Our reference: JCLP 4025 P-authorquery-v9

AUTHOR QUERY FORM

Journal: JCLP Please e-mail or fax your responses and any corrections to:

E-mail: corrections.esch@elsevier.tnq.co.in

Article Number: 4025 Fax: +31 2048 52789

Dear Author,

Please check your proof carefully and mark all corrections at the appropriate place in the proof (e.g., by using on-screen
annotation in the PDF file) or compile them in a separate list. Note: if you opt to annotate the file with software other than
Adobe Reader then please also highlight the appropriate place in the PDF file. To ensure fast publication of your paper please
return your corrections within 48 hours.

For correction or revision of any artwork, please consult http://www.elsevier.com/artworkinstructions.

Any queries or remarks that have arisen during the processing of your manuscript are listed below and highlighted by flags in
the proof.

Location Query / Remark: Click on the Q link to find the querys location in text
in article Please insert your reply or correction at the corresponding line in the proof
Q1 Please confirm that given names and surnames have been identified correctly.

,
Please check this box or indicate
your approval if you have no
corrections to make to the PDF file

Thank you for your assistance.

 JCLP4025_grabs 11 February 2014 1/1

Journal of Cleaner Production xxx (2014) 1

5
Contents lists available at ScienceDirect 6
7
8
Journal of Cleaner Production 9
10
journal homepage: www.elsevier.com/locate/jclepro 11
12
13
14
15
1 16
2 17
3 Highlights
4
 Innovative procedure was applied to recycle glass in ceramic tiles.
 Ceramic frits were fully replaced by recycled glass in engobe formulation.
 Statistical techniques were used to nd an optimal composition.
 Properties of engobe could be described by mathematical models.

http://dx.doi.org/10.1016/j.jclepro.2014.01.088
0959-6526/ 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088
 JCLP4025_proof 11 February 2014 1/7

Journal of Cleaner Production xxx (2014) 1e7

55
Contents lists available at ScienceDirect 56
57
58
Journal of Cleaner Production 59
60
journal homepage: www.elsevier.com/locate/jclepro 61
62
63
64
65
1 Formulation of ceramic engobes with recycled glass using mixture 66
2 67
3 design 68
4 69
5 Q1 Marcelo Dal B a, c, *, Adriano Michael Bernardin b, Dachamir Hotza c 70
6 a 71
Federal Institute of Rio Grande do Sul (IFRS), Campus Feliz, 95770-000 Feliz, RS, Brazil
7 b 72
Graduate Program on Materials Science and Engineering (PPGCEM), Group of Ceramic and Glass Materials, University of Southernmost Santa Catarina
8 (UNESC), 88806-000 Cricima, SC, Brazil 73
9 c
Graduate Program on Materials Science and Engineering (PGMAT), Department of Chemical Engineering (EQA), Federal University of Santa Catarina 74
10 (UFSC), 88040-900 Florianpolis, SC, Brazil
75
11 76
12 77
13 a r t i c l e i n f o a b s t r a c t 78
14 79
15 Article history: In this work glass frits were fully replaced by recycled glass in engobe compositions. Engobe is an
Received 31 May 2013
80
16 interlayer between glaze and substrate in traditional ceramic tiles. Five ceramic raw materials were used
Received in revised form 81
17 to formulate engobes. Linear shrinkage, water absorption, whiteness, water stain and coefcient of
22 January 2014
thermal expansion were measured in designed mixtures and described by mathematical models. A
82
18 Accepted 22 January 2014 83
Available online xxx
statistical analysis enabled describing the engobe properties at a 95% condence level. A major inuence
19 84
of the raw materials interaction on the engobe properties was determined as a result of the chemical
20 reactions during ring. According to the response surface analysis performed, it was possible to nd an 85
21 Keywords:
Engobe
area in the triaxial diagram where all properties were optimized. 86
22 2014 Elsevier Ltd. All rights reserved. 87
Recycled glass
23 Ceramic tiles 88
24 Mixture design 89
25 Response surface
90
26 91
27 92
28 93
29 94
1. Introduction Bergmann, 2005; Raimondo et al., 2007; Bernardo et al., 2009),
30 95
cement (Ling and Poon, 2012) and glazes (da Silva et al., 2012), few
31 96
The engobe is an interlayer applied on a substrate under the references refer to recycled materials in engobes (Larichkin et al.,
32 97
glaze layer which is used in most ceramic tiles, Fig. 1. An engobe is 2008; Velho and Bernardin, 2011). However, to the best of our
33 98
used to opacify the substrate, to attenuate physical-chemical dif- knowledge, in no case the ceramic frit was replaced 100% by
34 99
ferences, to improve the tting between the substrate and the recycled glass.
35 100
glaze, and to reduce the defects on tile surface (Santos et al., 2007). Glass wastes can be fully recycled to manufacture new glass
36 101
Engobes are formulated from natural raw materials like clays, products (Tchobanoglous and Kreith, 2002). Nevertheless, some
37 102
quartz and feldspars and synthetic raw materials as ceramic frits. A technological constraints hinder the recycling of glasses in ceramic
38 103
frit is a glass manufactured in a large furnace at high temperatures tile production. Generally, ceramic frits used in engobes have a
39 104
(around 1500
C) which is submitted thereafter to a thermal shock coefcient of thermal expansion between 6 and 7$106
C1, while
40 105
to produce irregularly shaped vitreous particles. Hence, this recycled glass from containers and at glass have a coefcient of
41 106
ceramic material is cost intensive and presents some negative im- thermal expansion higher than 9$106
C1(Navarro, 2003). In this
42 107
pacts on the environment. regard, to replace frits by recycled glass in engobes a new formu-
43 108
In order to reduce the environmental impact of glass residues, lation is a challenge.
44 109
the reuse of this solid waste can contribute to reduce and/or avoid In ceramic materials and processes, statistical tools have been
45 110
its discard, decreasing the use of natural raw materials. Despite the widely used to optimize compositions. The design of experiments
46 111
use of recycled glass in ceramic substrates (Bragana and (DoE) can help the development of formulations and the inter-
47 112
pretation of the results obtained from testing (Cornell, 2002). Sta-
48 113
tistical techniques applied to DoE have been widely used in various
49 * Corresponding author. Federal Institute of Rio Grande do Sul (IFRS), Campus 114
elds such as pharmacology (Cafaggi et al., 2003), food (Kumar
50 Feliz, 95770-000 Feliz, RS, Brazil. Tel.: 55 51 81201427; fax: 55 48 37217615. 115
E-mail addresses: marcelodalbo@hotmail.com, marcelo.dalbo@feliz.ifrs.edu.br et al., 2010) and materials science (Menezes et al., 2008; Correia
51 116
(M. Dal B). et al., 2009).
52 117
53 118
http://dx.doi.org/10.1016/j.jclepro.2014.01.088
54 0959-6526/ 2014 Elsevier Ltd. All rights reserved. 119

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088
 JCLP4025_proof 11 February 2014 2/7

2 M. Dal B et al. / Journal of Cleaner Production xxx (2014) 1e7

1 Table 2 66
2 Restrictions in contents of raw materials. 67
3 Raw material Minimum value (wt%) Maximum value (wt%) 68
4 Clay 10.0 40.0
69
5 Feldspar 0.0 40.0 70
6 Recycled glass 10.0 40.0 71
7 Quartz 0.0 40.0 72
Zircon 0.0 30.0
8 73
9 74
10 75
Table 3
11 76
Chemical analysis (wt%) of the raw materials as measured by XRF.
12 77
13 Raw SiO2 Al2O3 Na2O K2O CaO MgO ZrO2 TiO2 Fe2O3 LoI 78
material
14 79
15 Quartz 98.0 0.2 e e 0.8 e e <0.1 <0.1 0.9 80
16 Zircon 32.5 0.9 e e e e 66.2 0.1 <0.1 0.3 81
Feldspar 66.2 18.7 2.9 11.4 0.2 <0.1 e <0.1 0.1 0.5
17 82
Recycled 70.4 1.3 14.2 0.3 8.9 3.9 e <0.1 0.4 0.6
18 glass 83
19 Fig. 1. Microscopic image showing the constitutive layers of ceramic tiles. Clay 42.4 35.1 0.1 1.1 e 0.2 e 1.2 1.1 17.8 84
20 85
21 86
22 Mixture design has been employed as a tool for the optimization In this work, those properties were investigated and a com- 87
23 of mechanical and esthetical properties of ceramic products, parison to a currently used engobe was done. Some properties were 88
24 enabling the description of the property under analysis according measured in a standard engobe composition and taken as target 89
25 to the composition. Furthermore, this technique helps to reduce the values, which include LS 3.9%, WA 15.0%, L* 90.02, DE* 2.5 90
26 number of experiments in laboratory. When the processing con- and CTE 8.0$106
C1. 91
27 ditions are kept constant, as sought in the ceramic industry, the According to a previous study (Dal B et al., 2011), LS and CTE are 92
28 nal product properties are determined by the combination of raw properties of ceramic tiles to be adjusted to decrease the stresses 93
29 materials. between the layers during the cooling step. The residual stresses 94
30 Triaxial experimental designs, which can be seen as a special after ring can promote some defects in ceramic tiles such as cur- 95
31 case of DoE with only three variables, have been extensively used to vature and deformation (Cantavella et al., 2000; Melchiades et al., 96
32 study the properties of ceramic products (Schabbach et al., 2001; 2000). Therefore, the characterization of those properties repre- 97
33 Correia et al., 2006). However, the description of the materials sents an important support to industrial processes. Water stains are 98
34 properties with only three factors (triaxial designs) can lead to associated to some engobe characteristics (Melchiades et al., 2002), 99
35 simplistic predictions considering the complexity of ceramic ma- and may be expressed in terms of tonality difference between the 100
36 terials, particularly engobes. dry and wet engobe (DE*) (Quinteiro et al., 2010). Moreover, the 101
37 A ceramic engobe must present suitable values in key properties chromatic coordinate L*, which expresses the whiteness of the 102
38 such as linear shrinkage (LS), water absorption (WA), tonality, engobe, must be also taken into account particularly when the 103
39 whiteness expressed as lightness, L*, according to the CIELAB color ceramic glaze is transparent (Melchiades et al., 2002). 104
40 space (Schanda, 2007), water stain (WS), expressed as color dif- Thus, the aim of this work is replacing frits by recycled glass in 105
41 ference, DE*, according to the CIELAB color space (Witt, 2007) and ceramic engobes. A design of experiments is performed and 106
42 the coefcient of thermal expansion (CTE). response surfaces are build to analyze relevant properties of 107
43 engobes as a function of the composition. 108
44 Table 1 109
45 Mixture design with ve factors. 110
46 Formulation Clay Feldspar Recycled glass Quartz Zircon 2. Experimental 111
47 112
1 40.0 40.0 20.0 0.0 0.0
48 2 10.0 40.0 10.0 40.0 0.0 2.1. Engobe formulation and sample preparation 113
49 3 40.0 40.0 10.0 10.0 0.0 114
50 4 40.0 40.0 10.0 0.0 10.0 For the mixture design ve factors corresponding to ve raw 115
5 40.0 20.0 40.0 0.0 0.0
51 materials were applied, resulting in 21 formulations and one 116
6 40.0 0.0 40.0 20.0 0.0
52 7 20.0 0.0 40.0 40.0 0.0 genuine replica (central point or centroid). The software Statistica 117
53 8 10.0 10.0 40.0 40.0 0.0 7.0 (Statsoft) was used to assist the design and the analysis of the 118
54 9 40.0 20.0 10.0 0.0 30.0 experiments. Some restrictions on raw materials content were 119
55 10 40.0 0.0 10.0 20.0 30.0 imposed according to industrial practice. 120
11 10.0 40.0 10.0 10.0 30.0
56 Table 1 presents the mixture design matrix, i.e., the composition 121
12 20.0 40.0 10.0 0.0 30.0
57 13 10.0 10.0 10.0 40.0 30.0 of each formulation, showing the content (wt%) of the raw mate- 122
58 14 20.0 0.0 10.0 40.0 30.0 rials selected. 123
59 15 10.0 40.0 20.0 30.0 0.0 Engobe raw materials were obtained from a local ceramic in- 124
60 16 10.0 20.0 40.0 30.0 0.0 125
dustry (Endeka, Mogi Guau, SP, Brazil). Recycled glass was ob-
17 20.0 30.0 10.0 40.0 0.0
61 18 40.0 30.0 10.0 20.0 0.0 tained from a recycling company (Masx, So Paulo, SP, Brazil). 126
62 19 20.0 40.0 20.0 0.0 20.0 Table 2 and Equations (1)e(3) are related to the restrictions 127
63 20 10.0 40.0 20.0 10.0 20.0 regarding the raw materials content (wt%). 128
64 21a 24.5 25.0 19.5 19.5 11.5 129
65
21b 24.5 25.0 19.5 19.5 11.5 %Feldspar %Recycled glass  60% (1) 130

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088
 JCLP4025_proof 11 February 2014 3/7

M. Dal B et al. / Journal of Cleaner Production xxx (2014) 1e7 3

1 Table 4 66
2 Evaluated characteristics of the engobes. 67
3 Formulation LS (%)  0.3 WA Whiteness, WS, CTE 68
4 (%)  0.5 L*  0.02 DE*  0.1 (106
C1)  0.1 69
5 1 3.0 15.9 92.90 8.4 8.9 70
6 2 1.3 16.3 93.37 7.4 10.1 71
7 3 1.2 19.4 94.87 6.9 8.2 72
4 1.5 19.3 94.76 5.6 7.0
8 73
5 9.8 0.3 78.90 3.5 9.3
9 6 10.9 0.0 76.53 3.3 11.2 74
10 7 9.2 0.8 77.57 3.7 13.7 75
11 8 9.1 0.4 75.40 4.4 13.8 76
12 9 1.3 19.4 95.70 3.4 6.9 77
10 0.9 18.7 95.69 2.4 9.5
13 11 1.5 14.0 94.33 2.0 7.0
78
14 12 0.9 16.9 94.98 2.2 6.4 79
15 13 0.4 16.2 94.97 3.6 10.5 80
16 14 0.5 17.6 93.98 3.1 10.8 81
15 7.9 0.2 82.58 5.4 9.7
17 82
16 7.3 0.4 76.12 2.6 10.4
18 17 0.5 18.1 94.63 7.4 10.9 83
19 18 1.0 20.3 92.97 8.2 9.1 84
20 19 3.2 8.7 92.49 1.7 7.6 85
21 20 8.1 0.2 91.63 0.9 7.7 86
21a 2.5 13.0 93.27 3.1 9.7
22 21b 3.0 13.0 93.49 3.6 9.7
87
23 88
24 89
25 %Clay %Quartz  60% (2) 90
26 91
27 92
%Recycled glass %Zircon  40% (3)
28 93
29 The engobe formulations (according to Table 1) were weighed 94
30 (300.00  0.01 g), and milled with 35 wt% water in a planetary mill 95
31 (55 vol% alumina balls) during 15 min to form suspensions with 96
32 densities between 1.85 and 1.87 kg L1 and sieved in a 45 mm mesh 97
33 (0.5e1.0 wt% residue). 98
34 To measure CTE, LS, WA and WS, the engobe suspensions were 99
35 cast in plaster molds, so that 1  1  3 cm3 specimens were ob- 100
36 tained. After forming, the samples were dried at 110
C during 4 h 101
37 and red at 1050
C in a laboratory furnace (INTI FQ-1300) with a 102
38 heating rate of 25
C min1 and 3 min of holding time at the 103
39 maximum temperature. The furnace was then turned off and the 104
40 samples were cooled down to room temperature. 105
41 To measure L* and DE*, the engobe suspensions were cast on an 106
42 unred ceramic substrate using a doctor blade, resulting a thick- 107
43 ness of 0.6 mm (ISO, 1997). The samples were submitted to the 108
44 same drying and ring conditions than those used in the mea- 109
45 surement of CTE, LS, DE*, WA and WS. 110
46 111
47 2.2. Characterization of raw materials and sintered samples 112
48 113
49 The chemical characterization of the raw materials was per- 114
50 formed by X-ray uorescence (XRF) using Bruker Explorer S4 115
51 spectrometer. The linear shrinkage was measured using a caliper. 116
52 The thermal expansion analysis was performed using a contact 117
53 dilatometer (BP Engenharia RB 3000) with the specimens previ- 118
54 ously sintered. 119
55 120
56 121
Table 5
57 122
Analysis of variance of the engobe properties.
58 123
59 Variable Model Analysis of variance Lack of t 124
60 p-value R2 R2 adj p-value 125
61 LS Linear 0.0000 0.853 0.819 0.165
126
62 WA Special cubic 0.0049 0.9999 0.9989 0.0628 127
63 L* Special cubic 0.0396 0.9995 0.9950 0.1356
Fig. 2. a. Response surface for linear shrinkage. b. Response surface for water ab- 128
64 DE* Quadratic 0.0202 0.9836 0.9510 0.5124
sorption. c. Response surface for whiteness. 129
CTE Full cubic 0.0202 0.9999 0.9999 e
65 130

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088
 JCLP4025_proof 11 February 2014 4/7

4 M. Dal B et al. / Journal of Cleaner Production xxx (2014) 1e7

1 Table 6 66
2 Most adequate models for the engobe properties. 67
3 Models Equation for the engobe propriety 68
4 Linear LS 2:74x2 24:08x3
69
5 Special WA 104:5x1  229:8x2  123:9x4 375:8x5 901:6x1 x2 909:6x1 x4  797:8x1 x5 883:9x2 x4  500:2x4 x5  3109x1 x2 x4  1240x1 x3 x4 2667x1 x3 x5 70
6 cubic 71
7 Special L* 169:6x5 599:6x1 x2 451:9x2 x4 493:4x3 x5  1425x1 x2 x4 72
cubic
8 73
Quadratic DE* 32:88x1  54:2x3 30:73x4  70:41x1 x4  86:13x1 x5 106:2x2 x3 56:21x3 x4  70:97x4 x5
9 Full cubic CTE 9:73x1  30:86x3 19:86x4 90:90x1 x3 44:34x2 x3 20:22x2 x5 97:06x3 x4 70:50x3 x5 19:32x4 x5  66:58x1 x2 x5 100:50x1 x3 x5 74
10 75
11 76
12 77
13 The chromatic analysis was carried out on the engobe layer after contents of recycled glass, probably due to the liquid phase for- 78
14 ring. The colorimetric coordinates L* and DE* were measured by a mation during ring. The formulations with high contents of clay, 79
15 spectrophotometer (Minolta CM 2600d) in the CIELAB color space. quartz and zircon showed low shrinkage when sintered at 1050
C. 80
16 The EN 10545/1997 Standard (ISO, 1997) was taken as a guideline The inset in Fig. 2a shows the Pareto chart for LS, describing the 81
17 for assessing the water absorption. effect of each raw material on LS. The result showed that the 82
18 The measured variables (LS, WA, L*, DE* and CTE) were used to recycled glass content presented the higher effect on the linear 83
19 calculate interactively the equation coefcients with statistical shrinkage. 84
20 signicance (mathematical models) and subsequently the response LS was lower for formulations with higher contents of feldspar 85
21 surface. Thereafter, the engobe properties could be described as a than that for formulations with a higher glass content. These results 86
22 function of the raw materials content (wt%). probably are due to the higher content of liquid phase formed in the 87
23 engobes with recycled glass, for which the softening temperature is 88
24 3. Results and discussion around 660
C (Wang et al., 2010). The potassium feldspar starts the 89
25 liquid phase formation at >1000
C (Das and Dana, 2003). Previous 90
26 The chemical analysis of the raw materials is shown in Table 3. studies showed that soda-lime glass acted as a good uxing agent, 91
27 The soda-lime recycled glass is a mixture of lamps, jars, containers i.e. lowered the ring temperature (Tucci et al., 2004). 92
28 and at glass. Thus, it presents a high content of sodium and cal- Fig. 2b shows the WA as a function of the contents of quartz, clay 93
29 cium oxides. Feldspar can be considered orthoclase, according to and feldspar. The mass concentrations of glass and zircon were kept 94
30 the potassium oxide content showed by the chemical analysis. The constant at 20% and 15%, respectively. WA was zero for the com- 95
31 other raw materials show a high degree of purity considering their positions with a high content of feldspar due to the high content of 96
32 main mineral constituents (Table 4). ux materials (feldspar-glass-zircon) as can be observed in 97
33 The statistical models (equations) obtained from the response formulation 20, Table 1. On the other hand, high values of WA, 98
34 surfaces are shown in Table 5 as a function of the raw material around 20%, were obtained with mixtures of quartz and clay, as can 99
35 contents (variables). The condence level used for the statistical be observed in formulations 3 and 18. Higher water absorption in 100
36 analysis was 95% and the adjusted models present a p-value less engobes can contribute to the appearance of water stains (higher 101
37 than 5% (Myers and Montgomery, 2002), meaning that they are DE*) in ceramic tiles, requiring the use of crystalline phases that 102
38 suitable to describe the variables under study. The R2 and adjusted increase the light scattering, like zirconium compounds 103
39 R2 ttings are also given for each property, showing that the (Melchiades et al., 2002). The inset in Fig. 2b shows the t between 104
40 mathematical models agree with the experimental values for each the observed and the predicted values when using the special cubic 105
41 property under analysis. model showed in Table 6. This tting shows the good agreement 106
42 Table 5 shows that p-values of the lack-of-t were higher than between the experimental results and the mathematical model. 107
43 5% (signicance level), allowing to describe these variables with The response surface for the L*chromatic coordinate is shown in 108
44 reliability. The p-value for the CTE was not determined because the Fig. 2c, L* is an indicative of the whiteness degree of a material, 109
45 number of parameters to be estimated is larger than the number of varying from 0 (black) to 100 (white). The results show that a high 110
46 observations, i.e. the model is said to lack degrees of freedom content of recycled glass in the engobe formulation results in lower 111
47 (Table 5). However, the level of signicance (a < 5%) and the tting values for L* and, therefore, dark tonalities. On the other hand, a 112
48 (R2 and adjusted R2) suggest that this model is suitable to describe higher content of clay and zircon increases L* (around 100), indi- 113
49 CTE. cating lighter tonalities. In Fig. 2c, the contents of feldspar and 114
50 According to the variance analysis and model validation, a set of quartz were kept both constant at 20%. The inset in Fig. 2c shows 115
51 equations can be sought to describe the engobe properties. the t between the observed and the predicted values using the 116
52 Table 6 shows the equations for the engobe properties as a special cubic model. This tting shows the good agreement be- 117
53 function of the engobe raw materials, where x1, x2, x3, x4, and x5 are tween the experimental results and the mathematical model. 118
54 the clay, feldspar, recycled glass, quartz and zircon weight fractions, The response surface for water stain is shown in Fig. 3a. Lower 119
55 respectively. values of DE* for water stain were obtained for higher contents of 120
56 Table 6 shows that for LS, only the contents of feldspar (x2) and glass and zircon. The recycled glass increases the sinterability of the 121
57 recycled glass (x3) were signicant to describe the behavior of this engobes and decreases the porosity after ring. On the other hand, 122
58 variable. In contrast, for the remaining properties the interaction zircon improves the degree of opacication in ceramic materials 123
59 between the raw materials presented high signicance. This result due to the high refractive index of this silicate. Higher contents of 124
60 conrms that the chemical reactions during ring are very impor- clay and quartz result in higher DE* values for the engobe because 125
61 tant to describe the properties of ceramic products. clay and quartz particles present no crystals with high refractive 126
62 Fig. 2a presents the linear behavior of the engobe shrinkage index. Similar results were found when investigating water stain in 127
63 when the contents of clay, feldspar and recycled glass were varied, ceramic engobes (Melchiades et al., 2002). The inset in Fig. 3a 128
64 but the contents of quartz and zircon were held constant at 20% and shows the t between the observed and the predicted values using 129
65 15%, respectively. A higher shrinkage was observed for high the quadratic model shown in Table 6. 130

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088
 JCLP4025_proof 11 February 2014 5/7

M. Dal B et al. / Journal of Cleaner Production xxx (2014) 1e7 5

1 Thermal expansion coefcient is a relevant variable to evaluate 66

2 the internal tensions between layers, which compose a ceramic tile 67
3 (Paganelli and Venturelli, 2005). Those internal tensions can be 68
4 related to the curvature and/or deformation after ring (Cantavella 69
5 et al., 2000). Fig. 3b shows the effect of raw materials variation on 70
6 the thermal expansion coefcient of the engobe. The contents of 71
7 feldspar and zircon were kept constant at 20% and 15%, respectively. 72
8 Low values for CTE (w6$106
C1) were obtained with high con- 73
9 tents of glass in the engobe composition. On the other hand, high 74
10 contents of quartz increase the thermal expansion to values up to 75
11 w12$106
C1. The inset in Fig. 3b shows the t between the 76
12 observed and the predicted values using the cubic full model 77
13 shown in Table 6. 78
14 The analysis of the best results obtained by using the mixture 79
15 design models relating the engobe properties with the raw material 80
16 contents resulted in an optimized area in the triaxial diagram. Fig. 4 81
17 shows the area in the triaxial diagram formed by glass, clay and 82
18 feldspar where all engobe properties were optimized, i.e., low water 83
19 absorption, water stain, and linear shrinkage; high whiteness; and 84
20 thermal expansion coefcient suitable to the ceramic engobe 85
21 86
22 Fig. 4. Optimized area in the triaxial diagram of clay-feldspar-recycled glass. 87
23 88
24 89
currently used. In this case, the contents of quartz and zircon were
25 90
kept constant at 0% and 20%, respectively. Table 7 shows some
26 91
properties of two different engobes. Formulation 19 presented
27 92
similar properties when compared to the standard engobe.
28 93
According to the results obtained, it was possible to analyze the
29 94
relationship between the engobe properties. Fig. 5a shows the
30 95
relationship between water stain and zircon content in engobe
31 96
formulations. The results showed that there is an optimal zircon
32 97
content, around 20%, to minimize the water stain at engobes. Fig. 5b
33 98
shows the relationship between WS and WA in engobes. The results
34 99
showed that high WA increases WS. Interestingly, some high values
35 100
for WS were also found in engobes with WA near zero, showing
36 101
that only ux materials do not guarantee a low water stain in
37 102
engobes. The use of raw materials with a high refraction index
38 103
becomes therefore important.
39 104
Fig. 5c shows the relationship between L* and LS. Engobes with
40 105
high linear shrinkage present dark tonalities (low L*) and engobes
41 106
with white tonalities were found with low LS. Fig. 5d correlates WA
42 107
to LS. High LS decreases the WA, what is usual for ceramic mate-
43 108
rials. Finally, Fig. 5d shows WA near zero for engobes with LS
44 109
around 9%.
45 110
46 111
47 4. Conclusions 112
48 113
49 This work showed that the substitution of the ceramic frit by 114
50 recycled glass in ceramic engobes is a viable alternative, enabling 115
51 an environmental solution to reuse of waste glass. 116
52 The mixture design and response surface methodologies were 117
53 used in order to model the engobe properties according to the 118
54 content of raw materials, showing predicted values for the exper- 119
55 imental results with a condence level of 95%; when the processing 120
56 121
Table 7
57 122
Comparative properties of standard and optimal designed formulations of engobe.
58 123
59 Variable Engobe 124
60 Standard Formulation 19 125
61 LS (%) 3.9  0.3 3.2  0.3
126
62 WA (%) 15.0  0.5 8.7  0.5 127
63 L* 90.02  0.02 92.49  0.02 128
64 DE* 2.5  0.1 1.7  0.1 129
Fig. 3. a. Response surface for water stain. b. Response surface for thermal expansion
CTE (106
C1) 8.0  0.1 7.6  0.1
65 coefcient. 130

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088
 JCLP4025_proof 11 February 2014 6/7

6 M. Dal B et al. / Journal of Cleaner Production xxx (2014) 1e7

1 66
2 67
3 68
4 69
5 70
6 71
7 72
8 73
9 74
10 75
11 76
12 77
13 78
14 79
15 80
16 81
17 82
18 83
19 84
20 85
21 86
22 87
23 88
24 89
25 90
26 91
27 Fig. 5. Correlation between some engobe properties: (a) DE* vs. zircon content, (b) DE* vs. WA, (c) L* vs. LS and (d) WA vs. LS. 92
28 93
29 conditions were kept unchanged. The interactions between the raw Cafaggi, S., Leardi, R., Parodi, B., Caviglioli, G., Bignardi, G., 2003. An example of 94
application of a mixture design with constraints to a pharmaceutical formula-
30 materials showed a high signicance level (p-value < 5%). 95
tion. Chemom. Intelligent Lab. Syst. 65 (1), 139e147.
31 The results showed that the linear shrinkage was increased with Cantavella, V., Snchez, E., Garca-Ten, J., Feliu, C., Snchez, J., Soler, C., Portols, J., 96
32 a high content of recycled glass in the engobe formulation. The Sales, J., 2000. Residual stresses in porcelain tiles formed by two-charge 97
water absorption was related to the sintering degree of the raw pressing. In: VI World Congress on Ceramic Tile Quality, Qualicer 2000.
33 98
Cmara Ocial de Comercio, Industria y Navegacin, QUALICER, Castelln De La
34 materials. Therefore, a high content of feldspar and glass in the Plana e Spain, pp. 335e349. 99
35 engobe formulation decreased water absorption. Cornell, J.A., 2002. Experiments with Mixtures: Designs, Models and the Analysis of 100
36 The whiteness of the ceramic engobe was increased with higher Mixture Data, third ed. Wiley, New York. 101
Correia, S.L., Dienstmann, G., Folgueras, M.V., Segadaes, A.M., 2009. Effect of quartz
37 contents of refractory raw materials like clay and zircon. The water sand replacement by agate rejects in triaxial porcelain. J. Hazard. Mater. 163 (1), 102
38 stain was decreased with high contents of glass and zircon because 315e322. 103
39 the use of these raw materials can provide a high sintering degree Correia, S.L., Oliveira, A.P.N., Hotza, D., Segades, A.M., 2006. Properties of triaxial 104
porcelain bodies: interpretation of statistical modeling. J. Am. Ceram. Soc. 89
40 and a high refraction index, respectively. (11), 3356e3365. 105
41 The study of the thermal expansion coefcient in engobes da Silva, R.C., Pianaro, S.A., Tebcherani, S.M., 2012. Preparation and characterization 106
42 showed that high contents of quartz increased the thermal of glazes from combinations of different industrial wastes. Ceram. Int. 38 (4), 107
2725e2731.
43 expansion. On the other hand, high contents of glass and clay Dal B, M., Hotza, D., Melchiades, F.G., Boschi, A.O., 2011. Interaction between
108
44 decreased the thermal expansion coefcient. support-glaze and support-engobe in ceramic tiles. In: ECerS XII e 12th Con- 109
45 The superposition of the response surfaces enabled the determi- ference of the European Ceramic Society. ECerS, Stockholm e Sweden. 110
Das, S.K., Dana, K., 2003. Differences in densication behaviour of K- and Na-
46 nation of an area in the triaxial diagram where all properties presented 111
feldspar-containing porcelain bodies. Thermochim. Acta 406 (1e2), 199e206.
47 optimal values. The relationship between the engobe characteristics ISO, BS EN ISO10545e3, 1997. Method for Determination of Water Absorption. 112
48 showed that an optimum zircon content decreased the water stain. On British Standards Institution - BSI. 113
the other hand, not only engobes with low water absorption guar- Kumar, S.B., Ravi, R., Saraswathi, G., 2010. Optimization of fruit punch using mixture
49 114
design. J. Food Sci. 75 (1), S1eS7.
50 antee a low occurrence of water stain, but raw materials with high Larichkin, V.V., Khlebnikov, B.M., Kuzmina, N.G., 2008. Prospects of generation of 115
51 refraction index play also an important role in this case. high technologies of recycling of ash-and-slag waste of thermal power stations, 116
52 strategic Technologies, 2008. In: IFOST 2008. Third International Forum on 117
Strategic Technologies, Novosibirsk-tomsk, Russia, pp. 648e650.
53 Acknowledgements Ling, T.-C., Poon, C.-S., 2012. A comparative study on the feasible use of recycled 118
54 beverage and CRT funnel glass as ne aggregate in cement mortar. J. Clean. 119
55 The authors appreciate the nancial support received from the Prod. 29e30, 46e52. 120
Melchiades, F.G., Barbosa, A.R.D., Boschi, A.O., 2000. Relationship between tile
56 Brazilian Research Agency CAPES (Coordenao de Aperfeioa- curvature and engobe layer characteristics. In: VI World Congress on Ceramic 121
57 mento de Pessoal de Nvel Superior) in the form of a Ph.D. grant (M. Tile Quality, Qualicer 2000. Cmara Ocial de Comercio, Industria y Navegacin, 122
58 Dal B). The authors also thank Endeka Ceramics for providing raw QUALICER, Castelln De La Plana, Spain. 123
Melchiades, F.G., Silva, L.L., Silva, V.A., Romachelli, J.C., Vargas, D.D.T., Boschi, A.O.,
59 materials. 2002. About engobes and water stain. Cermica Ind. 7 (4), 31e39 (in
124
60 Portuguese). 125
61 References Menezes, R.R., Neto, H.G.M., Santana, L.N.L., Lira, H.L., Ferreira, H.S., Neves, G.A., 126
2008. Optimization of wastes content in ceramic tiles using statistical design of
62 mixture experiments. J. Eur. Ceram. Soc. 28 (16), 3027e3039.
127
63 Bernardo, E., Esposito, L., Rambaldi, E., Tucci, A., 2009. Glass based stoneware as 128
Myers, R.H., Montgomery, D.C., 2002. Response Surface Methodology: Process and
a promising route for the recycling of waste glasses. Adv. Appl. Ceram. 108
64 (1), 2e8.
Product Optimization Using Designed Experiments. Wiley, New York. 129
Navarro, J.M.F., 2003. The Glass, third ed. National Research Council (in Spanish).
65 Bragana, S.R., Bergmann, C.P., 2005. Waste glass in porcelain. Mater. Res. 8, 39e44. 130

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088
 JCLP4025_proof 11 February 2014 7/7

M. Dal B et al. / Journal of Cleaner Production xxx (2014) 1e7 7

1 Paganelli, M., Venturelli, C., 2005. Integrated dilatometer-tensiometer to measure Schanda, J., 2007. CIE Colorimetry. In: Colorimetry. John Wiley & Sons, Inc., 13
the state of tension of glazed bodies. LIndustrie Cramique Verrire 1000, 76e pp. 25e78
2 14
80 (mai-juin). Tchobanoglous, G., Kreith, F., 2002. Handbook of Solid Waste Management.
3 Quinteiro, E., Menegazzo, A.P.M., Paschoal, J.O.A., Gibertoni, C., Neto, O.T., 2010. Mcgraw-hill. 15
4 Engobes staining in the ceramic tiles e Part 1: the water stain. Cermica Ind. 15 Tucci, A., Esposito, L., Rastelli, E., Palmonari, C., Rambaldi, E., 2004. Use of soda-lime 16
5 (3), 19e23 (in Portuguese). scrap-glass as a uxing agent in a porcelain stoneware tile mix. J. Eur. Ceram. 17
Raimondo, M., Zanelli, C., Matteucci, F., Guarini, G., Dondi, M., Labrincha, J.A., 2007. Soc. 24 (1), 83e92.
6 Effect of waste glass (TV/PC cathodic tube and screen) on technological prop- Velho, P.L.T., Bernardin, A.M., 2011. Reuse of sludge treatment plant efuent for 18
7 erties and sintering behaviour of porcelain stoneware tiles. Ceram. Int. 33 (4), industrial production of engobes. Cermica Ind. 16 (2), 20e23 (in Portuguese). 19
8 615e623. Wang, M., Cheng, J., Li, M., 2010. Effect of rare earths on viscosity and thermal 20
Santos, G.R.d., Melchiades, F.G., Boschi, A.O., 2007. Development a methodology for expansion of soda-lime-silicate glass. J. Rare Earths 28 (Supplement 1), 308e
9 monitoring the evolution of ripening engobes during burn. Cermica Ind. 12 (5), 311.
21
10 22e27 (in Portuguese). Witt, K., 2007. CIE Color Difference Metrics. In: Colorimetry. John Wiley & Sons, Inc, 22
11 Schabbach, L.M., Fredel, M.C., Hotza, D., 2001. Three-component Lead Borosilicate pp. 79e100. 23
frit. Am. Ceram. Soc. Bull. 80, 7.
12 24

Please cite this article in press as: Dal B, M., et al., Formulation of ceramic engobes with recycled glass using mixture design, Journal of Cleaner
Production (2014), http://dx.doi.org/10.1016/j.jclepro.2014.01.088