Recently, several countries have committed to international agreements to reduce greenhouse gases (GHG) emissions. Therefore, it is necessary to carry out an evaluation of the most polluting processes. The brick kiln industry represents one of the major small-scale industries which fulfill the growing demand for urban expansion. This industry generates negative impacts on the environment in respect of air quality, human health and vegetation in particular (Skinder et al. 2014). Ceramic bricks are one of the most impacting materials in the whole building process (Martínez-Rocamora et al. 2016), even though this appears to be a very profitable business (Singh and Asgher 2005). China tops global production and its manufacturing process is increasingly dominated by modern technologies (Schmidt 2013). About 250 billion bricks are produced in the Indian industry, which is the world’s second largest-brick producer overall and the largest artisanal producer (Rajarathnam et al. 2014). Nevertheless, the producers of Latin America tend to rely on the most primitive and most polluting types of kilns (Schmidt 2013). There are more than 17 000 artisan bricklayers in Mexico according to the Instituto Nacional de Ecología y Cambio Climático (National Institute of Ecology and Climate Change (INECC 2013).

The brick kilns may be grouped into two broad categories known as intermittent and continuous kilns. Most of the kilns around the world that contribute to the pollution of environments are traditional intermittent (Rajarathnam et al. 2014). It seems therefore fundamental to asses and quantify the environmental impact of this production process (Hoxha et al. 2016).

Life cycle assessment (LCA) is a methodological tool used in industry to evaluate the environmental impacts of products, and improve operation and manufacturing efficiency (Güereca et al. 2015). Actually, the most frequently addressed topics are waste management, construction, agriculture, transport, and chemical sector, among others (Bovea et al. 2017). Some international LCA studies have been conducted to analyze the environmental impacts of buildings (Bribián et al. 2009, Saner et al. 2013, Zhang et al. 2013, Hoxha et al. 2016) and construction materials (Buyle et al. 2013, Castell et al. 2013, Ibbotson and Kara 2013, Lasvaux et al. 2015). Recently, it has been found that construction materials can be responsible for 20-90 % of the global warming potential (GWP) indicator and the variation is mainly due to the building projects and the embodied energy of the materials employed (Hoxha et al. 2016). However, information available on the life-cycle impacts of the traditional brick manufacturing process is limited. The aim of this work is to quantify the environmental impacts of Mexican brick manufacturing scenarios, evaluating an improvement in performance.

LCA studies could present the limitations of their application in different countries, with the exception of the resources damaged and the damage created by climate change, ozone layer depletion, air emissions of persistent carcinogenic substances and inorganic air pollutants with long-range dispersion. Lopsik (2013) concluded that the limitations and the assumptions made during the LCA study should be taken into account for interpreting impact assessment results.

Also, when using LCA, it is important to make a conscious choice of the database and LCIA methods to be used, knowing the consequences of decisions made in the study (Zhang et al. 2013). According to Rørbech et al. (2014), the selection of a method in specific LCAs may significantly affect its results and comparability with other studies. Thus, a sensitivity analysis was carried out based on comparing the variation of results using the LCIA methods. The analyses of the three scenarios are discussed below.

Eco-indicator 99. The results are shown in figure 3 , where a substantial decrease is observed (scenarios I-III) in impact categories like carcinogens, respiratory inorganic, ecotoxicity acidification/eutrophication and climate change. This last category is influenced by GHG emissions. Otherwise, the categories radiation, ozone layer, minerals and fossil fuels show an increase of less than 6.5 % caused by electricity consumption. The land use category remained constant regardless of the presence of the HB device.

Fig. 3 Fig. 3 Average contributions of life cycle impact assessment using Eco-indicator 99 damage approach for scenario I (traditional manufacturing), scenario II (one hopper-blower) and scenario III (two hopper-blowers). CA: carcinogens, RO: respiratory organics, RI: respiratory inorganics, CC: climate change, Ra: radiation, OL: ozone layer, ET: ecotoxicity, A/E: acidification/eutrophication, LU: land use, Mi: minerals, FF: fossil fuels (nomenclature in table IV)

Impact 2002+. Figure 4 shows a decrease through scenarios II and III in the categories carcinogens, no carcinogens, respiratory inorganics, aquatic and terrestrial ecotoxicity, terrestrial and aquatic acidification as well as global warming, which includes GHG emissions. The ionizing radiation, ozone layer depletion and respiratory organics categories present a slight increase of ~3 %, attributed to the addition of HB devices. On the other hand, there is an increase of ~46 % in the score of scenario II comparing to scenario I, whereas scenario III is ~63 % higher tan scenario I. This slight recovery percentage is attributed to the electricity demand characteristic of the fuel consumption (natural gas) of the combined cycle plant under study.

Fig. 4 Fig. 4 Average contributions of life cycle impact assessment using IMPACT 2002 midpoint damage approach for scenario I (traditional manufacturing), scenario II (one hopper-blower) and scenario III (two hopper-blowers). CA: carcinogens, N-CA: non-carcinogens, RI: respiratory inorganics, IR: ionizing radiation, OLD: ozone layer depletion, RO: respiratory organics, AE-Tox: aquatic ecotoxicity, TE: terrestrial ecotoxicity, T-A/N: terrestrial acidification/nutrification, LO: land occupation, AA: aquatic acidification, AE: aquatic eutrophication, GW: global warming, N-RE: non-renewable energy, Mex: mineral extraction (nomenclature in table V)

CML 2001. Both CML 2001 and the ReCiPe methods ( Figs. 5 and 6 ) consider the emissions in a greater number of categories and at different periods of time (10 and 20 yrs). In this method, a decrease in impact categories of ~50 % (scenarios I-III) is observed in the categories acidification, eutrophication, global warming (all of the time lapses presented), human toxicity, fresh water aquatic ecotoxicity (all of the time lapses), marine ecotoxicity infinite, terrestrial ecotoxicity (20a and infinity), marine sediment ecotoxicity (20a and infinity), fresh water sediment ecotoxicity (20a and infinity), and average European (kg of NO_x and SO_x). An increase greater than 8 % in Ozone layer depletion (for 10a, 20a and steady state), ionizing radiation, photochemical oxidation, malodorous air, equal benefit incremental reactivity, maximal incremental reactivity and maximal ozone incremental reactivity categories, can be seen in figure 5 . The categories related to ozone layer depletion remain constant trough the scenarios. In the photochemical oxidation category, a substantial decrease was observed between scenarios I and II; however, comparing scenarios II and III, a moderate increase was detected. The behavior in the photochemical oxidation midpoint category is due to carbon monoxide emissions. Scenario III has a higher percentage of carbon monoxide emissions than scenario II.

Fig. 5 Fig. 5 Average contributions of life cycle impact assessment using CML 2001 midpoint approach for scenario I (traditional manufacturing), scenario II (one hopper-blower) and scenario III (two hopper-blower). AC: acidification, EU: eutrophication, GW-20: global warming 20 years, ULGW: upper limit of net global warming, LLGW: lower limit of net global warming, OLD-10: ozone layer depletion 10 years, OLD-20: ozone layer depletion 20 years, OLD-SS: ozone layer depletion steady state, HT-20: human toxicity 20 years, HT-IN: human toxicity infinite, FAE-20: freshwater aquatic ecotoxicity 20 years, MAE-20: marine aquatic ecotoxicity 20 years, MAE-Inf: marine aquatic ecotoxicity infinite, TE-20: terrestrial ecotoxicity 20 years, TE-Inf: terrestrial ecotoxicity infinite, MSE-20: marine sediment ecotoxicity 20 years, MSE-Inf: marine sediment ecotoxicity infinite, FSE-20: freshwater sediment ecotoxicity 20 years, FSE-Inf: freshwater sediment ecotoxicity infinite, AE-NOx: average European (kg NOx eq), AE-SOx: average European (kg SO2 Eq), LC: land competition, IR: ionizing radiation, PhO: photochemical oxidation, PhO_lNOx: photochemical oxidation (low NOx), MA: malodorous air, EBIR: equal benefit incremental reactivity, MaxIR: maximum incremental reactivity, MaxOIR: maximum ozone incremental reactivity (nomenclature in table VI)

Fig. 6 Fig. 6 Average contributions of life cycle impact assessment using ReCiPe midpoint approach for scenario I (traditional manufacturing), scenario II (one hopper-blower) and scenario III (two hopper-blower). CC: climate change, OD: ozone depletion, TA: terrestrial acidification, FE: freshwater eutrophication, Meu: marine eutrophication, HT: human toxicity, POF: photochemical oxidant formation, PMF: particulate matter formation, TE: terrestrial ecotoxicity, FE-Tox: freshwater ecotoxicity, MEc: marine ecotoxicity, IR: ionizing radiation, ALO: agricultural land occupation, ULO: urban land occupation (nomenclature in table VII)

ReCiPe. In figure 6 , a significant reduction of over 50 % (scenarios I-III) is observed in terrestrial acidification, marine eutrophication, human toxicity, photochemical oxidant formation, particulate matter formation, terrestrial ecotoxicity, fresh water ecotoxicity, marine ecotoxicity, and fossil depletion. The other categories remain practically constant. The climate change category decreases 74 % when comparing scenarios I and II, while the comparison between scenario II and III shows an increase of 7 %, which is caused by the use of two HB. The ReCiPe method does not consider the CO gas in its GHG analysis, although this contributes indirectly to the greenhouse effect and also indicates incomplete combustion.

Each LCA database is developed by an organism in a specific location, and the modeled processes are based on its manufacturing characteristics (Martínez-Rocamora et al. 2016). Even when ecoinvent was developed by the Swiss Centre for Life Cycle Inventories, the present study complements this database by including more adequate inventories for Mexico.

A significantly decreasing trend was observed in most of the potential impact categories analyzed using the selected methods. The lower level of emissions arises from the inclusion of one HB device due to the improvement of the manufacturing process efficiency in firing time ( Table III ). According to Kumbhar et al. (2014) the highly energy-intensive process of traditional brick production needs an efficient fuel burning method to reduce the environmental impact of the production process. Additionally, it was observed that a prolonged firing time had no significant effect on the mechanical and physical properties of bricks.

In order to validate this work, table VIII shows a comparison of traditional brick kiln technology (sawdust as fuel) and tunnel brick technology (pet coke as fuel) applied in Greece (Koroneos and Dompros 2007). Even though biomass (sawdust) is supposed to be a cleaner fuel than pet coke, both studies show similar CO₂ emissions. Moreover, any of the reviewed LCA studies consider the category of carcinogens (or impacts related to polychlorinated emissions) in their impact analysis (Khan et al. 2007, Cappuyns and Kessen 2012, Quirós et al. 2015). Table IX shows a comparison of the carcinogens and ecotoxicity categories of the present study with an LCA of traditional soil brick carried out on 16 traditional brick kilns located in India (Kumbhar et al. 2014).

TABLE VIII TABLE VIII COMPARISON OF CO2 EMITTED BY TRADITIONAL BRICK KILNS (MEXICO) AND TUNNEL BRICK KILNS (GREECE)

TABLE IX TABLE IX COMPARISON OF THE CATEGORIES RELATED TO POTENTIAL POLYCHLORINATED EMISSIONS

Pt: points of life cycle assessment

In this study case, the GHG emissions in CO₂ equivalents were calculated from the energy inputs of the various processes in the LCI and they were used as an indicator of energy use. The LCA results identified an improvement opportunity for the energy efficiency of the process by using one HB device in the clamp kiln.

It has been verified that the obtained information is relevant and reliable for the environmental impact comparison, since testing different assessment methods helps to determine the influence on data variations and the sensitivity of processes (Lopsik 2013).

This study assessed the environmental impacts of three clamp kiln scenarios applying an LCA. It is a contribution to the application of LCA to the environmental assessment and the improvement of traditional processes in Mexico, where the evolution of LCA has been slower than in developed countries.

The LCIA results previously discussed were found to be quite sensitive to changes in GHG emissions, which were used as an indicator of energy consumption. An improved environmental performance was achieved when one device was used, as evidenced by the decrease in emissions.

It was observed that there are residual impacts that increase by adding blowers, such as radiation, eutrophication or depletion of the ozone layer. This study identifies the importance of the life-cycle impacts by improving the environmental performance of traditional technologies, especially in countries where economic and social aspects do not allow the complete replacement or technological upgrade of these processes.

Future studies are recommended to analyze particle and NO_x emissions using the HB device during the production stage, even though the firing time decreased.

The inventory information contributes to the development of the national life-cycle inventory (LCI) database and it could be applicable to others countries where the manufacturing process is similar.