Plant photosynthesis generates photosynthetic electrons (PEs) by splitting water into protons, oxygen, and electrons with the aid of solar energy. Solar photons energize the PEs and excite them to a higher energy level, and PEs are transferred through various photosynthetic apparatus in thylakoid membranes with a series of redox reactions (Figure 1
). There have been many attempts to directly or indirectly extract PEs from microbial and photosynthetic bacteria [1
]. Photosystem II (PSII) have been isolated from plant cells and experimented to assess the feasibility as a stand-alone molecular complex that can continuously split water molecules and generated PEs [3
]. Photosystem I (PSI)-based systems demonstrated stable and enhanced performance as bio-solar energy systems with various anode materials [6
]. However, the poor stability of the isolated PSII complexes or need for additional electron donors other than water for PSI-based systems still remain as challenges for their further development. More recently, direct extraction of PEs from living algal cells by nanoelectrode insertion was also demonstrated with high efficiency and long-term stability [9
]. However, the difficulty of scale-up of this approach also needs to be further investigated.
Thylakoids contain all the photosynthetic apparatus, such as PSII, a plastoquinone pool, cytochrome b6
f complex, plastocyanin, and PSI. They are relatively easy to be isolated from plant cells and remain stable after isolation compared to other isolated photosynthetic protein complexes. With these advantages, there have been many investigations that utilized thylakoids for bio-solar energy conversion. A key challenge of thylakoid-based bio-solar energy conversion is maximizing electrical connections between thylakoids and main anodes of a PE harvesting system. With this regard, many nanomaterials, including carbon nanotubes (CNTs) [11
], nanowires [13
], and two-dimensional materials, such as graphene [14
], were incorporated with isolated thylakoids, and demonstrated their enhanced performance compared to systems without such nanomaterials. However, since these approaches rely on secondary connections by the nanomaterials between thylakoids and an anode, a way to increase direct contact between thylakoids and an anode may further enhance the performance of PE harvesting.
In this work, we propose a novel design of an anode to accommodate thylakoids more efficiently and maximize direct contact between thylakoids and the anode surface. Micro-pillar (MP)-shaped electrodes were designed and fabricated based on the morphology of isolated thylakoids. It was hypothesized that three-dimensional MP electrodes could accommodate thylakoids between the micro-pillars such that their direct contact area could be increased compared to that of flat electrodes. Thylakoids isolated from spinach leaves were carefully analyzed to obtain their morphology and dimensions. Based on the thylakoid geometry, the optimum dimension of MP electrodes (diameters and spacing between MPs) was determined. MP electrodes were fabricated using metal-assisted chemical (MAC) etching. The controlled amount of thylakoids was drop cast on both flat and micro-pillar electrodes. Photosynthetic (PS) currents from the thylakoid-cast MP electrodes were analyzed under various conditions. Finally, a prototype of a PS bio-solar fuel cell was prepared and their performance was further studied.
2. Materials and Methods
4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES, H3375), sodium chloride (NaCl, S7653), magnesium chloride (MgCl2, M8266), D-sorbitol (S3889) were purchased from Sigma-Aldrich (St. Louis, MO, USA) for isolation of thylakoids. Hydrogen peroxide (H2O2, HDE0-18005) and hydrogen fluoride (HF, HDC0-00101) were purchased from Duksan (Cheonan, Korea), gold etchant (standard, 651818) was purchased from Sigma-Aldrich for metal-assisted chemical (MAC) etching. Nafion 212 and carbon paper were purchased from the Nanoholdings Ltd. (Seoul, Korea). Potassium ferricyanide (K3Fe(CN)6, 702587) and potassium nitrate (KNO3, P8394) as redox mediators, were purchased from Sigma-Aldrich. All aqueous solutions were prepared by using ultrapure de-ionized water (>18 MΩ·cm−1).
2.2. Thylakoids Isolation
Thylakoids were isolated from spinach by following the protocols as described previously [16
]. Spinach was purchased from a local market and washed with tap water first and then distilled water. After removing stems from the spinach, the leaves were cut into small pieces with scissors. They were ground in a phosphate buffer solution (pH 7, 50 mM sodium phosphate, 10 mM NaCl, 5 mM MgCl2
, 300 mM D-sorbitol) using a laboratory blender (Cole-Parmer, 8011EG, Waring) at a speed of 22,000 rpm for 20 s. The ground spinach leaves were filtered through a 20 μm nylon mesh and centrifuged at 3000× g
for 10 min. After the supernatants were removed, the pellets were re-suspended with a 5 mM MgCl2
solution and gently pipetted for 10 s to induce osmotic shock to rupture the chloroplast membranes. A wash medium (pH 7, 50 mM sodium phosphate, 10 mM NaCl, 5 mM MgCl2
, 300 mM D-sorbitol) was added to the solution and centrifuged at 4000× g
for 5 min. The isolated thylakoid membranes were suspended in the wash medium and stored in a refrigerator at 4 °C before use. All centrifugation processes were conducted at 4 °C.
2.3. Measurement of Chlorophyll Concentration
The concentration of the chlorophyll inside the isolated thylakoids was spectroscopically determined as described previously [17
]. Four microliters of isolated thylakoids were suspended in 2 mL of 80% (v
) acetone and the suspended solution was gently pipetted, and the absorbance of the solution was measured at 645 nm and 663 nm. Using the conversion equation by the previous work, the chlorophyll concentration in the isolated thylakoids was determined as 4 to 9 mg·chl mL−1
2.4. SEM Imaging of Thylakoids
Isolated thylakoids were diluted with distilled water to image the precise morphology of single or multiple vesicles. A diluted thylakoid solution was dropped on a silicon wafer piece whose surface was treated with oxygen plasma to make the surface hydrophilic for thin and uniform spreading of thylakoids on the substrate. The thylakoid-deposited substrate was dried in a convection oven at 36 °C for two days. A few nanometer thick Pt layer was coated before SEM imaging (JSM-7001F, JEOL Ltd., Akishima, Japan).
2.5. Fabrication of MP Electrode
MP electrodes were fabricated by the MAC etching process (Figure 2
). First, photoresist (PR) (TDUR-P902, TOKYO OHKA KOGYO CO., Ltd., Kawasaki, Japan) was coated on a silicon wafer using a spin coater (Figure 2
b). Photolithography was conducted to obtain a micro circle array pattern (Figure 2
c) using the KrF stepper (PAS5500, ASML, Veldhoven, The Netherlands). An E-beam evaporator (UEE, ULTECH CO., Ltd., Daegu, Korea) was used to deposit a Ti adhesion layer (2 nm thick) and a Au catalyst layer (20 nm thick) at a deposition rate of 0.1 nm/sec (Figure 2
d). After the metal deposition, a lift-off process was conducted to obtain a desired pattern of Au catalyst on the silicon wafer (Figure 2
e). MAC etching was conducted using a mixed solution comprised of 12.5% (v
) HF, and 1.5% (v/v
f). A gold etchant was used to remove the Au catalyst layer (Figure 2
g). Finally, a Au layer (50 nm thick) was deposited over the MP structures using a RF magnetron sputter (KOVA, Daegu, Korea) (Figure 2
h). The metalized MP electrodes were wire-bonded using silver paste and insulated with epoxy glue (Figure 2
2.6. Measurement of Photosynthetic Currents
Electrochemical measurement was conducted in a three-electrode setup comprised of MP electrode as a working electrode, Ag/AgCl as a reference electrode, and a Pt mesh as a counter electrode based on electrochemical analyzsis (CompactStat, Ivium, Eindhoven, The Netherlands). The entire measurement setup was placed in a custom-built Faraday cage with a built-in halogen lamp (LG-PS2, Olympus, Shinjuku, Japan) that was controlled from the outside. Unless mentioned otherwise, a bias potential of 0.4 V was applied between an MP anode and Pt mesh counter electrode. In a standard measurement, light intensity was 10 mW cm−2. The amount of thylakoid coating was maintained at a value of 2.96 mg·chl/mL. PCs were measured in a manner of chronoamperometry (CA) to monitor light-triggered PC values, while cyclic voltammetry (CV) was performed to identify which photosynthetic apparatus contribute to the light-triggered PC generation.
The dimensions of MP electrodes were determined based on the size of isolated thylakoids. The thylakoids isolated from spinach leave had the average size of 1 μm (Figure 3
). Small features are typically employed to increase the surface-to-volume ratio of an electrode, therefore, nanoscale features are often used in many applications. However, when the size of a target object (thylakoids in our study) is not small enough compared to the feature size of an electrode, more careful design of the electrode feature is necessary. Since we aim to maximize the direct contact area between thylakoids and an electrode, use of very thin nanowire-like structure may not be the best choice for our purpose. In fact, when the area ratio of an electrode surface over flat electrode was calculated for varying dimensions of the pillar or wire structures, the maximum area ratio was predicted for the pillar diameter of 1 μm (Figure 3
d inset). With micro-pillars 1 μm in diameter, the surface area of an MP array electrode is about 1.9 times larger than that of a flat electrode. A similar amount of enhancement in PC collection was observed in Figure 6
Fabrication of such MP array electrodes with 1 μm features is non-trivial. There have been various fabrication methods that can create nanostructures, such as nanowires smaller 100 nm or MPs larger than a few micrometers. However, creation of MPs with a 1 μm diameter with a precise spacing of 1 μm between each MP requires fabrication approaches different from well-established bottom-up vapor-liquid-solid (VLS) methods or top-down deep reactive ion etching (DRIE) techniques. MAC-etching was recently introduced to the field as a simple, but robust, fabrication method that can create highly-dense array of nanoscale to a few micrometer features with high A
]. As shown in Figure 2
, the photolithographically-patterned metal layer enabled precise placement of MPs with high A
. Unlike DRIE-created structures, the side walls of MPs by MAC-etching have smooth surfaces (Figure 4
a,b) and the fabricated MPs are highly uniform over an entire substrate (Figure 4
c). Although MPs with A/R of up to 1.5 were fabricated in this study, higher A
above 10 can be easily achieved by increasing etching time.
Use of electrochemical mediators facilitates electron transport between thylakoids and the anodes of photosynthetic fuel cells. Typically, this leads to extraction of much larger amount of photosynthetic electrons than direct electron extraction without mediators. However, the use of mediators is less desired due to its environmental toxicity and degrading performance over time. The degrading performance of mediators can cause the voltage loss of the systems and their instability under light and in high temperatures [27
]. As another approach, linker molecules or electrically-conducting nanomaterials were employed to immobilize or connect thylakoids and metal electrodes. Thylakoids were immobilized on glassy carbon electrodes using carboxyphenyl groups [28
], multi-wall carbon nanotubes (MWCNT) using a tethering molecule [11
], 1-pyrenebutanoic acid succinimidyl ester (PBSE) or thylakoids were electrically-wired with osmium redox polymers [29
]. Although these approaches demonstrated much enhanced performance of thylakoid-decorated anodes, they rely on electron transport through the secondary materials to main electrodes.
On the other hand, there have been a relatively small number of previous works that focused on extraction of PEs from the direct contact between thylakoids and metal electrodes. Photocurrents of 100 nA/cm2
were measured from thylakoid-deposited indium tin oxide (ITO) electrodes without a redox mediator [30
]. In another work [27
], thylakoid membranes were deposited on carbon paper and stabilized by forming a thin layer of silicon oxide. From this system, a light-triggered photocurrent of ca. 150 nA/cm2
was observed. In this work, as shown in Figure 6
a, thylakoid-deposited flat Au electrodes collected ca. 120 nA/cm2
, which is comparable to the photocurrents from the above-mentioned works. As predicted from our hypothesis, use of thylakoid-deposited MP electrodes resulted in enhanced photocurrents of ca. 280 nA/cm2
. This is strongly related to the increase in the ECSA of MP electrodes compared to the flat electrodes. Although MP electrodes with A/R of 1.5 have been tested in this work, the further increase of the A/R of MP electrodes may further enhance the magnitude of photosynthetic currents without sacrificing biocompatibility, stability, and voltages.
In the second cycle of light on/off for the measurement of photosynthetic currents, the photosynthetic currents from thylakoid-deposited MP electrodes started decreasing (Figure 6
b). This is likely due to the presence of hydrogen peroxide that is generated during photosynthesis. Hydrogen peroxide acts as an oxidant that can damage the photosynthetic apparatus in the photosynthetic electron transport (PET) chains. Although the chloroplast has a self-eliminating system for hydrogen peroxide, isolated thylakoids lack this self-eliminating function [31
] and continuous generation of hydrogen peroxide damages photosynthetic functions of the isolated thylakoids. To minimize such performance degradation by hydrogen peroxide, use of catalase, which can oxidize hydrogen peroxide into water and electrons, was proposed and stable, non-degrading performance of a thylakoid-deposited electrode was demonstrated [27
There is a limited number of previous works of PFCs based on direct electron transfer. Most previous PFCs used mediator molecules to enhance the performance (indirect electron transfer). In this study, PFCs based on the MP electrode without any mediator, generated a maximum current density of about 1 μA/cm2
with OCV of 407 mV, which is slightly lower than 1.5–3 μA/cm2
with OCV of 470–650 mV measured from other systems based on carbon paper electrodes [27
]. However, as discussed above, the anode current densities from the MP electrodes are similar to, or greater than, the values from other systems. Thus, the lower current density and OCV of the PFCs based on the MP electrodes are likely to be due to the less-optimized cathodes and cell configuration. In addition, the carbon paper electrodes of the other systems were decorated with quantum dots or stabilizing enzymes, such as catalase for enhanced performance, while the PFC in this study did not employ any secondary material. There has been no report of durability of PFCs yet, other than monitoring of PE currents for different illumination cycles as discussed above. Although a similar trend to the PE currents is expected, durability of full PFCs based on MP electrodes needs to be investigated in the near future.