Environmental Life Cycle Assessment and Techno-Economic Analysis of Triboelectric Cite This: Energy Environ

Energy & Environmental Science

View Article Online ANALYSIS View Journal | View Issue

Environmental life cycle assessment and techno-economic analysis of triboelectric Cite this: Energy Environ. Sci., 2017, 10,653 nanogenerators†

Abdelsalam Ahmed,‡ab Islam Hassan,‡bc Taofeeq Ibn-Mohammed,‡de Hassan Mostafa,fg Ian M. Reaney,h Lenny S. C. Koh,de Jean Zub and Zhong Lin Wang*ai

As the world economy grows and industrialization of the developing countries increases, the demand for energy continues to rise. Triboelectric nanogenerators (TENGs) have been touted as having great potential for low-carbon, non-fossil fuel energy generation. Mechanical energies from, amongst others, body motion, vibration, wind and waves are captured and converted by TENGs to harvest electricity, thereby minimizing global fossil fuel consumption. However, only by ascertaining performance efficiency along with low material and manufacturing costs as well as a favorable environmental profile in comparison with other energy harvesting technologies, can the true potential of TENGs be established. This paper presents a detailed techno-economic lifecycle assessment of two representative examples of TENG modules, one with a high performance efficiency (Module A) and the other with a lower efficiency (Module B) both fabricated using low-cost materials. The results are discussed across a number of sustainability metrics in the context of other energy harvesting technologies, notably photovoltaics.

Module A possesses a better environmental profile, lower cost of production, lower CO2 emissions and shorter energy payback period (EPBP) compared to Module B. However, the environmental profile of Module B is slightly degraded due to the higher content of acrylic in its architecture and higher electrical energy consumption during fabrication. The end of life scenario of acrylic is environmentally viable given its recyclability and reuse potential and it does not generate toxic gases that are harmful to humans and

Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. the environment during combustion processes due to its stability during exposure to ultraviolet radiation. Despite the adoption of a less optimum laboratory manufacturing route, TENG modules Received 17th January 2017, generally have a better environmental profile than commercialized Si based and organic solar cells, but Accepted 22nd February 2017 Module B has a slightly higher energy payback period than PV technology based on perovskite- DOI: 10.1039/c7ee00158d structured methyl ammonium lead iodide. Overall, we recommend that future research into TENGs should focus on improving system performance, material optimization and more importantly improving rsc.li/ees their lifespan to realize their full potential.

a School of Materials Science & Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0245, USA. E-mail: [email protected] b NanoGenerators & NanoEngineering Laboratory, School of Mechanical & Industrial Engineering, University of Toronto, Toronto, M5S 3G8, Canada c Design & Production Engineering Department, Faculty of Engineering, Ain Shams University, Cairo, 11535, Egypt d Centre for Energy, Environment & Sustainability, The University of Sheffield, Sheffield, S10 1FL, UK e Advanced Resource Efficiency Centre, The University of Sheffield, Sheffield, S10 1FL, UK f Department of Electronics and Communications, Faculty of Engineering, Cairo University, Giza, Egypt g Center of Nanoelectronics and Devices (CND) at Zewail City and AUC, Egypt h Departments of Materials Science & Engineering, University of Sheffield, Sheffield, S1 3JD, UK. E-mail: i.m.reaney@sheffield.ac.uk i Beijing Institute of Nanoenergy & Nanosystems, Chinese Academy of Sciences, Beijing, 100083, China † Electronic supplementary information (ESI) available. See DOI: 10.1039/c7ee00158d ‡ A. Ahmed, I. Hassan, and T. Ibn-Mohammed contributed equally to this work.

This journal is © The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10,653--671| 653 View Article Online

Analysis Energy & Environmental Science

Broader context The ability of triboelectric nanogenerators (TENGs) to convert mechanical energy from various sources into useful electrical energy has drawn attention in recent years. Given their potential for low cost energy generation for self-powered applications, it is important to assess their environmental profile and cost viability by carrying out a detailed techno-economic lifecycle assessment. This will provide an indication as to whether they constitute new environmental challenges or not. In this paper, a robust environmental life cycle assessment within a techno-economic framework is carried out for two TENG modules in the context of other energy harvesting technologies. 11 lifecycle environmental metrics as well as the energy payback periods and carbon dioxide emission factors are determined. Although the environmental impact of both modules is lower compared to traditional PV technologies, the higher quantities of acrylic in one of the modules along with energy-intensive fabrication led to a slightly higher environmental burden. Material optimization focusing on reduced material utilization as well as better fabrication processes should, however, improve their environmental profile. Uncertainty sensitivity analysis is conducted to provide deeper insights into TENGs. The current work therefore lays the foundation for future investigations into the profile of TENGs for environmentally friendly innovation in the energy sector. To have a significant impact, technological solutions capable of harvesting electricity from mechanical energy must also be competitive within the marketplace.

1. Introduction physical contact; the contact induces triboelectric charges and generates a potential drop when the two surfaces are separated The burning of fossil fuels is responsible for 480% of primary by a mechanical force, causing electrons to flow between the energy demands and current profiles reveal that the world two electrodes built on the two surfaces.3,16 Following the first remains highly dependent on carbon-based power generation publication on TENGs in 2012, huge progress has been recorded. 1 resulting in the emission of record levels of carbon dioxide (CO2). For instance, by the year 2015, the areal power density had The growth of the world economy, coupled with industrialization reached 500 W m2,17 and the volume power density attained of the developing world, has resulted in a demand for energy that was 15 MW m3, with an instantaneous conversion efficiency of continues to increase.2 Given the growing demand for energy and around 70%.18 TENGs boast a wide range of applications, given dwindling oil reserves, the development of alternative sustainable their capability to harvest mechanical energy from a variety of energy is of paramount importance. Energy from solar, wind and sources, including body motions, vibrations, wind and waves.19 tidal waves has the potential to be integrated with electrical power Additionally, the successful application of TENGs in self-powered grids to meet mega- to gigawatt power requirements.3 The overall chemical sensors has recently been demonstrated20–22 for driving requirements for harvesting these forms of energy are based on electrochemical processes23–25 and commercial light-emitting a number of factors including low-cost, high stability and high diodes (LEDs).26–30 efficiency.3 Several fabrication processes for TENGs have been described An increasingly wide range of mobile electronic devices in the extant literature. Specifically, four modes of operation of often connected to the Internet of Things (IoT) have not only TENGs, including vertical contact-separation mode, in-plane sliding modified our way of life but also have created the need for a mode, single-electrode mode and free-standing triboelectric- Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. highly diversified energy platform.3 For applications such as layer mode were extensively described by Wang et al.3 In this medical care,4 healthcare monitoring, infrastructure monitoring, paper, attention is focused on two fabricated modules. The first environmental protection and security,manysensors,computer is a thin-film-based micro-grating triboelectric nanogenerator control circuits and antennas are required. Although the power (MG-TENG). The operation principle of the MG-TENG relies on for driving each miniature system is relatively small (from micro the coupling between electrostatic induction and the triboelectric to milli-Watt range),3 the collective number of units is forecasted effect.17,31–34 Consisting of two sets of complementary micron by Cisco (the worldwide leader in information technology) to be sized electrode gratings on thin-film polymers, the MG-TENG trillions by the year 2020.5 The use of batteries to power these harvests energy by sliding these surfaces.17 Based on previous units is currently the default solution but this is not sustainable research on this technology, a 0.6 g MG-TENG with an overall area given the large number required and their limited lifespan. of 60 cm2 and a total volume of 0.2 cm3 achieves an average Moreover, the concept of the IoT will be rendered meaningless output power of 3 W (50 mW cm2 or 15 W cm3) and an overall without the inherent ability of devices to be self-powered. This conversion efficiency of roughly 50%, which is sufficient to power challenge has prompted the development of nanogenerators that regular electronics such as light bulbs.17 These performance harvest mechanical energy from the surrounding environment. parameters highlight that MG-TENGs are a promising and effi- Nanogenerators were first developed based on two effects namely, cient solution for harvesting energy from mechanical energy piezoelectricity6–12 and triboelectricity,13–15 with the intention of under ambient conditions. The second module is a TENG based harvesting energy from activities such as walking, talking, typing on two radially arrayed fine electrodes that generates periodically and breathing. A string of groundbreaking research advances changing triboelectric potential and induces alternating currents have subsequently been reported since the landmark publications between the two electrodes. As presented in previous work, at a by Wang and Song.13 rotation rate of 3000 rpm (rotation per minute), a TENG with a

The concept of the triboelectric nanogenerator (TENG) is diameter of 10 cm achieves an output open-circuit voltage (Voc)of

based on the use of the electrostatic charges created on the around 850 V and a short-circuit current (Isc)ofaround3mAata surfaces of two dissimilar materials when they are brought into frequency of 3 kHz. Additionally, with a load of 0.8 M, the TENG

654 | Energy Environ. Sci., 2017, 10,653--671 This journal is © The Royal Society of Chemistry 2017 View Article Online

Energy & Environmental Science Analysis

provides an average output power of 1.5 W (19 mW cm2), and the energy harvesting technologies are presented in Section 3. In efficiency to an external load reaches 24%.26 Here, we fabricated Section 4, the key findings from the LCA and techno-economic the same structure using a copper electrode on both the rotator and analysis are discussed leading to the summary and final con- stator layers instead of a gold electrode, which was used as the stator clusions in Section 5. electrode in the previous work.26 In addition, other devices with the same structure fabricated based on copper instead of gold are used,25,35 and the average output power density of the fabricated 2. Fabrication route of a micro-grating device used in this study is the same as the previous one. Moreover, triboelectric nanogenerator (MG-TENG) thesmallvolume,lightweight,lowcost,andhighscalability characteristics make the TENG a suitable solution for harvesting To manufacture the TENG modules, roll-to-roll (R2R) processing is mechanical energy for both small-scale self-powered electronics and used. R2R processing is a cheap and fast substrate-based manufac- potentially in future larger scale energy generation. turing process,42,43 which can build structures in a continuous Given the potential of TENGs for low cost energy generation manner and has become an important manufacturing technology for self-powered applications, it is important to assess their for a wide range of new environmentally friendly and energy-efficient environmental profile and carbon footprint by carrying out a products. Roller-based R2R lines consist of a series of sequential detailed lifecycle assessment (LCA). This will provide an indication processing steps which begin by feeding input materials and as to whether they constitute new environmental challenges or not. culminate in winding of the finished material. It is often chosen A great deal of work has been published on the LCA of energy because it can make a sheet or roll at high volume and relatively low harvesting technologies. However, to the best of our knowledge, cost, a desired attribute for the concepts discussed inthispaper.In other than the comparative LCA of lead zirconate titanate (PZT) vs. addition, it is used globally to fabricate high volume commercial potassium sodium niobate (KNN), both potential materials for products such as flexible electronics, chemical separation piezoelectric energy harvesters,36 no LCA work currently exists on membranes and multilayer capacitors.44 mechanical energy harvesters such as the TENG. Given the limited Fig. 1(A–C) shows the architectures of Modules A and B, environmental information on TENGs, LCA is undertaken within which were assembled with series connections. Module A17 was the context of other energy harvesting technologies. LCA involves developed using a new type of electricity-generation method that the evaluation of the complete environmental impact of a material takes advantage of triboelectrification, a universal phenomenon or product from the raw material extraction phase, through the created upon contact between two materials. Based on polymer processing as well as the usage phases, and the final disposal.37 It is thin films that have complementary linear electrode arrays, the an important technique that should be adopted to highlight the MG-TENG (Module A) effectively produces electricity that is environmental hotspots in the production of consumer goods and sufficient for powering regular electronics as the two contacting their global environmental impact.38 The use of LCA, therefore, surfaces slide with respect to each other. The shape-adaptive defines and addresses environmental sustainability issues that are design of Module A suggests that it may be ideal for harvesting essential for future development and upscaling. Significantly per- energy from a wide variety of mechanical motions. Given its high Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. haps, it steers us clear of paths that will create new environmental electric output power and other significant advantages in terms problems while providing the necessary information with respect to of weight, volume, cost, scalability and adaptability, Module A is the consequences of material or device substitution. a practically promising approach in harvesting mechanical We live in a world dominated by networked product supply motions for self-powered electronics. chains, complex production technologies, and nonlinear con- Module B was developed with a new type of planar-structured, sumption patterns.39,40 It is essential therefore, for consumers, electricity-generation method to convert mechanical energy industries and policy makers, to have the right information in using the triboelectrification effect. Based on a stator–rotator the course of evaluating the environmental consequences of structure that has arrays of micron sized radial sectors, Module B substitute materials (from extraction, design and fabrication produces output power sufficient for conventional consumer processes to usage).36 To date, a detailed cost estimation and electronics. It also has the potential to harvest energy from a techno-economic evaluation and analysis of TENG modules has variety of types of ambient energy from motions such as air flow, not been carried out. Such an evaluation is vital regarding the water flow and even body motion. The fabrication of Module B future of TENG technology due to the urgent need to build a requires a series of finely controlled processes and production of cost-efficient industry that can survive with minimal government patterns with lasers, while DC sputtering is used to produce Cu intervention.41 Accordingly, the power conversion efficiencies electrodes. The high precision of the fabrication processes may and the ensuing financial costs of two TENG module designs result, however, in a prohibitively high manufacturing cost. were analyzed and compared with existing energy harvesting The main functional differences between Modules A and B technologies. are their mode of operation, performance efficiency and In light of the above, the rest of the paper is structured as potential applications. Whereas Module A operates in a sliding follows. In Section 2, a brief description of the fabrication processes free standing mode, Module B operates in a rotating free of both TENG modules under consideration is presented. Details of standing mode. The performance efficiency of A was experi- the overall methodological LCA principles and the techno-economic mentally determined to be 50% with a resulting power output of framework for comparative cost-benefit analysis with existing 500 W m2 and an area of 60 cm2 (see Table 1). For Module B,

This journal is © The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10,653--671| 655 View Article Online

Analysis Energy & Environmental Science Published on 22 February 2017. Downloaded 29/03/2017 04:57:27.

Fig. 1 (A) Structural design of TENG Module A; (B) structural design of TENG Module B and (C) fabrication steps for both TENG modules.

the calculated conversion efficiency is 24% (78.95 cm2), with a bodies, wind and body motion under ambient conditions. On corresponding power output of 190 W m2 (see Table 1). In the other hand, Module A boasts higher conversion efficiency terms of their applications, Module B offers more robust and compared to Module B, but offers less practical applications reliable applications regarding energy harvesting from water compared to TENG B.17,26

656 | Energy Environ. Sci., 2017, 10,653--671 This journal is © The Royal Society of Chemistry 2017 View Article Online

Energy & Environmental Science Analysis

Table 1 Differences between two TENG modules The goal of this study is to assess the potential life cycle impacts of two TENG modules (A and B). The overall assessment includes Module parameters Module A Module B five main steps: (i) gaining an understanding of the TENG 2 2 ‘‘TENG’’ module size 60 cm 78.95 cm technology in terms of raw material requirements, and production Distance between TENG unit 1 cm 1 cm Module efficiency 50% 24% and fabrication processes of the modules; (ii) characterization of Power output of one piece of module W 3 1.5 the system (i.e., establishing systems boundaries, the functional 2 Power output of one piece of TENG W m 500 190 unit, modular components, material composition, operational efficiencies, etc.); (iii) construction of the system inventory (i.e., 3. Materials and methods input requirements (physical units), process flow, energy flow, material flow, and reference flow); (iv) overall impact assessment In the preceding sections, the phenomenon of triboelectricity as and environmental profile evaluations across multiple sustain- a potential effect for energy harvesting is highlighted. Against ability metrics; and (v) performance evaluation and techno- this backdrop, a detailed environmental profile evaluation and economic analysis. techno-economic analysis of TENG modules are carried out In this work, the functional unit is set as 1 m2 of the TENG based on the framework schematically illustrated in Fig. 2. module and all of the inventories generated are converted by aligning them to conform to the functional unit based on the 3.1 Life cycle analysis framework defined system boundaries, as schematically illustrated in LCA can be used as a decision-making tool for the systematic Fig. 3. tracking of a wide spectrum of environmental impacts across the The TENG module is assembled by depositing the components entire value chain of the development of a product,45 identifying onto the substrate. The manufacturing process consumes energy baskets of interventions for reducing the environmental impact and produces emissions. After the TENG module is utilized without burden shifting.38,46 LCA entails the gathering and and decommissioned, the waste modules are landfilled in the evaluation of the inputs, outputs, and potential environmental disposal stage. Disposal mechanisms including incineration impacts of a product system throughout its lifespan and involves and waste recycling are not taken into consideration within the four key steps namely: (i) goal and scope definitions, where system boundaries drawn due to the dearth of data regarding questions such as what, how and why regarding the LCA work combustion processes or waste recycling for TENG modules. are asked and where the systems boundaries and functional unit Modular use phases and transportation are also excluded are set; (ii) inventory analysis, where input and output data of from the system boundaries in line with assumptions made each process in the life cycle are collated, adding them across the in a number of LCA studies for energy harvesting technologies entire system; (iii) evaluation of the environmental effects, such as photovoltaics.49–51 Although input–output data can be detailing LCA results through classification and characterization augmented with process-based data within a hybrid LCA frame- for comparative analysis; (iv) the interpretation of the inventory work36 to complete the system boundaries based on missing and impact assessment of results and the identification of issues data, such an approach is not considered in the current work. Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. that are of significant importance.37,47,48 The balance of system (BoS) is omitted as part of the overall

Fig. 2 Schematic representation of the overall framework for life cycle assessment (LCA) and techno-economic analysis (TEA) of TENG modules.

This journal is © The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10,653--671| 657 View Article Online

Analysis Energy & Environmental Science

Fig. 3 System boundary considered in the LCA, showing the material composition and energy flows associated with the fabrication steps captured within the inventory.

system boundaries to ensure fair comparison with those of Table 2 Material inventory of 1 m2 of TENG Module A with 90% active other energy harvesting technologies. area 3.1.1 Life cycle inventory. The construction of the life cycle Raw materials Mass (kg) Usage Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. inventory (LCI) is central to any LCA work. Based on the system Substrate patterning boundaries described in Fig. 3 we classified the LCI of each Acrylic sheet E 1.18 10 Substrate module into two categories, namely material inventory and Ethanol 1.00 105 Substrate cleaning solvent 5 energy inventory. A material inventory table consists of the Deionized water 1.00 10 Substrate cleaning solvent Acetone 1.00 105 Substrate cleaning solvent mass of raw materials, direct emission during manufacturing, and disposal materials per functional unit of the module. In Grating patterning 1 this analysis, the focus is on two TENG modules (i.e., Module A PTFE film E 1.10 10 3*Layer thickness 25 mm Ethanol 6.00 105 Grating cleaning solvent and Module B), used as representative solutions for TENG Deionized water 6.00 105 Grating cleaning solvent modules. The major differences between the two modules are Acetone 6.00 105 Grating cleaning solvent listed in Table 1. 2 Electrode deposition The material inventory of a 1 m functional unit of Module A Copper ETH U 2.24 102 5*Layer thickness 500 nm is shown in Table 2. The active area ratio and the module Titanium I 4.43 104 5*Layer thickness 20 nm efficiency are 90% and 50%, respectively.17 The masses of the Electrode wires cleaning solvents, PTFE, and acrylic are obtained from the Lead ETH U 2.30 103 Wire diameter 0.0100, length 4 m literature.17,26 The masses of the electrode layer copper and titanium are derived based on the thickness of the corres- Direct emission Ethanol 7.00 105 Cleaning solvent ponding layers, the active area ratio of the module, and the Acetone 7.00 105 Cleaning solvent material utilization efficiency. Since the material utilization Disposal materials 1.32 10 To landfill efficiencies are not reported for TENG modules, we assume that the material utilization efficiencies for laser cutting and sputtering are 30% and 75%, respectively. The mass of direct The energy inventory of 1 m2 of the TENG Module A is shown emission is determined as the mass of the cleaning solvents of in Table 3. As shown, all the operations are performed using ethanol, acetone and deionized water. electric equipment. Therefore, energy consumption is evaluated

658 | Energy Environ. Sci., 2017, 10,653--671 This journal is © The Royal Society of Chemistry 2017 View Article Online

Energy & Environmental Science Analysis

Table 3 Energy consumption for manufacturing 1 m2 of TENG Module A Table 5 Energy consumption for manufacturing 1 m2 of Module B with with 90% active area 78% active area

Power (W) Time (S) Electricity (MJ) Power (W) Time (s) Electricity (MJ) Substrate cutting Substrate cutting Laser cutter machine 1.50 103 8.50 101 1.28 101 Laser cutter machine 1.50 103 3.50 103 5.25 10

Grating patterning Grating patterning Laser cutter machine 1.50 103 2.55 102 3.83 101 Laser cutter machine 1.50 103 2.00 101 3.00 102 Drilling 2.20 103 2.00 101 4.40 102 2 1 3 Electrode deposition Air/O2 plasma 1.00 10 6.00 10 6.00 10 Titanium coating/sputtering 1.50 103 4.00 102 6.00 101 Copper coating/sputtering 1.50 103 2.00 103 3.00 10 Electrode deposition Titanium coating 1.50 103 5.00 102 7.50 101 Total 1.14 kW h Copper coating 1.50 103 5.00 103 7.5 10 1.36 101

Total 3.78 kW h by multiplying equipment power by corresponding operating time. We apply the same energy consumption as that evaluated by 50 Espinosa et al. The total electricity consumption for manufacturing the materials used for the fabrication of the TENG modules 2 1m oftheTENGmoduleis1.14kWh.Wetranslatetheelectricity over their life cycle. Each entry of life cycle inventory developed consumptioninmanufacturingthe TENG modules to the equivalent for this work is matched with an appropriate unit process in primary energy consumption assuming that the electricity conformity with the functional unit. Using life cycle inventories, 52 applies to the average electricity mix in the US. The end-of- the environmental impacts are calculated as follows:36,55 life primary energy consumption accounts for the energy usage XI involved in landfilling the waste modules. Bj ¼ bj;i xi j ¼ 1; 2; ...; J (1) Tables 4 and 5 summarize the material and the energy i¼1 inventories of 1 m2 of Module B, respectively. The mass of Module 26 B is evaluated from the data reported in the literature. XJ 3.1.2 Life cycle impact assessment modelling. The overall Ek ¼ ek; j Bj k ¼ 1; 2; ...; K (2) impact assessment based on the LCI above is performed following j¼1 the guidelines provided in the International Organization where b is the environmental burden j per unit activity i, with 53 54 j,i for Standardization (ISO) 14040 and 14044. This allows for burdens constituting raw materials and energy consumed the appropriate data management of life cycle inventory and within the system and emissions to air, land and water. These assessment of environmental impacts stemming from each of parameters are obtained from LCA software and databases such 56 as SimaPro and Ecoinvent. xi is the mass or energy flow Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. associated with unit activity i. e is the relative contribution Table 4 Material inventory of 1 m2 of TENG Module B with 78% active k,j area of the total burden Bj to impact Ek as defined by the CML 2001 method.57 Raw materials Mass (kg) Usage The overall focus of the current work is on global warming Substrate patterning potential (GWP). However, the need to consider multiple sustain- Acrylic sheet E 2.27 10 Substrate 5 ability metrics when analyzing the environmental profile of a Ethanol 3.00 10 Substrate cleaning solvent 36 Deionized water 3.00 105 Substrate cleaning solvent product or process was demonstrated by Ibn-Mohammed et al. Acetone 3.00 105 Substrate cleaning solvent This will, for environmental trade-oﬀ analysis, ensure that green- house gas (GHG) emissions are not minimized at the expense of Grating patterning FEP film E 7.05 102 Layer thickness 25 mm other indicators including human toxicity, acidification, eutro- Ethanol 3.00 105 Cleaning solvent phication, material use, fossil fuel and ozone layer depletion. Deionized water 3.00 105 Cleaning solvent 5 Acetone 3.00 10 Cleaning solvent 3.2 Techno-economic evaluation of TENG modules Adhesive 2.50 104 Epoxy resin 3.2.1 Module cost estimation. To assess the cost of fabricat- Electrode deposition ing the modules, we assumed the production capacity of both Copper 5.374 103 Layer thickness 200 nm & layer thickness 100 nm routes to be 100 MW per year. As shown in Fig. 4, the module cost Titanium 8.86 105 Layer thickness 10 nm consists of the capital, the materials and the overhead cost. The capital cost is calculated based on the depreciation of capital Electrode wires Lead 2.30 103 Wire diameter 0.0100, length 4 m investment (CI). Given that the complete process of Module A was based on the fabrication steps in Fig. 1, the CI is taken to be $7 Direct emission million for a production capacity of 100 MW (see Tables S1 and Ethanol 6.00 105 Cleaning solvent Acetone 6.00 105 Cleaning solvent S2 in the ESI†). Module B has an eﬃciency of 24% which is lower Disposal materials 2.35 10 To landfill than that of Module A (50%), as such, the capital investment for

This journal is © The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10,653--671| 659 View Article Online

Analysis Energy & Environmental Science

Fig. 4 Cost parameters considered for the techno-economic analysis of TENG modules detailing the relevant materials, overhead costs, capital costs and levelized cost.

Module B (CI Module B) for a 100 MW capacity per year is where i is the annual interest and n is the 5-year equipment estimated to be $14 million per year (see Tables S1 and S3, ESI†). lifetime. An annual interest of 5% is assumed for 2020, based Details of how the cost estimates are carried out are presented in on current low global interest rates. The costs of materials for Tables S2 and S3 of the ESI.† Modules A and B are estimated to be US$ per 0.617 W and US$ Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. The depreciation of the facility results in a decrease of the per 2.56 W, respectively, based on the ratio of investment in capital investment from year to year according to eqn (3):41 materials to the output power with a material usage of 80%. The overhead costs consist of labor, renting facilities and (CI) =CI bn (3) n utilities. The labor cost of US$ 0.0304 per W was estimated where n is the number of years after construction and b is the based on the flexible electronics industry average (see Table S7, depreciation ratio, which is assumed to be 0.5 based on ESI† for details). Based on a similar industry of DSCs and thin- information from the nascent industry of TENG developers. films, silicon solar cell manufacturing lines, the rents for Depreciation of an investment should cease when bn o 0.1. Modules A and B are estimated to be US$ 0.00792 per W After four years, there was no further depreciation of the and US$ 0.022 per W, respectively. The costs of utilities for investment because (0.5)4 = 0.063. The capital costs of Modules Modules A and B are estimated to be US$ 0.00792 per W and A and B are based on the ratio of capital investment to the US$ 0.022 per W, respectively. After adding 1% of the capital output power, which changed from US$ 0.07 per W to US$ costs for maintenance fees (US$ 0.0016 million per year and 0.004375 per W and from US$ 0.14 per W to US$ 0.00875 per W, US$ 0.0016 million per year for Modules A and B (Table S8, respectively, during the first five years (see Tables S4 and S5 in ESI†)), the overhead costs of Modules A and B are calculated the ESI† for details). The module cost is calculated by summing to be US$ 0.04784 per W and US$ 0.075 per W, respectively the capital amortization, materials, and overhead costs. The (Table S8, ESI†). capital amortization costs for Modules A and B are taken to be The resultant module costs, calculated based on our US$ 0.016 per W and US$ 0.032 per W, respectively, based on assumptions, are US$ 0.68084 per W and US$ 2.667 per W for the annual worth of CI (1.6 million USD for Module A and 3.2 Modules A and B, respectively (Table S9, ESI†). These are the million USD for Module B); they were equated to: baseline values used in the sensitivity analysis (Section 4.5.2). Estimations of the levelized cost of electricity are based on the i ð1 þ iÞn CI (4) total cost of a solar cell system, including the costs of the ð1 þ iÞn1 module, balance of systems (BoS), land, support structures,

660 | Energy Environ. Sci., 2017, 10,653--671 This journal is © The Royal Society of Chemistry 2017 View Article Online

Energy & Environmental Science Analysis

wiring, power conditioning and installation,41 and summed 4. Results and discussion 58,59 according to eqn (5): 4.1 Primary energy consumption and carbon footprint Primary energy demand and, correspondingly, the carbon ICC 1000CRF LCOE ¼ þ 0&M (5) footprint due to the fabrication of both TENG modules is the CF 8760 focus of the current LCA work with a view to identify hotspots in the entire supply chain of these modules. Based on the con- where ICC is the Installed Capacity Cost ($ per W DC) = BoS cost + structed LCIs in Tables 2–5, the primary energy consumption module cost; CRF is the Capital Recovery Factor, expressed as: and the corresponding carbon footprint distributions for TENG modules A and B are evaluated and depicted in Fig. 5 and 6, respectively. As indicated in Fig. 5, about 90% of primary energy ið1 þ iÞn CRF ¼ (6) consumed in both modules is attributed to raw material require- ð1 þ iÞn 1 ments. A disaggregation of the material embodied energy high- lights the key differences between the TENG modules. For where i is the discount rate and n is the useful lifetime (i.e., instance,inModuleA,acrylic(78.18%) and polytetrafluoroethylene, lifetime of the system), CF is the alternating current capacity PTFE (20.48%), are the major contributors to the material embodied factor, calculated as 0.8 the renewable energy source (i.e., energy. Similarly, the distribution of embodied material energy is wind energy/8760 hours). This factor is reduced by 37% due to dominated by acrylic, fluorinated ethylene propylene (FEP film), the losses in the conversion process from direct current to and copper with each contributing shares of 96.88%, 2.87% and alternating current. O&M is the operation and maintenance 0.25%, respectively. As indicated in Tables 2 and 4, the quantities cost expressed in $ per kW h. of acrylic in the materials composition of both modules A and B The following assumptions are made. BoS is US$ 75 per m2 are 1.18 kg and 2.27 kg, explaining their dominance in the total based on the projected long term goal of silicon based solar mass of the modules (78.18% for Module A and 96.88% for cells in 2020.41 BoS costs at an eﬃciency of 50% and 40% for Module B). Module A are US$ 0.15 per W and US$ 0.1875 per W, respec- In terms of electrical energy consumption (also expressed in tively. For Module B, with eﬃciencies of 24% and 20%, the MJ m2 to conform to the unit of material embodied energy), corresponding costs are US$ 0.394 per W and US$ 0.474 per W, electrode deposition of copper coating/sputtering consumed respectively. By using BoS cost = 75 US$ per m2; O&M = the largest amount of energy (B73%) due to the length of time $0.001 per kW h; i = 5%, and n = 20 (no tax credits and no associated with carrying out such a process during the fabrication accelerated depreciation), from these values, CRF (i =5%,n = 15) = of Module A. Electrical energies consumed by sputtering for 0.1. In order to derive the energy produced per year due to 1 W of titanium coating deposition and the laser cutting machine installed TENGs, a CF of 37% is assumed. constitute roughly 15% and 13%, respectively. Overall, electrode Published on 22 February 2017. Downloaded 29/03/2017 04:57:27.

Fig. 5 Distributions of the primary energy consumption for fabricating two TENG modules.

This journal is © The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10,653--671| 661 View Article Online

Analysis Energy & Environmental Science

Fig. 6 Distributions of the carbon footprint of TENG Modules A and B.

sputtering consumes B85% of the electrical energy for the fabrica- environmental impact metrics are normalized to 100% with the tion of Module A. Adoption of alternative deposition techniques for view that the sum of the impact of each of the contributing the copper and titanium coating would go a long way in minimizing processes or materials is 100%. As indicated in Fig. 7, the acrylic the overall electrical energy consumption. As for Module B, the is the most significant contributor for carcinogens (82%), increased number of operations involved in its fabrication results in respiratory organics (85%), respiratory inorganics (73%), climate higher electrical energy consumption compared to Module A. As change (74%), acidification/eutrophication (76%), fossil fuels with Module A, sputtering of titanium and copper consumes B62% (81%), and ecotoxicity (33%). Although the intensity of material

of the electrical energy and the laser machining and associated embodied energy and CO2-eq of copper, lead and titanium are drilling activities consume 38%. Sputtering as a means of depositing numerically higher than that of acrylic, given that the quantity of thin films of the metals in the modules guarantees high quality but acrylic in the material composition is the largest, its overall Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. comesattheexpenseofhighcost.60 Overall, Module B consumes impact across the aforementioned impact categories outweighs more electrical energy during fabrication compared to Module A. other materials. Sputtering due to electrical energy consumption Fig. 6 shows the distribution of carbon footprint from which also has a great influence on radiation (96%), the ozone layer the major contributors of the substrate, the copper electrode, (83%), and land use (83%). The use of acrylic, however, oﬀers an sputtering and laser cutting can be established. The distribution advantage in the fabrication of the modules. For instance, acrylic of primary energy consumption during fabrication indicates has very good structural properties such as lightweight, ease of similar patterns to the carbon footprint because diﬀerent fabri- fabrication, impact resistance and ability to withstand poor cation operations consume only electricity and their conversion weather conditions. Its high strength and durability are also

to carbon dioxide equivalent (CO2-eq) is based on appropriate important advantages. Furthermore, acrylic sheets are fabricated characterization factors in the evaluation process. Not only do using fabrication processes in facilities that are certified by the distributions of primary energy consumption and the carbon ISO-14001. More importantly, the scenario of their end of life footprint exhibit similar patterns, but also the impact of other is environmentally viable given their recyclability and reuse categories remains identical, provided that the steps involved in potential. Additionally, compared to other plastics that produce the fabrication process remain constant. A resemblance can be toxic gases that are harmful to humans and to the environment found between the distributions of the material embedded during combustion processes, acrylic does not pose such threats primary energy consumption and the carbon footprint, which due to its stability during exposure to ultraviolet radiation. suggests similar strategies for optimizing both modules for AsshowninFig.8,forTENGModuleB,thepresenceofacrylicas improved environmental performance should be adopted. in Module A also constitutes the major influence across a number of indicators. For instance, acrylic is the most significant contributor 4.2 Environmental profile assessment of contributing for carcinogens (83.1%), respiratory organics (B92%), respiratory components of TENG Modules A and B across multiple indicators inorganics (80.8%), climate change (81%), acidification/eutrophica- Fig. 7 and 8 show the environmental profiles of 1 m2 of TENG tion (79.5%) and fossil fuels (88%). The reason for this is similar to Module A and 1 m2 of TENG Module B, respectively. All 11 that of Module A (i.e., the quantity of acrylic used dominates those of

662 | Energy Environ. Sci., 2017, 10,653--671 This journal is © The Royal Society of Chemistry 2017 View Article Online

Energy & Environmental Science Analysis

Fig. 7 Environmental profile of 1 m2 of TENG Module A. Published on 22 February 2017. Downloaded 29/03/2017 04:57:27.

Fig. 8 Environmental profile of 1 m2 of TENG Module B.

other materials in the structure). Sputtering due to electrical energy Comparative life cycle impact assessment results between consumption also has a great influence on radiation (88%), the the two TENG modules are depicted in Fig. 9–11. Module A is ozone layer (70%), and land use (70%). used as the standard for normalization. In Fig. 9, Module A

This journal is © The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10,653--671| 663 View Article Online

Analysis Energy & Environmental Science

Fig. 9 Comparison per damage category, by summation of individual impacts, the higher impact set equal to 100, using Eco-indicator 99 Europe E/E methodology.

performs better environmentally than Module B except in one impact category, minerals. This is attributed to the higher Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. quantities and triple layer thickness of polytetrafluoroethylene (PTFE) used in Module A compared to the single layer thickness of fluorinated ethylene propylene (FEP film E) used in Module B. PTFE is generated through polymerization of tetrafluoro- ethylene using free radicals, and hence has high mineral resource requirements. The uniformity of its material structure (i.e., PTFE), its excellent chemical, electrical and physical properties, its tightly controlled thickness and its inherent capabilities to serve as a semi-permeable membrane render it applicable for TENG and biomedical applications.61 On the other hand, the compatibility of FEP with various chemicals, its reliable electrical properties, its mechanical toughness anditsbroadthermalrangemakeitsuitableforTENG Fig. 10 Endpoint comparison after weighing, using Eco-indicator 99 applications.62 Europe E/E methodology. Fig. 10 displays proportions between impacts of the two types of TENG modules with respect to Eco-indicator 99 under human health, resources and ecosystem quality. As shown, Module B results in more damage compared to Module A. 4.3 Comparison with existing energy harvesting technologies Single score comparison by impact category based on Eco-indicator 4.3.1 Eco-indicator. Eco-indicator 99 results, across eco- 99 is depicted in Fig. 11, where the environmental impact of system quality, resources and human health for eight variants Module B also surpasses that of Module A. For further comparative of energy harvesting technologies, notably PV modules, are results of the environmental profile of TENG modules, we refer compared with the TENG modules as depicted in Fig. 12. In readers to the ESI.† all three damage categories, both TENG modules achieve the

664 | Energy Environ. Sci., 2017, 10,653--671 This journal is © The Royal Society of Chemistry 2017 View Article Online

Energy & Environmental Science Analysis

Fig. 11 Single score comparison by impact category, using Eco-indicator 99 Europe E/E methodology.

lowest points and are one order of magnitude lower than those technologies, organic solar cells and perovskite solar cells). The

of c-Si, a-Si, ribbon-Si, CdTe, CIS, OPV TiO2 and ZnO PV EPBP is given by: modules. This clearly demonstrates the overall environmental Embodied energy kW h m2 edge of the TENG modules when compared to PV technologies. EPBP ¼ (7) Energy outputðÞ kW h m2 year1 Therefore, a more environmentally sustainable energy harvesting technology could potentially be developed based on TENG The result of the comparison is shown in Fig. 13. As shown, modules, although this requires switching to greener sub- Module A has a shorter nominal EPBP than the other technologies strates and reducing the consumption of organic solvents as at 0.05 years. Module B also has a shorter EPBP compared to well as the use of eﬃcient fabrication processes. traditional PV technologies but higher than those of organic and 4.3.2 Energy payback period. In this section, the energy perovskite solar cells. The reason for TENGs outperforming payback periods (EPBPs) of Modules A and B are compared with silicon and CdTe based PV cells is because their fabrication existing PV technologies (i.e., silicon technologies, thin-film does not have high energy intensity requirements associated Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. with silicon or rare element purification and processing that causes a higher environmental impact.64 This is largely due to the eﬃcient fabrication routes based on R2R processing. It is important to note that the EPBP of Module B is higher than those of OPV and perovskite solar cells, attributed to its lower

Fig. 12 Eco-indicator 99 results for 1 m2 of each module. The data for c-Si, a-Si, ribbon-Si, CdTe, and CIS are extracted from the study of Laleman et al.63 The data for OPV are extracted from the study of Espinosa Fig. 13 Comparison of energy payback time for 7 PV modules with TENG 50 et al. The data for the TiO2 perovskite module and the ZnO perovskite modules. The data for the energy payback period of all the PV modules module were based on the work of Gong et al.64 were based on the work of Gong et al.64

This journal is © The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10,653--671| 665 View Article Online

Analysis Energy & Environmental Science

output energy compared to the aforementioned technologies. demonstrate a wide range, with the highest bars representing Nevertheless, in the future, by leveraging optimal and eﬃcient the values of the highest probabilities. The asymmetric profile processing technologies, the EPBP of TENGs can be further of both distributions results from the nonlinear relationship reduced significantly. Overall, the favorable environmental pro- between the input parameters and the sustainability indicators. file and EPBP of the TENG modules compared to the traditional The simulation results are shown in Fig. 15 and 16. The single energy harvesting technologies suggest that in the future, they cores in both cases are comparatively robust when the key speci- will challenge the existing technologies, whilst contributing fications of the modules are subject to uncertainty. The low single immensely towards addressing global energy problems. core points for the entire 95% confidence regions demonstrate that

4.3.3 CO2 emission factor. The CO2 emissions factor (CEF) TENGs are already environmentally competitive. is given by: A sensitivity analysis is also conducted to estimate how the environmental performance of Modules A and B alter if the Carbon footprint ðkg CO2 eqÞ CEF ¼ (8) consumption of materials and energy during manufacturing is Energy output across the lifespanðÞ kW h varied, given the dominating influence of some input para- To apply eqn (8), the lifespan of the TENG system under meters across all the considered impact categories. For each consideration should be established. The lifespan of other existing parameter, two scenarios were modeled and then compared with PV technologies is already well-established. Likewise, assumptions the baseline, i.e.,a10% variations in the total consumption. have been made about the lifespan of perovskite solar cells. Given that TENGs are still in their infancy, no exact value in terms of lifespan has yet been reported for them. Fig. 14 shows the compar-

ison of CO2 emission factors for existing energy harvesting techno-

logies to TENG modules. As indicated, the CO2 emission factor for TENG Module B is higher, similar to that of CdTe, Ribbon-Si and P-Si (TENG Module A shows a significantly lower CEF). This suggests

that currently the associated cost of CO2 is currently high due to their shorter lifespan (assumed to be 2 years). In the future, it is expected that the lifespan of TENGs will increase considerably due to advancement in material optimization, thereby lowering their CEF. These results deliver an important message for the development of otherenergyharvestingdevicessuchasTENGsaspotentialenvir- onmentally viable energy harvesters. The TENG is the youngest among the energy harvester technologies with enormous potential for better manufacturing processes with improved eﬃciency, more stable performance, and a longer operational lifetime. Published on 22 February 2017. Downloaded 29/03/2017 04:57:27.

4.4 Sensitivity analysis Fig. 15 Probability distributions for the single core impact category of The probability distributions of the two forecasts for the TENG Module A. modules are shown in Fig. 15 and 16. Both distributions

Fig. 14 CO2 emission factor for selected PV modules and 2 TENG

modules A and B. The data for the CO2 emission factor of all the PV Fig. 16 Probability distributions for the single core impact category of modules were based on the work of Gong et al.64 Module B.

Energy & Environmental Science Analysis

Table 6 Sensitivity analysis results for Module A

Resp. Resp. Climate Ozone Acidification/ Land Fossil Input Variation (%) Carcinogens Organics Inorganics change Radiation layer Ecotoxicity eutrophication use Minerals fuels Acrylic 10 +8.2 +8.4 +7.4 +7.5 0.0 0.0 +3.3 +7.5 0.0 0.0 +8.1 +10 8.2 8.4 7.4 7.5 0.0 0.0 3.3 7.5 0.0 0.0 8.1 PTFE 10 0.0 +0.7 +0.5 +0.5 0.0 0.0 +0.9 +0.5 0.0 0.0 +0.5 +10 0.0 0.7 0.5 0.5 0.0 0.0 0.9 0.5 0.0 0.0 0.5 Electrode 10 +0.3 +0.1 +0.3 +0.1 +0.3 +0.9 +1.6 +0.2 +0.4 +9.7 +0.1 deposition +10 0.3 0.1 0.3 0.1 0.3 0.9 1.6 0.2 0.4 9.7 0.1 Manufacturing 10 +1.4 +0.8 +1.8 +1.9 +9.7 +9.1 +2.8 +1.8 +9.6 +0.2 +1.3 +10 1.4 0.8 1.8 1.9 9.7 9.1 2.8 1.8 9.6 0.2 1.3

Table 7 Sensitivity analysis results for Module B

Resp. Resp. Climate Ozone Acidification/ Land Fossil Input Variation (%) Carcinogens Organics Inorganics change Radiation layer Ecotoxicity eutrophication use Minerals fuels Acrylic 10 +8.5 +9.1 +8.0 +8.3 0.0 0.0 +3.6 +8.1 0.0 0.0 +8.9 +10 8.5 9.1 8.0 8.3 0.0 0.0 3.6 8.1 0.0 0.0 8.9 FEP 10 0.0 +0.3 +0.2 +0.2 0.0 0.0 +0.3 +0.2 0.0 0.0 +0.2 +10 0.0 0.3 0.2 0.2 0.0 0.0 0.3 0.2 0.0 0.0 0.2 Electrode 10 0.0 0.0 0.0 0.0 0.0 +0.1 0.2 0.0 0.0 +4.8 0.0 deposition +10 0.0 0.0 0.0 0.0 0.0 0.1 0.2 0.0 0.0 4.8 0.0 Manufacturing 10 +1.4 +0.7 +1.7 +1.5 +10.0 +9.9 +5.1 +1.7 +10 +4.9 +0.9 +10 1.4 0.7 1.7 1.5 10.0 9.9 5.1 1.7 10 4.9 0.9

As shown in Tables 6 and 7, the variation of the acrylic consumption has the highest influence on the Resp. organics. For instance, a 10% variation of cement consumption leads to 8.4% and 9.1% changes in the Resp. organics impact in Modules A and B, respectively. As expected, the mineral impact is most sensitive to the variation of electrode deposition consumption, and a 10% decrease of electrode deposition leads to 9.7% and 4.8% corresponding decreases of this indicator for Modules A and B. The fluctuation of manufacturing during construction and operation leads to the largest value change of Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. radiation and about 9.7% and 10% variation occurs for Modules A and B, respectively, if the former changes by 10%.

4.5 Techno-economic analysis

4.5.1 Estimation of costs of TENG modules. Fig. 17 shows Fig. 17 Calculated module costs of TENGs for the first year, the fifth year the cost of Modules A and B in the 1st and 5th year and and amortizing over 5 years by taking depreciation and amortizing capital amortization capital cost over 5 years. The module cost can cost into consideration. The depreciation rate was 50% per year and the be divided into the materials, the overhead, and the capital capital cost was assumed to remain constant after the five-year period. cost. The capital costs for Modules A and B are calculated based on the capital costs of TENGs fabricated using the data in After that time, the contribution of the capital cost to the total Tables S2 and S3 (ESI†), respectively. The cost of the materials cost is lowered, so that the module cost is determined mainly by is estimated based on the amount used. The overhead cost is the overhead and the materials costs. Fig. 18 presents the estimated based on reasonable assumptions (see Table S8, distribution of the materials cost for TENG production routes. ESI†). The details of the calculations are shown in the Methods DSM layers represent device structural materials, D/EM repre- section and the ESI.† The relatively high module cost in the first sents electrode dielectric materials and LW represents electrode year is due to the high depreciation rate (50%) of the capital wire. Other materials costs in Fig. 17 include the expense of investment. The calculated capital costs in the first year are 0.07 Ti/CU deposition. The total calculated cost of materials for and 0.14 US$ per W for Modules A and B, respectively. The initial Module A is 0.617 US$ per W which is lower than the cost for capital cost of Module A is lower because the capital investment Module B that is 2.56 US$ per W (Table S6, ESI†). The higher cost associated with the use of large eﬃciency is higher than that in of materials for Module B is because both the output power and Module B. However, the capital cost rapidly decreased because of eﬃciency are higher in Module A than in Module B. depreciation and there is a monotonic decrease of the total Based on thin film silicon solar cell production,65,66 Module module cost during the first 5 years (Tables S4 and S5, ESI†). A and Module B have total overhead costs of US$ 0.04784 per W

Analysis Energy & Environmental Science

Fig. 18 Cost of material distribution for Module A (left) and Module B (right). The values of material costs are assumed by the real amount of material used in both structures and the wholesale price. The 80% material usage ratio has been considered.

and US$ 0.075 per W, respectively (see Table S8, ESI†). There- line corresponds to the efficiency of the present research status. fore, the total production costs of Modules A and B are similar. The efficiency of Module A is assumed to be 20–50% based on In order to compare the costs between different energy harvest- a current device efficiency of 40–50%. The corresponding ing technologies and to calculate the costs for electricity gen- estimated module cost is 0.8308–0.86834 US$ per W. And the eration and amortization, module costs are also calculated by efficiency is assumed to be 15–24% based on a current device amortizing total capital cost by the lifetime of the devices. The efficiency of 20–24% for Module B. The calculated module cost results show that Module A’s amortization cost is US$ 0.68084 is 4.731–4.811 US$ per W. If we further extend the solid line, the whereas Module B’s amortization cost is US$ 2.667, as shown in module costs of Module B decrease dramatically while Module Fig. 18. These results are used in the sensitivity analysis and the A decreases only slightly. This result reveals that the module estimation of the leveled cost of electricity to obtain an estimate efficiency acts as an important factor for module cost no matter of the cost of electricity generation. which route is used for the manufacturing process. Improve- 4.5.2 Sensitivity analysis of module cost. It is noteworthy ment of the TENG efficiency and active area, by upgrading the that these cost estimates are based on assumptions about the precision of deposition methods, will further increase the two kinds of TENG structures. However, the assumed para- module efficiency and therefore it will be an effective way to meters might vary when TENGs are commercialized. Hence, we reduce the cost of Module B. performed further sensitivity analyses to consider the effect of 4.5.3 Levelized cost of electricity produced with TENGs. TENGs on module costs. The module costs increase exponen- The LCOE is typically used to compare system costs of electricity tially as their module efficiency decreases (Fig. 19). The solid produced using different sources of energy. The LCOEs of Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. traditional energy sources are 7.04–11.90 US cents per kW h, and the costs of solar PV technologies are 9.78–19.33 US cents per kW h as reported in Levelized Cost and Levelized Avoided Cost of New Generation Resources in the Annual Energy Outlook 2015.67 The LCOE is calculated according to eqn (5) (Section 3), and the output is affected mainly by the module cost, efficiency, and lifetime. In our module cost analysis, both Modules A and B are estimated to produce TENG energy harvesting modules at a cost in the range of 0.68084–2.667 US$ per W. We calculate theLCOEofaTENGenergyharvestingmodulebyassuminga module cost of 0.68084 US$ per W for Module A and 2.667 US$ per W for Module B and a lifetime of 15 years. The LCOEs are 2.569 US cents per kW h and 2.681 US cents per kW h corresponding to module efficiencies of 50% and 20%, respectively for Module A. On the other hand, the costs are 9.198 US cents per kW h and 9.43 US cents per kW h corresponding to module efficiencies of 24% and 20%, respectively for Module B, which are lower than those of Fig. 19 Module cost of TENGs as a function of module efficiency. Except the traditional energy sources (Fig. 19) for Module A and in for the independent variables in these figures, the other parameters the same range of wind power for Module B. Details of the associated with Module A and Module B were fixed. The solid lines were calculated based on the range of reported efficiencies; the dashed lines calculations are shown in the Methods section and Table S10 are based on calculations assuming high module efficiencies that are (ESI†). This analysis indicates that the module efficiency has a expected but not yet achieved. significant influence on the LCOE.

Energy & Environmental Science Analysis

of the efficiency and the lifetime of TENGs are urgent tasks from the perspective of cost, and more efforts should be devoted to this field. The LCOE is calculated according to eqn (5) (Method part), and it is affected mostly by the module cost, efficiency, and lifetime. In our cost analysis, the module cost is estimated to be in the range of 0.68–2.667 US$ per W corresponding to TENG Module A and TENG Module B, respectively. We calculate the LCOE of TENG Module A by assuming a module cost of 0.68 US$ per W and a lifetime of 15 years. While, TENG Module B is calculated by assuming a module cost of 2.667 US$ per W and a lifetime of 15 years. The TENG Module A LCOEs are 2.569 US cents per kW h and 2.681 US cents per kW h for efficiencies of 50% and 40%, respectively. On the other hand, the TENG Module B LCOEs are 9.198 US cents per kW h and 9.43 US cents per kW h corresponding to module efficiencies of Fig. 20 The relationship between the LCOE and the lifetime. A system 24% and 20%, respectively, which are lower than other energy lifetime o10 years was not considered in our analysis. sources (Fig. 21). Details of the calculation are shown in the Methods section and Table S10 (ESI†). Consequently, module Fig. 20 shows the effect of lifetime on the LCOE of TENGs for efficiency has a significant influence on the LCOE. wave energy harvesting. The LCOEs estimate 50% and 40% efficiency for Module A, and 24% and 20% for Module B but each decreases exponentially with the extension of the system 5. Summary and concluding remarks lifetime in the range of 10–30 years. For high efficiency (50%) modules, a lifetime of 10 years can lead to an LCOE of 3.42 US Mechanical energy is available in abundant quantities every- cents per kW h. The low-efficiency (40%) modules require a where around us and is completely independent of weather, short lifetime (12 years) to achieve a similar LCOE. A conserva- day/night or even season. This abundant source of energy tive estimate of the discount rate of 5% is used above. Based on remains largely untapped but with continuous and improved the above analysis, the module efficiency and lifetime are the power conversion efficiencies reported in the past few years, most sensitive factors for the LCOE of TENGs. The ultra-low triboelectric nanogenerators (TENGs) have been touted as LCOE of TENGs is achieved to be 2.569–2.68 US cents per kW h highly promising sources of electricity generation from with 15 years of lifetime, surpassing the United States ‘‘Sun Shot mechanical energy. In this paper, a cradle-to-grave life cycle Initiative’’ target of 6.0 US cents per kW h. Hence, improvements assessment of two TENG modules is performed. The life cycle Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. environmental impact assessment involves 11 midpoint impact categories, and an endpoint evaluation by following the Eco- indicator 99 methodology. We shed light on two important sustainability indicators and find that TENG modules have the shortest EPBT among existing PV technologies. In addition, we find that the environmental hotspots come from the use of acrylic (both Modules A and B), PTFE (Module A) and FEP (Module B). As such, for future development of this technology, material optimization should be advanced. Moreover, we evaluated the sustainable indicators considering the uncertainties of the major input parameters. The resulting probability distributions demonstrate that for TENGs at the current stage,

EPBTs are stable and competitive, while CO2 emission factors are less stable. Lastly, through sensitivity analysis, we find that TENG modules are potentially one of the most environmentally sustainable energy harvesters if future development confirms a larger performance ratio and a longer lifetime. To this end, a comparative techno-economic analysis of the TENG modules has been performed based on an annual capacity of 100 MW. Fig. 21 The comparison of LCOEs based on coal, natural gas, nuclear, We find that the cost of Module A is much lower than other wind, commercialized solar PV, hydropower, PSC and TENG modules. The LCOE values are referenced to the Levelized Cost and Levelized Avoided technologies when fully operational, while the cost of Module B Cost of New Generation Resources in the Annual Energy Outlook 2015 is found to be comparable to the cost of hydropower technol- reported by the United States Energy Information Administration. ogies. The results of the sensitivity analysis show that improved

Analysis Energy & Environmental Science

performance efficiency reduces significantly the module cost. 16 F.-R. Fan, Z.-Q. Tian and Z. L. Wang, Nano Energy, 2012, 1, The fabrication of high-efficiency modules through the adop- 328–334. tion of high precision fabrication processes is the most promis- 17 G. Zhu, Y. S. Zhou, P. Bai, X. S. Meng, Q. Jing, J. Chen and ing approach for further cost reduction. The results indicate an Z. L. Wang, Adv. Mater., 2014, 26, 3788–3796. estimated levelized cost of Module A and Module B to be US 18 W. Tang, T. Jiang, F. R. Fan, A. F. Yu, C. Zhang, X. Cao and 2.681 cents per kW h and US 9.43 cents per kW h, respectively. Z. L. Wang, Adv. Funct. Mater., 2015, 25, 3718–3725. The LCOE of TENGs is also very sensitive to the module 19 G. Zhu, B. Peng, J. Chen, Q. Jing and Z. L. Wang, Nano efficiency and is expected to be lower than that of other energy Energy, 2015, 14, 126–138. technologies if the module efficiency and lifetimes exceed 25% 20 Z. H. Lin, G. Zhu, Y. S. Zhou, Y. Yang, P. Bai, J. Chen and and 15 years, respectively. To achieve these targets, more efforts Z. L. Wang, Angew. Chem., Int. Ed., 2013, 52, 5065–5069. should be made to improve the lifetime and the efficiency of 21 Z. Wen, J. Chen, M.-H. Yeh, H. Guo, Z. Li, X. Fan, T. Zhang, TENGs rather than to identify cheaper materials and fabrica- L. Zhu and Z. L. Wang, Nano Energy, 2015, 16, 38–46. tion processes. 22 H. Zhang, Y. Yang, T.-C. Hou, Y. Su, C. Hu and Z. L. Wang, Nano Energy, 2013, 2, 1019–1024. 23 Y. Yang, H. Zhang, Y. Liu, Z.-H. Lin, S. Lee, Z. Lin, Competing financial interest C. P. Wong and Z. L. Wang, ACS Nano, 2013, 7, 2808–2813. The authors declare no competing financial interest. 24 Z. Li, J. Chen, J. Yang, Y. Su, X. Fan, Y. Wu, C. Yu and Z. L. Wang, Energy Environ. Sci., 2015, 8, 887–896. 25 S. Chen, N. Wang, L. Ma, T. Li, M. Willander, Y. Jie, X. Cao Acknowledgements and Z. L. Wang, Adv. Energy Mater., 2016, 1501778–1501787. 26 G. Zhu, J. Chen, T. Zhang, Q. Jing and Z. L. Wang, Nat. This research was supported by KAUST and the Hightower Commun., 2014, 5, 3426–3435. Chair foundation. The support provided for completing this 27 K. Y. Lee, J. Chun, J. H. Lee, K. N. Kim, N. R. Kang, J. Y. Kim, research is gratefully acknowledged. In addition, a helpful M. H. Kim, K. S. Shin, M. K. Gupta and J. M. Baik, Adv. discussion with Professor Heather MacLean is gratefully Mater., 2014, 26, 5037–5042. acknowledged. 28 L. Zhang, B. Zhang, J. Chen, L. Jin, W. Deng, J. Tang, H. Zhang, H. Pan, M. Zhu and W. Yang, Adv. Mater., 2016, References 28, 1650–1656. 29 Z.-H. Lin, G. Cheng, X. Li, P.-K. Yang, X. Wen and Z. L. 1 EIA, US Energy Information Administration, Washington, DC, Wang, Nano Energy, 2015, 15, 256–265. 2013. 30 Z. L. Wang, L. Lin, J. Chen, S. Niu and Y. Zi, Green energy and 2 IPCC, Climate change 2014: mitigation of climate change, technology, 2016. Cambridge University Press, 2015. 31 F.-R. Fan, L. Lin, G. Zhu, W. Wu, R. Zhang and Z. L. Wang, Published on 22 February 2017. Downloaded 29/03/2017 04:57:27. 3 Z. L. Wang, J. Chen and L. Lin, Energy Environ. Sci., 2015, 8, Nano Lett., 2012, 12, 3109–3114. 2250–2282. 32 G. Zhu, C. Pan, W. Guo, C.-Y. Chen, Y. Zhou, R. Yu and 4 G.-T. Hwang, Y. Kim, J.-H. Lee, S. Oh, C. K. Jeong, D. Y. Park, Z. L. Wang, Nano Lett., 2012, 12, 4960–4965. J. Ryu, H. Kwon, S.-G. Lee and B. Joung, Energy Environ. Sci., 33 G. Zhu, Z.-H. Lin, Q. Jing, P. Bai, C. Pan, Y. Yang, Y. Zhou 2015, 8, 2677–2684. and Z. L. Wang, Nano Lett., 2013, 13, 847–853. 5 D. Evans, How the Next Evolution of the Internet is Changing 34 G. Zhu, J. Chen, Y. Liu, P. Bai, Y. S. Zhou, Q. Jing, C. Pan and Everything, Whitepaper, Cisco Internet Business Solutions Z. L. Wang, Nano Lett., 2013, 13, 2282–2289. Group (IBSG), 2011, vol. 1, pp. 1–12. 35 S. Chen, C. Gao, W. Tang, H. Zhu, Y. Han, Q. Jiang, T. Li, 6 Y. Gao and Z. L. Wang, Nano Lett., 2007, 7, 2499–2505. X. Cao and Z. Wang, Nano Energy, 2015, 14, 217–225. 7 Y. Qin, X. Wang and Z. L. Wang, Nature, 2008, 451, 809–813. 36 T. Ibn-Mohammed, S. Koh, I. Reaney, A. Acquaye, D. Wang, 8 X. Wang, J. Song, J. Liu and Z. L. Wang, Science, 2007, 316, S. Taylor and A. Genovese, Energy Environ. Sci., 2016, 9, 102–105. 3495–3520. 9 S. Xu, B. J. Hansen and Z. L. Wang, Nat. Commun., 2010, 1,93. 37 T. Ibn-Mohammed, R. Greenough, S. Taylor, L. Ozawa- 10 S. Xu, Y. Qin, C. Xu, Y. Wei, R. Yang and Z. L. Wang, Nat. Meida and A. Acquaye, Energy Build., 2013, 66, 232–245. Nanotechnol., 2010, 5, 366–373. 38 S. Hellweg and L. M. I. Canals, Science,2014,344, 1109–1113. 11 R. Yang, Y. Qin, L. Dai and Z. L. Wang, Nat. Nanotechnol., 39 S. Koh, T. Ibn-Mohammed, A. Acquaye, K. Feng, I. Reaney, 2009, 4, 34–39. K. Hubacek, H. Fujii and K. Khatab, Sci. Rep., 2016, 6, 39514. 12 G. T. Hwang, M. Byun, C. K. Jeong and K. J. Lee, Adv. 40 A. Acquaye, K. Feng, E. Oppon, S. Salhi, T. Ibn-Mohammed, Healthcare Mater., 2015, 4, 646–658. A. Genovese and K. Hubacek, J. Environ. Manage., 2016, 187, 13 Z. L. Wang and J. Song, Science, 2006, 312, 242–246. 571–585. 14 Z. L. Wang, G. Zhu, Y. Yang, S. Wang and C. Pan, Mater. 41 M. Le, Solar Energy Technologies Office, U. S. Department of Today, 2012, 15, 532–543. Energy, 2015, pp. 1–121. 15 Z. Lina´Wang, Faraday Discuss., 2014, 176, 447–458. 42 F. C. Krebs, Sol. Energy Mater. Sol. Cells, 2009, 93, 394–412.

Energy & Environmental Science Analysis

43 R. R. Søndergaard, M. Ho¨sel, N. Espinosa, M. Jørgensen and 55 A. Azapagic, C. Pettit and P. Sinclair, Clean Technol. Environ. F. C. Krebs, Energy Sci. Eng., 2013, 1, 81–88. Policy, 2007, 9, 199–214. 44 K. Hwang, Y. S. Jung, Y. J. Heo, F. H. Scholes, S. E. Watkins, 56 Ecoinvent, Ecoinvent database, http://www.ecoinvent.org/, J. Subbiah, D. J. Jones, D. Y. Kim and D. Vak, Adv. Mater., Accessed 20th November 2016, 2016. 2015, 27, 1241–1247. 57 J. B. Guineé, Int. J. Life Cycle Assess., 2002, 7, 311–313. 45G.Rebitzer,T.Ekvall,R.Frischknecht,D.Hunkeler,G.Norris, 58 W. Kellogg, M. Nehrir, G. Venkataramanan and V. Gerez, T.Rydberg,W.-P.Schmidt,S.Suh,B.P.Weidemaand IEEE Trans. Energy Convers., 1998, 13, 70–75. D. W. Pennington, Environ. Int., 2004, 30, 701–720. 59 K. Branker, M. Pathak and J. M. Pearce, Renewable Sustain- 46 T. Ibn-Mohammed, R. Greenough, S. Taylor, L. Ozawa- able Energy Rev., 2011, 15, 4470–4482. Meida and A. Acquaye, Build. Environ., 2014, 72, 82–101. 60 D. L. Smith and D. W. Hoffman, Phys. Today, 1996, 49, 60. 47 A. A. Acquaye, T. Wiedmann, K. Feng, R. H. Crawford, J. Barrett, 61 Polyflon Technology Limited, 100 PTFE Film (Polytetrafluoro- J.Kuylenstierna,A.P.Duffy,S.L.KohandS.McQueen-Mason, ethylene), http://www.polyflon.co.uk/products/fluoropolymer- Environ. Sci. Technol., 2011, 45, 2471–2478. film/ptfe-film, Accessed 16th December, 2016. 48 D. Collado-Ruiz and H. Ostad-Ahmad-Ghorabi, J. Cleaner 62 Teflon, Fluoroplastic Film-Properties Bulletin, https://www. Prod., 2010, 18, 355–364. chemours.com/Teflon_Industrial/en_US/assets/downloads/ 49 R. Garcıá-Valverde, J. A. Cherni and A. Urbina, Prog. Photo- teflon-fep-film-properties.pdf, Accessed, 16th, December, voltaics, 2010, 18, 535–558. 2016. 50 N. Espinosa, M. Ho¨sel, D. Angmo and F. C. Krebs, Energy 63 R. Laleman, J. Albrecht and J. Dewulf, Renewable Sustainable Environ. Sci., 2012, 5, 5117–5132. Energy Rev., 2011, 15, 267–281. 51 D. Yue, P. Khatav, F. You and S. B. Darling, Energy Environ. 64 J. Gong, S. B. Darling and F. You, Energy Environ. Sci., 2015, Sci., 2012, 5, 9163–9172. 8, 1953–1968. 52 J.J.Conti,P.D.Holtberg,J.A.Beamon,A.Schaal,J.Ayouband 65 J. Poortmans and V. Arkhipov, Thin film solar cells: fabrication, J. T. Turnure, United States of America Department of Energy characterization and applications, John Wiley & Sons, 2006. Information. Office of Integrated and International Energy 66 C. Becker, D. Amkreutz, T. Sontheimer, V. Preidel, D. Lockau, Analysis. Available at http://www.eia.gov/neic/speeches/ J. Haschke, L. Jogschies, C. Klimm, J. J. Merkel and P. Plocica, newell_12162010.pdf, 2011. Sol. Energy Mater. Sol. Cells, 2013, 119, 112–123. 53 I. ISO, London: British Standards Institution, 2006. 67 Levelized Cost and Levelized Avoided Cost of New Genera- 54 I. ISO, International Organization for Standardization, Geneva, tion Resources in the Annual Energy Outlook, U.S. Energy Switzerland, 2013. Information Administration, 2015. Published on 22 February 2017. Downloaded 29/03/2017 04:57:27.