Photoelectric conversion of bacterial reaction centres on IO-mesoITO electrodes.

(1)

Photoelectric conversion of bacterial reaction

centres on IO-mesoITO electrodes.

Including a review on biophotovoltaics using PSI & PSII

Sofie Klabbers

Report Bachelor Project Physics and Astronomy, size 15 EC, conducted between 3

– 04 – 2017 and 13 – 07 – 2017

Wis- en natuurkunde faculteit, Vrije Universiteit Amsterdam

Date of submission: 11 – 07 – 2017

Daily supervisor: Vincent Friebe

Assessor: Raoul Frese

Second assessor: Marloes Groot

(2)

Popular scientific summary

In een wereld waar duurzaamheid steeds belangrijker wordt is het van belang om de duurzame alternatieve energiebronnen die er nu zijn eens goed onder de loep te nemen. Zo zijn er bijvoorbeeld zonnecellen die zonne-energie omzetten in chemische energie, die wij dan weer kunnen gebruiken om onze telefoon op te laden. Maar wat vaak vergeten wordt is dat deze zonnecellen niet van duurzaam materiaal gefabriceerd worden. Zo zit er in de gemiddelde zonnecel bijvoorbeeld silicium. Door een grote toename in de fabricatie van zonnecellen is de silicium voorraad al flink geslonken (Takiguchi & Morita, 2009). Daarnaast is het niet over de hele wereld aanwezig wat ook politieke gevolgen heeft. De huidige zonnecellen worden niet op een duurzame manier gefabriceerd, wat er toe leid dat ze niet hergebruikt worden aan het eind van hun leven en allemaal op een grote stapel eindigen. De zon daarentegen is een zeer duurzame bron van energie, en als het mogelijk is een techniek te vinden om deze energie ook op een duurzame manier om te zetten, dan zou dat een uitkomst kunnen zijn voor het CO2 probleem.

vb. biofotovoltaïsche cel

Nu wordt er natuurlijk hard aan deze technieken gewerkt en één van de mogelijke alternatieven voor de huidige zonnecel zou de biofotovoltaïsche cel kunnen zijn. Dit zijn zonnecellen die gebruik maken van het fotosynthese systeem van planten en bacteriën. Fotosynthese is een zeer effectief systeem wat miljarden jaren door evolutie geperfectioneerd is. Bij fotosynthese kunnen elektronen in eiwitten in de chloroplasten van een plant door een zonnestraal worden geraakt. Als dat gebeurt nemen ze de energie uit de zonnestraal op en komen in een hoger energielevel, waarna ze van molecuul tot molecuul kunnen stromen. Biofotovoltaïsche cellen maken gebruik van dit systeem door de elektronenstroom op te vangen en door een batterij te sturen. Op die manier kan energie worden opgewekt.

In dit verslag is een overzicht gegeven van al het onderzoek wat in de afgelopen jaren in dit gebied is uitgevoerd. Hierdoor is het inzichtelijker geworden welke technieken een grote verbetering

bewerkstelligen en welke technieken minder goede resultaten geven. Met behulp van dit overzicht is een techniek geselecteerd die probeert het oppervlak van de zonnecel te vergroten door een soort gatenkaas-structuur te fabriceren. Deze techniek werd gecombineerd met het fotosynthetische systeem uit blauwalgen. In het huidige

onderzoek is een poging gedaan om ditzelfde systeem te combineren met een

fotosynthetisch systeem uit paarse bacteriën. De resultaten uit dit onderzoek waren

vergelijkbaar met het eerder gedane onderzoek met behulp van blauwalg. Dit is veelbelovend, want er zijn nog veel mogelijkheden om dit systeem te verbeteren.

(4)

(5)

Abstract

This paper presents an overview on the current state of research on biophotovoltaics. Comparing studies with the photosynthetic proteins PSI, PSII and RC-LH1 , the most promising techniques to increase the photocurrent are highlighted. This overview showed that the currently highest

performing technique available using PSI or PSII, and no photoactive substrates, is done by Mersch et al. 2015 using a IO-mesoITO substrate. This substrate is combined with a RC-LH1 complex using cytochrome as an electron donor and quinone as an electron acceptor. In order to increase the photocurrent, multilayers are tested and different polystyrene bead sizes are compared. This resulted in a photocurrent density peak of -927 ± 38 μA cm-2_{. Multilayers did not increase the photocurrent,} but comparable results are achieved comparing the photocurrents of one-layer systems to the result achieved by Mersch et al. (2015).

Introduction

The world is changing and the need for renewable energy is getting more and more pressing. The sun sends more energy to the earth every hour than we use in a year (Olmos & Kargul, 2015). We just need to find a way to collect that energy more effectively, and with sustainable technologies. Fortunately evolution has already found an incredibly efficient way to collect that energy:

photosynthesis. Photosynthesis converts solar energy into chemical energy which plants use to grow. If we use this system to convert the solar energy we have so abundantly on earth to electricity, we can stop pumping greenhouse gases into our atmosphere by burning fossil fuels. In this paper I will give an overview of all the research that has been done into solar cells made from photosynthetic materials, so-called biophotovoltaics, and try to discover the paths that may lead to a more efficient photovoltaic cell.

Motivation

With the application of solar cells rapidly growing it is important to be critical about their

sustainability. The energy source, in this case the sun, is perfectly sustainable because there is more than enough and it will not die out anytime soon. But the materials from which the commonly used solar cells are made are a lot less sustainable. Most commercialized solar cells, or photovoltaic cells, are made with silicon. The increase in photovoltaic cells has already led to a tightened supply of silicon (Takiguchi & Morita, 2009). There are several solutions for this problem; making solar cells that use less energy and are less costly, making a photovoltaic that is recyclable or reusable or making it more efficient to develop photovoltaic cells from other materials. This paper will focus on the last option, making photovoltaics from readily available materials in our environment, especially pigment proteins.

Theoretical introduction

Biophotovoltaics

Biophotovoltaics consist of a substrate, typically made from a metal or semiconductor, combined with a photoactive element derived from green plants or photosynthetic bacteria. These photoactive elements, typically derived from spinach or purple bacteria, convert solar energy into chemical energy. Photosynthetic complexes have been improved over billions of years of evolution, making for an efficient system. In these systems almost 100% of the excited electrons are used for energy conversion (Kothe et al., 2014). Another reason to construct photovoltaic cells from photosynthetic

(6)

complexes is that they are abundant in our environment and can quickly and easily be

(bio)synthesized, in contrast to silicon for example. This also means that manufacturers are not dependent on resources from other countries to produce the cells. In most of the research into photovoltaic cells only a half-cell is made instead of a complete cell with an anode and cathode. This way optimizing the half-cell of interest is easier and the results will not be limited by the counter electrode. To do measurements on a half-cell you need a counter electrode and a reference electrode in a three-electrode configuration in conjunction with a potentiostat to control and measure the voltage and current respectively.

Photosynthesis

There are two different photosynthetic systems that are important for making biophotovoltaic cells. First there is the system used by green plants, algae and some bacteria (e.g. cyanobacteria), it combines photosystem I and photosystem II to perform a function described in the Z-scheme. Secondly there is the bacterial reaction centre in green and purple bacteria (Bartlett, 2008). We will look at these systems separately.

Photosynthetic Z-Scheme

The Z-Scheme consist of two systems, photosystem I (PSI) and photosystem II (PSII) (Figure 1). They are located in the thylakoid membrane of the chloroplast. The thylakoid membrane tends to form into pancake-like structures that are called grana. These grana then stick together and form stacks (Bartlett, 2008). Both photosystem complexes are surrounded by a light-harvesting complex

containing chlorophyll a, chlorophyll b and carotenoid pigments. Because of these different pigments the complex can absorb a wider range of wavelengths so more energy can be captured. This energy is then transferred to the reaction centres (Bartlett, 2008).

Figure 1. The Z-scheme in the thylakoid membrane, on the left PSII and on the right PSI. Retrieved from https://sites.google.com/site/accessrevision/biology/cell-form-and-function/photosynthesis? tmpl=%2Fsystem%2Fapp%2Ftemplates%2Fprint%2F&showPrintDialog=1

The general process that takes place can be described as follows: PSII takes an electron from water and with the energy from the sun excites that electron into a higher energy state, than transports that electron to PSI where it gets excited again by a photon and converts NADP+ to NADPH. In order to use this system to make a solar cell instead of NADPH or ATP, some adjustments have to be made. PSI and PSII work differently and will be explained separately.

(7)

Photosystem II

Inside the light harvesting complexes a special pair of chlorophyll a molecules is found (P680). When

light shines on PSII the special pair gets excited (P680*/P680+). This is followed by a series of reactions

where the electron gets transferred to plastoquinone A (QA). Potential energy is sacrificed in this electron transfer steps to ensure efficient separation of charge. PSII can receive an electron from splitting water into oxygen and protons reducing the special pair to P680 making it neutral again. In

research other electron donors are also used, all listed in the appendix (A1). This electron has to tunnel to the reaction centre, which is an important step because it can slow down the process immensely. Often a mediator is used to accelerate the transfer, for example DCBQ or hydrogel, all examples are listed in the Appendix (A1) (Plumeré & Nowaczyk, 2016).

Theoretically photosystem II could have a turnover frequency (TOF) of 500 electrons per second. Unfortunately PSII has a short lifetime (in the range of minutes) so the turnover number (TON), the total amount of electrons that go through the reaction centre, is relatively small (in the range of 104₎ (Plumeré & Nowaczyk, 2016).

Photosystem I

Similarly to PSII, PSI has a special pair (P700) where an electron is excited to a higher energy level

leading to the radical pair (P700+/P700*) (Plumeré & Nowaczyk, 2016). The reaction centre of PSI

absorbs light at a wavelength of 700 nm, which is slightly red shifted from the reaction centre of PSI which absorbs light at 680 nm. Subsequently the electron is transferred by different reactions to Ferredoxin (Fd) (Bartlett, 2008)(Plumeré & Nowaczyk, 2016). In a Z-scheme with both PSI and PSII, PSI receives an electron from PSII, but when you make a half-cell the electron has to come from an electron donor. Several different kinds have been used in experiments (see A1). Once more the electron has to tunnel to the reaction centre where it reduces the special pair back to P700.

In a complete system with PSI and PSII the TOF of PSI is limited by the rate at which PSII provides electrons and the amount of PSII complexes available. Because PSI has a longer lifetime of about 40 hours, it has a much higher TON (in the range of 7 x 106_{). In a half-cell where the amount of electrons} available is not the rate limiting step, a theoretical TOF of about 106_e-_{/s is possible (Plumeré &} Nowaczyk, 2016). These characteristics make PSI more suitable for solar cell applications. Bacterial Photosynthesis

Bacterial photosynthesis (Figure 2) is similar to these processes but differs slightly. These systems do not produce oxygen like the photosynthetic Z-scheme does (Plumeré & Nowaczyk, 2016). In this case the reaction centre is not located in the thylakoid membrane of a chloroplast but in a lipid bilayer cytoplasmic membrane. The special pair, just like PSI and PSII, is surrounded by light harvesting complexes, but in this case there are two complexes. The light is harvested by the peripheral antenna complex (LH2) with transfers it to the core antenna (LH1) surrounding the bacterial reaction centre (Plumeré & Nowaczyk, 2016). When the special pair (P870) gets excited, a charge separation occurs

across the bacterial membrane where it takes electrons from cytochrome c, instead of water, and transfers the electron to the electron acceptor: quinone (Bartlett, 2008).

In the bacterial photosynthesis system the rate limiting step is the release of electrons to quinone. The resulting TOF is around 5000 è/s, 10 times the TOF of PSII (Plumeré & Nowaczyk, 2016).

(8)

Figure 2. The bacterial reaction centre in a lipid bilayer cytoplasmic membrane. Retrieved from https://www.studyblue.com/notes/note/n/bio-2051-study-guide-2013-14-brininstool/deck/8695959

Review research

In this paper an overview is given of the research that has been done on biophotovoltaics with PSI and PSII. A review of the research that has been done on biophotovoltaics with RC-LH1 was done by Friebe et al. 2017. The goal was to find an approach that works well with PSI or PSII, meaning it results in high photocurrent densities, and subsequently combine this approach with RC-LH1. This led to a table where the studies were compared on photocurrent density and other specifics (A1). In order to compare these numbers some had to be converted to the right unit, and some information was read from a graph. An overview of this table is given in Figure 3. This table provided insight in the different strategies that are used and what the effects on the photocurrent density are. Several articles use hydrogel to improve stability, there has been experimented with different kinds of hydrogel but it looks like an Os-complex hydrogel works the best. Another technique is to use Self Assembled Monolayers (SAM’s), of which there are also several variations that are used. The SAM’s form a connection between the substrate and the photosynthetic complexes and make these connections more stable. Also the electron donors and acceptors are diverse. For example DCBQ (2,6-dichloro-1,4-benzoquinone), MV (Methyl Viologen) and DCPIP (2,6-dichloorindofenol) are commonly used acceptors and water, Os complex, DCIPH2, NaAsc (Sodium ascorbate) are common donors.

Another way to increase the photocurrent is to increase the surface area of the substrate. With PSII this has been done with IO-mesoITO (inverse opal mesoporous indium tin oxide) substrates, by making a structure with holes so more photosynthetic complexes can bind to the electrode. This increases the protein loading. With RC-LH1 this has been done on rough silver, making the surface uneven.

Other substrates that are commonly used in research are Au (gold), GC (glassy carbon), graphene, Al (aluminum) and ITO (indium tin oxide), but there is also a lot of research done on substrates made of Si (silicon) and TiO2 (titanium dioxide). These elements are photoactive which means that they can induce a photocurrent without the use of photosynthetic complexes like PSI and PSII. The

photovoltaic cells that are currently used are mostly made of silicon so it is widely known that this substrate works very well, but they are not biophotovoltaics. The reason for research on

biophotovoltaics is that these elements (Si and Ti) are rare and not sustainable, so the goal is to make a solar cell without these elements. In the articles in this overview Si and TiO2 are combined with PSI or PSII, increasing the photocurrents.

(9)

Figure 3. Recent progress in biophotovoltaics with PSI or PSII as photoactive elements . The

corresponding papers are listed in the table. Only papers with a photocurrent of above 3 μA cm-2_are included. a) The whole table, b) excluding papers that used photoactive metals, Si or TiO2.

(10)

At the moment the highest photocurrent densities are achieved by Shah, (Shah et al., 2015), with a substrate of TiO2 nanostructured films deposited onto tin-doped ITO-coated aluminosilicate glass. In the article they make a distinction between this substrate without PSI and with PSI which shows a doubling of the photocurrent with the use of PSI (from 1.38 to 2.51 mA cm-2_{). From this they} conclude that half the photocurrent is due to PSI. But in this article they also did a measurement of the substrate with only ITO and PSI, no TiO2, this resulted in a photocurrent of 0.05 mA cm-2, which is almost nothing in comparison. So in conclusion this is not a very promising technique to make

biophotovoltaics.

From all the research that has been done on PSI and PSII the study with the highest photocurrent density, without a photoactive substrate, was selected.

Conclusion review

From the available literature we can conclude that electrodes made from inverse opal mesoporous Indium tin oxide (IO-mesoITO) increase the photocurrent significantly with PS II (A1) . This is because the structure of IO-mesoITO increases the surface of the electrode, so more photosystem complexes can bind to the electrode (Figure 4). A mixture of polystyrene beads and ITO nanoparticles is

deposited on a FTO-coated glass slide. FTO was selected as surface material because of its transparent and conductive properties. The ITO nanoparticles are also conductive, so after the polystyrene is burned away the structure that is left is completely conductive. The structure also enables light to scatter through due to the disorganized structure of the holes. This increases the chance of exciting an electron in the photosynthetic proteins. (Mersch et al., 2015)

Figure 4. First a mixture of Polystyrene and ITO nanoparticles is deposited on a FTO-coated glass slide (a). After it dries (b), it is sintered, the polystyrene beads melt away and become holes, enlarging the surface area (c). The melting point of ITO lies much higher than polystyrene so it will not melt in this step. If a photosynthetic reaction center is deposited on the IO-mesoITO surface it can travel through the structure and bind inside the holes (Mersch et al., 2015).

IO-mesoITO electrodes

In this experiment the same technique was used as in the paper of Mersch et al. 2015 to make the IO-mesoITO electrodes. These electrodes were then combined with the technique to make biophotovoltaic cells with RC-LH1 (Friebe et al., 2016), using quinone as an electron acceptor and Cytochrome c as an electron donor.

Several experiments were done to improve the IO-mesoITO electrode. Three different polystyrene beads sizes (0.75 μm, 0.8 μm, 1.1 μm) where mixed with ITO nanoparticles (<50 nm). Two different techniques were used to increase the protein loading, and experiments with quinone solutions of pH 7.0 and pH 8.0 were compared.

(11)

Experimental section

Preparation of IO-mesoITO electrodes

For the preparation of Inverse Opal mesoporous Indium Tin Oxide electrodes the protocol published by Mersch was adapted (Mersch et al., 2015). A mixture of ITO nanoparticles and polystyrene beads was deposited on FTO-coated glass slides. In order to get the right ratio of pure ITO and polystyrene in a MeOH/water solution the ITO particles were diluted and the polystyrene was purified. The mixture was made as follows: 35 mg of ITO nanoparticles < 50 nm diameter were mixed with 300 μL of a MeOH/water mixture of (6:1 v/v) and then sonicated for 3 hours. 0.26 mL of the polystyrene beads (0.8 nm and 1.1 nm, 10% w/v suspension in water) was centrifuged at 12 000 RPM for 5 min and after that the supernatant was poured out in order to remove the water. Subsequently the residue was redispersed in 1 mL pure MeOH, vortexed and centrifuged again at the same speed and the supernatant poured out, to ensure all the water was eliminated. The MeOH evaporated quickly resulting in a residue of only polystyrene beads. The ITO mixture was added to the polystyrene and vortexed thoroughly until all the residue was redispersed. This mixture was sonicated for at least 5 minutes in ice cold water so it could be mixed thoroughly without the polystyrene melting.

The FTO-coated glass (8 Ω cm-2 _{) was cut in 0.5 x 2 cm rectangles. To clean them they were sonicated} in acetone (in the protocol isopropyl alcohol was used but this was not present and acetone has the same function) and afterwards with ethanol for at least 10 min each and stored at 150 °C so all the leftover fluids evaporated. To predefine the surface area of the IO-mesoITO layer a squared area of 0.25 cm2_{was defined by a small piece of scotch tape. In the protocol a parafilm ring was used to} define the surface area but scotch tape had the same effect. Then 4 μL of the polystyrene-ITO mixture was pipetted onto the surface area of the FTO-coated glass. To make multilayers 4 μL were pipetted on top with a drying period of 3 hours in between each layer. In the protocol a drying period of at least 4 hours was adhered to make sure all the solvent would be evaporated. This period was shortened a bit because of a lack of time. For some electrodes 8 or 12 μL of the polystyrene-ITO mixture was pipetted onto the FTO with a drying period of 8 hours.

The IO-mesoITO electrodes were then heated with 1 °C per minute from room temperature to 500 °C and kept at this temperature for 20 minutes. During this sintering the polystyrene beads were burned away but the ITO, having a much higher melting point, did not, leaving a structure with holes. To clean the electrodes they were put in a mixture of 30%H2O2/H2O/30%NH4OH (1:5:1 v/v) at 70 °C for 15 minutes, rinsed with water, and heated for 1 hour at 180 °C to evaporates all residues.

Preparation of RC-LH1 on electrodes

The LH1 preparation was done according to the protocol of a previously published paper of RC-LH1 on nanoporous silver ((Friebe et al., 2016). The RC-RC-LH1 (66.9 μM) was diluted with Tris + b DSN 0.04% from 200 μL to 800 μL. This was divided over small cuvettes each containing 150 μL. The electrodes were stored in the cuvettes to incubate for at least 30 minutes at 5 °C. On rough silver the optimum of RC-LH1 binding to the surface is around 1 hour but after 15 minutes 95% of the optimum is achieved (unpublished data, Vincent M. Friebe 2016). After the incubation they were dipped in a reservoir of Tris (20 mM, pH 8) and then stored in the same solution for at least 5 min. At last they were stored in cytochrome c (100 μM) for 5 minutes.

(12)

Measurements

The electrodes were measured in a photoelectrochemical cell fitted with a Ag/AgCl reference electrode and a platinum counter electrode in a solution of 5mM Quinone in 20mM Tris (either pH 8.0 or pH 7.0). A PGSTAT128N potentiostat (Metrohm Autolab) was used to control the three electrode cell, a bias potential of −50 mV versus Ag/AgCl being applied for all experiments

(approximating to the dark open circuit potential). The entire working electrode was homogeneously illuminated with a 870 nm LED with an intensity of 46 mW cm−2_.

Results/discussion

In the following results only a quinone solution of pH 8.0 was used. Experiments were done with a solution of pH 7.0 but the resulting photocurrents were significantly lower than in a pH 8.0 solution (A5). After about five hours of measuring the quinone solution was refreshed with the same solution but not used. This had a significant reducing effect on the photocurrent which could be explained by salt leaking from the reference electrode into the quinone solution. This would imply that the presence of salt increases the photocurrent. In further research this could be examined by testing the electrodes in various salt densities. The remaining electrodes were measured in the old quinone solution.

Difference in polystyrene bead sizes

Three different polystyrene beads sizes were measured, mean particle size; 0.75 μm, 0.8 μm and 1.1 μm. A difference in structure was expected because bigger polystyrene beads give bigger holes in the IO-mesoITO. Bigger holes can result in more accessibility for the RC-LH1 to enter the structure, but it could also reduce the surface area. It was found that the electrodes with a particle size of 0.8 μm provided the highest current density peak compared both to 0.75 μm beads (0.8 μm induced a photocurrent of -927 ± 38 μA cm-2_{and 0.75 μm induced a photocurrent of -819 ± 63 μA cm}-2_{) and to} 1.1 μm beads (0.8 μm induced a photocurrent of -866 ± 271 μA cm-2_{and 1.1 μm induced a}

photocurrent of -608 ± 196 μA cm-2_{) see (Figure 5). Also the steady state of the IO-mesoITO structure} with 0.8 μm polystyrene beads was shown to be largest (-191 ± 18 μA cm-2

_{), with the 0.75}

_{μm beads} (-156 ± 14 μA cm-2

₎

_{, and the 1.1 μm beads (-142 ± 40 μA cm}-2_{). Because the difference between 0.8} µm and 0.75 µm is smaller than the difference between 0.8 µm and 1.1 µm, a smaller difference in photocurrent was to be expected.

Figure 5. Different bead sizes with one layer, corresponding to 4 μL of ITO/polystyrene mixture. a) 0.8 and 0.75 μm polystyrene beads, b) 0.8 and 1.1 μm polystyrene beads. Photocurrents were performed under -50 mV vs NHE, at 46 mW cm-2_{illumination (870 nm LED) and in a solution containing 5 mM} UQ0.

(13)

The different polystyrene bead sizes caused very different IO-mesoITO surfaces. There is a difference visible with the naked eye (A2) but also SEM (Scanning Electron Microscope) pictures show

differences (Figure 6)(A4).

Figure 6. SEM pictures of different polystyrene bead sizes. Magnification 3000x.

The SEM pictures show that the surface of the IO-mesoITO structure (with polystyrene bead sizes 0.8 μm and 1.1 μm) has a very little amount of holes compared to the inner structure of the IO-mesoITO. This would make it hard for the RC-LH1 complexes to enter the inner structure. In the IO-mesoITO with 3.0 μm polystyrene beads the surface of the IO-mesoITO has a lot of half open shells, forming a structure which is easy to penetrate and has a lot of surface area for the RC-LH1 to connect to. This would make it a promising structure to induce high photocurrents. Unfortunately first measurements on these electrodes did not show very high photocurrents (-269 ± 133 μA cm-2 _{see A5). But more} research has to be done to confirm this data. Earlier research on IO-mesoITO (Mersch et al., 2015) showed that the top surface can have a more open structure than the electrodes with 0.8 μm and 1.1 μm polystyrene beads in this research, looking similar to the electrodes with 3.0 μm beads. In the current density figures a back current is visible after the light is turned off. This is likely due to the recombination of the QH2 with the electrode. The steady state of the photocurrent is relatively low compared to the peak. This could imply that the limiting factor is the release of electrons to the quinone, in the first instance the quinone is all around the photosynthetic complexes but once they have formed QH2 a new quinone has to diffuse to the RC-LH1 complex and this takes time, especially in a porous film. A way to increase the release of electrons would be to add a rotator in the quinone solution making the passing of quinone quicker, or a counter electrode which converts the QH2 back into Qo2, thereby replenishing the analyte. The other possibility is that there is charge recombination occurring, as apparent in the large anodic (positive) spike, which would indicate QH2 is back reacting with the working electrode. This could be reduced by adding some sort of blocking layer to prevent charge recombination of the product with the electrode.

Multilayers

In previous experiments multilayers of the IO-mesoITO have shown to increase the photocurrent density in combination with PSII (Mersch et al., 2015). Multilayers increase the thickness of the IO-mesoITO layer, increasing the surface area for the RC-LH1 to bind to so the protein loading increases. More layers do not infinitely increase the photocurrent due to an increasing resistance, mass

transport limitations and the fact that in a thick layer it is more difficult for light to scatter all the way to the first layer (Mersch et al., 2015). In the previously mentioned paper an optimum of

(14)

photocurrent density was found with four layers corresponding to a 40 μm thick IO-mesoITO structure.

Figure 7. Different bead sizes and techniques to make multilayers and increase protein loading. a) droplet method with polystyrene beads (0.8 μm), b) droplet method with polystyrene beads (1.1 μm), c) layer method with polystyrene beads (0.8 μm), d) layer method with polystyrene beads (1.1 μm). Photocurrents were performed under -50 mV vs NHE, at 46 mW cm-2_{illumination (870 nm LED) and in} a solution containing 5 mM UQ0.

In the experiment with multilayers combined with RCL-H1 it is found that a thicker IO-mesoITO structure does not increase the photocurrent density, neither with the 0.8 μm polystyrene beads nor with the 1.1 μm polystyrene beads (Figure 7). Two techniques to make a thicker IO-mesoITO

structure were used, in the layer method several layers of 4 μL were deposited on the FTO with a three hour drying time in between every layer. In the droplet method a thicker structure was obtained by directly depositing a bigger amount of the ITO/polystyrene mixture, to make 2 layers 8 μL and to make 3 layers 12 μL was deposited.

When making thicker IO-mesoITO structures, the layers became very fragile and did sometimes crack. The thicker the layer, the more fragile it became. There was a difference between the structures with different polystyrene bead sizes, the electrodes with the 1.1 μm bead sizes showed less cracking. Nevertheless there was no increase in photocurrent upon adding more layers. There was a decrease in photocurrent visible from the electrodes with very bad damage, but little damage did not seem to have a significant influence on the photocurrent (for specification of damage see A2). It should be possible to make multilayers without cracking because it has been done before (Mersch et al., 2015). This should be investigated in further research.

(15)

Improvements

To improve the IO-mesoITO structure other polystyrene bead sizes and nanoparticle sizes could be experimented with. It is clear from the SEM pictures that the relative size of the ITO and the

polystyrene had an influence on the structure. Also the relative amount could make a big difference. If there would be too much polystyrene compared to ITO there could be not enough ITO to make the shells and the structure would fall apart. If there would be too little polystyrene compared to the ITO, there would not be enough holes and the photosynthetic complexes would not be able to penetrate through the structure. Somewhere in the middle there is an optimum that has to be found.

A control measurement showed that electrodes either without IO-mesoITO or without RC-LH1 induced no significant photocurrent. This means that the photocurrent is induced by RC-LH1 and that the IO-mesoITO increased the photocurrent significantly (Figure 8).

Figure 8. Control with one layer and polystyrene beads (0.8 μm) Photocurrents were performed under -50 mV vs NHE, at 46 mW cm-2_{illumination (870 nm LED) and in a solution containing 5 mM UQ0.}

A pigment extraction was performed for a few electrodes according to Friebe et al. 2016. This showed a protein loading of (ΓRC-LH1= 231 pmol cm-2) for electrodes with one layer and (ΓRC-LH1= 392 pmol cm-2_{) for electrodes with two layers (Figure 9a). This is consistent with the pigment loading} shown by Mersch et al. 2015 (1020 pmol cm-2_{for four layer). But the increase in pigment loading did,} in this experiment, not increase the photocurrent, which it did do in the research by Mersch et al. 2015.

Figure 9. a) The absorbance spectrum at different wavelengths, b) a durability plot presenting only the steady state photocurrent, showing a half-life of approximately 6 hours.

(16)

Combining the RC-LH1 complex solution with hydrogel could improve the photocurrent. This has been done before in an IO-mesoITO structure (Sokol et al., 2016). Hydrogel improves the stability of the photosynthetic complexes and makes it easier for electrons to be transported from the reaction centre to the electrode. But it could also congest the IO-mesoITO structure and make it harder for the photosynthetic complexes to enter the structure.

Conclusion

With the interest in sustainable development increasing the field of biophotovoltaics is also growing. Over the last fifteen years al lot of techniques have been tried, not all research has led to high photocurrents but the main trend has been a significant increase in photocurrent densities. Going from 7 E-6_{mA cm}-2 _{to 1.4 mA cm}-2 _{in just a few years. There are still a lot of techniques that have to} be researched and combinations of these techniques that can be made, but there are also still a lot of obstacles that have to be cleared. For example the half-life of biophotovoltaics is still very short and the steady state of the photocurrent is still always lower than the peak. These are things that have to be solved before they can compete with the current solar cells and make the solar cell industry more sustainable.

For the IO-mesoITO electrodes examined in this paper specifically, there is a lot of research that has to be done to use the complete potential of this technique. But a significant increase in photocurrent density peaks has been reached for the biophotovoltaics using RC-LH1 (unpublished data, Vincent M. Friebe 2017). Even without multilayers and a quarter of the protein loading a comparable result was achieved as in the research with PSII on mesoITO (Mersch et al., 2015). This makes the IO-mesoITO a promising substrate, especially combined with RC-LH1.

(17)

References

Badura, A., Esper, B., Ataka, K., Grunwald, C., Wöll, C., Kuhlmann, J., … Rögner, M. (2006). Light-driven water splitting for (bio-)hydrogen production: photosystem 2 as the central part of a

bioelectrochemical device. Photochemistry and Photobiology, 82(5), 1385–1390. https://doi.org/10.1562/2006-07-14-RC-969

Badura, A., Guschin, D., Esper, B., Kothe, T., Neugebauer, S., Schuhmann, W., & Rögner, M. (2008). Photo-induced electron transfer between photosystem 2 via cross-linked redox hydrogels.

Electroanalysis, 20(10), 1043–1047. https://doi.org/10.1002/elan.200804191

Badura, A., Guschin, D., Kothe, T., Kopczak, M. J., Schuhmann, W., & Rögner, M. (2011). Photocurrent generation by photosystem 1 integrated in crosslinked redox hydrogels. Energy &

Environmental Science, 4(7), 2435. https://doi.org/10.1039/c1ee01126j

Bartlett, P. (Ed.). (2008). Bioelectrochemistry Fundamentals, Experimental Techniques and

Applications. John Wiley & Sons, Ltd.

Beam, J. C., LeBlanc, G., Gizzie, E. A., Ivanov, B. L., Needell, D. R., Shearer, M. J., … Cliffel, D. E. (2015). Construction of a Semiconductor-Biological Interface for Solar Energy Conversion: P-Doped Silicon/Photosystem I/Zinc Oxide. Langmuir, 31(36), 10002–10007.

https://doi.org/10.1021/acs.langmuir.5b02334

Calkins, J. O., Umasankar, Y., O’Neill, H., & Ramasamy, R. P. (2013). High photo-electrochemical activity of thylakoid–carbon nanotube composites for photosynthetic energy conversion.

Energy & Environmental Science, 6(6), 1891. https://doi.org/10.1039/c3ee40634b

Ciesielski, P. N., Faulkner, C. J., Irwin, M. T., Gregory, J. M., Tolk, N. H., Cliffel, D. E., & Jennings, G. K. (2010). Enhanced photocurrent production by Photosystem i multilayer assemblies. Advanced

Functional Materials, 20(23), 4048–4054. https://doi.org/10.1002/adfm.201001193

Ciesielski, P. N., Hijazi, F. M., Scott, A. M., Faulkner, C. J., Beard, L., Emmett, K., … Kane Jennings, G. (2010). Photosystem I - Based biohybrid photoelectrochemical cells. Bioresource Technology,

101(9), 3047–3053. https://doi.org/10.1016/j.biortech.2009.12.045

Darby, E., Leblanc, G., Gizzie, E. A., Winter, K. M., Jennings, G. K., & Cliffel, D. E. (2014). Photoactive films of photosystem I on transparent reduced graphene oxide electrodes. Langmuir, 30(29), 8990–8994. https://doi.org/10.1021/la5010616

Das, R., Kiley, P. J., Segal, M., Norville, J., Yu, A. A., Wang, L., … Baldo, M. (2004). Integration of photosynthetic protein molecular complexes in solid-state electronic devices. Nano Letters,

4(6), 1079–1083. https://doi.org/10.1021/nl049579f

Feifel, S. C., Lokstein, H., Hejazi, M., Zouni, A., & Lisdat, F. (2015). Unidirectional Photocurrent of Photosystem i on π-System-Modified Graphene Electrodes: Nanobionic Approaches for the Construction of Photobiohybrid Systems. Langmuir, 31(38), 10590–10598.

https://doi.org/10.1021/acs.langmuir.5b01625

Feifel, S. C., Stieger, K. R., Lokstein, H., Lux, H., & Lisdat, F. (2015). High photocurrent generation by photosystem I on artificial interfaces composed of π-system-modified graphene. J. Mater.

Chem. A, 3(23), 12188–12196. https://doi.org/10.1039/C5TA00656B

Friebe, V. M., Delgado, J. D., Swainsbury, D. J. K., Gruber, J. M., Chanaewa, A., Van Grondelle, R., … Frese, R. N. (2016). Plasmon-Enhanced Photocurrent of Photosynthetic Pigment Proteins on Nanoporous Silver. Advanced Functional Materials, 26(2), 285–292.

https://doi.org/10.1002/adfm.201504020

Friebe,V.M., PhD Thesis entitled: ‘’Hacking Photosynthesis,” 2017.

Gizzie, E. A., Leblanc, G., Jennings, G. K., & Cliffel, D. E. (2015). Electrochemical preparation of photosystem I-polyaniline composite films for biohybrid solar energy conversion. ACS Applied

Materials and Interfaces, 7(18), 9328–9335. https://doi.org/10.1021/acsami.5b01065

(18)

Cliffel, D. E. (2015). Photosystem I-polyaniline/TiO 2 solid-state solar cells: simple devices for biohybrid solar energy conversion. Energy Environ. Sci., 8(12), 3572–3576.

https://doi.org/10.1039/C5EE03008K

Gunther, D., LeBlanc, G., Cliffel, D. E., & Jennings, G. K. (2013). Pueraria lobata (Kudzu) Photosystem I Improves the Photoelectrochemical Performance of Silicon. Industrial Biotechnology, 9(1), 37– 41. https://doi.org/10.1089/ind.2012.0036

Hartmann, V., Kothe, T., Pöller, S., El-Mohsnawy, E., Nowaczyk, M. M., Plumeré, N., … Rögner, M. (2014). Redox hydrogels with adjusted redox potential for improved efficiency in Z-scheme inspired biophotovoltaic cells. Phys. Chem. Chem. Phys., 16(24), 11936–11941.

https://doi.org/10.1039/C4CP00380B

Kato, M., Cardona, T., Rutherford, A. W., & Reisner, E. (2012). Photoelectrochemical water oxidation with photosystem II integrated in a mesoporous indium-tin oxide electrode. Journal of the

American Chemical Society, 134(20), 8332–8335. https://doi.org/10.1021/ja301488d

Kothe, T., Pöller, S., Zhao, F., Fortgang, P., Rögner, M., Schuhmann, W., & Plumeré, N. (2014). Engineered electron-transfer chain in photosystem 1 based photocathodes outperforms electron-transfer rates in natural photosynthesis. Chemistry - A European Journal, 20(35), 11029–11034. https://doi.org/10.1002/chem.201402585

Leblanc, G., Chen, G., Gizzie, E. A., Jennings, G. K., & Cliffel, D. E. (2012). Enhanced photocurrents of photosystem i films on p-doped silicon. Advanced Materials, 24(44), 5959–5962.

https://doi.org/10.1002/adma.201202794

Leblanc, G., Winter, K. M., Crosby, W. B., Jennings, G. K., & Cliffel, D. E. (2014). Integration of photosystem i with graphene oxide for photocurrent enhancement. Advanced Energy

Materials, 4(9), 1301953. https://doi.org/10.1002/aenm.201301953

Mersch, D., Lee, C. Y., Zhang, J. Z., Brinkert, K., Fontecilla-Camps, J. C., Rutherford, A. W., & Reisner, E. (2015). Wiring of Photosystem II to Hydrogenase for Photoelectrochemical Water Splitting.

Journal of the American Chemical Society, 137(26), 8541–8549.

https://doi.org/10.1021/jacs.5b03737

Mershin, A., Matsumoto, K., Kaiser, L., Yu, D., Vaughn, M., Nazeeruddin, M. K., … Zhang, S. (2012). Self-assembled photosystem-I biophotovoltaics on nanostructured TiO2 and ZnO. Scientific

Reports, 2, 1–7. https://doi.org/10.1038/srep00234

Ocakoglu, K., Krupnik, T., Van Den Bosch, B., Harputlu, E., Gullo, M. P., Olmos, J. D. J., … Kargul, J. (2014). Photosystem I-based biophotovoltaics on nanostructured hematite. Advanced

Functional Materials, 24(47), 7467–7477. https://doi.org/10.1002/adfm.201401399

Olmos, J. D. J., & Kargul, J. (2015). A quest for the artificial leaf. International Journal of Biochemistry

and Cell Biology, 66, 37–44. https://doi.org/10.1016/j.biocel.2015.07.005

Peters, K., Lokupitiya, H. N., Sarauli, D., Labs, M., Pribil, M., Rathouský, J., … Fattakhova-Rohlfing, D. (2016). Nanostructured Antimony-Doped Tin Oxide Layers with Tunable Pore Architectures as Versatile Transparent Current Collectors for Biophotovoltaics. Advanced Functional Materials,

26(37), 6682–6692. https://doi.org/10.1002/adfm.201602148

Plumeré, N., & Nowaczyk, M. M. (2016). Biophotoelectrochemistry of Photosynthetic Proteins. In

Advances in biochemical engineering/biotechnology (Vol. 123, pp. 127–141). Springer Berlin

Heidelberg. https://doi.org/10.1007/10_2016_7

Saboe, P. O., Lubner, C. E., McCool, N. S., Vargas-Barbosa, N. M., Yan, H., Chan, S., … Kumar, M. (2014). Two-dimensional protein crystals for solar energy conversion. Advanced Materials,

26(41), 7064–7069. https://doi.org/10.1002/adma.201402375

Shah, V. B., Henson, W. R., Chadha, T. S., Lakin, G., Liu, H., Blankenship, R. E., & Biswas, P. (2015). Linker-free deposition and adhesion of photosystem i onto nanostructured TiO2 for biohybrid photoelectrochemical cells. Langmuir, 31(5), 1675–1682. https://doi.org/10.1021/la503776b Sokol, K. P., Mersch, D., Hartmann, V., Zhang, J. Z., Nowaczyk, M. M., Rögner, M., … Reisner, E. (2016).

Rational wiring of photosystem II to hierarchical indium tin oxide electrodes using redox polymers. Energy Environ. Sci., 9(12), 3698–3709. https://doi.org/10.1039/C6EE01363E

(19)

Stieger, K. R., Feifel, S. C., Lokstein, H., Hejazi, M., Zouni, A., & Lisdat, F. (2016). Biohybrid

architectures for efficient light-to-current conversion based on photosystem I within scalable 3D mesoporous electrodes. J. Mater. Chem. A, 4(43), 17009–17017.

https://doi.org/10.1039/C6TA07141D

Takiguchi, H., & Morita, K. (2009). Sustainability of silicon feedstock for a low-carbon society.

Sustainablilty Science, 4(1), 117–131. https://doi.org/10.1007/s11625-009-0069-1

Yang, S., Robinson, M. T., Mwambutsa, F., Cliffel, D. E., & Jennings, G. K. (2016). Effect of Cross-linking on the Performance and Stability of Photocatalytic Photosystem I Films. Electrochimica Acta,

222, 926–932. https://doi.org/10.1016/j.electacta.2016.11.059

Zhao, F., Conzuelo, F., Hartmann, V., Li, H., Nowaczyk, M. M., Plumeré, N., … Schuhmann, W. (2015). Light Induced H2 Evolution from a Biophotocathode Based on Photosystem 1 - Pt Nanoparticles Complexes Integrated in Solvated Redox Polymers Films. Journal of Physical Chemistry B,

119(43), 13726–13731. https://doi.org/10.1021/acs.jpcb.5b03511

Zhao, F., Sliozberg, K., Rogner, M., Plumere, N., & Schuhmann, W. (2014). The Role of Hydrophobicity of Os-Complex-Modified Polymers for Photosystem 1 Based Photocathodes. Journal of the

(20)

Appendix

(21)

Figure A1. Table with all research with reported photocurrents above 3 µA cm-2_{. References: [1] (Badura et al.,} 2006); [2] (Badura et al., 2008); [3] (Badura et al., 2011); [4] (Hartmann et al., 2014); [5] (Kothe et al., 2014); [6] (Zhao, Sliozberg, Rogner, Plumere, & Schuhmann, 2014); [7] (Zhao et al., 2015); [8] (Kato, Cardona, Rutherford, & Reisner, 2012); [9] (Mersch et al., 2015); [10] (Sokol et al., 2016); [11] (S. C. Feifel, Stieger, Lokstein, Lux, & Lisdat, 2015); [12] (Sven C. Feifel, Lokstein, Hejazi, Zouni, & Lisdat, 2015); [13] (Stieger et al., 2016); [14] (Ciesielski, Hijazi, et al., 2010); [15] (Ciesielski, Faulkner, et al., 2010); [16] (Leblanc, Chen, Gizzie, Jennings, & Cliffel, 2012); [17] (Gunther, LeBlanc, Cliffel, & Jennings, 2013); [18] (Leblanc, Winter, Crosby, Jennings, & Cliffel, 2014); [19] (Darby et al., 2014); [20] (E. A. Gizzie, Leblanc, Jennings, & Cliffel, 2015); [21] (Beam et al., 2015); [22] (E. a. Gizzie et al., 2015); [23] (Yang, Robinson, Mwambutsa, Cliffel, & Jennings, 2016); [24] (Das et al., 2004); [25] (Mershin et al., 2012); [26] (Calkins, Umasankar, O’Neill, & Ramasamy, 2013); [27] (Saboe et al., 2014); [28] (Ocakoglu et al., 2014); [29] (Shah et al., 2015); [30] (Peters et al., 2016).

(22)

A2 – Photographic images of IO-mesoITO electrodes

Figure A2. Images of the IO-mesoITO electrodes. The difference in multilayers is shown, the degrees of cracking that occurred after sintering and the difference between 0.8 µm polystyrene beads and 1.1 µm polystyrene beads.

(23)

Figure A3. a) Defining the surface area, b) cleaning the electrodes, c) Experimental setup, three electrodes in a Quinone solution, lamp shining from the side.

(24)

A4 – SEM pictures

(25)

A5 – Other experiments

(26)

Figure A6. Experiment B: with quinone solution 7.0, beads 0.8 and 0.75.

Figure A7. Experiment A: different cooking times. (Faster to 500 °C), 0.8 beads, pH8 quinone, incorrect amount of ITO and polystyrene

(27)

Photoelectric conversion of bacterial reaction centres on IO-mesoITO electrodes.