The International Energy Agency reports than more than 1500 million people live without access to electricity 1, at the same time there exists a growing concern about climate change and green house gas emissions, and therefore the unavoidable need to think about sustainable means of electricity supply to those deprived of it. Renewable energy technologies in general and photovoltaic in particular will be part of the solution to this complex problem.
In the past few years, there has been a remarkable improvement in the efficiency of photovoltaic (PV) devices and in the deployment of PV systems; nevertheless, the share of PV in world electrification is still very limited. Major barriers have been: low energy density of the solar resource and high cost of the photovoltaic modules. Even with high power conversion efficiencies obtained routinely for recognized PV technologies, the cost per kWh generated by solar electricity is still very high.
New technologies that could create a breakthrough in the photovoltaic industry are now in the research stage and an increase in amount of resources are invested in their development. In recent years huge research efforts are being devoted to new promising photovoltaic technologies, which forecasts further and significant cost reduction in the photovoltaic module manufacturing. In particular, organic solar cells have been the subject of attention for some time because they offer the prospect of low cost active layer material, low-cost substrates, low energy input and easy industrial production up-scaling. The latter offers the possibility of printing active layers, thereby boosting production throughputs typically by a factor 10 to 100 compared to other thin-film technologies. Modules that use organic solar cells technology could cost less than 0.5€/Wp and easily reach the GW production scale. This paper focuses on a most promising technology: plastic solar cells, in which the devices use conjugated polymers as active layer, it aims to study the environmental impact of PV organic technologies using a life cycle assessment (LCA) methodology, with special focus on the material inventory and the energy payback time calculation 2.
This paper will perform a life cycle assessment of the fabrication of a typical laboratory plastic solar cell. We have selected, as representative of a broad range of different approaches, organic solar cells whose active layer is made of a blend of a fullerene derivative (phenyl-C61-butyric acid methylester, PCBM) and a conjugated polymer (poly-3-hexil-thiophene, P3HT). The proposed substrates and electrodes are also typical of this organic PV technology. The material inventory and the energy embedded in the process will, however, not differ from other similar approaches which use blends of other functionalized carbon-based nanoparticles and conjugated polymers of other kind, such as derivatives of poly-para-phenylene-vinylene (PPV). For this technology, drawing on existing literature and the experience arising from our own device preparation method 3–7, a material inventory has been calculated, and the environmental impact of module fabrication (i.e. energy payback time and green house gas emissions) extrapolating the laboratory procedures towards an assumed large scale industrial production has also been calculated. In the Methodology section, the assumptions and scope of the work are stated. In the discussion, results for the organic photovoltaic technology will be compared to other PV technologies. The study introduces a preliminary approach to calculate the environmental impact of the production of a promising technology that, if the forecasted cost reduction is achieved, could soon reach the GW scale.
To summarize, it is important to develop a method that calculates the parameters used in life cycle analysis for this organic PV technology. This paper highlights the LCA issues such as material inventory, energy embedded in the material processing and in the fabrication procedure, and finally energy payback time and avoided emissions for a given irradiance, module life-time and power conversion efficiency. These parameters will be calculated for a typical organic solar cell and compared to other inorganic technologies.
LIFE CYCLE ANALYSIS OF PHOTOVOLTAIC TECHNOLOGIES
From an environmental point of view, as it emerges from all LCA assessments of electricity production from different energy systems, photovoltaic technologies, with a range of 21–37 g of CO2 eq. emissions per kWh of produced energy for an average southern European isolation have a clear advantage over coal (900), gas combined cycle (439) and even nuclear (40) (see e.g. References 8, 9). Also a detailed comparison of energy payback times and emissions for different photovoltaic technologies has been carried out where most attention has been devoted to monocrystalline (mono-Si) and multicrystalline (multi-Si) silicon modules since they are the most often used PV technologies with a 93% share of the market in 2006 9–11. Also thin film technologies of amorphous silicon (a-Si:H), cadmium telluride (CdTe) and copper indium diselenide (CIS) modules have been assessed 12, 13. All thin film technologies show better values of energy payback time and avoided emissions, and despite their lower power conversion efficiency, their main advantage is the use of much less material in the active layer (200–300 µm for multi-Si and mono-Si, compared with around 10 and 1 µm for CIS and CdTe, respectively). Lower material and energy consumption in the manufacturing process is likely to bring, in the near future, a reduction in the cost of the modules as the technology progresses through its learning curve 14, 15. Yet these cutbacks will not necessarily cause a clear breakthrough that will allow a massive deployment of photovoltaic systems, especially in developing countries where they could be very important but where the availability of capital is small.
The extremely high manufacturing throughput potential of organic solar cells could have a positive impact in the markets and therefore is attracting many investors. The European Photovoltaic Technology Platform, supported by the Sixth European Framework Programme for Research and Technological Development has included the organic photovoltaic technologies on its roadmap, which indicates that organic PV is on the verge of commercialization and is already producing a rapidly developing global manufacturing base 16.
Plastic solar cells can be flexible and lightweight, and though not as efficient or long-lived as solid-state PV devices, they can reach a much better performance in terms of cost per energy produced during its lifetime than any other inorganic technology.
Manufacturability of the organic devices is appalling since low-temperature fabrication steps can be used, such as screen-printing, spray-coating or ink-jet printing. Furthermore, such deposition routes are compatible with heat-sensitive substrates such as plastics, allowing high throughput reel-to-reel manufacturing of flexible and lightweight modules by integrating thermoplastic processing technologies. For all of the approaches in this category of organic solar cells, the active layer consists, at least partially, of an organic dye, small, volatile organic molecules or polymers suitable for liquid processing.
Dye-sensitized solar cell (DSSC), which could supply solar electricity on a large scale, has attracted most research in recent years. Nanostructured metal oxide films formed by an assembly of nanoparticles sintered over a conducting substrate provide a large internal area that can perform different functions, especially when combined with organic hole-conductors that fill the voids in the nanostructure 17. DSSCs can achieve efficiencies higher than 11% 18, and it has triggered the first steps towards commercialization of the devices and the associated machinery by companies such as Konarka, Dyesol, Solaronix and G24I 19–22. Through a detailed material inventory approach, the environmental impacts of electricity generated from dye sensitized solar cell systems have been analysed; it shows a process energy requirement of 100–280 kWh/m2 of active solar cell area and the related green house gas emissions ranges from 19 to 47 g CO2 eq. per kWh of energy produced for different assumptions of isolation, efficiency and lifetime of the DSSC modules 23.
Early in the 90s, several groups also developed ‘full-organic’ photovoltaic devices called ‘plastic solar cells’ 24. They are based on initial work by A. J. Heeger, University of California (Santa Barbara, USA) 25, 26 and by R. H. Friend, University of Cambridge (UK) 27. In 1995, it was observed that the mixing of two polymers formed a three-dimensional heterojunction which led to efficient charge generation within the whole film. Solar cells based on blends of methanofullerenes with derivates of conjugated polymers have been under investigation for several years since those initial and promising discoveries. Now, some groups report the fabrication of devices with efficiencies higher than 4% 28–30. Tandem devices with polymer/fullerene derivative blends have given efficiencies just above 6% 31. With this trend in the power conversion efficiencies, the full organic polymeric technology could represent a breakthrough in the learning curve if the durability of the devices is guaranteed for more than 5 years. Partial cost analysis and forecast for this technology has been recently presented, however, it does not include an environmental analysis of the energy payback times and green house gas emissions 14.
The Life Cycle Analysis (LCA) approach measures the environmental impact of products over their entire life cycle, from cradle to grave. This study will focus on some aspects of the LCA for a typical organic solar photovoltaic module. The study carries out the analysis of the laboratory process for the fabrication of organic solar cells, then extrapolates frame and module process from other well-known technologies, and finally makes some assumptions about the cap that could be set for extrapolating the laboratory process to the large-scale industrial process. Usually, in an LCA analysis, a separated calculation is performed for the balance of system (BOS) components, which are strongly dependent on the final use of the PV system, it is expected that big differences in the energy payback time will arise between grid-connected and stand-alone systems. A detailed calculation for a typical solar home system considering the environmental impacts and the economical costs will be presented in another paper. The aim of this work is to analyse the different stages of the organic solar cell module fabrication process. The steps for our calculation are explained in the following paragraphs.
The methodology is comprised of three approaches that are linked between them in a sequential order. Each stage produces a partial outcome that is used in the following step. A global view of the environmental impact of any technology should take into account at least the following parameters.
It is the amount of raw materials per square metre of PV module. This calculation is the first step in order to obtain the energy embedded in a defined surface of the active solar cell, plus the non-active surface, which contributes to the total module surface. The subsequent energy calculations will take into account this material inventory. Also, some materials used in PV technologies are scarce and this fact could become an important problem for a large scale deployment of a particular PV technology, the material inventory is useful in order to identify possible bottlenecks for this up-scaling in the production.
Energy pay back time (EPBT)
It is the time required for the solar PV system to generate the equivalent amount of energy consumed in the construction and decommissioning phases. First, we calculate the energy embedded in the production process of all the materials identified in the inventory, and then we add the energy process of the organic solar cell fabrication and the module fabrication. The output of the calculation is the energy embedded per surface of PV module.
In order to calculate the energy produced by the PV system during its operative lifetime, some assumptions have to be made: first, the efficiency of the produced organic PV modules is set to 5% (efficiencies already obtained for laboratory cells) and 10% (forecast for module efficiency on industrial level in 2020); second, the isolation level used for the calculation of the energy produced by the PV system during its operation is set to 1700 kWh/m2, typical of European southern countries and representative of a world average, and finally, we consider a 15 years lifetime for the system, which is low compared with other PV technologies (usually 20–30 years), but realistic in terms of the aims for 2013 of the European Photovoltaic Technology Platform for organic photovoltaics. The forecasted costs for this figures are ∼0.5–0.6€/Wp by 2013–2020 and < 0.5€/Wp after 2020 16.
Life cycle of green house gas (GHG) emissions
Photovoltaic systems does not generate green house gas emissions during its electricity production stage, but if we consider the entire life cycle, then the pollutant emissions during the fabrication, decommissioning and recycling stages have to be taken into account. The method here used is that the emissions are calculated using as starting point the embedded energy which is one of the outputs of the previous energy payback time calculation. Then we will assume a determined energy mix for the location of the facility where we supposedly produce the modules. It will be given by the UCPTE database (Union of the Coordination of Production and Transmission of Electricity in Europe). This allows us to calculate the CO2 emissions per square metre of module during the lifetime of the system, also called the embodied CO2.
The CO2 payback time (CO2PBT) is the ratio of the total embodied CO2 in the facility, to the total avoided CO2 emissions. These avoided emissions are calculated taking into account a determined energy mix of the location where the system has been installed and the amount of emission generated in order to produce the same amount of electricity produced by the PV system during its lifetime. This calculation is strongly dependent on the geographical location of the PV system, since the embodied and avoided emission are usually taken as the emissions of the energy mix of the selected country for the amount of energy used in fabrication and produced during the operational lifetime of the installation.
LIFE CYCLE OF ORGANIC PHOTOVOLTAIC MODULES
In this section the life cycle of organic photovoltaic modules is addressed. Since there is still no commercial production of full organic photovoltaic cells, our calculation is based in the laboratory process currently used. The typical process used by most groups and ourselves to prepare laboratory samples is analysed here. First the fabrication steps are explained and the energy input of these steps is calculated. Then, the life cycle of raw materials is analysed. With the output of this analysis a final material inventory and energy embedded in both material and cell processing is calculated. Finally, a comparison with other photovoltaic technologies is presented.
Fabrication of a typical organic solar cell
Our laboratory process for cell fabrication is explained in Figure 1. A glass substrate covered with indium tin oxide is carefully cleaned with acetone, isopropanol and oxygen plasma. On the cleaned substrate a thin film of poly-(3,4-ethylene-dioxythiophene)/poly-styrene-sulphonate (PEDOT:PSS, 1:2.5 wt) is spin-cast (1500 rpm, 180 s) in air from a filtered (0.45 µm) water solution (1.3 wt% in water, previously heated at 50°C and stirred for 30 min). Then, the devices are transferred to a nitrogen glove box, and annealed at 70°C for 2 h. Then, a second layer of a blend of P3HT and PCBM (1:1 wt) is spin cast (1500 rpm, 180 s) from a toluene solution (3 wt%), finally a Ca(60 nm)/Al(150 nm) electrode is evaporated on top of the P3HT. All the polymers and fullerene derivatives that we use in the preparation process are commercial products (Sigma-Aldrich).
We use this process as a guide for the calculation of the material inventory and the energy embedded during manufacture of the organic cell. It is a cap for the total amount of energy and materials, since it is clear that any large scale industrial process will optimize the amount of raw material and the consumption of energy used in the process.
ITO patterning and cleaning
ITO covered glass substrate is the preferred front transparent electrode. At laboratory level ITO patterning can be made masking with adhesive tape and exposing the part of ITO that should be etched (e.g. with HCl and HNO3 aqueous solutions). Once the traces of acid are removed (e.g. using a NaHCO3 (aq.) solution) the adhesive is peeled off. We have not considered energy requirements or the etching inputs for the patterning step since our cells are made individually for the moment.
The ITO substrates are then cleaned by ultrasonication in isopropanol and acetone for 10 min each. An oxygen plasma treatment is then applied immediately prior to spin-casting the PEDOT:PSS layer. Conditions and material inventory for ultrasonication and oxygen plasma are shown in Tables I and II, respectively. Note that the calculations include a use factor of the laboratory equipment, since during the processing of 1 cm2 solar cell, they are not used at their maximum capacity.
Table I. Ultrasonication conditions and material inventory for 1 cm2 organic solar cell.
Estimated from the use factor and the maximum capacity.
Spin casting is a procedure used to apply uniform thin layers to flat substrates. An excess amount of a solution is placed on the substrate, which is then rotated at high speed in order to spread the fluid by centrifugal force. The applied solvent is usually volatile, and simultaneously evaporates. The higher the angular speed of spinning is, the thinner the film which is obtained. The thickness of the film also depends on the concentration of the solution and the solvent.
During the spin-casting process the substrate must be fit to the plate and it is made applying a small vacuum to the bottom of the substrate.
Previous to spin casting, the aqueous solution of PEDOT:PSS, 1:2.5 (1.3 wt%) is heated and stirred (50°C for 30 min) (Table III). Then it is filtered to remove dust (usually with a 0.45 µm filter). Spin-casting conditions are shown in Table IV. A material recovery system is considered, since the material use efficiency for spin casting is extremely low. To deposit 120 nm on 1 cm2 surface, 0.2 ml of solution must be used; it means that only the 0.46% of PEDOT/PSS 1:2.5 (solid content) is deposited on the substrate with the spin-casting method.
Table III. Heating and stirring of PEDOT/PSS conditions for 1 cm2 organic solar cell.
Estimated from the use factor and the maximum capacity.
Then the substrates are transferred to a nitrogen glove box in order to prevent the effects of the moisture and the oxygen on the active layer. The energy requirements and the N2 consumption for the glove box are treated separately below, since it affects several steps.
Annealing and active layer coating
Once the substrates are transferred to the nitrogen atmosphere, they are annealed during 2 h at 70°C. The annealing process is usually made using an oven; however in a laboratory scale it is common the use of a simple hot-plate previously introduced into the glove box. Table V shows the considered conditions for substrate annealing.
Table V. Substrate annealing conditions for 1 cm2 organic solar cell.
Hot plate mean power
Hot plate use factor
Maximum applicable surface
The active layer is a blend of the P3HT and PCBM (1:1) in toluene. The solution is spin-cast on the substrate. The spin-casting conditions are the same that in the case of PEDOT:PSS.
For the materials that are deposited using the spin-coating method (PEDOT:PSS and P3HT:PCBM layers for our cells) we give two values for the material contained in the cell: (1) the minimum amount of material that should be used in order to fill the volume (area of cell times the thickness of the layer), taking into account for the calculation the reported densities of the materials: PCBM, with Mw = 910.88, has a density of ρPCBM = 1.5 g/cm3; P3HT (regioregular head-to-tail) with Mw = 45000–65000, with ρP3HT = 1.11 g/cm3 and PEDOT:PSS (1:25) with ρPEDOT ≈ 1 g/cm3; and (2) the material amount used to prepare the solutions (3 wt%) in toluene (ρtoluene = 0.862 g/cm3), which includes a big loss of material during the spin-casting process that is scattered from the spinning substrates and not used in the final active layers. From comparison of this value with that given in Table XIX, and since we use 1 ml of solution to prepare five cells of 1 cm2 in our experimental procedure, we can deduce a value for the ‘material usage efficiency’ of the spin-casting process, which results in a meagre 0.36% for the P3HT and PCBM, and a slightly better 0.45% for the PEDOT:PSS. Both figures demonstrate that the spin-coating processing method is not suitable for large-scale industrial production, even if a ten-fold improvement of material usage is achieved in a new developed ad hoc spinner.
Note that material inventory calculations (Table XIX) have been done assuming a spin-coating process with material recovery system (both for PEDOT:PSS and active layer deposition), so the material usage efficiency is assumed to be 30% (Table IV). This ‘recovery’ is used because of the extremely low efficiency of the use of materials by the spin-coating deposition method, which gives further support to the need of exploring alternative methods for massive manufacturing.
Metal evaporation of the back electrode
Finally, two thin layers of metals (Al and Ca) are evaporated on the top of the active layer. They are used as back electrode. Thermal metal evaporation is an energetically intensive process; apart from the electric source to heat the filament the evaporation system includes a vacuum pump and a cooling system. The evaporation of a metal layer requires around 60–90 min: previous to evaporation 15 min are needed to create the required vacuum, then the specific time to deposition of the metal (the deposition rate is controlled by the base pressure applied and the current through the filament) and finally around 15 min are needed to cool the filament. The conditions and equipment followed in our process are presented in Table VI. The material use efficiency is worked out based on our own experience. Metal evaporation is a widely used technique in the electronic manufacturing industry and our data must be taken as a laboratory estimation.
Table VI. Metal vacuum evaporation conditions for 1 cm2 organic solar cell.
Base pressure for Al
2.8 × 10−6mbar
Base pressure for Ca
1.8 × 10−6mbar
Al deposition rate
Ca deposition rate
Vacuum pump working time
Cooling system working time
Evaporator mean power
Cooling system mean power
Vacuum pump mean power
System use factor
Maximum vacuum pressure
Maximum evaporator power
Maximum applicable surface
Cooling system power
Vacuum pump maximum power
Material use efficiency
N2 glove box atmosphere
As it was mentioned above, part of the processing of the organic solar cells must be done under nitrogen atmosphere. The use of the nitrogen glove box is required because the active layer is affected from the exposure to oxygen and moisture. The estimation of energy requirements and nitrogen consumption for the glove box has been done using a factor of use from the maximum working capacity of the glove box during 1 week. The only continuous input of energy is the control and maintenance of the N2 atmosphere (pump and instrumentation), since the auxiliaries have been already considered in the integrated evaporator and spin coater. An auxiliary pump to control the antechamber is also taken into account with an estimated mean working time of 7 h/week. Table VII resumes the used conditions and equipment for N2 glove box consumption.
Table VII. N2 glove box consumption for 1 cm2 organic solar cell.
Equivalent energy for 1 cm2 OSC
Equivalent N2 consumption for 1 cm2 OSC
3.93 × 10−3 Nm3
System use factor for 1 cm2 OSC
M Braum 1.8 m3
Mean power of consumption
Antechamber pump mean power
150 l/week (200 bar)
Maximum surface of the solar cells
Estimated number of 100 cm2 cells per week
Life cycle of raw materials
The life cycle of the raw materials used in the preparation of the cell must be analysed in order to calculate the full-embedded energy per square metre of module.
Electrodes: front ITO transparent oxide electrode and Ca/Al back electrode,
Front electrode: layer of PEDOT:PSS,
Active layer: blend of polymer and fullerene derivative (for this selected bulk-heterojunction realization of the organic solar cell we use P3HT:PCBM),
Encapsulation and frame (usually aluminium frame, plus additional EVA layer, but if encapsulation is rigid enough, the additional frame is not needed).
Glass, encapsulation and frame are the same as that for other well-known technologies; therefore, it is not necessary to remake the inventory here. We can use data from other references for these components.
Energy requirements for chemical synthesis steps
Estimation of energy requirements in the production of speciality chemicals is quite complex, since information on such production processes is scarce. Moreover, many of the chemical processes presented in this work are not yet in a full production scale, but we presume that the volume of production is still not at laboratory scale. We have evaluated several approaches to this problem 32–35.
The production routes of P3HT, PEDOT/PSS and PCBM include several chemical synthesis steps. The production processes and the production volumes of the last steps before final product is obtained are often intellectual property protected by patents granted to the companies and the only available information is the laboratory synthesis found in research articles and run-out patents. In this sense, we have proceeded retrieving the most cited synthesis as reference in articles and patents for those last steps where there are no previous LCA studies.
However, estimation of energy requirements based on a laboratory scale seems not to be very realistic. Geisler proposed a methodology and default values for utility inputs requirement in fine chemical synthesis steps derived from on-site data on chemical production processes and from heuristics 32. In our study, this methodology has been applied. We have used the available information about the investigated chemical synthesis steps in order to extrapolate to a bigger scale production and to estimate the energy requirements and main material inputs. Table VIII shows the default values of the utility inputs for a generic chemical synthesis step. For N2 and steam inputs, we have quantified their embodied energies according to the values shown in Table IX.
Table VIII. Estimated best-case and worst-case default values of the energy inputs for one chemical synthesis step 32.
n.a., not applicable because no solvent regeneration assumed.
Utility inputs for reaction and workup
Utility inputs for solvent regeneration
Table IX. Embodied primary energy in the utility inputs.
For the sake of simplicity we do not include solvents or reagents on the material inventory. Moreover, we have considered the worst-case in the calculations since these chemical products have been only produced since a few years and probably the full-scale production is not completely optimized.
Processing of the active layer
Poly-3-hexylthiophene (P3HT) is amongst the common poly-3-alkylthiophenes (P3ATs) used nowadays as donor material in the active layer of the organic solar cells 36–38. Polythiophenes (PTs) result from the polymerization of thiophenes, a sulphur heterocycle that can become conducting when electrons are added or removed from the conjugated-orbitals via doping. PTs are insoluble, so in the quest for soluble and procesable conducting polythiophenes, alkylthiophenes have been polymerized. PTs can be synthesized electrochemically by applying a potential across a solution of the monomer to be polymerized, or chemically using oxidants or cross-coupling catalysts. Chemical synthesis offers two advantages compared with electrochemical synthesis of PTs: a greater selection of monomers, and, using the proper catalysts, the ability to synthesize perfectly regioregular substituted PTs. Regioregularity has shown to be determinant for electrical conductivity in PATs. It is shown how the optical and electronic properties vary with regioregularity of the PAT, i.e. the larger the percentage of HT–HT (head to tail energetic conformation) couplings the more conjugated and more conductive the PAT will be 39.
At the present, the main suppliers of commercial PATs (e.g. Rieke Metals Inc., Plextronics, Merck, Honeywell) use the Grignard metathesis (GRIM) 40, 41 or the Rieke method 42, 43 as base to achieve regioregular ∼100% HT-coupled structures of polythiophene derivates.
Figure 2 shows the processing route from thiophene and bromine to ∼100% HT Regioregular P3HT. Reagents and solvents are not shown in the scheme.
We have followed the GRIM polymerization method as reference in this study. Before the polymerization the initial monomer 3-hexylthiophene (3HT) must be activated to 2,5-dibromo-3-hexylthiophene 44–46. 3HT is derived from 3-bromothiophene (a brominated thiophene) and bromohexane 47.
Brominated thiophenes are derivated from thiophene and bromine; however their synthesis process usually contains various chemical steps (bromination and debromination with different solvents and conditions), which are known as halogen dance. The synthesis of 3-bromothiophene involves the bromination of thiophene to 2,5-dibromothiphene 48 and posterior debromination 49.
Alkenes bromination can be done by a great number of different chemical reactions; however, substitution and addition reactions (substitutive and additive bromination) are the most common methods employed in industrial processes 50, 51. Hexane production is considered to be similar to the pentane production process 52.
Thiophene and its derivatives exist in petroleum or coal. Thiophene derivatives are also found in natural plant pigments. Commercially, thiophene can be prepared by the reaction of butane and sulphur 51. Due to the lack of data about the energy requirements of this process on industrial producers, we have considered the thiophene production similar to toluene (both are aromatic rings produced from oil refining), for which there exists published LCI data 52. The bromine industrial production involves the treatment of bromide-rich brines with sulphuric acid and chlorine gas, flushing through with air. We have considered the energy inputs in this process similar to the brine purification (previous process for the chlorine production) 52.
The results for energy and material requirements for P3HT processing are presented in Tables X and XI. The functional unit is 1 kg of the 99% HT-couple regioregular P3HT.
Table X. Cumulative energy requirements for 1 kg of 99% HT-coupled P3HT production.
Thermal energy (MJth)
Electrical energy (MJel)
Bold denotes the Total value.
Table XI. Input materials inventory for 1 kg of 99% HT-coupled P3HT production.
PCBM, reported in 1995 26, is a soluble derivative of fullerenes. The addition of diazoalkanes to fullerenes is used to overcome the limited solubility of pure C60 and C70 in organic solvents and its tendency to crystallize during film formation in high-concentration blends.
Fullerene production methods can be sorted into two categories based on the type of carbon feedstock: methods that decompose a liquid or gaseous feedstock to obtain the atomic carbon, and methods that vaporize a pure carbon source into high temperature plasma. Among the first ones, pyrolytic processes decomposing hydrocarbons, such a benzene or toluene, are use to create most C6053 All fullerene synthesis production methods typically produce unwanted impurities and most require purification. Kushmir and Sandén have presented recently an interesting and complete approximation to the energy requirements of carbon nanoparticle production 54. According to them the pyrolytic processes are simpler to implement and scaling them up is a matter of increasing fuel flow.
In order to get the final PCBM from C60 or C70, a functional group is attached to the cluster. The addition of diazoalkanes to C60, proposed by Hummelen et al. 55, is the functionalization method most cited in the literature. Their one-pot procedure starts from the in situ generation of the diazo compound (methyl 4-benzoylbutyrate p-tosylhidrazone) from the tosylhydrazone anion and 4-benzoylbutyric acid.
4-Benzoylbutyric acid is a keto-acid, which can be prepared from 1-phenycyclopentene through ozonolysis and deprotonation substeps 56. 1-phenylcyclopentene is prepared from cyclopentanone 57. Cyclopentanone is derived from the oxidation of cyclopentene/cyclopentane mixtures 58. The PCBM processing route until this step is depicted in Figure 3.
The tosylhidrazide anion is added in form of p-toluene sulphonyl hydrazide (tosylhidrazide) to obtain the final diazocompound. Tosylhydrazide can be prepared by the reaction of p-toluene sulphonyl chloride with hydrazine 59. A common hydrazine synthesis is the Olin Raschig process 52 (based on the oxidation of ammonia with sodium hypochlorite). A possible route to p-toluene sulphonyl is from chlorosulphonic acid and toluene 60. Chlorosulphuric acid is manufactured by direct union of equimolar quantities of sulphur trioxide and dry hydrogen chloride gas chlorosulphuric acid.
Toluene, ammonia, sodium hypochlorite and hydrogen chloride life cycle inventories are published in Reference 52. Sulphur trioxide can be obtained as by-product from sulphuric acid production or from desulphuration process during the oil refining. Due to lack of reliable data the Naphtha production energy requirement are considered as reference. Figure 4 illustrates the initial steps in the PCBM processing route.
The results for energy and material requirements for PCBM processing are presented in Tables XII and XIII. The functional unit is 1 kg of the PCBM.
Table XII. Cumulative energy requirements for 1 kg of PCBM production.
Thermal energy (MJth)
Electrical energy (MJel)
Bold denotes the Total value.
It does not include the toluene production for fullerene preparation, which already embedded in the pyrolysis fullerene production, as well as the feedstock energy.
Indium tin oxide (ITO) is a tin-doped indium transparent oxide semiconductor material. An ITO film exhibits both high transmittance in the visible region and good electrical conductivity. It is used widely as a transparent conducting electrode in opto-electronic applications such as flat panel displays, solar cells and heat reflecting glass. Production of indium tin oxide (ITO) thin-film coatings continues to be the leading end use of indium and accounts for approximately 84% of global indium consumption 61. Indium is a very scarce and expensive metal, therefore their use in large scale production of organic solar cells should be eliminated in order to avoid a future bottleneck in the material stock. Alternative transparent conducting layers are subject of extensive research in recent years, the most successful candidates so far have been high conductivity formulations of PEDOT:PSS 62, zinc oxide (ZnO) nanostructured layers 63 or carbon nanotubes deposited or grown on the substrates 64, 65.
A typical ITO transparent electrode production process is show in Figure 5. ITO electrodes are implemented as ITO coated glass substrates. A SiO2 passivation layer is previously deposited on the glass in order to prevent reaction between ITO and glass. The preferred ITO deposition method for thin films is radiofrequency (RF) sputtering. The process is performed in an Argon atmosphere, where the working pressure is around 1 mtorr. The addition of O2 increases the crystallinity and the conductivity of the ITO layer.
Table XIV shows the assumed RF sputtering conditions for our estimation of the process. These assumptions are based on typical values for ITO sputtering found in literature 66 and commercial sources 67, 68.
Table XIV. Sputtering assumed conditions.
RF reactive sputtering
Fused quartz (10 cm2)
Substrate heater maximum power
Vacuum pump maximum power
Heater and vacuum pump factor use
Pre-sputter time (heating substrate)
Argon flow rate
O2 reactive gas
Efficiency of ITO use (including recovery)
Previously to the ITO deposition, a SiO2 layer is deposited as passivation layer. For sake of simplicity the same sputtering system is considered for this layer and the same energy requirements for the target preparation. Energy requirements for sand extraction can be found in Reference 69.
The base materials for ITO target are indium and tin. Indium is recovered as a by-product of zinc production from the fumes, dusts, slags, and residues in zinc smelting. Tin is obtained chiefly from cassiterite mining, where it occurs as an oxide. Energy requirements for Tin and Indium metals production are found in References 69, 70.
The ITO powder production can be done by different routes, but normally it involves the preparation of a solution containing a mixture of Indium and Tin (typically with ratio 10:1). The liquid is evaporated and the powders are fired in an oxygen atmosphere 71. Then, the mixture of oxide powders is hot pressed, hot isostatically pressed or slip cast and sintered 67. Energy requirements of both processes are expected to be highly energetic intensive, our information source (personal communication with an ITO target producer, Hunan Boyun New Materials Co. Ltd. 2008) estimated 72 and 108 MJel/kg, respectively.
Pure and compressed argon, oxygen and nitrogen are commonly used gases in many chemical synthesis and laboratory processes. The most common production process of these gases is the full distillation of air (three distillation towers). From personal communication with industrial and lab gases producer (Praxair, 2008), we have estimated the mean energy requirements per Nm3 of N2, Ar and O2 as 3.12, 231.73 and 9.23 MJel, respectively. This estimation considers the precompression of air (5–6bars), the auxiliary equipment consumption, the ratio of recovery for each gas and the final compression to 200 bar.
The results for energy and material requirements for ITO transparent electrodes are presented in Tables XV and XVI. We have used a functional unit of 1 m2 of area with 180 nm of ITO and 25 nm of SiO2 layer. The considered substrate is fused quartz with a final thickness of 0.7 mm. This functional unit is based on the typical ITO transparent electrodes (Präzisions glass and optic) that we use in the preparation of our solar cells.
Table XV. Cumulative energy requirements for 1 m2 of ITO transparent electrode production.
Thermal energy (kJth)
Electrical energy (kJel)
Bold denotes the Total value.
SiO2 passivation (sputtering)
ITO ceramic target production
SiO2 target production
Table XVI. Input materials inventory for 1 m2 of ITO transparent electrode production.
Raw materials inventory
Argon (in normal litres)
Oxygen (in normal litres)
Processing of PEDOT:PSS
Poly-(3,4-ethylenedioxythiophene) (PEDOT) is an intrinsic conducting polymer which currently plays a dominant role in antistatic, electric and electronic applications 72. In polymeric photovoltaic cells it is used as hole injecting layer, it has also been investigated as an ITO replacement option for the transparent electrode 62, 73. The PEDOT is prepared using standard oxidative chemical or electrochemical polymerization methods and it was initially found to be an insoluble polymer. The solubility problem was subsequently circumvented by using a water-soluble polyelectrolyte, poly(styrene sulphonic acid) (PSS).
The commercial form of the polymer is an aqueous dispersion (1.3–3% of solid content) of the complex-mixture PEDOT:PSS. The ratio PEDOT:PSS depends on the requirements of the application (a balance between conductivity and solubility).
The investigated route in this study for the PEDOT:PSS processing is shown in Figure 6.
EDOT monomer is polymerized in the aqueous polyelectrolyte (PPS solution). We estimate that the oxidative chemical polymerization is the most common technique in the industry. The oxidizing agent can vary according to the polymerization procedure. We have followed a typical procedure presented in Reference 74 to approximate the input reactants.
The EDOT monomer is usually chemically synthesized from 3,4-dimethoxylthiophene, using ethylene glycol and p-toluene sulphuric acid as main solvents 74. 3,4-dimethoxylthiophene is obtained from 3,4-dibromothiophene, which is a brominated thiophene. The considered synthesis of 3,4 dibromothiophene from thiophene and bromine involves three bromination–debromination steps 49, 50.
According to Reference 75 the most frequently quoted process for the preparation of a lightly sulphonated polystyrene (PSS) is the method of homogeneous sulphonation of polystyrene with acetyl sulphate in a solution of dichloroethane described in Reference 76. Polystyrene is a common polymer of fossil origin and a life cycle inventory for this polymer can be found in Reference 36.
The results for energy and material requirements for PEDOT:PSS are presented in Tables XVII and XVIII. We have used a functional unit of 1 kg of aqueous dispersion of PEDOT:PSS (1:2.5 and 1.3 wt%), based on the typical PEDOT:PSS solution we use in the preparation of the solar cells. Note that only 13 g of this functional unit is the real product.
Table XVII. Cumulative energy requirements for 1 kg of PEDOT:PSS (1:2.5 1.3 wt%) processing.
Thermal energy (kJth)
Electrical energy (kJel)
Bold denotes the Total value.
Oxidative chemical polymerization
Halogen dance 3rd step
Halogen dance 2nd step
Halogen dance 1st step
Table XVIII. Input materials inventory for 1 kg of PEDOT:PSS (1:2.5 1.3 wt%) processing.
The two main processing stages for aluminium production are alumina refining (Bayer process) and aluminium smelting (Hall–Heroult process). The Hall–Heroult electrolytic process is the only process for Al production in commercial use today. The energy consumption for Al production from virgin metals is calculated as 211 MJ/kg in Reference 77. The last version of the inventory of the University of Bath 69 accounts 62 records for virgin Al production with an average of 224 MJ/kg. However, the recycling rates of Al are quite high (a 33% of recycled content, which is the worldwide average, was assumed from the IAI, International Aluminium Institute) and the involved energy for the production from recycled Al is much lower. As a reference we used the average value of 157.1 MJ/kg 69, which takes into account some records of production from recycled metal.
Calcium metal is typically produced by high-temperature vacuum reduction of calcium oxide in the aluminothermal process. A typical flow sheet for the process is shown in Figure 7. Large amounts of energy are required by this method, partially because of the high temperatures of the process and partially because of the energy-intensive raw materials employed, i.e. the calcinated CaO and electrolytically produced aluminium 52. For certain applications, especially those involving reduction of other metal compounds, better than 99% purity is required. This can be achieved by redistillation, which implies lower pressures and heating in the last step.
We have not found any available references about the energy requirements (MJ/kg of Ca) of this energetically intensive process. We take as simplification the involved energy production of Aluminium (from virgin raw materials) as similar (224 MJ/kg 69).
Material inventory and embedded energy (per square metre of module)
Our laboratory process allowed us to prepare organic solar cells of 1 cm2. This process is scalable, using the same methods (the main constrainer in the spin-casting procedure) to 100 cm2 cells, and the organic modules should be composed of cells of this size connected in series and/or parallel using additional metal contacts. For the final size of the modules, we consider a surface of 1 m2, with 90% coverage of active area of cells, we use this correction factor for the calculated material inventory of the final ‘organic’ module.
The organic photovoltaic technology will scale-up its production process not using spin-casting, which is an inefficient preparation method, but using roll-to-roll industrial processes, well known from the plastic industry. Methods such as ink-jet printing, spray pyrolysis, screen printing etc. will be less energy demanding and will make a better use of the raw materials that our spin-coating method. Therefore, we consider our material inventory and the embedded energy calculation as a cap for any future large-scale industrial process of an organic photovoltaic technology.
In Table XIX we show the calculated material inventory for our particular cells. The results are presented for a solar cell of 100 cm2 and a hypothetical module of 1 m2. The thicknesses are given in the second column of the table, which describes our particular cells. Similar values are found in the literature and the corresponding calculation will be quantitatively very similar to those shown in the table.
Table XIX. Material inventory for organic solar cell and module.
100 cm2 cell
1 m2 module (90% active area)
For clarity the front electrode components are presented separately, however the input material is the coated and pasivated transparent substrate.
1.3 wt% aqueous dispersion of PEDOT:PSS (1:2.5). The solid content is indicated in brackets.
For the module inventory, we have included the input material for encapsulation (typically a 0.5 mm layer of ethylene vinyl acetate (EVA) is considered) and for the frame (Aluminium) a typical value of 280 g/m2, typical for thin film modules 12, is assumed.
Finally, using the energy input of the cell fabrication method described in Section 3.1 (spinner, glove box, heater etc. energy consumption), the energy input required for the raw material processing given in Section 3.2 and the material inventory indicated in Table XIX, we can obtain the embedded energy per square metre of solar module, Table XX summarizes our results. These values are estimations of the involved energy during the extraction and processing of the different materials. Other factors, like energy in transport or involved energy inputs during decommissioning phase, are not included in the study. The depicted total will be used for the calculation of the energy payback time and for comparison with other photovoltaic technologies.
Table XX. Embodied energy (primary energy) for 1 m2 organic solar module (90% active area).
Equivalent primary energy (MJ)
The energy for oxygen production is included in the ITO clearing energy requirement.
As it is recommended in the reference literature, all the energy results are converted to Equivalent Primary Energy† per surface in order to make the comparison with previous PV energy requirement values easier.
The results are graphically presented in Figure 8. Figures 9 and 10 have been included to easily identify the most energetically intensive processes and materials.
Laboratory versus industrial production
The aim of our paper is to give a life cycle assessment calculation for laboratory production of organic solar cells. This fabrication procedure that we know in detail is the starting point for the calculations. Any step towards massive pre-industrial or industrial production will involve a much more efficient manufacturing procedure (in terms of number of cells output) and therefore the values given in our paper should be considered as a cap to any further LCA study regarding more automated industrial production.
Before any company can develop a final industrial process and reach the market with a commercial product some issues have to be addressed.
Flexible substrates: Conductivity, transmissivity and impermeability to oxygen and water are required. Inflexible ITO-coated glass substrates are typical for laboratory cells. ITO coated poly(ethylene terephthalate) (PET) is already used as a substrate for flexible amorphous silicon solar cells. However, the conductivity of the ITO layer decreases on plastic substrates and it has been shown that bending the device processed on a plastic substrate could damage the brittle inorganic oxide layers, thereby reducing their electrical properties 78. Highly conductive PEDOT as a transparent contact could be used as a replacement material for ITO.
Ease of deposition: For research devices, spin coating is commonly used, but this is uneconomical for large areas because of the material waste. Some of the following production techniques could be considered:
Doctor blading or ‘wire blading’: a rod is drawn over a line of the polymer solution, to spread it out into a thin sheet.
Screen printing: a fine mesh screen loaded with polymer solution is brought into temporary contact with the substrate and polymer solution flows across 79.
Inkjet printing and spray coating: A fine jet of polymer solution is sprayed on to the substrate in a required pattern 80. A simpler approach is printing polymer and contact materials using and embossed stamp 81. This may be compatible with reel-to-reel processing.
Patterning: An organic photovoltaic module implies series and/or parallel connected single cells. According to the deposition technique well designed patterns or masks must be investigated in order to avoid undesirable shortcuts.
Chemical stability: Inert and dry atmosphere for large-scale production is costly. Costs could be reduced understanding the chemical instability of the active layer and finding materials which are sufficiently stable so that devices can be prepared in ambient conditions before encapsulation.
Encapsulation: Coatings are needed which are sufficiently impermeable to moisture and oxygen. Typical encapsulation materials and techniques used in inorganic photovoltaic modules should be modified in order to offer flexibility and specific protection properties.
Processing temperature: High temperature processing stages are costly and damaging to organic films. Common metallization techniques used for large production of circuit boards and inorganic photovoltaic modules have to be revised.
Energetic and economic costs: Cost of each stage also has to be envisaged. Vacuum evaporation is the most expensive stage, since many expensive vacuum components are required in order to maintain high vacuum (∼10−6 torr). An alternative to vacuum deposition for small molecules is organic vapour-phase deposition (OVPD) 82.
Energy payback times and avoided emissions
The following paragraphs describe the calculation to obtain significant parameters that allow us to obtain a global view of the environmental assessment of the organic photovoltaic technology; these parameters will be compared with the published values for other well known photovoltaic technologies.
A metric used in evaluating the importance of embodied energy is the embodied energy coefficient, defined as the ratio of the embodied energy in GJ of equivalent primary energy to the generated peak power in kWp. For this calculation we have to use a nominal efficiency forecast for the power conversion that the industrial modules could yield. As explained in the Methodology section, we consider two values for the efficiencies of the organic photovoltaic modules: 5 and 10%. From these assumptions we therefore obtain two reference values of organic photovoltaic modules‡: 56.02 GJ/kWp and 28.01 GJ/kWp, respectively.
Also, we perform the Energy Payback Time (EPBT) for an organic PV module as the ratio calculated embodied energy to annual generated energy by the module. The generated energy is very much dependent on the irradiation and overall system performance. We assume the average South European irradiance, which is 1700 kWh/m2/year and a performance ratio of 0.8. The EPBT is additive, it means that EPBT values for the rest of the PV system components can simply be added up to obtain the total PV system EPBT. However, EPBT does not give an indication of the net energy balance over the module's life time. We introduce the Energy Return Factor, defined as the ratio life time of the module to EPBT 83. Table XXI shows the EPBT and ERF of the organic photovoltaic module, a module lifetime of 15 years is assumed.
Table XXI. Energy payback times and energy return factors of organic photovoltaic modules. Two values of nominal module efficiency are considered. Module utilization is assumed for the South Mediterranean situation.
Embodied energy in 1 m2 module (MJ)
Equivalent primary energy annual saved by electricity supplya from 1 m2 module (MJ/year)
The conversion efficiency from primary energy to electricity for the end-user is assumed 35%.
The CO2 emissions due to the production of a PV module can be obtained by multiplying all energy and material inputs with their corresponding CO2 emission factor. Most of the energy inputs to the module manufacturing are electricity inputs, therefore the equivalent primary energy is converted to electricity. The CO2 emission factor for electrical energy is highly dependent of the fuel mix of the considered utility system; we have used the average European electricity mix (411.44 g-eq. CO2/kWhel in 2006 84, 85. We obtain a CO2 emission per kWp of organic solar module given by 1120.35 and 2240.70 kg of eq.-CO2/kWp (for 10 and 5% of efficiency, respectively).
Another significant parameter for any electricity generation technology is the CO2 emission factor (equivalent CO2 emissions per kWh of generated electricity). It is calculated as the total CO2 emissions due to the embodied energy of the module (Embodied CO2) divided by the total generated electricity by the module during its lifetime (Table XXII).
Table XXII. CO2 emission factor of organic photovoltaic modules. Two values of nominal efficiency are considered. The 1 m2 module has 90% of active area and its utilization is assumed for the South Mediterranean situation (1700 kWh/m2/year of irradiance), a life time of 15 years and a performance ratio of 0.8).
Embodied CO2 (kg eq-CO2)
Generated electricity during the module life time (kWhel)
CO2 emission factor (g eq.-CO2/kWhel)
Comparison of organic solar cells with other PV technologies
Numerous studies have been carried out to estimate the energy consumption in the manufacturing of solar PV modules. In order to compare them with the results for organic solar modules, we present in Table XXIII some of the values from the most relevant studies. Note that we have included only those studies where the embodied energy is referred only to the process energy and energy embedded in consumed materials of the PV module (for cases where frame is not included, the estimation for energy frame is indicated). The values have been converted to equivalent primary energy per module rated peak power (kWp) and per m2 of module, indicating always the considered efficiency for the PV technology.
Table XXIII. Energy requirements in equivalent primary energy (GJ) per kWp for the actual PV technologies.
The module area for 1 m2 of active area is 1.43 m2, we take 10% efficiency of active area (7% module efficiency)
Laboratory organic PV
5% module efficiency
Laboratory organic PV
10% module efficiency
From the embodied energy values for different PV modules technologies, we compare now the EPBT for them in Figure 11. The Mediterranean irradiance conditions (1700 kWh/m2/year), a performance ratio of 0.8 and 0.35, conversion efficiency from primary energy to electricity are assumed.
Organic photovoltaic technologies in general, and full organic polymeric technology in particular compare favourably with other PV technologies regarding the embedded energy of the PV module even for a far from optimum laboratory fabrication procedure. The main reason is that there are not high temperatures involved in the process. For laboratory cell production, the final embedded energy per square metre of module is 2800.79 MJ/m2. This value is half of the average value calculated for crystalline silicon technologies, it is of the same order of magnitude of thin film technologies and slightly higher than dye-sensitized solar modules. We can expect bigger reductions for the organic technology in the large scale industrial process, therefore, giving a clear advantage to organic technologies provided that the efficiency of these industrial modules is similar to that of actual laboratory cells (about 5% now and an expected 10% in the coming years).
For a typical organic solar module, the energy embodied in the materials is 726.26 MJ/m2 including both the materials of the solar cell and the materials used in the process. The direct process energy is 1973.78 MJ/m2, which is more than double, indicating that there is plenty of room for further reduction if the optimization of an industrial production process is accomplished. In particular, the nitrogen use is a key factor: in the laboratory-fabrication process it accounts for 48.19% of the embodied energy as an input material and for 38.56% of the process energy required to keep the N2 atmosphere in the globe box. The spin-casting method widely used in the laboratory is very inefficient regarding the usage of materials. More than 99% of the polymer is misused with this coating technology. Also the lack of control of the nanostructure of the spin cast layer should be avoided. This result makes compulsory the shift to ink-jet, spray-coating or screen-printing technologies for a scale-up in production.
The fabrication of the electrodes is the most energy costly factor. For ITO, it accounts for 50.39% of total material embodied energy, which adds to the fact that Indium is a scarce and expensive material. Alternative materials for cathodes are therefore needed to make this technology more competitive, the problem risen by the use of ITO in organic low-cost technologies is also enhanced by the fact that it accounts for more than half of all the materials embodied energy. The evaporation of Calcium and Aluminium for the anode fabrication accounts jointly for 33.77% of direct process energy input in the fabrication of laboratory cells. Alternative methods for anode deposition should be developed.
We point out that throughout our calculations of estimated energy output of the systems the distinction between direct and diffuse irradiation has not been taken into account. When a crystalline silicon PV system is compared with organic PV for lower levels of irradiance, a higher ratio of diffuse versus direct solar irradiance is produced and the output of the organic PV technology is underestimated if this issue is not taken into account. If low irradiance levels are considered, the organic technologies are less energy-costly per installed capacity than any other PV technology even with actual laboratory processing methods and power conversion efficiencies.
Finally, we would like to include a few comments regarding the applications of organic photovoltaic modules. The companies that are about to begin the commercialization of this technology are aimed to a market of gadget applications or building integration. Nevertheless, the main interest of this technology should swiftly move towards applications in Solar Home Systems (SHS 89) for rural electrification in rural livelihoods 90. SHS are strongly in need of low-cost technologies, which enable for massive deployment in developing countries. A huge market could be expected for this technology in the near future. The reduced cost of this technology as well as their lower environmental impact regarding the embedded energy could be a significant breakthrough to create a market for this organic technology in rural electrification.
The authors wish to thank Nicoletta Marigo and Chiara Candelise (ICEPT, Imperial College and UK Energy Research Centre) for enlightening and helpful discussions. Also the financial support from the Spanish Ministry of Science and Innovation through projects MAT2006-12970-C02-02 and HOPE-CSD2007-00007 (Consolider Ingenio) and from the Comunidad Autónoma de la Región de Murcia (CARM-D429-2008) and grants FPU AP2005-2271 (R. G.-V.) and PR2008-0272 (A. U.) are acknowledged.
For the conversion efficiency for electricity production we used a value of 0.35. In case thermal energy consumption we used 0.80 for the conversion efficiency to primary energy equivalent.