Alternative Materials for Printed Circuit Board Production: An Environmental Perspective

: This article investigates the potential environmental impacts of four-layer printed circuit board (PCB) production from cradle to grave. The study starts with a lifecycle assessment of conventional PCB production. Then, the alternative materials of polyethylene terephthalate (PET), polylactic acid (PLA)/glass ﬁber composite and paper are investigated for the substrate. A conventional PCB adopts copper as the conductive material and requires an etching process. The environmental impacts of changing the conductive deposition method to an additive method by printing silver nanoparticles is studied. In a conventional PCB, electricity generation contributes 41% of the global warming potential (GWP) and 38% of the abiotic resource depletion (ADP), in the fossil category. By applying an additive manufacturing method, the GWP of PCB manufacturing can be reduced to 14% of that of the conventional method. A sensitivity analysis of silver recycling illustrates that a 40% higher silver recycling rate would decrease the GWP of silver material by about 48–60%. Uncertainty in the energy consumption of PCB production would alter the environmental impacts; however, even with the most conservative energy consumption in a conventional PCB production method, the environmental impacts of the additive method are about ﬁve times lower than those of conventional PCB production. Conceptualization, M.N.N. M.N.N.; software, M.N.N.; validation, I.D., M.H.; investigation, M.N.N.; writing—original preparation, M.N.N.; writing— review editing, I.D. M.H.; M.N.N.;


Introduction
By 2021, the market for the Internet of Things (IoT) has been forecast to expand to billions of devices [1], which will consequently produce a considerable amount of electronic waste. One requirement of digitalization is that electronics are embedded in products, which, at some point in their lifecycle, may be released into the natural environment. Examples of applications for embedded electronics include sensors in disposable smart packaging, flexible and printable electronics embedded in consumer products, health care products and disposable sensors for weather and crop monitoring. On the other hand, the increasing amount of electronic equipment increases the environmental impacts of their materials and production.
New trends and applications in the electronic industry require novel developments and the re-designing of sustainable alternatives to the existing production stream. This new market opens a discussion on how conventional practices in electronics manufacturing are in accordance with the sustainability principles and how the environment may benefit from the transformation of production practices in the electronics industry.
Printed circuit boards (PCBs) are a fundamental component in almost all electronic products, and circuits, devices and chips are built on PCBs. Conventional multilayer PCBs consist of a core dielectric layer (substrate) clad with copper as a conductive material on

Materials and Methods
The current article is part of the ECOtronics project, in which different research units investigate the electrical circuits, material behavior and commercial aspects of developing printed electronics. The objective of this study is to demonstrate the potential of printed electronics to decrease the environmental impacts throughout the products' lifecycle. The scenarios in this article were based on the best commercially available materials and could help electronic circuit designers to understand the environmental profile of their products.
The data used to conduct a lifecycle assessment were gathered by searching through the following sources: (1) LCA databases (Ecoinvent database or GaBi software); (2) Literature study; (3) Contacting specialists in the field.
A lifecycle assessment (LCA) quantifies the potential environmental impacts throughout the products' lifecycle [23]. The LCA framework is implemented in four phases: goal and scope definition, lifecycle inventory (LCI), lifecycle impact assessment (LCIA) and interpretation. In addition, a sensitivity analysis is recommended to enhance the creditability of the LCA study [23].

Goal and Scope Definition
The goal of this LCA study was to assess and compare the environmental impact of alternative production methods and materials for PCB manufacturing. The conventional substrate in this study refers to fiber-glass-reinforced brominated epoxy resin (FR4) and the comparative cases are PET, PLA/GF composite and paper. The conductive material in the baseline is copper and, in other cases, the baseline is an electrically conductive adhesive, which is a combination of 50% resin and 50% silver nanoparticles.
The manufacturing method for the baseline scenario is conventional copper treatment (chemical etching is involved), and, in the other cases, the method is lab-scale PCB production, which is mainly based on the experimental procedure described in [6].
The functional unit in this LCA study was set as one m 2 of four-layer PCB. The thickness of the conventional PCB substrate is 1.6 mm, whereas, for the alternative cases, the thickness of the substrate material was retrieved from [8], which has two layers of 125 µm and one layer of 36 µm.
The system boundary of this study is depicted in Figure 1, covering cradle-to-grave processes. It includes the raw material acquisition, manufacturing, and waste disposal of PCB. Transport and use phases are excluded from the system boundaries because they are assumed to be the same for all alternatives. Most of the utilized data represent global conditions; however, due to the lack of global data, some data sources correspond to more specific geographical locations.

Scenario Description
Several possible alternative materials were studied for the PCB substrate. PET, PLA/GF and paper were selected in this study, and etched copper was substituted with ECA containing silver nanoparticles (Ag NPs). Five study scenarios are listed in Table 1. Paper Ag NPs S1 represents the baseline scenario for a conventional FR4-based PCB, and S2 is a hypothetical scenario designed to check the environmental impacts when only the conductive material is changed to Ag NPs. The intention of introducing S2 was to demonstrate the environmental benefits when only the manufacturing process (not the substrate material) is changed from a subtractive method to an additive method. Then, in scenarios S3, S4 and S5, the substrate material is also changed. In S2-S5, Ag NPs were utilized as the conductive material, deposited on the substrate using the screen-printing method. S3 represents a scenario related to the PET-based substrate. S4 depicts the scenario derived from the PLA/GF composite. The substrate material in scenario S5 is normal printing paper.
From the functional perspective, the function of S3, S4 and S5 is, to some extent, different; PET plastic is applicable, whereas a transparent substrate is needed. PLA/GF is a bio-compostable plastic and would be used in bioplastic packaging. Using paper as the substrate has the advantage of biodegradability and is suitable in paper and cardboard packaging. The objective of this article was not to compare different substrate materials, but to show the environmental impacts of moving toward sustainable electronics.

Scenario Description
Several possible alternative materials were studied for the PCB substrate. PET, PLA/GF and paper were selected in this study, and etched copper was substituted with ECA containing silver nanoparticles (Ag NPs). Five study scenarios are listed in Table 1. Paper Ag NPs S1 represents the baseline scenario for a conventional FR4-based PCB, and S2 is a hypothetical scenario designed to check the environmental impacts when only the conductive material is changed to Ag NPs. The intention of introducing S2 was to demonstrate the environmental benefits when only the manufacturing process (not the substrate material) is changed from a subtractive method to an additive method. Then, in scenarios S3, S4 and S5, the substrate material is also changed. In S2-S5, Ag NPs were utilized as the conductive material, deposited on the substrate using the screen-printing method. S3 represents a scenario related to the PET-based substrate. S4 depicts the scenario derived from the PLA/GF composite. The substrate material in scenario S5 is normal printing paper.
From the functional perspective, the function of S3, S4 and S5 is, to some extent, different; PET plastic is applicable, whereas a transparent substrate is needed. PLA/GF is a bio-compostable plastic and would be used in bioplastic packaging. Using paper as the substrate has the advantage of biodegradability and is suitable in paper and cardboard packaging. The objective of this article was not to compare different substrate materials, but to show the environmental impacts of moving toward sustainable electronics.

Lifecycle Inventory (LCI)
LCI is an LCA stage quantifying the inputs and outputs of the studied system, and is performed for each scenario. Table 2 shows the data source for each unit process. The LCIs for processes in Scenarios 1 and 2 were mainly obtained from the Ecoinvent 3.6 database, with the activity name "printed wiring board production, for surface mounting, Pb free surface". This process represents the manufacturing of 1 m 2 of a six-layer PCB based on the FR4 substrate. The dataset was compiled using the data collected during 2003-2006 from environmental reports of three PCB manufacturers in Europe and it covers the manufacturing processes of a printed wiring (circuit) board: raw materials, energy consumption, emission to air and water and waste amount from the production of 1 m 2 of PCB [24].
The literature has reported controversial results regarding the electricity and heat consumption involved in the manufacturing of conventional FR4-based PCBs (S1 and S2). The electricity and heat consumption values in the Ecoinvent database for a six-layer PCB are 326 and 105 kWh. The authors in [6] collected the inventory data from a manufacturing plant in China and reported a total energy consumption for the manufacturing process of 20 kWh. The authors in [25] collected the inventory data from a manufacturing plant in Turkey and reported the total energy consumption for the manufacturing process of 1 m 2 of a 2-layer PCB as 7.15 kWh. The authors in [5] reported the energy consumption of a 1 m 2 PCB substrate (only the FR4) as 13.4 kWh. The authors in [26] studied the energy consumption of PCB manufacturing and calculated the energy demand for each unit process of a four-layer PCB production. They reported the total energy demand for the manufacturing of a panel of a 4-layer PCB as about 6 kWh. Considering the typical size of a PCB panel (18 × 24", 18 × 21", 21 × 24") [27], the energy demand for 1 m 2 of a 4-layer PCB would be 18-25 kWh. Therefore, in Scenario 1 of this study, the total energy consumption for manufacturing 1 m 2 of a 4-layer PCB is assumed to be 20 kWh and, in Scenario 2, 10 kWh is taken as the electricity consumption. To assess the impact of uncertainty regarding this parameter on the results, a sensitivity analysis was carried out.
As mentioned above, the functional unit of this study was set to 1 m 2 of a four-layer PCB. However, in the Ecoinvent database, LCI only exists for a two-layer and six-layer PCB. Thus, it is necessary to convert the inventory dataset of a 6-layer PCB into a four-layer PCB. The structures of the multilayer PCBs produced by different manufacturers are roughly similar. As an example, the researchers in [28] described the structure of typical four-and six-layer PCBs. The quantity of sheet rolling and the mass of copper used in the dataset for a six-layer PCB was modified by factor of 66% to represent the inventory to produce 1 m 2 of a four-layer PCB.
The inventory data for the manufacturing of a PCB in Scenarios 3-5 were mainly obtained from [6], since, at the date of this study, [6] is the only article which reported a full inventory of materials and energy consumption for the manufacturing of a multilayer PCB based on a paper substrate. In Scenarios 3 and 4, the thickness of substrates was retrieved from [8] and the weight of each material in the substrate was calculated using the average density of 1.38 g/cm 3 and 1.5 g/cm 3 for PET and PLA/GF substrates, respectively.
The waste management of a conventional PCB is a complicated and multi-process procedure, and an environmental impact assessment of the end-of-life (EoL) stage of a conventional PCB is not the focus of this study. Therefore, in Scenarios 1 and 2, the environmental impact corresponding to Scenario 1 of [29] is imported to the model in the GaBi program. In the EoL stage of the other scenarios, it is assumed that 50% of the silver contained in the used PCB can be separated and recycled [30], and the rest of the PCB material is incinerated. In Scenarios 3, 4 and 5, for the purposes of modeling, each material fraction of PCB (PET, resin, paper, PLA and GF) was modeled separately in its corresponding incineration unit process. * Amount for: S1; ** Amount for: S2.
The authors in [31] studied the possible routes for the manufacturing of silver nanoparticles. Ag NPs are produced through several synthesis methods with different reactions, inputs and particle sizes. The authors in [32] conducted a comprehensive LCA study for various Ag NPs' production routes. They selected three production methods of chemical reduction with trisodium citrate (CR-TSC) as applicable for the electronics industry. The inventory data regarding the manufacturing of Ag NPs were taken from [31].

Lifecycle Impact Assessment (LCIA)
LCA studies were carried out in GaBi 9 software. The CML 2001-August 2016 method was selected to assess the environmental impacts. The selected environmental impact categories used to assign LCI results to specific environmental issues were as follows: Abiotic Depletion Potential (ADP) (fossil), Acidification Potential (AP), Eutrophication Potential (EP), Freshwater Aquatic Ecotoxicity Potential (FAETP), Global Warming Potential (GWP), Human Toxic Potential (HTP), Ozone Layer Depletion Potential (ODP), Photochemical Ozone Creation Potential (POCP) and Terrestrial Eco-toxicity Potential (TETP). These categories can describe the environmental impact of PCB production.

Results
The results presented in Figure 2 are based on the environmental assessment of each scenario described in the lifecycle inventory section. They depict the environmental impacts for all impact categories. S1 has the highest environmental impacts in all categories. Comparing the environmental impacts of S1 with S3, S4 and S5 in all impact categories, the additive manufacturing of PCBs using AgNPs and PET, PLA/GF and paper as substates has less than 80% of the impact of the baseline scenario. The primary reason for the lower impacts in the latter scenarios stems from the simplicity of the inventory used in S3, S4 and S5, and the lower material intake. For instance, about 2.87 kg of copper was consumed to manufacture 1 m 2 of the baseline PCB. However, in S3, S4 and S5, only 0.0103 kg of AgNPs was used. Furthermore, in S1, several chemical substances were consumed, mainly for the treatment of etched copper. Finally, S1 follows the subtractive method, which is energy-intensive and consumes a large amount of chemicals, whereas scenarios S3, S4 and S5 follow the additive method, which is material-efficient and requires considerably less energy. The environmental impacts of scenarios S3, S4 and S5 in all the categories are almost equal, and the variation is within the range of 4% of the S1 amount. In ADP fossil and GWP categories, S2 has a 48% and 49% lower impact. Scenario S2 has the same substrate as S1; thus, by changing the material from copper to silver nanoparticles, there would be a greater than 50% decrease in the environmental impacts of S1.     Figure 2 shows the relative environmental impacts of these scenarios and is suitable for comparing the scenarios; however, a contribution analysis sheds light onto the reasons for the different impacts of the scenarios. Figure 3 shows the contribution of different processes and materials to the environmental impacts of each scenario. The vertical axis corresponds to the share of each process from the total environmental impacts. The absolute value of the environmental impacts for each process is also denoted in Figure 3. The inventory processes are aggregated into four groups of "Electricity", "Conductive", "Substrate" and "Chemicals" to distinguish the contribution of each group of materials. The "Electricity" group represents the electricity consumption for each manufacturing scenario; the "Conductive" group corresponds to the processes related to conductive metal, which in S1 is copper use and in S2, S3, S4 and S5 is silver from the raw material intake; and the material deposition to the EoL stage.
The "Chemicals" group in S1 represents the unit processes of iron(III)chloride, sodium chloride, sodium hydroxide, water, hydrogen peroxide, sulfuric acid, dipropylene glycol monomethyl ether and hydrochloric acid. These substances are mainly consumed during etching and copper recovery at the production plant. The "Substrate" group represents the materials incorporated in the substrate of PCBs. Figure 3 demonstrates the aggregated environmental impacts of both the production and EoL stages for substrate and conductive groups.

Sensitivity Analysis
3.1.1. Sensitivity Analysis of Silver Recovery Rate for S3, S4 and S5 Figure 4 shows the sensitivity of the LCIA results to the recycling rate for silver in S3, S4 and S5. As shown in Figure 4, increasing the silver recovery rate to 70% reduces GWP and AP by about 30% and EP and HTP by approximately 40%. Thus, elevating the silver recycling rate from 50% to 90% causes a 48% reduction in the GWP impact category.  In Scenario 1, electricity is the main contributor to GWP, at 41%, followed by chemical processes in manufacturing and the FR4 substrate. However, in EP and AP, copper treatment and chemicals are the main sources of emissions.
In Scenarios 3, 4 and 5, the conductive material (Ag NPs) is the most significant contributor in all environmental impact categories, even though it accounts for less than 3% of the total substrate weight. This is due to the extensive material and energy consumption involved in the production of silver material. However, the substrate materials in scenarios S3, S4 and S5 have a 14%, 12% and 2% contributions to the GWP category.
In the HTP category, the environmental impacts of the conductive material are the dominant burden for all scenarios. Notably, the HTP value for S1 is more than 10 times higher than the other scenarios. The high difference in the HTP value of copper mainly stems from the higher metal (copper) consumption in S1.
The silver recovery rate was assumed to be 50% in this study. However, due to the high impact of silver in the latter cases, the higher silver recovery rate might diminish the environmental burdens of these scenarios. Therefore, a sensitivity analysis of the silver recovery rate was carried out.
3.1. Sensitivity Analysis 3.1.1. Sensitivity Analysis of Silver Recovery Rate for S3, S4 and S5 Figure 4 shows the sensitivity of the LCIA results to the recycling rate for silver in S3, S4 and S5. As shown in Figure 4, increasing the silver recovery rate to 70% reduces GWP and AP by about 30% and EP and HTP by approximately 40%. Thus, elevating the silver recycling rate from 50% to 90% causes a 48% reduction in the GWP impact category. tronic products per capita [33].
As this analysis illustrates, there would be environmental benefits from the recovery of silver content from electronic waste; however, even in Europe, the lack of recycling facilities hinders the recycling of metal contents of scrap electronics [30]. The aim of this sensitivity analysis was to depict the potential benefits of a higher recycling rate of silver and examine the uncertainty of the results. To conduct a comprehensive environmental study, the impacts of the silver recovery process should be also included.  Figure 5 demonstrates the sensitivity analysis of the energy consumption in Scenario 1. Regarding energy (electricity and heat) consumption, in S1, as mentioned in the LCI section, the Ecoinvent 3.6 database reported higher values for PCB production. Although, in this study, a conservative value of 20 kWh was adopted for electricity consumption in S1, the environmental impacts of a higher energy consumption are shown in Figure 5. The recovery rate of silver depends on the silver content in the scrap electronics and recycling practices, which varies in different geographical locations. The Nordic region is considered a high-ranked, affluent part of Europe, due to its high consumption of electronic products per capita [33].

Sensitivity Analysis of Energy Consumption
As this analysis illustrates, there would be environmental benefits from the recovery of silver content from electronic waste; however, even in Europe, the lack of recycling facilities hinders the recycling of metal contents of scrap electronics [30]. The aim of this sensitivity analysis was to depict the potential benefits of a higher recycling rate of silver and examine the uncertainty of the results. To conduct a comprehensive environmental study, the impacts of the silver recovery process should be also included. Figure 5 demonstrates the sensitivity analysis of the energy consumption in Scenario 1. Regarding energy (electricity and heat) consumption, in S1, as mentioned in the LCI section, the Ecoinvent 3.6 database reported higher values for PCB production. Although, in this study, a conservative value of 20 kWh was adopted for electricity consumption in S1, the environmental impacts of a higher energy consumption are shown in Figure 5.

Sensitivity Analysis of Energy Consumption
In case the electricity consumption increases by 10%, the GWP of S1 will increase by 25%. The electricity consumption in this scenario causes high uncertainty in the environmental impact assessment of scenarios 1 and 2. As depicted in Figure 5, the reduction in the electricity consumption in S1 to 10 kWh did not result in a lower impact compared to scenarios S3-S5. Sustainability 2021, 13, x FOR PEER REVIEW 10 of 13 In case the electricity consumption increases by 10%, the GWP of S1 will increase by 25%. The electricity consumption in this scenario causes high uncertainty in the environmental impact assessment of scenarios 1 and 2. As depicted in Figure 5, the reduction in the electricity consumption in S1 to 10 kWh did not result in a lower impact compared to scenarios S3-S5.

Sensitivity Analysis of Ag NP Thickness and Net Area Coverage
The thickness of the ink film on the substrate may vary in different PCB designs to fulfill the electronical requirements of the product. In addition, the net area covered by the silver ink depends on the design parameters of the PCB. Variations in these two parameters affect the mass of the silver ink, and consequently change the mass of silver and the resin consumption in S2-S5. A sensitivity analysis was carried out to determine the effect of variations in Ag NP thickness and net area coverage. Thus, in scenarios S2, S3, S4 and S5, three ink thicknesses of 5, 10 and 15 μm and two values for net area coverage (50% and 80%) were utilized to examine the dependency of four LCI categories of GWP, AP, EP and HTP on the thickness and net area coverage of Ag NP. Figure 6 shows the sensitivity analysis of the thickness and net area coverage of ink. The environmental impacts of S1 for the inventory listed in Table 2 were selected as the reference for other cases. The baseline in this case represents a thickness of 5 μm and area coverage of 50%. The charts provide data for the five scenarios described in Table 1, and six thickness-covered area combinations for Ag NPs.
Regarding the GWP of S2, when the ink layer thickness increases from 5 to 15 μm and the net ink coverage increases from 50% to 80%, the environmental impacts are doubled and the GWP of S2 is the same as that of S1. The variation in environmental impacts in S3, S4 and S5 for the selected impact categories are approximately similar and about four times higher when the thickness of ink changes from 5 to 15 μm and the net ink coverage increases from 50% to 80%. The thickness and area coverage of silver have a direct effect on the mass of Ag NP: a thicker ink or higher area coverage leads to a higher consumption of Ag NP.

Sensitivity Analysis of Ag NP Thickness and Net Area Coverage
The thickness of the ink film on the substrate may vary in different PCB designs to fulfill the electronical requirements of the product. In addition, the net area covered by the silver ink depends on the design parameters of the PCB. Variations in these two parameters affect the mass of the silver ink, and consequently change the mass of silver and the resin consumption in S2-S5. A sensitivity analysis was carried out to determine the effect of variations in Ag NP thickness and net area coverage. Thus, in scenarios S2, S3, S4 and S5, three ink thicknesses of 5, 10 and 15 µm and two values for net area coverage (50% and 80%) were utilized to examine the dependency of four LCI categories of GWP, AP, EP and HTP on the thickness and net area coverage of Ag NP. Figure 6 shows the sensitivity analysis of the thickness and net area coverage of ink. The environmental impacts of S1 for the inventory listed in Table 2 were selected as the reference for other cases. The baseline in this case represents a thickness of 5 µm and area coverage of 50%. The charts provide data for the five scenarios described in Table 1, and six thickness-covered area combinations for Ag NPs.
Regarding the GWP of S2, when the ink layer thickness increases from 5 to 15 µm and the net ink coverage increases from 50% to 80%, the environmental impacts are doubled and the GWP of S2 is the same as that of S1. The variation in environmental impacts in S3, S4 and S5 for the selected impact categories are approximately similar and about four times higher when the thickness of ink changes from 5 to 15 µm and the net ink coverage increases from 50% to 80%. The thickness and area coverage of silver have a direct effect on the mass of Ag NP: a thicker ink or higher area coverage leads to a higher consumption of Ag NP. Sustainability 2021, 13, x FOR PEER REVIEW 11 of 13 Figure 6. Sensitivity analysis of Ag NP ink thickness and net area coverage; at the baseline (BL) of S2-S5, the thickness of ink is 5 μm and 50% of the substrate is covered by silver ink.

Conclusions
In this study, we investigated the environmental impacts of the manufacturing of conventional printed circuit boards (PCBs) and the environmental benefits of alternative manufacturing methods. The focus of this study was the utilization of PET, PLA/GF and paper as the PCB substrates and silver nanoparticles as the conductive material. The results of the environmental lifecycle assessment (LCA) showed a significant decline in all the studied environmental impact categories for the scenarios with PET, PLA/GF and paper substrates.
In conventional PCB manufacturing, electricity and chemicals had the highest impacts for GWP and accounted for 41% and 25% of the total impacts. In scenarios 3, 4 and 5, silver nanoparticles had the highest contribution; silver accounted for 83%, 84% and 94% of the GWP in scenarios S3, S4 and S5, respectively.
There is a very large range in the electricity consumption values for S1 in the literature, and so it is a sensitive and uncertain modeling parameter. Our selection for the electricity consumption followed a conservative approach and complied with the lowest reasonable value; however, as shown in the sensitivity analysis section, the conclusions of the study remained the same.
In the PCB manufacturing industry, the transition from subtractive manufacturing to additive manufacturing may significantly decrease the environmental impacts. When using heavy metals such as silver as the conductive material, it is crucial to minimize the material consumption to diminish the environmental impacts of PCB production. On the other hand, a higher recycling rate of metals in the EoL stage could reduce the total environmental impacts.

Conclusions
In this study, we investigated the environmental impacts of the manufacturing of conventional printed circuit boards (PCBs) and the environmental benefits of alternative manufacturing methods. The focus of this study was the utilization of PET, PLA/GF and paper as the PCB substrates and silver nanoparticles as the conductive material. The results of the environmental lifecycle assessment (LCA) showed a significant decline in all the studied environmental impact categories for the scenarios with PET, PLA/GF and paper substrates.
In conventional PCB manufacturing, electricity and chemicals had the highest impacts for GWP and accounted for 41% and 25% of the total impacts. In scenarios 3, 4 and 5, silver nanoparticles had the highest contribution; silver accounted for 83%, 84% and 94% of the GWP in scenarios S3, S4 and S5, respectively.
There is a very large range in the electricity consumption values for S1 in the literature, and so it is a sensitive and uncertain modeling parameter. Our selection for the electricity consumption followed a conservative approach and complied with the lowest reasonable value; however, as shown in the sensitivity analysis section, the conclusions of the study remained the same.
In the PCB manufacturing industry, the transition from subtractive manufacturing to additive manufacturing may significantly decrease the environmental impacts. When using heavy metals such as silver as the conductive material, it is crucial to minimize the material consumption to diminish the environmental impacts of PCB production. On the other hand, a higher recycling rate of metals in the EoL stage could reduce the total environmental impacts.