Can photosynthesis enable a global transition from fossil fuels to solar fuels, to mitigate climate change and fuel-supply limitations?

Renewable and Sustainable Energy Reviews 62 (2016) 134–163

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews journal homepage: www.elsevier.com/locate/rser

Can photosynthesis enable a global transition from fossil fuels to solar fuels, to mitigate climate change and fuel-supply limitations? Andrew K. Ringsmuth a,b,1, Michael J. Landsberg a,2, Ben Hankamer a,n a b

The University of Queensland, Institute for Molecular Bioscience, St Lucia, Queensland 4072, Australia The University of Queensland, ARC Centre of Excellence for Engineered Quantum Systems, St Lucia, Queensland 4072, Australia

art ic l e i nf o

a b s t r a c t

Article history: Received 24 September 2015 Accepted 6 April 2016

This review article considers Earth as an energy-storing (photosynthetic) and energy-consuming (metabolic) system. We evaluate whether and how photosynthetic, solar fuel-production systems can be engineered and deployed sufficiently rapidly to supplant enough fossil fuel supply to sustain a complex human economy and natural ecosystems over the long term. Geophysical, ecological, economic, technological and political constraints are quantified. We consider the potential to innovate and scale up promising systems such as microalgal and artificial photosynthetic systems to economic viability within a time frame meaningful for mitigating the effects of climate change and fuel-supply limitations. A future global society powered sustainably by solar fuels is forecast to require increased global photosynthetic productivity, through increased photon-conversion efficiency and production area. Increasing the efficiency of socioeconomic energy utilisation is also important. Meeting these challenges on the required time scale demands historically unprecedented technical progress, highlighting the need for both advanced international policy frameworks and scientific excellence. Based on evidence from a broad range of fields, a multiscale systems optimisation approach is identified as important, to integrate analyses from the scale of the global climate, economy and energy systems, down to the nanoscale of light-harvesting and carbon-fixing machinery that drives photosynthesis. & 2016 Elsevier Ltd. All rights reserved.

Keywords: Photosynthesis Solar fuel Climate change Fuel security Light harvesting Multiscale analysis

Contents 1. 2.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Global energy systems and society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Earth’s solar energetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1. Solar radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2. Earth’s radiation fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Photosynthetic primary production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Global net primary production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4. Areal productivities and fuel production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5. Fossil fuels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1. Coals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.2. Crude oils (petroleum) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

135 136 136 136 137 137 138 139 139 139 139

Abbreviations: PAR, Photosynthetically active radiation; AM0, Air Mass 0 reference solar spectrum; AM1.5, Air Mass 1.5 reference solar spectrum; TPES, Total primary energy supply; GPP, Gross primary production; NPP, Net primary production; HHV, Higher heating value; 1P reserves, reserves with a 90% probability of recovery; URR, Ultimately recoverable resources; IEA, International Energy Agency; HANPP, Human appropriation of NPP; NPPLC, NPP lost due to land conversion; NPP0, Total potential NPP; NPPact, Actual NPP; NPPh, NPP harvested; NPPt, NPP not appropriated by humans; PPR, Primary production required; GDP, Gross domestic product; NGL, Natural gas liquid; CTL, Coal to liquid; GTL, Gas to liquid; EROI, Energy return on energy investment; EV, Electric vehicle; LHC, Light-harvesting complex; PPC, Pigment–protein complex; EET, Excitation energy transfer; PSI, Photosystem I; PSII, Photosystem II; BTL, Biomass to liquid; GWP, Global warming potential; NPQ, Nonphotochemical quenching; HVP, Highvalue product; LCA, Life cycle analysis; GHG, Greenhouse gas; AP, Artificial photosynthesis; OEC, Oxygen-evolving complex; XFEL, X-ray free-electron laser n Corresponding author. Tel.: þ 61 7 334 62012. E-mail address: [email protected] (B. Hankamer). 1 Current address: VU University Amsterdam, Department of Physics and Astronomy, and Solardam Amsterdam Solar Energy Research Initiative, De Boelelaan 1081, 1081 HV Amsterdam, The Netherlands. 2 Current address: The University of Queensland, School of Chemistry and Molecular Biosciences, St Lucia, Queensland 4072, Australia. http://dx.doi.org/10.1016/j.rser.2016.04.016 1364-0321/& 2016 Elsevier Ltd. All rights reserved.

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

2.5.3. Natural gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photosynthates in fossil-fuelled society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.1. Human appropriation of current net primary production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.2. Human appropriation of terrestrial net primary production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.3. Human appropriation of aquatic net primary production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.4. Human appropriation of ancient net primary production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.5. Limits to sustainable HANPP, and bioenergy production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.6. Sustainable HANPP for agro-bioenergy production in terrestrial ecosystems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.7. Sustainable HANPP in aquatic ecosystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7. Time scales of constraints to fossil-fuelled society . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7.1. Anthropogenic climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7.2. Fuel supply limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8. Transitioning to energy systems based on current photosynthetic primary production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8.1. The importance of fungible fuels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8.2. Time scales of global energy transitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. Photosynthetic energy systems for solar fuel production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Photosynthesis in higher plants and green algae. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Solar fuels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1. Ethanol . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2. Biodiesel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3. Biogas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.4. Hydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.5. Hydrocarbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Photosynthetic energy systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1. Kinetic rate limitations to productivity in biological photosynthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.2. Limitations to supply and utilisation of CO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.3. Limitations to supply and utilisation of light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Microalgal cultivation systems – potential as scalable energy systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1. Areal productivities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.2. Global-scale resource constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.3. Technoeconomic and lifecycle analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.4. Key innovation pathways for microalgal cultivation systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5. Artificial photosynthetic systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.1. Key innovation pathways for artificial photosynthetic systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Grand challenges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Engineering next-generation catalysts for water splitting and carbon fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.1. H2O oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.2. H2 synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1.3. Carbon fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Light harvesting in the chloroplast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.1. Quantum coherence in excitation energy transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2. Nonphotochemical quenching mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.3. Structure determination of the light-harvesting machinery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Complex-systems analysis and optimisation of photosynthetic systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Policy frameworks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Summary and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.

1. Introduction Photosynthesis is nature's solar energy conversion technology, refined through natural selection over approximately three billion years in higher plants, eukaryotic algae, a few genera of bacteria and a few species of animals that have harnessed these processes for their own benefit [1–3]. The photosynthetic machinery captures energy from electromagnetic radiation and stores it as chemical energy by breaking and creating molecular bonds against the chemical equilibrium. In biological systems, oxygenic photosynthesis builds relatively energydense organic products such as carbohydrates and lipids out of relatively diffuse radiation, carbon dioxide (CO2) and water (H2O) from the environment, while oxygen (O2) and heat are emitted as by-products (see Section 3.1). This process has been responsible for almost all carbon fixation and oxygen evolution on Earth. It therefore provides most of the energy driving the biosphere and (through fossil fuels) ultimately also the human economy. Due to growing concerns over fuel security

135

140 140 140 140 140 140 141 141 141 141 141 142 143 143 144 144 144 145 145 145 146 146 146 146 146 147 147 149 149 150 151 151 152 152 153 153 153 154 154 154 155 155 155 156 157 157 158 158

and the consequences of anthropogenic climate change, there is increasing interest in the potential for renewable solar fuel-production technologies to meet society’s energy needs sustainably. Such technologies may harness higher plants, algae, cyanobacteria, artificial photosynthetic systems or combinations of these. This review article considers Earth as an energy-storing (photosynthetic) and energy-consuming (metabolic) system. The coupling between these processes is taken to have two components: (1) a short time-scale component through which current photosynthetic systems supply primary photosynthates to energise natural ecosystems and a fraction of the human economy (e.g. food, bioenergy); and (2) a long time-scale component through which photosynthetically-derived fossil fuels meet most of the human economy’s energy needs. The central question addressed is whether and how photosynthetic systems can be engineered and rapidly deployed to supplant enough fossil fuel supply to sustain a complex human economy while also sustaining natural ecosystems over the long term. Geophysical, ecological, economic,

136

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

technological and political constraints to this possibility are quantified. Microalgal cultivation systems and artificial photosynthetic systems are identified as promising photosynthetic energy technologies. We consider the potential for innovating these to economic scalability within a time frame meaningful for mitigating climate change and fuel-supply limitations. The aim is to bring together information from many fields in a synthesis that is useful for researchers and policymakers. Wherever possible in our analysis of global systems, we focus on data for the year 2010. This is for consistency and because it is both recent enough to be relevant and long enough ago to have been well studied from

diverse perspectives. Salient numbers later described in detail are collated in Table 1.

2. Global energy systems and society 2.1. Earth’s solar energetics 2.1.1. Solar radiation Photosynthesis on Earth is constrained by available sunlight. The Sun’s radiant power or luminosity is 3.846  1026 W [32]. Its emission

Table 1 Photosynthesis in global energy systems. Energy flow Energy flow (ZJ a  1) Carbon flow (W) (GtC a  1)

Quantity

Solar irradiance Global

Local, peak power density

Top of atmosphere Earth’s surface PAR at surface

1.74  1017 9.60  1016 4.12  1016

Top of atmosphere 1366 m  2 (AM0) Earth’s surface (AM1.5) 1000 m  2 PAR at surface 430 m  2

Energy stock (ZJ)

Carbon stock (GtC)

5490 3020 1300

References

[4] [4] [4,5]

4.31  10  11 m  2  11

[6]

2

3.16  10 m 1.36  10  11 m  2

[7] [5,7]

188 m  2 81 m  2

5.93  10  12 m  2 2.56  10  12 m  2

Net photosynthetic primary production Global NPP Terrestrial NPP Cropping NPP Forestry NPP Grazing land NPP Aquatic NPP

1.4  1014 6.8  1013 6.2  1012 3.2  1013 1.2  1013 7.4  1013

4.5 2.1 0.20 0.52 0.38 2.3

110 55 5.0 13 9.8 55

[8–10] [8,9] [9] [9] [9] [10]

Primary energy supply Total (2010)

1.63  1013

0.503

10.0 (in CO2)

[11,12]

Fossil fuels total

1.36  1013

0.438

Oil Coal Natural gas Biomass fuels

5.36  1012 4.72  1012 3.80  1012 1.06  1012

0.169 0.149 0.120 3.35  10  2

9.1 (incl. cement production) 3.28 3.64 1.65 0.903

9.89  1011

3.13  10  2

0.844

[18]

Total 6.85  1010 Bioethanol 5.13  1010 Biodiesel 1.71  1010

2.16  10  3 1.62  10  3 5.40  10  4

3.95  10  2 2.84  10  2 1.11  10  2

[18] [18] [18]

0.606 0.302 6.42  10  2 0.173 0.147 2360 0.044–0.133

15.6 7.77 1.65 4.45 3.46 6.06  104 1.19–3.59

[9] [9] [9] [9] [21,22] [23,24,11] [20]

Surface-averaged, timeaveraged power density with average atmospheric losses

Earth’s surface PAR at surface

Total Traditional fuels Refined agrobiofuels

biomass

Human appropriation of net primary production Total terrestrial Cropping Forestry Grazing land PPR to support world fisheries Ancient NPP (via fossil fuels) Estimated sustainable limit to dedicated agrobiofuel cropping Atmospheric carbon dioxide Current Estimated safe limit Estimated limit to cumulative emissions from 2000 to 2050 with confidence P 40.9 Quantity that would be generated by complete combustion of fossil fuels total URR

(HANPP) 1.92  1013 9.57  1012 2.03  1012 5.48  1012 4.66  1012 7.48  1016 1.39– 4.21  1012

4.2 (2 ppmv)

1P 32.6

URR 96.9

1P 855

URR 1844

7.90 17.7 7.06

31.1 28.1 37.7

160 593 102

581 718 545

[11,12]

[11,13,14] [11,15,16,13,17] [11,14,18,19] [20]

833 (392 ppmv) o 956 (o 450 ppmv) 170

[25,26] [27,28] [29]

6215

[14]

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

137

Fig. 1. Terrestrial solar irradiance and photosynthetic absorption spectra. AM0 and standard solar spectra (see Section 3.2) are shown. Atmospheric absorption bands are visible in the AM1.5 spectrum. Inset shows in vivo absorption spectra for pigments from higher plants and green algae. Inset adapted from [30]. Overall figure adapted from [30,31].

spectrum (Fig. 1) is well approximated by a blackbody at temperature 5780 K, with a peak in the green at  500 nm [7,33]. Most solar energy (52%) is in the visible spectrum (390–750 nm), with a large fraction (42%) also in the infra-red (IR) (0.7–300 mm). Ultraviolet (UV) radiation (10–390 nm) accounts for most of the remaining 6% [7]. The spectrum conventionally defined as photosynthetically active radiation (PAR) is 400–700 nm [5] (Fig. 1 inset), which corresponds closely but not exactly to the visible spectrum and contains 43% of the total solar energy in the commonly used Air Mass 1.5 (AM1.5) reference spectrum (Fig. 1 main, black curve). This corresponds to spectral filtering through 1.5 cloudless atmospheric depths, delivering  1000 W m  2 of irradiance [7]. An estimate of time-averaged solar irradiance at the top of Earth’s atmosphere (the ‘solar constant’) can be obtained from the quotient of the Sun’s luminosity to the surface area of a sphere with radius equal to Earth’s average distance from the Sun (149.6 g [34]), or 1,367.5 W m  2. Comparatively, time-averaged satellite measurements [6] give  1366 W m  2. The total power incident on the cross-sectional area of Earth (mean radius 6371 km) before atmospheric interference can therefore be calculated to be 174 PW. Over one year this delivers 5490 ZJ of energy. In 2010 this represented 11,000 times the total primary energy supply3 (TPES) of the global economy, which was 0.503 ZJ [11]. 2.1.2. Earth’s radiation fluxes 2.1.2.1. Transmission to surface. The solar spectrum incident at the upper atmosphere is known as the Air Mass 0 (AM0) spectrum (Fig. 2 main, white curve) [7]. The atmosphere and clouds reflect approximately 26% of solar irradiance back into space and absorb a further 19% of the total [4], leaving 3020 ZJ a  1 (96 PW average) available at Earth’s surface. Approximately 80% (2420 ZJ a  1 or 77 PW) is incident on the oceans, which cover 70.8% of the planet’s surface [33,35]. Twenty percent (600 ZJ a  1 or 19 PW) is incident on land (29.2% of the surface) [33]. Currently, Earth’s surface 3 ‘Primary energy’ refers to energy harvested from nature, prior to any humaninduced conversions. Examples include raw fossil and nuclear fuels, solar radiation, biomass sources and wind.

reflects around 4% of the total energy incident at the upper atmosphere [4]. 2.1.2.2. Surface insolation. Solar irradiance at Earth’s surface (termed insolation) varies with location and time. Radiation falling within the planet’s cross-sectional area is projected onto its approximately spherical surface, reducing average irradiance by a factor of four to 342 W m  2, even without atmospheric interference [33]. Including atmospheric effects, average insolation is reduced to 188 W m  2, although instantaneous, local insolation varies widely. The AM1.5 spectrum is a commonly used standard for midday insolation and, at the opposite extreme, night-time irradiance does not exceed 2.4  10  4 W m  2 under a full moon on a clear night [36]. Daily, local insolation (averaged over 24 h) depends on location and weather, and typically ranges from 12 to 405 W m  2 [37]. Annual, local insolation in Australia, averaged over one year at different sites, falls between 69 and 278 W m  2 [37]. These time-averaged, local insolation values are an order of magnitude lower than typical power production densities of thermal power plants, such as coal-fired or nuclear [33]. Furthermore, efficiency losses during solar energy conversion through existing photosynthetic, photovoltaic or solar–thermal systems widen the disparity by  1–2 orders of magnitude. These comparably low power densities and the intermittency of solar power currently challenge the economic competitiveness of solar energy, despite its abundant total supply across Earth's surface [33]. Technologies for concentrated, stable storage of solar energy, for example in chemical fuels, are needed to address this challenge. 2.2. Photosynthetic primary production Primary production is chemical synthesis of new biomass from inorganic precursors. Though this may occur through either photosynthesis or chemosynthesis, the latter happens on much smaller scales and in largely inaccessible areas such as the deep oceans. Primary production and photosynthetic primary production are therefore often referred to interchangeably, as is done here. Three conventions for quantifying primary production are dry biomass, assimilated carbon (C) and stored energy. Gross primary production

138

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

(NPP), which is the most commonly used metric of primary production. NPP represents the resource of photosynthates available to other organisms, including humans [9,38]. The rate of primary production (usually annualised) – primary productivity – depends on both the environment and organism (see Section 3.3), and is ultimately limited by PAR insolation. Approximately 43% of the 1000 W m  2 insolation under AM1.5 conditions (430 W m  2) falls within the spectral range absorbed by chlorophyll and carotenoid chromophores and is thus referred to as photosynthetically active radiation [7] (Fig. 1, inset). Since global insolation delivers 3020 ZJ a  1, a simple estimate for the energy available to drive biological photosynthesis is 0.43  3020 ZJ a  1 ¼1300 ZJ a  1. However, actual global primary production is limited by geographical variations in climate, availabilities of water, CO2 and nutrients, type and abundance of photosynthetic systems (e.g. plant species), and energy losses within the photosynthetic machinery (see Section 3.3.3). 2.3. Global net primary production

Fig. 2. Global energy resources and consumption. Cubic volumes correspond to magnitudes of labelled quantities. The volume of the blue cube was calculated as follows: According to the International Energy Agency (IEA), in 2010, 18% of global final energy consumption was supplied by electricity, 2% by heat, and the remaining 80% directly by fuels [50]. Assuming efficiencies of 38% for electricity generation [11], and 93% for electricity transmission and distribution [74], gives an estimate that 37% and 2% of global primary energy respectively are currently used to supply final electricity and heat consumption; the remaining 61% of primary energy supply, therefore, meets demand directly with chemical fuels. Further assumptions: (1) 95% is taken as a representative value for transportation and distribution efficiency of fuels, based on [75]; and (2) transmission and distribution efficiency of heat is taken to be 90%, based on [76]. Horizontal bars show decompositions of global bioenergy consumption (light green), global fuel consumption (blue), total primary energy supply (red), global net primary productivity (dark green), 1P reserves (light grey), and ultimately recoverable resources (URR – dark grey). Abbreviations: ‘Trad. Bm’ – traditional biomass, ‘Mod. Bm’ – modern biomass, ‘E’ – ethanol, ‘D’ – biodiesel, ‘Bm’ – biomass, ‘R’ – all sources of renewable energy combined, ‘Ur’ – uranium, ‘Wild terrest.’ – wild terrestrial, ‘Cr’ – cropping, ‘Frst’ – forestry, ‘Grz’ – grazing land. ‘URR’ – ultimately recoverable resources (i.e. technically recoverable down to a 5% certainty at current fuel prices). ‘1P’ – 1P reserves having a 90% probability of recovery. Uranium reserves (1.59 ZJ) and resources (5.04 ZJ conventional; 2.79 ZJ unconventional) are given by [77]. Traditional biomass is defined as biomass consumption in the residential sector in developing countries and refers to the often unsustainable use of wood, charcoal, agricultural residues and animal dung for cooking and heating [18]. Modern biomass refers to consumption of biomass fuels other than traditional biomass fuels, ethanol or biodiesel. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(GPP) is the total amount of the chosen measure stored after losses due to photorespiration during the dark reactions of photosynthesis. Autotrophic respiration reduces GPP to give net primary production

NPP is difficult to measure on geographical scales. Estimates are increasingly satellite-based and incorporate data from more traditional land-based sources. A recent analysis [8] of nearly 30 years of satellite and weather data [39,40] found that global terrestrial NPP has been stable over that period, with less than 2% annual variation despite deforestation and increased agriculture. Mean global NPP was estimated at 53.6 GtC a  1, in reasonable agreement with another influential study [9] that employed more land-based methods, combining vegetation modelling, agricultural and forestry statistics and geographical information systems data on land use to obtain an estimate of 59.22 GtC a  1. A metaanalysis of literature values from 1862 to 2011, for global terrestrial NPP, found mean 7standard deviation and median values of 56.27 14.3 GtC a  1 and 56.4 GtC a  1 respectively. Annual, terrestrial NPP is therefore here estimated to be  55 GtC. Based on approximate conversion factors4 for dry phytomass, 1 kg dry phytomass  0.45 kgC  17.5 MJ [33,41] (compared with 1 kg dry woody phytomass  0.5 kgC  19 MJ [33,41–43] – see Table 2). Annual, terrestrial NPP corresponds to 2.1 ZJ, which is approximately four times global TPES in 2010. Approximately 30 GtC a  1 is the above-ground fraction of NPP, with an energy content of 1.1 ZJ [44], equal to roughly twice global TPES in 2010. Recent studies of aquatic NPP place it within a similar range, 557  5 GtC a  1 [10]. This equates to annual energy storage of 2.370.2 ZJ a  1, assuming aquatic biomass energy density and carbon content of 20 MJ kg  1 [45] and 0.47 kgC kg  1 [46] respectively, as for typical green microalgae (see Table 2). Annual, global NPP is therefore estimated at 4.4 ZJ (110 GtC), which exceeds global TPES ninefold (based on 2010 figures) [11]. On these estimates, photosynthetic biota currently store only 0.3% of the  1300 ZJ of PAR annually incident at Earth’s surface, or 0.1% of the 3020 ZJ of annual, global insolation. Global NPP is distinct from the global long-term biological CO2 sink because the former is mostly counterbalanced by heterotrophic respiration and fires, leaving only a small fraction to enter long-term carbon-storage reservoirs such as permanent forests and the deep ocean. Currently, the net global terrestrial and oceanic sinks are approximately 2.5 GtC a  1 [47] and 2.3 GtC a  1 [26] respectively. The total sink, 4.8 GtC a  1 equates to 4.4% of global NPP and 48% of anthropogenic CO2 emissions in 2010 (see Section 2.8.1). 4 For energy density the conventional ‘heat of combustion’ or ‘higher heating value’ (HHV) is used here. This is the energy released as heat when a compound undergoes complete combustion with oxygen at standard ambient temperature and pressure.

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

139

Table 2 Photosynthate fuel properties at standard ambient temperature and pressure (STP). Hydrogen, methane and ethanol are pure compounds with invariant properties at STP. All other fuels listed are variable mixtures of compounds and their stated properties therefore represent a range. HHV refers to ‘higher heating value’, defined in Section 2.3. Fuel

Phase at STP

Composition

Mass density (kg L  1)

Energy density (HHV) (MJ kg  1)

Carbon content (kgC kg  1)

CO2 emissions (kg kg  1)

CO2 emissions (kg MJ  1)

References

Hydrogen Methane Natural gas Biogas Crude oil Gasoline Kerosene-type aviation fuel (Jet A-1) Petrodiesel Biodiesel

Gas Gas Gas Gas Liquid Liquid Liquid

H2 CH4 CH4, C2H6, etc. CH4, CO2, etc. Hydrocarbons (C5–C20) Hydrocarbons (C4–C12) Hydrocarbons (C6–C16)

9.0  10  5 6.6  10  4 7.8  10  4 4 6.6  10-4 0.88 0.75 0.81

142 55.6 52.2 o 55.5 45.5 46.5 43.1

0.0 0.75 0.75 o 0.75 0.85 0.85 0.86

0.0 2.75 2.75 2.75 3.12 3.12 3.37

0.0 4.95  10  2 5.3  10  2 44.95  10  2 6.85  10  2 6.70  10  2 7.31  10  2

[51–53] [54] [53] [54–56] [33,51,53] [51,57,58] [59–62]

Liquid Liquid

Hydrocarbons (C8–C21) Long-chain, mono-alkyl esters C2H5OH Macerals, ash, H2O Macerals, ash, H2O Cellulose, lignin, ash Cellulose, lignin, ash Woods, charcoal, agricultural residues, dung, ash Proteins, carbohydrates, lipids, ash

0.84 0.88

45.8 37.5

0.87 0.77

3.18 2.82

6.94  10  2 7.52  10  2

[51,58,63] [57,64]

0.79 1.5 1.3 0.6 0.6 0.6

29.6 30 15 19 17.5 18

0.52 0.9 0.5 0.5 0.45 0.49

1.91 3.30 1.84 1.83 1.65 1.79

6.45  10  2 0.11 0.12 9.6  10  2 0.11 9.9  10  2

[57,65] [33,66–68] [66,67] [33,69,70] [33,71] [33,72]

0.6

20

0.47

1.72

8.6  10  2

[45,46]

Ethanol Coal (Anthracite) Coal (Lignite) Dry wood Dry phytomass Traditional biomass fuel

Liquid Solid Solid Solid Solid Solid

Dry green microalgal biomass

Solid

2.4. Areal productivities and fuel production Mean areal productivities of terrestrial and aquatic NPP across Earth’s surface respectively are estimated at 0.45 W m  2 of icefree land and 0.13 W m  2 of ocean [33]. However, these vary with geography and ecosystem type. Mean areal productivity of tropical forests is estimated at 1.3 W m  2, while temperate and boreal forests respectively average 0.5 W m  2 and 0.2 W m  2 [33]. The highest annual-average areal productivity of any known vegetation is 5.0 W m  2, for natural stands of the C4 grass, Echinochloa polystachya on the Amazon floodplain [48]. This sets a benchmark for engineered photosynthetic energy systems. Agriculture, which now appropriates 38% of Earth’s ice-free land area [8,49], often delivers lower NPP than the natural ecosystems it replaced but concentrates growth in biomass components valued by humans [8]. Global cropland accounted for 9.5% of global terrestrial NPP in the year 2000, while forestry and grazing land accounted for 24.8% and 18.1% of terrestrial NPP, respectively [20]. Crop productivities differ substantially between species, conditions and cultivation methods but average global areal productivity is estimated at 0.4 W m  2 [33]. NPP constrains metabolism of photosynthates within ecosystems and the economy. Humans appropriate photosynthates for many uses beyond their indispensible role as food, and traditional biomass fuels (defined in Fig. 2) remain an essential driver of economic activity in pre-industrial societies. However, in 2010 such fuels supplied only 6.2% of TPES [50]. Most primary energy in the modern, industrialised world comes from a large repository of geochemically processed, ancient photosynthates: fossil fuels [18]. 2.5. Fossil fuels Under favourable conditions during some geological periods, photosynthates entered geological processes that initiated longterm storage of solar energy in finite stocks of fossil fuels. These fuels have powered human civilisation’s rapid expansion since the industrial revolution and in 2010 accounted for 87% of TPES [11]. Table 2 compares physicochemical properties of different fuels and Fig. 2 compares current global energy consumption with available resources.

2.5.1. Coals Coals are sedimentary rocks composed of heterogeneous organic compounds derived from woody phytomass, minerals and water. Collectively, coals are thought to be the most abundant fossil fuel, with 17.7 ZJ of proven (‘1P’) reserves in 2010 [11,50] and worldwide ultimately recoverable resources (URR) of coal estimated at 28.1 ZJ [14,17]. Coal compositions span a wide continuum that is conventionally divided into four classes by gravimetric energy and carbon content (in decreasing order): anthracite, bituminous, sub-bituminous and lignite. Ash is the collective term for the incombustible minerals, which by mass range from negligible to 40% [33,66]. In 2010, humans harvested 0.149 ZJ of energy, or 29.6% of TPES, from coal [11]. Sixty-five percent of coal consumption was for the purpose of generating electricity, and coal was the largest source of electrical power at 41% [50]. One fifth of coal was consumed by industry and the remainder mostly in the agriculture and buildings sectors [50]. Coal contributed 44% of fossil-fuel CO2 emissions in 2010 at 13.8 GtCO2 (3.9 GtC), making it the largest source of emissions [50,73]. 2.5.2. Crude oils (petroleum) Crude oils are liquid mixtures of hydrocarbons with various structures and chain lengths (generally C5–C20). The chemical composition of crude oil is particular to each deposit, with its own unique geological history. Natural crude oils arose following fossilisation of large quantities of aquatic organisms, such as algae [33,78]. In 2010, proven global reserves of oil, conventional and unconventional, totalled 7.90 ZJ [11]. Ultimately recoverable resources of conventional and unconventional oil were respectively estimated at 10.0 ZJ and 21.1 ZJ [14,17]. Global oil consumption in 2010 was 0.169 ZJ, accounting for 33.6% of TPES and constituting the largest source of primary energy [11]. In 2011, 54% of oil consumed was used in the transport sector [50]. Globally this sector is powered  95% by oil-based fuels [79,80]. The industry and buildings sectors were the nextlargest oil consumers in 2011, each with an approximately 8% share [50]. Oil contributed 36% of fossil-fuel CO2 emissions in 2010, at  11 GtCO2 (3.2 GtC) [73].

140

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

2.5.3. Natural gases Energy-dense natural gases are predominantly mixtures of the three smallest alkanes, methane (CH4) (73–95%), ethane (3–13%) and propane (0.1–1.3%), though they can also contain butane, pentane and trace amounts of larger alkanes. Contaminants are present as hydrogen sulphide, nitrogen, helium and water vapour [33], lowering the energy density, which is typically  52.2 MJ kg  1 [53], compared with 55.6 MJ kg  1 for pure CH4 [53]. Natural gases are formed by the same processes that form crude oil, though gas genesis requires temperatures and pressures sufficiently high to crack longer hydrocarbons into short-chain hydrocarbons (oC5). In 2010, proven natural gas reserves were 7.06 ZJ [11]. Ultimately recoverable resources of conventional and unconventional natural gas were respectively estimated at 15.6 ZJ and 22.1 ZJ [14,17]. The world consumed 0.120 ZJ of energy from natural gas, or 23% of TPES [11] in 2010. The sector with the highest demand was electricity generation at 39% [50]. The second largest gasconsumption component was from the buildings sector, followed by industry [50]. In the same year, natural gas combustion emitted 6.2 GtCO2 (1.8 GtC), 20% of global emissions [73]. 2.6. Photosynthates in fossil-fuelled society 2.6.1. Human appropriation of current net primary production In 1973, Whittaker and Likens reported the first prominent estimate of the total biomass consumed by humans, accounting for only the harvest of food and wood used directly by humans, and concluded that this appropriated 3% of Earth’s annual NPP [42,81]. Subsequent developments in the study of human appropriation of net primary production (HANPP) [9,38,82–84] have produced a conventional definition (Fig. 3) [42] that includes two interdependent processes: (1) land-use changes that modify the NPP of vegetation (ΔNPPLC) given by the formula, ΔNPPLC ¼NPP0–NPPact where NPPact is the actual NPP of the human-modified land and NPP0 represents the vegetation that would have persisted in the absence of human interference; and (2) extraction or destruction of a share of NPP for human purposes (ΔNPPh), such as through biomass harvest or livestock grazing. HANPP can be considered as the sum of these two components (i.e. HANPP¼ΔNPPh þΔNPPLC) and indicates land use intensity, explicitly linking natural and socioeconomic processes. Depending on the precise definitions used for components 1 and 2, a third component, human-induced destruction of NPP without purpose (e.g. by human-induced fires), may also be included. Even without including this, the conventional definition of HANPP accounts for the fraction of NPP which, in strongly human-controlled ecosystems such as plantation forests or grazing pastures, is not appropriated by humans (NPPt ¼NPP0  HANPP), though its inclusion completes this account [42]. 2.6.2. Human appropriation of terrestrial net primary production Employing data from vegetation modelling, geographical information systems data on land use, land cover and soil degradation, and agricultural and forestry statistics collected for the year 2000, global terrestrial HANPP has been estimated at 0.606 ZJ [9] or 28.3% of global terrestrial NPP (NPPact). Of this HANPP, harvest (NPPh) contributed 53%, land conversion (NPPLC) 40%, and human-induced fires 7%. Notably, the total annual harvest of primary production by humans (NPPh) contains approximately 0.32 ZJ of energy, which was equivalent to 64% of TPES in 2010. Contributions of cropping and forestry to HANPP are of interest in considering large-scale agro-bioenergy production. Respectively, they contributed 49.8% and 10.6% of terrestrial HANPP, and grazing land an additional 28.5%. Remarkably, global terrestrial HANPP only doubled over the 20th century while the human population grew fourfold and economic output seventeenfold. This is because of decreased

Fig. 3. The standard definition of human appropriation of net primary production (HANPP), graphically represented. Terms are defined in Section 2.6.1. Figure adapted from [42].

reliance on bioenergy due to the proliferation of fossil fuel use, and higher conversion efficiencies of primary photosynthates to products [85]. 2.6.3. Human appropriation of aquatic net primary production Published estimates of HANPP in aquatic ecosystems are fewer and less sophisticated than those for terrestrial ecosystems, focusing on ΔNPPh through fish harvest and largely neglecting other impacts (the ‘aquatic equivalent of ΔNPPLC’). Vitousek et al. [38] estimated HANPP in aquatic ecosystems at 2.2%. Pauly and Christensen [22] introduced the metric, ‘primary production required’ (PPR) to sustain the world’s fish harvest, giving a detailed estimate based on multiple trophic models of various aquatic ecosystem types. Results suggest that 8% of global aquatic primary production was required to support the harvest and discarded by-catch of world fisheries, averaged over the years 1988–1991. Only 2% of NPP was needed to support fisheries in open ocean systems but 24–35% was required in fresh-water systems [22]. A more recent study [21] applied the same metric but with more sophisticated analysis to find annual, global PPR at 6.3% of global, aquatic NPP, averaged over the years 2000–2004. Other metrics indicate strong human impacts on aquatic ecosystems beyond fish harvest throughout the world’s oceans [86]. This suggests that a comprehensive assessment of aquatic HANPP should include a measure of eco-systemic degradation analogous to ΔNPPLC, and may yield an estimate significantly larger than those considering only fish harvest and by-catch. In other words, the estimate of PPR to support global fisheries in Table 1 likely represents a “lowest case” scenario. 2.6.4. Human appropriation of ancient net primary production Using published data on energy conversion efficiencies of the biological, geochemical and industrial steps required for formation and extraction of coal, oil and gas, the ancient NPP that supplies modern human energy demand can be estimated [23]. The efficiency of formation and extraction of coal, oil and gas were respectively found to be  9%,  0.009% and  0.008% [23] (starting from biomass; a factor for photosynthetic efficiency (see Section 3.3.3) is not included here). Based on these efficiencies, fossil fuel properties and consumption rates, Dukes concluded that the 0.318 ZJ of fossil fuels globally consumed in 1997 [24] were derived from 1710 ZJ of ancient biomass energy. Scaling this result for fossil fuel consumption in 2010 (0.438 ZJ [11]) indicates a requirement of 2360 ZJ of ancient biomass

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

energy, which is equivalent to 550 years of Earth’s current annual NPP. In summary, due to the energy losses at each step of formation and human extraction of fossil energy originally stored through photosynthesis, 550 years worth of solar energy photosynthetically captured by the whole planet is required every year to supply current global demand for fossil fuels.5 Though stark, this result does not rule out the possibility of supplying global energy demand sustainably into the future via natural or artificial photosynthesis because global insolation outstrips current fossil fuel demand by three orders of magnitude (see Section 2.2). However, it clearly shows that this will be possible only if solar fuels can be brought to market at dramatically higher efficiencies than those of fossil fuels, including their formation. 2.6.5. Limits to sustainable HANPP, and bioenergy production The ratio, HANPP:NPP0 quantifies the potential biospheric energy flow diverted by humans. To maintain the biosphere requires this ratio be less than one but other species need NPP too, and so to sustain robust, biodiverse ecosystems this ratio must be substantially less than one. Bioenergy makes up a fraction of HANPP, so its quantity is constrained by the limit to sustainable HANPP and magnitudes of other essential HANPP fractions such as food, materials and other ecosystem services (e.g. sustaining food chains, waste decomposition, climate regulation and provision of clean drinking water). Based on data of Haberl et al. [9], Bishop et al. [87] estimated the maximum level of HANPP that could be sustainably supported and found that, to prevent interference with ecosystem services essential to humanity, the desirable upper limit of HANPP was a figure 60% less than the actual HANPP in 2005. Similarly, other global sustainability assessments based on different but related metrics such as the ecological footprint have found that global economic metabolism has already overshot the Earth’s regenerative capacity [88,89]. This strongly suggests that energy systems requiring increased HANPP must achieve either efficiency gains in socioeconomic utilisation of biomass (NPPh – see Fig. 3) [20], or increase Earth’s photosynthetic productivity beyond that of current biota (NPPact), or be based on some combination of the two. In terms of the components of NPP0 (Fig. 3), how might this be achieved? The need for sustainability restricts increases in NPPact through expanded agriculture (expanded NPPLC and NPPh, and correspondingly reduced NPPt) and/or further intensification of cropping and forestry (increased NPPh, with corresponding reduction in NPPt) [20,87,90,91]. The remaining option is to increase NPP0 by deploying photosynthetic energy systems (e.g. plants, microalgal cultivation systems and/or artificial photosynthetic systems) on otherwise-unproductive or low-productivity lands (including existing built environments), or on productive lands in ways that do not degrade existing ecosystems. Overall, this suggests that a global society powered sustainably by solar fuels will require a combination of increased photosynthetic productivity through increased production area and/or photonconversion efficiency per production area, and increased socioeconomic energy utilisation efficiency. 2.6.6. Sustainable HANPP for agro-bioenergy production in terrestrial ecosystems The maximum physical potential of the world’s total land area outside croplands, infrastructure, wilderness and denser forests to deliver plant-based agro-bioenergy is estimated at approximately 0.19 ZJ a  1 (38% of current TPES) [44]. Achieving this potential would require the complete conversion of pasture lands, woodlands, savannahs and tundras, which are already used heavily for grazing 5 One must add to this number the vast time (107–108 yr) required for geochemical processing of ancient biomass to fully capture the time investment required to supply fossil fuels to the modern economy.

141

and store abundant carbon, to agro-bioenergy and intensive forage production. This would roughly double the global human terrestrial biomass harvest, strongly affecting biodiversity, ecosystems and food supply [44]. This estimate of physical potential corresponds to 31% of terrestrial HANPP in 2000, leaving approximately one quarter of the sustainable limit to HANPP estimated by Bishop et al. [87] (i.e. 41% of terrestrial HANPP in 2000) available for other purposes such as food production. However, global energy demand is projected to rise by approximately 44–70% by 2035 [92], and food demand 60–100% by 2050 [44]. A continuation of current crop yield growth trends until 2050 will not suffice to meet the rise in food demand without added cropland areas and it therefore seems unrealistic to expect that food crop yield growth will free up land for agro-bioenergy production [44]. These findings indicate that agro-bioenergy from plant crops and forestry (e.g. ethanol derived from corn, sugar and lignocellulosic conversion of crop or forestry products) can be expected to sustainably supply only a small fraction of future global energy demand. A thorough exploration of this issue can be found elsewhere [90]. 2.6.7. Sustainable HANPP in aquatic ecosystems Although methods for studying terrestrial HANPP have yet to be adapted for aquatic ecosystems, studies using other metrics raise concerns over aquatic NPP reductions due to human-induced biodiversity loss. Chassot et al. [21] provide strong evidence for the basic claim that aquatic NPP ultimately constrains fishery catches. There is also evidence that increasing diversity of fish is associated with fishery yields [93]. However, the effect of fishery depletion on overall aquatic NPP is unclear, as it is difficult to predict the complex effects across trophic levels resulting from changes in predator populations at higher levels [94]. The current lack of comprehensive measures for HANPP in aquatic ecosystems may be of limited consequence to the potential for replacing fossil fuels with fuels derived from current NPP, since almost all proposals for the latter involve land-based production systems. However, [95] gives an ambitious proposal for large-scale domestication of open ocean for bio-energy production using aquatic micro-organisms. In light of the available evidence for strong pre-existing human interference in aquatic ecosystems, such a proposal suggests that a comprehensive study of aquatic HANPP is increasingly important. As part of this, the value of increased macroalgae production in coastal regions [96] and the use of closed marine microalgae production systems [97] could be considered. 2.7. Time scales of constraints to fossil-fuelled society 2.7.1. Anthropogenic climate change The components of the global climate system (atmosphere, ocean, land, ice and biosphere) interact through biogeochemical cycles including the carbon cycle, in which photosynthetic NPP is a major driver [26,47]. Human interference in the carbon cycle is very likely now sufficient to be causing lasting, global climatic changes, chiefly through global warming due to the greenhouse effect [98,99]. Conventionally, it is accepted that the ‘safe’, mean global warming limit is 2 °C above preindustrial temperatures [27,28,100]. However, it is now recognised that large-scale, abrupt disruptions to the climate system could occur before warming reaches 2 °C, and it has been suggested that this target may be better described as ‘the threshold between dangerous and extremely dangerous climate change’ [27–29]. By capping cumulative emissions of CO2 from 2000 to 2500 at 170 GtC, the probability of stabilizing mean global temperatures at 2 °C warming would be at least 90% (P(〈ΔT〉r2 °C) Z0.9). Considered in terms of the carbon content of remaining fossil fuel resources, this suggested upper

142

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

level of cumulative emissions equates to the consumption of 20% of 1P reserves and 9% of URR. Confidently avoiding dangerous climate change will therefore require that the remaining fossil fuel resources be left largely unproduced unless effective carbon capture and storage technologies can be implemented at a global scale in the required time frame [47,101]. Rogelj et al. [102] reanalyzed a large set of published emissions scenarios from integrated techno-enviro-economic assessment models. In a less stringent scenario than that considered by Zickfeld et al. [29] (P(〈ΔT〉r2 °C) Z0.66), greenhouse gas emissions were predicted to peak between 2010 and 2020, falling to a median level 8.3% below the 2010 rate by 2020. In contrast, emissions have in fact grown rapidly since the 2008–2009 global financial crisis [12]. Global energy demand is projected to rise from 0.5 ZJ to 0.72–0.85 ZJ by 2035, with economic growth rates ranging from 2.5% of gross domestic product (GDP) (low), to 3.9% GDP (medium) and 5.3% GDP (high) by the US Department of Energy [92]. This suggests an urgent need for a global atmospheric CO2 mitigation scheme, through both emissions reductions and scalable, sustainable carbon capture and storage systems. Robust evidence from a range of climate–carbon cycle models shows that the maximum warming relative to pre-industrial times caused by CO2 emissions is almost directly proportional to the total amount of emitted anthropogenic carbon [103]. The peak warming is determined by four variables: current atmospheric CO2 concentration, current rate of emissions increase, starting time of a global mitigation scheme, and the rate of emissions reduction realised by this scheme. Stocker determined that to limit CO2induced global warming to 2 °C, annual emissions reductions of 2.7% will be required if mitigation is begun immediately, rising to 3.2% if mitigation is begun in 2020, and almost 6% if the implementation of a global mitigation scheme is delayed until 2030. Economic models suggest that feasible maximum rates of emissions reduction are limited to about 5% per annum, meaning that under this assumption the 2 °C target will become unachievable if mitigation is not begun aggressively by 2027 [103]. Given that it will inevitably take decades to scale CO2-neutral energy systems to a globally significant level (see Section 2.8.2), the data suggest that it is important to begin this process as soon as possible. Atmospheric CO2 is now concentrated at 400 ppm by volume (ppmv) [104] which, by a conversion factor of 2.124 GtC ppmv  1 [105], equates to a carbon pool of  850 GtC. As a result of all CO2 sources and sinks, net growth in atmospheric CO2 was  3.9 GtC in 2008 [13], a rate that has grown following the global financial crisis of 2008–2009 [12]. In 2010, fossil fuel combustion and cement production together added  9.1 GtC to the atmosphere [12]. Combined with 0.9 GtC emissions from land-use change, this put total anthropogenic carbon emissions for 2010 at  10.0 GtC [12]. Lenton considered the potential for terrestrial, photosynthetic primary production (based on existing species and cultivation methods) to lower future atmospheric CO2 concentration on a meaningful time scale [47] and concluded that this would require diverting primary photosynthates into long-lived storage reservoirs including permanent forests, buried biomass (either deep in soil or in the deep ocean), biochar in soils and CO2 stored in geological formations. CO2 removal flux is constrained by the availability of land area (see Sections 2.2.1 and 3.4.2), the yield of carbon per area (which is constrained by photosynthetic light-tobiomass conversion efficiency – see Section 3.3.3), and the efficiency of conversion to long-lived carbon stores [47]. Lenton found that by 2050, the potential terrestrial CO2 removal flux is forecast to be 4–6 GtC a  1 which, together with natural sinks, could match current total anthropogenic CO2 emissions, thus stabilizing the atmospheric CO2 concentration and lowering peak global warming [47]. Following a business-as-usual trajectory, however, emissions would grow commensurately with global energy demand,

meaning that a larger global CO2 removal flux would be needed by 2050. This highlights the importance not only of transitioning away from fossil fuels to renewable energy sources, but also of engineering photosynthetic systems with elevated carbon-capture efficiency, to remove more atmospheric CO2 per land area. 2.7.2. Fuel supply limitations The future of fossil fuel supplies is a complex subject, influenced by many variables, and often controversial. Estimates of fuel reserves and resources are frequently revised due to price changes, improvements in exploration and production (extraction) technologies, and classification changes (e.g. what constitutes conventional vs unconventional reserves) [106,107]. Production rates can depend on geophysical constraints, demand, consumption efficiency improvements, sociopolitical dynamics and other factors [106–109]. The often-used reserves-to-production (R/P) ratio is at best a crude measure of fuel security because economic effects are likely to arise long before reserves are actually depleted, if production rate cannot keep pace with demand [79,108,110–113]. ‘Peak oil’ describes the time at which the production rate of an oil field or producing region peaks before going into decline. Predictions of peak oil were first made by Hubbert in 1956, using a model that did not depend directly on economic variables but rather only the oil field’s time of discovery, production history, an estimate of its ultimately recoverable resources (URR), and the assumption that production would peak when half of the resource had been depleted [107,108,114]. This model predicted a peak date for the USA in close agreement with that eventually observed in 1970 [107,108]. Similar peaks have since occurred in more than 60 (of 95) other oil-producing nations [108]. Hubbert’s work has been extended to a variety of modelling techniques, each with strengths and weaknesses, reviewed by Sorrell and Speirs [115]. Oil is the largest source of primary energy and tightly coupled to the dynamics of the modern economy. History suggests that recessions typically occur when oil expenditures exceed  5.5% of GDP [108]. Ten out of the eleven recessions in the USA since the second World War, including the most recent ‘Global Financial Crisis’ beginning in 2008, were preceded by spikes in oil prices [108,113]. Unsurprisingly, there is both wide interest in and controversy surrounding Hubbertian forecasts for a worldwide oil production peak. Predictions of a global oil peak ultimately rest on the assumption that all oil fields eventually peak, and global production equates to a sum over individual fields [107]. Peak-oil critics emphasise the theory’s reliance on uncertain URR estimates and other limitations of resource-peak models [107,115,116]. Hughes and Rudolph [107] collated 40 published forecasts for global peak oil; 31 predict a global peak between 2000 and 2020 while nine predict a peak as late as the 2030s or argue that production will grow indefinitely. Oil extracted from wells drilled in the traditional manner (called ‘conventional’) currently supplies  80% of all liquid fuel energy globally [117]. Conventional oil production has tracked a fluctuating plateau since 2005 [108,113,118,119] despite generally high (though volatile) oil prices. This contrasts with a history of strong correlation between price and production [113]. Moreover, production at existing fields is declining at rates of 4.5–6.7% a  1. Only by adding new fields is global production holding steady [113,118], yet oil discovery peaked in 1963 and has fallen short of production since 1980, suggesting that this is unsustainable [79,115]. Murray and King argued [113] that since 2005 the oil market has undergone a ‘phase transition’ to a new state in which global production is inelastic to demand, causing increased price volatility. Critics of peak oil theory predict that declining production at conventional fields will be more than offset by adding new fields yet to be discovered (e.g. deep water) plus natural gas liquids

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

(NGLs) and ‘unconventional’ sources of oil, such as ‘heavy’ oil, ‘tight’ (shale) oil, tar sands, coal-to-liquid (CTL) and gas-to-liquid (GTL) conversions. In 2012 all unconventional oil sources together supplied 5.5% of the global market, with a further 14.5% coming from NGLs [117]. Only 0.5% came from CTL and GTL in 2010 [120]. However, a recent boom in oil and gas production rates from shale in the US [121] has led to a prediction that these will respectively grow three- and sixfold from 2011 levels by 2030 [122]. The IEA has forecast a net increase of 15.5% in global oil production by 2035, driven entirely by unconventional oil [50]. Some analysts describe these forecasts as overly optimistic [113,118,123]. Hughes suggests [123] that industry-standard productivity models of shale oil and gas fields systematically underestimate their decline rates which, at 440% a  1 are in fact very rapid compared with 4.5–6.7% a  1 for conventional fields [113]. Consequently, the high drilling rates required to compensate for such rapid decline are likely to hasten production peaks compared with industry predictions [123]. More time is needed for data collection to settle these uncertainties. Importantly however, aside from questions of production rate, shale oil and gas extraction is more expensive, more resource intensive and more environmentally damaging than conventional oil and gas drilling on land [123]. This is also generally true of other unconventional oil production methods as well as deep-water drilling for conventional oil [106,113,118,123]. Moreover, these costs are likely to increase because the economic incentives are to exploit the highest-quality, lowest-cost resources first [106]. As exploration and production become more difficult, the energy return on (energy) investment (EROI) declines, driving up the price of oil and oil-dependent products and services [124]. Hall et al. [125] estimated that an EROI value of approximately 3 is required to just barely maintain an organised society, allowing 'only for energy to run transportation or related systems, but [leaving] little discretionary surplus for all the things we value about civilisation: art, medicine, education and so on'. It has been further argued that an EROI value of up to 10 may be needed for the primary energy supply of a modern industrial society, to sustain societal complexity [124]. Table 3 compares these EROI requirements with values that have been estimated for various fossil and solar fuels. The EROI of conventional oil itself has declined from 100 in 1930 to 20 currently (‘at the wellhead’, before transportation and/or conversions) [110,126,127]. This value nonetheless remains high compared with equivalent EROI estimates for shale oil and tar sands (after extraction, before transportation and conversion), and CTL; these are respectively 2 [127], 2–4 [128] and  2 [105]. The EROI of deep-water oil is below 10 as well [126]. These relatively low EROI values limit the capacity of unconventional oil to economically substitute for conventional oil and mitigate the effects of its peaking [107,108,111]. Naturally, the limitation of EROI applies to all energy sources, and is discussed for photosynthetic energy systems in Sections 3.4 and 3.5. Verbruggen and Al Marchohi report carbon intensities for unconventional oil sources that are 10–80% higher than conventional oil [106]. This is closely related to the low EROI values for these fuels since exploration and production are powered mainly by fossil fuels. Unconventional oils are therefore prone to deepening the already serious problem of anthropogenic climate change and are also vulnerable to the impacts of carbon pricing, compounding their costliness [107]. Although the precise future trajectory of global oil production remains uncertain, it is clear that scalable substitutes for conventional oil are now required. The high financial and environmental costs associated with unconventional oil cast doubt on its capacity to fill this role sustainably. According to Hughes and Rudolph [107], ‘It is no longer a case of ‘if’ or ‘when’ there will be a plateau and eventual peak in world oil production, it is now a question of how jurisdictions can prepare

143

Table 3 Energy return on energy investment (EROI) requirements for society, and current values for fossil and solar fuels. Energy requirement

Approximate EROI

References

Sustaining a modern, industrial society Minimally sustaining an organised society Fuel Conventional oil in 1930 Conventional oil Deepwater oil Shale (‘tight’) oil Tar sands oil Coal-to-liquid (CTL) Agro ethanol at commercial scale Agro biodiesel at commercial scale Microalgal biofuels at pilot scale Fuels from artificial photosynthesis

10 3

[124] [125]

100 o 20 o 10 o7 o4 2 mean 5 (0.64-48) mean 2 1 (0.001-5) None published

[125] [125,126] [126] [127,129] [128,129] [105] [129] [129] [130,131]

Numbers in parentheses indicate ranges of literature values from which estimates were calculated.

for a world with less oil, one in which energy security and climate policy play a dominant role.’ 2.8. Transitioning to energy systems based on current photosynthetic primary production 2.8.1. The importance of fungible fuels Although electricity is the most versatile energy carrier, its market share at final consumption is substantially less than that of fuels. While estimates from different sources vary, worldwide,  61% of total primary energy is finally consumed in the form of carbon-based fuels, without electricity as an intermediary (Fig. 2). The share of fuels converted to electricity is projected to continue growing from 18% to  22% of global final energy consumption by 2030 [18]. Electrification has the substantial benefit that electricity can in many applications be converted to mechanical work with efficiencies approximately three times higher than those of fuelpowered heat engines [132,133]. Nonetheless, it is thought that electricity cannot foreseeably replace chemical fuels in some sectors, particularly aviation and long-distance road transportation, due to the relatively low energy densities of existing and foreseeable charge-storage systems [134–137]. Existing batteries have energy densities two orders of magnitude lower than liquid fuels [134] and this has improved only sixfold since the earliest rechargeable batteries in the 1900s [138], despite decades of intensive research. This rate of improvement is more than 20 times slower than Moore’s law for digital computing power, indicating that caution is required when extending techno optimism inspired by the latter to battery technology. Shorter-range road transportation is more amenable to electrification [134], though electric vehicles (EVs) remain a negligibly small fraction of the roughly billion-strong global on-road fleet. Technoeconomic models suggest that market penetration of pure (batteryonly) EVs in developed economies is unlikely to exceed 10% by 2030 [139,140]. Including plug-in hybrid EVs, which rely on liquid fuels for long-range travel, total market penetration is predicted to not exceed 30% by 2030 [139,140]. Moreover, existing high-efficiency, gasolineonly and gasoline-hybrid vehicles emit less CO2 than coal-fired electricity-powered EVs [91]. Consequently, to mitigate carbon pollution, the adoption of EVs would need to be accompanied by decarbonisation of the consumer electricity supply [91], which is projected to take place over several decades [91,141,142]. Even near-term breakthroughs in electricity-generation technology, such as practical nuclear fusion reactors, may not accelerate this significantly (see Section 2.8.2). An alternative that may help to accelerate a large-scale transition away

144

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

from fossil fuels is to supply the economy with non-polluting, fungible fuels (compatible with existing or minimally modified fuel distribution infrastructure and fuel-powered technologies). In summary, due to time constraints it is important to focus on sustainably powering the global carbon fuel-based economy that actually exists, in parallel to preparations for a cleaner, future economy powered by different energy carriers such as renewably generated electricity and hydrogen. This is because a transition to any such future economy will inevitably be complex, unpredictable and gradual, spanning several decades at least (see Section 2.8.2) [133,143]. This is very probably longer than the time available for effective climate-change and energy-insecurity mitigation. Importantly for the rapid development of viable photosynthetic energy systems, this suggests a need to focus on systems that produce fungible fuels such as hydrocarbons, biodiesel and ethanol now, while also ramping up the development of future fuels such as hydrogen. Anticipating postcarbon fuel systems, Toyota for example is planning to commercialise hydrogen fuel cell vehicles from 2015 [144]. This is supported by early-stage investments in refuelling infrastructure in the UK, Germany, Denmark, Japan and the US [144], and can benefit significantly from increasing economies of scale. Although existing renewable electricity-generation technologies can reduce CO2 emissions, they cannot directly sequester atmospheric CO2. Conversely, renewable fuel-production cycles can be made carbon negative, such that they divide the captured carbon between producing CO2-neutral fuels and a carbon-rich material appropriate for sequestration in long-term stores. For example, biomass components not used for biofuel production can be pyrolysed into so-called bio-char, which can be added to soil or buried [47,145]. In the case of microalgal biomass, as much as 55% of the carbon can be recovered as biochar [146]. The enhancement of soil carbon also offers the potential for increased NPP on marginal land and so could theoretically further expand biological carbon sequestration and contribute to a gradual reversal of NPP loss through land conversion [147]. 2.8.2. Time scales of global energy transitions According to Smil, ‘An energy transition encompasses the time that elapses between the introduction of a new primary energy source (coal, oil, nuclear electricity, wind captured by large turbines) and its rise to claiming a substantial share of the overall market.’ [91]. Smil argues that at least 15% constitutes a ‘substantial share’ and claims that for an energy source to be an absolute leader, it must contribute more than 50% of TPES [91]. The large scale of global fossil fuel consumption, and urgency with which a transition away from it is required, raise the question of how rapidly such a transition could be made to happen. Previous global energy transitions lend insight. Coal replaced biomass in 1885 as the energy source having the largest share of TPES, and oil replaced coal by 1960 [33,133]. Marchetti found that individual energy source market shares approximated Gaussian (bell-shaped) curves over time [148]. For more than a century (1850–1970), successive transitions appeared to follow a predictable pattern, robust to major economic perturbations such as wars, booms and depressions [33,148]. This suggested a characteristic time scale of almost a century, for the emerging energy source to achieve 50% market share [133]. Applying his model to more recently adopted energy sources, natural gas and nuclear electricity, Marchetti projected similar, near-century transition times. In reality these transitions have been slower. For example, since 1970, different energy source market shares have stabilised and so diverged from Marchetti’s model, with coal and oil retaining large shares for longer than predicted. Unruh argues [149] that industrial economies have been locked into fossil fuel-based energy systems through a process of technological and institutional coevolution driven by path-dependent increasing returns to scale. It

is asserted that this ‘carbon lock-in’ creates persistent market and policy failures that can inhibit the diffusion of carbon-saving technologies despite their environmental and economic advantages [149]. Smil writes, ‘There is only one thing that all large-scale energy transitions have in common: Because of the requisite technical and infrastructural imperatives and because of numerous (and often entirely unforeseen) social and economic implications (limits, feedbacks, adjustments), [such transitions] are inherently protracted affairs. Usually they take decades to accomplish, and the greater the degree of reliance on a particular energy source …, the longer [the substitution] will take.’ [133]. Evidence reviewed in Section 2.8 indicates that the time remaining in which to reduce atmospheric CO2 concentrations to safe levels and improve energy security by replacing heavily entrenched fossil fuels with sustainable energy sources is now short when compared with previous global energy transitions. Moreover, previous global energy transitions have all: (1) taken place in a world economy that was at least an order of magnitude smaller, in GDP terms, than today’s; (2) transitioned from a lowerquality6 to a higher-quality dominant primary energy source; (3) been largely market based. The transition needed by today’s much larger and more interconnected economy is from higherquality fossil fuels to a suite of lower-quality, low-carbon primary energy sources – primarily solar radiation and its conversion products – within a time frame of a few decades. This suggests a need for unprecedented sociopolitical leadership and international cooperation, as well as strong incentives for energy technology innovation while avoiding ‘picking winners’ prior to commercialisation, to accelerate the transition away from fossil fuels as much as possible. Technologies based on photosynthesis are likely to be integral to this strategy as they provide a nexus between the huge solar energy resource and economically useful chemical fuels.

3. Photosynthetic energy systems for solar fuel production 3.1. Photosynthesis in higher plants and green algae Higher plants, green algae and cyanobacteria conduct oxygenic photosynthesis via evolutionarily-related photosynthetic machinery [151]. Fig. 4 summarises the photosynthetic machinery in higher plants and green algae, which are increasingly of interest for the development of bioengineered photosynthetic energy systems and provide a blueprint for bio-inspired, artificial systems. Cyanobacterial systems differ particularly in their light-harvesting antenna complexes. Where these consist of light-harvesting membrane proteins in higher plants and green algae, cyanobacteria have extensive, extrinsic phycobilisomes. In biological systems photosynthesis can be conceptually partitioned into two stages: the ‘light’ and ‘dark’ reactions (Fig. 4). During the light reactions, pigment-binding protein complexes (PPCs) absorb light and transfer the derived excitation energy through a network of pigment molecules to photochemical reaction centres (RCs) of photosystems I and II (PSI, PSII). Here charge separation yields energised electrons and protons derived from water splitting. These are subsequently transferred via the photosynthetic electron-transport chain and associated protontransport pathways to produce the energy carrier, ATP and the reducing agent, NADPH. The dark reactions, which do not require light, then use these products to reduce CO2 in the Calvin–Benson– Bassham cycle (more commonly, the Calvin cycle), producing sugars, starch, oils and other bio-molecules [159]. Although the 6 Cleveland et al. define energy quality as the relative economic usefulness per heat equivalent unit of different fuels and electricity [150].

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

two stages include a complex web of photochemical and biochemical reactions [159], the overall chemistry is summarised by Eq. (1) Light energy

Carbon dioxide þ Water - Carbohydrate þ Oxygen Photons

2nCO2 þ 2nH2 O-2ðCH2 OÞn þ 2nO2

ð1Þ

The process is comparable to a photovoltaic cell (light reactions) coupled to an electrochemical cell (dark reactions), which together use photons to drive chemical synthesis. 3.2. Solar fuels Currently, the only industrial-scale, renewable fuel-synthesis processes are agro-biofuel systems. In 2010, 0.0027 ZJ of fuel ethanol and biodiesel were produced globally [50]. At 0.5% of TPES, these two fuels constitute almost the entire global solar fuel sector, though demand for others is growing, especially biogas [160]. This section summarises properties and production of different solar fuels. Except ethanol and hydrogen, the fuels discussed are not pure compounds but rather variable mixtures of similar compounds falling within a range. Physicochemical properties of different fuels are given in Table 2.

145

3.2.1. Ethanol Fuel ethanol is either pure and anhydrous or may contain a small fraction (4–5%) of water, depending on production methods and intended use. The stages of large-scale renewable ethanol production are: Photosynthetic primary production, sugar fermentation, distillation and dehydration. The last stage may be omitted if complete purity is not required. Purification by distillation alone limits ethanol purity to 95–96%, due to the formation of an azeotrope. The resulting solution can be used as a fuel alone but is immiscible in gasoline and so cannot be used in gasoline blends [161]. In 2010, 0.0021 ZJ of fuel ethanol were produced worldwide [50]. Numerous studies, collated in [129] have evaluated the EROI of fuel ethanol from different production methods based on feedstocks from plants. Numbers vary widely, from 0.64 for ethanol produced from wood cellulose to 48 for ethanol produced from molasses in India. Hall et al. found a mean EROI value of roughly 5 from 74 studies [129]. The broad range of EROI values is partly attributable to different assumptions made by different analysts; care must be taken in comparing values and future work would benefit from methodological standardisation [162–164]. 3.2.2. Biodiesel Biodiesel is a variable mixture of alkyl (methyl, propyl or ethyl) esters of fatty acids. Traditionally, biodiesel is produced by

Fig. 4. Photosynthesis in green algae and higher plants. Incident light (photon wavelengths shown are reduced by 2–3 orders of magnitude for practicality) is absorbed by photosystem core complexes I and II (PSI and PSII) and their coupled antenna light-harvesting complexes (LHCs), embedded in the thylakoid membrane within the chloroplast. PSI and PSII, together with their LHCs, form the PSI–LHCI and PSII–LHCII supercomplexes, in which the primary role of the LHCs is light absorption and excitation energy transfer (EET) to reaction centres (RCs) within PSI and PSII. There the energy drives the light reactions, which are linked to form the photosynthetic electron transport chain [152]. The first step is catalysed by PSII–LHCII [153]. Light is predominantly absorbed by the major LHCII proteins [154], which bind chlorophyll a, chlorophyll b and carotenoid chromophores. This creates electronic excitations which are transferred non-radiatively through the minor LHCII proteins (CP29, CP26 and CP24, the latter being absent from algae) [155–157] to PSII where they drive the splitting of water molecules, yielding protons, electrons and oxygen. The electrons are passed via plastoquinone (PQ), cytochrome b6f (Cyt–b6f) and plastocyanin (PC), on to PSI where the associated LHCI proteins transfer electronic photoexcitations to PSI. Electrons are passed from PSI to ferredoxin (Fd) where they are ultimately used in the production of nicotinamide adenine dinucleotide phosphate (NADPH), a reaction catalysed by the ferredoxinNADPþ oxidoreductase. Linear flow of electrons through the electron transport chain (cf. cyclic electron flow [158]) is accompanied by simultaneous release of protons into the thylakoid lumen by PSII, and the PQ/plastoquinol (PQH2) cycle and Cyt–b6f. This generates a proton gradient across the thylakoid membrane, driving subsequent adenosine triphosphate (ATP) production via the proton-pumping ATP synthase (ATPase). NADPH and ATP are subsequently used in the Calvin–Benson–Bassham cycle and other biochemical pathways to produce sugars, starch, oils and other biomolecules.

146

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

transesterification of biological oil feedstock such as vegetable or animal oil. In 2010, 0.00063 ZJ of biodiesel were produced worldwide [50]. Hall et al. found a mean EROI value of approximately 2 from 28 studies of EROI in biodiesel production from plant-based feedstocks [129]. 3.2.2.1. Biomass-to-liquid (BTL) fuels. Oil-based fuels can also be produced by a multi-stage BTL process, in which biomass from a wide variety of sources is gasified7 and the gas then polymerised through Fischer–Tropsch synthesis to produce diesel-range hydrocarbons. Whereas traditional biodiesel production uses only a part of the available biomass (oil), BTL allows all of the biomass to be used, thus using NPP more efficiently. Hydrothermal liquefaction is a different BTL process under development [165]. It is a medium-temperature, high-pressure thermochemical process, which produces a liquid product, often called bio-oil or bio-crude. The biomass is hydrolysed into smaller compounds, many of which are unstable, reactive and can recombine into larger compounds. During this process, a substantial part of the oxygen in the biomass is removed by dehydration or decarboxylation. The chemical properties of bio-crude are highly dependent on the biomass substrate composition. Hydrothermal liquefaction is seen as a promising means for converting microalgae to fuel [166], in particular because it uses wet biomass as an input (reducing preprocessing costs compared with dehydrated feedstock) and because the bulk of the biomass, and not only the oil fraction, is converted to bio-crude. 3.2.3. Biogas Biogas is produced by fermentation or anaerobic digestion of organic materials such as biomass, manure, sewage and municipal waste [54]. Its composition varies depending on production process but is predominantly CH4 (40–75%, typically  65%) and CO2 (25–60%), with trace amounts of water vapour, hydrogen and hydrogen sulphide [54,55]. The useful energy content of biogas is contained almost exclusively in its CH4, which under standard conditions has an energy density of 55.6 MJ kg  1 [53]. The energy density of biogas scales linearly with CH4 content; for a typical mole fraction of 65% it is  30 MJ kg  1. Carbon emissions per volume of biogas combustion are effectively independent of composition since the CH4 fraction is converted to CO2 and the initial CO2 fraction is unchanged [53]. However, due to the scaling of energy density with composition, CO2 emissions per MJ also depend on composition. For 65 mol% CH4 biogas, CO2 emissions per MJ are equivalent to typical biodiesel (Table 2). Notably, the global warming potential (GWP) of biogas is reduced through combustion, since the 100-year GWP of CO2 is approximately 25 times weaker than that of CH4. 3.2.4. Hydrogen Biological systems provide various methods for hydrogen (‘biohydrogen’) production, including indirect water photolysis, direct water photolysis, photofermentation and dark fermentation. Cyanobacteria evolve hydrogen through indirect water photolysis, by first photosynthesising carbohydrates and then fermenting them to produce H2 and CO2 [167]. Photoheterotrophic species can photoferment carbohydrates provided by autotrophs such as cyanobacteria [167]. These species are of interest for converting waste organic compounds into hydrogen. During dark fermentation, anaerobic bacteria ferment carbohydrates to produce a biogas mixture which is predominantly H2 and CO2 but which can also contain trace amounts of CH4, carbon monoxide and/or hydrogen 7 Reacted with a controlled amount of oxygen and/or steam at high temperature to produce syngas, a mixture of carbon monoxide and hydrogen.

sulphide [167]. Green microalgae such as the model species C. reinhardtii can produce hydrogen by direct water photolysis under anaerobic conditions, typically achieved through a process of sulphur-deprivation through which the rate of photosynthetic O2 production is balanced with metabolic O2 consumption [168]. For energy applications, hydrogen is useful both directly as a fuel and as a feedstock for the production of carbon-based fuels such as methanol, ideally in combination with CO2 extracted from the atmosphere [169]. 3.2.5. Hydrocarbons ‘Biohydrocarbons’ are hydrocarbons produced by living organisms or from biological feedstock. They offer a directly fungible substitute for fossil hydrocarbon fuels. In principle, biohydrocarbon fuel mixtures can be produced that exactly mimic petroleum-derived gasoline, diesel and aviation fuel such as kerosene-type BP Jet A-1 [170]. Biohydrocarbons can be produced by chemically processing biological feedstocks including cellulosic biomass, bio-oils and sugars [170]. Production may also occur through direct biosynthesis, as recently characterised for alkanes in cyanobacteria [171], although this remains in the early stages of research. Biohydrocarbons have garnered increasing interest and funding in recent years, particularly because they are compatible with existing infrastructure [172]. 3.3. Photosynthetic energy systems Generically, photosynthetic fuel production requires [173]: (1) an energy source in the form of light; (2) a substrate that can be oxidised to emit electrons (e.g. H2O); and (3) a substrate that can be reduced by those electrons (e.g. CO2 or H þ ) to produce a fuel such as those described above, or feedstock (e.g. biomass) for later processing into fuel. Broadly, there are two approaches to combining these elements into photosynthetic energy systems with the potential to be both environmentally sustainable and economically scalable. The first involves bioengineering natural photosynthetic systems to reduce inefficiencies and re-purpose overall system functioning for cost-effective fuel production rather than survival and reproduction in native ecosystems. The second involves bio-inspired design of artificial photosynthetic systems based on components already well optimised by evolution in natural systems together with novel components designed to improve upon natural systems. Sections 3.3–3.5 introduce and compare photosynthetic energy systems based respectively on these two approaches: microalgal cultivation systems and artificial photosynthetic systems. These both offer promising alternatives to unscalable agro-biofuel production, though they differ substantially in their current levels of development. Comprehensive reviews of these systems’ technical details can be found in the referenced literature; our goal here is to summarise the state of the art for each type and identify key challenges for future development. Fundamental challenges facing both types of system are described in Section 4. 3.3.1. Kinetic rate limitations to productivity in biological photosynthesis Limitations to photosynthetic productivity in biological systems are major obstacles to both bioengineered and bio-inspired photosynthetic energy systems. We describe them here as a foundation for discussing engineering objectives for both types of systems. Fuel production through biological photosynthesis has three chemical steps: (1) Energy absorbed from radiation generates electrochemical excitations/redox equivalents; (2) the light-induced redox potential catalyses water oxidation (e.g. in photosystem II), generating protons, electrons stored as reducing equivalents, and oxygen; and (3) another catalytic system (e.g. RuBisCO for carboxylation using CO2, or a hydrogenase for H2 synthesis) uses the reducing equivalents

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

Table 4 Functions and kinetic rates of key catalysts involved in oxygenic photosynthesis. Catalyst

Primary function

Turnover rate (s  1)

References

Photosystem II FeFe Hydrogenase RuBisCO

H2O oxidation H2 synthesis Carboxylation

100–400 6000–10000 3–10

[252] [256,257] [177]

to produce high-energy, low-entropy chemicals such as carbohydrates, lipids, hydrocarbons or H2 [173]. The particular realisation of these steps found in higher plants and green algae is detailed in Section 3.1. Assuming optimal temperature conditions (i.e. most physiological conditions), as well as adequate supplies of inorganic nutrients (e.g. nitrogen, phosphorus, sulphur) and water, limitations to the rate of photosynthesis fall into two classes: (1) supply or utilisation of CO2; and (2) supply or utilisation of light [174]. Despite the complexity of chemical pathways involved in photosynthesis, under most conditions the process is selectively rate limited by either CO2 or light, and regulation mechanisms ensure that all other steps proceed commensurately with the rate-limiting process [174]. Under CO2-limited conditions, photosynthetic rate is insensitive to rising PAR irradiance, and under light-limited conditions it is insensitive to increases in CO2 concentration. In an ideal photosynthetic energy system, supplies of CO2 and light are matched to the system's capacity to utilise them and this capacity, which depends on the system’s structure and material composition at multiple scales, is maximised [175,176]. 3.3.2. Limitations to supply and utilisation of CO2 Limitations to supply and utilisation of CO2 in the chloroplast are usually discussed in terms of how the rate of carboxylation by the RuBisCO enzyme varies with the partial pressure of CO2. This is a primary limitation to photosynthetic rate under most physiological conditions since light saturation occurs at PAR irradiances prevalent between approximately 7 a.m. and 5 p.m. in sunny climates (IPAR 4400 μmol photons m  2 s  1). The RuBisCO enzyme, of which there are many structural variants across species, is plagued by three main limitations: Its extremely slow catalytic turnover rate of a few per second (compared with thousands per second for fast, naturally occurring enzymes – see Table 4), its low affinity for atmospheric CO2 and its use of O2 as an alternative substrate for the competing process of photorespiration [177,178]. RuBisCO evolved three billion years ago when Earth's atmosphere lacked oxygen and was vastly richer in CO2 than today. As the atmosphere slowly changed, the enzyme gradually evolved into its current-day forms, constrained by a complex reaction mechanism that evolved under very different conditions [177,179]. Different organisms have evolved strategies to compensate for RuBisCO's low efficiency. First, most organisms simply produce the enzyme in large quantities; it is probably the most abundant protein on Earth [180]. Cyanobacteria and C4 plants have developed systems for concentrating CO2 above atmospheric levels in the vicinity of RuBisCO and this elevates photosynthetic activity [178,181]. Crassulacean-acid-metabolism plants concentrate CO2 on a diurnal basis to conserve water under dry conditions; they absorb and concentrate CO2 at night and close their stomata during the day to avoid transpiration losses [182]. However, as carboxylation by RuBisCO remains the rate-limiting step in all of these organisms under PAR-limited conditions, the enzyme has long been viewed as an important target for genetic manipulation.

147

Section 4.1.3 explores the potential for this and for engineering artificial carboxylation enzymes more effective than RuBisCO. 3.3.3. Limitations to supply and utilisation of light Light-harvesting kinetics and efficiency are usually assessed through linear-process analysis in which a model system is partitioned stepwise into subsystems that each complete one of the energy transfer and/or conversion steps spanning the free-energy gradient from incident light to biochemical products. Fig. 5 illustrates this for a system cultivating wild-type green microalgae, under light-limited and light-supersaturated conditions. The proportion of incident light energy remaining after each process step is shown. Similar analyses for higher plants can be found in References [183,184]. 3.3.3.1. Photosynthetically active radiation. Forty-three percent of the energy in standard AM1.5 solar spectral irradiance is in photosynthetically active radiation (PAR) (see Section 2.1) and is therefore taken as the maximum available for harvesting by natural microalgae. 3.3.3.2. Light transfer to pigment–protein complexes. PAR incident on a microalgal cultivation system must be transferred to the cells and PPCs within them. The first step is transmission into the aqueous culture. Weyer et al. [185] calculate that for an open-pond system, 5% of PAR or 2% of total incident radiation is reflected from the culture surface over a day, leaving 41% of the total incident energy available for absorption within the culture. Assuming limited absorption and scattering by components other than the PPCs, almost all PAR entering the culture can be absorbed by the latter [185]. In a photobioreactor however, light may encounter multiple culturecontainer air interfaces, affecting reflection losses. Furthermore, the spectrum of the light eventually incident on cells depends on the path length travelled because water absorbs red light more strongly than blue light. These effects depend on the specific photobioreactor design so, for simplicity, reflection losses are here assumed for the ‘base case’ of an open-pond system. 3.3.3.3. Excitation energy transfer to photochemical reaction centres. Electronic excitations resulting from light absorption by the PPCs are transferred non-radiatively through the chromophore network of LHCI, LHCII and minor antenna proteins to RCs in the PSI and PSII core complexes. There the energy is used by P680 (in PSII) and P700 (in PSI) to drive primary charge separation reactions, generating the electron and proton flows of the photosynthetic light reactions. Under both light-limited and lightsupersaturated conditions, some energy loss to other excitation decay pathways such as fluorescence and internal conversion is unavoidable. A minimum of 0.5% of absorbed light energy is reported to be wasted as fluorescence under light-limited conditions, and a maximum of 5% under light-supersaturated conditions [191]. Each charge-separation event requires an input of 1.80 eV (the energy contained in a 680 nm, red photon) in the case of PSII and 1.75 eV (700 nm, red photon) for PSI [186]. Additional energy absorbed from shorter-wavelength photons is wasted as heat through internal conversion during excitation energy transfer (EET) [186,192]. This effective down conversion (for example, of blue photons to red photons) during EET through the ‘energy funnel’ landscape of the antenna is a source of energy inefficiency even though the quantum efficiency of EET approaches unity under light-limited conditions [193,195]. The maximum photonutilisation efficiency can be estimated as the ratio of the average energy used for charge separation (1.78 eV, red) to the average absorbed photon energy (2.34 eV, green), giving 76% [185,188]. Further, thermodynamically dictated losses occur during charge separation itself [196], leaving on average 1.6 eV stored per

148

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

Fig. 5. Process analysis of photosynthetic energetics in wild-type green microalgae under light-limited and light-supersaturated conditions. See Section 3.3.3 for explanation. Abbreviations: PAR – photosynthetically active radiation; PPC – pigment-protein complex; EET – excitation energy transfer; RC – reaction centre; CS – charge separation; NPQ – nonphotochemical quenching; Fl – fluorescence; D-C – down-conversion; FEL – free energy loss [4,5,7,185-194].

primary charge separation event. This reduces the estimated maximum photon-utilisation efficiency to 68% and further small losses are incurred by occasional charge recombination in the RC [193]. Overall therefore, photon utilisation losses account for 13% of the incident solar energy under light-limited conditions (i.e. in Fig. 5, 41% of total irradiance induces chromophore excitations, while only 28% of total irradiance induces charge separation at RCs). Under light-supersaturated conditions the light-harvesting complexes (LHCs) absorb energy from radiation faster than RCs can utilise the resulting excitations, due to the kinetic limitations of downstream carboxylation by RuBisCO. This leads to so-called ‘closed’ RCs. To avoid photoinhibition (oxidative damage at the RCs [189]) under these conditions, excess excitations are dissipated as heat through the processes of nonphotochemical quenching (NPQ) [189]. Melis et al. estimated that in a typical high-light environment, a microalgal mass culture or dense plant foliage can overabsorb and dissipate approximately 60% of the daily irradiance through NPQ [188,190]. Cells at the directly illuminated surface can dissipate or ‘waste’ over 80% of absorbed irradiance through NPQ. On a daily basis, taking account of average light fluctuations, this is equivalent to  25% of total solar energy being lost via NPQ (Fig. 5). The precise amount of wastage through various nonphotochemical relaxation pathways (i.e. fluorescence and downconversion losses) depends on species, acclimation state, irradiance and spectral quality of incident light, temperature, and system macrostructure. Accordingly, a broad range of estimates exists in the literature for absorbed photon utilisation efficiency

under physiological, light-supersaturated conditions; Weyer et al. state a range from 10–30% [185]. In Fig. 3, 27% (based on daily NPQ losses of 60% of absorbed light) has been chosen as representative (Fig. 3 – Light supersaturated condition: 41% of solar energy inducing chromophore excitation – 24.6% NPQ). 3.3.3.4. Chemistry and metabolism. The balanced equation of CO2 fixation via the Calvin–Benson–Bassham cycle (see Section 3.1) is [197]: 6CO2 þ18ATP þ12NADPH þ 12H2OC6H12O6 þ18ADP þ18Pi þ12NADP þ þ 6H þ Three ATP and two NADPH molecules are required to fix one molecule of CO2 into a hexose sugar such as glucose. The reduction of NADP þ is a two-electron process. To transfer two electrons to NADP þ via linear electron transport requires four photons (two to transfer two electrons from water through PSII and two more to transfer them through PSI). Ideally, eight photons would therefore be required to produce the two NADPH molecules needed to fix one molecule of CO2 into one hexose sugar molecule. Three ATP molecules are produced simultaneously. The measured number of photons per O2 evolved or CO2 fixed is  9.5 [188]. Thus, 84% of the energy used to drive charge separation in PSII and PSI is stored as carbohydrates [188]. Weyer et al. [185] quote literature values ranging from 11 to 89% for the fraction of energy captured by photosynthesis that algal cells require for respiration and housekeeping. We choose 50% of 9.5 photons as a representative value,

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

for illustrative purposes (Fig. 5). This effectively raises the number of photons required per CO2 fixed to  14. 3.3.3.5. Hydrogen. The H2-producing hydrogenase (HydA) is tightly coupled to the photosynthetic electron transport chain (Fig. 4). This reduces energy losses associated with the more extensive biochemical pathways involved in the synthesis of carbohydrates and lipids, resulting in higher theoretical light-tobiofuel conversion efficiencies. The metabolic load of maintaining the cell is of course still present. The theoretical upper limit of efficiency in microalgal H2 production is estimated to be 12–14% [188]. 3.3.3.6. Carbohydrate and oil-rich biomass. Due to the different energy requirements of different biochemical pathways, the number of photons required to store one mole of carbon in a given molecule varies. The more energy dense the final biomass product, the higher the number of photons required to produce it. Thus, while low-lipid biomass typically needs 14 photons,  20 photons are typically needed to achieve high lipid content. Overall theoretical maximum light-to-biomass production efficiencies are calculated to be 12% and  5% respectively for the light-limited and light-supersaturated conditions shown in Fig. 3 [184,185,188,194]. Only approximate values are provided because the photosynthetic machinery differs between systems and species. Importantly, the theoretical upper limits stated are higher than the 2–4% photon conversion efficiencies currently reported at the pilot and demonstration scale over an annual cycle [356]. 3.3.3.7. Key efficiency limitations. Fig. 5 shows that the greatest inefficiencies in photosynthetic light harvesting are non-absorption of incident wavelengths falling outside the PAR spectrum (i.e. UV and IR) and the dissipation of unusable electronic excitations through NPQ. Under light-limited conditions, non-absorption of non-PAR is the dominant inefficiency. Under light-supersaturated conditions, NPQ dominates. The challenge of maximising light-harvesting efficiency is complicated by the fact that light-supersaturated and lightlimited conditions coexist in real systems in a typical high-light environment. Light-supersaturated conditions exist near illuminated surfaces, and light-limited conditions in shaded regions. The above analysis shows the importance of matching PAR irradiance to the level that saturates photosynthetic kinetics; in both light regimes a mismatch between these rates underlies the dominant source of inefficiency. Light-harvesting efficiency would be maximised if every photosynthetically active system component received PAR in balance with its maximumachievable productivity, determined only by thermodynamic limitations of its internal machinery (assuming that CO2 and other inputs are not limiting) [176]. This highlights the importance of integrated systems engineering, which balances detailed focus on system components with holistic system design. This applies not only to microalgal cultivation systems but also to artificial photosynthetic systems, and is discussed in Sections 3.4.3, 3.4.4, 3.5.1 and 4.3. 3.4. Microalgal cultivation systems – potential as scalable energy systems Microalgae (Fig. 6c,d) are unicellular eukaryotes, typically 1–30 mm in size, ubiquitous across Earth’s ecosystems [198]. Most are photoautotrophic, requiring PAR, H2O, CO2, nitrogen, phosphorous and a set of additional nutrients including K, Ca, Mg, S, Na, Cl, Fe, Mn, Zn, Cu, B, Mo, Si, Co, Se, V, I, Ni, and in some cases amino acids, vitamins and/or other additives [199]. Their ancient lineage (1.5–2 billion years since divergence from plants) and ability to colonise a broad range of

149

ecological niches has led to the evolution of an estimated 350,000 species or more [200]. Microalgae show great metabolic diversity and plasticity, and can synthesise a wide range of chemical compounds despite their simple nutritional needs. This enables them to adapt and survive under varied environmental stresses such as heat, cold, drought, salinity, photo-oxidation, anaerobiosis, osmotic pressure and UV radiation, and so access habitats inhospitable to higher plants [168,198]. Consequently, microalgae, cyanobacteria and other plankton are currently responsible for around 90% of aquatic NPP globally [21], or 45% (2.0 ZJ) of total global NPP. Microalgal cultivation systems are either open ponds, closed photobioreactors or hybrid systems combining elements of the two [201–203]. There is widespread interest in the potential for microalgal cultivation systems to produce fuels and other products economically, due to beneficial properties described in many recent review articles: [198,201,202,204–209,31,35,203,210–215]. Microalgae offer higher areal productivities than higher plants and, unlike higher-plant cropping, algaculture does not require arable land. Microalgal systems can also be coupled to multiple waste streams (wastewater, saline water [including seawater], CO2 sources), combining waste reduction (wastewater treatment, carbon sequestration) with valueadding processes (production of fuel, food, high-value-products [HVPs] such as pharmaceuticals). Flexible microalgal metabolism can be tuned to produce a wide spectrum of solar fuels (Section 3.2), either directly or through biomass used as feedstock for chemical processing. In contrast to higher plants, microalgae can be harvested on a daily or weekly basis, due to their short life cycles, giving potential for a more continuous year-round product stream. A growing number of green algal genomes (either complete or as scaffolds) are reported in the National Centre for Biotechnology Information (NCBI) genome database [216] and many diverse toolkits to engineer these are being developed. In terms of global primary productivity, microalgal cultivation systems have the potential to offer increased NPPh simultaneously with other economic and environmental services. However, they remain at an early stage of development, and important research and commercialisation milestones are still to be met before economic scalability can be achieved. This section considers some important constraints on the environmental and economic viability of scaling up microalgal cultivation systems to a globally significant level, and identifies key innovation pathways. 3.4.1. Areal productivities Literature surveys [185,217] of daily biomass areal productivities in microalgal cell cultures grown in outdoor open ponds8 under nutrient-replete conditions report areal productivities9 of 2.3–11.1 W m  2 (daily average) (10–48 g m  2 day  1). Weyer et al. [185] estimate 7.6–9.7 W m  2 (33–42 g m  2 day  1) as a ‘best case’ range, representing an optimistic target for production, based on realistic efficiencies and geographical locations with favourable conditions. Stephens et al. use 4.6 W m  2 (20 g m  2 day  1) as conservatively representative of existing open-pond systems, and 10.4 W m  2 (45 g m  2 day  1) as achievable with existing photobioreactors [35]. These figures compare favourably with the global average for agricultural crops,  0.4 W m  2 [33], and even the conservative productivity estimate for open ponds approaches the highest recorded plant growth rate of 5.0 W m  2 [48]. Assuming average insolation of 230 W m  2 (20 MJ m  2 day  1), which is realistic for a well-selected cultivation site [35,185], current average areal productivities of 4.6 W m  2 and 10.4 W m  2 8 Open pond cultivation systems are a more mature technology than photobioreactor systems, and are accordingly better represented in the literature. 9 Assuming biomass energy density of 20 MJ kg  1.

150

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

Fig. 6. Multiscale, hierarchical partitioning of higher plant and microalgal cultivation systems. Each row shows components within the holon, energy carriers that mediate transfer between components, and whether the composition and structure of the system are strongly or weakly interdependent at that scale. The particular structures shown at each scale a–j each represent a range of possible structures. a, (left) Schematic plant stand; (right) schematic photobioreactor array. b, (left) Higher-plant phytoelements; (right) photobioreactor phytoelements. Imaged reproduced from [320]. c, (left) Higher-plant chloroplasts within cells [321]; (right) microalgal chloroplasts within cells (image courtesy of G. Jakob, used with permission). d, (left) Transmission electron micrograph of a sugarcane chloroplast. Densely stacked grana (dark) regions of the thylakoid membrane are visible (image courtesy of R. Birch, used with permission); (right) transmission electron micrograph of cell from the green microalga largeantenna mutant strain stm3. Densely stacked pseudograna (dark) regions are visible in the chloroplast (image courtesy of E. Knauth, used with permission). e, (left) Schematic higher-plant ‘thylakoid domain’, comprising granum and surrounding stroma lamellae; (right) transmission electron micrograph of a typical ‘thylakoid domain’ from stm3 [322], comprising a pseudogranum and adjacent stroma lamellae (image courtesy of E. Knauth, used with permission). f, Schematic semicrystalline arrays of PSII (blue)–LHCII (green, brown) supercomplexes within the thylakoid. g, (left) C2S2M2-type PSII–LHCII supercomplex comprising reaction centre core dimer and peripheral LHCs forming a light-harvesting antenna; (right) C2S2-type PSII–LHCII supercomplex comprising reaction centre core dimer and peripheral LHCs forming a light-harvesting antenna functional domain. Nomenclature for different PSII–LHCII supercomplex types is explained in [323]. PSII–LHCII complexes with different antenna sizes can coexist in a membrane. The connectivity between such supercomplexes can vary depending on ultrastructural conformation of the membrane. h, Trimeric LHCII: one type of pigment– protein complex (PPC). i, Monomeric subunit of LHCII, with chlorophyll crystal structure overlaid. This complex binds 14 chlorophyll (8 chlorophyll-a, 6 chlorophyll-b) and 4 carotenoid chromophores. j, Strongly coupled dimer (a613–a614) of chlorophyll-a chromophores. k, Chlorophyll-a chromophore. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

respectively equate to average photosynthetic efficiencies of 2% and 4.5%. This is well within what is conventionally accepted to be the theoretical efficiency limit of  12% for biomass production (Section 3.3.3) and more realistically achievable limits [35], leaving room for engineering improvements.

3.4.2. Global-scale resource constraints With average insolation of 230 W m  2 and photosynthetic conversion efficiency of 2%, an area of 3.44  106 km2 would be required to displace global TPES in 2010 (0.503 ZJ). This equates to approximately 2.3% of total land, 3.4% of non-agricultural land, or

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

31.3% of estimated marginal land, based on existing productivities [35], suggesting that global scale-up of microalgal energy systems is not strictly land-limited. However, a comprehensive analysis must account for geographical availabilities of essential resources other than light, such as H2O, CO2, nutrients and funds. The total amount of primary production harvested annually by humans (global NPPh) at present contains energy equivalent to only 64% of TPES in 2010 (Section 2.6), and yet human interference in the global nitrogen cycle, principally through agriculture, is reported to have far exceeded the sustainable planetary boundary proposed by Rockström et al. [218]. Interference in the phosphorus cycle and global freshwater use are also rapidly nearing their safe limits [218]. Producing microalgae at scales comparable to present (let alone future) TPES will place further large demands on water (either fresh, saline or wastewater) and nutrient cycles, including nitrogen and phosphorus. If these were simply released as wastes, scaling up the microalgae industry would not be a sustainable option. However, compared with plant-based agriculture, microalgal systems are much better suited to nutrient recycling (at some energetic cost). Consequently, nutrient recycling is central to the optimisation of next-generation system designs [219]. Sequestration of CO2 extracted from the atmosphere at ambient concentration is an important goal for a global-scale microalgal cultivation industry, to assist with climate change mitigation (Section 2.7.1). However, colocation of cultivation systems with concentrated CO2 emitters such as fossil-fuelled power stations may provide an economic advantage during early industry expansion. This is because the primary limitation to photosynthetic rate under most physiological conditions is the rate of carboxylation by the RuBisCO enzyme, which depends on the partial pressure of CO2 (Section 3.3.2); raising this increases photosynthetic productivity (Section 4.1.3). Global-scale resource constraints to microalgal industry scale-up are further discussed by Stephens et al. [220]. The global potential of microalgae-based energy systems was recently assessed by rigorously linking global-scale resource flows to 4388 local-scale photobioreactor systems in a computational model that accounted for both biological and meteorological constraints [221]. This improved on previous studies, which have assessed global feasibility by simply scaling up lab- and pilot-scale models. It was found that maximum annual average lipid yields of 5.5–6.4 W m  2 (13–15 g m  2 day  1) are possible in the highproductivity countries including Australia, Brazil, Columbia, Egypt, Ethiopia, India, Kenya and Saudi Arabia. These productivities were integrated with geography-specific fuel consumption and land availability data to perform a scalability assessment. It was estimated that to replace 30% of national oil consumption with fuels from microalgae, assuming water, nutrients, and CO2 are not limiting, Brazil, Canada, China, Japan and the United States would respectively require 16%, 2%, 11%, 10,000%, and 49% of their non-arable land. This suggests that many regions can generate significant fractions of their transportation fuel requirements through microalgae production without being limited by land availability. Some countries (e.g. Australia) may also have spare capacity to produce fuels for export to countries without adequate land area to supply their own microalgal fuel needs. However, a fully comprehensive analysis must address the question of whether water, nutrients, CO2 or other inputs are limiting at the local scale of each cultivation system, as well as assess the system’s waste outputs. This is addressed by techno-economic and lifecycle analyses, which assess a system’s economic and environmental feasibility. 3.4.3. Technoeconomic and lifecycle analysis Technoeconomic analysis integrates simple physicochemical models of a system’s operation with economic models of its interactions with the market [222]. It is used to estimate whether a given

151

system design can achieve its economic targets (e.g. return on financial investments through cost-competitive fuel production) under the physical constraints of the system (e.g. local climatic conditions and light-to-biomass conversion efficiency of the photosynthetic machinery). Technoeconomic analysis is especially valuable in the early stages of system development, allowing designers to determine whether a design might be viable and, if so, which parameters give the greatest leverage over system performance (sensitivity analysis). This provides a basis for engineering improvements and the resulting system can be thoroughly assessed later through lifecycle analysis (LCA). LCA evaluates a system’s environmental and economic sustainability in more detail through comprehensive accounting of all energy, material and financial inputs and outputs associated with a particular product or process over all stages of its life cycle (e.g. extraction of raw materials, manufacturing, transport, use and recycling or disposal) [223]. This provides a measure of energy system performance that is more comprehensive than other measures commonly appearing in the literature, such as energy-conversion efficiency and financial cost per energy output. Furthermore, whereas ‘first-order’ LCA considers only the production and consumption flows of system operation, ‘second-order’ LCA also includes resources embedded in materials used for system construction and deconstruction [224]. Similar to techno-economic analysis, LCA sensitivity analyses give insight into which parameters can most influence system performance, and therefore engineering targets. It must be emphasised, however, that energy systems are complex systems with interdependencies and feedbacks between many variables. This presents a major challenge to designers, even with the aid of established system-analysis tools. Results from LCAs are highly sensitive to system particulars, definitions of system boundaries, lifecycle inventories, process efficiencies and functional units, and therefore often cannot be meaningfully compared between systems [130,223,225]. Moreover, design refinements based on LCA results can affect system performance in unpredictable ways. This is an irreducible property of any complex system; in general there is no single parameter that will reliably predict optimal performance of the overall system (see Section 4.3). Most LCA studies focus on the EROI (or, reciprocally, the net energy ratio (NER)) of a system, as well as its greenhouse gas emissions per unit of energy produced. Consumption of fresh water, nutrients and funds per unit of energy produced are sometimes also quantified. A recent review of the literature by Quinn and co-workers reported GHG values ranging from  75.29 g-CO2-eq MJ  1 to 534 g-CO2-eq MJ  1 and EROI values ranging from an extremely low  0.001 to 5 [130]. Most current studies report EROI values near or below unity [131] which, notwithstanding the complex effects of particulars in different studies, shows a clear need for system optimisation to elevate the EROI into the sustainable range above at least  3 and preferably closer to 10 (see Table 3). 3.4.4. Key innovation pathways for microalgal cultivation systems Every production system will have components important for achieving favourable EROI and GHG emission values, starting from current designs. For example, recent LCA results have been strongly affected by the efficiencies with which high-energy biomass components are harvested and the residual low-energy biomass recycled [130,226–228]. In particular, wet extraction and hydrothermal liquefaction or solar drying have been suggested as favourable harvesting methods compared with established techniques such as centrifugation. Additionally, anaerobic digestion is an effective mechanism for nutrient recycling [130,226,229]. These techniques and a variety of others are currently under development. However, even successfully deploying these techniques will not guarantee optimality within the space of possible system designs.

152

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

More general insights into improving lifecycle measures of microalgal systems may require meta-analysis of many LCA studies rather than detailed focus on individual studies, due to the sensitive dependence of LCA results on system particulars. As the field progresses and greater standardisation becomes possible, for example through universal adoption of a particular LCA software package and other standards, results of individual studies may become more broadly applicable. Nonetheless, due to the strong interconnections and feedbacks between system components, an approach is needed that balances reductionism with holism and allows all components to be optimised together. The goal is effective model-guided design, which allows for rigorous, rapid and cost-effective exploration of the design space. Optimising designs in silico may help to inform engineering that is integrated across scales, from large-scale properties of overall photobioreactor arrays down to the nanoscale properties of the chloroplast. One important area for improvement is the initial light-harvesting stage of photosynthesis. Section 3.3.3 identified the key limitations to light harvesting in microalgal photosynthesis as non-absorption of incident wavelengths outside of the PAR spectrum under light-limited conditions, and the dissipation of unusable electronic excitations through NPQ under light-supersaturated conditions. Both of these conditions occur simultaneously under most physiological conditions, in shaded and directly illuminated regions respectively. This seemingly paradoxical problem of absorbing both too little and too much light demands that systems be engineered to better distribute and utilise available light. A strategy with potential for achieving this is genetic down regulation of targeted LHC antenna proteins to minimise the light-harvesting antenna size, or genome editing (e.g. through the use of transcription activator-like effector nucleases [TALENs] and clustered regularly interspaced short palindromic repeats [CRISPR] technologies). This has the potential to reduce light supersaturation of cells and energy wastage near the culture surface, increasing light penetration to cells deeper in the culture that would otherwise be light limited, thereby improving overall light-conversion efficiency of the culture [230,231]. However, there remain many open questions about interrelationships between thylakoid membrane protein composition, structure and energetics across scales [175]. Addressing these questions is central to comprehensively modelling photosynthetic light harvesting for targeted engineering of energy systems. Naturally, this modelling must also account for processes of mass transfer (supplies of H2O, CO2, nutrients, pH conditions, sterility, and disposal of O2 and other waste metabolites) as well as heat transfer (temperature control). Achieving this will require a framework that links system-scale objectives and constraints with engineerable parameters across multiple scales in the system, in a coordinated way. This may need to integrate biological, chemical, physical, environmental and economic models, which presents a major theoretical challenge, discussed in Section 4.3. Meeting this challenge may provide a strategy for process scale-up, help to de-risk commercialisation and to fast track systems deployment. Ultimately, the aim is to deliver a positive triple-bottom-line outcome in terms of social, environmental and economic benefits. 3.5. Artificial photosynthetic systems Biological photosynthesis has inspired the design of artificial solar fuel-production systems intended to incorporate the best features of natural systems and, where possible, improve upon them. According to Gust et al. [232] ‘artificial photosynthetic systems’ (AP systems) include ‘photovoltaic cells based on inorganic semiconductors for electricity production, dye-sensitised solar cells, photovoltaics based on organic semiconductors, systems for fuel production based on these types of devices, a large variety of devices for direct photochemical conversion of light excitation energy to a fuel using organic or inorganic molecular

photocatalysts, natural organisms or their components interfaced to synthetic materials, and combinations of these approaches.’ Here the term is taken to refer to any predominantly abiological system that photocatalyses chemical fuel production without intermediate electricity generation. While some systems fitting this definition, such as concentrated solar thermal collectors coupled to thermochemical fuel production cycles [233], do not closely mirror biological photosynthesis, other systems are more directly biomimetic. These include nanoscale systems containing large organic molecules, metals, semiconductors, nano-structured materials, or combinations of these [232–237]. Such systems often mimic natural photosynthetic systems, for example by interfacing light-harvesting antennae with photochemically-active reaction centres that energise water oxidation, the reduction of protons (to H2) and/or CO2 (to carbohydrates, lipids or hydrocarbons). Sophisticated systems can also include quenchers to mitigate oxidative photodamage, similar to NPQ mechanisms in biological photosynthesis [232]. Furthermore, artificial and often self-assembling scaffolds (or maquettes) have been developed to provide the 3D coordination environment needed to maximise the efficiency of various light-harvesting and redox-active components required for solar fuel production [238]. It is reported that AP systems will eventually be able to exceed natural photon-to-chemical energy conversion efficiencies. This is partly because it is often assumed that no energy is required to cover the metabolic cost of living in an AP system [194,233]. However, these systems do have initial and ongoing energetic costs including embodied energy, energy of production, and energy required for maintenance, which can be likened to the metabolic energy required to grow and maintain algal cells. In natural photosynthetic organisms, this loss is typically on the order of 50% of the chemical energy stored by photosynthesis (Section 3.3.3). So, although it has been argued that systems using photovoltaic cells to power electrochemical cells for fuel production have a significant efficiency advantage over natural and bioengineered photosynthetic systems [194], this is based on a first-order energetic analysis of isolated components and not a full second-order lifecycle analysis of an integrated system. This is largely because these systems do not yet exist in a state ready for commercial deployment; as development proceeds, fairer comparisons between system types will likely become possible. Although it is widely acknowledged that scalable solar energy systems will need to be sustainably constructed from low-cost (financially and energetically), abundant materials, some AP systems and high-efficiency photovoltaics currently still depend on scarce, expensive materials, and/or energy-intensive construction [233,239]. In contrast, photosynthetic organisms self-assemble from abundant materials at ambient temperatures, and the abiological components of existing microalgal cultivation systems comprise relatively simple, abundant materials (e.g. Mg, Ca, Fe) instead of rare-Earth metals. Importantly, however, rapid progress is now being made on artificial catalysts that are similarly based on abundant materials and have greater potential for sustainability [240–244]. The development of bio-inspired AP systems is at an earlier stage than that of microalgal cultivation systems. Whereas microalgal systems are now at demonstration scale, to the authors’ knowledge, no AP systems have yet been proven feasible outside of the laboratory, so no fair comparison can yet be made between lifecycle measures for the two types of systems. 3.5.1. Key innovation pathways for artificial photosynthetic systems Most AP development to date has focused on photocatalytic water splitting for hydrogen production. Although this constitutes fuel production and a step towards mimicking the multi-step carbon-fixation of biological photosynthesis, today’s carbon-based economy as yet has relatively low demand for pure hydrogen fuel.

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

As discussed in Section 2.8.2, this is unlikely to change within the crucial period for dealing with climate change and fuel supply limitations (i.e. the next few decades). Carbon-fixing AP systems are therefore of significant interest and while they remain at a very early stage of development [233– 235], significant progress is being made [245]. Bringing such systems to maturity may provide a sustainable source of fungible carbon-based (and potentially carbon-neutral) fuels and/or a means for fixing atmospheric carbon for sequestration. A central challenge facing AP development is to progress from basic science, largely focused on individual system components such as molecular catalysts, to optimised, integrated systems that are environmentally sustainable and economically scalable, on a time scale meaningful to the mitigation of climate change and fuel-supply limitations. Substantial work remains to be done, both on system components and on their integration into useful systems, including the synthesis of more efficient water oxidation catalysts and electron mediators, design of robust light-harvesting assemblies, direct injection of photo-excited electrons into enzymatic redox centres, application of a wider spectrum of redox biocatalysts (e.g. mono-oxygenases, CO2 fixing enzymes) to artificial photosynthesis, and nanoscale assemblies of artificial PS I and II systems [235]. According to Berardi and co-workers, the main challenge ahead for the construction of technologically useful, artificial photosynthetic systems is the harmonisation of all the numerous reactions that occur in each compartment, such that the kinetics of all the processes involved facilitate the desired reactions [234]. It is clear that integrated systems engineering is essential, balancing the detailed focus on individual components with holistic design. Biological photosynthetic systems provide a guide to meeting this challenge, since they have also needed to strike this balance through evolution. AP systems show much promise as scalable solar fuel-production technologies for the long term. However, the present state of AP development suggests that aggressive investment, extremely rapid technical progress and expeditious technological adoption are required if commercial-scale systems are to protect against climate change and provide fuel security on a meaningful time scale . Despite the protracted nature of previous global energy transitions (see Section 2.8.2), Sovacool argues that ‘Future transitions may… become a social or political priority in ways that previous transitions have not been’ and ‘Indeed, although previous, historical transitions may have taken a great deal of time,… we have learned a sufficient amount from them so that contemporary, or future, energy transitions can be expedited.’ [355]. Moreover, there are some historical examples [355] of uncharacteristically fast national-scale transitions to new energy technologies. This may suggest the possibility of a substantial impact from commercial AP systems in the nearer term, if sufficiently rapid technical progress can be made and effective policy frameworks adopted (see Section 4.4) [355] them in a coordinated way. Morever, because these parameters also underly other process steps, tuning them to optimise just one step can have unexpected results for the process overall. For example, LHCs affect light distribution, light absorption, EET, NPQ, quinone transfer through the thylakoid, and possibly also have other functions not yet understood.

4. Grand challenges This section reviews current limitations to the development of scalable, sustainable photosynthetic energy systems, which apply to both microalgal-cultivation and AP systems.

153

4.1. Engineering next-generation catalysts for water splitting and carbon fixation 4.1.1. H2O oxidation Photocatalytic water-splitting (H2O oxidation followed by H2 and O2 synthesis) systems are based on one of two main approaches [246,247]. The first splits water completely into H2 and O2, using a single photocatalyst that absorbs light in the visible spectrum to generate a potential sufficient for overall water splitting. Photocatalysts that have these properties and are also reproducible are rare because of their stringent requirements for a thermodynamic potential suitable for water splitting, a sufficiently narrow band gap to absorb visible photons, and the required stability in terms of photocorrosion [247]. The second alternative is a two-step excitation mechanism using two different photocatalysts, as is seen in the Z-scheme of the light reactions in biological photosynthesis (see Section 3.1). This has the advantage that it can utilise a broader spectrum of visible light because the change in free energy needed for each photocatalyst is less than for a onestep water-splitting catalyst. It has the added advantage that the evolved H2 and O2 can be separated [247]; while this might seem trivial, it is a major design consideration for AP systems because even slight leakage could result in the production of an explosive gas mixture and undesirable product recombination. Photosystem II catalyses the most oxidising reaction known to biology [248], to carry out efficient water oxidation and oxygen evolution in the biological Z-scheme, and is the only biological photosystem capable of doing so [249]. Its active site, in the oxygen-evolving complex (OEC), is a Mn4CaO5 complex bound to residues of the D1 protein within the PSII core complex [250]. Through successive one-electron events, this so-called Mn cluster extracts electrons from water and donates them to the P680 þ ‘special pair’ of chlorophylls within D1. Four such events result in the oxidation of water and release of O2 [251]. The PSII reaction centre’s large reduction potential of  1.25 V and remarkable turnover rate of 100–400 s  1 in vivo under light-saturated conditions [252] have made its Mn cluster the focus of intensive study. Table 4 gives a simple comparison of key catalysts involved in oxygenic photosynthesis. Although the Mn cluster’s structure has recently been determined to 1.9 Å resolution [253], several different mechanisms of water oxidation and O–O bond formation remain plausible based on the current structure. Determining the exact reaction mechanism and all possible structural rearrangements of the OEC during operation will require further study [250]. X-ray free-electron lasers (XFELs) [254], which facilitate rapid trapping of structural intermediates [255], may be particularly useful towards this end. Despite its high catalytic performance, PSII is prone to oxidative damage, particularly under high light conditions [258]. This eventually degrades the D1 protein and oxidises pigments, causing photobleaching [259]. Conversely, when a kinetic bottleneck downstream in the electron transport chain causes a build-up of reduced plastoquinone (PQH2) near PSII, electron donation from the reaction centre is inhibited and charge recombination can occur. This can cause the reaction centre to interact with atmospheric triplet oxygen, producing singlet oxygen, which in turn bleaches the reaction centre. These damaging processes, together known as photoinhibition (see Section 4.2.2), reduce the number of active PSII units and, therefore, photosynthetic activity [259]. Evolution has compensated for PSII’s weaknesses through complex mechanisms of acclimation (slower) and photoprotection (faster), of two types: those that help to prevent photoinhibition by managing light absorption and the resulting electronic excitations in the light-harvesting machinery, and those that repair damaged D1 proteins after photoinhibition. Although acclimation and photoprotection are not discussed exhaustively here,

154

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

photoprotective nonphotochemical quenching mechanisms are described in Section 4.2.2. PSII serves as an example that highlights the difficulty of constructing an effective and robust water-oxidising complex. Ideally, such a complex would be a self-assembling and self-healing material, cheaply constructed from abundant elements and exhibiting, for example, a consecutive, four-step, proton-coupled electron-transfer pathway to circumvent high-energy intermediates during the fourelectron abstracting process of water oxidation [260]. Although no such ideal material has yet been found, various attempts have been made to produce a catalyst that mimics PSII in activity and efficiency, and substantial progress has been made in recent years. These are reviewed elsewhere, alongside efforts to combine them with effective light-harvesting and charge-separating components in integrated photocatalytic systems [242,246,247,260,261]. Of particular note are recently reported cobalt-based catalysts [245,262,263]. To different extents, these share important features with PSII, including structure, ability to self-assemble, self-repair, manage the proton-coupled electron transfer chemistry of water splitting, and capability for water oxidation at low overpotential, in neutral pH ranges (6–8) [245]. 4.1.2. H2 synthesis In biological systems, proton reduction and hydrogen oxidation are typically performed by hydrogenase enzymes, most of which can be classified into two families, based on the metal content of their respective dinuclear catalytic centres: nickel-iron (NiFe) and ironiron (FeFe) hydrogenases. These differ functionally in that NiFe hydrogenases tend to be more involved in hydrogen oxidation and FeFe hydrogenases in hydrogen synthesis. Moreover, FeFe hydrogenases show 102 times less affinity for hydrogen than NiFe hydrogenases, but are more sensitive to inhibition by oxygen and carbon monoxide [257]. They also have turnover numbers in the range of 6000–10,000 [256,257] (Table 4), which is approximately 101–102 times more active than NiFe hydrogenases. In some cases, FeFe hydrogenases reportedly achieve catalytic efficiencies rivalling platinum electrodes [264,265] despite exclusively incorporating abundant base metals. Consequently, hydrogenases, and the FeFe family in particular, are of great interest with a view to genetic engineering for enhanced photobiological H2 synthesis, as well as designing bioinspired artificial H2-synthesis catalysts able to replace expensive and resource-constrained platinum catalysts. Similarly to PSII, most hydrogenases are plagued by oxidative damage. Whereas inhibition by CO is reversible, inhibition by O2, even at trace levels, results in irreversible destruction of the active site [256,266]). An exception to this is a subtype of NiFe hydrogenases, which sustain H2 conversion in the presence of O2, called O2-tolerant hydrogenases [266]. There is substantial interest in developing hydrogenases that combine the high activities of the FeFe type with this capacity to tolerate O2 [256,266,267]. Recent developments have elucidated the mechanisms permitting oxygen tolerance. Notably, the X-ray crystal structure of an O2-tolerant NiFe hydrogenase was reported to contain an iron–sulphur centre near the active site, which can act as an electron acceptor during H2 oxidation and an electron-delivering protective device when O2 is incident on the active site [268]. Other features characteristic of O2-tolerant hydrogenases, and their implications for developing technologies based on O2-tolerant hydrogenases, are reviewed in [266], while [269] describes recent developments in the de novo design of artificial hydrogenases. Whereas it had previously been assumed that the assembly of functional hydrogenases required the maturase pathways, it was recently shown that the Fe–S cluster of an FeFe hydrogenase can be reconstituted with hydrogenase protein in a relatively straightforward manner to generate active hydrogenases [270]. The importance of this is that it opens up a pathway for much more cost-effective industrial production of hydrogenases,

artificial H2-synthesis catalysts, and potentially also similar pathways for the assembly of carboxylases and other enzymes involved in the production of carbonaceous fuels. 4.1.3. Carbon fixation Despite the substantial limitations of carbon-fixing RuBisCO enzymes described in Section 3.3.2, it has been suggested that RuBisCOs in different organisms may already be nearly perfectly optimised for their specific cellular microenvironments [271]. Their catalytic properties vary considerably, however, suggesting that changes in turnover rate, affinity or specificity for CO2 might be achieved by design, to improve RuBisCO performance in particular organisms or environments [272]. Bioengineering progress has so far been modest despite extensive effort [178,272], with targeted changes to the catalytically active large subunit through chloroplast transformation being the most successful approach [272]. However, there is also strong evidence – coming in part from computational models [273] – that the small subunits, which are distant from the active site, can fine-tune the dynamic structure of the overall enzyme and influence catalysis [272]. In light of ongoing progress in molecular bioengineering, computing power and molecular modelling techniques, there is potential to move towards directed bioengineering of carboxylating enzymes such as RuBisCO, based on modelling studies. However, analysis and optimisation of such complex systems sufficiently rigorous to inform directed bioengineering poses a substantial theoretical and computational challenge (see Section 4.3). So far, the most successful engineering approaches to increasing carbon fixation rate in autotrophs have involved modification of neighbouring pathways and enzymes, rather than of RuBisCO itself [274]. By extension, there is interest in the design of synthetic carbonfixing biological systems based on novel combinations of enzymes from different organisms [274,275]. The goal is to establish novel carbon-fixation pathways with improved overall carbon-fixation performance compared with the Calvin–Benson cycle. Bar-Even and coworkers used the entire repertoire of approximately 5,000 metabolic enzymes known to occur in nature, and computationally identified alternative carbon fixation pathways that combine existing metabolic building blocks from various organisms [275]. One such cycle was predicted to be two to three times faster than the Calvin– Benson cycle, although it is not clear whether the formidable challenges facing implementation of this or alternative pathways in a practical system can be met [275]. ‘Biohybrid’ electrocatalytic systems for carbon fixation are also being developed. In these systems, high-performing biological redox enzymes are applied to the surfaces of electrodes, which are then exposed to the chemical substrate of interest, such as CO2 [276]. The enzymes catalyse electrochemical half-cell reactions, linking interfacial electron transfers with chemical reactions. Of particular interest for carbon fixation are carbon monoxide dehydrogenases, which catalyse electrochemical interconversion of CO2 and CO with turnover rates comparable to hydrogenases (103– 104 s  1), and high catalytic efficiencies [276]. Studies of biohybrid artificial systems have demonstrated drastically improved catalytic performance at the laboratory scale compared with RuBisCO. Although such systems are not yet scalable to industrial levels, they provide proof of principle that the carboxylation step of photosynthesis need not be rate limiting [276]. Readers seeking a comprehensive review of options for fixing CO2 from the atmosphere, in the context of climate change mitigation, are directed to reference [277]. 4.2. Light harvesting in the chloroplast Even after decades of study, many questions remain about the chloroplast’s intricate light-harvesting mechanisms. Light

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

absorption, EET, charge separation, mass transfer, NPQ and heat dissipation mechanisms in vivo all depend on complex interdependencies between the composition, structure and dynamics of components within the thylakoid membrane and its surrounding aqueous phase [175]. The dynamics and interactions of these components show characteristic features across a spectrum of length and time scales, from chromophores to protein subunits, proteins, protein supercomplexes, aggregates of supercomplexes and overall membrane ultrastructure. These features can also change significantly in response to environmental conditions. For example, expression levels of photosystem and antenna proteins within the thylakoid vary with irradiance, temperature and even nutrient supply, which affects the 2D arrangement of these proteins and 3D arrangement of the overall membrane (e.g. layers stacked into ‘grana’ vs unstacked layers) [278]. These arrangements in turn affect how light, excitons, heat and mass flow through the membrane, including whether some excitons are nonphotochemically quenched and dissipated as heat [259]. They also affect how the overall membrane interacts with light, heat and mass in its environment [175]. Quantifying compositionstructure-dynamics relationships in the chloroplast is a major current research effort spanning a range of scientific disciplines. 4.2.1. Quantum coherence in excitation energy transfer An extensive body of recent research has addressed the question of whether long-lived quantum mechanical effects may be present in EET mechanisms and, if so, whether they can improve light-harvesting performance when compared with lightharvesting systems that are well described by classical physics [279–283]. Results have so far indicated that quantum effects may provide small improvements in photon conversion efficiency and/ or robustness to environmental noise or damage. The possibility that such improvements may affect an organism’s fitness has placed photosynthetic light harvesting at the centre of the emerging field of quantum biology [284–286]. Traditional models of EET, such as Förster theory [287], make restrictive assumptions such that some properties of excitons, like electronic quantum coherence, cannot be accommodated over time scales significant to EET [283,288]. Experimental evidence for longlived quantum coherence in EET came from spectroscopic observations of light-harvesting proteins [289] prepared and observed under conditions significantly different from their natural conditions in vivo [281,282]. Subsequent experiments under similar conditions have provided more evidence for long-lived quantum coherence in EET [290,291] and also, recently, in photosystem reaction centres [292], including at room temperature. Theoretical work has given insight into mechanisms that may preserve quantum coherence for remarkably long times (picoseconds) in the experiments, as well as its interactions with other features of the light-harvesting dynamics, and its possible benefits for efficiency and robustness (reviewed in [279,281–285,288,293]). However, the details of these mechanisms and their functional significance are not yet fully understood. It remains particularly controversial whether they might also operate in vivo, under natural sunlight illumination [281,282]. Irrespective of their significance in biological systems, the quantum dynamical effects currently under investigation may provide design principles for quantum-enhanced, bio-inspired, light-harvesting systems, and possibly for quantum-engineered systems more generally. These could include AP systems or bio-hybrid systems that combine elements from biological and artificial systems. Importantly however, successful exploitation of quantum coherence in practical, scalable light-harvesting systems will require a design approach that considers EET as only one of many processes essential to system operation and is focused on global system optimality rather than optimality of only the initial stage of light harvesting (see Section 4.3).

155

4.2.2. Nonphotochemical quenching mechanisms The mechanisms of excess energy dissipation through NPQ are another subject of ongoing research. NPQ is not a single process but rather a collection of processes that operate under lightsupersaturated conditions and determine the fates of excess excitons neither used to drive charge separation in reaction centres (‘photochemical quenching’) nor lost through fluorescence or internal conversion. Conventionally, NPQ is divided into three component processes that operate on different time scales as determined by spectroscopic experiments: so-called qI, qT and qE [189]. qI or photoinhibition is caused by either inactivation/damage of PSII reaction centres (see Section 3.3.3) or by a stable quencher in the PSII antenna. This process relaxes slowly, on the time scale of hours, after the period of light supersaturation largely due to the repair of PSII. qT refers to the so-called state transitions, in which a fraction of the major and minor LHCII proteins are shuttled reversibly from PSII to PSI complexes on a time scale of tens of minutes. PSI is downstream from PSII in the electron transport chain (Section 3.1) and is also kinetically slower, so the reallocation of absorbed energy to PSI that is facilitated by qT mitigates the kinetic ‘bottleneck’ that can occur at PSI under high-light conditions. qE or ‘high-energy quenching’ is the most rapid and typically most active NPQ mechanism, activating on a time scale of minutes, redirecting excitons in the PSII antenna to quenching sites thought to be in the antenna. The precise identity of the quenching sites, however, and the exact sequence of events leading to their activation, are yet to be determined. qE is known to be triggered when rapid charge separation by PSII generates a pH gradient across the thylakoid, as this indicates an imbalance between the rates of water oxidation by PSII and H þ flow from the lumen to the stroma via ATP synthase (Fig. 4), which, if not corrected, can result in photoinhibition. Different lines of evidence indicate that qE depends on the presence of xanthophyll molecules such as lutein and zeaxanthin, bound to the PSII antenna proteins, as well as interactions between those proteins and the PsbS protein in vascular plants and the LhcSR proteins in green algae [189,294–296]. These interdependencies are sufficiently well understood that a detailed kinetic model of qE in higher plants has been able to accurately predict experimental spectroscopic signatures of qE under different light conditions [297]. Nonetheless, mechanistic details underlying the kinetics remain an area of active investigation. Since all photosynthetic energy systems under realistic conditions are subject to varying light levels and susceptible to oxidative damage in their light-harvesting machinery, a complete understanding of NPQ mechanisms and their interactions would be useful in modelling and optimising robust biological and artificial energy systems. This is particularly important for high-density microalgal cell cultures in which cells cycle between light and dark zones, typically on a timescale of seconds. 4.2.3. Structure determination of the light-harvesting machinery The dependence of energetic processes such as EET and NPQ on thylakoid structure makes accurate structural determination important for system engineering. Resolving the 3D organisation of the thylakoid and its nested subsystems requires spatial information over at least five orders of magnitude (10  10–10  6 m) and no single technique can resolve across this range. Therefore, an integrated, multi-technique approach is required [175,298]. Mature techniques exist to enable the collection of structural data from the cellular level (  10–100 nm resolution) to the atomic ( 1–3 Å resolution), in the form of optical microscopy [299], scanning and transmission electron microscopy [300], X-ray and electron crystallography [301], high-resolution single-particle analysis [302–304], NMR [305] and XFELS [255]. Collectively these techniques, together with advanced computational tools for interpreting and integrating multiscale structural data, are

156

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

opening up the possibility of resolving ‘pseudo-atomic’ resolution models of the photosynthetic machinery [298,303]. A multiscale structural atlas of the photosynthetic machinery can help to determine design principles for improved-efficiency bioengineered cells, as well as bio-inspired artificial systems. Due to the huge diversity of structures possible at each characteristic scale in different species and under different environmental conditions, high-throughput methods for determining multiple structures are a significant future goal. Ultimately, for system engineering, general principles relating energetics and structure within and between scales under a variety of conditions are more important than exhaustive characterisation of a single system under particular conditions. Therefore, in addition to cataloguing structures, it is important to study the interactions underlying those structures as this helps to reveal their significance within the space of all possible structures. Efforts to characterise the complex interrelationships between thylakoid composition, multiscale structure and energetics may benefit from an approach that integrates experimental and theoretical studies in the context of complex systems theory. 4.3. Complex-systems analysis and optimisation of photosynthetic systems Generically, a complex system consists of multiple components whose interactions give rise to collective behaviour that is not predictable from properties of the individual components. Such systems typically display multiscale, hierarchical structures, made up of recursively nested complex subsystems [306–308]. System structure and dynamics are usually governed by multiple interacting mechanisms (e.g. light, heat and mass transfer in the case of an energy system) [175,309]. Photosynthetic systems satisfy these criteria, and a systems approach to optimisation has increasingly been advocated [175,310–314]. Beginning at the system scale of a plant stand or algal pond/photobioreactor array (Fig. 6a), hierarchical subsystem nesting proceeds to individual plants, photobioreactors or ponds (Fig. 6a), system elements (e.g. tubes in a tubular photobioreactor) (Fig. 6b), cells (Fig. 6c), chloroplasts (Fig. 6d), thylakoid domains – each comprising a stacked ‘granum’ and its surrounding stroma lamellae (Fig. 6e), photosystem supercomplex arrays/aggregates (Fig. 6f), photosystem supercomplexes (Fig. 6g), pigment–protein complexes (Fig. 6h), protein subunits (Fig. 6i), clusters of strongly coupled chromophores (pigments – Fig. 6j) and finally, individual chromophores (Fig. 6k). Light, heat and mass transfer processes, as well as system growth and maintenance processes operate across these scales. System composition and structure are therefore optimised not for a single process but rather for coexistence and cooperation between these many processes. In a standard linear-process analysis of photosynthetic energetics such as that in Section 3.3.3, the system is partitioned stepwise into subsystems that each complete one of the energy transfer and/or conversion steps spanning the free-energy gradient from incident light to biochemical products. Each process step suffers inefficiencies, dissipating useful energy carriers and/or converting them into waste heat for subsequent dissipation. It is typically assumed that these inefficiencies add linearly to yield the net system inefficiency in light-tochemical energy conversion (Fig. 5). However, the standard linearprocess analysis takes no account of how processes such as light distribution depend simultaneously on parameters spanning a range of scales (e.g. optical properties of pigments, proteins, cells, photobioreactor components and the overall photobioreactor), or of how those parameters depend on each other. The simplified account of light transfer to PPCs given in Section 3.3.3, which assumes a particular, simple system geometry and refers to optical properties at only one scale in the system, is typical of state-of-the-art accounts found in

the literature [184,185,188]. Based on this, strategies to improve light distribution have tended to focus at a single scale (or a relatively small subset of the scales of organisation present in the system), independently tuning compositional and structural parameters such as LHC antenna protein composition [231,278,315], cell culture density [316– 318], and photobioreactor structure and surface properties [278,316– 319], rather than optimising them in a coordinated way. Morever, because these parameters also underly other process steps, tuning them to optimise just one step can have unexpected results for the process overall. For example, LHCs affect light distribution, light absorption, EET, NPQ, quinone transfer through the thylakoid, and possibly also have other functions not yet understood. Furthermore, engineering efforts to date have typically begun from a model algal species and/or photobioreactor design and improved system performance incrementally by tuning only a small number of parameters. The predominance of approaches focusing on particular parameters at particular scales in particular systems has resulted in a lack of general theory enabling rigorous system optimisation using system-scale objectives and constraints (e.g. optimising the lifecycle analysis under local conditions) and engineering parameters at different scales in a coordinated way (e.g. LHC antenna composition, culture cell density, photobioreactor structure and surface properties). Ideally, such a theory would allow the designer to start from a minimal number of fixed parameters and let system configuration emerge organically from an optimisation algorithm. The result may be beyond what is achievable based on intuition, starting from an existing model [324]. AP systems particularly stand to benefit from this approach, since their design is not constrained to start from a naturally occurring organism and, as emphasised in Section 3.5.1, a critical challenge facing AP systems development and commercialisation is integration of catalytic and redox components within macroscopic systems under real environmental and economic conditions. The ongoing rapid development of tools for engineering microalgal systems at all scales, from the thylakoid to overall cultivation systems, means that these systems too are increasingly amenable to model-guided multiscale optimisation. Theoretical tools are available for complex system analysis and optimisation, which naturally suit the hierarchical, multiscale structure of photosynthetic systems. So-called hierarchy theory partitions a complex dynamical system into a hierarchy of recursively nested subsystems with well-separated characteristic length and time scales, to simplify system analysis [175,325,326]. At each scale, the encompassing subsystem is termed the ‘holon’, and its constituent subsystems the ‘components’, which are simultaneously holons at the next scale down (Fig. 6). Multiscale analysis of the hierarchy is done by studying internal properties and dynamics of holons at different scales independently (e.g. the optical properties of a photobioreactor tube filled with cell culture, and the optical properties of an individual cell), as well as interscale couplings between holons (e.g. interdependencies between the optical properties of the photobioreactor tube and the cell). This provides a framework for linking system-scale objectives and constraints to engineerable parameters at different scales within a system. The conceptual framework of hierarchy theory also underpins a range of quantitative multiscale analysis techniques [327–337], which are under ongoing development. Moreover, multiscale analysis has in recent years been extended for multiscale optimisation of complex systems [338–343]. An important concept in this theory is that global system optimality may require local suboptimality of subsystems at some scale(s). This is due to nonlinear scaling; global system performance is often not simply a linear sum over the performances of all subsystems. These and other sophisticated theoretical tools for model-guided design may become increasingly important as researchers grapple with integrating engineering across the broad range of scales characteristic

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

of photosynthetic energy systems, and the need to fast-track system optimisation and scale up. 4.4. Policy frameworks All energy research is conducted within policy frameworks of the international community, national governments, universities and companies, while also being under the scrutiny of public perception, and these can all strongly affect technical outcomes. Policy frameworks based on rigorous science and economics, designed to deliver results over the long term (decades) rather than only the short-term (e.g. electoral cycle), and that have multipartisan support, would be of considerable value because they would be less sensitive to policy changes driven by electoral cycles. In particular, policy developers should consider strategically adjusting drivers to promote sustainable profitability, energy security and safe levels of greenhouse gas emissions, to achieve triple-bottom-line outcomes over the long term. Understanding technology adoption barriers and supporting a technologyagnostic approach focused on achieving these goals would likely be helpful, given the enormity of transitioning to a sustainable world within the tight timelines posed by climate change and fuelsupply limitations. Such long term-focused policies should take seriously the scale and urgency of these problems, and the broader problem of limits to physical growth of the human enterprise on a finite planet. They should consider the implementation of economic tools needed for effective action, such as carbon pricing and the reduction of structural protection for unsustainable industries (e.g. large subsidies to the fossil-fuel industries) by reallocating funds to enterprises that promote sustainability. Such frameworks should assist researchers in addressing the enormous technical challenges described in this section in a more effective and timeefficient manner. With a focus on international agreements and significant stated positions on CO2 emission reductions by the European Union, USA, China, Russia, Mexico, Norway, Switzerland and Gabon, subject to specific terms and agreements, there now appears to be an increasing focus on addressing Climate Change [344,345]. It is also of note that an increasing number of organisations and regions are divesting from fossil fuels [346–348] and that there are reports, such as by HSBC [349], that investments into fossil fuels are becoming less attractive. This, together with reports that, in the UK for example, up to 90% of people would prefer renewable energy [350], and the emergence of social enterprises to facilitate this while incorporating social outcomes into their business models [351–353], are encouraging. However, the scale of the challenge and the rapidly closing window of opportunity for change are daunting and demand urgent action [357–359]. The recent UNFCC COP21 Paris 2015 Climate Change Agreement, although important, is only a step on the path to expediting the necessary development and commercialisation of sustainable energy technologies. Photosynthetic energy systems exist within the suite of such technologies, and face some unique challenges. In particular, it is important to actively interlink research efforts between the broad range of disciplines covered in this article, to fast track development and scale up. This is because photosynthesis research is arguably the best possible example of a scientific field that is intrinsically interdisciplinary [186], and the context of energy systems development adds yet further disciplinary perspectives to consider. This suggests a need for interdisciplinary research centres and collaborations, as well as ‘specialist–nonspecialist’ individual researchers who can work comfortably across disciplinary lines. Leading photosynthesis researchers have recently made a case for a global collaboration tasked with rapid development of high-performance AP systems [354]. This would be comparable to the international project on fusion energy, ITER, or earlier intense research efforts such as the Apollo project. Although our findings

157

here suggest that a broader focus on both bio-engineered and artificial photosynthesis may be in order, we strongly agree that the scale and urgency of the challenge are consistent with a need for a focused global effort.

5. Summary and conclusions This review article has brought together evidence from a broad range of fields, to describe the roles of photosynthesis in current global energy systems and society, as well as its potential future roles in mitigating the effects of climate change, fuel-supply limitations, and aiding the transition to a more sustainable world. Microalgal cultivation systems and artificial photosynthetic systems are promising photosynthetic energy technologies, and we have considered the potential for innovating these to economic scalability within a time frame meaningful to global climate and energy challenges. Solar power is available at Earth’s surface far in excess of global human demand but is diffuse and intermittent. Concentrated, stable storage media such as chemical fuels help to improve solar power’s economic utility. Globally at present, carbon-based fuels directly supply four times more energy at final consumption than electricity. While there is an ongoing transition to electricity, this remains gradual compared with the speed required if electricity is to be the main player in mitigating climate change and energy insecurity. Consequently, economic sectors such as transportation will likely depend on chemical fuels for the foreseeable future. Some analysts are optimistic that growth in unconventional oil will offset inevitable declines in conventional oil. However, the high financial, energetic and environmental costs associated with unconventional oil bring into serious question its capacity to fill this role sustainably. EROI values for all current forms of unconventional oil are already below 10, which is the estimated requirement for a modern, industrial society, and are likely to decline into the future because economic incentives have driven exploitation of the ‘easiest’ oil first. A large-scale transition to fungible, sustainable fuels, such as those derived from current NPP, is therefore critically important. Humanity already appropriates roughly one third of current terrestrial NPP, and this HANPP is estimated to be 60% higher than the sustainable limit. The maximum physical potential of the world’s total land area outside croplands, infrastructure, wilderness and denser forests, to deliver plant-based agro-bioenergy is estimated at 38% of TPES, based on 2010 estimates and current levels of photosynthetic efficiency. Achieving this would roughly double the already-unsustainable global human terrestrial biomass harvest, strongly affecting biodiversity, ecosystems and food supply. A future global society powered sustainably by solar fuels will therefore likely require a combination of increased photosynthetic productivity, through increased production area (eg. in urban areas and on non-arable land), increased photon-conversion efficiency per production area, and increased efficiency of socioeconomic energy utilisation. NPP based on existing plant species and cultivation methods could also be used for carbon-capture-and-storage. In this scenario, by 2050 the potential terrestrial flux, together with natural sinks, is forecast to be able to match current total anthropogenic CO2 emissions and so provide a path to stabilizing atmospheric CO2 concentrations and lower peak global warming. However, as global energy and food demand are projected to rise sharply under a business-as-usual scenario, much larger NPPh and carboncapture-and-storage fluxes will likely be required, for which additional photosynthetic productivity on non-arable land and in existing built up areas could be important. Broadly, there are two approaches to engineering photosynthetic energy systems with the potential to be both

158

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

environmentally sustainable and economically scalable: (1) bioengineering natural photosynthetic systems to reduce inefficiencies and re-purpose overall system functioning for cost-effective fuel production and (2) bio-inspired design of artificial photosynthetic systems based on components already well optimised by evolution in natural systems. Microalgal cultivation systems are rapidly emerging as solar fuel production systems with the potential to offer increased biomass harvest (NPPh) simultaneously with other economic and environmental services. Areal productivities at the lab and pilot scales routinely outperform the highest values recorded for plants, and recent global-systems modelling suggests that many regions of Earth can make significant fractions of their transportation fuel requirements through microalgae production without being limited by land availability. However, recent lifecycle analyses at the scale of demonstration systems have found EROI values around 1. This shows a clear need for system optimisation, to elevate the EROI to a sustainable level above at least 3 and preferably closer to 10. General insights into improving system LCA measures may require meta-analysis of many LCA studies rather than a detailed focus on individual studies, due to the sensitive dependence of LCA results on system particulars. Ultimately, a design approach that allows all components to be optimised together, for overall system performance, would be beneficial. Model-guided in silico design may help to enable rigorous, rapid and cost-effective exploration of the design space, integrated across scales. Integrated systems engineering is also important for AP systems, which remain at an earlier stage of development than microalgal systems but show promise for the longer term. Most AP development to date has focused on photocatalytic water splitting to produce hydrogen, for which there will likely be slow uptake as an energy source (spanning decades), relative to carbon-based fuels. Carbon-fixing AP systems remain at a very early stage of development but are making progress. A central challenge facing AP development is to progress rapidly from basic science to integrated systems engineering. Substantial work remains to be done, both on redox and catalytic components, and their integration. Since biological photosynthetic systems, such as higher plants, have also faced these developmental challenges through evolution, better understanding relationships between their compositions, structures and energetics across scales of organisation may help to enable guided design of practical AP systems. Extremely rapid technical progress is required if commercial-scale AP systems are to protect against climate change and provide fuel security on a meaningful time scale. Quantifying composition-structure-dynamics relationships in the chloroplast is a major current research effort spanning a range of scientific disciplines. Current investigations into the presence and significance of quantum coherence in excitation energy transfer (EET), and the mechanistic details of nonphotochemical quenching (NPQ) processes, seek to uncover new design principles for improvedperformance light-harvesting systems. The dependence of EET, NPQ and other energetic processes on thylakoid structure makes accurate structure determination important for system engineering. The ultimate aim is to complete pseudoatomic-resolution 3D atlases of the photosynthetic machinery and photosynthetic organisms, bridging scales from individual atoms to the whole cell, as these could provide detailed parameters for multiscale energetic modelling. The multiscale, hierarchical structures of photosynthetic systems, which extend from the nanoscale of the thylakoid to the macroscale of an overall plant or cultivation system, are naturally suited to multiscale analysis and optimisation techniques that have been developed within the field of complex systems science and engineering. In such systems, composition and structure are optimised not for a single process or mechanism at some particular scale, but rather for coexistence and cooperation between all components and processes serving overall

system functioning. System optimisation therefore requires methods that can quantitatively link global system-scale objectives and constraints with engineerable parameters at different scales within the system, in a coordinated way. These and other sophisticated theoretical tools for model-guided design may become increasingly important as researchers grapple with integrating engineering across the broad range of scales inherent in photosynthetic energy systems. Efforts to address the technical challenges outlined in this review article will inevitably happen under international, governmental and institutional policies that strongly affect technical outcomes. There is a need for rigorously science-based policies with multipartisan support, which incentivise technological innovation in pursuit of triplebottom-line outcomes over the long term. The enormous scale and urgency of threats posed by climate change and fuel-supply limitations may further demand a global collaboration tasked with rapid development of high-performance engineered photosynthetic systems, comparable in scale to the Apollo or ITER projects. At the institutional level there is a need to embrace the intrinsically interdisciplinary nature of photosynthesis research and development by fostering interdisciplinary research collaborations as well as individual researchers who can work comfortably across traditional disciplinary lines. There is much potential for technical breakthroughs in this field and these will be helped by policy settings that allow researchers to make the best possible use of the limited time [31] remaining for meaningful progress.

Acknowledgements We thank Melanie Oey, Ian Ross, and Ivan Kassal for comments on an early draft of the manuscript. Financial support was provided by Australian Research Council Grants CE110001013, FF0776191, DP1093287 and DP0986352, the National Health & Medical Research Council (APP1047243) and via a Dan David Prize doctoral scholarship.

References [1] Graham ER, Fay SA, Davey A, Sanders RW. Intracapsular algae provide fixed carbon to developing embryos of the salamander Ambystoma maculatum. J Exp Biol 2013;216:452–9. [2] Rumpho ME, Pelletreau KN, Moustafa A, Bhattacharya D. The making of a photosynthetic animal. J Exp Biol 2011;214:303–11. [3] Venn AA, Loram JE, Douglas AE. Photosynthetic symbioses in animals. J Exp Bot 2008;59:1069–80. [4] NASA. Earth's energy budget. 〈http://education.gsfc.nasa.gov〉; 2014 [accessed 03.10.14]. [5] Alados I, Foyo-Moreno Iy, Alados-Arboledas L. Photosynthetically active radiation: measurements and modelling. Agric For Meteorol 1996;78:121–31. [6] NASA. Active cavity radiometer irradiance monitor (ACRIM). 〈http://www. acrim.com〉; 2012 [accessed 02.03.13]. [7] USNREL. Reference Solar Spectral Irradiance: Air Mass 1.5. 〈http://rredc.nrel. gov/solar/spectra/am1.5/〉; 2010 [accessed 02.03.13]. [8] Running SW. A measurable planetary boundary for the biosphere. Science 2012;337:1458–9. [9] Haberl H, Erb KH, Krausmann F, Gaube V, Bondeau A, Plutzar C, et al. Quantifying and mapping the human appropriation of net primary production in earth's terrestrial ecosystems. Proc Natl Acad Sci 2007;104:12942–7. [10] Taucher J, Oschlies A. Can we predict the direction of marine primary production change under global warming? Geophys Res Lett 2011;38 (L02603):1–6. [11] BP. Statistical Review of World Energy; June 2011. [12] Peters GP, Marland G, Le Quéré C, Boden T, Canadell JG, Raupach MR. Rapid growth in CO2 emissions after the 2008–2009 global financial crisis. Nat Clim Change 2012;2:1–3. [13] Le Quéré C, Raupach MR, Canadell JG, Marland G, et al. Trends in the sources and sinks of carbon dioxide. Nat Geosci 2009;2:831–6. [14] Ward JD, Mohr SH, Myers BR, Nel WP. High estimates of supply constrained emissions scenarios for long-term climate risk assessment. Energy Policy 2012;51:598–604. [15] Mohr SH, Evans GM. Forecasting coal production until 2100. Fuel 2009;88:2059–67. [16] U.S. Energy Information Administration. International energy outlook; 2010.

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

[17] Mohr SH. Projection of world fossil fuel production with supply and demand interactions [Ph.D. thesis]. University of Newcastle; 2010. [18] International Energy Agency. World energy outlook 2010. Paris: IEA Publications; 2010. [19] International Energy Agency. World energy outlook 2009. Paris: IEA Publications; 2009. [20] Haberl H, Beringer T, Bhattacharya SC, Erb K-H, Hoogwijk M. The global technical potential of bio-energy in 2050 considering sustainability constraints. Curr Opin Environ Sustain 2010;2:394–403. [21] Chassot E, Bonhommeau S, Dulvy NK, Mélin F, Watson R, Gascuel D, et al. Global marine primary production constrains fisheries catches. Ecol Lett 2010;13:495–505. [22] Pauly D, Christensen V. Primary production required to sustain global fisheries. Nature 1995;374:255–7. [23] Dukes JS. Burning buried sunshine: human consumption of ancient solar energy. Clim Change 2003;61:31–44. [24] International Energy Agency. World energy outlook 2000. Paris: IEA Publications; 2000. [25] McGee M. CO2 Now. 〈http://co2now.org〉; 2015 [accessed 21.06.15]. [26] Lal R. Carbon sequestration. Philos Trans R Soc B: Biol Sci 2008;363:815–30. [27] Anderson K, Bows A. Beyond ‘dangerous’ climate change: emission scenarios for a new world. Philos Trans R Soc A 2011;369:20–44. [28] Lenton T. 2 °C or not 2 °C? That is the climate question Nature 2011;473:7. [29] Zickfeld K, Eby M, Matthews HD, Weaver AJ. Setting cumulative emissions targets to reduce the risk of dangerous climate change. Proc Natl Acad Sci 2009;106:16129–34. [30] Govindjee. Chlorophyll a fluorescence: a bit of basics and history. In: Papageorgiou GC, Govindjee, editors. Chlorophyll a fluorescence: a signature of photosynthesis. Springer; 2004. p. 1–42. [31] Kruse O, Rupprecht J, Mussgnug JH, Dismukes GC, Hankamer B. Photosynthesis: a blueprint for solar energy capture and biohydrogen production technologies. Photochem Photobiol Sci 2005;4:957–70. [32] NASA. Sun Fact Sheet. 〈http://nssdc.gsfc.nasa.gov〉; 2014 [accessed 07.04.14]. [33] Smil V. Energy in nature and society: general energetics of complex systems. MIT Press; 2008. [34] Standish EM, Newhall XX, Williams JG, Yeomans DK. Orbital ephemerides of the Sun, Moon, and planets. In: Seidelmann PK, editor. Explanatory supplement to the astronomical almanac. University Science Books; 1992. p. 279–323. [35] Stephens E, Ross IL, Mussgnug JH, Wagner LD, Borowitzka MA, Posten C, et al. Future prospects of microalgal biofuel production systems. Trends Plant Sci 2010;15:554–64. [36] Scholes GD, Fleming GR, Olaya-Castro A, van Grondelle R. Lessons from nature about solar light harvesting. Nat Chem 2011;3:763–74. [37] Australian Bureau of Meteorology. Average daily solar exposure. 〈http:// www.bom.gov.au〉; 2013 [accessed 07.04.13]. [38] Vitousek PM, Ehrlich PR, Ehrlich AH, Matson PA. Human appropriation of the products of photosynthesis. Bioscience 1986;36:368–73. [39] Nemani RR, Keeling CD, Hashimoto H, Jolly WM, Piper SC, Tucker CJ, et al. Climate-driven increases in global terrestrial net primary production from 1982 to 1999. Science 2003;300:1560–3. [40] Zhao M, Running SW. Drought-induced reduction in global terrestrial net primary production from 2000 through 2009. Science 2010;329:940–3. [41] Jenkins B, Baxter L, Miles T. Combustion properties of biomass. Fuel Process Technol 1998;54:17–46. [42] Erb K-H, Krausmann F, Gaube V, Gingrich S, Bondeau A, Fischer-Kowalski M, et al. Analyzing the global human appropriation of net primary production – processes, trajectories, implications. An introduction. Ecol Econ 2009;69:250–9. [43] Koppejan J, van Loo S. The handbook of biomass combustion and co-firing. Routledge; 2012. [44] Haberl H, Erb K-H, Krausmann F, Running S, Searchinger TD, Kolby Smith W. Bioenergy: how much can we expect for 2050? Environ Res Lett 2013;8:031004. [45] Illman AM, Scragg AH, Shales SW. Increase in Chlorella strains calorific values when grown in low nitrogen medium. Enzym Microb Technol 2000;27:631–5. [46] Geider RJ, Roche JL. Redfield revisited: variability of C:N:P in marine microalgae and its biochemical basis. Eur J Phycol 2002;37:1–17. [47] Lenton TM. The potential for land-based biological CO2 removal to lower future atmospheric CO2 concentration. Carbon Manag 2010;1:145–60. [48] Somerville C, Youngs H, Taylor C, Davis SC, Long SP, et al. Feedstocks for lignocellulosic biofuels. Science 2010;329:790–2. [49] Foley JA, Ramankutty N, Brauman KA, Cassidy ES, Gerber JS, Johnston M, et al. Solutions for a cultivated planet. Nature 2012;478:337–42. [50] International Energy Agency. World energy outlook 2012. Paris: IEA Publications; 2012. [51] Supple D. Units and conversions fact sheet. MIT Energy Club; 2007. [52] US Department of Energy. Hydrogen fuel cell engines and related technologies; module 1: hydrogen properties. 〈http://www1.eere.energy.gov/hydro genandfuelcells/tech_validation/pdfs/fcm01r0.pdf〉; 2001 [accessed 22.04.13]. [53] Oak Ridge National Laboratory. Biomass energy data book: edition 4. US Department of Energy; 2011. [54] Cuéllar AD, Webber ME. Cow power: the energy and emissions benefits of converting manure to biogas. Environ Res Lett 2008;3:034002.

159

[55] Klass DL. Biomass for renewable energy, fuels, and chemicals. Academic Press Limited; 1998. [56] Walsh JL, Ross CC, Smith MS, Harper SR, Wilkins WA. Biogas utilization handbook. Georgia Tech Research Institute; 1988. [57] Coronado CR, de Carvalho Jr JA, Silveira JL. Biodiesel CO2 emissions: a comparison with the main fuels in the Brazilian market. Fuel Process Technol 2009;90:204–11. [58] US Department of Energy Alternative Fuels and Advanced Vehicles Data Center. Properties of fuels data table. 〈http://www.afdc.energy.gov〉; 2011 [accessed 22.04.13]. [59] Air BP. Handbook of products; 2008. [60] Nojoumi H, Dincer I, Naterer GF. Greenhouse gas emissions assessment of hydrogen and kerosene-fueled aircraft propulsion. Int J Hydrogen Energy 2009;34:1363–9. [61] Starik AM. Gaseous and particulate emissions with jet engine exhaust and atmospheric pollution. Adv Propuls Technol High-Speed Aircr 2002;15:1–22. [62] Edwards T. Liquid fuels and propellants for aerospace propulsion: 1903– 2003. J Propuls Power 2003;19:1089–107. [63] US Environmental Protection Agency. Emission facts: average carbon dioxide emissions resulting from gasoline and diesel fuel. 〈http://www.epa.gov/oms/ climate/420f05001.htm〉 ; 2011 [accessed May, 2012]. [64] US National Biodiesel Board. Biodiesel energy content fact sheet. 〈http:// www.biodiesel.org/pdf$_$files/fuelfactsheets/BTU$_$Content$_$Final$_ $Oct2005.pdf〉; 2011 [accessed May, 2012]. [65] Shapouri H, Duffield JA, Wang M. The energy balance of corn ethanol revisited. Trans Am Soc Agric Eng 2003;46:959–68. [66] Lett RG, Ruppel TC. Coal, chemical and physical properties. Encycl Energy 2004;1:411–23. [67] Keeling CD. Industrial production of carbon dioxide from fossil fuels and limestone. Tellus 1973;25:174–98. [68] Perry RH, Green D. Perry's chemical engineer's handbook. 6th ed.McGrawHill; 1984. [69] Henry M, Besnard A, Asante WA, Eshun J, Adu-Bredu S, Valentini R, et al. Wood density, phytomass variations within and among trees, and allometric equations in a tropical rainforest of Africa. For Ecol Manag 2010;260:1375–88. [70] Lamlom SH, Savidge RA. A reassessment of carbon content in wood: variation within and between 41 North American species. Biomass Bioenergy 2003;25:381–8. [71] Witkowski ETF, Lamont BB. Leaf specific mass confounds leaf density and thickness. Oecologia 1991;88:486–93. [72] Victor NM, Victor DG. Macro patterns in the use of traditional biomass fuels. Program on Energy and Sustainable Development, Publication Number WP10, Stanford University; 2002. p. 1–35. [73] US Energy Information Administration. International energy outlook 2013. Washington, DC: US Department of Energy; 2013. [74] US Energy Information Administration. Frequently asked questions. 〈http:// www.eia.gov/tools/faqs/〉; 2014 [accessed 16.05.14]. [75] Huang W-D, Zhang Y-HP. Energy efficiency analysis: biomass-to-wheel efficiency related with biofuels production, fuel distribution, and powertrain systems. PLoS One 2011;6:e22113. [76] Norwegian Water Resources and Energy Directorate. Energy in Norway 2009. 〈http://www.nve.no/Global/Energi/Analyser/Energi i Norge folder/Energy in Norway 2010 edition.pdf〉; 2009 [accessed July, 2012]. [77] Rogner H-H. Energy resources. In: Toth FL, editor. Energy for development. Springer Netherlands; 2012. p. 149–60. [78] Carbon Dioxide Information Analysis Center. Carbon Dioxide Information Analysis Center – Conversion Tables. 〈http://cdiac.ornl.gov/pns/convert. html〉; 2013 [accessed February, 2013]. [79] Miller RG, Sorrell SR. The future of oil supply. Philos Trans R Soc A: Math, Phys Eng Sci 2013;372:20130179. [80] Goswami DY. A review and future prospects of renewable energy in the global energy system. Proceedings of ISES World Congress 2007, I–V. Berlin, Heidelberg: Springer; 2009. [81] Whittaker RH, Likens GE. Primary production: the biosphere and man. Hum Ecol 1973;1:357–69. [82] Wright DH. Human impacts on the energy flow through natural ecosystems, and implications for species endangerment. Ambio 1990;19:189–94. [83] Rojstaczer S, Sterling SM, Moore NJ. Human appropriation of photosynthesis products. Science 2001;294:2549–52. [84] Imhoff ML, Bounoua L, Ricketts T, Loucks C, Harriss R, Lawrence WT. Global patterns in human consumption of net primary production. Nature 2004;429:870–3. [85] Krausmann F, Erb K-H, Gingrich S, Haberl H, Bondeau A, Gaube V, et al. Global human appropriation of net primary production doubled in the 20th century. Proc Natl Acad Sci 2013;110:10324–9. [86] Halpern BS, Walbridge S, Selkoe KA, Kappel CV, Micheli F, D'Agrosa C, et al. A global map of human impact on marine ecosystems. Science 2008;319: 948–52. [87] Bishop JDK, Amaratunga GAJ, Rodriguez C. Quantifying the limits of HANPP and carbon emissions which prolong total species well-being. Environ Dev Sustain 2009;12:213–31. [88] Rockström J, Steffen W, Noone K, Persson A, Chapin FS, Lambin EF, et al. A safe operating space for humanity. Nature 2009;461:472–5. [89] Wackernagel M, Schulz NB, Deumling D, Linares AC, Jenkins M, Kapos V, et al. Tracking the ecological overshoot of the human economy. Proc Natl Acad Sci USA 2002;99:9266–71.

160

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

[90] Giampietro M, Mayumi K. The biofuel delusion: the fallacy of large scale agro-biofuel production. Earthscan; 2009. [91] Smil V. Energy myths and realities: bringing science to the energy policy debate. AEI Press; 2010. [92] US Energy Information Administration. International energy outlook 2011. Washington, DC: US Department of Energy; 2011. [93] Cardinale BJ, Duffy JE, Gonzalez A, Hooper DU, Perrings C, Venail P, et al. Biodiversity loss and its impact on humanity. Nature 2012;486:59–67. [94] Bruno JF, O'Connor MI. Cascading effects of predator diversity and omnivory in a marine food web. Ecol Lett 2005;8:1048–56. [95] Wang C. Connecting carbon capture with oceanic biomass production. Nat Precedings 2010. [96] Wei N, Quarterman J, Jin Y-S. Marine macroalgae: an untapped resource for producing fuels and chemicals. Trends Biotechnol 2013;31:70–7. [97] Trent J, Wiley P, Tozzi S, McKuin B, Reinsch S. Research spotlight: the future of biofuels: is it in the bag? Biofuels 2012;3:521–4. [98] Alexander L, Allen S, Bindoff NL, Bréon FM, Church J, Cubasch U, et al. Working Group I contribution to the IPCC fifth assessment report. Climate Change; 2013. [99] Houghton J. Global warming: the complete briefing. Cambridge University Press; 2009. [100] Parry M, Palutikof J, Hanson C, Lowe J. Squaring up to reality. Nat Rep Clim Change 2008:68–71. [101] Lackner KS, Brennan S, Matter JM, Park AHA, Wright A, Van Der Zwaan B. The urgency of the development of CO2 capture from ambient air. Proc Natl Acad Sci 2012;109:13156–62. [102] Rogelj J, Hare W, Lowe J, van Vuuren DP, Riahi K, Matthews B, et al. Emission pathways consistent with a 2 °C global temperature limit. Nat Clim Change 2011;1:413–8. [103] Stocker TF. The closing door of climate targets. Science 2013;339:280–2. [104] McGee M. CO2 Now. 〈http://co2now.org〉 [accessed]. [105] Kreider JF, Curtiss PS. Comprehensive evaluation of impacts from potential future automotive fuel replacements. In: Proceedings of the ASME 2007 energy sustainability conference. American Society of Mechanical Engineers; 2007. [106] Verbruggen A, Al Marchohi M. Views on peak oil and its relation to climate change policy. Energy Policy 2010;38:5572–81. [107] Hughes L, Rudolph J. Future world oil production: growth, plateau, or peak? Curr Opin Environ Sustain 2011;3:225–34. [108] Hall CAS, Klitgaard KA, Hall CAS, Klitgaard KA. Energy and the wealth of nations: understanding the biophysical economy. New York, NY: Springer; 2011. [109] Heinberg R. The party's over: oil, war and the fate of industrial societies. Clairview Books; 2005. [110] Fantazzini D, Höök M, Angelantoni A. Global oil risks in the early 21st century. Energy Policy 2011;39:7865-–73. [111] Chapman I. The end of peak oil? Why this topic is still relevant despite recent denials Energy Policy 2014;64:93–101. [112] Lutz C, Lehr U, Wiebe KS. Economic effects of peak oil. Energy Policy 2012;48:829–34. [113] Murray J, King D. Climate policy: oil's tipping point has passed. Nature 2012;481:433–5. [114] Hubbert MK. Nuclear energy and the fossil fuels. Drilling and production practice. American Petroleum Institute; 1956. [115] Sorrell S, Speirs J. Using growth curves to forecast regional resource recovery: approaches, analytics and consistency tests. Philos Trans R Soc A: Math, Phys Eng Sci 2013;372:20120317. [116] Meng QY, Bentley RW. Global oil peaking: responding to the case for ‘abundant supplies of oil’. Energy 2008;33:1179–84. [117] International Energy Agency. World energy outlook 2013. Paris: IEA Publications; 2013. [118] Kerr RA. Peak oil production may already be here. Science 2011;331:1510. [119] Kumhof M, Muir D. Oil and the world economy: some possible futures. Philos Trans R Soc A: Math, Phys Eng Sci 2013;372:20120327. [120] Höök M, Fantazzini D, Angelantoni A, Snowden S. Hydrocarbon liquefaction: viability as a peak oil mitigation strategy. Philos Trans R Soc A: Math, Phys Eng Sci 2013;372:20120319. [121] Kerr RA. Natural gas from shale bursts onto the scene. Science 2010;328:1624–6. [122] Rü hl C, Giljum J. BP energy outlook 2030. Energy 2011;2030:2. [123] Hughes JD. Energy: a reality check on the shale revolution. Nature 2013;494:307–8. [124] King CW, Hall CA. Relating financial and energy return on investment. Sustainability 2011;3:1810–32. [125] Hall CAS, Balogh S, Murphy DJR. What is the minimum EROI that a sustainable society must have? Energies 2009;2:25–47. [126] Murphy DJ. The implications of the declining energy return on investment of oil production. Philos Trans R Soc A: Math, Phys Eng Sci 2013;372:20130126. [127] Cleveland CJ, O’Connor PA. Energy Return on Investment (EROI) of oil shale. Sustainability 2011;3:2307–22. [128] Murphy DJ, Hall CA. Year in review – EROI or energy return on (energy) invested. Ann N Y Acad Sci 2010;1185:102–18. [129] Hall CAS, Lambert JG, Balogh SB. EROI of different fuels and the implications for society. Energy Policy 2013:1–12. [130] Quinn JC, Smith TG, Downes CM, Quinn C. Microalgae to biofuels lifecycle assessment – multiple pathway evaluation. Algal Res 2014;4:116–22.

[131] Oehlschlaeger MA, Wang H, Sexton MN. Prospects for biofuels: a review. J Therm Sci Eng Appl 2013;5:021006-1–9. [132] Pickard WF. Energy return on energy invested (EROI): a quintessential but possibly inadequate metric for sustainability in a solar-powered world? Proc IEEE 2014;102:1118–22. [133] Smil V. Energy transitions: history, requirements, prospects. ABC-CLIO; 2010. [134] Fischer M, Werber M, Schwartz PV. Batteries: higher energy density than gasoline? Energy Policy 2009;37:2639–41. [135] Hedenus F, Karlsson S, Azar C, Sprei F. Cost-effective energy carriers for transport – the role of the energy supply system in a carbon-constrained world. Int J Hydrogen Energy 2010;35:4638–51. [136] Rye L, Blakey S, Wilson CW. Sustainability of supply or the planet: a review of potential drop-in alternative aviation fuels. Energy Environ Sci 2010;3: 17–27. [137] Warshay B, Pan J, Sgouridis S. Aviation industry’s quest for a sustainable fuel: considerations of scale and modal opportunity carbon benefit. Biofuels 2011;2:33–58. [138] Van Noorden R. The rechargeable revolution: a better battery. Nat Chem 2014;507:26–8. [139] Higgins A, Paevere P, Gardner J, Quezada G. Combining choice modelling and multi-criteria analysis for technology diffusion: an application to the uptake of electric vehicles. Technol Forecast Soc Change 2012;79:1399–412. [140] Nemry F, Brons M. Plug-in hybrid and battery electric vehicles. Market penetration scenarios of electric drive vehicles. European Commission Joint Research Centre – Institute for Prospective Technological Studies; 2010. [141] Battaglini A, Lilliestam J, Haas A, Patt A. Development of SuperSmart grids for a more efficient utilisation of electricity from renewable sources. J Clean Prod 2009;17:911–8. [142] Verbruggen A, Lauber V. Basic concepts for designing renewable electricity support aiming at a full-scale transition by 2050. Energy Policy 2009;37:5732–43. [143] Service RF. The hydrogen backlash: toward a hydrogen economy. Science 2004;305:958–61. [144] Hotten R. Paris motor show: toyota's plans for a fuel cell future. BBC News Business: BBC; 2014. [145] Lehmann J. A handful of carbon. Nature 2007;447:143–4. [146] Sayre R. Microalgae: the potential for carbon capture. BioScience 2010;60:722–7. [147] Ke YJ, Hu XY, Yi Q, Yu Z. [Impacts of rice straw biochar on organic carbon and CO2 release in arable soil]. Huan jing ke xue ¼ Huanjing 2014;35:93–9. [148] Marchetti C. Primary energy substitution models: on the interaction between energy and society. Technol Forecast Soc Change 1977;10:345–56. [149] Unruh GC. Understanding carbon lock-in. Energy Policy 2000;28:817–30. [150] Cleveland CJ, Kaufmann RK, Stern DI. Aggregation and the role of energy in the economy. Ecol Econ 2000;32:301–17. [151] Hankamer B, Barber J, Nield J. Structural analysis of the photosystem II coreantenna holocomplex by electron microscopy. In: Wydrzynski TJ, Satoh K, editors. Photosystem II. Netherlands: Springer; 2005. p. 403–24. [152] Rochaix J-D. Regulation of photosynthetic electron transport. Biochim Biophys Acta (BBA) – Bioenerg 2011;1807:375–83. [153] Renger G, Renger T, Photosystem II. The machinery of photosynthetic water splitting. Photosynth Res 2008;98:53–80. [154] Barros T, Kühlbrandt W. Crystallisation, structure and function of plant lightharvesting Complex II. Biochim Biophys Acta (BBA) – Bioenerg 2009;1787:753–72. [155] de Bianchi S, Dall'Osto L, Tognon G, Morosinotto T, Bassi R. Minor antenna proteins CP24 and CP26 affect the interactions between photosystem II subunits and the electron transport rate in grana membranes of arabidopsis. Plant Cell 2008;20:1012–28. [156] Ginsberg NS, Davis JA, Ballottari M. Solving structure in the CP29 light harvesting complex with polarization-phased 2D electronic spectroscopy. Proc Natl Acad Sci 2011;108:3848–53. [157] Tokutsu R, Kato N, Bui KH, Ishikawa T. Revisiting the supramolecular organization of photosystem II in Chlamydomonas reinhardtii. J Biol Chem 2012. [158] Iwai M, Takizawa K, Tokutsu R, Okamuro A. Isolation of the elusive supercomplex that drives cyclic electron flow in photosynthesis. Nat Chem 2010. [159] Taiz L, Zeiger E. Plant physiology. Sinauer Associates Incorporated; 2010. [160] International Energy Agency. World energy outlook 2011. Paris: IEA Publications; 2011. [161] Madson PW, Monceaux DA. Fuel ethanol production. Katzen International, Inc.; 2000. [162] Mulder K, Hagens NJ. Energy return on investment: toward a consistent framework. AMBIO: J Hum Environ 2008;37:74–9. [163] Triana CAR. Energetics of Brazilian ethanol comparison between assessment approaches. Energy Policy 2011;39:4605–13. [164] Hall CAS, Dale BE, Pimentel D. Seeking to understand the reasons for different energy return on investment (EROI) estimates for biofuels. Sustainability 2011;3:2413–32. [165] Toor SS, Rosendahl L, Rudolf A. Hydrothermal liquefaction of biomass: a review of subcritical water technologies. Energy 2011;36:2328–42. [166] Barreiro DL, Prins W, Ronsse F, Brilman W. Hydrothermal liquefaction (HTL) of microalgae for biofuel production: state of the art review and future prospects. Biomass Bioenergy 2013;53:113–27. [167] Levin DB, Pitt L, Love M. Biohydrogen production: prospects and limitations to practical application. Int J Hydrogen Energy 2004;29:173–85.

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

[168] Melis A, Zhang L, Forestier M, Ghirardi ML, Seibert M. Sustained photobiological hydrogen gas production upon reversible inactivation of oxygen evolution in the green alga Chlamydomonas reinhardtii. Plant Physiol 2000;122:127–36. [169] Benson EE, Kubiak CP, Sathrum AJ, Smieja JM. Electrocatalytic and homogeneous approaches to conversion of CO2 to liquid fuels. Chem Soc Rev 2008;38:89. [170] Regalbuto JR. Cellulosic biofuels – got Gasoline? Science 2009;325:822–4. [171] Schirmer A, Rude MA, Li X, Popova E, del Cardayre SB. Microbial biosynthesis of alkanes. Science 2010;329:559–62. [172] Regalbuto JR. The sea change in US biofuels funding: from cellulosic ethanol to green gasoline. Biofuels, Bioprod Biorefin 2011;5:495–504. [173] Gust D, Moore TA, Moore AL. Solar fuels via artificial photosynthesis. Acc Chem Res 2009;42:1890–8. [174] Sharkey TD. Photosynthesis in intact leaves of C3 plants: physics, physiology and rate limitations. Bot Rev 1985;51:53–105. [175] Ringsmuth AK. Multiscale analysis and optimisation of photosynthetic solar energy systems [Ph.D. thesis]. The University of Queensland; 2013. [176] Terashima I, Hikosaka K. Comparative ecophysiology of leaf and canopy photosynthesis. Plant, Cell Environ 1995;18:1111–28. [177] Ellis RJ. Biochemistry: tackling unintelligent design. Science 2010;463:164–5. [178] Spreitzer RJ, Salvucci ME. RUBISCO: structure, regulatory interactions, and possibilities for a better enzyme. Annu Rev Plant Biol 2002;53:449–75. [179] Parry M, Andralojc PJ, Scales JC, Salvucci ME, Carmo-Silva AE, Alonso H, et al. Rubisco activity and regulation as targets for crop improvement. J Exp Bot 2013;64:717–30. [180] Raven JA. Rubisco: still the most abundant protein of Earth? New Phytol 2013;198:1–3. [181] Zarzycki J, Axen SD, Kinney JN, Kerfeld CA. Cyanobacterial-based approaches to improving photosynthesis in plants. J Exp Bot 2013;64:787–98. [182] Davis SC, LeBauer DS, Long SP. Light to liquid fuel: theoretical and realized energy conversion efficiency of plants using Crassulacean Acid Metabolism (CAM) in arid conditions. J Exp Bot 2014;65:3471–8. [183] Zhu X-G, Long SP, Ort DR. What is the maximum efficiency with which photosynthesis can convert solar energy into biomass? Curr Opin Biotechnol 2008;19:153–9. [184] Zhu X-G, Long SP, Ort DR. Improving photosynthetic efficiency for greater yield. Annu Rev Plant Biol 2010;61:235–61. [185] Weyer KM, Bush DR, Darzins A, Willson BD. Theoretical maximum algal oil production. Bioenergy Res 2010;3:204–13. [186] Blankenship RE. Molecular mechanisms of photosynthesis. John Wiley & Sons; 2013. [187] Jennings RC, Engelmann E, Garlaschi F, Casazza AP, Zucchelli G. Photosynthesis and negative entropy production. Biochim Biophys Acta (BBA) – Bioenerg 2005;1709:251–5. [188] Melis A. Solar energy conversion efficiencies in photosynthesis: minimizing the chlorophyll antennae to maximize efficiency. Plant Sci 2009;177:272–80. [189] Ruban AV, Johnson MP, Duffy CDP. The photoprotective molecular switch in the photosystem II antenna. Biochim Biophys Acta 2012;1817:167–81. [190] Melis A, Neidhardt J, Benemann JR. Dunaliella salina (Chlorophyta) with small chlorophyll antenna sizes exhibit higher photosynthetic productivities and photon use efficiencies than normally pigmented cells. J Appl Phycol 1998;10:515–25. [191] Henriques FS. Leaf chlorophyll fluorescence: background and fundamentals for plant biologists. Bot Rev 2009;75:249–70. [192] Jennings RC, Bassi R, Garlaschi FM, Dainese P, Zucchelli G. Distribution of the chlorophyll spectral forms in the chlorophyll-protein complexes of photosystem II antenna. Biochemistry 1993;32:3203–10. [193] Fan D-Y, Jia H, Barber J, Chow WS. Novel effects of methyl viologen on photosystem II function in spinach leaves. Eur Biophys J 2009;39:191–9. [194] Blankenship RE, Tiede DM, Barber J, Brudvig GW, Fleming G, Ghirardi M, et al. Comparing photosynthetic and photovoltaic efficiencies and recognizing the potential for improvement. Science 2011;332:805–9. [195] Melis A. Excitation energy transfer: functional and dynamic aspects of Lhc (cab) proteins. Oxygenic photosynthesis: the light reactions. Springer; 2004. p. 523–38. [196] Kern J, Renger G. Photosystem II: structure and mechanism of the water: plastoquinone oxidoreductase. Photosynth Res 2007;94:183–202. [197] Stryer L, Berg JM, Tymoczko JL. Biochemistry. W.H. Freeman; 2007. [198] Larkum AWD, Ross IL, Kruse O, Hankamer B. Selection, breeding and engineering of microalgae for bioenergy and biofuel production. Trends Biotechnol 2012;30:198–205. [199] Radzun KA, Wolf J, Jakob G, Zhang E, Stephens E, Ross I, et al. Automated nutrient screening system enables high-throughput optimisation of microalgae production conditions. Appl Microbiol Biotechnol 2015;8:31. [200] Brodie J, Lewis J. Unravelling the algae: the past, present, and future of algal systematics. CRC Press; 2007. [201] Singh B, Guldhe A, Rawat I, Bux F. Towards a sustainable approach for development of biodiesel from plant and microalgae. Renew Sustain Energy Rev 2014;29:216–45. [202] Bahadar A, Khan MB. Progress in energy from microalgae: a review. Renew Sustain Energy Rev 2013;27:128–48. [203] Brennan L, Owende P. Biofuels from microalgae – a review of technologies for production, processing, and extractions of biofuels and co-products. Renew Sustain Energy Rev 2010;14:557–77. [204] Ziolkowska JR, Simon L. Recent developments and prospects for algae-based fuels in the US. Renew Sustain Energ Rev 2014;29:847–53.

161

[205] Makareviciene V, Skorupskaite V, Andruleviciute V. Biodiesel fuel from microalgae-promising alternative fuel for the future: a review. Appl Microbiol Biotechnol 2013;12:119–30. [206] Kirrolia A, Bishnoi NR, Singh R. Microalgae as a boon for sustainable energy production and its future research & development aspects. Renew Sustain Energy Rev 2013;20:642–56. [207] Day JG, Slocombe SP, Stanley MS. Overcoming biological constraints to enable the exploitation of microalgae for biofuels. Bioresour Technol 2012;109:245–51. [208] Stephenson PG, Moore CM, Terry MJ, Zubkov MV, Bibby TS. Improving photosynthesis for algal biofuels: toward a green revolution. Trends Biotechnol 2011;29:615–23. [209] Singh A, Nigam PS, Murphy JD. Renewable fuels from algae: an answer to debatable land based fuels. Bioresour Technol 2011;102:10–6. [210] Mata TM, Martins AA, Caetano NS. Microalgae for biodiesel production and other applications: a review. Renew Sustain Energy Rev 2010;14:217–32. [211] Larkum A. Limitations and prospects of natural photosynthesis for bioenergy production. Curr Opin Biotechnol 2010;21:271–6. [212] Barber J. Photosynthetic energy conversion: natural and artificial. Chem Soc Rev 2008;38:185–96. [213] Pienkos PT, Darzins A. The promise and challenges of microalgal-derived biofuels. Biofuels, Bioprod Biorefin 2009;3:431–40. [214] Beer LL, Boyd ES, Peters JW, Posewitz MC. Engineering algae for biohydrogen and biofuel production. Curr Opin Biotechnol 2009;20:264–71. [215] Hambourger M, Moore GF, Kramer DM, Gust D, Moore AL, Moore TA. Biology and technology for photochemical fuel production. Chem Soc Rev 2008;38:25–35. [216] National Centre for Biotechnology Information. Genome. 〈http://www.ncbi. nlm.nih.gov/genomes〉; 2014 [accessed 18.11.14]. [217] Griffiths MJ, Harrison STL. Lipid productivity as a key characteristic for choosing algal species for biodiesel production. J Appl Phycol 2009;21:493–507. [218] Rockström J, Steffen W, Noone K, Persson Å, Chapin FS, Lambin EF, et al. A safe operating space for humanity. Nature 2009;461:472–5. [219] Biller P, Ross AB, Skill SC, Lea-Langton A, Balasundaram B, Hall C, et al. Nutrient recycling of aqueous phase for microalgae cultivation from the hydrothermal liquefaction process. Algal Res 2012;1:70–6. [220] Stephens E, Wagner L, Ross IL, Hankamer B. Microalgal production systems – global impact of industry scale up. In: Posten C, Walter C, editors. Microalgal biotechnology: integration and economy. Berlin/Boston: De Gruyter; 2012. p. 267–301. [221] Moody JW, McGinty CM, Quinn JC. Global evaluation of biofuel potential from microalgae. Proc Natl Acad Sci 2014;111:8691–6. [222] Stephens E, Ross IL, King Z, Mussgnug JH, Kruse O, Posten C, et al. An economic and technical evaluation of microalgal biofuels. Nat Biotechnol 2010;28:126–8. [223] Colosi LM, Zhang Y, Clarens AF, White MA. Will algae produce the green? Using published life cycle assessments as a starting point for economic evaluation of future algae-to-energy systems Biofuels 2012;3:129–42. [224] Beal CM, Stillwell AS, King CW, Cohen SM, Berberoglu H, Bhattarai RP, et al. Energy return on investment for algal biofuel production coupled with wastewater treatment. Water Environ Res 2012;84:692–710. [225] Grierson S, Strezov V. Life cycle assessment of the microalgae biofuel value chain: a critical review of existing studies. Bionature 2012. In: Proceedings of the third international conference on bioenvironment, biodiversity and renewable energies; 2012. p. 1–6. [226] Azadi P, Brownbridge G, Mosbach S, Smallbone A, Bhave A, Inderwildi O, et al. The carbon footprint and non-renewable energy demand of algaederived biodiesel. Appl Energy 2014;113:1632–44. [227] Zaimes GG, Khanna V. Microalgal biomass production pathways: evaluation of life cycle environmental impacts. Environ Prog Sustain Energy 2013;6:88. [228] Vasudevan V, Stratton RW, Pearlson MN. Environmental performance of algal biofuel technology options. Environ Sci Technol 2012;46:2451–9. [229] Orfield ND, Fang AJ, Valdez PJ, Nelson MC, Savage PE, Lin XN, et al. Life cycle design of an algal biorefinery featuring hydrothermal liquefaction: effect of reaction conditions and an alternative pathway including microbial regrowth. ACS Sustain Chem Eng 2014;2:867–74. [230] Oey M, Ross IL, Stephens E, Steinbeck J, Wolf J, Radzun KA, et al. RNAi knockdown of LHCBM1, 2 and 3 increases photosynthetic H2 production efficiency of the green alga Chlamydomonas reinhardtii. PLoS One 2013;8:e61375. [231] Ort DR, Zhu X, Melis A. Optimizing antenna size to maximize photosynthetic efficiency. Plant Physiol 2011;155:79–85. [232] Gust D, Moore TA, Moore AL. Realizing artificial photosynthesis. Faraday Discuss 2012;155:9–26. [233] Magnuson A, Styring S. Molecular chemistry for solar fuels: from natural to artificial photosynthesis. Aust J Chem 2012;65:564–72. [234] Berardi S, Drouet S, Francàs L, Gimbert-Suriñach C, Guttentag M, Richmond C, et al. Molecular artificial photosynthesis. Chem Soc Rev 2014. [235] Kim JH, Nam DH, Park CB. Nanobiocatalytic assemblies for artificial photosynthesis. Curr Opin Biotechnol 2014;28:1–9. [236] Sherman BD, Vaughn MD, Bergkamp JJ, Gust D. Evolution of reaction center mimics to systems capable of generating solar fuel. Photosynth Res 2014. [237] Barber J, Tran PD. From natural to artificial photosynthesis. J R Soc Interface 2013;10:20120984. [238] Dutton PL, Moser CC. Engineering enzymes. Faraday Discuss 2011;148:443–8. [239] Peter LM. Towards sustainable photovoltaics: the search for new materials. Philos Trans R Soc A: Math, Phys Eng Sci 2011;369:1840–56.

162

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

[240] Schreier M, Curvat L, Giordano F, Steier L, Abate A, Zakeeruddin SM, et al. Efficient photosynthesis of carbon monoxide from CO2 using perovskite photovoltaics. Nat Commun 2015;6:7326. [241] Luo J, Im J-H, Mayer MT, Schreier M, Nazeeruddin MK, Park N-G, et al. Water photolysis at 12.3% efficiency via perovskite photovoltaics and Earthabundant catalysts. Science 2014;365:1593–6. [242] Du P, Eisenberg R. Catalysts made of earth-abundant elements (Co, Ni, Fe) for water splitting: recent progress and future challenges. Energy Environ Sci 2012;5:6012–21. [243] Reece SY, Hamel JA, Sung K, Jarvi TD, Esswein AJ, Pijpers JJ, et al. Wireless solar water splitting using silicon-based semiconductors and earth-abundant catalysts. Science 2011;334:645–8. [244] Dincă M, Surendranath Y. Nickel-borate oxygen-evolving catalyst that functions under benign conditions. Proc Natl Acad Sci 2010;107: 10337–41. [245] Torella JP, Gagliardi CJ, Chen JS, Bediako DK, Colón B, Way JC, et al. Efficient solar-to-fuels production from a hybrid microbial-water-splitting catalyst system. Proc Natl Acad Sci 2015. [246] Tachibana Y, Vayssieres L, Durrant JR. Artificial photosynthesis for solar water-splitting. Nat Photon 2012;6:511–8. [247] Maeda K, Domen K. Photocatalytic water splitting: recent progress and future challenges. J Phys Chem Lett 2010;1:2655–61. [248] McEvoy JP, Brudvig GW. Water-splitting chemistry of photosystem II. Chem Rev 2006;106:4455–83. [249] Dau H, Zaharieva I, Haumann M. Recent developments in research on water oxidation by photosystem II. Curr Opin Chem Biol 2012;16:3–10. [250] Shen J-R. The structure of photosystem II and the mechanism of water oxidation in photosynthesis. Annu Rev Plant Biol 2015;66 150306101137005. [251] McEvoy JP, Brudvig GW. Water-splitting chemistry of photosystem II; 2006. [252] Duan L, Bozoglian F, Mandal S, Stewart B, Privalov T, Llobet A, et al. A molecular ruthenium catalyst with water-oxidation activity comparable to that of photosystem II. Nat Chem 2012;4:418–23 [Nature Publishing Group]. [253] Umena Y, Kawakami K, Shen J-R, Kamiya N. Crystal structure of oxygenevolving photosystem II at a resolution of 1.9 Å. Nature 2012;473:55–60. [254] Waldrop MM. X-ray science: the big guns. Nature 2014;505:604–6. [255] Chapman HN, Fromme P, Barty A, White TA, Kirian RA, Aquila A, et al. Femtosecond X-ray protein nanocrystallography. Nat Chem 2011;470:73–7. [256] Friedrich B, Fritsch J, Lenz O. Oxygen-tolerant hydrogenases in hydrogenbased technologies. Curr Opin Biotechnol 2011;22:358–64. [257] Frey M. Hydrogenases: hydrogen-activating enzymes. ChemBioChem 2002;3:153–60. [258] Nath K, Jajoo A, Poudyal RS, Timilsina R, Park YS, Aro E-M, et al. Towards a critical understanding of the photosystem II repair mechanism and its regulation during stress conditions. FEBS Lett 2013;587:3372–81. [259] Johnson MP, Goral TK, Duffy CDP, Brain APR, Mullineaux CW, Ruban AV. Photoprotective energy dissipation involves the reorganization of photosystem II light-harvesting complexes in the grana membranes of spinach chloroplasts. Plant Cell 2011;23:1468–79. [260] Joya KS, Joya YF, Ocakoglu K, van de Krol R. Water-splitting catalysis and solar fuel devices: artificial leaves on the move. Angew Chem-Int Ed 2013;52:10426–37. [261] Andreiadis ES, Chavarot-Kerlidou M, Fontecave M, Artero V. Artificial photosynthesis: from molecular catalysts for light-driven water splitting to photoelectrochemical cells. Photochem Photobiol 2011;87:946–64. [262] Li H, Opgenorth PH, Wernick DG, Rogers S, Wu TY, Higashide W, et al. Integrated electromicrobial conversion of CO2 to higher alcohols. Science 2012;335:1596. [263] McCool NS, Robinson DM, Sheats JE. A Co4O4 “cubane” water oxidation catalyst inspired by photosynthesis. J Am Chem Soc 2011;133:11446–9. [264] Lubitz W, Reijerse EJ, Messinger J. Solar water-splitting into H2 and O2: design principles of photosystem II and hydrogenases. Energy Environ Sci 2008;1:15–31. [265] Tye JW, Hall MB, Darensbourg MY. Better than platinum? Fuel cells energized by enzymes Proc Natl Acad Sci 2005. [266] Fritsch J, Lenz O, Friedrich B. Structure, function and biosynthesis of O2-tolerant hydrogenases. Nat Rev Microbiol 2013;11:106–14. [267] Mulder DW, Shepard EM, Meuser JE, Joshi N, King PW, Posewitz MC, et al. Insights into [FeFe]-hydrogenase structure, mechanism, and maturation. Structure 2011;19:1038–52. [268] Fritsch J, Scheerer P, Frielingsdorf S, Kroschinsky S, Friedrich B, Lenz O, et al. The crystal structure of an oxygen-tolerant hydrogenase uncovers a novel iron–sulphur centre. Nature 2011;479:249–52. [269] Faiella M, Roy A, Sommer D, Ghirlanda G. De novo design of functional proteins: toward artificial hydrogenases. Pept Sci 2013;100:558–71. [270] Esselborn J, Lambertz C. Spontaneous activation of [FeFe]-hydrogenases by an inorganic [2Fe] active site mimic. Nat Chem Biol 2013;9:607–10. [271] Tcherkez GG, Farqhuar GD, Andrews TJ. Despite slow catalysis and confused substrate specificity, all ribulose bisphosphate carboxylases may be nearly perfectly optimized. Proc Natl Acad Sci 2006;103:7246–51. [272] Parry M, Andralojc PJ, Scales JC. Rubisco activity and regulation as targets for crop improvement. J Exp Bot 2013. [273] van Lun M, van der Spoel D, Andersson I. Subunit interface dynamics in hexadecameric rubisco. J Mol Biol 2011;411:1083–98. [274] Ducat DC, Silver PA. Improving carbon fixation pathways. Curr Opin Chem Biol 2012;16:337–44. [275] Bar-Even A, Noor E, Lewis NE, Milo R. Design and analysis of synthetic carbon fixation pathways. Proc Natl Acad Sci 2010;107:8889–94.

[276] Woolerton TW, Sheard S, Chaudhary YS, Armstrong FA. Enzymes and bioinspired electrocatalysts in solar fuel devices. Energy Environ Sci 2012;5:7470. [277] Mikkelsen M, Jørgensen M, Krebs FC. The teraton challenge. A review of fixation and transformation of carbon dioxide. Energy Environ Sci 2010;3:43. [278] Mussgnug JH, Thomas-Hall S, Rupprecht J, Foo A, Klassen V, McDowall A, et al. Engineering photosynthetic light capture: impacts on improved solar energy to biomass conversion. Plant Biotechnol J 2007;5:802–14. [279] Chenu A, Scholes GD. Coherence in energy transfer and photosynthesis. Annu Rev Phys Chem 2015;66:69–96. [280] Anna JM, Scholes GD, van Grondelle R. A little coherence in photosynthetic light harvesting. BioScience 2014;64:14–25. [281] Fassioli F, Dinshaw R, Arpin PC, Scholes GD. Photosynthetic light harvesting: excitons and coherence. J R Soc Interface 2014;11:20130901. [282] Kassal I, Yuen-Zhou J, Rahimi-Keshari S. Does coherence enhance transport in photosynthesis? J Phys Chem Lett 2013;4:362–7. [283] Ishizaki A, Fleming GR. Quantum coherence in photosynthetic light harvesting. Annu Rev Condens Matter Phys 2012;3:333–61. [284] Mohseni M, Omar Y, Engel GS, Plenio MB. Quantum effects in biology. Cambridge University Press; 2014. [285] Huelga SF, Plenio MB. Vibrations, quanta and biology. Contemp Phys 2013;54:181–207. [286] Lambert N, Chen Y-N, Cheng Y-C, Li C-M, Chen G-Y, Nori F. Quantum biology. Nat Phys 2012;8:1–9. [287] Forster T. Energiewanderung und fluorescenz. Naturwissenschaften 1946;33:166–75. [288] Olaya-Castro A, Scholes GD. Energy transfer from Förster–Dexter theory to quantum coherent light-harvesting. Int Rev Phys Chem 2011;30:49–77. [289] Engel GS, Calhoun TR, Read EL, Ahn T-K, Mančal T, Cheng Y-C, et al. Evidence for wavelike energy transfer through quantum coherence in photosynthetic systems. Nature 2007;446:782–6. [290] Panitchayangkoon G, Hayes D, Fransted KA, Caram JR, Harel E, Wen J, et al. Long-lived quantum coherence in photosynthetic complexes at physiological temperature. Proc Natl Acad Sci 2010;107:12766–70. [291] Collini E, Wong CY, Wilk KE, Curmi PMG, Brumer P, Scholes GD. Coherently wired light-harvesting in photosynthetic marine algae at ambient temperature. Nature 2010;463 644-U69. [292] Romero E, Augulis R, Novoderezhkin VI, Ferretti M, Thieme J, Zigmantas D, et al. Quantum coherence in photosynthesis for efficient solar-energy conversion. Nat Phys 2014;10:676–82. [293] Ringsmuth AK, Milburn GJ, Stace TM. Multiscale photosynthetic and biomimetic excitation energy transfer. Nat Phys 2012;8:562–7. [294] Rochaix J-D. Regulation and dynamics of the light-harvesting system. Annu Rev Plant Biol 2014;65:287–309. [295] Croce R, van Amerongen H. Natural strategies for photosynthetic light harvesting. Nat Chem 2014;10:492–501. [296] Niyogi KK, Truong TB. Evolution of flexible non-photochemical quenching mechanisms that regulate light harvesting in oxygenic photosynthesis. Curr Opin Plant Biol 2013;16:307–14. [297] Zaks J, Amarnath K, Kramer DM, Niyogi KK, Fleming GR. A kinetic model of rapidly reversible nonphotochemical quenching. Proc Natl Acad Sci 2012;109:15757–62. [298] Alber F, Förster F, Korkin D, Topf M. Integrating diverse data for structure determination of macromolecular assemblies. Annu Rev Biochem 2008. [299] Schermelleh L, Heintzmann R, Leonhardt H. A guide to super-resolution fluorescence microscopy. J Cell Biol 2010;190:165–75. [300] Engel BD, Schaffer M, Cuellar LK, Villa E, Plitzko JM, Baumeister W, et al. Native architecture of the Chlamydomonas chloroplast revealed by in situ cryo-electron tomography. eLife 2015;4:e04889. [301] Kühlbrandt W. Combining Cryo-EM and X-ray crystallography to study membrane protein structure and function. In: Kühlbrandt W, editor. NATO science for peace and security series A: chemistry and biology. Dordrecht: Springer; 2011. p. 93–101. [302] Cheng Y, Grigorieff N, Penczek PA, Walz T. A primer to single-particle cryoelectron microscopy. Cell 2015;161:438–49. [303] Orlova EV, Saibil HR. Structural analysis of macromolecular assemblies by electron microscopy. Chem Rev 2011;111:7710–48. [304] Liao M, Cao E, Julius D, Cheng Y. Structure of the TRPV1 ion channel determined by electron cryo-microscopy. Nature 2013. [305] Sakakibara D, Sasaki A, Ikeya T, Hamatsu J, Hanashima T, Mishima M, et al. Protein structure determination in living cells by in-cell NMR spectroscopy. Nat Chem 2009;458:102–5. [306] Clauset A, Moore C, Newman MEJ. Hierarchical structure and the prediction of missing links in networks. Nature 2008;453:98–101. [307] Sales-Pardo M, Guimera R, Moreira AA, Amaral LAN. Extracting the hierarchical organization of complex systems. Proc Natl Acad Sci 2007;104: 15224–9. [308] Ravasz E, Barabási A-L. Hierarchical organization in complex networks. Phys Rev E 2003;67:026112. [309] Li J, Zhang J, Ge W, Liu X. Multi-scale methodology for complex systems. Chem Eng Sci 2004;59:1687–700. [310] Long SP, Marshall-Colon A, Zhu X-G. Meeting the global food demand of the future by engineering crop photosynthesis and yield potential. Cell 2015;161:56–66. [311] Peers G. Increasing algal photosynthetic productivity by integrating ecophysiology with systems biology. Trends Biotechnol 2014:1–5.

A.K. Ringsmuth et al. / Renewable and Sustainable Energy Reviews 62 (2016) 134–163

[312] Weston DJ, Hanson PJ, Norby RJ, Tuskan GA, Wullschleger SD. From systems biology to photosynthesis and whole-plant physiology: a conceptual model for integrating multi-scale networks. Plant Signal Behav 2012;7: 260–2. [313] Band LR, Fozard JA, Godin C, Jensen OE, Pridmore T, Bennett MJ, et al. Multiscale systems analysis of root growth and development: modeling beyond the network and cellular scales. Plant Cell 2012;24:3892–906. [314] Zhou H, Li X, Fan T, Osterloh FE, Ding J, Sabio EM, et al. Artificial inorganic leafs for efficient photochemical hydrogen production inspired by natural photosynthesis. Adv Mater 2010;22:951–6. [315] Murphy TE, Berberoglu H. Effect of algae pigmentation on photobioreactor productivity and scale-up: a light transfer perspective. J Quant Spectrosc Radiat Transf 2011;112:2826–34. [316] Dillschneider R, Posten C. Closed bioreactors as tools for microalgae production. Adv Biofuels Bioprod 2013:629–49. [317] Posten C. Design principles of photo-bioreactors for cultivation of microalgae. Eng Life Sci 2009;9:165–77. [318] Ugwu CU, Aoyagi H, Uchiyama H. Photobioreactors for mass cultivation of algae. Bioresour Technol 2008;99:4021–8. [319] Gordon JM. Tailoring optical systems to optimized photobioreactors. Int J Hydrogen Energy 2002;27:1175–84. [320] Solar Biofuels Consortium. 〈http://solarbiofuels.org〉; 2008 [accessed 04.05.14]. [321] Peters K. Chloroplasts visible in the cells of Plagiomnium affine, the manyfruited thyme moss. In: laminazellen.jpeg Pa, editor. Reproduced persuant to creative commons license. Wikipedia: Wikimedia Commons; 2006. [322] Mussgnug JH. NAB1 is an RNA binding protein involved in the light-regulated differential expression of the light-harvesting antenna of Chlamydomonas reinhardtii. The Plant Cell 2005;17:3409–21. [323] Croce R, van Amerongen H. Light-harvesting and structural organization of photosystem II: from individual complexes to thylakoid membrane. J Photochem Photobiol B: Biol 2011;104:142–53. [324] Wang C, Yu S, Chen W, Sun C. Highly efficient light-trapping structure design inspired by natural evolution. Sci Rep 2013;3:1025-1–8. [325] Lischke H, Loeffler TJ, Thornton PE, Zimmermann NE. Model up-scaling in landscape research. In: Kienast F, Wildi O, Ghosh S, editors. A changing world: challenges for landscape research. Springer; 2007. [326] Ahl V, Allen TFH. Hierarchy theory: a vision, vocabulary, and epistemology. Columbia University Press; 1996. [327] Hoekstra A, Chopard B, Coveney P. Multiscale modelling and simulation: a position paper. Philos Trans R Soc A: Math, Phys Eng Sci 2014;372:20130377. [328] Chopard B, Borgdorff J, Hoekstra AG. A framework for multi-scale modelling. Philos Trans R Soc A: Math, Phys Eng Sci 2014;372:20130378. [329] Kirschner DE, Hunt CA, Marino S, Fallahi-Sichani M, Linderman JJ. Tuneable resolution as a systems biology approach for multi-scale, multi-compartment computational models. WIREs Syst Biol Med 2014. [330] Yang A. On the common conceptual and computational frameworks for multiscale modeling. Ind Eng Chem Res 2013;52:11451–62. [331] Zhao Y, Jiang C, Yang A. Towards computer-aided multiscale modelling: an overarching methodology and support of conceptual modelling. Comput Chem Eng 2012;36:10–21. [332] Vlachos DG. Multiscale modeling for emergent behavior, complexity, and combinatorial explosion. AIChE J 2012;58:1314–25. [333] Visscher L, Bolhuis P, Bickelhaupt FM. Multiscale modelling. 2011;13:10399. [334] Dada JO, Mendes P. Multi-scale modelling and simulation in systems biology. Integr Biol 2011;3:86–96. [335] Chung PS, Jhon MS, Biegler LT. The holistic strategy in multi-scale modeling. Adv Chem Eng 2011;40:59–118. [336] Engquist B, Lötstedt P, Runborg O. Multiscale modeling and simulation in science. Springer Science & Business Media; 2009.

163

[337] Woods CJ, Mulholland AJ. Multiscale modelling of biological systems. Chem Model 2008;5:13–50. [338] Ouyang Q, Chen X, Yao W. Sequential probabilistic analytical target cascading method for hierarchical multilevel optimization under uncertainty. Struct Multidisc Optim 2014;49:267–80. [339] Gardenghi M, Wiecek MM, Wang W. Biobjective optimization for analytical target cascading: optimality vs. achievability. Struct Multidiscip Optim 2012;47:111–33. [340] Sakalkar V, Hajela P. Multilevel decomposition based nondeterministic design optimization for structural systems. Adv Eng Softw 2011;42:1–11. [341] de Wit A, van Keulen F. Framework for multi-level optimization of complex systems. In: Borst Rd, Ramm E, editors. Multiscale methods in computational mechanics. Dordrecht: Springer Netherlands; 2010. p. 347–77. [342] Xiong F, Yin X, Chen W, Yang S. Enhanced probabilistic analytical target cascading with application to multi-scale design. Eng Optim 2010;42:581–92. [343] Allison JT, Kokkolaras M, Papalambros PY. Optimal partitioning and coordination decisions in decomposition-based design optimization. J Mech Des 2009;131:081008. [344] United Nations Framework Convention on Climate Change. INDCs (Intended Nationally Determined Contributions) as communicated by Parties. Bonn: UNFCCC Newsroom; 2015. [345] United Nations Framework Convention on Climate Change. Intended Nationally Determined Contributions (INDCs). 〈http://unfccc.int/focus/indc_ portal/items/8766.php〉; 2015 [accessed 10.08.15]. [346] Goldenberg S. Rockefeller brothers fund: it is our moral duty to divest from fossil fuels. The Guardian. New York: Guardian Media Group; 2015. [347] FossilFree. Divestment Commitments. 〈http://gofossilfree.org/commitments/ 〉; 2015 [accessed 02.08.15]. [348] Wikipedia. Fossil fuel divestment. 〈https://en.wikipedia.org/wiki/Fossil_fuel_ divestment〉; 2015 [accessed 05.08.15]. [349] Paun A, Knight Z, Chan W-S. Stranded assets: what next? HSBC Glob Res 2015. [350] Staff of Business Green. Nine out of 10 people want more renewable energy. The Guardian; 2012. [351] Case study: solar social enterprise in the UK sees positive growth. Renewable Energy Focus. Elsevier Ltd; 2012. [352] Field A. Startup social entrepreneurs help finance solar power social enterprises. Forbes: Forbes, Inc.; 2013. [353] McEachran R. African social enterprises pave the way for solar power while stimulating the local economy. The Guardian: Guardian Media Group; 2013. [354] Faunce TA, Lubitz W, Rutherford AWB, MacFarlane D, Moore GF, Yang P, et al. Energy and environment policy case for a global project on artificial photosynthesis. Energy Environ Sci 2013;6:695–8. [355] Sovacool, Benjamin K. How long will it take? Conceptualizing the temporal dynamics of energy transitions Energy Res. Soc. Sci. 2016. [356] Wolf J, Stephens E, Steinbusch S, Yarnold J, Ross IL, Steinweg C, Doebbe A, Krolovitsch C, Müller S, Jakob G, Kruse O, Posten C, Hankamer B. Multifactorial comparison of photobioreactor geometries in parallel microalgae cultivations. Algal Res. 2016;15:187–201. [357] Wagner L, Ross I, Foster J, Hankamer B. Trading off global fuel supply, CO2 emissions and sustainable development. PLoS One 2016;11(3):e0149406. http://dx.doi.org/10.1371/journal.pone.0149406. [358] Stocker TF. The closing door of climate targets. Science 2013;339:280–2. [359] Schramski John R, Gattie, David K, Brown James H. Human domination of the biosphere: rapid discharge of the earth-space battery foretells the future of humankind. Proc. Natl. Acad. Sci. 2015;112(31):9511–7.