 10 20 30 40 50 60 70 80 90
fraction was used in the mix. The results for drying times
2q (degree)
showed an initial decrease for formulations WP-5 and WP-
Figure 2: X-ray diffraction pattern of the sanitary ware waste. 10; this drying time remained constant for WP-25, increased
[Figura 2: Difratograma de raios X do resíduo de louça cerâmica.]
for WP-50 and reached the same value for the control slip
in WP-100. Moreover, during the tactile evaluation of these
by the WP in all formulations tested, the amount of residue ceramic pastes, all samples were considered good and firm,
retained on the 44 µm sieve was less than that observed which indicated that the suspensions did not become soft,
for the control, possibly due to WP being pulverized prior pasty, or dry immediately after casting, and some plasticity
to its characterization. Measurements of the viscosity and remained in the mixes. It is worth noting that, for this later
rheological properties showed significant deviation from the evaluation, the ceramic slip that most resembled the control

a) b) c)

Figure 3: Scanning electron microscopy images of the sanitary ware waste.

[Figura 3: Imagens de microscopia eletrônica de varredura do resíduo de louça cerâmica.]

Table III - Characterization of ceramic slips and sanitary ware pieces.

[Tabela III - Caracterização das massas cerâmicas e peças de louça sanitária.]
Attribute Control WP-5 WP-10 WP-25 WP-50 WP-100
Density (g/cm ) 3
1.81 1.81 1.81 1.80 1.81 1.80
Residue 44 µm (%) 9.1 6.7 8.5 4.7 5.6 4.9
Viscosity 50 rpm (cP) 640.0 429.6 480.0 496.8 464.8 724.8
Viscosity 2.5 rpm (cP) 4000 2452 2496 2416 2080 3088
Thixotropic index 6.2 5.7 5.2 4.9 4.5 4.3
Setting rate (g/cm²/min) 0.152 0.153 0.163 0.157 0.158 0.194
Drying time (min:sec) 1:40 1:35 1:25 1:25 1:29 1:39
Good and firm Good and firm Good and firm Good and Good and firm Good and firm
Tact and detachment
(plastic) (plastic) (plastic) firm (plastic) (little plastic) (little plastic)
Drying shrinkage (%) 2.1 2.1 2.1 2.1 2.1 2.1
Firing shrinkage (%) 9.5 10.2 10.2 10.2 10.6 10.6
Total shrinkage (%) 11.4 12.1 12.1 12.1 12.5 12.5
Pyroplastic deformation (mm) 63 64 66 58 57 55
Water absorption (%) 0.09 0.05 0.10 0.11 0.06 0.11
 T. H. Silva et al. / Cerâmica 65 (2019) 426-431 430

was the WP-5. powder (WP). Although drying time, residue percentage and
While the drying shrinkage remained the same for linear shrinkage were similar to the control, flexural strength
all specimens, the total shrinkage increased for the WP showed a decrease as more WP was added to the ceramic
formulations as a result of higher firing shrinkages. These slip. On the other hand, the rheology tests showed that
results reflected the ones from firing cone tests, restating that adding WP boosted rheological performance as a result of a
the WP acted as a flux material, which formed liquid-phase lower thixotropy index. The flexural strength tests indicated
during sintering and filled open pores, thus increasing the that all formulations had values above the minimum strength
ceramic densification. Although the granite is also used as required, with an excellent performance by the WP-5, which
a fluxing agent by the industry, the smaller particles of the was statistically equivalent to the control. This is a decisive
pulverized WP could have been responsible for easing liquid- factor when it comes to applying recycled products for the
phase formation [23]. The results for pyroplastic deformation production of new materials, where the goal is to produce
showed no significant deviations from the control for a similar or better product. The results showed that the WP
formulations WP-5 and WP-10, while WP contents of 25% can replace up to 100% of the granite in formulations and
and above diminished the pyroplastic deformation. The fact still meet the industry’s minimum requirements; however,
that water absorption results did not show clear correlation to a 5% substitution exhibited the best results and reflected
WP levels was due to standardized test of water absorption, the highest similarity to the mix design employed. Given
which involves keeping samples in boiling water, not that the fraction of sanitary ware that is discarded is usually
being able to ensure complete saturation of test specimens slightly less than 5% of total production, applying the
[24]; therefore, the values obtained from this test did not WP-5 formulation as a recycling process could significantly
necessarily reflect a difference in surface porosity among the reduce the volume of broken pieces in the crushing yard and
formulations. Nevertheless, all specimens were far below the landfill; in addition, higher degrees of substitution could be
limit of 0.5% for water absorption established by standard enough to consume all the discarded pieces, thus reducing
organizations [11]. environmental impacts and costs of disposal.
Regarding three-point flexural strength of raw pieces, the
incorporation of WP reflected a decrease in the average values ACKNOWLEDGMENTS
(Fig. 4). Among the tested WP formulations, the one with the
lowest WP content (WP-5) presented the highest flexural The authors would like to express their gratitude to
strength, which was statistically the same as the one obtained Minas Gerais State Research Foundation (FAPEMIG
by the control (4.5 MPa). In order to avoid breakages during APQ3739-16), the National Council for Scientific and
handling, the industry usually adopts 2 MPa as the minimum Technological Development (CNPq), to the Brazilian
flexural strength for raw pieces [9]; consequently, regardless Federal Agency for Support and Evaluation of Graduate
of the decreasing in strength, all WP formulations complied Education (Coordenação de Aperfeiçoamento de Pessoal de
with the minimum requirement for flexural strength in an Nível Superior - Brasil - CAPES) for their financial support
industrial scenario. in scientific initiation and master’s degree scholarships
(Finance Code 001); and to Santa Clara Louças Sanitárias
Ltda for providing the materials and some of the equipment
3-point flexural strength (MPa)

5.0 used in this study.

4.5
4.0
REFERENCES
3.5
[1] A. Halicka, P. Ogrodnik, B. Zegardlo, Constr. Build.
3.0
Mater. 48 (2013) 295.
2.5
[2] V. Medri, E. Landi, Ceram. Int. 40, 1 (2014) 307.
2.0 [3] C. Medina, M.I. Sánchez de Rojas, M. Frias, Cem. Concr.
1.5 Comp. 34, 1 (2012) 48.
1.0 [4] NBR 10004, “Resíduos sólidos - classificação”, Ass.
0.5 Bras. Norm. Técn., Rio Janeiro (2004).
0.0
Control WP-5 WP-10 WP-25 WP-50 WP-100 [5] DECA, “Reaproveitamento de resíduos e a economia
circular”, 21ª ed. Prêmio Mérito Amb., Fed. Ind. Est. S.
Figure 4: Flexural strength of test specimens. Paulo (2015).
[Figura 4: Resistência à flexão dos corpos de prova.] [6] I. Guerra, I. Vivar, B. Llamas, A. Juan, J. Moran, Waste
Manage. 29, 2 (2009) 643.
CONCLUSIONS [7] C. Medina, M. Frías, M.I. Sánchez de Rojas, Constr.
Build. Mater. 31 (2012) 112.
This study showed that some parameters that define [8] F. Pacheco-Torgal, S. Jalali, Constr. Build. Mater. 24
the quality of ceramic slips for use as sanitary ware were (2010) 832.
dependent on the degree of substitution of granite by waste [9] E. Martini, D. Fortuna, A. Fortuna, G. Rubino, V.
431 T. H. Silva et al. / Cerâmica 65 (2019) 426-431

Tagliaferri, Cerâmica 63, 368 (2017) 542. Ceram. 38 (1981) 194.

[10] NBR NM24, “Materiais pozolânicos - determinação [17] S.M. Antao, I. Hassan, J. Wang, P.L. Lee, B.H. Toby,
do teor de umidade”, Ass. Bras. Norm. Técn., Rio Janeiro Can. Mineral. 46 (2008) 1501.
(2003). [18] D. Balzar, H. Ledbetter, Am. Mineral. 78 (1993) 1192.
[11] NBR 15097-1, “Aparelhos sanitários de material [19] A.E. Souza, S.R. Teixeira, G.T.A. Santos, E. Longo,
cerâmico - parte 1: requisitos e métodos de ensaios”, Ass. Cerâmica 59, 349 (2013) 147.
Bras. Norm. Técn., Rio Janeiro (2011). [20] H. Zhao, D. Darwin, Cem. Conc. Res. 22 (1992) 695.
[12] I.R. Ganenko, Z.Y. Kovaleva, Glass Ceram. 46 (1989) [21] A.C. Morelli, J.B. Baldo, Cerâm. Ind. 8, 3 (2003) 42.
36. [22] A.K. Bougher, Cerâm. Ind. 151, 1 (2001) 25.
[13] Brookfield viscometer, “Operat. Instruct., Manual no. [23] G.S. Feng, S.L. Wu, H.L. Han, L.W. Ma, W.Z. Jiang,
M/85-150-P700”. X.Q. Liu. Int. J. Miner. Metal. Mat. 18 (2010) 270.
[14] Brookfield, “More solutions to sticky problems”, [24] A.W. Vieira, M.D.M. Innocentini, E. Mendes, T. Gomes,
https://www.brookfieldengineering.com. A. Demarch, O.R.K. Montedo, E. Angioletto, Mater. Res. 20
[15] D.Y. Tunçel, M.K. Kara, E. Özel, IOP Conf. Ser. Mater. (2017) 637.
Sci. Eng. 18 (2011) 222025. (Rec. 30/10/2017, Rev. 26/11/2018, 26/01/2019, 04/03/2019,
[16] N.S. Belostotskaya, V.V. Kargina, N.S. Lykhina, Glass Ac. 07/03/2019)