How can the European ceramic tile industry meet the EU's low-carbon targets? A life cycle perspective

Ceramic tile manufacturing is deemed to be an energy intensive industry, mainly based on combustion processes and, therefore, subject to European policies aiming at reducing greenhouse gas emissions. The “Roadmap for moving to a competitive low-carbon economy in 2050”, approved by the European Commission, calls for sectoral strategies to reduce CO2 emissions by 20% by 2020 and by 83–87% by 2050, compared to 1990 CO2 emissions. This study included up to 17 technological alternatives and their combination, resulting in 25 technological scenarios associated to the life cycle of porcelain stoneware tiles. In this regard, a high parametrized LCA model was developed to allow for the required flexibility. The scenario analysis can be used: a) to estimate the degree of technological innovation required; b) to define and to focus strategies and; c) to devise the lines of technological development that need to be implemented in the ceramic tile manufacturing sector in the coming years. The alternatives consisted of endogenous and exogenous sectoral technologies. The technologic alternatives involved changes in product design (thickness and decoration), changes in the manufacturing process (preparation of raw material by dry or wet route, and simultaneous implementation of thermal energy efficiency techniques), and changes in the energy sources (hybrid and/or electric driers, and kilns and decarbonization of the power grid mix). It was clearly proven that the wider the scope of the Life Cycle Assessment study is, the greater eco-innovations are necessary. In all the studied scenarios, the manufacturing stage was always the most significant from the global warming point of view. Finally, regarding the achievability of EU objectives, the results of this study show that the implementation of widespread technologies suffice for fulfilling 2020 targets; nevertheless, only some limited combinations of both widespread and ambitious breakthrough technologies may achieve the 2050 reduction targets.

legal framework based on European Directives, etc. (Gabaldon-Estevan et al., 2016). Nevertheless, there are differences to be considered in other countries, such as the degree of cogeneration systems implementation (Confindustria Ceramica, 2018;Pardo, 2018) or the raw material origin. The study is focused on PST, since it is the type of ceramic tile with the greatest commercial and innovation interest due to its high technical, functional and aesthetic versatility (ASCER, 2011;Sánchez et al., 2010).

Research method 2.1 Goal definition
On the account of the introduction above, the aim of this study was to determine whether and how it is possible to achieve the emission reduction objectives set by the European Commission through a roadmap for moving to (and surviving in) a competitive low carbon economy in 2020 and 2050, using the LCA methodology on different technological scenarios of the PST life cycle.
When performing LCA, several standards (EN 15804:2012+A1:2013ISO 14040:2006;ISO 14044:2006) and the ILCD handbook (Wolf et al., 2010) recommendations were followed. The EeBGuide was consulted as well (Lasvaux et al., 2014). In addition, scenario analyses were used: a) to estimate the degree of technological innovation required; b) to define and to outline the strategies and; c) to devise the lines of technological development which need to be implemented in the ceramic tile manufacturing sector in the coming years.

System definition and functional unit
The analyses considered the entire life cycle, i.e. from cradle to grave, although other scopes were used in the discussion of results. Life cycle modules were those used in CEN/TC 350 standards. The system boundary included the raw materials supplied for the body and glaze manufacturing, the raw materials transport means and distances, and each stage of the ceramic tile manufacturing process. Once the tiles are packaged, they are worldwide distributed; then, the tiles are duly unpacked for installation with fast-setting mortars. In this paper, a residential scenario was considered with a lifespan of 50 years. Afterwards, 70% of the removed tiles are deemed to be recovered as a filler, and 30% landfilled.
The Functional Unit (FU) was defined as "covering 1 m 2 of household floor surface for 50 years with an average PST". The characteristics of an average fired PST were defined as: water absorption <0.5%; 23.2 kg/m 2 weight; and 10.4 mm thickness with 0.76 kg/m 2 of glazes (Ros-Dosdá et al., 2017).

Baseline scenario and latest scenario
The study was built on a compilation of environmental information from 26 Spanish companies to obtain 14 Environmental Product Declarations (EPD) of PST. Therefore, inventory data, which corresponded to the period 2010-2015, were verified by independent third parties. The study includes companies from the whole value chain: elaborated raw material producers (both spray-dried granulated and glazes) and ceramic tile manufacturers. Moreover, some generic data, such as the type of means of transport or the type of waste management processes, were taken from a Spanish sectoral LCA study, carried out in 2007(Ros-Dosdá et al., 2017. In order to define the reference scenario, i.e. 1990, inputs and outputs of energy (thermal and electrical) for that year were taken from sectoral historical data (Celades et al., 2012), as they constitute more than 80% of the GHG emitted throughout the manufacturing process of the PST (Almeida et al., 2016;Benveniste et al., 2011;Bovea et al., 2010;Ibáñez-Forés et al., 2011;Ros-Dosdá et al., 2017). Table 1

OUTPUTS
Electrical energy sold to the grid (MJ/m 2 ) 0 14.6 a Including natural gas to feed the combined heat and power cogeneration system (CS) installed in the spray drier.

Limitations
No reference to the technical or economic viability of the proposals is discussed in this paper. In fact, the authors recognize that, in some cases, the development and implementation of the breakthrough technology will still be required (e.g. 100% of renewable energy used in kilns and driers), and they are aware that some studies (Wesselling et al., 2017) claimed that the integration of breakthrough innovations in traditional industries (as the studied ones) is typically slower than in high-tech industries for different reasons, such as the need for long-term investments, low risk managerial decisions or lack of market incentives. Nevertheless, the results of this study provide an idea of the level of theoretical demand set by the European Union and the technological trends to fulfil the required targets.

Selection of environmental impact categories
The life cycle impact assessment was carried out applying the CML 2001 impact assessment method (Guinée et al., 2002), as suggested in EN 15804:2012+A1:2013, updated to 2015. Although special attention was given to the Global Warming Potential (GWP), other environmental categories were also analyzed under the CML 2001 method (Guinée et al., 2002), as shown in Table 2, to study potential environmental burden shifting.

LCA model
The LCA model was developed in GaBi software (PE International, 2008b;Thinkstep, 2016b) and the bundled professional databases PE International 2008a, Thinkstep 2016a and ELCD 3.2 (JRC- IES, 2015). in order to obtain the background data. A total of 194 variables were parameterized to allow for the scenario analyses. Some parameters served to define the technological route when multiple alternatives were possible (e.g. dry milling vs. wet milling) and some others were used to input process values (e.g. thermal energy needed in the firing stage). The latter parameters could be grouped into different categories: consumption of raw and auxiliary materials; water and energy consumption; emissions to air; waste generated; distances and types of transport between the different life cycle stages; etc. Table 3 presents a summary of the different parameter categories applied in this study.

Type of parameters Life cycle processes Features
Types and quantities of raw materials This high parameterization of the model made programming more difficult at the beginning, but it gave much more flexibility to build scenarios in the long run. The modelling applied a modular approach, i.e. each process was modelled separately to facilitate the definition of process routes and technological and managerial options, providing high flexibility in the scenario simulation process.

Technological alternatives
A literature review of the technological alternatives and innovation trends allowed us to identify a set of the technologies which are the most likely to be applied in the horizon 2020 and 2050. They may be classified as: widespread (ready to be implemented) or breakthrough (further research is needed) and exogenous (outside the scope of the ceramic tile industry) or endogenous (sector specific) technologies. In Table 4, the characteristics of each technological alternative are summarized, as well as the reference source used to identify these technological alternatives and their representative values for energy and material.  (Table 4 is very wide, better to edit horizontally, using the width of two columns) alt-text:

Technological scenarios
From the combination of the different technological alternatives identified in the previous chapter, a total of 25 technological scenarios were formulated. Inventory data for each scenario was adapted from the average PST inventory (Ros-Dosdá et al., 2017) to the technological alternatives applied in each case.
In Table 5, the scenarios are listed from A to Z, showing the different technological alternatives chosen from the ones listed in the columns, grouped in three major classes: product design, manufacturing processes, and energy sources. In the scenario simulation, it should be considered that reference scenario A compiles technological alternatives which are the most likely to be applied in 1990, taken as a baseline by the Kyoto protocol and EU Directives, and scenario C represents the most likely technological situation in the ceramic tile industry in 2015.

Results and discussion
The emissions of CO 2 equivalent (CO 2 eq.) associated to all technological alternatives and constructed technological scenarios in the lifecycle of PST were quantified and analyzed with the support of GaBi Analyst (Thinkstep, 2016b). It should be noted that the same distribution destinations, maintenance operations and end-of-life management were considered in all technological alternatives and scenarios. Furthermore, other environmental impact categories were evaluated to identify possible burden shifting. In the figures, both the potential CO 2 eq. emitted in 1990 (A, baseline scenario) and the reduction considered by the EU objective for 2020 and 2050 are especially highlighted.
The sensitivity analysis performed assessed the influence of inventory data related to technology alternatives on the Global Warming Potential impact category. The results showed that the parameters related to the use of energy (directly or indirectly) were the most critical ones; these parameters included the consumption of thermal and electrical energy, cogeneration systems and the thickness of the tiles, but also the quantity of glazes and frits. The results of the sensitivity analysis are presented as supplementary information.

Product design alternatives
In the product design, two major factors were considered: quantity of glaze (GL100/GL50/GL0) and thickness of the tile (TH100/TH50). Regarding decoration materials, the extraction and transport of raw materials and the manufacturing of 1 kg of solid glaze with 33% of frit content applied on PST involved the emission of 0.8 kg CO 2 eq. In this regard, it should be pointed out that the frit is the glaze component with a higher carbon footprint, since its manufacturing process includes the fusion of the raw materials at around 1500 °C (Gómez-Tena et al., 2009). Consequently, the reduction in the amount of glaze or frit content would entail an almost proportional reduction of CO 2 eq. emissions.
On the other hand, lightening the tile by reducing the thickness of the ceramic body is possible as long as the technical and functional performance of the final product are not compromised (da Silva et al., 2014a;Girao et al., 2009). Fig. 2 (left) shows a potential reduction of 36% of CO 2 eq. emissions corresponding to a 50% thickness reduction (C-TH50) when compared to the latest scenario (C-TH100). This high dependence was due to the influence of thickness along the life cycle through the reduction of raw materials extraction, transportation, energy demand, etc. (da Silva et al., 2014a;Pini et al., 2014;Ros-Dosdá et al., 2017). No burden shifting in other impact categories were identified, as shown in Fig. 2 (right) where both technological alternatives are compared with relative values in 7 environmental impact categories. Indeed, the only impact category that exhibited a non-proportional reduction was the ADP element category, which is much more influenced by glaze components than by body thickness.

Manufacturing process alternatives
Two main process alternatives were analyzed: body raw materials preparation process (WCS/DRY) and energy efficiency technologies implemented in driers and kilns (CTT/WDS).
In the European ceramic tile industry, body raw materials are commonly prepared following a wet route (EIPPCB, 2007), because it facilitates the production of higher quality ceramic tiles in larger sizes. However, some studies claim that the new dry route developments may allow similar results to be obtained in a more sustainable way (Bonucchi, 2012;Mezquita et al., 2017;Shu et al., 2012aShu et al., , 2012b. Nevertheless, these studies have not taken a life cycle approach. Consequently, a specific analysis has been included in this work. To perform the comparison in this section, three alternatives were considered: wet route (WET); wet route using combined heat and power cogeneration systems (WCS); and dry route (DRY).
The WCS is very popular in the Spanish ceramic tile industry. A combined heat and power cogeneration system (CS) installed in the spray-drier allows for the simultaneous production of electric and thermal energy with high efficiency, but it entails more natural gas consumption than the thermal process itself (Caglayan and Caliskan, 2018;Mezquita et al., 2017;Monfort at al., 2010). The surplus of electricity cogenerated was usually sent to the power grid. To analyze this co-product from a life-cycle perspective, an expansion of the system was applied instead of an allocation method, due to the lack of data and high level of uncertainties to represent the physical causalities of this process (Azapagic and Clift, 1999) and the difficulty to apply economic allocation since only the electricity which is sold to the grid has an economic value. Then, it was considered that only flexible technologies of the SGM would be displaced by the system (Weidema, 2000). Fig. 3 (left) shows a comparison of CO 2 eq. emissions associated to the preparation of body raw materials following different routes and considering the implementation of CS. Fig. 3 (right) shows the relative contribution to other environmental impact categories and provides the evidence that potential transferring of environmental loads occurs when these processes were assessed from a life cycle perspective. Thus, the achievement of a slightly reduction of 2% of CO 2 eq. with the total implementation of DRY milling instead of the WCS would imply increasing the emissions of acidification substances and photochemical oxidants by 9% and 12%, respectively. alt-text: Fig. 2 Regarding the energy efficiency technologies in thermal processes (driers and kilns), Fig. 4 reveals that the current scenario (scenario C) with an average thermal efficiency of 15%, almost fulfils the EU's objectives for 2020. Mezquita et al., 2014aMezquita et al., , 2014b showed that the simultaneous implementation of the available widespread technologies allows for a maximum thermal efficiency of around 45% to be achieved (scenario D). Fig. 4 indicates that this does not suffice to attain the EU's objectives for 2050; therefore, these outcomes suggest that the combination of widespread and breakthrough technologies will be needed.

Energy source alternatives
The effect of the evolution of the Spanish power grid mix (SGM) until 2050 was based on Capros et al., 2013 who forecasted, in a study for the European Commission, the power and transport evolution with considerations regarding to market, economics, industry structure, demography and energy/environmental policies and regulations. The relation between the evolution of the SGM and its GHG emissions was obtained by programming the different mixes using the GaBi software (Thinkstep, 2016b).  (2015), and that the SGM evolves according to the forecast made by Capros et al. (2013). Moreover, an additional alternative is presented in this figure: scenario C with a 100% renewable scenario in the SGM for 2050 (C-REN50). The figure also shows the content of energetic renewable sources and the GHG reduction targets set by the EU. REN50 was built on the percentage of renewable sources foreseen for 2050 and calculating an extrapolation to cover 100%, thereby maintaining the proportions of the different technologies. REN50 was composed then by 54% wind power, 24% solar power, 15% hydropower, 6.5% biomass and 0.5% geothermal and other renewable energies. Fig. 3 Effect of the preparation of raw material on the associated CO 2 eq. emissions of PST life cycle (left) and relative contribution in each impact category (right).
alt-text: Fig. 3   Fig. 4 Effect of the energy efficiency of thermal processes on the associated CO 2 eq. emissions of PST life cycle.
alt-text: Fig. 4 In Fig. 5, a very slight decrease in CO 2 eq. emissions can be observed; consequently, it may be concluded that the accomplishment of the EU target in 2050 relying solely on the evolution of the SGM (exogenous factor) was not realistically affordable, supporting Gabaldón-Estevan et al., 2016. Fig. 6 shows the CO 2 eq. emissions associated with the life cycle of the scenario C when electric driers and kilns were used in the manufacturing process. An increase in emissions was detected due to the nature of the SGM. Fig. 6 evidences that SGM 2015 has a bigger carbon footprint than natural gas, i.e. 0.09 and 0.07 kg CO 2 eq./MJ, respectively and, therefore, from the global warming impact point of view, it does not make sense to devote efforts to develop electrification technologies, if electricity from renewable energy sources is not assured. It means that ceramic industry "per se" cannot meet the EU CO 2 emissions objectives, if exogenous technologies are not implemented in reducing the SGM carbon footprint. The implementation of renewable sources at sector or plant scale does not seem to be sufficient to supply the required energy, hence this option has not been considered in this study.

Potential reduction by technological scenarios
This section presents the results of the LCA of 25 technological scenarios. These 25 technological scenarios were obtained by the combination of those previous technological alternatives (see Table 5) which delivered substantial improvements in reducing CO 2 eq. emissions. Fig. 7 depicts the results of CO 2 eq. emissions of each PST technological scenario, differentiating each module of the life cycle of PST, from the raw material supply (stage A1) to the end-of-life (stage C4). In Fig. 8 a gate-to-gate scope (life cycle stage A3) is shown. In each scope, the CO 2 eq. emissions of the reference scenario and the correspondent EU objective reductions for 2020 and 2050 are marked.   Fig. 7 shows that, according to the proposed simulation, the CO 2 eq. emissions associated with A1 and A2 stages (extraction and transport of raw materials, respectively) are mostly affected by the body thickness (tile weight), being the latter much less significant in absolute value. In this regard, it should be pointed out that the effect of the thickness on the emissions beyond the manufacturing process, from A4 (tile distribution) to C4 (end-of-life), also shows slight differences in absolute values among the studied scenarios. Even some of them (specifically A5 and B2, installation and maintenance respectively) are practically independent of the studied technological scenarios.
The main effect of technological scenarios on CO 2 eq. emissions can be clearly observed in stage A3 (manufacturing process). To highlight this, an explicit figure (Fig. 8) with a gate-to-gate scope (i.e. manufacturing stage) has been produced.
For the sake of simplicity, in both Figs. 7 and 8, the CO 2 eq. emissions associated with the studied scenarios are presented in a decreasing series of data. This allows for a better comparison with the EU objectives, and how the scope of the LCA influences the emission values. The comparison of Figs. 7 and 8 points out the need to clearly refer the targets and emission values to a specific scope to avoid misunderstandings and unfair comparisons among products and sectors.  alt-text: Fig. 8 3.2.1 Technological scenarios fulfilling EU's targets for 2020 (Merge the paragraphs of this section 3.2.1 into a single paragraph) Figs. 7 and 8 clearly show that all the simulated technological scenarios fulfilled the EU 's targets for 2020, i.e. a 20% reduction in CO 2 eq. emissions by 2020 compared to 1990.
The current scenario (scenario C) fulfilled the 2020 targets owing to the technological changes experienced in recent years (Celades et al., 2012), but it was slightly lower than the EU's target, therefore any eco-innovation that could be included would ensure compliance with a greater margin.
The technological scenarios with an electric energy source defined for 2020 according to Capros et al. (2013) (Replace this reference: Capros et al. (2013) with " Capros et al., 2013) are the following ones: E; G; I; K; O; and R.
It is worth noting the results obtained with scenario E, which allowed for a significant reduction of the CO 2 eq. emissions with a relatively low innovation effort. This scenario consisted of the simultaneous implementation of the widespread technologies in the drying and firing stages (increasing the thermal energy efficiency), while the rest of the alternatives remained unchanged.
Scenario R, the one with the greatest reductions, using the electric energy sources for 2020 (SGM20) required greater effort and more limitations, because it consisted of manufacturing unglazed lightened tiles (GL0, TH50) with the simultaneous implementation of widespread technologies in the drying and firing stage (WDS), but it was really close to the EU 'targets for 2050, particularly when a gate-to-gate approach was applied.

Technological scenarios fulfilling EU's targets for 2050
Figs. 7 and 8 show that few simulated scenarios fulfilled the EU 's targets for 2050, i.e. an 85% reduction in CO 2 eq. emissions by 2050 compared to 1990, especially when a cradle-to-grave approach was employed.
Effectively, when reduction targets were applied to the entire product life cycle, only four scenarios (W, X, Y, and Z) could fulfil the requirements of the Commission (Fig. 7). The common characteristics of these technological scenarios are: changes in product design (reductions of the thickness of the body (TH50) and reduction to half the quantity of glaze (GL50)); implementation of widespread technologies in thermal energy efficiency (WDS). Furthermore, full electrification of thermal processes (NG0) from renewable sources (REN50) would be jointly needed. It is interesting to point out that the route to prepare the body raw materials did not seem to have a significant effect on the carbon footprint, as explained in 3.1.2.
If the objectives were only focused on the ceramic tile manufacturing stage (A3), a less demanding implementation of technological alternatives would be needed to achieve the objectives of the European Commission. The technological scenarios that met the requirements of the roadmap were eleven: M; N; Q; S; T; U; V; W; X; Y; and Z (see Fig. 8). These objectives could be achieved either by modifying the product design (removal of the glaze (GL0) and reduction of the thickness of the ceramic body (TH50)) or by electrifying 50% of the thermal processes (NG50) through renewable sources (REN50). In all these cases, the implementation of widespread technologies in thermal energy efficiency (WDS) was considered.

Conclusions
It has been proven that the greater the scope of the LCA study is, the greater eco-innovations are needed. The main environmental advantages appear in the use and end-of-life stages of the PST. This happens thanks to their long lifespan, easy cleaning and maintenance, the inert nature of the end-of-life waste flows, and their simplicity or open loop recycling. It is necessary then to clarify the targets and emission values to a specific scope, in order to avoid misunderstandings and unfair comparisons with products or sectors (wood floorings, carpets, etc.), which may have fewer production impacts but higher ones when it comes to use or disposal.
From the studied technological alternatives, the incorporation of widespread technologies to increase energy efficiency of the thermal processes up to 45% (WDS) and the thickness reduction of the ceramic body (TH50) were the alternatives that implied the higher reductions of CO 2 eq. emissions. It has been ensured that none of the alternatives implied burden shifting among the different environmental impact categories. On the other hand, the increase of renewable energies in the SGM entailed relevant improvements.
Neither the electrification of thermal processes (NG50/NG0) nor the preparation of raw materials following the dry route (DRY) seemed to be interesting measures when an overall life cycle approach was used, unless the electric generation was dramatically decarbonized using renewable sources of energy, since the environmental burdens would shift to other categories and processes.
In the evaluation of the different technological scenarios, the objectives for 2020 were found to be almost fulfilled thanks to the technological advances already being implemented in the European sector of ceramic tile manufacturing. However, the objectives for 2050 are far from being met, and the implementation of endogenous widespread technology will not be enough. Therefore, a combination of endogenous and exogeneous breakthrough technologies must be applied. These breakthrough technologies mainly lie on a decrease in the dependence on non-renewable fuels, the implementation of highly efficient energy measures and the application of product eco-design innovations.
Consequently, to support the ceramic tile industry in this context, it is necessary to find realistic solutions without jeopardizing its survival in a low-carbon economy. In this regard, further research is needed to evaluate technical and economic feasibility of the studied scenarios. In addition, other technical alternatives could be studied, such as using new glaze compositions or evaluating measures to promote the transition of the ceramic tile sector to a circular economy, among others.