Development of Nanoporous Gold Based Bioelectrodes

Development of nanoporous gold based bioelectrodes

Thesis presented for the award of Doctor of Philosophy (Ph.D.)

Xinxin Xiao

Under the supervision of Professor Edmond Magner

Submitted to the Faculty of Science and Engineering University of Limerick, Ireland August 2018

Submitted: August 2018

Declaration

I declare that this thesis is entirely my own work carried at the

University of Limerick and has not been previously submitted to this or any other university.

Xinxin Xiao

Abstract

Dealloyed nanoprous gold (NPG) is a porous material that possesses three dimensional frameworks of bicontinuous pores and ligaments, fabricated by electro/- chemical dissolution of the less noble component from an Au alloy. In this thesis, bioelectrodes were prepared by immobilising enzymes onto NPG and characterised in detail. For redox enzymes, osmium polymers were used to host the enzyme and as electron transfer mediators. Enzymatic biofuel cells (EBFCs) were assembled with redox enzyme modified bioelectrodes and characterised for a number of applications.

A glucose/O2 EBFC device has been developed that can harvest electricity in nonaqueous solvents, which may inspire new applications of EBFCs in bioelectrosynthesis. The EBFC was comprised of a NPG/[Os(4,4′-dimethyl-2,2′-

+/2+ bipyridine)2(polyvinyl -imidazole)10Cl] (Os(dmbpy)2PVI)/glucose oxidase (GOx)

+/2+ bioanode and a NPG/[Os(2,2′-bipyridine)2(polyvinylimidazole)10Cl]

(Os(bpy)2PVI)/bilirubin oxidase (BOx) biocathode. The power output of the cell decreased with increasing solvent hydrophobicity in the alcohols examined and the response of each electrode was restored when the electrodes were placed in phosphate buffer solution after operation in organic solutions.

To further expand the range of EBFC, a proof-of-concept “self-powered pulse generator” based on a supercapacitor/EBFC hybrid device has been developed. The device was prepared by immobilising redox enzymes with electrodeposited poly(3,4- ethylenedioxythiophene) (PEDOT) and Os(bpy)2PVI on NPG. Once charged by the internal EBFC, the device can be discharged as a supercapacitor at a current density of

2 mA cm-2 providing a maximum power density of 608.8 μW cm-2, an increase of a factor of 468 when compared to the power output from the EBFC itself.

To address the constrained oxygen supply that occurs at the biocathode, an oxygen-independent and membrane-less glucose biobattery was prepared by replacing the BOx based biocathode, with a solid-state NPG/MnO2 cathode. The potential of the

discharged MnO2 could be recovered, enabling the development of a proof-of-concept biobattery/supercapacitor hybrid device. The resulting device exhibited a stable performance for 50 cycles of self-recovery and galvanostatic discharge as a supercapacitor at 0.1 mA cm-2.

Wearable EBFCs are emerging as potential power sources for wearable microelectronic devices. A key requirement of such cells is the need for flexible electrodes.

Mechanically stable and flexible NPG electrodes were prepared using an electrochemical dealloying method consisting of a pre-anodization process and a subsequent electrochemical cleaning step. A flexible lactate/O2 EBFC consisting of a lactate oxidase based bioanode using electrodeposited Os(bpy)2PVI, and a BOx biocathode was placed between two commercially available contact lenses to avoid direct contact with the eye. When tested in air-equilibrated artificial tear solutions (3 mM lactate), a maximum power density of 1.7±0.1 μW cm-2 and an open-circuit voltage of 380±28 mV was obtained, values slightly lower than in phosphate buffer solution

(2.4±0.2 μW cm-2 and 455±21 mV, respectively). The decrease was mainly attributed to interference from ascorbate. After 5.5 h of operation, the EBFC retained 20% of its initial power output.

Finally, the utilization of NPG in fluidic biocatalysis was investigated. An electrochemically triggered sol-gel process was used to generate a thin silica layer for the immobilisation of lipase onto dealloyed NPG. The catalytic response of the entrapped lipase was examined using the hydrolysis of 4-nitrophenyl butyrate (4-

NPB) as a model reaction. A deposition time of 180 s and a lipase concentration of 3 mg/mL was used to prepare the optimised electrode. The operational stability of the silica immobilised enzyme was enhanced on NPG in comparison to that on planar gold, which may arise from confinement of the enzyme in the porous structure. The modified electrodes were incorporated into a 3D printed flow cell with conversion efficiencies of up to 100% after 8 cycles.

iii

ACKNOWLEDGEMENTS

I would like to take this opportunity to thank my supervisor Prof. Edmond

Magner for the great support on my IRC scholarship application and bring me to study abroad in Ireland. I am grateful for Edmond’s mentoring, encouragement, inspiration and never-ending patience on correcting my writings throughout the past four years. Most importantly, Edmond offers me the maximum freedom to conduct my research, making me an independent researcher.

I am very grateful to be an Associated Fellow of the BIOENERGY Marie

Curie ITN programme coordinated by Prof. Wolfgang Schuhmann at Ruhr University

Bochum. I benefit from the useful seminars, workshop, knowledge transfer and discussion with all the PIs and fellows. Among them, Prof. Dónal Leech and Dr. Peter

Ó Conghail at National University of Ireland Galway for providing Os polymers and

Dr. Roland Ludwig from BOKU-University of Natural Resources and Life Sciences for supplying glucose dehydrogenase are greatly appreciated.

I would like to thank my collaborators and friends for their kind support and hosting for my short-term visit: Assoc. Prof. Pengchao Si and Prof. Lijie Ci at

Shandong University, Prof. Jens Ulstrup, Prof. Jingdong Zhang and Assoc. Prof. Qijin

Chi at Technical University of Denmark, Prof. Yi Ding at Tianjin University of

Technology, Prof. Aihua Liu at Qingdao University.

My sincere thanks also goes to instrument scientists at Bernal Institute: Dr.

Serguei Belochapkine for performing sputtering of Au/Ag alloy films, Dr. Yina Guo for training me to use TEM and SEM, Dr. Lekshmi Kailas for AFM and FTIR training, Dr. Fathima Laffir for Raman spectroscope training, Dr. Wynette Redington for TGA training. Internal collaborators, Dr Micheál D. Scanlon for nice discussion,

Dr. Tadhg Kennedy for Autolab test and Robert Whelan for 3D printing of flow cells, are also acknowledged.

Many thanks to members of “Edmond’s research warriors” for their significant help: Alessandro Serleti, Cristina Carucci, Fernando Otero Diez, Dr.

Marcela Salazar Alvarez, Dr. Noreldeen Abdallah, Simon White, Dr. Till

Siepenkoetter, Dr. Urszula Salaj-Kosla, Dr. Victoria Gascón Pérez. I would also like to thank my friends from the “Chinese Community” in Limerick for the homely type company: Dr. Feng Chen, Dr. Fengwei Sun, Haiyang Zhang, Huan Ren, Dr. Lei Shi,

Dr. Xiaoming Ju, Dr. Yina Guo.

Finally, I express my profound gratitude to my parents for their unfailing love and moral support. I thank Ms. Yan Li, who was my girlfriend when I just came to Ireland and is my wife now, for her long-distance accompany and continuous encouragement.

Long, long had been my road and far, far was the journey;

I would go up and down to seek my heart's desire.

-Authored by Qu Yuan (c. 340-278 BC) -Translated by David Hawkes

Publications and conference presentations

Publications

1. Xinxin Xiao, Peter Ó Conghaile, Dónal Leech, Edmond Magner; Use of polymer coatings to enhance the response of redox-polymer-mediated electrodes,

ChemElectroChem 2018, DOI: 10.1002/celc.201800983

2. Xinxin Xiao, Edmond Magner; A quasi-solid-state and self-powered biosupercapacitor based on flexible nanoporous gold electrodes, Chemical

Communications 2018, 54, 5823-5826.

3. Xinxin Xiao, Till Siepenkoetter, Peter Ó Conghaile, Dónal Leech,

Edmond Magner; Nanoporous gold based biofuel cells on contact lenses, ACS

Applied Materials & Interfaces 2018, 2018, 10 (8), 7107–7116

4. Ciaran Lyons, Pratibha Dev, Pathik Maji, Neelima Rathi, Praveen K.

Surolia, Owen Byrne, Xinxin Xiao, Niall J. English, Edmond Magner, J. M. D.

MacElroy, K. Ravindranathan Thampi; Silicon-bridged triphenylamine-based organic dyes for efficient dye-sensitised solar cells, Solar Energy 2018, 160, 64-75

5. Francesca Lopez, Till Siepenkoetter, Xinxin Xiao, Edmond Magner,

Wolfgang Schuhmann, Urszula Salaj-Kosla; Potential pulse-assisted immobilization of Myrothecium verrucaria bilirubin oxidase at planar and nanoporous gold electrodes, Journal of Electroanalytical Chemistry 2018, 812, 194-198

6. Xinxin Xiao, Till Siepenkoetter, Robert Whelan, Urszula Salaj-Kosla,

Edmond Magner; A continuous fluidic bioreactor utilising electrodeposited silica for lipase immobilisation onto nanoporous gold, Journal of Electroanalytical Chemistry

2018, 812, 180-185

7. Xinxin Xiao, Peter Ó Conghaile, Dónal Leech, Roland Ludwig, Edmond

Magner; An oxygen-independent and membrane-less glucose

biobattery/supercapacitor hybrid device, Biosensors and Bioelectronics 2017, 98,

421-427

8. Xinxin Xiao, Christian Engelbrekt, Minwei Zhang, Zheshen Li, Jens

Ulstrup, Jingdong Zhang, Pengchao Si; A straight forward approach to electrodeposit tungsten disulfide/poly (3, 4-ethylenedioxythiophene) composites onto nanoporous gold for the hydrogen evolution reaction, Applied Surface Science 2017, 410, 308-

314

9. Xinxin Xiao, Peter Ó Conghaile, Dónal Leech, Roland Ludwig, Edmond

Magner; A symmetric supercapacitor/biofuel cell hybrid device based on enzyme- modified nanoporous gold: an autonomous pulse generator, Biosensors and

Bioelectronics 2017, 90, 96-102

10. Till Siepenkoetter, Urszula Salaj-Kosla, Xinxin Xiao, Serguei

Belochapkine, Edmond Magner; Nanoporous Gold Electrodes with Tuneable Pore

Sizes for Bioelectrochemical Applications, Electroanalysis 2016, 28, 2415-2423

11. Till Siepenkoetter, Urszula Salaj-Kosla, Xinxin Xiao, Peter Ó Conghaile,

Marcos Pita, Roland Ludwig, Edmond Magner; Immobilisation of redox enzymes on nanoporous gold electrodes: applications in biofuel cells, ChemPlusChem 2016, 81,

1-9

12. Xinxin Xiao, Pengchao Si, Edmond Magner; An overview of dealloyed nanoporous gold in bioelectrochemistry, Bioelectrochemistry 2016, 109, 117-126

13. Xinxin Xiao, Edmond Magner; A biofuel cell in non-aqueous solution,

Chemical Communications 2015, 51, 13478-13480

Poster presentations

1. Xinxin Xiao and Edmond Magner; Fluidic enzymatic biofuel cells based on screen-printed electrodes, 6th Annual NUI Galway-University of Limerick

Alliance Research Day (Apr. 29, 2016, Limerick, Ireland)

vii

2. Xinxin Xiao and Edmond Magner; The construction of enzymatic fuel cells operating in organic media, Nanoweek Conference 2015 (Oct. 21-22, 2015,

Limerick, Ireland)

3. Xinxin Xiao and Edmond Magner; The construction of enzymatic fuel cells operating in organic media, XXIII International Symposium on

Bioelectrochemistry and Bioenergetics (Jun. 14-18, 2015, Malmö, Sweden)

4. Xinxin Xiao and Edmond Magner; The construction of enzymatic fuel cells operating in organic media, 5th Annual NUI Galway-University of Limerick

Alliance Research Day (Apr. 21, 2015, Galway, Ireland)

5. Xinxin Xiao, Pengchao Si, and Edmond Magner; The application of microscopes (scanning electron microscope and transmission electron microscope) for identification of nanoporous gold based nanocomposites interfaces, The first joint meeting of the Scottish Microscopy Group (SMG) and the Microscopy Society of

Ireland (MSI) (Nov. 27-28, 2014, Glasgow, Scotland)

Oral presentations

1. Xinxin Xiao, and Edmond Magner; Contact lenses supported enzymatic biofuel cells; ISE Satellite Student Regional Symposium on Electrochemistry, (Oct.

27, 2017, Cork, Ireland)

2. Xinxin Xiao, Peter Ó Conghaile, Dónal Leech, Roland Ludwig, Edmond

Magner; An energy-harvesting device based on supercapacitive enzyme-modified nanoporous gold electrodes: an autonomous pulse generator; 68th Annual Meeting of the International Society of Electrochemistry (Agu. 28- Sep. 1, 2017, Rhode Island,

U.S.A)

3. Xinxin Xiao, Peter Ó Conghaile, Dónal Leech, Roland Ludwig, Edmond

Magner; An energy-harvesting device based on supercapacitive enzyme-modified nanoporous gold electrodes: an autonomous pulse generator; XXIV International

Symposium on Bioelectrochemistry and Bioenergetics (July 3-7, 2017, Lyon, France) viii

4. Xinxin Xiao, Peter Ó Conghaile, Dónal Leech, Roland Ludwig, Edmond

Magner; An energy-harvesting device based on supercapacitive enzyme-modified nanoporous gold electrodes: an autonomous pulse generator; 7th Annual NUI

Galway-University of Limerick Alliance Research Day (Apr. 20, 2017, Galway,

Ireland); Best paper award (science section)

5. Xinxin Xiao, Edmond Magner; Symmetric supercapacitor/biofuel cell hybrid device using enzyme-modified nanoporous gold: an autonomous pulse generator, 2016 Bioinorganic and Catalysis meeting (Sep. 15-16, 2016, Marseille,

France)

6. Xinxin Xiao, Kanso Hussein, Peter Ó Conghaile, Dónal Leech, Edmond

Magner; A capacitor/biofuel cell hybrid device using enzyme-modified nanoporous gold, 67th Annual Meeting of the International Society of Electrochemistry (Aug. 21-

25, 2016, The Hague, Netherlands)

Table of Contents

Table of Contents Declaration ...... i

Abstract ...... ii

ACKNOWLEDGEMENTS ...... iv

Publications and conference presentations ...... vi

Publications ...... vi

Poster presentations ...... vii

Oral presentations ...... viii

Table of Contents ...... x

LIST OF FIGURES ...... xv

LIST OF TABLES ...... xxii

Chapter 1. Literature review ...... - 1 -

1.1 Enzymes and electrochemistry ...... - 1 -

1.1.1 Electrochemistry and bioelectrochemistry ...... - 1 - 1.1.2 Enzyme immobilisation ...... - 2 - 1.1.3 Electron transfer between an enzyme and electrode surface ...... - 4 - 1.1.4 Common enzymes ...... - 8 -

1.1.4.1 Glucose oxidase ...... - 8 -

1.1.4.2 Cellobiose dehydrogenase ...... - 9 -

1.1.4.3 Bilirubin oxidase ...... - 10 -

1.1.4.4 Cytochrome c ...... - 10 -

1.1.5 Biosensors ...... - 11 - 1.1.6 Biofuel cells ...... - 13 - x

1.1.6.1 Power density of a biofuel cell...... - 14 -

1.1.6.2 Open circuit voltage of a biofuel cell ...... - 15 -

1.1.6.3 Recent advances of biofuel cells ...... - 17 -

1.2 Nanoporous gold ...... - 18 -

1.2.1 Nanomaterials in bioelectrochemistry ...... - 18 - 1.2.2 Dealloying of nanoporous gold ...... - 19 - 1.2.3 Applications of dealloyed nanoporous gold ...... - 20 - 1.2.4 Dealloyed nanoporous gold in bioelectrochemistry ...... - 21 -

1.2.4.1 NPG based glucose biosensors ...... - 22 -

1.2.4.2 NPG based BFCs ...... - 24 -

1.3 Scope of this project...... - 26 -

1.4 References ...... - 27 -

Chapter 2. A biofuel cell operating in nonaqueous solutions ...... - 43 -

2.1 Introduction ...... - 43 -

2.2 Experimental section ...... - 46 -

2.2.1 Materials ...... - 46 - 2.2.2 Enzyme immobilization ...... - 47 - 2.2.3 Electrochemical measurements ...... - 47 -

2.3 Results and discussion ...... - 48 -

2.4 Conclusions ...... - 54 -

2.5 References ...... - 55 -

Chapter 3. A symmetric supercapacitor/biofuel cell hybrid device based on enzyme- modified nanoporous gold: an autonomous pulse generator ...... - 57 -

3.1. Introduction ...... - 57 -

3.2. Experimental section ...... - 60 -

3.2.1. Materials ...... - 60 - 3.2.2. Enzyme immobilisation procedures ...... - 61 - 3.2.3. Morphology characterisation ...... - 61 - 3.2.4. Electrochemical measurements ...... - 62 - 3.2.5. Calculation of device performance ...... - 62 -

3.3. Results and discussion ...... - 64 -

3.3.1 Electrochemical characterisation of the capacitive bioelectrodes ...... - 64 - 3.3.2 Morphology characterisation ...... - 69 - 3.3.3 Hybrid device testing ...... - 70 - 3.3.4 A proof-of-concept pulse generator ...... - 77 -

3.4. Conclusions ...... - 78 -

3.5 References ...... - 79 -

Chapter 4. An oxygen-independent and membrane-less glucose biobattery/supercapacitor hybrid device ...... - 83 -

4.1. Introduction ...... - 83 -

4.2. Experimental section ...... - 86 -

4.2.1. Materials and apparatus ...... - 86 -

4.2.2. Preparation of the enzyme modified anode and NPG/MnO2 cathode ... - 87 - 4.2.3. Electrochemical measurements ...... - 87 -

4.3. Results and discussion ...... - 88 -

4.3.1 Electrochemical performance of NPG/MnO2 ...... - 88 - 4.3.2 Electrochemical performance of the bioanode and assembled biobattery - 92 - 4.3.3 Electrochemical performance of the hybrid device ...... - 93 -

4.3.4 Potential recovery of NPG/MnO2 ...... - 97 -

4.4. Conclusions ...... - 99 -

4.5 References ...... - 100 -

xii

Chapter 5. Nanoporous gold based biofuel cells on contact lenses ...... - 104 -

5.1. Introduction ...... - 104 -

5.2. Experimental section ...... - 107 -

5.2.1. Materials and apparatus ...... - 107 - 5.2.2. Electrochemical dealloying ...... - 108 - 5.2.3. Enzyme immobilisation ...... - 108 - 5.2.4. Electrochemical measurements ...... - 110 -

5.3. Results and discussion ...... - 110 -

5.3.1. Electrochemical dealloying ...... - 110 - 5.3.2. Characterisation of the bioelectrodes ...... - 118 - 5.3.3. Performance of EBFC ...... - 123 -

5.4. Conclusions ...... - 128 -

5.5 References ...... - 128 -

Chapter 6. A continuous fluidic bioreactor utilising electrodeposited silica for lipase immobilisation onto nanoporous gold ...... - 133 -

6.1. Introduction ...... - 133 -

6.2. Experimental section ...... - 136 -

6.2.1. Materials and apparatus ...... - 136 - 6.2.2. Enzyme immobilisation procedures and activity measurement ...... - 137 - 6.2.3. Flow cell design ...... - 138 - 6.2.4. Determination of immobilised protein concentration ...... - 139 -

6.3. Results and discussion ...... - 139 -

6.4. Conclusions ...... - 145 -

6.5 References ...... - 146 -

Chapter 7. Conclusions and recommendations ...... - 149 -

7.1. Conclusions ...... - 149 - xiii

7.2. Recommendations ...... - 151 -

7.3 References ...... - 151 -

Appendix 1: ...... - 154 -

xiv

LIST OF FIGURES

Figure 1.1 Schemes of different immobilisation methods...... - 4 -

Figure 1.2 Schematic diagram of different electron transfer mechanisms: A) mediated electron transfer (MET); B) direct electron transfer (DET)...... - 5 -

Figure 1.3 Effects of enzyme configuration on the electrode...... - 7 -

Figure 1.4 Structures of common enzymes: (A) Aspergillus niger glucose oxidase

(PDB ID: 3QVP), (B) Myricoccum thermophilum cellobiose dehydrogenase (PDB ID:

4QI6), (C) Myrothecium verrucaria bilirubin oxidase (PDB ID: 2XLL), (D) cytochrome c (PDB ID: 3CYT)...... - 8 -

Figure 1.5 Scheme of the structure of a biosensor...... - 12 -

Figure 1.6 Scheme of a biofuel cell...... - 13 -

Figure 1.7 (A) Polarisation curves of a bioanode and biocathode. (B) Voltage-current profile (B) and power density-voltage profile of an EBFC. Key parameters of an

EBFC are highlighted. Reprinted with permission from [123]. Copyright (2008)

American Chemical Society...... - 14 -

Scheme 2.1 The underlying mechanism and employed polymer structure on the proposed anode and cathode...... - 45 -

Scheme 2.2 A typical NPG electrode preparation process...... - 47 -

Figure 2.1 (A) Schematic diagram of the biofuel cell. Cyclic voltammograms (CVs) of NPG/Os(dmbpy)2PVI/GOx modified electrodes in PBS (B) and 95% ACN (C) at a

-1 scan rate of 5 mVs . (C) CVs of NPG/Os(bpy)2PVI/BOx electrode in PBS (D) and

95% ACN (E) at a scan rate of 5 mVs-1. (F) Polarization and power curves for the

EBFC in O2 bubbled PBS (initial curve: solid line; after testing in 95% ACN: dotted line) and 95% ACN (dashed line)...... - 48 -

Figure 2.2 Chronoamperometry response of the NPG/Os(dmbpy)2PVI/GOx bioanode at +0.2 V vs. SCE in PBS (A) and 95% ACN (B)...... - 49 -

Figure 2.3 Chronoamperometry response of the NPG/Os(bpy)2PVI/BOx biocathode at +0.1 V vs. SCE in PBS (A) and 95% ACN (B)...... - 50 -

Figure 2.4 Chronoamperometric response of blank electrodes without enzymes in

PBS: NPG/Os(dmbpy)2PVI at +0.2 V vs. SCE (A); NPG/Os(bpy)2PVI at +0.1 V vs.

SCE (B)...... - 51 -

Figure 2.5 Operational stability of EBFC in PBS (A) and 95% ACN (B)...... - 52 -

Figure 2.6 Storage stability of the proposed EBFC...... - 52 -

Figure 2.7 Power density curve of the EBFC in different percentages of O2 bubbled

ACN containing 5 mM glucose (A) and different organic solvents containing 5 mM glucose (B). (C) Plot of the power density versus log P (data points taken from reference [3]). The error bars correspond to the values recorded for three EBFCs.- 54 -

Figure 3.1 Cyclic voltammograms (CVs) of various electrodes (deposition time: 300 s)...... - 64 -

Figure 3.2 CVs of (B) NPG/PEDOT/Os(bpy)2PVI/FAD-GDH and (C)

-1 NPG/PEDOT/Os(bpy)2PVI/BOx electrodes at a scan rate of 5 mV s ...... - 65 -

Figure 3. 3 TEM of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (450 s deposition). . - 66 -

Figure 3.4 SEM images of the bare NPG (A) and NPG/PEDOT/Os(bpy)2PVI/FAD-

GDH (300 s deposition) (B)...... - 67 -

Figure 3.5 Optimisation curves for the anode (A) and cathode (B)...... - 67 -

Figure 3.6 (A) CVs of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (300 s deposition) at various scan rates. (B) Calibration plots of the oxidation/reduction peak current vs. scan rate...... - 68 -

Figure 3.7 Lineweaver–Burk plot for the NPG based FAD-GDH bioelectrode. ... - 69 -

Figure 3.8 (A) STEM dark-field micrograph of bare NPG at 300 kX. (B) TEM image of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (300 s deposition)...... - 70 -

Figure 3.9 (A) Schematic diagram of the BFC. (B) Polarisation and power curve for the BFC consisting of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH bioanode and

NPG/PEDOT/Os(bpy)2PVI/BOx biocathode...... - 71 -

xvi

Figure 3.10 Schematic diagrams of the hybrid device working at the self-charging

(A) and galvanostatic discharging mode (B) (with simplified charge-discharge description on the capacitive NPG/PEDOT hybrid). (C) Charge/discharge curves of the as-assembled biocapacitor (black line); Experimental setup: reset at open-circuit and cutoff at 0.4 V, followed by discharging at 0.2 mA cm-2 for 0.5 s (red line). (D)

Magnified image the first discharge segment...... - 72 -

Figure 3.11 Galvanostatic charge/discharge at 10 µA cm-2 of the symmetric supercapacitor consisting of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH and

NPG/PEDOT/Os(bpy)2PVI/BOx in a blank PBS solution...... - 73 -

Figure 3.12 (A) Charge/discharge curves of the biocapacitor for 50 cycles;

Experimental setup: reset at open-circuit and cutoff at 0.4 V, followed by discharging at 0.2 mA cm-2 for 0.5 s. (B) Charge/discharge curves of the biocapacitor upon various discharging current densities; Experimental setup: reset at open-circuit and cutoff at 0.4 V, followed by discharging at 0.005 (a), 0.01 (b), 0.02 (c), 0.05 (d), 0.1

(e), 0.2 (f), 0.5 (g), 1 (h), 2 (i) mA cm-2 for 0.2 s...... - 74 -

Figure 3.13 (A) Polarisation and power curves for the BFC (initial curves: solid line; after long-term operation: dashed line). (B) CVs of the bioanode in PBS before and after long-term testing...... - 75 -

Figure 3. 14 (A) Polarisation and power curve for a planar Au based BFC. (B) Long- term testing of the biocapacitor for 50 cycles; Experimental setup: reset at OCV and cutoff at 0.35 V, followed by discharging at 0.2 mA cm-2 for 0.5 s and cutoff at 0 V. . -

76 -

Figure 3.15 Charge/discharge curves of the series connection of three biofuel cells

(see Fig. S9); Experimental setup: the connected cells were allowed to reset at open- circuit for 2 hours, followed by discharging at 10 μA for 0.5 ms every 5 s reset. (A) is the first measurement of 1500 discharging pulses; (B) is for the measurement of 1300 discharging pulses upon refilling of fresh solutions; insets show zooms at the specific cycles...... - 78 -

xvii

Figure 4.1 The relationship between deposition time and specific capacitance of the corresponding NPG/MnO2 at a potential of 0.45 V vs. SCE...... - 88 -

Figure 4.2 SEM images of the bare NPG (A) and NPG/MnO2 obtained by electrodeposition for 30 s (B), and NPG/MnO2 for 300 s (C)...... - 89 -

Figure 4.3 Raman spectrum of the electrodeposited MnO2 film on the Au film

(deposition time: 180 s)...... - 90 -

Figure 4.4 SEM (A) and TEM (C) image of NPG/MnO2 (deposition time: 180 s). (B)

EDX spectra of bare NPG and NPG/MnO2 (deposition time: 180 s). (D) LSV of

-1 NPG/MnO2 (deposition time: 180 s) in 0.1 M pH 7.0 PBS at a scan rate of 2 mV s . . -

90 -

Figure 4.5 TEM images of the bare NPG...... - 91 -

Figure 4.6 Stability of NPG/MnO2 and Au/MnO2 (deposition time: 180 s) in 0.1 M

PBS, pH 7.0 in a potential range from 0 to 0.5 V vs. SCE at a scan rate of 100 mV s-1.

...... - 92 -

Figure 4.7 (A) CVs of the NPG/PEDOT/Os(bpy)2PVI/FAD-GDH bioanode. (B) The performance of the biobattery in the presence of 10 mM glucose...... - 93 -

Charge/discharge curves of the biocapacitor upon various discharging current densities; Experimental setup: reset at open-circuit for 30 min, followed by discharging at 0.005 (a), 0.01 (b), 0.02 (c), 0.05 (d), 0.1 (e), 0.2 (f), 0.5 (g), 1 (h), 2 (i) mA cm-2 for 0.2 s and cutoff at 0 V...... - 95 -

Figure 4.9 (A) OCP of NPG/MnO2 in the presence of N2 or O2. The electrode was discharged by scanning potential from 0.5 to 0 V vs. SCE at a scan rate of 1 mV s-1.

(B) OCP of the bioanode upon the addition of 10 mM glucose...... - 98 -

Figure 4.10 Potential profile of NPG/PEDOT (negative electrode of the capacitor)//NPG/MnO2 (positive electrode of the capacitor) over 50 cycles.

xviii

Experimental procedure: reset at open-circuit for 30 min, followed by discharging at

0.1 mA cm-2 and cutoff at 0 V...... - 99 -

Scheme 4.1 Schematic diagrams of the hybrid device working at the reset (left) and galvanostatic discharging mode (right). The scheme in the middle depicts the relevant potential differences, with potential shifts caused by galvanostatic discharging (blue arrows) and on the recovery of the potential during the quiescent step (red arrows). ... -

98 -

Scheme 5.1 Schematic diagram of the assembly of the modified contact lens (A) and the configuration of the EBFC (B)...... - 107 -

Scheme 5.2 Electron transfer route between the electrode surface and LOx mediated by Os(bpy)2PVI...... - 120 -

Figure 5.1 Cyclic voltammogram of a NPG electrode in an ice-bath cooled solution of 2 mL 2 M HCl containing 2 mM NaNO2 and 2 mL solution of 20 mM NA in acetonitrile. Scan rate: 200 mV s-1...... - 109 -

Figure 5.2 (A) Linear sweep voltammogram of Ag70/Au30 alloy in 0.5 M NaF. (B)

Cyclic voltammograms of the as-anodised NPG (1.5 V) in 1 M H2SO4...... - 112 -

Figure 5.3 (A) Schematic diagram of the electrochemical dealloying process. (B-E)

SEM images of the porous structure of NPG obtained at different conditions. B: anodisation in 0.5 M NaF at 1.5 V vs. SCE for 10 min, anodisation and cycling potential in 1 M H2SO4 for 1 (C), 2 (D) and 15 (E) potential cycles...... - 114 -

Figure 5.4 SEM images the porous structure of NPG obtained with different conditions. A: anodisation in 0.5 M NaF at 1.05 V vs. SCE for 10 min, anodisation and cycling potential in 1 M H2SO4 for 1 (B), 2 (C) and 15 (D) potential cycles.- 115 -

Figure 5.5 SEM images the porous structure of NPG obtained by cycling potential in

1 M H2SO4 for 2 (A), 15 (B) and 30 (C) potential cycles without anodisation. ... - 116 -

Figure 5.6 Plots of (A) roughness factor; (B) residual silver content; (C) pore size;

(D) crack width obtained after potential cycling of as-anodised NPG...... - 116 -

xix

Figure 5.7 (A) Digital photo of the PET supported NPG obtained via electrochemical dealloying. (B) SEM images of the electrochemically dealloyed NPG, and the corresponding microstructure under a 40-degree (C) and 60-degree bend (D). ... - 118 -

Figure 5.8 (A) Schematic drawing of the experimental setup for resistance measurements according to reference [56]. (B) The relationship between bending angles of the NPG film and sheet resistance...... - 118 -

Figure 5.9 SEM image of the deposition layer of NPG/Os(bpy)2PVI/LOx (360 s deposition)...... - 120 -

Figure 5.10 Operational stability of NPG/Os(bpy)2PVI/LOx (360 s deposition) in air- equilibrated PBS...... - 122 -

Figure 5.11 (A) Effect of the deposition time for NPG/Os(bpy)2PVI/LOx on the catalytic response towards 3 mM lactate in air-equilibrated 0.1 M pH 7.0 PBS at 250 mV vs. SCE. Blue line indicates the surface coverages of the Os polymer obtained by various deposition times. (B) CVs of NPG/Os(bpy)2PVI/LOx (360 s deposition) in air-equilibrated 0.1 M pH 7.0 PBS at a scan rate of 5 mV s-1. (C) Catalytic response of

NPG/Os(bpy)2PVI/LOx (360 s deposition) towards various concentrations of lactate in air-equilibrated 0.1 M pH 7.0 PBS at 250 mV vs. SCE. (D) CVs of the BOx modified electrode in 0.1 M pH 7.0 PBS at a scan rate of 5 mV s-1...... - 123 -

Figure 5.12 Photograph of the contact lens encapsulated EBFC (A) and testing setup

(B). (C) Polarisation and power curves for the EBFC consisting of

NPG/Os(bpy)2PVI/LOx bioanode and NPG-BOx biocathode. (D) Operational stability of the EBFC at 150 mV in artificial tears...... - 124 -

Figure 5.13 Effects of the presence of 0.18 mM ascorbate towards bioanode (A) and biocathode (B). Inset of (A) shows the response on a bare NPG...... - 126 -

Scheme 6.1 Schematic drawing of the electrodeposition of silica for enzyme immobilisation at a constant negative potential...... - 134 -

Scheme 6.2 (A) CAD drawing of the flow cell consisting of top-plate and base. (B)

Sectional view of the base (unit: mm). (C) Detailed view of part 3 from (B). (D)

Photograph of the flow cell during operation; the arrows indicate the direction of flow...... - 135 -

Scheme 6.3 Hydrolysis reaction of 4-NPB catalysed by lipase...... - 136 -

Figure 6.1 (A) The effect of deposition time on the catalytic performance of

NPG/silica/lipase obtained in 1 mg mL-1 lipase (measured by immersion in 2 mL of

75 μM 4-NPB for 0.5 h). (B) The effect of lipase concentration on the catalytic performance of NPG/silica/lipase obtained by depositing for 180 s (measured by immersion in 2 mL of 75 μM 4-NPB for 1 h)...... - 140 -

Figure 6.2 TEM images showing the NPG/silica/lipase obtained in 1 mg mL-1 lipase with various deposition durations: (A): 60 s, (B): 180 s, (C): 360 s; the arrow in (A) and (B) indicates the silica/lipase layers. The pores are filled with silica/lipase in (C);

Scale bars at the right-bottom of (A), (B) and (C) indicate 30 nm...... - 140 -

Figure 6.3 SEM image of a bare NPG and NPG/silica/lipase (180 s deposition). . - 141

Figure 6.4 Storage stability of NPG/silica/lipase; Response was measured by immersion in 2 mL of 75 μM 4-NPB for 1 h...... - 142 -

Figure 6.5 The effect of flow rate at the catalytic behavior of NPG/silica/lipase towards 75 μM 4-NPB...... - 143 -

Figure 6.6 Conversion ratio of 2 mL of 75 μM 4-NPB by cycling in a loop at a flow rate of 0.05 mL min-1...... - 144 -

Figure 6.7 Regeneration of an NPG/silica/lipase electrode. The response was measured by immersion of the electrode in 2 mL of 75 μM 4-NPB for 1 h...... - 145 -

xxi

LIST OF TABLES

Table 3.1 Electrochemical capacitances of various modified electrodes (obtained from Figure 3.1A)...... - 65 -

Table 3.2 Cell performance of the hybrid device upon various discharging current densities (jpulse) (obtained from Figure 3.11B)...... - 75 -

Table 4.1 List of properties of enzymatic power sources utilising glucose as substrate.

...... - 96 -

Table 5.1 Summary of results obtained by cycling potential in 1 M H2SO4 of as- anodised NPG at 1.05 V...... - 114 -

Table 5.2 Summary of results obtained by cycling potential in 1 M H2SO4 of as- anodised NPG at 1.5 V...... - 115 -

Table 5.3 Summary of results obtained by cycling potential in 1 M H2SO4 of

Ag70/Au30 alloy without anodisation...... - 116 -

Table 5.4 List of properties of tear based EBFCs...... - 127 -

xxii

CHAPTER 1:

Literature review

Chapter 1

Chapter 1. Literature review

1.1 Enzymes and electrochemistry

1.1.1 Electrochemistry and bioelectrochemistry Electrochemistry is the branch of chemistry that deals with the interrelation of electrical and chemical processes [1]. The concept of “electrochemistry" originated in 1791 when Luigi Galvani studied how frogs’ nerves work. Electrochemistry kept developing significantly during electrification era when Michael Faraday discovered Faraday’s law in 1831. In the 20th century, great advancements have been made with birth of new theories and techniques. The Nernst equation, which describes the relationship between the voltage of an electrochemical cell and its properties, was formulated by Walther Hermann Nernst, bringing him the 1920 Nobel Prize in Chemistry. Around 1922, Jaroslav Heyrovsky, regarded as the “father of electroanalytical chemistry", developed a method of analysis named as polarography, the current-voltage curve, making him the winner of 1959 Nobel Prize in Chemistry. Rudolph A. Marcus received the 1992 Nobel Prize in Chemistry for his contributions to the theory of electron transfer reactions in chemical systems, proposed in 1956. Nowadays, the scope of electrochemistry includes a plurality of chemical processes (e.g., metal corrosion, electrophoresis), devices (e.g., electrochemical sensors, batteries, fuel cells and electrochromic displays) and engineering (e.g., metal electrodeposition). It is an inter-disciplinary science that involves materials science, physics, chemistry, biology, medicine, energy and environmental science and earth sciences. It allows us to detect diseases quickly, use smart phones everywhere with the support of high-capacity and light batteries, protect cars and ships from corrosion, find out cheap and efficient catalysts to accelerate chemical reactions, and decrease our dependence on fossil energy and following environmental crisis and global climate change, etc. It is no exaggeration to say that electrochemistry has a profound impact on all aspects of modern human life.

- 1 -

Chapter 1

Bioelectrochemistry, as a branch of electrochemistry, focus on the study of biological electron transfer (ET) processes, especially at interfaces [2]. The general research subjects are biological molecules with sizes ranging from large proteins, to smaller species such as nicotinamide adenine dinucleotide (NADH). To mimic biological processes, biological molecules, mainly enzymes, are immobilised and oriented onto various designed electrode surfaces to enable ET to occur. Bioelectrochemistry allows us to develop biosensors that monitor physiological parameters in terms of health care and biofuel cells that harvest electricity from physiological fluids in order to power implantable [3] and wearable devices [4] in the future, as well as enlargement of our scope such as bioelectrosynthesis [5].

1.1.2 Enzyme immobilisation Enzyme immobilisation is a necessary and important step for bioelectrochemistry [6] and biocatalysis [7]. Immobilising enzymes provides advantages including recycling, enhanced stability, prevention of protein contamination, possible control of catalytic properties and electron transfer, etc. [7, 8] While some disadvantages caused by immobilisation should also be kept in mind. For example, immobilisation may cause conformational changes [9] leading to activity losses of enzymes [10]. Moreover, immobilisation may decay ET and causes mass diffusion problems of substrate or cofactors [11, 12]. General methods of immobilizing enzymes on an electrode surface are summarised below (Figure 1.1). i) Adsorption. By directly and simply contacting an enzyme solution with the solid support for a specific duration, enzyme molecules can be adsorbed via weak interactions involving of Van der Waal's forces and electrostatic or hydrophobic interactions [6]. This method is generally non-destructive for enzyme activity, but allows biomolecule leaching. Layer-by-layer (LBL) adsorption technique has been developed for multilayers of one or more species of enzymes by alternative deposition of oppositely charged species [13]. Another technique is Langmuir–Blodgett (LB)

- 2 -

Chapter 1 method based on the enzymes association with a monomolecular film (self-assembled by amphiphilic biomolecules) formed at an air/water interface [14]. ii) Covalent attachment. Covalent coupling between functional groups of the enzyme and support can result in stable bonds that prevent the detachment of enzymes. On the other hand, irreversible deactivation of the enzyme caused by chemical modification or changing of conformation may occur. Generally, the primary amine group of lysine side chains are coupled to amino or carboxyl groups on the supports with the assistance of glutaraldehyde [15, 16] or carbodiimide coupling [12, 17]. Supports with functional groups include self-assembled monolayer (SAM) modified noble metal electrodes [18] and predesigned polymeric supports [19]. iii) Entrapment. Entrapment means incorporation of enzyme within three dimensional matrices such as a gel, paste or polymer. Typically, entrapment allows relatively large amount of enzyme to be immobilized. However, the physical restraints might allow enzyme leakage. Conducting polymers (CPs) such as polyaniline, polypyrrole or polythiophene have been widely used for entrapment of enzymes during polymerisation [20, 21]. The thickness of the polymer film can be easily controlled to reduce the diffusion barrier with tuning deposition parameters [11]. Sol-gel process involves hydrolysis of alkoxide precursors such as tetramethoxysilane (TMOS) or tetraethoxysilane (TEOS) followed by condensation of the hydroxylated units, which leads to the formation of a gel matrix for enzyme entrapment [22, 23]. Recently, an electrochemical assisted sol-gel process has been developed for facile, one-step decoration [24]. In this case, the electrogenerated OH- at negative potential act as catalysts to gelify alkoxide precursors onto the electrode surface. Thus, co-existing enzymes in the electrolyte could be simultaneously entrapped in the electrogenerated silica film. iv) Cross-linking. Immobilisation of enzymes by cross-linking with each other or in the presence of a functionally inert protein such as bovine serum albumin (BSA) is also widely adopted due to its simplicity and the strong chemical binding. The constraints are similar to that of covalent bonding, resulting in irreversible

- 3 -

Chapter 1 deactivation of the enzyme. The bifunctional agent, glutaraldehyde [25], is widely used as a cross linker. Another important cross-linking agent is poly(ethylene glycol) diglycidyl ether (PEGDGE) introduced by A. Heller et al. [26], with epoxide groups that react with the amine groups of enzymes or imidazole groups of redox polymers at room temperature. By enveloping the redox enzymes with Os redox polymers, electron communication of redox centres of enzymes and polymers with electrode surfaces can be built, irrespective of the spatial orientation and connect to electrode redox centres of multiple enzyme layers [27]. v) Affinity. Affinity bonds, but not covalent bonds, between an activated support with specific functional groups and affinity tags on a protein sequence allow controlled and oriented enzyme immobilisation. Several affinity methods, such as avidin-biotin [28], lectin-carbohydrate [29] and metal cation-chelator [30] interactions, have been described. The main drawback is the relatively high complexity of those procedures.

Figure 1.1 Schemes of different immobilisation methods.

1.1.3 Electron transfer between an enzyme and electrode surface Enzyme catalysed electron transfer (ET) between enzymes and the electrode surface is crucial for bioelectrodes (Figure 1.2). The basic model of ET was developed by Marcus et al. in the1950s [31]. According to Marcus’ theory, the rate of - 4 -

Chapter 1

ET is determined by parameters such as the reorganisation energy and the distance between the donor and the acceptor. ET rate through a saturated protein backbone decreases by an order of magnitude for every 2.3 Å [32]. To achieve sufficient ET, an upper threshold of 14 Å has been recommended [33] and the optimal electrode configuration should ensure that the ET distance is as short as possible [34]. Generally the redox centres of enzymes (such as glucose oxidase) are buried deeply within protein shells, making direct ET impossible as they are too far away. For this reason mediated ET is generally employed.

Figure 1.2 Schematic diagram of different electron transfer mechanisms: A) mediated electron transfer (MET); B) direct electron transfer (DET).

ET between electrodes and active sites of enzymes occurs via direct ET (DET) and mediated ET (MET). Redox mediators can be dioxygen and artificial molecules including ferrocene derivatives, ferricyanide, conducting organic salts, quinone compounds, transition-metal complexes, phenothiazine and phenoxazine compounds [34]. They are capable of promotion of ET at different potentials, which are related to the operation potentials of the resulting biosensors. Dioxygen, commonly available in physiological fluids, although not in a high concentration (0.14 mM in arterial blood and 0.08 mM in intestinal tissue [35, 36]), has been used as

- 5 -

Chapter 1 the electron acceptor since the first glucose biosensor was reported in 1962 [37]. The electron carrying mediators can operate in solution or co-immobilized onto the electrode. The former case is not suitable for in vivo applications and the latter case makes bioelectrode fabrication more complicated. Redox polymers introduced by Heller et al. [27, 38, 39] are the most important group of mediators for the construction of bioelectrodes [40]. The redox polymer also acts as the host matrix to immobilise enzymes via electrostatic interactions, entrapment and/or chemically cross-linking, resulting in a catalytic film permeable to the fuels and necessary ions [27]. Polymer backbones bearing organometallic groups (e.g. Os complex [40], ferrocene [41-43], cobaltocene [44]), organic dyes (e.g. viologen [44, 45], phenothiazine [46, 47]), and quinone [48, 49] have been synthesized for mediated bioelectrodes. Their redox potentials are determined by the intrinsic nature of the anchored redox mediator and can be tuned by changing the binding status. For example, introducing electron-withdrawing/donating groups into the ligand sphere allows the change of the redox potential of Os complex [46, 50]. DET, which eliminates issues associated with the mediator system, was first reported in 1970s by studying the reversible ET of cytochrome c on different electrodes [51-53]. The DET of some copper blue proteins (e.g. laccase [54], bilirubin oxidase [55]), flavin proteins (glucose oxidase [56], fructose dehydrogenase [57], cellobiose dehydrogenase [58]) have been reported [59]. DET capable enzymes generally harbor a catalytic cofactor reacting with oxidant and an intra-enzyme ET relay [60]. For example, D-Fructose dehydrogenase (FDH) is a flavohemoprotein with three subunits [61, 62]: subunit I with covalently bound FAD showing a pH-dependent formal redox potential, E0’, of -0.231 V vs. Ag/AgCl at pH 5.5 for catalytic oxidation of D-fructose; subunit II containing three heme c moieties with E0’ of -0.062, 0.054 and 0.34 V at pH 5.5 and the heme with the lowest E0’ is suggested to be the exit site for ET pathway [63]; subunit III does not carry any redox centers, whose function is still unclear. Bacterial derived, hetero-

- 6 -

Chapter 1 oligomertric FAD dependent GDH (FAD-GDH) consisting of a FAD based catalytic subunit, a small chaperone subunit and a multi-haem ET subunit [64], is capable of DET. 3Fe-4S cluster has been identified in this type FAD-GDH, acting as a ET bridge between FAD and the multi-heme cytochrome c subunit [65]. Quinohaemoprotein- type pyrroloquinoline quinone (PQQ)-dependent enzymes (e.g. GDH [66-68], alcohol dehydrogenase (ADH) [69] and lactate dehydrogenase (LDH) [70]) contain a PQQ prosthetic group that is coordinated with the apoenzyme with Ca2+ and heme-c moieties performing as ET relay [71]. Hydrogenases catalysing H2 oxidation possess a [NiFe] or [FeFe] catalytic cofactor accompanied with Fe-S clusters for facilitated intermolecular ET between the catalytic cofactor and the b-type cytochrome (figure x) [72-74]. Although not widely applied but interesting to be included, deglycosylated enzymes permit DET showing decreased molecular weight and molecular size and thus shorten ET distance [75]. Mano et al. deglycosylated Aspergillus niger GOx preserving its activity with the direct redox reaction of FAD occuring at -0.49 V vs. Ag/AgCl [76]. Pyranose dehydrogenase (PDH) carries covalently bound FAD as the cofactor and its deglycosylated form showed DET on graphite electrode [77]. As discussed, an optimally designed electrode configuration has to ensure the ET distance is as short as possible, indicating why the appropriate orientation of immobilised biomolecules is so important for DET (Figure 1.3).

Figure 1.3 Effects of enzyme configuration on the electrode.

- 7 -

Chapter 1

1.1.4 Common enzymes Enzymes are biological catalysts that maintain living systems, e.g. respiratory and energy-transducing processes, and marcomolecules that only function when maintaining structural integrity. Among them, oxidoreductases, catalysing oxidation and reduction processes, are of vital importance and great interests. The properties of some enzymes and proteins commonly used in bioelectrochemistry are reviewed.

Figure 1.4 Structures of common enzymes: (A) Aspergillus niger glucose oxidase (PDB ID: 3QVP), (B) Myricoccum thermophilum cellobiose dehydrogenase (PDB ID: 4QI6), (C) Myrothecium verrucaria bilirubin oxidase (PDB ID: 2XLL), (D) cytochrome c (PDB ID: 3CYT).

1.1.4.1 Glucose oxidase

Glucose oxidase (β-D-glucose:oxygen 1-oxidoreductase, GOx; EC 1.1.2.3.4) is a flavoprotein that catalyses the two-electron involving oxidation of β-D-glucose [78]. GOx is a dimeric protein with a molecular weight of 16 kDa and contains a cofactor of flavin adenine dinucleotide (FAD) [79]. In nature, GOx uses molecular oxygen as the electron acceptor. Briefly, with FAD reduced to FADH2, GOx catalyses the oxidation of β-D-glucose to D-glucono-δ-lactone, which is subsequently hydrolysed to gluconic acid. To convert FADH2 back to FAD, O2 acts as the oxidant to reduce H2O2. Notably, FAD is deeply buried by the protein surrounding layer. According to X-ray structure studies, there is a minimal distance of 13-18 Å between the periphery of GOx and the N 7 nitrogen of the isoalloxazine rings of FAD (Figure 1.4 A) [79]. Only the ends of the coenzyme directly contact with the solvent, resulting

- 8 -

Chapter 1 in FAD being tightly attached. FAD can be removed from the enzyme in saturated

(NH4)2SO4 solution that is acidified to pH 1.4 with concentrated H2SO4 (7% v/v) [80]. GOx has been purified from various microorganisms, including Aspergillus and Penicillium [78]. The most commercialised and commonly utilized one is A. niger GOx, with an overall dimensions of 70×55×80 Å [79]. GOx has been extensively used in biosensors and biofuel cells (BFCs) since Clark & Lyons described the first example of glucose electrode relying on GOx in 1962 [37] and the first enzymatic BFC was reported by Yahiro et al. in 1964, where GOx served as the anodic catalyst [81]. GOx is highly specific for β-D-glucose and shows negligible activities with other sugars, rendering it the main enzyme used for diabetes monitoring. For ET between electrodes and GOx to occur, various electron acceptors (e.g. ferrocene derivatives, quinone compounds and transition-metal complexes etc.) have been proposed, and DET is difficult to achieve due to the long distance for electrons to travel. Even though, attempts including penetration the protein with molecular wires [82] have been made to establish DET of GOx. New carbon materials, i.e. carbon nanotubes (CNTs) [83] and graphene [84], based DET have been claimed, while arguments recently arose that observed redox peaks of FAD are from adsorbed apo-GOx (i.e. deflavined GOx that is not enzymatic active) [85, 86].

1.1.4.2 Cellobiose dehydrogenase

Cellobiose dehydrogenase ((cellobiose:acceptor) 1-oxidoreductase, CDH; EC 1.1.99.18) is a flavocytochrome that catalyses the oxidation of different carbohydrates (e.g. cellobiose, lactose and glucose) [87, 88]. Molecular masses of CDH vary from 90 to 100 kDa depend on the fungal source (e.g. P. chrysosporium, T. villosa, Myriococcum thermophilum and Corynascus thermophilus, etc.). Accordingly, substrate specificity, optimal pH, DET efficiency and surface binding affinity also differ. CDH possesses two separate domains connected via a polypeptide linker region. The flavodehydrogenase domain (DHCDH) is catalytically active and the cytochrome domain (CYTCDH) has a haem b that can operate as an electron relay, i.e.

- 9 -

Chapter 1 an in-built mediator. The dual-domain feature enables efficient DET between the active sites and electrode surfaces (Figure 1.4 B). The DET of CDH was first documented in 1996 by adsorption of CDH on graphite electrode [89]. Based on the fact that recently discovered ascomycete CDHs show high turnover rates for glucose and works at neutral pH, increased interests arose in CDH based biosensors and BFCs exhibiting simple configurations by circumventing the use of mediators. Currently, significant efforts have been made to develop new nanostructured electrode materials to increase DET by appropriate enzyme orientation [58, 90], as well as modification of CDH via enzyme engineering to increase faradaic currents of CDH-based electrodes.

1.1.4.3 Bilirubin oxidase

Bilirubin oxidase (BOx), which belongs to the multicopper oxidase (MCO) family, utilises four Cu ions to reduce dioxygen into water [91]. Fungal BOxs have been identified in Pleurotus ostreatus, Trachyderma tsunodae, Myrothecium verrucarria, etc. [92] According to their spectroscopic and magnetic characters, the four Cu centres can be classified as type I Cu, type II Cu and a pair of type III coppers (Figure 1.4 C) [93]. Type I Cu accepts electrons transferred from substrates such as bilirubin to the final electron acceptor O2, which is converted to H2O, without releasing activated oxygen species. The O2 reduction site is a trinuclear Cu centre formed by type II Cu and the pair of type III Cu’s. Tremendous efforts have been made since 2001 when BOx was first reported for dioxygen reduction [94]. Beneficial from its tolerance towards halide ions, thermostability, high activity and operational stability at neutral pH, as well as relative ease to establish DET on electrodes [95], BOx is a reasonable candidate as biocathode catalyst of BFC working in physiological conditions.

1.1.4.4 Cytochrome c

Cytochrome c (cyt c) contains a heme group, which is covalently linked to the cyt c peptide chain through thioether bonds with cysteine residues 14 and 17. The

- 10 -

Chapter 1 crystallographic dimension of yeast cyt c is 2.5×2.5×3.7 nm [96]. The heme iron is hexacoordinated with His18 and Met80 (Figure 1.4 D), with a redox potential of ~260 mV vs. NHE [52, 97] [98]. Aliphatic and aromatic amino acid side chains render the heme group in a very hydrophobic microenvironment. Cyt c plays an important role in mitochondrial electron transport and intrinsic type II apoptosis [98]. Cyt c has relatively high thermodynamic stability, small size, high solubility in water, commercial availability. It can undergo DET at modified electrodes with a favourable orientation of the heme facing the electrode that allowed rapid electron transfer. Self-assembled monolayers (SAMs) on gold electrodes are universally used for cyt c immobilisation. Immobilised cyt c shows a quasi-reversible, heterogeneous one electron transfer process [99].Choosing a SAM requires a balance [100]: If the interaction between the surface and the protein are weak, cyt c adsorption on the SAM would be incomplete; if the interaction were strong, the protein might denature on the surface. DET of cyt c is easily to be achieved for electrochemists. For biochemists, the high helical content and heme cofactor of cyt c allow studies by different spectroscopes. Based on the reasons above, cyt c has been widely studied [101]. Cyt c is widely investigated as a model redox protein [102-105].

1.1.5 Biosensors A biosensor is an analytical device that transforms biological information (typically the concentration of analyte) into a quantifiable and processable signal that can be collected [106]. Two main components of a biosensor are a biorecognition element (e.g. antibodies, nucleic acids, enzymes, microorganisms) and a transducer (Figure 1.5). Biosensors can be classified into affinity sensors, enzymatic sensors and microbial sensors according to the type of biological components used [107]. Electrochemical biosensors can be categorized based on signal transduction mode, i.e. amperometry, potentiometry, conductimetry, AC impedance spectroscopy, etc. [2].

- 11 -

Chapter 1

Figure 1.5 Scheme of the structure of a biosensor.

Sensitivity, referring to the signal intensity upon substrate concentration and selectivity, the capability that distinguishes target species with interferences, are two important and basic parameters for evaluation of biosensors [108]. Sensitivity, response time, detecting/linear range and detection limit can be determined by performing calibration curves that are obtained by dosing standard solutions of the analyte and by plotting steady-state responses. The linear range of glucose biosensors should cover the concentration range of blood glucose (i.e. 2-10 mM [109]) in order to void a dilution step. From the viewpoint of application, reproducibility and stability must also be considered. The reproducibility of biosensor fabrication based on nanomaterials should be emphasized as differences arise from lab to lab even using the same procedures. The relatively poor stability of biosensors (in the range from days to few months) restricts their application. The lifetime can be extended by the judicious choice of enzyme immobilisation method and genetic engineering [110]. The applications of biosensors can be found in clinical, food and environmental areas. Notably, in a commercial viewpoint, around 85% of world market of biosensors is blood glucose measurement for the management of diabetes [111]. Trends in the development of biosensor include the development of cheap, miniaturized and disposable strip biosensors [112], implantable sensors for continuously in vivo measurement [113], and non-invasive biosensors that analyse tears [114], urine [115], sweat [116] and saliva [117].

- 12 -

Chapter 1

1.1.6 Biofuel cells Broadly speaking, biofuel cells or biological fuel cells comprise microbial fuel cells and enzymatic fuel cells that utilize microorganisms and enzymes as catalysts, respectively [118]. Enzymatic biofuel cell (EBFC) for converting the chemical energy of fuels into electrical current is an electrochemical device that utilises enzyme modified electrodes, i.e. bioanodes and biocathodes (Figure 1.6) [119]. Typically, but not always, enzymes such as glucose oxidase (GOx), alcohol dehydrogenase (ADH) and hydrogenase for the oxidation of glucose, alcohol and hydrogen, respectively are utilised at the anodes and multi-copper oxidases (e.g. laccase and bilirubin oxidase) that catalyse the reduction of oxygen are used at the cathode. The first EBFC was reported by Yahiro et al. in 1964, where glucose oxidase served as the anodic catalyst [81]. Compared to conventional fuel cells, EBFCs show advantages such as operating at mild condition (20-40℃, near-neutral pH [120]), high selectivity and following exemption of membrane, making them emerging alternatives for portable and miniaturized batteries [2]. The bottlenecks are limited stability and relatively low power outputs (in the scale of µW cm-2).

Figure 1.6 Scheme of a biofuel cell.

- 13 -

Chapter 1

1.1.6.1 Power density of a biofuel cell

EBFCs are generally characterised by current density (i.e. polarisation curve) and power density profiles obtained by recording the steady state current by gradually imposing a constant potential or linear sweep voltammetry at a low scan rate (<1 mV s-1) to maintain at a steady-state status and exclude the non-faradaic current contributed from high-capacitance electrode materials (Figure 1.7) [121]. It’s noteworthy that mostly reported current density and power density in literature are respected to the projected geometric surface area [122]. In the case for an EBFC utilising a bioanode and biocathode with different surface areas, the geometric surface area of the limiting bioelectrode displaying lower current can be applied for power density calculation.

Figure 1.7 (A) Polarisation curves of a bioanode and biocathode. (B) Voltage-current profile (B) and power density-voltage profile of an EBFC. Key parameters of an EBFC are highlighted. Reprinted with permission from [123]. Copyright (2008) American Chemical Society.

- 14 -

Chapter 1

A typical bioelectrocatalytic reaction comprises i) mass transport of the reactant from the bulky solution to the active site on the solid surface, ii) enzymatic reaction with the reactant, iii) electron transfer between the active site and the electrode, and iv) diffusion of the products into the solution from the solid-liquid interface ([124]. For a comprehensive study, it’s important to identify the limiting step of the bioelectrocatalytic process. In the point view of kinetics, improvement on the rate-limiting step leads to an increased current density and, consequently, power density of the EBFC [124, 125]. On one hand, general strategies (including utilisation of highly-active enzymes, using nanomaterials based electrodes, enhancement of substrate diffusion etc.) can be employed to increase the current density for a single bioelectrode based on the principle to improve the turnover rate and density of catalytic active site [126]. On the other hand, improved fuel cell designs for considerable overall current/power density should be taken into account.

1.1.6.2 Open circuit voltage of a biofuel cell

One critical challenge of EBFCs is their output voltages are generally incompatible with that are required for commercial available microelectronic devices (1-3 V [127]), although transistors requiring an operation voltage of 0.5 V and even lower have been developed [128]. The open circuit voltage (OCV) of a biofuel cell is

0 limited by the thermodynamic value E , which can be calculated from the overall reaction occurring on both bioelectrodes. E0 is a measurement of thermodynamic driving force allowing the spontaneous reaction with a negative Gibbs free-energy change ∆G. In the standard states, the relationship between standard Gibbs free

0 -1 0 energy change ∆G (kJ mol ) and E (V) can be expressed by the equation [129]: |∆G0| = 푛FE0 (1.1) ∆G0 of biochemical reactants have been summarized by Alberty et al. [130]

A glucose/O2 EBFC using glucose oxidase (GOx) or glucose dehydrogenase (GDH) as bioanode catalysts undergoes an overall reaction: - 15 -

Chapter 1

1 β − D − Glucose (C H O ) + O → D − Gluconolactone (C H O ) + 6 12 6 2 2 6 10 6

H2O (1.2)

0 -1 The |∆G | of the reaction is 227.23 kJ mol , n = 2 and Ecell is calculated to be

1.18 V according to eq. 1.1. Based on the same calculation, H2/O2 and lactate/O2 EBFCs present thermodynamic limits of 1.23 [126] and 1.0 V [131], respectively. In practice, the registered OCV of a EBFC is much lower than E0 due to the presences of three types of potential losses, namely kinetic related (ηact), ohmic losses

(I∑R) and mass transport losses (ηdiff), in the system (Figure 1.7B). The relationship between registered OCV and E0 can be determined by [122]:

0 OCV = E − η푎푐푡 − I∑R − η푑푖푓푓 (1.3)

where ηact is the overpotential required to overcome energy barriers on the electrode-electrolyte interfaces, ∑R is the sum of all resistances associated with current I that flow through solid electrodes, electrolyte and various interconnections,

ηdiff is the mass transport based overpotential due to the reactant diffusion limitation as the reaction sites drain reactant rapidly at this stage (Figure 1.7B). Visually, the measured OCV can be read from the power density profile (Fig. 1.7C), which is consistent to the difference between the onset redox potentials of the bioanode and biocathode [123]. Although the term “onset potential” is quite fuzzy due to the difficulties in define the exact starting points for electrochemical oxidation or reduction [122], it can be obtained, in practice, by comparing the potential-current profiles of bioelectrode in the presence and absence of the substrate [132, 133]. Thus, the measured OCV can also be expressed as [126]:

표푛푠푒푡 표푛푠푒푡 OCV = E푐 − E푎 − I∑R푒 (1.4) 0′ 0′ OCV = (E푐 − η푐) − (E푎 − η푎) − I∑R푒 (1.5)

onset onset where Re is the experimental resistance, Ec and Ea are the observed onset

0’ 0’ potentials for cathode and anode, Ec and Ea are the thermodynamic onset potentials for cathode and anode, ηb and ηa are the overpotentials for the cathode and anode, respectively. Eq. 1.4 and 1.5 suggest the strategies to maximize OCV of a single

- 16 -

Chapter 1

EBFC via bring the starting potentials of both bioanode and biocathode closer to those of the enzymes/cofactors [134]. EBFCs containing enzymes with significantly high potential windows are thus favorable. MCOs based biocathodes undergoing DET are generally preferred. Laccase (Lac) exhibits much higher redox potential (up to 0.78 V vs. NHE [135]) of T1 Cu site than that of BOx (ca. 0.67 V vs. NHE [136]). However, BOx exhibits a high activity at physiological conditions (i.e. neutral pH) and less sensitive to chloride ions, making it a better candidate for implantable EBFCs. Lac is usually inhibited by chloride ions and mostly active in pH 4-5, making it a suitable choice for non- implantable applications. On the bioanode side, NAD dependent dehydrogenases

0’ present a low onset potential due to the low formal potential of NAD (E NADH/NAD+: - 0.33 V vs. SHE [137]), however, the cofactor is loosely bound. FAD dependent

0’ dehydrogenases (E0’FADH2/FAD: -0.18 V [138]) are preferred over PQQ (E PQQH2/PQQ: 0.12 V vs. SHE [139]) due to the lower redox potential.

1.1.6.3 Recent advances of biofuel cells

Significant inputs to the fields of EBFC in recent years have been roughly summarized below. i) Enzyme cascades have been used for deep oxidation of fuel so that increase the efficiency of bioanode, which is generally the limiting factor of EBFC [140]. ii) Air-breathing biocathode has been constructed to consume oxygen gas directly due to the low solubility and poor diffusion coefficients of dissolved oxygen in aqueous solutions [141]. iii) Miniaturized and implantable EBFCs in a range of plants and animals have been widely reported to date [142]. iv) Flexible EBFCs, that are bendable or stretchable, applicable toward flexible displays and portable biomedical devices with curved surface like electronic contact lenses [143], have also been developed [144]. v) Microfluidic EBFCs that work on microfluidic chips allow efficient mass transport by increasing flow rate and increased power density due to the high surface-to-volume ratio [145]. vi) Self-powered sensors that integrate sensors and EBFCs have been developed, avoiding any externally powered - 17 -

Chapter 1 multimeter or potentiostat for readout [146, 147]. vii) Hybrid devices that with dual- function of supercapacitor and EBFC, maybe called “biosupercapacitor”, enable high power discharge cycles while the capacitance is continuously recharged through the biocatalytic energy conversion thus overcoming the diffusion issues [148].

1.2 Nanoporous gold

1.2.1 Nanomaterials in bioelectrochemistry The basic requirements of an electrode material contain good conductivity, ease of modification, chemical stability, surface regeneration, etc. In order to load enzyme, materials possess high surface area, accessibility and biocompatibility, are expected. Nanomaterials, at least one dimension at nanoscales (between approximately 1 and 100 nm [149]), have promising physiochemical properties in terms of improved plasticity, noticeable thermal and optical property changes, higher reactivity and activity, faster electron/ion transport and novel quantum mechanical features [150]. During the last 20 years, there have been extensive studies of nanomaterials in bioelectrochemistry, in order to improve the performance of biodevices and allow miniaturisation and microfluidic integration. To briefly summarize nanomaterials applied in bioelectrochemistry, herein they are grouped according to their dimensions, i.e. 0 D (nanodots, nanoparticles [151], C60), 1D (nanowires [152], nanotubes [153], nanofibers [154]), 2D (nanosheets, graphene [155]) and 3D (nanoporous materials) materials. Among them, carbon nanotubes, graphene and gold nanoparticles are commonly used materials for bioelectrochemical applications due to their ease of preparation and intrinsic conductivity. The potential toxic effects of such nanomaterials should be considered when use in vivo [156]. Porous or nanoporous structured electrodes, a kind of 3D materials, are of significant interest for applications in bioelectrochemistry [157]. The available surface area (generally 2-1000 times larger than the geometric area) can host significantly higher amounts of biomolecules (enzymes, antibodies, etc.). Tremendous efforts have - 18 -

Chapter 1 been devoted to developing porous electrode materials with oriented or random pore distribution. Examples include mesoporous carbon [158], nanoporous noble metals [159], nanoporous metal oxides [160], etc. Among them, porous gold electrodes, featuring good electrical conductivity, chemical stability and biocompatibility, can be fabricated by dealloying [161-163], Au nanoparticle (AuNP) assembly [55, 95, 147, 164, 165], anodization [166], hard [167-169] and soft [170] template method. Au electrode can be easily modified with self-assembled monolayers (SAMs) formed by thiols [12, 171], diazonium grafting [172-174] and electropolymerization [11, 175, 176] for enzyme immobilization to achieve direct and mediated ET.

1.2.2 Dealloying of nanoporous gold The corrosion process of alloys is a very common phenomenon as it occurs all the time and all around us, such as the corrosion of steels and brasses which leads to undesirable materials failure. Indians of pre-Columbian Central America used a so- called “depletion gilding” technique to create a layer of pure gold artefact by etching copper away from Cu/Au alloys. The pioneering work in 1960-70s by Forty [177] and Pickering [178, 179] revealed that chemically removing less-noble component from a binary alloy (e.g. Cu/Au, Ag/Au, Si/Au) is a “dealloy” process, which leads to a bicontinuous nanoporous structures. In 2001, Erlebacher et al. proposed a continuum model underlying porosity evolution during dealloying [180, 181]. At the solid/electrolyte interface, more noble atoms (e.g. Au in Ag/Au alloy) are chemically driven to aggregate into clusters and form islands rather than spreading over the surface, thus forming pores. Continuously, newly formed pores open up regions of virgin alloy for further etching of less noble atoms (e.g. Ag in Au/Ag alloy) and dealloying continues. In 2004, Erlebacher et al. systematically demonstrated that dealloying can be used to fabricate nanoporous gold by dealloying commercially available 12 carat Ag/Au leaves with a thickness of approximately 100 nm [162]. Since then, dealloyed NPG has been widely studied for technological applications involving of catalysis,

- 19 -

Chapter 1 and sensing. Generally, NPG fabricated by electro-/chemically dissolving the less noble component from Au alloys (i.e. dealloying) is a kind of porous material containing three dimensional frameworks of bicontinuous pores and ligaments [162]. The pore size (~5-700 nm [182]) can be easily tailored by tuning alloy compositions and dealloying conditions.

1.2.3 Applications of dealloyed nanoporous gold The advantage of NPG include large surface-volume-ratio, high electrical conductivity, chemical stability, biocompatibility, permeability, etc. [183] Thus, various potential applications of NPG have been evaluated recent years. Some areas but not all have been summarized below. i) Optical applications [184]. In general, there are two kinds of surface plasmon resonance (SPR), i.e. localized and propagating SPR. Like other nanostructured metals, NPG shows localized SPR excitations as its characteristic sizes (ligaments and pores) are much smaller than the wavelength of visible light (400-700 nm). Moreover, the integrated skeletons allow NPG possessing the feature of planar metal films that exhibit propagating SPR excitations [185]. Plasmonic NPG can be used to provide quantitative information about the surface binding events on it, e.g. an biotin-streptavidin binding system [185]. NPG is also active for surface-enhanced Raman scattering (SERS) [186] and a pore size dependent relationship has been proposed [182]. For example, the Raman scattering intensities of Rhodamine 6G on NPG increase with pore size decreasing from 700 to 5 nm [182]. ii) Heterogeneous catalysis [187, 188]. NPG is reactive for a wide variety of oxidation reactions. The most surprising thing is that NPG, as a non-supported Au catalyst, is catalytically active for CO oxidation [189, 190]. To work, it was thought that Au should be smaller than 5 nm and supported on high-surface-area oxides, e.g.

TiO2 [191]. The phenomenon suggests that dealloyed NPG can activate dioxygen. Compared with the role of low-coordination Au sites [192] that activate molecular oxygen, the residual Ag or Cu atoms of dealloyed NPG are more important as they

- 20 -

Chapter 1 promote oxygen absorption [193]. For example, pure NPG without residual Ag or Cu fabricated by dealloying Au/Zr alloys shows poor activity comparing with dealloyed NPG with residual Ag or Cu [194]. Dealloyed NPG is thus also active for other oxygen-assisted coupling reactions including methanol oxidation [195] and bezyl alcohol oxidation [196]. iii) Energy devices. Due to its high conductivity, NPG has been used as a current collector (that collects electrons) material for energy devices. Platinum coated NPG membranes have been studied as proton exchange membrane fuel cell electrodes [197]. By plating Pt onto NPG via simple benchtop chemistry, it allows low loading of Pt. Pseudocapacitive transition metal oxides (e.g. MnO2 [198], RuO2 [199]), conducting polymers [200] or mixture of them can be easily deposited onto NPG, and used as electrochemical supercapacitors. For example, polypyrrole-decorated NPG leaves can be assembled as a flexible supercapacitor using solid electrolyte, which offers large specific capacitance even at various curvatures [200]. NPG has also been studied as current collectors of lithium ion batteries [201, 202]. It should been pointed out that NPG is still too expensive for large-scale applications. Alternatives might be other cheap dealloyed nanoporous metals (e.g. Cu, Ni, Ti). In addition to the applications mentioned above, NPG is also a good candidate for electrochemistry, particularly bioelectrochemistry, which will be introduced in more detail in the following section.

1.2.4 Dealloyed nanoporous gold in bioelectrochemistry Benefiting from its nanostructure, unique features of dealloyed NPG for bioelectrochemistry can be summarized: i) Enzymes confined in pores displays higher stability upon high temperature [203] and organic solvents [204] in contrast to free one; ii) High content of low index crystalline faces [192] on NPG surface promotes its higher electrochemical activity towards redox molecules with sluggish [205]; iii) Compared with planar Au, NPG can stabilize thiolate self-assembled monolayers (SAMs) [206, 207], beneficial from the defective sites, lattice strain and residual Ag

- 21 -

Chapter 1 on the ligament surface; iv) NPG preserves its electrochemical accessibility in biofouling environment by excluding large proteins like fibrinogen [208, 209]. The applications of dealloyed NPG in glucose biosensors and EBFCs are described below.

1.2.4.1 NPG based glucose biosensors

Glucose biosensors can be divided into three generations according to the ET mechanisms [34]. Using dioxygen as a mediator, first generation glucose biosensors rely on detection of H2O2 produced by the reduction of dioxygen. Planar gold shows no catalytic effect for the oxidation of H2O2 with an onset potential up to +0.7 V (vs. SCE), a rather high detection potential which may cause interferences from oxidising species. In contrast, NPG is electroactive to H2O2 oxidation with an onset potential of +0.2 V (vs. SCE) and oxidation peak at +0.4 V (vs. SCE) [210], indicating a possible alternative to Pt. A glucose biosensor has been developed by the physical adsorption of GOx within NPG and protected by Nafion film, which displayed a linear range up to 18 mM, a sensitivity of 0.049 µA mM-1 and a detection limit of 196 µM at an applied potential of +0.4 V vs. SCE [210]. NPG is also electroactive to reduction of

H2O2 with an onset of -0.1 V (vs. Ag/AgCl) [211, 212]. Responses of H2O2 to NPG with pore size of 18, 30, 40 and 50 nm have been compared and smaller pores possess higher sensitivity [212]. After immobilisation of GOx via thiol linking and applying a potential of -0.2 V (Ag/AgCl), the highest sensitivity (8.6 µA cm-2 mM-1) to glucose was obtained with 30 nm NPG, instead of 18 nm NPG showing the highest response to H2O2 [212]. The small pores of the 18 nm NPG are not large enough for more enzymes to enter. Further, Prussian Blue (PB) was electrodeposited onto NPG for to enhance the reduction of H2O2 [213]. Working at just 0 V (vs. Ag/AgCl), the biosensor fabricated by physical adsorption of GOx onto NPG/PB yielded a linear response up to 30 mM glucose and a sensitivity of 50 µA cm-2 mM-1, as well as remarkable tolerance to interferences including lactate, uric acid and ascorbic acid.

- 22 -

Chapter 1

First generation glucose biosensors rely on the use of dissolved oxygen in physiological fluids and suffer stoichiometric limitation as oxygen concentrations are around 1 order of magnitude lower than blood glucose [34]. Thus MET with synthetic electron acceptors was proposed in 1984 when ferrocene mediated electrochemistry of GOx was developed [214], which is called second generation glucose biosensors. SAM is generally used as the bridge for enzyme anchoring, whose thickness has been shown to give effects on NPG based glucose sensing [12]. SAMs formed by shorter carbon chain is helpful for efficient ET and mass transport of mediators. The electrochemical behaviors of two diffusing redox mediators, p-benzoquinone (BQ) and ferrocenecarboxylic acid (FCA), on NPG have been well studied [16]. BQ, which carries on “inner sphere” ET, displays higher interfacial electrochemical ET rates and more efficient function towards mediated GOx electrocatalysis than FCA, with “outer sphere” character. Notable, in both cases which undergo fast kinetic redox reactions [157], the deep inside the NPG electrode does not contribute to glucose sensing. Moreover, in the presence of excess mediator, the sensitivity of sensor is mainly decided by the amount of enzyme immobilized rather than the mediator concentration [16]. One-step electropolymerisation of hybrid film of 3,4-ethylenedioxythiophene (EDOT) and GOx on nanoporous gold (NPG) has been proposed [11]. Mediated with BQ, the biosensor, that obtained by two cycles deposition with sufficient enzyme loading and effective substrate diffusing, exhibits a sensitivity of 7.3 µA cm-2 mM-1 and a linear range of 0.1-15 mM. Encapsulation of GOx within Os-redox polymer onto NPG has also been employed as glucose biosensor with limit of detection of 2.0 (±0.1) μM [215]. NPG based third generation glucose biosensors, working by DET, have not been reported as far as we know. Glucose dehydrogenase (GDH) is another class of enzymes used for glucose sensing, while the regeneration of cofactor, nicotinamide adenine dinucleotide (NAD, NAD+/NADH for the oxidized/reduced form), making the situation more complicated. The oxidation of NADH requires a high overpotential thus mediators are

- 23 -

Chapter 1 required. NPG can enhance the sluggish ET kinetics of NADH [210]. The electrooxidation of NADH takes place on planar gold at +0.72 V (vs. SCE), while occurs on NPG at +0.52 V (vs. SCE), originated from its low coordinated gold atoms. For alcohol dehydrogenase (ADH) modified NPG, it can work at +0.5 V (vs. SCE) for ethanol sensing without using mediators [210]. One advantage of GDH biosensor is its oxygen-independent feature. The flavocytochrome cellobiose dehydrogenase (CDH) is another enzyme whose substrates are generally carbohydrates including but not restricted to glucose [87]. It’s fascinating because DET of CDH is relatively easy to achieve as it contains two separate domains that allow intramolecular ET. Immobilized in Os-redox polymer, NPG supported CtCDH displays limit of detection of 16 (±0.1) μM for the determination of glucose [215]. Despite massive attention and study on NPG based biosensors in the last decade, and in particular for glucose monitoring, it is still pending for commercial realisations with scientific and engineering challenges to be comprehensively addressed. Achievement of accessibility into the deep porous part would be a prior issue. Further characterisations of biomolecules distribution existing in pores need to be delivered, rather than very general information of activity assay.

1.2.4.2 NPG based BFCs

A glucose/O2 BFC has been reported in 2012 by immobilising glucose oxidase/laccase onto SAM modified NPGs, with a maximum power density of 52 µW cm-2 testing in pH 5.0 air saturated buffer solution containing 100 mM glucose [206]. Myrothecium verrucaria bilirubin oxidase (MvBOx) has been physically adsorbed onto NPG and obtained bioelectrocatalytic reduction of oxygen, undergoing efficient DET, can be used for biocathode [216]. In contrast, negligible catalytic response was obtained at a planar gold electrode. Another oxygen reduction enzyme, Trametes hirsute laccase (ThLc), shows well-defined DET at bare NPG, while no Faradaic response occurs at unmodified polycrystalline gold [217]. These cases indicate NPG providing a favourable microenvironment for intimate communication between the - 24 -

Chapter 1 gold walls of the pores and copper site, originating from synergistic effects of high active surface and preferential orientation of enzymes. NPG based EBFCs composed of GOx or CtCDH as bioanodes, combined with MvBOx or Melanocarpus albomyces laccaseas biocathodes, have been constructed and well evaluated, by modifying electrodes with hydrogels containing enzyme, Os-redox polymers and the cross-linking agent PEGDGE [215]. A maximum power density of 41 µW cm-2 was obtained in the case of CtCDH/MvBOx biofuel cell in 5 mM lactose and O2 saturated buffer (pH 7.4).

Siepenkoetter et. al. prepared a glucose/O2 EBFC consisting of a MvBOx- modified NPG biocathode (500 nm thickness) and a Glomorella cingulata GDH based bioanode (300 nm), registering maximum power densities of 17.5 and 7.0 μW cm-2 in phosphate buffer solution and artificial serum containing 5 mM glucose, respectively [173]. In this work, BOx was attached covalently via carbodiimide coupling to a diazonium-functionalised surface and GDH was immobilised onto the electrode by drop-casting of a mixture of enzyme, Os polymer and cross-linker. Pore size, inversely proportional to the specific surface area [212, 218], is an important factor for the performance of the bioelectrode. An optimal DET current of BOx was observed on porous gold electrodes with a pore size of ca. 20 nm which was large enough for enzyme accommodation and also high surface area for sufficient enzyme loading [173]. For the Os redox polymer mediated GDH bioelectrode prepared by drop-casting showed no dependence of biocatalytic response on pore size ranging from 9-60 nm, indicating a planar electrode like performance [173]. This is because of a “cap” like configuration of modification layer, hampering the full occupancy of deeper pores [215, 216]. This can be overcome by electrodeposited redox hydrogel/enzyme thin films in a stable and uniformly distributed manner [176]. Regarding to future practical applications, EBFC is progressing towards to flexible, fluidic and miniaturized type [120]. A thin film of Au/Ag can be sputtered onto bendable polymer film or micro channel, followed by dealloying to obtain NPG for further modification. A macro size gold wire can be rebuilt with external layer of

- 25 -

Chapter 1

NPG [219]. These formats of NPG are promising for high power density microelectronics, combining with its non-toxicity property.

1.3 Scope of this project

The aim of the project in this thesis is to explore the applications of nanoporous gold in bioelectrochemistry, especially in enzymatic based biofuel cells, with addressing critical problems.

In Chapter 2, a well-studied glucose/O2 enzymatic biofuel cell operating in organic solvents was investigated. The cell utilises glucose oxidase and bilirubin oxidase immobilised nanoporous gold electrodes with the assistance of redox polymers. Common organic solvents including methanol, ethanol, acetone, acetonitrile etc. containing a small amount water (≤5%) and conducting salt (tetraethylammonium p-toluenesulfonate, TEATS) have been used. The electrochemical catalytic response was compared in aqueous solution and organic solvents. A relationship between the power output of the cell and solvent hydrophobicity was found. In Chapter 3, the author integrated supercapacitors with enzymatic biofuel cells to prepare hybrid devices in order to harvest significantly higher power output. To overcome the detachment issue of drop-casted Os polymer/enzyme mixture, electrodeposition of Os polymer for enzyme immobilisation was used. Conducting polymer, poly(3,4-ethylenedioxythiophene) (PEDOT), was co-electrodeposited with Os polymer/enzyme onto NPG to improve the capacitance. The biofuel cell utilising capacitive electrodes was tested in a sequence of self-charging and discharging, exhibiting interesting behaviours with intermittent electric outputs. Control experiments of planar gold based cells were prepared and tested, highlighting the enhancement by NPG. The performance of three individual cells that were connected in series was also examined.

- 26 -

Chapter 1

To overcome the oxygen supply limitation occurring on biocathode, a consumable cathode using electrodeposited MnO2 on NPG was studied in Chapter 4. An oxygen-independent glucose biobattery was obtained by assembling the

NPG/MnO2 cathode with the bioanode developed in Chapter 3. A concept of “biobattery/supercapacitor” was proposed, as an interesting phenomenon of potential recovery on the discharged NPG MnO2 was found. The obtained cell can function following a sequence of reset and discharging, which is similar to the biofuel cell/supercapacitor hybrid device fabricated in Chapter 3. In Chapter 5, the author aimed to prepare a wearable biofuel cell. A new-type flexible NPG was fabricated using electrochemical dealloying treatment towards polyethylene terephthalate (PET) supported Au-Ag alloy films. The dealloy conditions were optimised according to surface characterisation of obtained NPG in terms of pore size, roughness factor and residual Ag content. Lactate oxidase and bilirubin oxidase were further immobilised onto the optimal NPG electrodes to prepare a lactate/O2 biofuel cell. To evaluate its potential application as a contact lens supported power source, artificial tears were used. The interference effect arising from co-existing ascorbate was examined. Chapter 6 reports an electrodeposition method to entrap lipase using sol-gel derived silica on NPG. The electrodeposition parameters including the electrolyte composition and deposition duration have been optimised. A 3D printed flow cell with a laminar flow channel of a depth of 500 μm was used to perform fluidic biocatalysis using the developed bioelectrodes. Chapter 7 summarises the achievement of the thesis, accompanying with recommendations for further investigation.

1.4 References

[1] A.J. Bard, L.R. Faulkner, Electrochemical Methods: Fundamentals and Applications, John Wiley & Sons Inc., New York (2001). [2] P.N. Bartlett, Bioelectrochemistry: Fundamentals, Experimental Techniques and Applications, John Wiley & Sons Inc., New York (2008).

- 27 -

Chapter 1

[3] S.C. Barton, J. Gallaway, P. Atanassov, Enzymatic Biofuel Cells for Implanted and Microscale Divces, Chem. Rev., 104 (2004) 4867-4886. [4] A.J. Bandodkar, J. Wang, Wearable Biofuel Cells: A Review, Electroanalysis, 28 (2016) 1188-1200. [5] R.D. Milton, S.D. Minteer, Enzymatic Bioelectrosynthetic Ammonia Production: Recent Electrochemistry of Nitrogenase, Nitrate Reductase, and Nitrite Reductase, ChemPlusChem, 82 (2017) 513-521. [6] A. Sassolas, L.J. Blum, B.D. Leca-Bouvier, Immobilization Strategies to Develop Enzymatic Biosensors, Biotechnol. Adv., 30 (2012) 489-511. [7] U.T. Bornscheuer, Immobilizing Enzymes: How to Create More Suitable Biocatalysts, Angew. Chem. Int. Ed., 42 (2003) 3336-3337. [8] C. Mateo, J.M. Palomo, G. Fernandez-Lorente, J.M. Guisan, R. Fernandez-Lafuente, Improvement of Enzyme Activity, Stability and Selectivity via Immobilization Techniques, Enzyme Microb. Technol., 40 (2007) 1451-1463. [9] J.M. Palomo, G. Muñoz, G. Fernández-Lorente, C. Mateo, M. Fuentes, J.M. Guisan, R. Fernández-Lafuente, Modulation of Mucor Miehei Lipase Properties via Directed Immobilization on Different Hetero-Functional Epoxy Resins: Hydrolytic Resolution of (R,S)-2-butyroyl-2-phenylacetic Acid, J. Mol. Catal. B: Enzym., 21 (2003) 201-210. [10] J.S. Tan, P.A. Martic, Protein Adsorption and Conformational Change on Small Polymer Particles, J. Colloid Interface Sci., 136 (1990) 415-431. [11] X. Xiao, M.e. Wang, H. Li, P. Si, One-Step Fabrication of Bio- Functionalized Nanoporous Gold/Poly (3, 4-Ethylenedioxythiophene) Hybrid Electrodes for Amperometric Glucose Sensing, Talanta, 115 (2013) 1054-1059. [12] X. Xiao, H. Li, K. Zhang, P. Si, Examining the Effects of Self- Assembled Monolayers on Nanoporous Gold Based Amperometric Glucose Biosensors, Analyst, 139 (2014) 488-494. [13] F. Lisdat, R. Dronov, H. Mohwald, F.W. Scheller, D.G. Kurth, Self- Assembly of Electro-Active Protein Architectures on Electrodes for the Construction of Biomimetic Signal Chains, Chem. Commun., (2009) 274-283. [14] A.P. Girard-Egrot, S. Godoy, L.J. Blum, Enzyme Association with Lipidic Langmuir–Blodgett Films: Interests and Applications in Nanobioscience, Adv. Colloid Interface Sci., 116 (2005) 205-225. [15] M.L.-D. Itamar Willner , Sharon Marx-Tibbon , Eugenii Katz, Bioelectrocatalyzed Amperometric Transduction of Recorded Optical Signals Using Monolayer-Modified Au-Electrodes, J. Am. Chem. Soc., 117 (1995) 6581–6592. [16] X. Xiao, J. Ulstrup, H. Li, J. Zhang, P. Si, Nanoporous Gold Assembly of Glucose Oxidase for Electrochemical Biosensing, Electrochim. Acta, 130 (2014) 559-567. [17] J. J. Gooding, L. Pugliano, D. B. Hibbert, P. Erokhin, Amperometric Biosensor with Enzyme Amplification Fabricated Using Self-Assembled Monolayers of Alkanethiols: The Influence of the Spatial Distribution of the Enzymes, Electrochem. Commun., 2 (2000) 217–221.

- 28 -

Chapter 1

[18] J. J. Gooding, N. Darwish, The Rise of Self-Assembled Monolayers for Fabricating Electrochemical Biosensors--An Interfacial Perspective, Chem. Rec., 12 (2012) 92-105. [19] S. Cantone, V. Ferrario, L. Corici, C. Ebert, D. Fattor, P. Spizzo, L. Gardossi, Efficient Immobilisation of Industrial Biocatalysts: Criteria and Constraints for the Selection of Organic Polymeric Carriers and Immobilisation Methods, Chem. Soc. Rev., 42 (2013) 6262-6276. [20] J.W. Mirtha Umana Protein-Modified Electrodes. The Glucose Oxidase/Polypyrrole System, Anal. Chem., 58 (1986) 2979–2983. [21] T. Ahuja, I. A. Mir, D. Kumar, Rajesh, Biomolecular Immobilization on Conducting Polymers for Biosensing Applications, Biomaterials, 28 (2007) 791-805. [22] S. Braun, S. Rappoport, R. Zusman, D. Avnir, M. Ottolenghi, Biochemically Active Sol-Gel Glasses: the Trapping of Enzymes, Materials Letters, 10 (1990) 1-5. [23] J. Niu, J.Y. Lee, Reagentless Mediated Biosensors Based on Polyelectrolyte and Sol–Gel Derived Silica Matrix, Sens. Actuat. B: Chem., 82 (2002) 250-258. [24] O. Nadzhafova, M. Etienne, A. Walcarius, Direct Electrochemistry of Hemoglobin and Glucose Oxidase in Electrodeposited Sol–Gel Silica Thin Films on Glassy Carbon, Electrochem. Commun., 9 (2007) 1189-1195. [25] O. Barbosa, C. Ortiz, A. Berenguer-Murcia, R. Torres, R.C. Rodrigues, R. Fernandez-Lafuente, Glutaraldehyde in Bio-Catalysts Design: A Useful Crosslinker and a Versatile Tool in Enzyme Immobilization, RSC Adv., 4 (2014) 1583-1600. [26] B.A. Gregg, A. Heller, Redox Polymer Films Containing Enzymes. 1. A Redox-Conducting Epoxy Cement: Synthesis, Characterization, and Electrocatalytic Oxidation of Hydroquinone, J. Phys. Chem., 95 (1991) 5970-5975. [27] A. Heller, Electron-Conducting Redox Hydrogels: Design, Characteristics and Synthesis, Curr. Opin. Chem. Biol., 10 (2006) 664-672. [28] C. Esseghaier, Y. Bergaoui, H. ben Fredj, A. Tlili, S. Helali, S. Ameur, A. Abdelghani, Impedance Spectroscopy on Immobilized Streptavidin Horseradish Peroxidase Layer for Biosensing, Sens. Actuat. B: Chem., 134 (2008) 112-116. [29] B. Bucur, A.F. Danet, J.-L. Marty, Versatile Method of Cholinesterase Immobilisation via Affinity Bonds Using Concanavalin a Applied to the Construction of a Screen-Printed Biosensor, Biosens. Bioelectron., 20 (2004) 217-225. [30] C.M. Halliwell, E. Simon, C.-S. Toh, P.N. Bartlett, A.E.G. Cass, Immobilisation of Lactate Dehydrogenase on Poly(Aniline)–Poly(Acrylate) and Poly(Aniline)–Poly(Vinyl Sulphonate) Films for Use in A Lactate Biosensor, Anal. Chim. Acta, 453 (2002) 191-200. [31] R.A. Marcus, Electron Transfer Reactions in Chemistry: Theory and Experiment (Nobel Lecture), Angew. Chem. Int. Ed., 32 (1993) 1111–1121. [32] S.L. Mayo, W.R. Ellis, R.J. Crutchley, H.B. Gray, Long-Range Electron Transfer in Heme Proteins, Science, 233 (1986) 948-952.

- 29 -

Chapter 1

[33] C.C. Page, C.C. Moser, X. Chen, P.L. Dutton, Natural Engineering Principles of Electron Tunnelling in Biological Oxidation-Reduction, Nature, 402 (1999) 47-52. [34] J. Wang, Electrochemical Glucose Biosensors, Chem. Rev., 108 (2008) 814–825. [35] A. Carreau, B.E. Hafny‐Rahbi, A. Matejuk, C. Grillon, C. Kieda, Why is the Partial Oxygen Rressure of Human Tissues a Crucial Parameter? Small Molecules and Hypoxia, J. Cell. Mol. Med., 15 (2011) 1239-1253. [36] S. Shleev, Quo Vadis, Implanted Fuel Cell?, ChemPlusChem, (2017) DOI: 10.1002/cplu.201600536. [37] L.C. Clark, C. Lyons, Electrode Systems for Continuous Monitoring in Cardiovascular Surgery, Ann. Ny. Acad. Sci., 102 (1962) 29-45. [38] A. Heller, Electrical Connection of Enzyme Redox Centers to Electrodes, J. Phys. Chem., 96 (1992) 3579-3587. [39] A. Heller, B. Feldman, Electrochemical Glucose Sensors and Their Applications in Diabetes Management, Chem. Rev., 108 (2008) 2482-2505. [40] V. Soukharev, N. Mano, A. Heller, A Four-Electron O2-Electroreduction Biocatalyst Superior to Platinum and a Biofuel Cell Operating at 0.88 V, J. Am. Chem. Soc., 126 (2004) 8368-8369. [41] S. Aquino Neto, D.P. Hickey, R.D. Milton, A.R. De Andrade, S.D. Minteer, High Current Density PQQ-Dependent Alcohol and Aldehyde Dehydrogenase Bioanodes, Biosens. Bioelectron., 72 (2015) 247-254. [42] K.L. Knoche, D.P. Hickey, R.D. Milton, C.L. Curchoe, S.D. Minteer, Hybrid Glucose/O2 Biobattery and Supercapacitor Utilizing a Pseudocapacitive Dimethylferrocene Redox Polymer at the Bioanode, ACS Energy Lett., (2016) 380- 385. [43] D.P. Hickey, A.J. Halmes, D.W. Schmidtke, D.T. Glatzhofer, Electrochemical Characterization of Glucose Bioanodes Based on Tetramethylferrocene-Modified Linear Poly(ethylenimine), Electrochim. Acta, 149 (2014) 252-257. [44] C. Tapia, R.D. Milton, G. Pankratova, S.D. Minteer, H.-E. Åkerlund, D. Leech, A.L. De Lacey, M. Pita, L. Gorton, Wiring of Photosystem I and Hydrogenase on an Electrode for Photoelectrochemical H2 Production by using Redox Polymers for Relatively Positive Onset Potential, ChemElectroChem, 4 (2017) 90-95. [45] N. Plumeré, O. Rüdiger, A.A. Oughli, R. Williams, J. Vivekananthan, S. Pöller, W. Schuhmann, W. Lubitz, A Redox Hydrogel Protects Hydrogenase from High-Potential Deactivation and Oxygen Damage, Nat. Chem., 6 (2014) 822. [46] P. Pinyou, A. Ruff, S. Pöller, S. Alsaoub, S. Leimkühler, U. Wollenberger, W. Schuhmann, Wiring of the Aldehyde Oxidoreductase PaoABC to Electrode Surfaces via Entrapment in Low Potential Phenothiazine-Modified Redox Polymers, Bioelectrochem., 109 (2016) 24-30. [47] A. Ruff, P. Pinyou, M. Nolten, F. Conzuelo, W. Schuhmann, A Self- Powered Ethanol Biosensor, ChemElectroChem, 4 (2017) 890-897.

- 30 -

Chapter 1

[48] S. Abdellaoui, R.D. Milton, T. Quah, S.D. Minteer, NAD-dependent dehydrogenase bioelectrocatalysis: the ability of a naphthoquinone redox polymer to regenerate NAD, Chem. Commun., 52 (2016) 1147-1150. [49] R.D. Milton, D.P. Hickey, S. Abdellaoui, K. Lim, F. Wu, B. Tan, S.D. Minteer, Rational Design of Quinones for High Power Density Biofuel Cells, Chem. Sci., 6 (2015) 4867-4875. [50] D.A. Guschin, J. Castillo, N. Dimcheva, W. Schuhmann, Redox Electrodeposition Polymers: Adaptation of the Redox Potential of Polymer-Bound Os Complexes for Bioanalytical Applications, Anal. Bioanal. Chem., 398 (2010) 1661- 1673. [51] M.J. Eddowes, H.A.O. Hill, Novel Method for the Investigation of the Electrochemistry of Metalloproteins: Cytochrome c, J. Chem. Soc., Chem. Commun., (1977) 771b-772. [52] P. Yeh, T. Kuwana, Reversible Electrode Reaction of Cytochrome C, Chem. Lett., 6 (1977) 1145-1148. [53] S.R. Betso, M.H. Klapper, L.B. Anderson, Electrochemical Studies of Heme Proteins. Coulometric, Polarographic, and Combined Spectroelectrochemical Methods for Reduction of the Heme Prosthetic Group in Cytochrome c, J. Am. Chem. Soc., 94 (1972) 8197-8204. [54] M. Dagys, K. Haberska, S. Shleev, T. Arnebrant, J. Kulys, T. Ruzgas, Laccase–Gold Nanoparticle Assisted Bioelectrocatalytic Reduction of Oxygen, Electrochem. Commun., 12 (2010) 933-935. [55] K. Murata, K. Kajiya, N. Nakamura, H. Ohno, Direct Electrochemistry of Bilirubin Oxidase on Three-Dimensional Gold Nanoparticle Electrodes and Its Application in a Biofuel Cell, Energy Environ. Sci., 2 (2009) 1280-1285. [56] D. Ivnitski, B. Branch, P. Atanassov, C. Apblett, Glucose Oxidase Anode for Biofuel Cell Based on Direct Electron Transfer, Electrochem. Commun., 8 (2006) 1204-1210. [57] K. Murata, M. Suzuki, K. Kajiya, N. Nakamura, H. Ohno, High Performance Bioanode Based on Direct Electron Transfer of Fructose Dehydrogenase at Gold Nanoparticle-Modified Electrodes, Electrochem. Commun., 11 (2009) 668- 671. [58] F. Tasca, W. Harreither, R. Ludwig, J.J. Gooding, L. Gorton, Cellobiose Dehydrogenase Aryl Diazonium Modified Single Walled Carbon Nanotubes: Enhanced Direct Electron Transfer Through A Positively Charged Surface, Anal. Chem., 83 (2011) 3042-3049. [59] M. Falk, Z. Blum, S. Shleev, Direct Electron Transfer Based Enzymatic Fuel Cells, Electrochim. Acta, 82 (2012) 191-202. [60] R.D. Milton, S.D. Minteer, Direct Enzymatic Bioelectrocatalysis: Differentiating between Myth and Reality, J. R. Soc. Interface, 14 (2017). [61] M. Ameyama, E. Shinagawa, K. Matsushita, O. Adachi, D-fructose Dehydrogenase of Gluconobacter Industrius: Purification, Characterization, and Application to Enzymatic Microdetermination of D-Fructose, J. Bacteriol., 145 (1981) 814-823. - 31 -

Chapter 1

[62] H.-q. Xia, Y. Hibino, Y. Kitazumi, O. Shirai, K. Kano, Interaction between D-Fructose Dehydrogenase and Methoxy-Substituent-Functionalized Carbon Surface to Increase Productive Orientations, Electrochim. Acta, 218 (2016) 41-46. [63] M. Kizling, R. Bilewicz, Fructose Dehydrogenase Electron Transfer Pathway in Bioelectrocatalytic Reactions, ChemElectroChem, (2017) DOI: 10.1002/celc.201700861. [64] Y. Yamashita, S. Ferri, M.L. Huynh, H. Shimizu, H. Yamaoka, K. Sode, Direct Electron Transfer Type Disposable Sensor Strip for Glucose Sensing Employing an Engineered FAD Glucose Dehydrogenase, Enzyme and Microb. Technol., 52 (2013) 123-128. [65] M. Shiota, T. Yamazaki, K. Yoshimatsu, K. Kojima, W. Tsugawa, S. Ferri, K. Sode, An Fe–S Cluster in the Conserved Cys-Rich Region in the Catalytic Subunit of FAD-Dependent Dehydrogenase Complexes, Bioelectrochem., 112 (2016) 178-183. [66] S. Babanova, I. Matanovic, M.S. Chavez, P. Atanassov, Role of Quinones in Electron Transfer of PQQ–Glucose Dehydrogenase Anodes—Mediation or Orientation Effect, J. Am. Chem. Soc., 137 (2015) 7754-7762. [67] J. Okuda, K. Sode, PQQ Glucose Dehydrogenase with Novel Electron Transfer Ability, Biochem. Biophys. Res. Commun., 314 (2004) 793-797. [68] D. Ivnitski, P. Atanassov, C. Apblett, Direct Bioelectrocatalysis of PQQ- Dependent Glucose Dehydrogenase, Electroanalysis, 19 (2007) 1562-1568. [69] M. Torimura, K. Kano, T. Ikeda, T. Ueda, Spectroelectrochemical Characterization of Quinohemoprotein Alcohol Dehydrogenase from Gluconobacter suboxydans, Chem. Lett., 26 (1997) 525-526. [70] B.L. Treu, S.D. Minteer, Isolation and Purification of PQQ-Dependent Lactate Dehydrogenase from Gluconobacter and Use for Direct Electron Transfer at Carbon and Gold Electrodes, Bioelectrochem., 74 (2008) 73-77. [71] V. Laurinavicius, J. Razumiene, A. Ramanavicius, A.D. Ryabov, Wiring of PQQ–Dehydrogenases, Biosens. Bioelectron., 20 (2004) 1217-1222. [72] I. Mazurenko, X. Wang, A. de Poulpiquet, E. Lojou, H2/O2 Enzymatic Fuel Cells: from Proof-of-Concept to Powerful Devices, Sustainable Energy Fuels, 1 (2017) 1475-1501. [73] W. Lubitz, H. Ogata, O. Rüdiger, E. Reijerse, Hydrogenases, Chem. Rev., 114 (2014) 4081-4148. [74] É. Lojou, X. Luo, M. Brugna, N. Candoni, S. Dementin, M.T. Giudici- Orticoni, Biocatalysts for Fuel Cells: Efficient Hydrogenase Orientation for H2 Oxidation at Electrodes Modified with Carbon Nanotubes, J. Biol. Inorg. Chem., 13 (2008) 1157-1167. [75] H.M. Kalisz, H.-J. Hecht, D. Schomburg, R.D. Schmid, Crystallization and Preliminary X-ray Diffraction Studies of a Deglycosylated Glucose Oxidase from Aspergillus niger, J. Mol. Biol., 213 (1990) 207-209. [76] O. Courjean, F. Gao, N. Mano, Deglycosylation of Glucose Oxidase for Direct and Efficient Glucose Electrooxidation on a Glassy Carbon Electrode, Angew. Chem. Int. Edit., 48 (2009) 5897-5899. - 32 -

Chapter 1

[77] M.E. Yakovleva, A. Killyéni, R. Ortiz, C. Schulz, D. MacAodha, P.Ó. Conghaile, D. Leech, I.C. Popescu, C. Gonaus, C.K. Peterbauer, L. Gorton, Recombinant Pyranose Dehydrogenase—A Versatile Enzyme Possessing both Mediated and Direct Electron Transfer, Electrochem. Commun., 24 (2012) 120-122. [78] S.B. Bankar, M.V. Bule, R.S. Singhal, L. Ananthanarayan, Glucose Oxidase - An Overview, Biotechnol. Adv., 27 (2009) 489-501. [79] H.J. Hecht, D. Schomburg, H. Kalisz, R.D. Schmid, The 3D Structure of Glucose Oxidase from Aspergillus Niger. Implications for the Use of God as a Biosensor Enzyme, Biosens. Bioelectron., 8 (1993) 197-203. [80] B.E.P. Swoboda, The Relationship between Molecular Conformation and the Binding of Flavin-Adenine Dinucleotide in Glucose Oxidase, Biochim. Biophys. Acta, 175 (1969) 365-379. [81] A.T. Yahiro, S.M. Lee, D.O. Kimble, Bioelectrochemistry: I. Enzyme Utilizing Bio-Fuel Cell Studies, Biochim. Biophys. Acta, 88 (1964) 375-383. [82] G. Liu, M.N. Paddon-Row, J. Justin Gooding, A Molecular Wire Modified Glassy Carbon Electrode for Achieving Direct Electron Transfer to Native Glucose Oxidase, Electrochem. Commun., 9 (2007) 2218-2223. [83] C. Cai, J. Chen, Direct Electron Transfer of Glucose Oxidase Promoted by Carbon Nanotubes, Anal. Biochem., 332 (2004) 75-83. [84] P. Wu, Q. Shao, Y. Hu, J. Jin, Y. Yin, H. Zhang, C. Cai, Direct Electrochemistry of Glucose Oxidase Assembled on Graphene and Application to Glucose Detection, Electrochim. Acta, 55 (2010) 8606-8614. [85] M. Wooten, S. Karra, M. Zhang, W. Gorski, On the Direct Electron Transfer, Sensing, and Enzyme Activity in the Glucose Oxidase/Carbon Nanotubes System, Anal. Chem., 86 (2014) 752-757. [86] B. Liang, X. Guo, L. Fang, Y. Hu, G. Yang, Q. Zhu, J. Wei, X. Ye, Study of Direct Electron Transfer and Enzyme Activity of Glucose Oxidase on Graphene Surface, Electrochem. Commun., 50 (2015) 1-5. [87] R. Ludwig, R. Ortiz, C. Schulz, W. Harreither, C. Sygmund, L. Gorton, Cellobiose Dehydrogenase Modified Electrodes: Advances by Materials Science and Biochemical Engineering, Anal. Bioanal. Chem., 405 (2013) 3637-3658. [88] R. Ludwig, W. Harreither, F. Tasca, L. Gorton, Cellobiose Dehydrogenase: A Versatile Catalyst for Electrochemical Applications, ChemPhysChem, 11 (2010) 2674-2697. [89] T. Larsson, M. Elmgren, S.-E. Lindquist, M. Tessema, L. Gorton, G. Henriksson, Electron Transfer between Cellobiose Dehydrogenase and Graphite Electrodes, Anal. Chim. Acta, 331 (1996) 207-215. [90] E. Malel, R. Ludwig, L. Gorton, D. Mandler, Localized Deposition of Au Nanoparticles by Direct Electron Transfer through Cellobiose Dehydrogenase, Chem.–Eur. J., 16 (2010) 11697-11706. [91] N. Mano, Features and Applications of Bilirubin Oxidases, Appl. Microbiol. Biotechnol., 96 (2012) 301-307. [92] N. Mano, L. Edembe, Bilirubin Oxidases in Bioelectrochemistry: Features and Recent Findings, Biosens. Bioelectron., 50 (2013) 478-485. - 33 -

Chapter 1

[93] T. Sakurai, K. Kataoka, Basic and Applied Features of Multicopper Oxidases, CueO, Bilirubin Oxidase, and Laccase, Chem. Rec., 7 (2007) 220-229. [94] S. Tsujimura, H. Tatsumi, J. Ogawa, S. Shimizu, K. Kano, T. Ikeda, Bioelectrocatalytic Reduction of Dioxygen to Water at Neutral pH Using Bilirubin Oxidase as an Enzyme and 2,2′-Azinobis (3-ethylbenzothiazolin-6-sulfonate) as an Electron Transfer Mediator, J. Electroanal. Chem., 496 (2001) 69-75. [95] X. Wang, M. Falk, R. Ortiz, H. Matsumura, J. Bobacka, R. Ludwig, M. Bergelin, L. Gorton, S. Shleev, Mediatorless Sugar/Oxygen Enzymatic Fuel Cells Based on Gold Nanoparticle-Modified Electrodes, Biosens. Bioelectron., 31 (2012) 219-225. [96] F.A. Armstrong, H.A.O. Hill, N.J. Walton, Reactions of Electron- Transfer Proteins at Electrodes, Q. Rev. Biophys., 18 (1985) 261-322. [97] J.K. Cullison, F.M. Hawkridge, N. Nakashima, S. Yoshikawa, A Study of Cytochrome c Oxidase in Lipid Bilayer Membranes on Electrode Surfaces, Langmuir, 10 (1994) 877-882. [98] M. Hüttemann, P. Pecina, M. Rainbolt, T.H. Sanderson, V.E. Kagan, L. Samavati, J.W. Doan, I. Lee, The Multiple Functions of Cytochrome C and Their Regulation in Life and Death Decisions of the Mammalian Cell: From Respiration to Apoptosis, Mitochondrion, 11 (2011) 369-381. [99] T.M. Nahir, R.A. Clark, E.F. Bowden, Linear-Sweep Voltammetry of Irreversible Electron Transfer in Surface-Confined Species Using the Marcus Theory, Anal. Chem., 66 (1994) 2595-2598. [100] X. Chen, R. Ferrigno, J. Yang, G.M. Whitesides, Redox Properties of Cytochrome c Adsorbed on Self-Assembled Monolayers: A Probe for Protein Conformation and Orientation, Langmuir, 18 (2002) 7009-7015. [101] I. Bertini, G. Cavallaro, A. Rosato, Cytochrome C: Occurrence and Functions, Chem. Rev., 106 (2006) 90-115. [102] A. Serleti, U. Salaj-Kosla, E. Magner, The Spatial and Sequential Immobilisation Of Cytochrome C at Adjacent Electrodes, Chem. Commun., 49 (2013) 8395-8397. [103] K.L. Davis, B.J. Drews, H. Yue, D.H. Waldeck, K. Knorr, R.A. Clark, Electron-Transfer Kinetics of Covalently Attached Cytochrome c/SAM/Au Electrode Assemblies, J. Phys. Chem. C, 112 (2008) 6571-6576. [104] D. Moss, E. Nabedryk, J. Breton, W. MÄNtele, Redox-Linked Conformational Changes in Proteins Detected by a Combination of Infrared Spectroscopy and Protein Electrochemistry, Eur. J. Biochem., 187 (1990) 565-572. [105] J.F. Cerda, C.X. Guzman, H. Zhang, E.J. Amendola, J.D. Castorino, N. Millet, A.L. Fritz, D.N. Houchins, M.H. Roeder, Spectroelectrochemical Measurements of Redox Proteins by Using a Simple Uv/Visible Cell, Electrochem. Commun., 33 (2013) 76-79. [106] C.R. Lowe, Biosensors, Trends Biotechnol., 2 (1984) 59-65. [107] D. Grieshaber, R. MacKenzie, J. Vörös, E. Reimhult, Electrochemical Biosensors-Sensor Principles and Architectures, Sensors, 8 (2008) 1400-1458.

- 34 -

Chapter 1

[108] D. R. Thévenota, K. Tothb, R. A. Durstc, G. S. Wilsond, Electrochemical Biosensors: Recommended Definitions and Classification, Biosens. Bioelectron., 16 (2001) 121–131. [109] K. E. Toghill, R. G. Compton, Electrochemical Non-Enzymatic Glucose Sensors: a Perspective and an Evaluation, Int. J. Electrochem. Sci., 5 (2010) 1246-1301. [110] S. Rubenwolf, S. Kerzenmacher, R. Zengerle, F. von Stetten, Strategies to Extend the Lifetime of Bioelectrochemical Enzyme Electrodes for Biosensing and Biofuel Cell Applications, Appl. Microbiol. Biotechnol., 89 (2011) 1315-1322. [111] A.P. Turner, Biosensors: Sense and Sensibility, Chem. Soc. Rev., 42 (2013) 3184-3196. [112] G.A.M. Mersal, M. Khodari, U. Bilitewski, Optimisation of the Composition of a Screen-Printed Acrylate Polymer Enzyme Layer with Respect to an Improved Selectivity and Stability of Enzyme Electrodes, Biosens. Bioelectron., 20 (2004) 305-314. [113] P.U. Abel, T. von Woedtke, Biosensors for In Vivo Glucose Measurement: Can We Cross The Experimental Stage, Biosens. Bioelectron., 17 (2002) 1059-1070. [114] K. Mitsubayashi, Y. Wakabayashi, S. Tanimoto, D. Murotomi, T. Endo, Optical-Transparent and Flexible Glucose Sensor with ITO Electrode, Biosens. Bioelectron., 19 (2003) 67-71. [115] T. Matsumoto, A. Ohashi, N. Ito, H. Fujiwara, T. Matsumoto, A Long- Term Lifetime Amperometric Glucose Sensor with a Perfluorocarbon Polymer Coating, Biosens. Bioelectron., 16 (2001) 271-276. [116] K. Mitsubayashi, M. Suzuki, E. Tamiya, I. Karube, Analysis of Metabolites in Sweat as A Measure of Physical Condition, Anal. Chim. Acta, 289 (1994) 27-34. [117] R.C.H. Kwan, H.F. Leung, P.Y.T. Hon, H.C.F. Cheung, K. Hirota, R. Renneberg, Amperometric Biosensor for Determining Human Salivary Phosphate, Anal. Biochem., 343 (2005) 263-267. [118] S. Shleev, A. Bergel, L. Gorton, Biological Fuel Cells: Divergence of Opinion, Bioelectrochem., 106, Part A (2015) 1-2. [119] M. Cooney, V. Svoboda, C. Lau, G. Martin, S. Minteer, Enzyme catalysed biofuel cells, Energy Environ. Sci., 1 (2008) 320-337. [120] R.A.S. Luz, A.R. Pereira, J.C.P. de Souza, F.C.P.F. Sales, F.N. Crespilho, Enzyme Biofuel Cells: Thermodynamics, Kinetics and Challenges in Applicability, ChemElectroChem, 1 (2014) 1751-1777. [121] M.J. Moehlenbrock, R.L. Arechederra, K.H. Sjöholm, S.D. Minteer, Analytical Techniques for Characterizing Enzymatic Biofuel Cells, Anal. Chem., 81 (2009) 9538-9545. [122] N. Mano, A. de Poulpiquet, O2 Reduction in Enzymatic Biofuel Cells, Chem. Rev., (2017).

- 35 -

Chapter 1

[123] J.A. Cracknell, K.A. Vincent, F.A. Armstrong, Enzymes as Working or Inspirational Electrocatalysts for Fuel Cells and Electrolysis, Chem. Rev., 108 (2008) 2439-2461. [124] J. Masa, W. Schuhmann, Electrocatalysis and Bioelectrocatalysis – Distinction without a Difference, Nano Energy, 29 (2016) 466-475. [125] A. Shukla, P. Suresh, S. Berchmans, A. Rajendran, Biological Fuel Cells and Their Applications, Curr. Sci., 87 (2004) 455-468. [126] R.A. Luz, A.R. Pereira, J.C. de Souza, F.C. Sales, F.N. Crespilho, Enzyme biofuel cells: thermodynamics, kinetics and challenges in applicability, ChemElectroChem, 1 (2014) 1751-1777. [127] M. Gamella, A. Koushanpour, E. Katz, Biofuel Cells – Activation of Micro- and Macro-Electronic Devices, Bioelectrochem., 119 (2018) 33-42. [128] S. Shleev, A. Bergel, L. Gorton, Biological fuel cells: Divergence of opinion, Bioelectrochemistry, 106 (2015) 1-2. [129] W. Winkler, P. Nehter, Thermodynamics of Fuel Cells, in: R. Bove, S. Ubertini (Eds.) Modeling Solid Oxide Fuel Cells: Methods, Procedures and Techniques, Springer Netherlands, Dordrecht, 2008, pp. 13-50. [130] R.A. Alberty, Calculating Apparent Equilibrium Constants of Enzyme‐ Catalyzed Reactions at pH 7, Biochem. Biophys. Res. Commun., 28 (2000) 12-17. [131] D. Pankratov, E. González‐Arribas, Z. Blum, S. Shleev, Tear based bioelectronics, Electroanalysis, 28 (2016) 1250-1266. [132] X. Xiao, E. Magner, A Biofuel Cell in Non-Aqueous Solution, Chem. Commun., 51 (2015) 13478-13480. [133] A. Suzuki, N. Mano, S. Tsujimura, Lowering the Potential of Electroenzymatic Glucose Oxidation on Redox Hydrogel-Modified Porous Carbon Electrode, Electrochim. Acta, 232 (2017) 581-585. [134] M. Rasmussen, S. Abdellaoui, S.D. Minteer, Enzymatic biofuel cells: 30 years of critical advancements, Biosens. Bioelectron., 76 (2015) 91-102. [135] S. Shleev, J. Tkac, A. Christenson, T. Ruzgas, A.I. Yaropolov, J.W. Whittaker, L. Gorton, Direct Electron Transfer Between Copper-Containing Proteins and Electrodes, Biosens. Bioelectron., 20 (2005) 2517-2554. [136] M. Falk, Z. Blum, S. Shleev, Direct electron transfer based enzymatic fuel cells, Electrochim. Acta, 82 (2012) 191-202. [137] D.J. Richardson, Bacterial respiration: a flexible process for a changing environment, Microbiology, 146 (2000) 551-571. [138] K. Rabaey, W. Verstraete, Microbial fuel cells: novel biotechnology for energy generation, Trends in Biotechnology, 23 (2005) 291-298. [139] E. Katz, D.D. Schlereth, H.-L. Schmidt, Electrochemical Study of Pyrroloquinoline Quinone Covalently Immobilized as a Monolayer onto a Cystamine- Modified Gold Electrode, J. Electroanal. Chem., 367 (1994) 59-70. [140] G.T.R. Palmore, H. Bertschy, S.H. Bergens, G.M. Whitesides, A Methanol/Dioxygen Diofuel Cell that Uses NAD+-Dependent Dehydrogenases as Catalysts: Application of an Electro-Enzymatic Method to Regenerate Nicotinamide

- 36 -

Chapter 1

Adenine Dinucleotide at Low Overpotentials, J. Electroanal. Chem., 443 (1998) 155- 161. [141] G. Gupta, C. Lau, V. Rajendran, F. Colon, B. Branch, D. Ivnitski, P. Atanassov, Direct Electron Transfer Catalyzed by Bilirubin Oxidase for Air Breathing Gas-Diffusion Electrodes, Electrochem. Commun., 13 (2011) 247-249. [142] M. Falk, C.W. Narváez Villarrubia, S. Babanova, P. Atanassov, S. Shleev, Biofuel Cells for Biomedical Applications: Colonizing the Animal Kingdom, ChemPhysChem, 14 (2013) 2045-2058. [143] M. Falk, V. Andoralov, Z. Blum, J. Sotres, D.B. Suyatin, T. Ruzgas, T. Arnebrant, S. Shleev, Biofuel Cell as a Power Source for Electronic Contact Lenses, Biosens. Bioelectron., 37 (2012) 38-45. [144] T. Miyake, S. Yoshino, T. Yamada, K. Hata, M. Nishizawa, Self- Regulating Enzyme−Nanotube Ensemble Films and Their Application as Flexible Electrodes for Biofuel Cells, J. Am. Chem. Soc., 133 (2011) 5129-5134. [145] T. Beneyton, I.P.M. Wijaya, C.B. Salem, A.D. Griffiths, V. Taly, Membraneless Glucose/O2 Microfluidic Biofuel Cells Using Covalently Bound Enzymes, Chem. Commun., 49 (2013) 1094-1096. [146] P. Pinyou, F. Conzuelo, K. Sliozberg, J. Vivekananthan, A. Contin, S. Pöller, N. Plumeré, W. Schuhmann, Coupling of an Enzymatic Biofuel Cell to an Electrochemical Cell for Self-Powered Glucose Sensing with Optical Readout, Bioelectrochem., 106, Part A (2015) 22-27. [147] M. Falk, M. Alcalde, P.N. Bartlett, A.L. De Lacey, L. Gorton, C. Gutierrez-Sanchez, R. Haddad, J. Kilburn, D. Leech, R. Ludwig, E. Magner, D.M. Mate, P.Ó. Conghaile, R. Ortiz, M. Pita, S. Pöller, T. Ruzgas, U. Salaj-Kosla, W. Schuhmann, F. Sebelius, M. Shao, L. Stoica, C. Sygmund, J. Tilly, M.D. Toscano, J. Vivekananthan, E. Wright, S. Shleev, Self-Powered Wireless Carbohydrate/Oxygen Sensitive Biodevice Based on Radio Signal Transmission, PLoS ONE, 9 (2014) e109104. [148] C. Agnes, M. Holzinger, A. Le Goff, B. Reuillard, K. Elouarzaki, S. Tingry, S. Cosnier, Supercapacitor/Biofuel Cell Hybrids Based on Wired Enzymes on Carbon Nanotube Matrices: Autonomous Reloading after High Power Pulses in Neutral Buffered Glucose Solutions, Energy Environ. Sci., 7 (2014) 1884-1888. [149] J. Wang, Nanomaterial-Based Electrochemical Biosensors, Analyst, 130 (2005) 421-426. [150] V. Scognamiglio, Nanotechnology in Glucose Monitoring: Advances and Challenges in the Last 10 Years, Biosens. Bioelectron., 47 (2013) 12-25. [151] S.A. Ansari, Q. Husain, Potential Applications of Enzymes Immobilized on/in Nano Materials: A Review, Biotechnol. Adv., 30 (2012) 512-523. [152] U. Yogeswaran, S.-M. Chen, A Review on the Electrochemical Sensors and Biosensors Composed of Nanowires as Sensing Material, Sensors, 8 (2008) 290-313. [153] J.J. Gooding, Nanostructuring Electrodes with Carbon Nanotubes: A Review on Electrochemistry and Applications for Sensing, Electrochim. Acta, 50 (2005) 3049-3060. - 37 -

Chapter 1

[154] Z.-G. Wang, L.-S. Wan, Z.-M. Liu, X.-J. Huang, Z.-K. Xu, Enzyme Immobilization on Electrospun Polymer Nanofibers: An Overview, J. Mol. Catal. B: Enzym., 56 (2009) 189-195. [155] Y. Shao, J. Wang, H. Wu, J. Liu, I.A. Aksay, Y. Lin, Graphene Based Electrochemical Sensors and Biosensors: A Review, Electroanalysis, 22 (2010) 1027- 1036. [156] A. Nel, T. Xia, L. Mädler, N. Li, Toxic Potential of Materials at the Nanolevel, Science, 311 (2006) 622-627. [157] M.D. Scanlon, U. Salaj-Kosla, S. Belochapkine, D. MacAodha, D. Leech, Y. Ding, E. Magner, Characterization of Nanoporous Gold Electrodes for Bioelectrochemical Applications, Langmuir, 28 (2012) 2251-2261. [158] A. Walcarius, Electrocatalysis, Sensors and Biosensors in Analytical Chemistry Based on Ordered Mesoporous and Macroporous Carbon-Modified Electrodes, TrAC, Trends Anal. Chem., 38 (2012) 79-97. [159] J. Zhang, C.M. Li, Nanoporous Metals: Fabrication Strategies and Advanced Electrochemical Applications in Catalysis, Sensing and Energy Systems, Chem. Soc. Rev., 41 (2012) 7016-7031. [160] Z. Dai, H. Ju, Bioanalysis Based on Nanoporous Materials, TrAC, Trends Anal. Chem., 39 (2012) 149-162. [161] M.D. Scanlon, U. Salaj-Kosla, S. Belochapkine, D. MacAodha, D. Leech, Y. Ding, E. Magner, Characterization of Nanoporous Gold Electrodes for Bioelectrochemical Applications, Langmuir, 28 (2011) 2251-2261. [162] Y. Ding, Y.J. Kim, J. Erlebacher, Nanoporous Gold Leaf: “Ancient Technology”/Advanced Material, Adv. Mater., 16 (2004) 1897–1900. [163] X. Xiao, P. Si, E. Magner, An Overview of Dealloyed Nanoporous Gold in Bioelectrochemistry, Bioelectrochem., 109 (2016) 117-126. [164] C. Hou, D. Yang, B. Liang, A. Liu, Enhanced Performance of a Glucose/O2 Biofuel Cell Assembled with Laccase-Covalently Immobilized Three- Dimensional Macroporous Gold Film-Based Biocathode and Bacterial Surface Displayed Glucose Dehydrogenase-Based Bioanode, Anal. Chem., 86 (2014) 6057- 6063. [165] H. Matsumura, R. Ortiz, R. Ludwig, K. Igarashi, M. Samejima, L. Gorton, Direct Electrochemistry of Phanerochaete chrysosporium Cellobiose Dehydrogenase Covalently Attached onto Gold Nanoparticle Modified Solid Gold Electrodes, Langmuir, 28 (2012) 10925-10933. [166] K. Nishio, H. Masuda, Anodization of Gold in Oxalate Solution to Form a Nanoporous Black Film, Angew. Chem. Int. Edit., 50 (2011) 1603-1607. [167] R. Szamocki, S. Reculusa, S. Ravaine, P. N. Bartlett, A. Kuhn, R. Hempelmann, Tailored Mesostructuring and Biofunctionalization of Gold for Increased Electroactivity, Angew. Chem. Int. Edit., 45 (2006) 1317–1321. [168] A. Karajić, S. Reculusa, M. Heim, P. Garrigue, S. Ravaine, N. Mano, A. Kuhn, Bottom-up Generation of Miniaturized Coaxial Double Electrodes with Tunable Porosity, Adv. Mater. Interfaces, 2 (2015) 1500192-n/a.

- 38 -

Chapter 1

[169] S. Boland, D. Leech, A Glucose/Oxygen Enzymatic Fuel Cell based on Redox Polymer and Enzyme Immobilisation at Highly-Ordered Macroporous Gold Electrodes, Analyst, 137 (2012) 113-117. [170] H. du Toit, M. Di Lorenzo, Continuous Power Generation From Glucose with Two Different Miniature Flow-Through Enzymatic Biofuel Cells, Biosens. Bioelectron., 69 (2015) 199-205. [171] A. Ciaccafava, P. Infossi, M. Ilbert, M. Guiral, S. Lecomte, M.T. Giudici-Orticoni, E. Lojou, Electrochemistry, AFM, and PM-IRRA Spectroscopy of Immobilized Hydrogenase: Role of a Hydrophobic Helix in Enzyme Orientation for Efficient H2 Oxidation, Angew. Chem. Int. Edit., 51 (2012) 953-956. [172] M. Pita, C. Gutierrez-Sanchez, M.D. Toscano, S. Shleev, A.L. De Lacey, Oxygen Biosensor based on Bilirubin Oxidase Immobilized on a Nanostructured Gold Electrode, Bioelectrochem., 94 (2013) 69-74. [173] T. Siepenkoetter, U. Salaj-Kosla, X. Xiao, P.Ó. Conghaile, M. Pita, R. Ludwig, E. Magner, Immobilization of Redox Enzymes on Nanoporous Gold Electrodes: Applications in Biofuel Cells, ChemPlusChem, 82 (2017) 553-560. [174] T. Siepenkoetter, U. Salaj-Kosla, E. Magner, The Immobilization of Fructose Dehydrogenase on Nanoporous Gold Electrodes for the Detection of Fructose, ChemElectroChem, 4 (2017) 905-912. [175] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, A Symmetric Supercapacitor/Biofuel Cell Hybrid Device Based on Enzyme-Modified Nanoporous Gold: An Autonomous Pulse Generator, Biosens. Bioelectron., 90 (2017) 96-102. [176] Z. Gao, G. Binyamin, H.H. Kim, S.C. Barton, Y. Zhang, A. Heller, Electrodeposition of Redox Polymers and Co‐Electrodeposition of Enzymes by Coordinative Crosslinking, Angew. Chem. Int. Ed., 41 (2002) 810-813. [177] A.J. Forty, Corrosion Micromorphology of Noble Metal Alloys and Depletion Gilding, Nature, 282 (1979) 597-598. [178] H.W. Pickering, C. Wagner, Electrolytic Dissolution of Binary Alloys Containing a Noble Metal, J. Electrochem. Soc., 114 (1967) 698-706. [179] H.W. Pickering, Characteristic Features of Alloy Polarization Curves, Corros. Sci., 23 (1983) 1107-1120. [180] J. Erlebacher, M.J. Aziz, A. Karma, N. Dimitrov, K. Sieradzki, Evolution of Nanoporosity in Dealloying, Nature, 410 (2001) 450-453. [181] J. Erlebacher, An Atomistic Description of Dealloying Porosity Evolution, the Critical Potential, and Rate-Limiting Behavior, J. Electrochem. Soc., 151 (2004) C614-C626. [182] L. Qian, X. Yan, T. Fujita, A. Inoue, M. Chen, Surface Enhanced Raman Scattering of Nanoporous Gold: Smaller Pore Sizes Stronger Enhancements, Appl. Phys. Lett., 90 (2007) 153120-153120. [183] M.M. Collinson, Nanoporous Gold Electrodes and Their Applications in Analytical Chemistry, ISRN Anal. Chem., (2013) 1-21. [184] Y. Ding, M. Chen, Nanoporous Metals for Catalytic and Optical Applications, MRS Bull., 34 (2009) 569-576. - 39 -

Chapter 1

[185] F. Yu, S. Ahl, A.-M. Caminade, J.-P. Majoral, W. Knoll, J. Erlebacher, Simultaneous Excitation of Propagating and Localized Surface Plasmon Resonance in Nanoporous Gold Membranes, Anal. Chem., 78 (2006) 7346-7350. [186] M.C. Dixon, T.A. Daniel, M. Hieda, D.M. Smilgies, M.H.W. Chan, D.L. Allara, Preparation, Structure, and Optical Properties of Nanoporous Gold Thin Films, Langmuir, 23 (2007) 2414-2422. [187] X. Zhang, Y. Ding, Unsupported Nanoporous Gold for Heterogeneous Catalysis, Catal. Sci. Technol., 3 (2013) 2862-2868. [188] A.W. Arne Wittstock, Marcus Bäumer, Nanoporous Gold as a Platform for a Building Block Catalyst, ACS Catalysis, 2 (2012) 2199-2215. [189] C. Xu, J. Su, X. Xu, P. Liu, H. Zhao, F. Tian, Y. Ding, Low Temperature CO Oxidation over Unsupported Nanoporous Gold, J. Am. Chem. Soc., 129 (2007) 42–43. [190] V. Zielasek, B. Jürgens, C. Schulz, J. Biener, M.M. Biener, A.V. Hamza, M. Bäumer, Gold Catalysts: Nanoporous Gold Foams, Angew. Chem. Int. Ed., 45 (2006) 8241-8244. [191] A. Corma, H. Garcia, Supported Gold Nanoparticles as Catalysts for Organic Reactions, Chem. Soc. Rev., 37 (2008) 2096-2126. [192] T. Fujita, P. F. Guan, K. McKenna, X. Y. Lang, A. Hirata, L. Zhang, T. Tokunaga, S. Arai, Y. Yamamoto, N. Tanaka, Y. Ishikawa, N. Asao, Y. Yamamoto, J. Erlebacher, M. W. Chen, Atomic Origins of the High Catalytic Activity of Nanoporous Gold, Nat. Mater., 11 (2012) 775–780. [193] J. Biener, M.M. Biener, R.J. Madix, C.M. Friend, Nanoporous Gold – Understanding the Origin of the Reactivity of a 21st Century Catalyst Made by Pre- Columbian Technology, ACS Catal., DOI: 10.1021/acscatal.5b01586 (2015). [194] T. Déronzier, F. Morfin, M. Lomello, J.-L. Rousset, Catalysis on Nanoporous Gold–Silver Systems: Synergistic Effects Toward Oxidation Reactions and Influence of the Surface Composition, J. Catal., 311 (2014) 221-229. [195] A. Wittstock, V. Zielasek, J. Biener, C. Friend, M. Bäumer, Nanoporous Gold Catalysts for Selective Gas-Phase Oxidative Coupling of Methanol at Low Temperature, Science, 327 (2010) 319-322. [196] D. Han, T. Xu, J. Su, X. Xu, Y. Ding, Gas-Phase Selective Oxidation of Benzyl Alcohol to Benzaldehyde with Molecular Oxygen over Unsupported Nanoporous Gold, ChemCatChem, 2 (2010) 383-386. [197] Y. Ding, M. Chen, J. Erlebacher, Metallic Mesoporous Nanocomposites for Electrocatalysis, J. Am. Chem. Soc., 126 (2004) 6876–6877. [198] X. Lang, A. Hirata, T. Fujita, M. Chen, Nanoporous Metal/Oxide Hybrid Electrodes for Electrochemical Supercapacitors, Nat. Nanotechnol., 6 (2011) 232–236. [199] L. Y. Chen, Y. Hou, J. L. Kang, A. Hirata, T. Fujita, M. W. Chen, Toward the Theoretical Capacitance of RuO2 Reinforced by Highly Conductive Nanoporous Gold, Adv. Energy Mater., 3 (2013) 851–856.

- 40 -

Chapter 1

[200] F. Meng, Y. Ding, Sub-Micrometer-Thick All-Solid-State Supercapacitors with High Power and Energy Densities, Adv. Mater., 23 (2011) 4098-4102. [201] Z. Peng, S.A. Freunberger, Y. Chen, P.G. Bruce, A Reversible and Higher-Rate Li-O2 Battery, Science, 337 (2012) 563-566. [202] Y. Yu, L. Gu, X. Lang, C. Zhu, T. Fujita, M. Chen, J. Maier, Li Storage in 3D Nanoporous Au-Supported Nanocrystalline Tin, Adv. Mater., 23 (2011) 2443-2447. [203] H. Qiu, C. Xu, X. Huang, Y. Ding, Y. Qu, P. Gao, Adsorption of Laccase on the Surface of Nanoporous Gold and the Direct Electron Transfer between Them, J. Phys. Chem. C, 112 (2008) 14781-14785. [204] X. Wang, X. Y. Liu, X. L. Yan, P. Zhao, Y. Ding, P. Xu, Enzyme- Nanoporous Gold Biocomposite: Excellent Biocatalyst with Improved Biocatalytic Performance and Stability, PLoS ONE, 6 (2011) e24207. [205] Z. Liu, L. Huang, L. Zhang, H. Ma, Y. Ding, Electrocatalytic Oxidation of D-glucose at Nanoporous Au and Au–Ag Alloy Electrodes in Alkaline Aqueous Solutions, Electrochim. Acta, 54 (2009) 7286-7293. [206] M. Hakamada, M. Takahashi, M. Mabuchi, Enzyme Electrodes Stabilized by Monolayer-Modified Nanoporous Au for Biofuel Cells, Gold Bull., 45 (2012) 9-15. [207] M. Hakamada, M. Takahashi, T. Furukawa, K. Tajima, K. Yoshimura, Y. Chino, M. Mabuchi, Electrochemical Stability of Self-Assembled Monolayers on Nanoporous Au, Phys. Chem. Chem. Phys., 13 (2011) 12277-12284. [208] J. Patel, L. Radhakrishnan, B. Zhao, B. Uppalapati, R.C. Daniels, K.R. Ward, M.M. Collinson, Electrochemical Properties of Nanostructured Porous Gold Electrodes in Biofouling Solutions, Anal. Chem., 85 (2013) 11610-11618. [209] S. Saraf, C.J. Neal, S. Park, S. Das, S. Barkam, H.J. Cho, S. Seal, Electrochemical Study of Nanoporous Gold Revealing Anti-Biofouling Properties, RSC Adv., 5 (2015) 46501-46508. [210] H. Qiu, L. Xue, G. Ji, G. Zhou, X. Huang, Y. Qu, P. Gao, Enzyme- Modified Nanoporous Gold-Based Electrochemical Biosensors, Biosens. Bioelectron., 24 (2009) 3014-3018. [211] F. Meng, X. Yan, J. Liu, J. Gu, Z. Zou, Nanoporous Gold as Non- Enzymatic Sensor for Hydrogen Peroxide, Electrochim. Acta, 56 (2011) 4657-4662. [212] L.Y. Chen, T. Fujita, M.W. Chen, Biofunctionalized Nanoporous Gold for Electrochemical Biosensors, Electrochim. Acta, 67 (2012) 1-5. [213] T. Li, F. Jia, Y. Fan, Z. Ding, J. Yang, Fabrication of Nanoporous Thin-Film Working Electrodes and Their Biosensingapplications, Biosens. Bioelectron., 42 (2013) 5-11. [214] A.E. Cass, G. Davis, G.D. Francis, H.A.O. Hill, W.J. Aston, I.J. Higgins, E.V. Plotkin, L.D. Scott, A.P. Turner, Ferrocene-Mediated Enzyme Electrode for Amperometric Determination of Glucose, Anal. Chem., 56 (1984) 667- 671.

- 41 -

Chapter 1

[215] U. Salaj-Kosla, M.D. Scanlon, T. Baumeister, K. Zahma, R. Ludwig, P.Ó. Conghaile, D. MacAodha, D. Leech, E. Magner, Mediated Electron Transfer of Cellobiose Dehydrogenase and Glucose Oxidase at Osmium Polymer-Modified Nanoporous Gold Electrodes, Anal. Bioanal. Chem., 405 (2013) 3823-3830. [216] U. Salaj-Kosla, S. Pöller, Y. Beyl, M.D. Scanlon, S. Beloshapkin, S. Shleev, W. Schuhmann, E. Magner, Direct Electron Transfer of Bilirubin Oxidase (Myrothecium Verrucaria) at an Unmodified Nanoporous Gold Biocathode, Electrochem. Commun., 16 (2012) 92-95. [217] U. Salaj-Kosla, S. Pöller, W. Schuhmann, S. Shleev, E. Magner, Direct Electron Transfer of Trametes Hirsuta Laccase Adsorbed at Unmodified Nanoporous Gold Electrodes, Bioelectrochem., 91 (2013) 15-20. [218] T. Siepenkoetter, U. Salaj‐Kosla, X. Xiao, S. Belochapkine, E. Magner, Nanoporous Gold Electrodes with Tuneable Pore Sizes for Bioelectrochemical Applications, Electroanalysis, 28 (2016) 2415-2423. [219] X.-Y. Lang, H.-Y. Fu, C. Hou, G.-F. Han, P. Yang, Y.-B. Liu, Q. Jiang, Nanoporous Gold Supported Cobalt Oxide Microelectrodes as High- Performance Electrochemical Biosensors, Nat. Commun., 4 (2013) 2169.

- 42 -

CHAPTER 2:

A biofuel cell operating in nonaqueous solutions

Chapter 2

Chapter 2. A biofuel cell operating in nonaqueous solutions

2.1 Introduction

Enzymes are located in and are catalytically active in a wide range of environments in nature, ranging from aqueous solutions to hydrophobic cell membranes [1]. The catalytic activity of enzymes in a range of nonaqueous solvents has been reported [2-4]. The use of enzymes in such media has a number of advantages that include increased substrate solubility, increased thermal stability, suppression of side reactions that can occur in water and changes in enzymatic selectivity [5, 6]. For example, transesterification reactions catalysed by lipase cannot occur in aqueous solution, but becomes favourable in organic solvents [2]. Enzymatic selectivity, including stereoselectivity and regioselectivity, can also be tuned by altering solvents. The enantioselectivity of γ-chymotrypsin varies greatly in the transesterification of methyl 3-hydroxy-2-phenylpropionate with propanol in a variety of organic solvents. The enzyme strongly prefers the S-enantiomer in some solvents (e.g. hexane, toluene), while the R-antipode is more reactive in others (e.g. acetone) [7]. Early nonaqueous enzymology studies focused on hyrdolyase [1, 8, 9]. Enzymatic reactions in nonaqueous solvents enable the production of a large range of useful chemicals including chiral drug molecules, biopolymers, peptides and proteins, modified fats and oils, and structured lipids etc. [10]. The properties of redox proteins containing transition metals as redox centers, such as cytochrome c, have been examined in organic solvents [11-13]. The redox potential (E°’) of the haem is sensitive to the polarity of the haem environment [14]. For example, O’Reilly et al. investigated the E°’ (ranging from +239 to +265 mV vs. SHE), enthalpy, and entropy of reduction of cytochrome c in various aqueous/organic solvent mixtures (100% aqueous buffer, 30% acetonitrile, 40%

- 43 -

Chapter 2 dimethyl sulfoxide, and 50% methanol) [11]. They concluded that the temperature- E°’ profiles of cytochrome c are likely to be determined by the temperature dependence of the dielectric constants of the solvent and of the protein. Enzymes are insoluble in nonaqueous media and as a consequence, electrochemical studies of redox enzymes in such media require that the enzyme be immobilized on electrodes. . Oxidoreductase, such as glucose oxidase (GOx), modified electrodes in organic solvents have been tested as biosensors to evaluate their capability to operate in aqueous organic mixtures [15]. Iwuoha et al. examined the effect of a series of organic solvents (acetonitrile, acetone, etc.) on the performance (e.g. maximum catalytic current response) of amperometric GOx modified electrodes [16-18]. Using ferrocenemonocarboxylic acid (FMCA) as a soluble electron-transfer mediator, a GOx electrode had a two-fold higher catalytic current in 90% (v/v) acetonitrile when compared to the response in phosphate buffer [17]. The enhanced activity was proposed to arise from desorption of water molecules from the active site of the enzyme to a favorable extent. Enzymatic biofuel cells (EEBFCs) that utilize oxidoreductases to generate electrical energy are of interests due to their potential applications as autonomous power suppliers [19, 20]. A wide range of reports have described the development of miniaturized EBFCs with extended lifetimes and increased power densities, focussing on screening enzymes from a variety of sources, developing more efficient methods of immobilization, the use of a range of electrode materials and the deployment of enzyme cascades [21, 22]. Due to the potential applications in biomedical devices, the properties of EBFCs are invariably examined in physiological conditions, with no reports on their use in nonaqueous solvents. Properties of organic solvents such as log P (a quantitative measure of solvent polarity) [3], dielectric constant [23, 24] and viscosity [25] can affect the enzymatic activity and specificity [26]. Laane et al. attempted to establish a correlation between log P and the activity of the enzyme as log P describes the ability of organic solvents to distort the essential water layer that stabilizes enzymes [3]. This

- 44 -

Chapter 2 is true for the reduction of prednisone catalysed by 20β-hydroxysterioid dehydrogenase. Affleck et al. investigated the dynamics of alpha-chymotrypsin in solvents with various dielectric constants using electron paramagnetic resonance (EPR) spectroscopy and molecular dynamics (MD) simulations [24]. They found that solvents with different dielectric constants can affect the motions of protein. Thus, the stereoselectivity of the enzyme could be altered by altering the dynamics of the protein. Konash et al. characterised the performance of a (HRP) based hydrogen peroxide sensor using free ferrocene dimethanol as a mediator in various solvents and observed a linear correlation between the sensor sensitivity and the reciprocal of the solvent viscosity [25]. The response of the sensor was mainly constrained by the rate of diffusion of the mediator.

Scheme 2.1 The underlying mechanism and employed polymer structure on the proposed anode and cathode.

In this chapter, the properties of a well-studied EBFC based on glucose and

O2 using glucose oxidase (GOx, Aspergillus niger) and bilirubin oxidase (BOx, Myrothecium verrucaria) were studied. GOx and BOx were entrapped with the osmium polymers [Os(4,4′-dimethyl-2,2′-bipyridine)2(polyvinylimidazole)10Cl]Cl

(Os(dmbpy)2PVI) and [Os(2,2′-bipyridine)2(polyvinylimidazole)10Cl]Cl

(Os(bpy)2PVI), respectively (Scheme 2.1, Figure 2.1A) [27]. Poly(ethylene glycol)diglycidyl ether (PEGDGE) was used as the cross-linking agent. The enzyme/redox polymer mixtures were immobilised onto dealloyed nanoporous gold (NPG) electrodes. NPG is a stable and conductive support [28] that enables the - 45 -

Chapter 2 polymer to be confined within the porous structure of the support, as well as improved rates of electron transfer between the enzymes and the electrode [29]. Due to the high selectivity of enzymes, the EBFC can be tested in a one-compartment cell. The EBFC operating in organic solvents described in the chapter is believed to be the first report on bioenergy conversion in a non-aqueous solution. Such cells may have application in bioelectrosynthesis in nonaqueous solutions. For example, an alcohol dehydrogenase based biocathode could be used to synthesis chiral hydrophobic alcohols in organic solvents.

2.2 Experimental section

2.2.1 Materials Sulfuric acid (95–98 %), nitric acid (70%),potassium phosphate monobasic (≥99 %) and dibasic (≥98 %), d-(+)-glucose (99.5 %), ethanol (EtOH, 96%), acetonitrile (ACN, ≥99.9%), acetone (AC, ≥99.8%), 1-propanol (PrOH, ≥99.5%), methanol (MeOH, ≥99.9%), 1-butanol(BuOH, ≥99.7%), 1-pentanol (PeOH, ≥99%), tetraethylammonium p-toluenesulfonate (TEATS), and poly(ethylene glycol)diglycidyl ether (PEGDGE) were obtained from Sigma-Aldrich Ireland, Ltd. Absolute ethanol was obtained from Lennox Ltd., Ireland. All solutions were prepared with deionised water (resistivity of 18.2 MΩ cm) from an Elgastat maxima-HPLC (Elga, UK). All experiments were carried out at room temperature (20±2 °C).

The complexes, [Os(2,2′-bipyridine)2(polyvinylimidazole)10Cl]Cl

o (Os(bpy)2PVI, E : 0.22 V vs. Ag/AgCl) and [Os(4,4′-dimethyl-2,2′-

o bipyridine)2(polyvinyllimida -zole)10Cl]Cl (Os(dmbpy)2PVI, E : 0.12 V vs. Ag/AgCl) were synthesized using published procedures [30]. GOx from Aspergillus niger (EC 1.1.3.4, type II, ≥15,000 U g-1) and BOx from Myrothecium verrucarria (EC 1.3.3.5, 2.63 U mg-1) were purchased from Sigma-Aldrich, Ireland, Ltd. and Amano Enzyme Inc. (Nagoya, Japan), respectively. NPG sheets were prepared by dealloying 100-nm-thick Au/Ag leaves (12- carat, Eytzinger, Germany) in concentrated HNO3 for 30 min at 30 ˚C and placed on - 46 -

Chapter 2 the surface of glassy carbon electrodes (GCE) (Scheme 2.2) [31]. Such dealloying conditions resulted in NPG with a pore size of ca. 30 nm (Scheme 2.2). Prior to using,

NPG electrodes were electrochemically cleaned by scanning potential in 1 M H2SO4.

Scheme 2.2 A typical NPG electrode preparation process.

2.2.2 Enzyme immobilization A 5.3 μl aliquot of a 6 mg ml-1 aqueous suspension of osmium-based redox

-1 polymer, Os(dmbpy)2PVI or Os(bpy)2PVI, was combined with 1.3 μl of a 15 mg ml aqueous solution of PEGDGE and, either 3.2 μl of a 10mg ml-1 solution of GOx or BOx. All the components were homogenously mixed using a vortex mixer. The surface of the NPG electrode was fully covered by a drop of the solution, and immediately placed in a vacuum desiccator connected to a vacuum pump for 10 min. The electrodes were then transferred into the fridge, allowed to dry overnight in the dark at 4oC. To elucidate the role of the enzymes on the catalytic response, NPG electrodes modified only with redox polymer were also prepared.

2.2.3 Electrochemical measurements Electrochemical studies were performed using a CHI802 potentiostat (CH Instruments, Austin, Texas) in a standard three-electrode electrochemical cell. Platinum wire and saturated calomel electrodes (SCE) were used as the counter and reference electrodes, respectively. Enzyme-modified electrodes were immersed in 0.1 M pH 7.0 phosphate buffer solution (PBS) for at least 20 min prior to electrochemical measurements to allow for film swelling.

The biofuel cell consists of a bioanode made of NPG/Os(dmbpy)2PVI/GOx and a NPG/Os(bpy)2PVI/BOx biocathode. The power density of biofuel cells was measured in different oxygen-bubbled organic solvents containing 5 mM glucose - 47 -

Chapter 2 using the bioanode as the working electrode and the biocathode as a combined counter/reference electrode. The potential was scanned at a scan rate of 1 mV s-1, while recording the current in the circuit. Storage stability was determined by storing the EBFC in buffer solution at 4oC and measuring the response for the required period of time. Nonaqueous solutions were prepared by the addition of the desired volume of buffer solution (4.4 mM phosphate, pH 7.0) to the organic solvent. 0.1 M TEATS was used as the electrolyte. Operational stability tests (for 5 hours in 95% ACN) were performed by continuously recording the current at a constant potential of 0.15 V, with O2 bubbling of the solution.

2.3 Results and discussion

Figure 2.1 (A) Schematic diagram of the biofuel cell. Cyclic voltammograms

(CVs) of NPG/Os(dmbpy)2PVI/GOx modified electrodes in PBS (B) and 95% ACN

-1 (C) at a scan rate of 5 mVs . (C) CVs of NPG/Os(bpy)2PVI/BOx electrode in PBS (D) and 95% ACN (E) at a scan rate of 5 mVs-1. (F) Polarization and power curves for

the EBFC in O2 bubbled PBS (initial curve: solid line; after testing in 95% ACN: dotted line) and 95% ACN (dashed line).

Firstly, the catalytic activities of the NPG/Os(dmbpy)2PVI/GOx and

- 48 -

Chapter 2

NPG/Os(bpy)2PVI/BOx electrodes were separately studied in a three-electrode cell.

Cyclic voltammograms (CVs) of NPG/Os(dmbpy)2PVI/GOx in 0.1 M pH 7.0 phosphate buffer (PBS) exhibited a pair of well-defined redox peaks corresponding to the conversion of Os2+/Os3+ at a low scan rate of 5 mV s-1 (Figure 2.1B, solid line).

The peak potential separation, ΔEp, of 15 mV, was indicative of a rapid and reversible electron transfer process. Upon addition of 5 mM glucose, a sigmoidal-shaped curve, characteristic of the bioelectrocatalytic oxidation of glucose, with an onset potential of

-0.1 V, was obtained (Figure 2.1B, dashed line). The response current density, jresponse, defined as the difference between the catalytic and background current density was 54 µA cm-2 in PBS.

Figure 2.2 Chronoamperometry response of the NPG/Os(dmbpy)2PVI/GOx bioanode at +0.2 V vs. SCE in PBS (A) and 95% ACN (B).

NPG/Os(dmbpy)2PVI/GOx was then transferred into acetonitrile (ACN) containing 5% added buffer. A reversible redox curve (Figure 2.1C, solid line) with a

-2 jresponse of 2.2 µA cm (4% of the original activity in PBS, Figure 2.1C, dashed line), with an increased onset potential of -0.03 V, was obtained. The catalytic response in both PBS and ACN was further confirmed by chronoamperometry, with a catalytic current clearly evident (Figure 2.2). This agrees with previous reports that enzymes in organic media possess only a fraction of the catalytic activity observed in water [15] with the decreased activity arising from a range of effects [8] (e.g. decrease in molecular flexibility [32], reduction in the amount of water bound to the enzyme). There was no perceptible change in the cathodic peak potential. It is noteworthy that - 49 -

Chapter 2 the response of the electrode was largely retained on re-immersion in aqueous buffer

-2 solution (jresponse of 52 µA cm , 96% of initial response) (Figure 2.1B, dotted line) indicating that GOx had not been denatured in the organic solution.

NPG/Os(bpy)2PVI/BOx cathodes also showed a pair of reversible and well- defined redox peaks in N2 bubbled PBS (Figure 2.1D, solid line). An initial jresponse of

-2 123 µA cm and an onset potential of 0.43 V in O2 bubbled aqueous solution were obtained (Figure 2.1D, dashed line). On switching to 95% ACN, a significantly lower

-2 jresponse of 5 µA cm (Figure 2.1E, dashed line) was observed, while the onset potential of O2 reduction decreased to approximate 0.33 V. Chronoamperometric data confirmed the activity of the enzyme in ACN (Figure 2.3). The recovery of activity

-2 (jresponse of 105 µA cm , 85% of initial response) in PBS indicated that BOx had not been significantly denatured on exposure to 95% ACN (Figure 2.1D, dotted line). No amperometric response was observed upon the addition of glucose or O2 to modified electrodes without GOx or BOx, indicating that the enzymes were catalytically active (Figure 2.4).

Figure 2.3 Chronoamperometry response of the NPG/Os(bpy)2PVI/BOx biocathode at +0.1 V vs. SCE in PBS (A) and 95% ACN (B).

- 50 -

Chapter 2

Based on the above results, GOx and BOx modified anodes and cathodes were subsequently assembled into EBFCs (Figure 2.1A), and the response of the cell monitored by linear sweep voltammetry (scan rate of 1 mV s-1). The EBFC displayed an open circuit voltage (OCV) of 0.56 V (the difference of the onset potentials associated oxidation of glucose and reduction of O2), a maximum current density of 21.2 µA cm-2, and a maximum power density of 3.65 µW cm-2 at a potential of 0.21 V in O2 bubbled PBS containing 5 mM glucose (Figure 2.1F, solid line).

Figure 2.4 Chronoamperometric response of blank electrodes without enzymes in PBS: NPG/Os(dmbpy)2PVI at +0.2 V vs. SCE (A); NPG/Os(bpy)2PVI at +0.1 V vs. SCE (B).

On replacement with 95% ACN, the performance of the EBFC decreased, with an OCV of 0.36 V, a maximum current density of 7.11 µA cm-2, and a maximum power density of 0.47 µW cm-2 at 0.12 V (Figure 2.1F, dashed line). The decrease in power arose from both changes in OCV and in current density (enzyme activity). The response of the cell was retained on re-immersion in PBS (OCV of 0.56 V), and a maximum power density of 3.44 µW cm-2 (94% of the original response, Figure 2.1F, dotted line). Leakage of redox polymer from the electrode surface was mainly responsible for the loss response [27] as evidenced by the decrease in the peak current in a blank buffer solution (data not shown).

- 51 -

Chapter 2

The operational stability of EBFC in 95% ACN and PBS was examined, with a half-life of ca. 3 and 5 h, respectively (Figure 2.5). On storage at 4 °C, the cell retained 75% of the initial response after storage for 12 h (Figure 2.6). A minor decrease of 15% in the OCV was observed after storage for 60 h (Figure 2.6).

Figure 2.5 Operational stability of EBFC in PBS (A) and 95% ACN (B).

Figure 2.6 Storage stability of the proposed EBFC.

The response of the EBFCs was examined in solutions with varying water content (1–5% (v/v)) in ACN (Figure 2.7A). The maximum power density 0.47 µW cm-2 was obtained with 5% added buffer decreasing to 0.13 µW cm-2 in 99% ACN, indicating as expected, that the enzymes are more active at higher water content. Water content used here was no higher than 5% with the concern that properties of organic solvents might be greatly altered. - 52 -

Chapter 2

The response of the EBFC was examined in a series of common solvents with 5% v/v added buffer (Figure 2.7B). The maximum power density was obtained in methanol and decreased in the sequence methanol> ethanol>propanol>butanol> pentanol. The response in these solvents decreased with increasing solvent hydrophobicity in an approximately linear manner (Figure 2.8C). The response in ACN and acetone (AC) was lower than in the short-chain alcohols (methanol and ethanol), and did not follow this linear relationship. Generally, the EBFC had a higher power density in organic solvents with lower values of log P. Similar trends were reported for single enzyme electrodes [24, 33]. The trend observed here is likely to arise from solvent based interactions at the enzymes active sites with either the enzymatic substrates or with the redox polymers.

- 53 -

Chapter 2

Figure 2.7 Power density curve of the EBFC in different percentages of O2 bubbled ACN containing 5 mM glucose (A) and different organic solvents containing 5 mM glucose (B). (C) Plot of the power density versus log P (data points taken from reference [3]). The error bars correspond to the values recorded for three EBFCs.

2.4 Conclusions

In conclusion, we describe the assembly of a membraneless EBFC that coupled an NPG/Os(dmbpy)2PVI/GOx bioanode with a NPG/Os(bpy)2PVI/BOx biocathode that operates in organic solvents. As far as we know, this is the first report

- 54 -

Chapter 2 of EBFCs operating in organic solvents. More importantly, both bioelectrodes displayed reversible recovery of their initial activities in PBS after operation in organic solutions. A well-defined trend with the maximum power density decreasing with increasing log P was obtained in straight-chain monohydric alcohols. The use of this EBFC is limited to a small range of solvents due to the low solubility of glucose in nonaqueous media (generally not higher than 5 mM glucose in 95% organic solvent). Applications for the EBFC described here may be possible in nonaqueous solvents with low water content.

2.5 References

[1] A.M. Klibanov, Improving Enzymes by Using Them in Organic Solvents, Nature, 409 (2001) 241-246. [2] A. Zaks, A.M. Klibanov, Enzyme-Catalyzed Processes in Organic Solvents, Proc. Natl. Acad. Sci., 82 (1985) 3192-3196. [3] C. Laane, S. Boeren, K. Vos, C. Veeger, Rules for Optimization of Biocatalysis In Organic Solvents, Biotechnol. Bioeng., 30 (1987) 81-87. [4] A. Zaks, A.M. Klibanov, The Effect of Water on Enzyme Action in Organic Media, J. Biol. Chem., 263 (1988) 8017-8021. [5] A.M. Klibanov, Asymmetric Enzymatic Oxidoreductions in Organic Solvents, Curr. Opin. Biotechnol., 14 (2003) 427-431. [6] E.I. Iwuoha, M.R. Smyth, M.E. Lyons, Organic Phase Enzyme Electrodes: Kinetics and Analytical Applications, Biosens. Bioelectron., 12 (1997) 53-75. [7] C.R. Wescott, H. Noritomi, A.M. Klibanov, Rational Control of Enzymatic Enantioselectivity through Solvation Thermodynamics, J. Am. Chem. Soc., 118 (1996) 10365-10370. [8] A.M. Klibanov, Why Are Enzymes Less Active in Organic Solvents than in Water?, Trends Biotechnol., 15 (1997) 97-101. [9] G. Carrea, G. Ottolina, S. Riva, Role of Solvents in the Control of enzyme Selectivity in Organic Media, Trends Biotechnol., 13 (1995) 63-70. [10] S. Hari Krishna, Developments and Trends in Enzyme Catalysis in Nonconventional Media, Biotechnol. Adv., 20 (2002) 239-267. [11] N.J. O’reilly, E. Magner, Electrochemistry of Cytochrome c in Aqueous and Mixed Solvent Solutions: Thermodynamics, Kinetics, and the Effect Of Solvent Dielectric Constant, Langmuir, 21 (2005) 1009-1014. [12] E.V. Ivanova, E. Magner, Direct Electron Transfer of Haemoglobin and Myoglobin in Methanol and Ethanol at Didodecyldimethylammonium Bromide Modified Pyrolytic Graphite Electrodes, Electrochem. Commun., 7 (2005) 323-327.

- 55 -

Chapter 2

[13] S.G. Sivakolundu, P.A. Mabrouk, Cytochrome c Structure and Redox Function in Mixed Solvents Are Determined by the Dielectric Constant, J. Am. Chem. Soc., 122 (2000) 1513-1521. [14] S. Crilly, E. Magner, Reversible Increase in the Redox Potential of Cytochrome c in Methanol, Chem. Commun., (2009) 535-537. [15] S. Kröger, S.J. Setford, A.P. Turner, Assessment of Glucose Oxidase Behaviour in Alcoholic Solutions Using Disposable Electrodes, Anal. Chim. Acta, 368 (1998) 219-231. [16] E.I. Iwuoha, M.R. Smyth, J.G. Vos, Amperometric Glucose Sensor Containing Nondiffusional Osmium Redox Centers: Analysis of Organic‐Phase Responses, Electroanalysis, 6 (1994) 982-989. [17] E.I. Iwuoha, M.R. Smyth, Organic-Phase Application of an Amperometric Glucose Sensor, Analyst, 119 (1994) 265-267. [18] E.I. Iwuoha, M.R. Smyth, M.E.G. Lyons, Solvent Effects on the Reactivities of an Amperometric Glucose Sensor, J. Electroanal. Chem., 390 (1995) 35-45. [19] D. Leech, P. Kavanagh, W. Schuhmann, Enzymatic Fuel Cells: Recent Progress, Electrochim. Acta, 84 (2012) 223-234. [20] S. Cosnier, A. Le Goff, M. Holzinger, Towards Glucose Biofuel Cells Implanted in Human Body for Powering Artificial Organs: Review, Electrochem. Commun., 38 (2014) 19-23. [21] Y.H. Kim, E. Campbell, J. Yu, S.D. Minteer, S. Banta, Complete Oxidation of Methanol in Biobattery Devices Using A Hydrogel Created from Three Modified Dehydrogenases, Angew. Chem. Int. Ed., 52 (2013) 1437-1440. [22] M. Cooney, V. Svoboda, C. Lau, G. Martin, S. Minteer, Enzyme catalysed biofuel cells, Energy Environ. Sci., 1 (2008) 320-337. [23] Q. Deng, S. Dong, The Effect of Substrate and Solvent Properties on the Response of an Organic Phase Tyrosinase Electrode, J. Electroanal. Chem., 435 (1997) 11-15. [24] R. Affleck, C.A. Haynes, D.S. Clark, Solvent Dielectric Effects on Protein Dynamics, Proc. Natl. Acad. Sci., 89 (1992) 5167-5170. [25] A. Konash, E. Magner, Characterization of an Organic Phase Peroxide Biosensor Based on Horseradish Peroxidase Immobilized in Eastman AQ, Biosens. Bioelectron., 22 (2006) 116-123. [26] T. Ke, C.R. Wescott, A.M. Klibanov, Prediction of the Solvent Dependence of Enzymatic Prochiral Selectivity by Means of Structure-Based Thermodynamic Calculations, J. Am. Chem. Soc., 118 (1996) 3366-3374. [27] U. Salaj-Kosla, M.D. Scanlon, T. Baumeister, K. Zahma, R. Ludwig, P.Ó. Conghaile, D. MacAodha, D. Leech, E. Magner, Mediated Electron Transfer of Cellobiose Dehydrogenase and Glucose Oxidase at Osmium Polymer-Modified Nanoporous Gold Electrodes, Anal. Bioanal. Chem., 405 (2013) 3823-3830. [28] Y. Ding, Y.J. Kim, J. Erlebacher, Nanoporous Gold Leaf: “Ancient Technology”/Advanced Material, Adv. Mater., 16 (2004) 1897–1900.

- 56 -

Chapter 2

[29] M.D. Scanlon, U. Salaj-Kosla, S. Belochapkine, D. MacAodha, D. Leech, Y. Ding, E. Magner, Characterization of Nanoporous Gold Electrodes for Bioelectrochemical Applications, Langmuir, 28 (2012) 2251-2261. [30] P.A. Jenkins, S. Boland, P. Kavanagh, D. Leech, Evaluation of Performance and Stability Of Biocatalytic Redox Films Constructed with Different Copper Oxygenases and Osmium-Based Redox Polymers, Bioelectrochem., 76 (2009) 162-168. [31] X. Xiao, J. Ulstrup, H. Li, J. Zhang, P. Si, Nanoporous Gold Assembly of Glucose Oxidase for Electrochemical Biosensing, Electrochim. Acta, 130 (2014) 559-567. [32] D.S. Hartsough, K.M. Merz Jr, Protein Flexibility in Aqueous and Nonaqueous Solutions, J. Am. Chem. Soc., 114 (1992) 10113-10116. [33] I. Cruz Vieira, O. Fatibello-Filho, Biosensor Based On Paraffin/Graphite Modified with Sweet Potato Tissue for the Determination of Hydroquinone in Cosmetic Cream in Organic Phase, Talanta, 52 (2000) 681-689.

- 57 -

CHAPTER 3:

A symmetric supercapacitor/biofuel cell hybrid device based on enzyme-modified nanoporous gold: an autonomous pulse generator

Chapter 3

Chapter 3. A symmetric supercapacitor/biofuel cell hybrid device based on enzyme-modified nanoporous gold: an autonomous pulse generator

3.1. Introduction

Enzymatic biofuel cells (BFCs) utilizing oxidoreductases as electrocatalysts can be used to generate electricity from fuels such as sugars or alcohols in combination with dioxygen [1-3]. BFCs are of interest as power sources for biosensors [4, 5], medical implants (e.g. insulin pumps, cardiac pacemakers [6]), and other devices [7, 8]. To be able to activate commonly used microelectronic devices (such as commercial pacemakers), appropriate output voltages (minimum of 1.4 V) are required [6]. The open circuit voltage (OCV) of glucose and oxygen BFCs is limited by the thermodynamic value of 1.179 V [9], and in practice by the difference between the onset redox potentials of the bioanode and biocathode [10]. The observed OCV can be increased by using direct electron transfer (DET) or by the use of redox mediators with redox potentials closer to those of the enzyme/cofactor [11]. The OCV can also be increased by using multiple cells connected in series [6]. However, due to the presence of conductive fluids within the body, implantable cell stacks suffer from the problem of short-circuits between individual cells [6, 12]. In such systems, isolation of the cells is essential. Another route is to couple BFCs with external electronic devices to increase the voltage. For example, using a charge pump and a DC-DC converter, a fluidic BFC utilizing PQQ-dependent glucose dehydrogenase and laccase with an intrinsic OCV of 0.47 V was sufficient to power a pacemaker [13]. Falk et al. presented a self-powered wireless lactose biosensing system, consisting of an energy harvesting module including a voltage amplifier and capacitor to build a

- 57 -

Chapter 3 power source based on a BFC using bilirubin oxidase (BOx) and cellobiose dehydrogenase (CDH) [14]. In addition to low voltage outputs, BFCs are also limited by their low current/power densities, which can be improved through efficient substrate diffusion [15], enhanced rates of electron transfer between enzymes and electrodes, improving catalytic activity [16] and loading of enzymes [17], as well as utilizing enzyme cascades for deep and complete oxidation pathways [18-20]. The introduction of capacitors into the BFC circuit enables the accumulation of charge, resulting in output pulses of higher power. Sode et al. proposed the concept of a “BioCapacitor” with the integration of a charge pump/capacitor and a BFC that resulted in higher voltages and currents [21, 22]. Electrochemical capacitors (known as supercapacitors) [23] take advantage of the electrical double layer capacitance attained via ion adsorption or pseudocapacitance achieved by fast and reversible faradaic reactions, offering high specific power density and great durability. Supercapacitors externally connected to a laccase-based cathode and zinc anode based biobattery, had higher power stability than the battery itself [24]. Recent progress has seen BFC assemblies with capacitive bioelectrodes [25-29]. These supercapacitor/BFC hybrids, or self-charging biocapacitors, are based on the fabrication of hybrid composite modified electrodes with integration of enzymes and capacitive materials. The main feature is their ability to generate cyclic, high power pulses from the discharge of the supercapacitor, which is recharged towards to OCV via the internal BFC in the following open-circuit mode [25]. In 2014, Agnès et al. developed an EBFC consisting of a compressed porous CNT matrix modified GOx bioanode and a Lac based biocathode, delivering 3 mA and 2 mW pulse with a short duration of 10 ms per 10 s for 5 days in the presence of glucose and O2 [25]. In this case, the electricity generated by the EBFC was stored continuously in the EDLC of CNTs. In parallel, Pankratov et al. reported a hybrid device based on flat graphite foil electrodes, with one side bearing an EBFC comprised of an AuNPs-CDH bioanode and an AuNPs-BOx biocathode, and the other

- 58 -

Chapter 3 side modified with capacitive materials (CNT/polyaniline) [27]. It displayed an initial power output of 1.2 mW cm-2 at 0.38 V which is 170-fold of that by the EFBC alone.

Kizling et al. reported a fructose/O2 EBFC/supercapacitor hybrid device comprised of a cellulose/polypyrrole/FDH bioanode [26, 28] and a naphthylated CNTs/Lac biocathode [28]. Three biodevices in a series could generate pulses of 45 s whose potentials were above 1 V. Villarrubia et al. prepared a glucose/O2 EBFC/supercapacitor based on a capillary-driven microfluidic system [30]. The bioanode was mediated NAD+-dependent GDH on buckypaper and the biocathode was Toray paper/buckypaper/BOx. The hybrid device could be self-charged and

-2 discharged by a current density as high as 4 mA cm for 0.01 s with a Pmax of 0.87 mW cm-2 (10.6 mW), 10-fold higher than that of the EBFC itself. Knoche et al. prepared a hybrid device consisting of a carbon felt/MWCNTs/dimethylferrocene- modified linear poly(ethylenimine) (FcMe2-LPEI)/FAD-GDH bioanode and a biocathode based on a carbon felt/anthracene terminated MWCNTs/BOx [31]. They found that the FcMe2-LPEI redox polymer serves as a mediator, enzyme immobilization matrix and also contributes as a supercapacitor whose capacitance increased with polymer loading. The device was able to generate 1 mA pulses for 1 s with 1 mW of power delivered. Pankratov et al. developed a capacitive EBFC using a same Os polymer on a GDH anode and BOx cathode [32]. The capacitance of the polymer was used for energy storage with an OCP up to 0.45 V, which could be discharged as a pulse with an 8-fold higher power output than that obtained in steady state. Further, Alsaoub et al. presented a hybrid device using different Os polymers on the bioanode and biocathode respectively [33]. The majority of supercapacitor/BFC systems have relied on the use of high-surface-area carbon nanomaterials (CNMs), such as CNT [25-28] and graphene [29]. However, the potential toxicity of CNMs [34] should be taken into account for in vivo applications and direct exposure to CNMs should be avoided in implantable devices [35]. Dealloyed nanoporous gold (NPG), a porous material in a self-supporting bulk form comprising three-dimensional frameworks of bicontinuous pores and ligaments [36],

- 59 -

Chapter 3 has been investigated as conductive and non-toxic supports for supercapacitors [37, 38] and enzyme immobilisation [39-42], separately. Mediators are required to enable efficient electron transfer between the cofactor of the enzyme and the NPG surface [43]. In this context, alternate potential pulses could be applied to electrodeposit osmium redox polymers with the co-immobilisation of enzymes onto electrode surface [44-46]. Unlike other soluble mediators that are prone to leakage [29], the electrodeposited redox polymer is robust, even in hydrodynamic conditions. For example, a laccase/redox polymer composite film showed little loss in response after rotation for 24 h at 2500 rpm [47]. In this chapter, we electrodeposited poly(3,4- ethylenedioxythiophene) (PEDOT) and the redox polymer [Os(2,2′-

+/2+ bipyridine)2(polyvinylimidazole)10Cl] (Os(bpy)2PVI) onto NPG electrodes with the co-immobilisation of enzymes. Flavin adenine dinucleotide-dependent glucose dehydrogenase (FAD-GDH, EC 1.1.99.10, D-glucose: acceptor 1-oxidoreductase) was used as an oxygen-insensitive enzyme at the anode [48], in contrast to glucose oxidase (GOx) which depletes dissolved oxygen and produces unwanted hydrogen peroxide [49]. BOx was immobilised at the cathode and the properties of the cell were characterised in detail. A proof-of-concept pulse generator for a pacemaker was demonstrated, which was able to deliver a 10 μA pulse at a frequency of 0.2 Hz. This chapter evaluates the electrodeposition of Os polymer is a feasible route to the deposition of a uniform biofilm onto NPG. This method is subsequently used throughout the thesis to fabricate bioelectrodes undergoing mediated electron transfer. Moreover, the self-powered pulse generator provides a new method of utilising EBFCs as power sources, which can be easily miniaturized for implantable applications.

3.2. Experimental section

3.2.1. Materials Potassium phosphate monobasic (≥99 %) and dibasic (≥98 %), D-(+)-glucose (99.5 %), 3,4-ethylenedioxythiophene (EDOT, 97%) were obtained from Sigma- - 60 -

Chapter 3

Aldrich Ireland, Ltd. Potassium chloride (KCl, ≥99 %) was purchased from Fisher Scientific Ireland, Ltd. All solutions were prepared with deionised water (resistivity of

18.2 MΩ cm) from an Elgastat maxima-HPLC (Elga, UK). Os(bpy)2PVI was synthesised using a published procedure [50]. BOx from Myrothecium verrucaria (EC 1.3.3.5, 2.63 U mg-1) was purchased from Amano Enzyme Inc., Japan. Recombinant, (in Pichia pastoris) expressed Glomerella cingulata GDH (EC 1.1.99.10) with a specific activity of 572 U mg-1 was prepared according to a published route [51]. NPG leaves were fabricated by dealloying ca. 100 nm thick Au/Ag leaf alloy

(12-carat, Eytzinger, Germany) in concentrated HNO3 (Sigma-Aldrich) for 30 min at 30 °C. The NPG films were then attached onto pre-polished glassy carbon electrodes (GCEs) with a diameter of 4 mm. Prior to use, cyclic voltammetry (CV) of NPG in

1 M H2SO4 were carried out to create clean surfaces and left to dry naturally.

3.2.2. Enzyme immobilisation procedures The electrodeposition solutions contained 0.1 M pH 7.0 phosphate buffer solution (PBS) with 2 mM polyethylene glycol 3400 (PEG3400), 20 mM EDOT, 0.5

-1 -1 mg ml Os(bpy)2PVI and either 0.5 mg ml of FAD-GDH or BOx. The presence of PEG enabled the dispersion of EDOT in aqueous media and increased the hydrophilicity of the polymer [52, 53]. A pulse sequence of 0.9 V (2 s) and -0.4 V (3 s) was used for deposition. The electrodes were then gently rinsed with PBS. For comparison purposes, films were also deposited onto polycrystalline planar Au electrodes.

3.2.3. Morphology characterisation Scanning electron microscopy (SEM) images were recorded using a Hitachi SU-70 microscope operated at 15 kV. Transmission electron microscopy (TEM) images were obtained using a JEOL JEM-2100 instrument at an acceleration voltage of 200 kV. The average pore size and deposition layer thickness were obtained by performing at least 30 different measurements with ImageJ software (National Institutes of Health, Bethesda, Maryland) [54].

- 61 -

Chapter 3

3.2.4. Electrochemical measurements Generally, electrochemical studies were performed using a CHI802 potentiostat (CH Instruments, Austin, Texas) in a standard three-electrode electrochemical cell containing 0.1 M pH 7.0 PBS and 0.1 M KCl. Enzyme-modified electrodes, a platinum wire and saturated calomel electrode (SCE) were used as the working, counter and reference electrodes, respectively. The polarisation and power curves of the assembled biofuel cells were measured using the bioanode as working electrode and the biocathode as a combined counter/reference electrode. The potential

-1 was scanned at a scan rate of 1 mV s in the presence of O2-bubbled 20 mM glucose, while recording the current in the circuit. All experiments were carried out at room temperature (20±2 °C). The current densities or power densities were calculated using the geometric surface area of the working electrode or bioanode unless stated otherwise. The charge/discharge performance of the hybrid devices in air-equilibrated buffer solution containing 20 mM glucose was examined with an Autolab PGSTAT100 potentiostat (Eco Chimie, Netherlands) using the biocathode as working electrode and the bioanode as a combined counter/reference electrode. Testing of the devices involved the test sequence: (i) charging at open-circuit mode using the BFC component and (ii) galvanostatic discharge of the capacitor at various current densities (Figure 4C).

3.2.5. Calculation of device performance The specific capacitance of an individual electrode in a three-electrode system was obtained from the current flowing in a region where no faradaic processes were occurring (Eq. 3.1). The specific capacitance of the assembled supercapacitor was obtained from galvanostatic discharge curves (Eq. 3.2).  The specific capacitance, C (unit: μF cm-2), can be obtained from the current using equation (3.1):

(푗 −푗 ) C = 푎 푐 (3.1) 퐶푉 (2×푣)

- 62 -

Chapter 3

-2 where ja and jc are the anodic and cathodic current densities (μA cm ), respectively, obtained in the potential range where no faradaic processes were occurring, and v the scan rate in V s-1.

 From galvanostatic discharge curves, the specific capacitance can be obtained using equation (3.2):

푗푝푢푙푠푒 C푔 = ∆V (3.2) [−( )] ∆t -2 where jpulse is the applied current density (in μA cm ) and ΔV/Δt is the slope of the discharge curve after the voltage drop (in V s-1).  For an assembled supercapacitor, the overall capacitance, C, is determined by:

1 1 1 = + (3.3) 퐶 퐶1 퐶2

where C1 and C2 are the individual capacitances of the electrode.

 The equivalent series resistance (ESR) in Ω, can be calculated according to Ohm's law: ∆V ESR = 표ℎ푚𝑖푐 (3.4) 푖푝푢푙푠푒

where ipulse is applied current (in A) and ∆Vohmic is the voltage drop in V.

-2  The maximum power density for each galvanostatic discharge, Pmax, in μW cm can be determined by:

P푚푎푥 = 푗푝푢푙푠푒 × 푉푚푎푥 (3.5)

where Vmax is the instant potential at the end of voltage drop and the beginning of each discharge.

 To describe the relationship between Vmax ,ESR and ∆Vohmic, one can conclude that:

V푐푢푡표푓푓 = V푚푎푥 + ∆V표ℎ푚푖푐 (3.6)

V푐푢푡표푓푓 = V푚푎푥 + ESR × i푝푢푙푠푒 (3.7)

- 63 -

Chapter 3

where Vcutoff is the maximum voltage attained for each reset.

The relationship between jpulse and Pmax can be expressed by combining Eq. (3.5) and (3.7):

P푚푎푥 = 푗푝푢푙푠푒 × (푉cutoff − 퐸푆푅 × 푗푝푢푙푠푒 × 퐴) (3.8) where A is the geometric surface area of the electrode in cm2.

3.3. Results and discussion

3.3.1 Electrochemical characterisation of the capacitive bioelectrodes

Figure 3.1 Cyclic voltammograms (CVs) of various electrodes (deposition time: 300 s).

Cyclic voltammograms (CVs) displayed an onset potential of 0.7 V (vs. SCE) for the growth of PEDOT on the NPG electrode in an aqueous solution [43]. The potentiostatic pulse comprised an anodic potential of 0.9 V (2 s) to generate the radical cation and a cathodic potential of -0.4 V (3 s) to enable the EDOT concentration in the proximity of the electrode surface to return to that in the bulk state, thus allowing polymer formation on the electrode surface [45]. In the presence of Os poly(N-vinylimidazole) redox polymer, the weakly coordinated chloride ions exchanged with more strongly coordinating pyridine or imidazole groups on proximal

- 64 -

Chapter 3 chains when Os3+ was reduced to Os2+ during the cathodic pulse [44]. This crosslinking effect led to irreversible polymer precipitation onto the electrode. The resting period at the anodic potential enabled the reestablishment of the bulk concentration of precursor at the electrode surface. Overall, the potential sequence led to the alternate deposition of PEDOT and Os(bpy)2PVI, which was confirmed by electrochemical studies (Figure 3.1). Enzymes in the deposition solution were physically and/or coordinately entrapped into the resulting films.

Figure 3.2 CVs of (B) NPG/PEDOT/Os(bpy)2PVI/FAD-GDH and (C)

-1 NPG/PEDOT/Os(bpy)2PVI/BOx electrodes at a scan rate of 5 mV s .

Table 3.1 Electrochemical capacitances of various modified electrodes (obtained from Figure 3.1A). Capacitancea Electrode Factor (μF cm-2)

Bare Au 27 1 Bare NPG 258 9.6 NPG/PEDOT 821 30.4

NPG/Os(bpy)2PVI/FAD-GDH 480 17.8

NPG/PEDOT/Os(bpy)2PVI/FAD- 42.5 1148 GDH a: calculated by Eq. 3.1.

- 65 -

Chapter 3

CVs of various modified electrodes (deposition time of 300 s) in PBS at 100 mV s-1 were compared to confirm the successful electrodeposition of the polymers

(Figure 3.1). CVs of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH electrodes showed the

2+/3+ faradaic redox reaction of Os (∆Ep of 76 mV) superimposed on the charge/discharge capacitive currents. The Os polymer modified NPG electrode without PEDOT displayed a pair of reversible redox peaks with a peak separation of 20 mV. NPG/PEDOT exhibited a rectangular charge/discharge curve without any redox peaks. Table 3.1 compares CV derived specific capacitances that are normalised with respect to the projected surface area. Bare NPG showed a 9.6-fold higher capacitance than that possible with the bare planar gold electrode, consistent with the surface roughness factor (the ratio between the electrochemically addressable and geometric surface areas) obtained from the outermost layer of Au oxide stripping (a specific charge of 390 μC cm-2 is required for gold oxide reduction [55]).

NPG/PEDOT and NPG/PEDOT/Os(bpy)2PVI had 3.2 and 4.4 times higher capacitance than that of bare NPG.

Figure 3.3 TEM of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (450 s deposition).

- 66 -

Chapter 3

Figure 3.4 SEM images of the bare NPG (A) and

NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (300 s deposition) (B).

Figure 3.5 Optimisation curves for the anode (A) and cathode (B).

The amount of deposited hybrid polymer, with the associated increase in the capacitance, and the enzyme loading increased with potential cycling before levelling off after a number of cycles (200 cycles for the case of [44]). On increasing the deposition time, the resulting film tended to block the pores (Figure 3.3) of the NPG electrode (Figure 3.4). For the FAD-GDH modified electrode, a deposition time of 300 s exhibited the optimal response to 10 mM glucose (Figure 3.5), attributed to a compromise between loading of biocatalyst and mass transport of substrate through the film. For the BOx modified electrode, a shorter pulse duration of 150 s afforded the highest electrocatalytic response to oxygen, with relatively low capacitance

- 67 -

Chapter 3

(Figure 3.5). A pulse of 300 s duration was chosen as a compromise between the electrochemical response and capacitance.

Both bioelectrodes were separately studied in detail at a scan rate of 5 mV s-1

(Figure 3.2A and B). As can be seen from Fig. 1B, NPG/PEDOT/Os(bpy)2PVI/FAD-

GDH displayed a pair of redox peaks with a midpoint potential, Em, of +191 mV (vs. SCE), in agreement with the reduction-oxidation of the Os2+/3+ couple. The ratio of the integrated area of the anodic to cathodic peak was ca. 1.1. The variation of peak current with scan rate was linear, indicative of a surface controlled process (Figure 3.7). In the presence of 10 mM glucose, a sigmoidal catalytic wave with an onset potential of -18±9 mV vs. SCE and a background-corrected limiting current density of 59.7±2.4 μA cm-2 was observed. These results were indicative of the successful immobilisation of FAD-GDH with high activity.

Figure 3.6 (A) CVs of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (300 s deposition) at various scan rates. (B) Calibration plots of the oxidation/reduction peak current vs. scan rate.

- 68 -

Chapter 3

app The apparent Michaelis-Menten constant, KM , of the enzyme modified electrode was 7.9 mM (Figure 3.7), lower than the value of 17.4 mM obtained from the same enzyme when chemically crosslinked onto graphite electrode [48]. This decrease may arise from improved substrate transport through the thin immobilizing layer. BOx based cathodes also showed a pair of redox peaks in N2 bubbled PBS (Figure 3.2B). The ratio of the integrated area of the anodic to cathodic peak was however less than 1, due to competition with residual O2 for the oxidation of BOx. The Em was +202 mV, a slight increase in comparison to that of FAD-GDH modified electrodes. In O2 bubbled solution, electrocatalytic reduction commenced at 387±13 mV and reached a maximum net catalytic current density of 65.2±4.5 μA cm-2.

Figure 3.7 Lineweaver–Burk plot for the NPG based FAD-GDH bioelectrode.

3.3.2 Morphology characterisation

NPG and NPG/PEDOT/Os(bpy)2PVI/FAD-GDH electrodes were examined by SEM (Figure 3.8A and B). Typical porous structures with bicontinuous pores/ligaments of NPG were observed. The average pore size was 30.6±4.7 nm for the bare NPG (Figure 3.8A and Figure 3.4A). The deposited layer uniformly grew along the pore surfaces, making the pores smaller and ligaments thicker, but not plugging the pores. The core-shell structure was clearly observed by TEM (Figure 3.8B), with the contrast between the modified film and the gold support clearly visible. The spatially homogeneous film was 7.4±1.4 nm in thickness, a size sufficient to encapsulate the enzyme.

- 69 -

Chapter 3

Figure 3.8 (A) STEM dark-field micrograph of bare NPG at 300 kX. (B)

TEM image of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (300 s deposition).

3.3.3 Hybrid device testing

NPG/PEDOT/Os(bpy)2PVI/FAD-GDH and NPG/PEDOT/Os(bpy)2PVI/BOx electrodes were subsequently assembled into a dual-functioning device comprising a

BFC and a capacitor. This type of device can perform as a glucose/O2 BFC when connected to a load in an external circuit (Figure 3.9A). The polarisation curve of the BFC was obtained with linear sweep voltammetry at a scan rate of 1 mV s-1, with the power curve calculated accordingly (Figure 3.9B). The BFC registered an OCV of 459.6±9.5 mV, a maximum current density of 28.9 µA cm-2, and a maximum power

-2 density of 1.3 µW cm at a potential of 0.09 V in O2 bubbled PBS containing 20 mM glucose. The assembled cell can also act as a supercapacitor, whose performance was examined by galvanostatic charge/discharge at a given external current density of 10 µA cm-2 (Fig. S7). A specific capacitance of 391.9±2.1 µF cm-2 was obtained (Eq. 3.2). The total capacitance of the supercapacitor is determined by the series

- 70 -

Chapter 3 connection of the two capacitive electrodes (Eq. 3.3) [56], leading to a lower overall capacitance compared with those of individual electrodes.

Figure 3.9 (A) Schematic diagram of the BFC. (B) Polarisation and power curve for the BFC consisting of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH bioanode and

NPG/PEDOT/Os(bpy)2PVI/BOx biocathode.

Recent reports described the underlying mechanism of a hybrid supercapacitor/ microbial fuel cell [57, 58]. The integration of a BFC with a capacitor enables the hybrid device to work as a self-powered capacitor, without the requirement for external input. In rest conditions, i.e. in open-circuit, the cell voltage tended to the equilibrium potential, i.e. OCV of the BFC. The existing potential difference between the two electrodes polarised the anode and cathode, leading the NPG backbones to be negatively or positively charged, respectively, and triggering the p-dopable PEDOT film to insert/deinsert anions (Figure 3.10A). In other words, the capacitive cell was electrostatically charged at the thermodynamically induced potential difference, driving its voltage profile close to the value of OCV. As shown in Figure 3.10C, the voltage increased with time initially rising rapidly before levelling off with time.

- 71 -

Chapter 3

Figure 3.10 Schematic diagrams of the hybrid device working at the self- charging (A) and galvanostatic discharging mode (B) (with simplified charge- discharge description on the capacitive NPG/PEDOT hybrid). (C) Charge/discharge curves of the as-assembled biocapacitor (black line); Experimental setup: reset at open-circuit and cutoff at 0.4 V, followed by discharging at 0.2 mA cm-2 for 0.5 s (red line). (D) Magnified image the first discharge segment.

The energy stored in the biocapacitor could be subsequently discharged at desired currents by releasing ions (Figure 3.10B). As can be seen in Figure 3.9C and D, a galvanostatic discharging current density of 0.2 mA cm-2, almost 7 times higher than the 28.9 µA cm-2 possible with the BFC mode, resulted in a rapid release of power. In the following cycle, the rest step at open-circuit mode without any external load enabled the recovery of the cell potential to OCV (0.45 V) of the BFC. The following cycles almost overlapped, indicative of the excellent stability (Figure 3.10C). On a closer examination of the first discharging segment (Figure 3.10D), a

-2 capacitance of 357 µF cm was calculated by dividing the given current density, jpulse, by the absolute value of the slope of the discharging curve (Eq. 3.2). A jpulse

- 72 -

Chapter 3 dependent voltage drop of 11 mV was observed due to the internal resistance (Eq. 3.4), probably assigned to the ohmic resistance of electrode, mass/charge transfer resistance, and/or low intrinsic biocatalytic activity [59]. The resistance was predominantly attributed to the internal resistance from the capacitor, instead of the biocatalytic processes, as the cell also showed a voltage drop when used only as a capacitor (Figure 3.11).

Figure 3.11 Galvanostatic charge/discharge at 10 µA cm-2 of the symmetric supercapacitor consisting of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH and NPG/PEDOT/Os(bpy)2PVI/BOx in a blank PBS solution.

The long-term operation (7 hours, 50 cycles) of the hybrid device was tested by recording the potential at the open-circuit with a cutoff at 0.4 V, followed by discharge at 0.2 mA cm-2 for 0.5 s (test sequence is shown as the red line of Figure 3.9C and D). For a period of 7 hours (50 cycles) (Figure 3.12A), the discharge finishing potentials remained constant at 0.07 V, demonstrating the stable capacitance of the supercapacitor for each discharge cycle. The reset time did increase, e.g. ca. 300 and 800 s for the first and final cycles, respectively. After ca. 7 hours of operation, the device exhibited a loss of 70% in maximum power density (0.38 µW cm-2) (Figure 3.13A) when tested as a BFC. Cyclic voltammograms of

- 73 -

Chapter 3

PEDOT/Os(bpy)2PVI showed little change, indicative of a stable modification layer (Figure 3.13B). Thus, decreased enzymatic activity, in particular of the FAD-GDH based bioanode (data not shown), was responsible for the extended self-charge time.

Figure 3.12 (A) Charge/discharge curves of the biocapacitor for 50 cycles; Experimental setup: reset at open-circuit and cutoff at 0.4 V, followed by discharging at 0.2 mA cm-2 for 0.5 s. (B) Charge/discharge curves of the biocapacitor upon various discharging current densities; Experimental setup: reset at open-circuit and cutoff at 0.4 V, followed by discharging at 0.005 (a), 0.01 (b), 0.02 (c), 0.05 (d), 0.1 (e), 0.2 (f), 0.5 (g), 1 (h), 2 (i) mA cm-2 for 0.2 s.

- 74 -

Chapter 3

Figure 3.13 (A) Polarisation and power curves for the BFC (initial curves: solid line; after long-term operation: dashed line). (B) CVs of the bioanode in PBS before and after long-term testing.

Table 3.2 Cell performance of the hybrid device upon various discharging

current densities (jpulse) (obtained from Figure 3.11B).

b Discharging Vmax Pmax

-2 jpulse (mV) (μW cm ) ( μA 5cm -2) 397.3 2.0 10 397.1 4.0 20 396.6 7.9 50 395.2 19.8 100 392.9 39.3 200 388.5 77.7 500 374.3 187.1 1000 352.1 352.1 2000 304.4 608.8 b: calculated by Eq. 3.5.

The discharge capability of the biocapacitor at various current densities was examined (Figure 3.12B), with a current density up to 2 mA cm-2. Generally, a larger discharge current density provided a larger power density (Eq. 3.8), as well as a longer recovery time. Table 3.2 compares the instant maximum power densities that

- 75 -

Chapter 3 can be delivered. For example, power pulses of 352 and 609 μW cm-2 at 1 and 2 mA cm-2 were achieved, 271 and 468 times higher than that obtained from a traditional

-2 BFC configuration (1.3 μW cm ). The maximum voltage output, Vmax, decreased with higher pulse current densities due to the potential loss caused by the equivalent series resistance (ESR) [57] (Eq. 3.7, Table 3.2). As a result, doubling the current density did not result in the same increase in the maximum power density. Decreasing the ESR could improve the maximum power density [57].

Figure 3. 14 (A) Polarisation and power curve for a planar Au based BFC. (B) Long-term testing of the biocapacitor for 50 cycles; Experimental setup: reset at OCV and cutoff at 0.35 V, followed by discharging at 0.2 mA cm-2 for 0.5 s and cutoff at 0 V.

To highlight the important role of the NPG substrate, a planar Au based hybrid electrode system was constructed using the same conditions. Au based BFC displayed a poor performance with an OCV of 365 mV, a maximum current density of 1.5 µA cm-2, and a maximum power density of 0.08 µW cm-2 at a potential of 0.11

- 76 -

Chapter 3

V in O2 bubbled PBS containing 20 mM glucose (Figure 3.14A). The internal resistance was larger, leading to a voltage drop of 84 mV (Figure 3.14B). A specific capacitance of 31.6 µF cm-2, 11 times lower than that reported on the NPG based device, was estimated.

3.3.4 A proof-of-concept pulse generator A cardiac pacemaker possesses dual-function of sensing and pacing the heart [60]. To be able to pace the heart, an electric stimulus generated by the pulse generator with a fixed pulse potential and width (i.e. threshold, the minimum voltage from the pacemaker to initiate a heartbeat) is required. Previous attempts proposed the possibility of using BFCs as power sources to replace lithium based batteries currently used in pacemakers [6, 13]. To increase the voltage to the required value, three individual cells were connected in series, giving a potential of ca. 1.24 V for 2 hours in the presence of 20 mM glucose. A test sequence of 5 s reset (i.e. a frequency of 0.2 Hz) and 0.5 ms discharge at 10 μA that matched typical pacemaker working characteristic was applied [61]. As shown in Figure 3.14A, the voltage dropped steadily to ca. 0.8 V in the initial 200-250 cycles, as 5 s was not long enough for voltage recovery (inset of Figure 3.15A), and then maintained at ca. 0.74 V in the following long-term testing cycles. Such an output voltage stabilised at 0.74 V is enough to exceed the mean pacing threshold (e.g. 0.51±0.22 V reported previously [62]). Upon refilling the solution, the series connected cells recovered to a potential of ca. 1.21 V after a 2-hour incubation period, followed by 1300 cycles of charge/discharge (Figure 3.15B). The output voltage gradually decreased for the initial 200 cycles due to the relatively short reset time of 5 s and attained a stable value of 0.7 V after the 220th cycle. The results demonstrate that three cells connected in series can mimic a pacemaker generating 0.2 Hz pulses (10 μA, 0.5 ms) with a stable output potential of 0.7 V.

- 77 -

Chapter 3

Figure 3.15 Charge/discharge curves of the series connection of three biofuel cells (see Fig. S9); Experimental setup: the connected cells were allowed to reset at open-circuit for 2 hours, followed by discharging at 10 μA for 0.5 ms every 5 s reset. (A) is the first measurement of 1500 discharging pulses; (B) is for the measurement of 1300 discharging pulses upon refilling of fresh solutions; insets show zooms at the specific cycles.

3.4. Conclusions

A supercapacitor/enzymatic biofuel cell hybrid device was prepared by a facile, one-step electrodeposition of PEDOT/Os polymer/enzyme onto dealloyed nanoporous gold electrodes. The dual-function properties of this hybrid device allowed the energy yielded by the biofuel cell to be stored in the supercapacitor and delivered at a significantly high power pulse. For instance, it permitted a pulse current density of 2 mA cm-2, with an instant maximum power density of 609 μW cm-2, 468 times higher than that of the BFC. The modification layer showed reasonable stability without visible leakage of the redox mediators after 50 cycles operation at 0.2 mA cm- 2 for approximately 7 hours. In contrast to the planar Au based system, nanoporous gold electrodes improved the performance in terms of lower resistance, higher - 78 -

Chapter 3 bioelectrochemical signal and capacitance. A proof-of-concept pulse generator (0.2 Hz pulse at 10 μA for 0.5 ms) to mimic a pacemaker was demonstrated using electrodes connected in series.

3.5 References

[1] S. Calabrese Barton, J. Gallaway, P. Atanassov, Enzymatic Biofuel Cells for Implantable and Microscale Devices, Chem. Rev., 104 (2004) 4867-4886. [2] D. Leech, P. Kavanagh, W. Schuhmann, Enzymatic fuel cells: Recent progress, Electrochim. Acta, 84 (2012) 223-234. [3] M. Cooney, V. Svoboda, C. Lau, G. Martin, S. Minteer, Enzyme catalysed biofuel cells, Energy Environ. Sci., 1 (2008) 320-337. [4] P. Pinyou, F. Conzuelo, K. Sliozberg, J. Vivekananthan, A. Contin, S. Pöller, N. Plumeré, W. Schuhmann, Coupling of an enzymatic biofuel cell to an electrochemical cell for self-powered glucose sensing with optical readout, Bioelectrochem., 106, Part A (2015) 22-27. [5] A. Zloczewska, A. Celebanska, K. Szot, D. Tomaszewska, M. Opallo, M. Jönsson-Niedziolka, Self-powered biosensor for ascorbic acid with a Prussian blue electrochromic display, Biosens. Bioelectron., 54 (2014) 455-461. [6] K. MacVittie, J. Halamek, L. Halamkova, M. Southcott, W.D. Jemison, R. Lobel, E. Katz, From "cyborg" lobsters to a pacemaker powered by implantable biofuel cells, Energy Environ. Sci., 6 (2013) 81-86. [7] M. Falk, V. Andoralov, Z. Blum, J. Sotres, D.B. Suyatin, T. Ruzgas, T. Arnebrant, S. Shleev, Biofuel cell as a power source for electronic contact lenses, Biosens. Bioelectron., 37 (2012) 38-45. [8] P. Ó Conghaile, M. Falk, D. MacAodha, M.E. Yakovleva, C. Gonaus, C.K. Peterbauer, L. Gorton, S. Shleev, D. Leech, Fully Enzymatic Membraneless Glucose|Oxygen Fuel Cell That Provides 0.275 mA cm–2 in 5 mM Glucose, Operates in Human Physiological Solutions, and Powers Transmission of Sensing Data, Anal. Chem., 88 (2016) 2156-2163. [9] D. Pankratov, E. González-Arribas, Z. Blum, S. Shleev, Tear Based Bioelectronics, Electroanalysis, 28 (2016) 1250-1266. [10] J.A. Cracknell, K.A. Vincent, F.A. Armstrong, Enzymes as Working or Inspirational Electrocatalysts for Fuel Cells and Electrolysis, Chem. Rev., 108 (2008) 2439-2461. [11] M. Rasmussen, S. Abdellaoui, S.D. Minteer, Enzymatic biofuel cells: 30 years of critical advancements, Biosens. Bioelectron., 76 (2015) 91-102. [12] V. Andoralov, M. Falk, D.B. Suyatin, M. Granmo, J. Sotres, R. Ludwig, V.O. Popov, J. Schouenborg, Z. Blum, S. Shleev, Biofuel cell based on microscale nanostructured electrodes with inductive coupling to rat brain neurons, Sci. Rep., 3 (2013) doi:10.1038/srep03270.

- 79 -

Chapter 3

[13] M. Southcott, K. MacVittie, J. Halamek, L. Halamkova, W.D. Jemison, R. Lobel, E. Katz, A pacemaker powered by an implantable biofuel cell operating under conditions mimicking the human blood circulatory system - battery not included, Phys. Chem. Chem. Phys., 15 (2013) 6278-6283. [14] M. Falk, M. Alcalde, P.N. Bartlett, A.L. De Lacey, L. Gorton, C. Gutierrez-Sanchez, R. Haddad, J. Kilburn, D. Leech, R. Ludwig, E. Magner, D.M. Mate, P.Ó. Conghaile, R. Ortiz, M. Pita, S. Pöller, T. Ruzgas, U. Salaj-Kosla, W. Schuhmann, F. Sebelius, M. Shao, L. Stoica, C. Sygmund, J. Tilly, M.D. Toscano, J. Vivekananthan, E. Wright, S. Shleev, Self-Powered Wireless Carbohydrate/Oxygen Sensitive Biodevice Based on Radio Signal Transmission, PLoS ONE, 9 (2014) e109104. [15] K. Murata, K. Kajiya, N. Nakamura, H. Ohno, Direct electrochemistry of bilirubin oxidase on three-dimensional gold nanoparticle electrodes and its application in a biofuel cell, Energy Environ. Sci., 2 (2009) 1280-1285. [16] E. Suraniti, O. Courjean, S. Gounel, E. Tremey, N. Mano, Uncovering and Redesigning a Key Amino Acid of Glucose Oxidase for Improved Biotechnological Applications, Electroanalysis, 25 (2013) 606-611. [17] V. Flexer, N. Brun, O. Courjean, R. Backov, N. Mano, Porous mediator- free enzyme carbonaceous electrodes obtained through Integrative Chemistry for biofuel cells, Energy Environ. Sci., 4 (2011) 2097-2106. [18] Y.H. Kim, E. Campbell, J. Yu, S.D. Minteer, S. Banta, Complete oxidation of methanol in biobattery devices using a hydrogel created from three modified dehydrogenases, Angew. Chem. Int. Ed., 52 (2013) 1437-1440. [19] S. Xu, S.D. Minteer, Enzymatic Biofuel Cell for Oxidation of Glucose to CO2, ACS Catal., 2 (2012) 91-94. [20] M. Shao, M. Nadeem Zafar, C. Sygmund, D.A. Guschin, R. Ludwig, C.K. Peterbauer, W. Schuhmann, L. Gorton, Mutual enhancement of the current density and the coulombic efficiency for a bioanode by entrapping bi-enzymes with Os-complex modified electrodeposition paints, Biosens. Bioelectron., 40 (2013) 308- 314. [21] T. Hanashi, T. Yamazaki, W. Tsugawa, S. Ferri, D. Nakayama, M. Tomiyama, K. Ikebukuro, K. Sode, BioCapacitor—A novel category of biosensor, Biosens. Bioelectron., 24 (2009) 1837-1842. [22] K. Sode, T. Yamazaki, I. Lee, T. Hanashi, W. Tsugawa, BioCapacitor: A novel principle for biosensors, Biosens. Bioelectron., 76 (2016) 20-28. [23] M. Winter, R.J. Brodd, What Are Batteries, Fuel Cells, and Supercapacitors?, Chem. Rev., 104 (2004) 4245-4270. [24] M. Skunik-Nuckowska, K. Grzejszczyk, K. Stolarczyk, R. Bilewicz, P.J. Kulesza, Integration of supercapacitors with enzymatic biobatteries toward more effective pulse-powered use in small-scale energy harvesting devices, J. Appl. Electrochem., 44 (2014) 497-507. [25] C. Agnes, M. Holzinger, A. Le Goff, B. Reuillard, K. Elouarzaki, S. Tingry, S. Cosnier, Supercapacitor/biofuel cell hybrids based on wired enzymes on

- 80 -

Chapter 3 carbon nanotube matrices: autonomous reloading after high power pulses in neutral buffered glucose solutions, Energy Environ. Sci., 7 (2014) 1884-1888. [26] M. Kizling, K. Stolarczyk, J.S.S. Kiat, P. Tammela, Z. Wang, L. Nyholm, R. Bilewicz, Pseudocapacitive polypyrrole–nanocellulose composite for sugar-air enzymatic fuel cells, Electrochem. Commun., 50 (2015) 55-59. [27] D. Pankratov, Z. Blum, D.B. Suyatin, V.O. Popov, S. Shleev, Self- Charging Electrochemical Biocapacitor, ChemElectroChem, 1 (2014) 343-346. [28] M. Kizling, S. Draminska, K. Stolarczyk, P. Tammela, Z. Wang, L. Nyholm, R. Bilewicz, Biosupercapacitors for powering oxygen sensing devices, Bioelectrochem., 106, Part A (2015) 34-40. [29] E. González-Arribas, D. Pankratov, S. Gounel, N. Mano, Z. Blum, S. Shleev, Transparent and Capacitive Bioanode Based on Specifically Engineered Glucose Oxidase, Electroanalysis, 28 (2016) 1290-1297. [30] C.W. Narvaez Villarrubia, F. Soavi, C. Santoro, C. Arbizzani, A. Serov, S. Rojas-Carbonell, G. Gupta, P. Atanassov, Self-feeding paper based biofuel cell/self-powered hybrid μ-supercapacitor integrated system, Biosens. Bioelectron., 86 (2016) 459-465. [31] K.L. Knoche, D.P. Hickey, R.D. Milton, C.L. Curchoe, S.D. Minteer, Hybrid Glucose/O2 Biobattery and Supercapacitor Utilizing a Pseudocapacitive Dimethylferrocene Redox Polymer at the Bioanode, ACS Energy Lett., (2016) 380- 385. [32] D. Pankratov, F. Conzuelo, P. Pinyou, S. Alsaoub, W. Schuhmann, S. Shleev, A Nernstian Biosupercapacitor, Angew. Chem. Int. Ed., 55 (2016) 15434- 15438. [33] S. Alsaoub, A. Ruff, F. Conzuelo, E. Ventosa, R. Ludwig, S. Shleev, W. Schuhmann, An Intrinsic Self-Charging Biosupercapacitor Comprised of a High- Potential Bioanode and a Low-Potential Biocathode, ChemPlusChem, 82 (2017) 576- 583. [34] A. Magrez, S. Kasas, V. Salicio, N. Pasquier, J.W. Seo, M. Celio, S. Catsicas, B. Schwaller, L. Forró, Cellular Toxicity of Carbon-Based Nanomaterials, Nano Lett., 6 (2006) 1121-1125. [35] T. Miyake, K. Haneda, N. Nagai, Y. Yatagawa, H. Onami, S. Yoshino, T. Abe, M. Nishizawa, Enzymatic biofuel cells designed for direct power generation from biofluids in living organisms, Energy Environ. Sci., 4 (2011) 5008-5012. [36] Y. Ding, Y.J. Kim, J. Erlebacher, Nanoporous Gold Leaf: “Ancient Technology”/Advanced Material, Adv. Mater., 16 (2004) 1897–1900. [37] F. Meng, Y. Ding, Sub-micrometer-thick all-solid-state supercapacitors with high power and energy densities, Adv. Mater., 23 (2011) 4098-4102. [38] X. Lang, A. Hirata, T. Fujita, M. Chen, Nanoporous metal/oxide hybrid electrodes for electrochemical supercapacitors, Nat. Nanotechnol., 6 (2011) 232–236. [39] M.D. Scanlon, U. Salaj-Kosla, S. Belochapkine, D. MacAodha, D. Leech, Y. Ding, E. Magner, Characterization of nanoporous gold electrodes for bioelectrochemical applications, Langmuir, 28 (2012) 2251-2261.

- 81 -

Chapter 3

[40] X. Xiao, J. Ulstrup, H. Li, J. Zhang, P. Si, Nanoporous gold assembly of glucose oxidase for electrochemical biosensing, Electrochim. Acta, 130 (2014) 559- 567. [41] X. Xiao, E. Magner, A biofuel cell in non-aqueous solution, Chem. Commun., 51 (2015) 13478-13480. [42] X. Xiao, P. Si, E. Magner, An overview of dealloyed nanoporous gold in bioelectrochemistry, Bioelectrochem., 109 (2016) 117-126. [43] X. Xiao, M.e. Wang, H. Li, P. Si, One-step fabrication of biofunctionalized nanoporous gold/poly (3, 4-ethylenedioxythiophene) hybrid electrodes for amperometric glucose sensing, Talanta, 115 (2013) 1054-1059. [44] Z. Gao, G. Binyamin, H.H. Kim, S.C. Barton, Y. Zhang, A. Heller, Electrodeposition of Redox Polymers and Co‐Electrodeposition of Enzymes by Coordinative Crosslinking, Angew. Chem. Int. Ed., 41 (2002) 810-813. [45] W. Schuhmann, C. Kranz, H. Wohlschläger, J. Strohmeier, Pulse technique for the electrochemical deposition of polymer films on electrode surfaces, Biosens. Bioelectron., 12 (1997) 1157-1167. [46] K. Habermüller, A. Ramanavicius, V. Laurinavicius, W. Schuhmann, An Oxygen-Insensitive Reagentless Glucose Biosensor Based on Osmium-Complex Modified Polypyrrole, Electroanalysis, 12 (2000) 1383-1389. [47] W. Shen, H. Deng, A.K.L. Teo, Z. Gao, An electrodeposited redox polymer–laccase composite film for highly efficient four-electron oxygen reduction, J. Power Sources, 226 (2013) 27-32. [48] M.N. Zafar, N. Beden, D. Leech, C. Sygmund, R. Ludwig, L. Gorton, Characterization of different FAD-dependent glucose dehydrogenases for possible use in glucose-based biosensors and biofuel cells, Anal. Bioanal. Chem., 402 (2012) 2069-2077. [49] R.D. Milton, K. Lim, D.P. Hickey, S.D. Minteer, Employing FAD- dependent glucose dehydrogenase within a glucose/oxygen enzymatic fuel cell operating in human serum, Bioelectrochem., 106, Part A (2015) 56-63. [50] P.A. Jenkins, S. Boland, P. Kavanagh, D. Leech, Evaluation of performance and stability of biocatalytic redox films constructed with different copper oxygenases and osmium-based redox polymers, Bioelectrochem., 76 (2009) 162-168. [51] C. Sygmund, P. Staudigl, M. Klausberger, N. Pinotsis, K. Djinović- Carugo, L. Gorton, D. Haltrich, R. Ludwig, Heterologous overexpression of Glomerella cingulata FAD-dependent glucose dehydrogenase in Escherichia coli and Pichia pastoris, Microb. Cell Fact., 10 (2011) 1-9. [52] S. Fabiano, C.Tran-Minh, B. Piro, L. A. Dang, M. C. Pham, O. Vittori, Poly 3,4-ethylenedioxythiophene as an entrapment support for amperometric enzyme sensor, Mater. Sci. Eng., C, 21 (2002) 61–67. [53] S. Reiter, K. Habermüller, W. Schuhmann, A reagentless glucose biosensor based on glucose oxidase entrapped into osmium-complex modified polypyrrole films, Sens. Actuat. B: Chem., 79 (2001) 150-156.

- 82 -

Chapter 3

[54] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, J. Schindelin, I. Arganda- Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, NIH image to imageJ: 25 years of image analysis, Nat. Methods, 9 (2012) 671. [55] S. Trasatti, O. A. Petrii, Real surface area measurements in electrochemistry, Pure Appl. Chem., 63 (1991) 711-734. [56] V. Khomenko, E. Frackowiak, F. Béguin, Determination of the specific capacitance of conducting polymer/nanotubes composite electrodes using different cell configurations, Electrochim. Acta, 50 (2005) 2499-2506. [57] C. Santoro, F. Soavi, A. Serov, C. Arbizzani, P. Atanassov, Self- powered supercapacitive microbial fuel cell: The ultimate way of boosting and harvesting power, Biosens. Bioelectron., 78 (2016) 229-235. [58] D. Pankratov, Z. Blum, S. Shleev, Hybrid Electric Power Biodevices, ChemElectroChem, 1 (2014) 1798-1807. [59] P. Liang, X. Huang, M.-Z. Fan, X.-X. Cao, C. Wang, Composition and distribution of internal resistance in three types of microbial fuel cells, Appl. Microbiol. Biotechnol., 77 (2007) 551-558. [60] R.S. Sanders, The Pulse Generator, in: F.M. Kusumoto, N.F. Goldschlager (Eds.) Cardiac Pacing for the Clinician, Springer US, Boston, MA, 2008, pp. 47-71. [61] V.S. Mallela, V. Ilankumaran, N. Rao, Trends in Cardiac Pacemaker Batteries, Ind. Pacing Electrophys. J., 4 (2004) 201-212. [62] P. Ritter, G.Z. Duray, C. Steinwender, K. Soejima, R. Omar, L. Mont, L.V. Boersma, R.E. Knops, L. Chinitz, S. Zhang, C. Narasimhan, J. Hummel, M. Lloyd, T.A. Simmers, A. Voigt, V. Laager, K. Stromberg, M.D. Bonner, T.J. Sheldon, D. Reynolds, Early performance of a miniaturized leadless cardiac pacemaker: the Micra Transcatheter Pacing Study, Eur. Heart. J., (2015).

- 83 -

CHAPTER 4:

An oxygen-independent and membrane-less glucose biobattery/supercapacitor hybrid device

Chapter 4

Chapter 4. An oxygen-independent and membrane-less glucose biobattery/supercapacitor hybrid device

4.1. Introduction

The use of enzymatic biofuel cells (EBFCs) is of promise in generating electricity from fuels [1, 2]. EBFCs function at physiological temperature and pH, in comparison to traditional fuel cells utilising abiotic catalysts which generally operate in harsh environments (e.g. strongly acidic or alkaline media). Immobilisation of enzymes at the anode and cathode can eliminate the requirement for membranes that are required in conventional fuel cells to separate the anode and cathode compartments. In vivo EBFCs utilising oxygen and glucose are of significant interest due to potential applications as miniaturised power sources for implantable medical devices [3] such as cardiac pacemakers [4] and insulin pumps. However, the successful application of autonomous biomedical devices is a significant challenge due to the requirements for high power density, biocompatibility and long lifetime [5]. The concentration of oxygen in vivo is significantly lower (0.14 mM in arterial blood and 0.08 mM in intestinal tissue [5, 6]) than that of glucose (3.3 and 4.8 mM in muscle and plasma, respectively [7]), together with possible mass transport limitation of oxygen, making oxygen reducing biocathode a significant limiting factor in the application of EBFCs. For example, the theoretical power output of an in vivo 1 cm long tubular glucose/oxygen EBFC is solely determined by the oxygen reduction reaction (ORR) at the cathode [8]. Moreover, the stability of the enzymes used, predominantly multi-copper oxidases such as laccase and bilirubin oxidase (BOx), needs to be considered. Laccase prefers a weakly acidic environment (ca. pH 4-5) and is inhibited by halide ions [9-12]. In comparison to laccases, BOx is more stable under physiological conditions (pH 7.4, no inhibition in the presence of Cl-). However, the operational stability of BOx based electrodes is limited, for example, an osmium - 83 -

Chapter 4 polymer “wired” Trachyderma tsunodae BOx displayed a current loss of 78% after 2 h rotation at 100 rpm, a loss that was mainly ascribed to the irreversible deactivation of BOx Cu-centers in the oxidised state [13]. Air-breathing biocathodes can be employed to circumvent limitations in the supply of oxygen, but can only be used in subcutaneous devices [14]. Recently, molecular oxygen-independent hybrid EBFCs or biobatteries relying on a combination of enzymatic anodes and solid-state cathodes have been proposed to address the underlying problems of enzymatic cathode based EBFCs. These abiotic cathodes utilise cheap and abundantly available materials such as Prussian Blue (PB)

[15], Ag2O/Ag [16] and MnO2 [16], which can be reduced/discharged via an external circuit, resulting in rechargeable biobatteries. For example, the oxidation of PB to Berlin Green (BG) occurs at a high potential of 0.87 V vs. SCE [17], exceeding the redox potentials of multi-copper oxidases. Minteer et al. developed a rechargeable ethanol biobattery based on an alcohol dehydrogenase (ADH) modified bioanode and a PB paste cathode that registered an open circuit voltage (OCV) of up to 1.2 V [15].

Dong et al. combined a glucose dehydrogenase (GDH) bioanode with an Ag2O/Ag

[16] or MnO2 cathode [16] to fabricate oxygen-independent recycled biobatteries with reported OCVs of 0.59 V and 0.43 V, respectively. Microbial biobatteries consisting of anodes colonized by microorganisms and reoxidisable solid-state cathodes such as

Ag2O/Ag [18] and PB [19] were stable, showing no loss of capacity over 20 cycles of operation [19]. Biofuel cell (BFC)/supercapacitor hybrid devices, or self-charging biocapacitors, utilising capacitive bioelectrodes are of great interest due to their ability to generate repeated electric pulses, with an instantaneous power density that is significantly higher than that from the BFC itself [20]. Biocapacitors taking advantage of enzymes [20-23], microbes [24] and thylakoids [5] have been presented. Recently, we described a supercapacitive EBFC prepared by the immobilisation of flavin adenine dinucleotide-dependent GDH (FAD-GDH) and BOx with electrodeposited poly(3,4-ethylenedioxythiophene) (PEDOT) and the redox polymer [Os(2,2′-

- 84 -

Chapter 4

+/2+ bipyridine)2(polyvinylimidazole)10Cl] (Os(bpy)2PVI) on dealloyed nanoporous gold (NPG) [25]. The device could operate as a pulse generator to mimic that in a cardiac pacemaker, producing 10 μA pulses for 0.5 ms at a frequency of 0.2 Hz. In this chapter, BOx biocathodes have been substituted with a non-enzymatic

MnO2 cathode to assemble an oxygen-independent glucose biobattery/supercapacitor hybrid device (Scheme 1). At neutral pH MnO2 only shows catalytic activity towards oxygen at negative potentials [26], outside the potential window needed in this work and is thus used as a consumed cathode. MnO2 has been selected based on several considerations: (i) a higher pseudo-capacitance in comparison to carbon materials

[27]. MnO2 is partially charged/discharged via the intercalation/deintercalation of electrolyte cations (e.g. Na+) and protons according to the reaction:

+ + − Mn(IV)O2 + 푥푁푎 + 푦퐻 + (푥 + 푦)푒 ↔

푀푛(퐼퐼퐼)(푥+푦)푀푛(퐼푉)1−(푥+푦)푂푂푁푎푥퐻푦 (1)

where 0 <(x+y)≤1. In this case, the discharged form is insoluble, avoiding issues with leakage. (ii) a moderate onset potential, resulting in a biobattery with a considerable OCV [16]. (iii) operation at neutral pH that is amenable to enzymes. (iv) inert to the oxidation of glucose, as confirmed by Dong et al. [16], resulting in a membrane-less biobattery.

In this chapter, MnO2 was grown onto NPG using anodic electrodeposition.

The effect of deposition time on the properties of the layer was evaluated. MnO2 deposited on NPG was more stable than that on a planar polycrystalline Au electrode. The additional stability is a result of confinement effects in the NPG support substrate. A spontaneous recovery of the potential of the discharged NPG/MnO2 was observed in open-circuit mode, similar to that reported with a pseudo-capacitive RuO2 electrode [28]. The assembled NPG/PEDOT/Os(bpy)2PVI/FAD-GDH//NPG/MnO2 biobattery/supercapacitor hybrid device delivered intermittent electric signals, with a power density much higher than that of the biobattery itself. MnO2 is thus believed to be a cost-efficient cathode candidate to overcome the constraints of biocathodes described earlier. - 85 -

Chapter 4

4.2. Experimental section

4.2.1. Materials and apparatus Sodium phosphate (monobasic dehydrate ≥99 % and dibasic ≥99 %), sodium sulfate (≥99.99 %), manganese(II) acetate tetrahydrate (99.99 %), D-(+)-glucose (99.5 %), 3,4-ethylenedioxythiophene (EDOT, 97%) were obtained from Sigma- Aldrich Ireland, Ltd. All solutions were prepared with deionised water (18.2 MΩ cm,

Elga Purelab Ultra, UK). Os(bpy)2PVI was prepared according to an established procedure [29, 30]. Oxygen-insensitive, recombinant Glomerella cingulata FAD- GDH (EC 1.1.99.10, D-glucose: acceptor 1-oxidoreductase) was expressed in Pichia pastoris and purified with a specific activity of 572 U mg-1 [31]. Dealloyed NPG leaves were obtained by floating ca. 100 nm thick Au/Ag leaves (12-carat, Eytzinger, Germany) on concentrated HNO3 (Sigma-Aldrich) for 30 min at 30 °C [32, 33]. And then placed on well-polished glassy carbon electrodes (GCEs, diameter: 4mm). The NPG electrodes were cleaned by scanning the potential

-1 over the range of -0.2 to 1.65 V in 1 M H2SO4 at a scan rate of 100 mV s for 15 cycles. Scanning electron microscopy (SEM) images were collected using a Hitachi SU-70 microscope (operating at 15 kV), equipped with an energy dispersive X-ray spectroscopy (EDX). Transmission electron microscopy (TEM, JEOL JEM-2100, operating voltage of 200 kV) images of the electrodes were obtained on samples mounted on 300-mesh copper grids (S147-3, Agar Scientific, UK). The average pore size and layer thickness were measured with ImageJ software (National Institutes of Health, Bethesda, Maryland) [34] using at least 30 measurement points. Raman spectra of MnO2 deposited on gold foils (thickness: 0.1 mm, purity: 99.9%) were recorded with a LabRAM 300 Raman spectrometer (Horiba Jobin Yvon) using an excitation source at 514 nm (Ar laser).

- 86 -

Chapter 4

4.2.2. Preparation of the enzyme modified anode and NPG/MnO2 cathode Electrodeposition was performed in solutions containing phosphate buffer solution (PBS, 0.1 M pH 7.0) containing 2 mM polyethylene glycol 3400 (PEG3400),

-1 -1 20 mM EDOT, 0.5 mg ml Os(bpy)2PVI and 0.5 mg ml of FAD-GDH using a pulse sequence of 0.9 V(2 s) and -0.4 V (3 s) for a total time of 300 s [25].

NPG/MnO2 was fabricated via potentiostatic electrodeposition in 0.1 M

Na2SO4 and 0.1 M Mn(CH3COO)2 solution at 0.45 V vs. SCE and 30 °C for certain durations. The as-prepared NPG/MnO2 electrodes were subsequently immersed in solutions of 1 M H2SO4 and deionized water.

4.2.3. Electrochemical measurements Electrochemical studies were performed with a CHI802 potentiostat (CH Instruments, Austin, Texas) in a three-electrode electrochemical cell, with the NPG electrode, platinum wire and saturated calomel electrode (SCE) as the working, counter and reference electrodes, respectively. To obtain the power density profile of the assembled biobattery, the bioanode and NPG/MnO2 cathode were used as the working electrode and combined counter/reference electrode in a two-electrode system. The current was recorded over the potential range open circuit voltage of the

-1 BFC to 0 V at a scan rate of 1 mV s in N2-bubbled 0.1 M pH 7.0 PBS containing 10 mM glucose. The power density curve was calculated accordingly. All experiments were carried out at room temperature (20±2 °C) unless stated otherwise. Testing of the charge/discharge properties of the biobattery was performed in a PBS solution in (0.1 M pH 7.0) containing 10 mM glucose using an Autolab

PGSTAT100 potentiostat (Eco Chimie, Netherlands). The NPG/MnO2 and bioanode were used as working and combined counter/reference electrodes, respectively. The testing sequence comprised (i) stand at open-circuit while recording the open circuit potential (OCP) and (ii) galvanostatic discharge at defined current densities.

- 87 -

Chapter 4

4.3. Results and discussion

4.3.1 Electrochemical performance of NPG/MnO2 Anodic deposition is a widely-used method to oxidise Mn(II) dissolved in solution to MnO2 which is deposited as a film on an electrode [35]. The specific capacitance of NPG/MnO2 electrodes increased linearly with deposition time (Figure 4.1), in agreement with previous reports [36]. The formation of a coating layer was verified by SEM (Figure 4.2 and Figure 4.4A) and the presence of Mn was confirmed by EDX (Figure 4.4B). A Raman band at 657 cm-1 was assigned to manganese oxide in the form of Mn(III) and Mn(IV) (Figure 4.3) [37]. Unmodified NPG had a typical porous structure comprising interconnected pores and ligaments [25], with a uniform diameter of 30.6±5 nm (Figure 4.2A). The coating layer obtained after 30 s deposition was not clearly visible in the SEM image (Figure 4.2B), but could be clearly identified after deposition for 180 (Figure 4.4A) and 300 s (Figure 4.2C).

Figure 4.1 The relationship between deposition time and specific capacitance

of the corresponding NPG/MnO2 at a potential of 0.45 V vs. SCE.

- 88 -

Chapter 4

Figure 4.2 SEM images of the bare NPG (A) and NPG/MnO2 obtained by

electrodeposition for 30 s (B), and NPG/MnO2 for 300 s (C).

- 89 -

Chapter 4

Figure 4.3 Raman spectrum of the electrodeposited MnO2 film on the Au film (deposition time: 180 s).

Figure 4.4 SEM (A) and TEM (C) image of NPG/MnO2 (deposition time:

180 s). (B) EDX spectra of bare NPG and NPG/MnO2 (deposition time: 180 s). (D)

LSV of NPG/MnO2 (deposition time: 180 s) in 0.1 M pH 7.0 PBS at a scan rate of 2 mV s-1.

- 90 -

Chapter 4

Using TEM, the NPG/MnO2 composite material could be distinguished by the contrast difference between the modified layer and the gold skeleton (Figure 4.4C, Figure 4.5). The electrodeposited layer along the pore surfaces showed a relatively uniform thickness of 5.4±1 nm. The previous report, where the same methodology was used but with much thicker coating layers, showed that electrodeposited MnO2 nanocrystals possess a spinel structure [36]. Long deposition times, e.g. 300 s, resulted in the formation of thick films that blocked the pores (Figure 4.2C), which were likely to be detached from the electrode, leading to significantly degraded operational stability. A deposition period of 180 s was chosen for further electrochemical study. NPG/MnO2 (180 s) showed an initial specific capacitance of

-2 1.5±0.1 mF cm , which was almost four times higher than that of MnO2 on planar gold obtained using the same procedure and six-fold higher than that of bare NPG

[25]. NPG/MnO2 retained 64% of its capacitance, while planar Au/MnO2 only retained 26% after 50 charge-discharge cycles (Figure 4.6) reflecting the role of the substrate NPG in stablising the coating layer due to the confinement effects.

Figure 4.5 TEM images of the bare NPG.

- 91 -

Chapter 4

Figure 4.6 Stability of NPG/MnO2 and Au/MnO2 (deposition time: 180 s) in 0.1 M PBS, pH 7.0 in a potential range from 0 to 0.5 V vs. SCE at a scan rate of 100 mV s-1.

Figure 4.4D shows a linear sweep voltammogram (LSV) of NPG/MnO2 in 0.1 M pH 7.0 PBS, exhibiting a cathodic reduction with an onset potential of ca. +433 mV and a net cathodic current density of 72 µA cm-2 at 0.15 V. This reaction was oxygen independent (eq. 1), undergoing insertion of H+ and Na+ [16]. The observed discharge ability enables NPG/MnO2 to act as a consumed solid-state cathode [16], a potential alternative to ORR active enzymes based biocathodes.

4.3.2 Electrochemical performance of the bioanode and assembled biobattery

A previously optimised NPG/PEDOT/Os(bpy)2PVI/FAD-GDH bioanode was prepared for the oxidation of glucose [25]. Briefly, a pulse sequence consisting of anodic (0.9 V for 2 s) and cathodic -0.4 V (3 s) potentials resulted in the successive deposition of PEDOT and Os(bpy)2PVI with the co-immobilisation of enzyme into the polymer matrix. CVs of the bi-functional electrode displayed a response corresponding to the charge/discharge currents from the capacitive materials and the redox reaction of Os2+/3+ (Figure 4.7A). The midpoint potential of the osmium redox couple was +210 mV vs. SCE, very close to its reported formal potential of +220 mV vs. Ag/AgCl [38]. On addition of 10 mM glucose, a sigmoidal response (Figure 4.7A) arising from the catalytic oxidation of glucose was observed (vide infra, indicative of - 92 -

Chapter 4 the immobilisation of FAD-GDH). An onset potential of -18±9 mV vs. SCE was observed.

The NPG/MnO2 cathode and FAD-GDH based bioanode were assembled and tested without using a membrane. For the first test (blank line, Figure 4.7B), the biobattery registered a maximum current density of 14 µA cm-2, a maximum power density of 2.3 µW cm-2 at 0.21 V and an OCV of 0.49 V. This performance is an improvement over an equivalent EBFC with a BOx cathode that had a maximum power density of 1.3 µW cm-2 and an OCV of 0.46 V [25]. A subsequent test (red line, Figure 4.7B) showed a decreased power density (max. 1.9 µW cm-2) and OCV

(0.39 V) due to partial discharge of MnO2. To demonstrate the recovery behavior of the cathode, NPG/MnO2 was then transferred into a three-electrode cell containing PBS and oxidised at 0.5 V vs. SCE for 120 s. The OCV was restored to 0.49 V (blue line, Figure. 4.7B), the same value of the initial test, with a maximum power density of 2.1 µW cm-2, approaching the initial value. The recovery of maximum power density and the OCV also implied that the Mn(IV) was reduced to Mn(III) which is insoluble and retained in the film, unlike Mn(II) which is soluble and could diffuse into solution causing unwanted side reactions.

Figure 4.7 (A) CVs of the NPG/PEDOT/Os(bpy)2PVI/FAD-GDH bioanode. (B) The performance of the biobattery in the presence of 10 mM glucose.

4.3.3 Electrochemical performance of the hybrid device

The cell was also tested as a hybrid device in N2-bubbled 10 mM glucose solution. It was reset at the open-circuit mode for 30 min (cut-off at 0.4 V) and

- 93 -

Chapter 4 subsequently galvanostatic discharged at 0.1 mA cm-2 (cut-off at 0 V), a level significantly higher than the discharge current of the biobattery mode (14 µA cm-2). Once the potential of the built-in asymmetric capacitor was discharged to a potential close to zero, interestingly, the potential recovered towards the OCV of the biobattery (Figure 4.8A). The mechanism of this is described in the next section. The device could be used for 50 cycles (25 h) of discharge with slight decreases in the onset potential for discharge (in the range of 0.35 and 0.39 V). The hybrid devices were discharged at various current densities up to 2 mA cm-2 (Figure 4.8B). Current densities of 1 and 2 mA cm-2 led to maximum instantaneous power densities of 378 and 676 μW cm-2, respectively, 164 and 294 times higher than that from a biobattery configuration (2.3 µW cm-2). The significantly improved instantaneous power density was attributed to the intrinsic nature of the supercapacitor. The specific capacitance of the asymmetric supercapacitor was 320 μF cm-2 according to the galvanostatic discharge curve. The ohmic resistance during discharge was estimated to be 458 Ω based on the observed voltage drop. The ability to generate high-power-density pules is promising in the development of a hybrid device as a miniaturised power source to generate electric stimuli (e.g. cardiac pacemakers).

- 94 -

Chapter 4

Figure 4.8 (A) Potential profile of the device for 50 cycles. Solution: 0.1 M 7.0 PBS and 10 mM glucose. Experimental protocol: reset at open-circuit for 30 min and cutoff at 0.4 V, followed by discharging at 0.1 mA cm-2 and cutoff at 0 V. (B) Charge/discharge curves of the biocapacitor upon various discharging current densities; Experimental setup: reset at open-circuit for 30 min, followed by discharging at 0.005 (a), 0.01 (b), 0.02 (c), 0.05 (d), 0.1 (e), 0.2 (f), 0.5 (g), 1 (h), 2 (i) mA cm-2 for 0.2 s and cutoff at 0 V. Table 4.1 summarises the performance of representative enzyme based power sources consuming glucose as substrate. In comparison to EBFCs utilising gold nanomaterials including gold nanoparticles (AuNPs) [39], highly-ordered macroporous gold (MPG) [1] and dealloyed NPG [25, 33] under similar testing conditions, the biobattery displays reasonable output in terms of maximum power density and OCV. Significantly higher power density was achieved with a carbon nanotube (CNT) based biobattery [16], however, this system suffers from the disadvantage that it requires the use of NAD+ as a cofactor. The performance of the biobattery/supercapacitor hybrid device compares well with that of a biosupercapacitor [25] and with an Os polymer based EBFC/supercapacitor hybrid [8]

- 95 -

Chapter 4

but is lower than that of CNT based hybrid devices in the presence of high glucose concentration (100 and 200 mM) [20, 24]. Table 4.1 List of properties of enzymatic power sources utilising glucose as substrate.

Power PMax Anode Cathode OCV (V) Stability Ref. source (µWcm-2)

Au/AuNPs/CtCD ∼20% drop in 12 h of Au/AuNPs/MvBOx 0.68 3.3 [39] H continuous operation

MPG/ MPG/Os(bpy)2PVI/M Os(dmbpy)2PVI/A ∼0.52 38 N/A [1] aLc Au electrode nGOx based EBFC NPG/Os(dmbpy)2 NPG/Os(bpy)2PVI/M 25% drop in 12 h of 0.56 3.65 [33] PVI/AnGOx vBOx storage

NPG/Os(bpy)2PV ∼40% drop in 8 h of NPG-MvBOx 0.45 17.5 [25] I/FAD-GDH continuous operation

CFP/NAD+-IL-

SWCNTs/GDH/C GF/MnO2 0.43 40.5 N/A [40] S Biobattery Less than 50% drop in 6 GCE/MWCNTs/ Ag2O/Ag 0.59 275 h of continuous [41] MDB/GDH operation

MWCNTs/AnGO EBFC: 16 mW Charge/discharge for 5 MWCNTs/Lc 1±0.1 [20] x/catalase Hybrid: N/A days at 3 mA

NPG/PEDOT/Os( 50 cycles NPG/PEDOT/Os(bpy EBFC: 1.3 bpy)2PVI/FAD- 0.46 charge/discharge for ∼7 [42] )2PVI/MvBOx Hybrid: 608.8 GDH h at 0.2 mA cm-2

Graphite/Os EBFC/SC Graphite/Os EBFC: ∼3 Charge/discharge for polymer/PQQ- 0.45 [43] hybrid polymer/MvBOx Hybrid: N/A ∼50 h GDH

BP/MWCNTs/pM EBFC: N/A Charge/discharge for 3 G/GDH BP/MWCNTs/BOx ∼0.56 [24] Hybrid: 1070 days at 0.4 mA cm-2 Diffusing NAD+

NPG/PEDOT/Os( 50 cycles Biobattery/S Biobattery: 2.3 This bpy)2PVI/FAD- NPG/MnO2 0.49 charge/discharge for C hybrid Hybrid: 676 work GDH ∼25 h at 0.1 mA cm-2

- 96 -

Chapter 4

CDH: cellobiose dehydrogenase; Os(dmbpy)2PVI: [Os(4,4′-dimethyl-2,2′- +/2+ bipyridine)2(polyvinylimidazole)10Cl] ; GOx: glucose oxidase; Lc: laccase; N/A: not available; SC: supercapacitor; MWCNTs: multi-walled carbon nanotubes; BP: buckypaper; pMG: polymerized methylene green; MDB: Meldola’s blue; CFP: carbonfiber paper; SWCNTs: single-walled carbon nanotubes; IL: ionic liquid; CS: chitosan; GF: graphite flake.

4.3.4 Potential recovery of NPG/MnO2 Conway et al. described the mechanism involved in the recovery of the electrode potential of RuO2 electrodes that had undergone discharge [28]. During discharge, the outer region of the metal oxide layer is reduced first, with reduction occurring at a much slower rate in the bulk material due to the limited rate of proton exchange [44]. The presence of abundant oxidised Ru species in the bulk region enables re-oxidation of the surface region via an electron-hopping/charge transfer mechanism. In an analogous manner, MnO2 may undergo a similar process. The OCP of the discharged NPG/MnO2 was examined in a solution that had been saturated with either N2 or O2 (Figure 4.9A). The potential slowly recovered to ca. 0.37 V in both cases, indicating that the potential recovery was not affected by O2 over a period of 30 min. Assuming that the electrodeposition of MnO2 is 100% faradaically efficient [36], the amount of deposited MnO2 can be calculated by integrating the i-t curve. Applying a potential of 0.45 V vs. SCE for 180 s resulted in the deposition of 30 nmol

-2 MnO2 onto the NPG. Galvanostatic discharge at 0.1 mA cm for a short period of 3 s reduced 0.39 nmol MnO2 on the surface, assuming that each MnO2 accepted a single electron (i.e. conversion from Mn(IV) to Mn(III)). This data indicates that each discharge step consumed a very small fraction (1.3%) of the bulk MnO2. In the open- circuit mode, redistribution of the concentrations of Mn(IV) in the bulk and Mn(III) at the surface leads to the recovery of the potential, in a similar manner as described with RuO2. In the discharge step, the OCP of the assembled device decreased rapidly to approach 0 V (Figure 4.8A), i.e. a low potential difference between the bioanode and cathode. In the reset step, the charge transfer within the MnO2 film enabled redistribution of oxidation states that resulted in the return of the potential of the

- 97 -

Chapter 4 cathode to 0.37 V vs. SCE (Figure 4.9A), while the catalytic oxidation of glucose by the bioanode caused the potential to decrease with time to 0.01 V vs. SCE (Figure 4.9B) [8]. Simultaneously, the potential difference allowed the capacitive material on the bioanode to be recharged, whose charge would be released together with the

MnO2 cathode in the next discharge step (Scheme 4.1).

Figure 4.9 (A) OCP of NPG/MnO2 in the presence of N2 or O2. The electrode was discharged by scanning potential from 0.5 to 0 V vs. SCE at a scan rate of 1 mV s-1. (B) OCP of the bioanode upon the addition of 10 mM glucose.

- 98 -

Chapter 4

To emphasise the role of the catalytic active bioanode, we tested a device comprising a NPG/PEDOT anode without FAD-GDH and a NPG/MnO2 cathode (Figure 4.10). The potential recovery was also observed, but with a maximum OCP no higher than 0.1 V, which is lower than that (0.4 V) in the presence of FAD-GDH. Therefore, we confirm that a bioanode is essential to harness the potential difference, making the output potential and power density of the hybrid device acceptable.

Figure 4.10 Potential profile of NPG/PEDOT (negative electrode of the

capacitor)//NPG/MnO2 (positive electrode of the capacitor) over 50 cycles. Experimental procedure: reset at open-circuit for 30 min, followed by discharging at 0.1 mA cm-2 and cutoff at 0 V.

4.4. Conclusions

An oxygen independent and membrane-less glucose biobattery/supercapacitor hybrid device delivering high-power-density pulses was presented. MnO2 could replace oxygen reducing enzymes as the cathode, due to its features including the capability to be discharged at a reasonable potential, inert to glucose and insoluble reduced state. Most importantly, when only a fraction was discharged in the pulse mode, the spontaneous potential recovery of MnO2 occurred due to the redistribution of the oxidation states. Coupled with a FAD-GDH based

- 99 -

Chapter 4 supercapacitive bioanode, the hybrid device function in the similar way of a biosupercapacitor. This biobattery based hybrid device holds promise as an intermittent power source, which can overcome the limited oxygen supply occurring on the conventional biofuel cells, for implanted medical devices.

4.5 References

[1] S. Boland, D. Leech, A glucose/oxygen enzymatic fuel cell based on redox polymer and enzyme immobilisation at highly-ordered macroporous gold electrodes, Analyst, 137 (2012) 113-117. [2] M. Rasmussen, S. Abdellaoui, S.D. Minteer, Enzymatic Biofuel Cells: 30 Years of Critical Advancements, Biosens. Bioelectron., doi:10.1016/j.bios.2015.06.029 (2015). [3] S. Calabrese Barton, J. Gallaway, P. Atanassov, Enzymatic Biofuel Cells for Implantable and Microscale Devices, Chem. Rev., 104 (2004) 4867-4886. [4] K. MacVittie, J. Halamek, L. Halamkova, M. Southcott, W.D. Jemison, R. Lobel, E. Katz, From "Cyborg" Lobsters to a Pacemaker Powered by Implantable Biofuel Cells, Energy Environ. Sci., 6 (2013) 81-86. [5] G. Pankratova, D. Pankratov, K. Hasan, H.-E. Åkerlund, P.-Å. Albertsson, D. Leech, S. Shleev, L. Gorton, Supercapacitive Photo-Bioanodes and Biosolar Cells: A Novel Approach for Solar Energy Harnessing, Adv. Energy Mater., (2017) DOI: 10.1002/aenm.201602285. [6] A. Carreau, B.E. Hafny‐Rahbi, A. Matejuk, C. Grillon, C. Kieda, Why is the partial oxygen pressure of human tissues a crucial parameter? Small molecules and hypoxia, J. Cell. Mol. Med., 15 (2011) 1239-1253. [7] D.G. Maggs, R. Jacob, F. Rife, R. Lange, P. Leone, M.J. During, W.V. Tamborlane, R.S. Sherwin, Interstitial fluid concentrations of glycerol, glucose, and amino acids in human quadricep muscle and adipose tissue. Evidence for significant lipolysis in skeletal muscle, J. Clin. Invest., 96 (1995) 370-377. [8] E. González-Arribas, D. Pankratov, S. Gounel, N. Mano, Z. Blum, S. Shleev, Transparent and Capacitive Bioanode Based on Specifically Engineered Glucose Oxidase, Electroanalysis, 28 (2016) 1290-1297. [9] D.J. Spira-Solomon, M.D. Allendorf, E.I. Solomon, Low-temperature magnetic circular dichroism studies of native laccase: confirmation of a trinuclear copper active site, J. Am. Chem. Soc., 108 (1986) 5318-5328. [10] F. Xu, Oxidation of Phenols, Anilines, and Benzenethiols by Fungal Laccases: Correlation between Activity and Redox Potentials as Well as Halide Inhibition, Biochemistry, 35 (1996) 7608-7614. [11] C. Vaz-Dominguez, S. Campuzano, O. Rüdiger, M. Pita, M. Gorbacheva, S. Shleev, V.M. Fernandez, A.L. De Lacey, Laccase electrode for direct

- 100 -

Chapter 4 electrocatalytic reduction of O2 to H2O with high-operational stability and resistance to chloride inhibition, Biosens. Bioelectron., 24 (2008) 531-537. [12] U. Salaj-Kosla, S. Pöller, W. Schuhmann, S. Shleev, E. Magner, Direct Electron Transfer of Trametes Hirsuta Laccase Adsorbed at Unmodified Nanoporous Gold Electrodes, Bioelectrochem., 91 (2013) 15-20. [13] C. Kang, H. Shin, A. Heller, On the stability of the “wired” bilirubin oxidase oxygen cathode in serum, Bioelectrochem., 68 (2006) 22-26. [14] T. Miyake, K. Haneda, N. Nagai, Y. Yatagawa, H. Onami, S. Yoshino, T. Abe, M. Nishizawa, Enzymatic Biofuel Cells Designed for Direct Power Generation from Biofluids in Living Organisms, Energy Environ. Sci., 4 (2011) 5008- 5012. [15] P.K. Addo, R.L. Arechederra, S.D. Minteer, Towards a rechargeable alcohol biobattery, J. Power Sources, 196 (2011) 3448-3451. [16] L. Yang, S. Cheng, J. Wang, X. Ji, Y. Jiang, M. Yao, P. Wu, M. Wang, J. Zhou, M. Liu, Investigation into the origin of high stability of δ-MnO2 pseudocapacitive electrode using operando Raman spectroscopy, Nano Energy, 30 (2016) 293-302. [17] V.D. Neff, Some Performance Characteristics of a Prussian Blue Battery, J. Electrochem. Soc., 132 (1985) 1382-1384. [18] X. Xie, M. Ye, P.-C. Hsu, N. Liu, C.S. Criddle, Y. Cui, Microbial battery for efficient energy recovery, Proc. Natl. Acad. Sci., 110 (2013) 15925-15930. [19] X. Xie, M. Ye, C. Liu, P.-C. Hsu, C.S. Criddle, Y. Cui, Use of low cost and easily regenerated Prussian Blue cathodes for efficient electrical energy recovery in a microbial battery, Energy Environ. Sci., 8 (2015) 546-551. [20] C. Agnes, M. Holzinger, A. Le Goff, B. Reuillard, K. Elouarzaki, S. Tingry, S. Cosnier, Supercapacitor/Biofuel Cell Hybrids Based on Wired Enzymes on Carbon Nanotube Matrices: Autonomous Reloading after High Power Pulses in Neutral Buffered Glucose Solutions, Energy Environ. Sci., 7 (2014) 1884-1888. [21] D. Pankratov, Z. Blum, S. Shleev, Hybrid Electric Power Biodevices, ChemElectroChem, 1 (2014) 1798-1807. [22] M. Kizling, S. Draminska, K. Stolarczyk, P. Tammela, Z. Wang, L. Nyholm, R. Bilewicz, Biosupercapacitors for Powering Oxygen Sensing Devices, Bioelectrochem., 106, Part A (2015) 34-40. [23] K.L. Knoche, D.P. Hickey, R.D. Milton, C.L. Curchoe, S.D. Minteer, Hybrid Glucose/O2 Biobattery and Supercapacitor Utilizing a Pseudocapacitive Dimethylferrocene Redox Polymer at the Bioanode, ACS Energy Lett., (2016) 380- 385. [24] C.W. Narvaez Villarrubia, F. Soavi, C. Santoro, C. Arbizzani, A. Serov, S. Rojas-Carbonell, G. Gupta, P. Atanassov, Self-Feeding Paper Based Biofuel Cell/Self-Powered Hybrid Μ-Supercapacitor Integrated System, Biosens. Bioelectron., 86 (2016) 459-465. [25] T. Siepenkoetter, U. Salaj-Kosla, X. Xiao, P.Ó. Conghaile, M. Pita, R. Ludwig, E. Magner, Immobilization of Redox Enzymes on Nanoporous Gold Electrodes: Applications in Biofuel Cells, ChemPlusChem, 82 (2017) 553-560. - 101 -

Chapter 4

[26] Z. Liu, L. Huang, L. Zhang, H. Ma, Y. Ding, Electrocatalytic Oxidation of D-glucose at Nanoporous Au and Au–Ag Alloy Electrodes in Alkaline Aqueous Solutions, Electrochim. Acta, 54 (2009) 7286-7293. [27] P. Simon, Y. Gogotsi, Materials for electrochemical capacitors, Nat. Mater., 7 (2008) 845-854. [28] T. Liu, W.G. Pell, B.E. Conway, Self-discharge and potential recovery phenomena at thermally and electrochemically prepared RuO2 supercapacitor electrodes, Electrochim. Acta, 42 (1997) 3541-3552. [29] E.M. Kober, J.V. Caspar, B.P. Sullivan, T.J. Meyer, Synthetic routes to new polypyridyl complexes of osmium(II), Inorg. Chem., 27 (1988) 4587-4598. [30] R.J. Forster, J.G. Vos, Synthesis, characterization, and properties of a series of osmium- and ruthenium-containing metallopolymers, Macromolecules, 23 (1990) 4372-4377. [31] C. Sygmund, P. Staudigl, M. Klausberger, N. Pinotsis, K. Djinović- Carugo, L. Gorton, D. Haltrich, R. Ludwig, Heterologous overexpression of Glomerella cingulata FAD-Dependent Glucose Dehydrogenase in Escherichia Coli and Pichia Pastoris, Microb. Cell Fact., 10 (2011) 1-9. [32] X. Xiao, H. Li, M. Wang, K. Zhang, P. Si, Examining the Effects of Self-Assembled Monolayers on Nanoporous Gold Based Amperometric Glucose Biosensors, Analyst, 139 (2014) 488-494. [33] X. Xiao, E. Magner, A Biofuel Cell in Non-Aqueous Solution, Chem. Commun., 51 (2015) 13478-13480. [34] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, J. Schindelin, I. Arganda- Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, NIH Image to ImageJ: 25 Years of Image Analysis, Nat. Methods, 9 (2012) 671. [35] D. Tench, L.F. Warren, Electrodeposition of Conducting Transition Metal Oxide/Hydroxide Films from Aqueous Solution, J. Electrochem. Soc., 130 (1983) 869-872. [36] L.Y. Chen, J.L. Kang, Y. Hou, P. Liu, T. Fujita, A. Hirata, M.W. Chen, High-energy-density nonaqueous MnO2@nanoporous gold based supercapacitors, J. Mater. Chem. A, 1 (2013) 9202-9207. [37] M.A. García, M.L. Ruiz-González, A. Quesada, J.L. Costa-Krämer, J.F. Fernández, S.J. Khatib, A. Wennberg, A.C. Caballero, M.S. Martín-González, M. Villegas, F. Briones, J.M. González-Calbet, A. Hernando, Interface Double-Exchange Ferromagnetism in the Mn-Zn-O System: New Class of Biphase Magnetism, Phys. Rev. Lett., 94 (2005) 217206. [38] P.A. Jenkins, S. Boland, P. Kavanagh, D. Leech, Evaluation of Performance and Stability Of Biocatalytic Redox Films Constructed with Different Copper Oxygenases and Osmium-Based Redox Polymers, Bioelectrochem., 76 (2009) 162-168. [39] X. Wang, M. Falk, R. Ortiz, H. Matsumura, J. Bobacka, R. Ludwig, M. Bergelin, L. Gorton, S. Shleev, Mediatorless Sugar/Oxygen Enzymatic Fuel Cells Based on Gold Nanoparticle-Modified Electrodes, Biosens. Bioelectron., 31 (2012) 219-225. - 102 -

Chapter 4

[40] Y. Yu, Y. Han, B. Lou, L. Zhang, L. Han, S. Dong, A miniature origami biofuel cell based on a consumed cathode, Chem. Commun., 52 (2016) 13499-13502. [41] Y. Yu, M. Xu, L. Bai, L. Han, S. Dong, Recoverable hybrid enzymatic biofuel cell with molecular oxygen-independence, Biosens. Bioelectron., 75 (2016) 23-27. [42] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, A Symmetric Supercapacitor/Biofuel Cell Hybrid Device Based on Enzyme-Modified Nanoporous Gold: An Autonomous Pulse Generator, Biosens. Bioelectron., 90 (2017) 96-102. [43] D. Pankratov, F. Conzuelo, P. Pinyou, S. Alsaoub, W. Schuhmann, S. Shleev, A Nernstian Biosupercapacitor, Angew. Chem. Int. Ed., 55 (2016) 15434- 15438. [44] S. Ardizzone, G. Fregonara, S. Trasatti, “Inner” and “outer” active surface of RuO2 electrodes, Electrochim. Acta, 35 (1990) 263-267.

- 103 -

CHAPTER 5:

Nanoporous gold based biofuel cells on contact lenses

Chapter 5

Chapter 5. Nanoporous gold based biofuel cells on contact lenses

5.1. Introduction

Enzymatic biofuel cells (EBFCs) have been extensively investigated [1-3], since the first prototype consisting of a glucose oxidase (GOx) modified anode and a Pt cathode was introduced by Yahiro et al. in 1964 [4]. EBFCs enjoy advantages such as ease of miniaturisation and the ability to operate at physiological conditions. EBFCs that mimic metabolic pathways (most interestingly, glucose oxidation) in the body have been envisioned as autonomous power supplies for implantable medical devices since the 1970s [5]. Significant efforts have been directed towards designing implantable EBFCs, with EBFCs examined in vivo in rats [6-9], snails [10], lobsters [11], etc. Recently, EBFCs have been tested ex vivo in human blood [12, 13]. However, practical applications of EBFCs operating in the human body have been hindered by i) the relatively large size of the devices [9], ii) the limited lifetime due to the deactivation and/or leakage of enzyme [13] and iii) low power density due to inefficient rates of electron transfer between the enzymes and the electrode surface, accompanied by the limited mass-transport of substrates. The supply of O2 in particular is a significant constraint for in vivo operation and thus oxygen-reducing biocathodes are a significant limiting factor [3, 12]. Non-invasive EBFCs utilising fuels in saliva [14, 15], sweat [16], tear [17], etc., offer an alternative that can circumvent issues caused by implantation. This type of EBFC avoids direct contact with the immune system and removes the necessity of a surgical procedure. Such devices can be easily discarded and replaced and can be used to activate wearable medical devices for continuous health monitoring and applications in sports science [18]. In comparison to implantable glucose/O2 EBFCs that utilise glucose and O2 in blood, lactate/O2 biofuel cells have more potential for use with wearable electronics due to the higher concentration of lactate in tears and in

- 104 -

Chapter 5 sweat. For example, the normal concentrations of glucose and lactate in human blood range from 3.3-6.5 mM and 0.5-0.8 mM, respectively, in comparison to 0.1-0.6 mM and 2-5 mM, respectively, in human tears [19]. Lactate is an important biomarker of metabolic efficiency during physical exercise. A correlation between the concentration of lactate in sweat and in blood has been described by Sakharov et al., indicating that sweat lactate levels can be measured to evaluate changes in blood lactate concentrations [20]. Self-powered lactate biosensors based on the construction of lactate EBFCs, have been demonstrated showing a linear increase in power density for lactate concentrations between 0 and 5 mM [21]. A promising type of wearable device is the temporary tattoo EBFC [22]. Wang et al. successfully fabricated an EBFC that can be attached on the skin to harvest energy from lactate present in human sweat during physical exercise [16]. The transferrable tattoo based EBFC employed a mediated lactate oxidase (LOx) bioanode and a platinum black cathode. Another interesting type of wearable EBFC is that incorporated on a contact lens [23]. Basal tears containing a range of species such as lactate, glucose, ascorbate, saturated air, etc., keep the cornea moist, making the preparation of continuous and self-sustained EBFCs possible during physical movement and under quiescent conditions. Contact lenses floating on the cornea for myopia correction are commercially available. Recently, new roles for such systems, including sensors [24-27], digital displays [28] and drug release [29] have been described. Falk et al. first reported experimental proof of a 3D nanostructured gold wire supported glucose/O2 EBFC that could be used on contact lenses [30]. Follow-up work from the same group described a hybrid EBFC relying on an abiotic anode to oxidise ascorbate and a bilirubin oxidase (BOx) based biocathode to reduce oxygen in real human tears [17]. Minteer et al. described the preparation of a buckypaper supported EBFC assembled with a lactate dehydrogenase (LDH) bioanode and a BOx biocathode that was deposited on a curved elastomeric substrate [31], registering a maximum power density of 2.4 ± 0.9 µW cm-2 [32].

- 105 -

Chapter 5

Electrode materials for tear based EBFCs should (i) be flexible enough to be seamlessly attached onto the eyeball, (ii) exhibit high surface areas to ensure high enzyme loadings for larger current densities and (iii) be biocompatible for use on the eye [23]. Dealloyed nanoporous gold (NPG) [33] possesses a three-dimensional porous structure fabricated via the etching of gold alloys and is a promising substrate for EBFCs [34-37]. The pore sizes can be finely tuned to accommodate enzymes in a manner that optimises the electrocatalytic response [36, 38, 39]. Thin (100 nm) NPG leaves have been used as the electrode material of flexible supercapacitors [40]. However, Au-Ag NPG prepared from these leaves are brittle [34], while the preparation conditions (concentrated nitric acid) used for chemical etching are highly corrosive [37, 41], with concomitant safety and environmental concerns. To overcome these issues, we electrochemically dealloyed Au-Ag alloys at neutral pH to fabricate mechanically robust NPG electrodes. A polyethylene terephthalate (PET) film was used to provide a flexible substrate. LOx and BOx were immobilised onto the

+/2+ electrode with the assistance of [Os(2,2′-bipyridine)2(polyvinylimidazole)10Cl]

(Os(bpy)2PVI) [35] and diazonium grafting [36], separately, for the bioanode and biocathode. The EBFC was enclosed between two commercially available contact lenses (Scheme 6.1) to avoid direct contact with the eye [31, 32]. Hydrophilic silicon- hydrogel contact lenses contain micro-channels to enable the transport of solutions and oxygen to the EBFC [42]. The performance of the EBFC was examined in solutions containing phosphate buffer solution (PBS) and artificial tears, exhibiting a maximum power density of 2.4±0.2 and 1.7±0.1 μW cm-2, respectively. This chapter highlights a new and green route to fabricate mechanically robust flexible NPG, which can find applications in wearable energy devices. The embedded bioelectrodes within silicon hydrogel contact lenses are believed to be more practical for use in next-generation “smart contact lenses”, rather than a supported micro-device supported on a surface of a contact lens.

- 106 -

Chapter 5

Scheme 5.1 Schematic diagram of the assembly of the modified contact lens (A) and the configuration of the EBFC (B).

5.2. Experimental section

5.2.1. Materials and apparatus Sodium phosphate (monobasic dehydrate ≥99% and dibasic ≥99%), sodium fluoride (NaF, 99.99%), hydrochloric acid (HCl, 37%), sulfuric acid (H2SO4, 95-

98%), D-(+)-glucose (99.5%), sodium nitrite (NaNO2, ≥99.999%), 6-amino-2- naphthoic acid (NA, 90%), 3-mercaptopropionic acid (MPA, ≥99%), L-ascorbic acid (≥99%), urea (≥99.5%), sodium L-lactate (≥99%), N-cyclohexyl-N’-(2- morpholinoethyl) carbodiimide metho-p-toluenesulfonate (CMC, ≥99%), lysozyme human (EC 3.2.1.17), bovine serum albumin (BSA), mucin from porcine stomach and Pediococcus sp. LOx (EC 1.13.12.4, ≥20 U mg-1) were purchased from Sigma- Aldrich Ireland, Ltd. Anhydrous acetonitrile (>99.8%) was obtained from Fisher Scientific, Ireland. Myrothecium verrucaria BOx (EC 1.3.3.5, 2.63 U mg-1) was obtained as a gift from Amano Enzyme Inc., Japan. Os(bpy)2PVI was synthesised using an established procedure [43, 44]. Silicon-hydrogel contact lenses (-0.5 and -9.0 diopter) were obtained locally. Deionised water (18.2 MΩ cm, Elga Purelab Ultra, UK) was used for all preparations.

- 107 -

Chapter 5

Morphology studies were performed with a Hitachi SU-70 scanning electron microscope (SEM, 10 kV), equipped with an energy dispersive X-ray spectroscopy (EDX) for residual Ag determination. ImageJ software (National Institutes of Health, Bethesda, Maryland) [45] was used to measure the average pore size and crack width of NPG by analysing at least 30 measurement points.

5.2.2. Electrochemical dealloying Magnetron sputtered Ag/Au alloy was prepared in an ultra-high vacuum chamber at room temperature according to a previous report [38]. Briefly, Ar plasma treated microscope glass slides or 100 µm thin PET substrates were coated with a 10 nm Ti adhesive layer, ca. 35 nm Au protective layer and 100 nm Ag70/Au30 (atomic %) alloy layer, subsequently. The glass supported alloy sheets were cut using a circular saw and painted with dielectric paste (Gwent Group, UK) to define an electroactive surface area of ca. 0.3 cm2. Cleanroom tape composed of a polyamide film (VWR, Ireland) was used to define the electrode area (0.35*0.35 cm2) of the PET supported alloy sheets. To prepare NPG, the alloy was anodized at +1.05 and +1.5 V vs. SCE in 0.5 M NaF at room temperature (20±2 °C) for 10 min and subsequently cleaned by

-1 scanning the potential from -0.2 to 1.65 V in 1 M H2SO4 at a scan rate of 100 mV s for a range of potential cycles (1 to 15). The electrochemically addressable surface area (Areal) of NPG and the roughness factor (Rf), i.e. the ratio of Areal to the geometric

-2 area (Ageo), was obtained by cyclic voltammetry using a value of 390 μC cm for the reduction of a single layer of gold oxide [46].

5.2.3. Enzyme immobilisation NPG based bioanodes were prepared by electrodeposition at -1.1 V for

-1 different durations (60-600 s) in 0.1 M pH 7.0 PBS containing 1 mg ml Os(bpy)2PVI and 1 mg ml-1 of LOx. The surface coverage (nmol cm-2) of Os polymer on the electrode was calculated according to the equation:

- 108 -

Chapter 5

푄 Surface coverage = (1) 푛퐹퐴 where Q (nC) is the charge regarding to the oxidation/reduction of Os polymer determined based on the cyclic voltammograms (CVs) in a blank PBS, n is the number of electrons involved, F is the Faraday constant (96485 C mol-1) and A (cm-2) is the geometric area of the electrode. BOx was covalently attached to NPG via a 2-carboxy-6-naphtoyl diazonium salt (NA-DS) modification layer, which was synthesised according to a previous report [36]. Briefly, a fresh NA-DS solution was obtained by dropwise addition of 2 mL of 2 M HCl containing 2 mM NaNO2 into a 2 mL solution of 20 mM NA in acetonitrile, in an ice bath. Electrografting was achieved by electrochemically reducing NA-DS at the electrode surface with a single potential scan over the potential range 0.6 to -0.6 V at a scan rate of 200 mV s-1 (Figure 5.1). The modified electrodes were immersed into a 1 mM MPA aqueous solution overnight to block any unmodified gold surface, followed by carefully rinsing with deionised water and dried in vacuum. A 20 μL aliquot of BOx (0.5 mg mL-1) was drop cast onto the surface of the electrode, incubated in a vacuum chamber for 5 min and then transferred to a fridge at 4 °C for 1 h. The modified electrodes were then immersed in a solution of CMC (5 mM) at 4 °C for 2 h to crosslink the enzyme molecules.

Figure 5.1 Cyclic voltammogram of a NPG electrode in an ice-bath cooled solution of 2 mL 2 M HCl containing 2 mM NaNO2 and 2 mL solution of 20 mM NA in acetonitrile. Scan rate: 200 mV s-1.

- 109 -

Chapter 5

5.2.4. Electrochemical measurements Electrochemical experiments were carried out with a CHI802 potentiostat (CH Instruments, Austin, Texas) in a three-electrode electrochemical cell consisting of the Au alloy or NPG based working electrode, a platinum counter electrode and saturated calomel electrode (SCE) as the reference electrode. The assembled EBFC was tested in a two-electrode system by using a LOx based bioanode as the working electrode and a BOx based biocathode as the combined counter/reference electrode, recording the current in the potential range between the open circuit voltage of the EBFC and 0 V at 1 mV s-1. The power density curve was calculated using the geometric area of the limiting electrode. All experiments were performed at room temperature (20±2 °C) unless stated otherwise. To test the contact lens supported EBFC, a contact lens (diopter -9.0) was first mounted onto the polycarbonate packing material used to package the contact lens. The material had the same curved surface shape of the lens. It was first treated with O2 plasma (30 s, Solarus 950 Advanced Plasma System, Gatan, USA) and the NPG based EBFC electrodes were then placed on the lens followed by a thinner contact lens (diopter -0.5) (Scheme 5.1). Artificial tear solutions [32] (a mixture of 50 µM glucose, 3 mM lactate, 180 µM ascorbate, 5.4 mM urea, 2.47 mg mL-1 lysozyme, 0.2 mg mL-1 BSA, and 0.15 mg mL-1 mucin in 0.1 M pH 7.0 PBS) were maintained at 35 °C and continuously dropped onto the contact lenses with a peristaltic pump (P720, Instech, USA).

5.3. Results and discussion

5.3.1. Electrochemical dealloying Dealloying of gold alloys is normally performed in concentrated nitric acid (ca. 15.7 M) at temperatures of 30 °C or higher [38, 47-49]. In contrast, there has only been a few reports describing the dealloying of gold alloys at neutral pH. Such methods are based on the electrochemical oxidation of the less noble element and subsequent removal of the oxidised product [50, 51]. Expensive salt solutions (such as - 110 -

Chapter 5

AgNO3) were used to dealloy Ag65/Au35 (atomic %) at an applied potential between 1.4 and 2 V vs. NHE [50]. Al/Au alloys were electrochemically dealloyed in solutions of NaCl [51]. However, in dealloying Ag/Au alloys, the use of NaCl as an electrolyte may result in the precipitation of insoluble AgCl [52]. Therefore, NaF was selected as a low cost alternative electrolyte for electrochemical dealloying.

Linear sweep voltammogram (LSV) of Ag70/Au30 alloy (sputtered on glass) in 0.5 M NaF (Figure 5.2A) displayed peaks corresponding to the oxidation of Ag (0.77 to 1.22 V vs. SCE and the evolution of oxygen (onset potential of 1.25 V vs. SCE), consistent with data from the Pourbaix diagram for Ag [53]. To illustrate the effect of oxidation potential, two representative potentials, 1.05 and 1.5 V in the region of Ag oxidation and OER respectively, were chosen (Figure 5.2A). After 10 min at room temperature, both potentials led to NPG with similar morphology (Figure 5.3B and Figure 5.4A) showing pinholes with very small average pore sizes (8.2±2 and ~5 nm at 1.05 and 1.5 V, respectively, Figure 5.6, Table 5.1 and 5.2) and wide cracks due to stress release on volume contraction during the dealloying process [49] (20±4.2 and 16.8±4.1 nm at 1.05 and 1.5 V, respectively Figure 5.6, Table 5.1 and 5.2). EDX analysis confirmed that a significant amount of Ag (12.2±0.3% at 1.05 V and 13.6±0.4% at 1.5V, Figure 5.6, Table 5.1 and 5.2) remained, due to the presence of residual silver oxide passivating further silver dissolution [50]. This was consistent with the observation that longer anodisation times (>10 min) resulted in no significant changes in terms of average pore size and Ag content. To obtain NPG electrodes with suitable pore sizes for enzyme immobilisation, the residual oxide was removed by cycling the applied potential in 1 M H2SO4 [50], which is an established protocol to create clean gold electrode surfaces [30, 41, 54]. As potential cycling continued (Figure 5.2B), the small peaks corresponding to removal of Ag at ca. +0.2 V started to disappear after the second potential cycle. Meanwhile, the reduction peak of gold oxide at ca. 0.85 V (Figure 5.2B) decreased due to coarsening of NPG [55] that was associated with an increase in the average pore size and a decrease in the specific surface area [41, 48]. SEM (Figure 5.3B, C and D, Figure 5.4) and EDX (Figure 5.6,

- 111 -

Chapter 5

Table 5.1 and 5.2) results reinforced these observations. NPG anodised at 1.05 and 1.5 V (Figure 5.6, Table 5.1 and 5.2) both showed decreases in the amounts of silver remaining, with increases in pore sizes and decreases in the surface roughness. The observed cracks grew in size with continuous potential cycling, until they were in the same size range of the pores and no longer distinguishable from them after the 15th scan [38]. The observed pore evolution process during potential scanning is quite similar to that reported on glass supported Ag/Au alloys during etching by concentrated nitric acid [38]. NPG prepared at potentials of 1.05 and 1.5 V both generated satisfactory nanoporous structures in terms of acceptable Rf and pore sizes. For further studies, 1.5 V was used as the optimal potential as the resulting pore sizes

(20.2±5.1 nm) and large Rf values (17.0±0.8) were suitable for enzyme immobilisation.

Figure 5.2 (A) Linear sweep voltammogram of Ag70/Au30 alloy in 0.5 M

NaF. (B) Cyclic voltammograms of the as-anodised NPG (1.5 V) in 1 M H2SO4.

Control experiments performed by cycling the potential in H2SO4 without anodisation also resulted in the appearance of nanoporous structures (Figure 5.5 and Table 5.3). After two cycles, only 4.7±0.4% Ag remained in the alloy. The electrodes had an average pore size of ca. 5 nm and also displayed large cracks (40.3±8.2 nm, Figure 5.5A, Table 5.3). In contrast, anodisation resulted in enriched Ag content (12.2±0.3% at 1.05 V and 13.6±0.4% at 1.5V, Figure 5.6, Table 5.1 and 5.2) on the surface and etching of Ag was relatively slower as it was impeded by the presence of - 112 -

Chapter 5 silver oxide, thus resulting in smaller crack sizes (20±4.2 nm for 1.05 V and 16.8±4.1 nm for 1.5 V, Figure 5.6, Table 5.1 and 5.2). Cyclic voltammograms of the electrodes in H2SO4 resulted in the extraction of the surface-enriched Ag and re-configuration of surface Au atoms (i.e. surface re-arrangement) [41, 50, 55]. Without anodisation, direct etching in H2SO4 stripped Ag from the bulk, immediately leading to rapid volume changes and resulting in large cracks. Additional potential scans (up to 15 cycles (Figure 5.5B, Table 5.3)) resulted in NPG with low Rf (4.0±0.2) and expanded crack sizes (44.6±5.7 nm). The obtained microstructure showed unconnected ligaments when scanned to the 30th cycle (Figure 5.5B). In conclusion, potential cycling in H2SO4 without anodisation resulted in NPG with pores that were too small for enzymes to enter (2 cycles) or unsatisfactory roughness (15 and 30 cycles) (Table 5.3). Thus, it is essential to first anodise the precursor to obtain the optimal NPG structure. Figure 5.3A illustrates a possible mechanism of the two-step electrochemical dealloying process. Anodisation generates silver oxide passivated

NPG with pinholes. Subsequent coarsening of the surface in H2SO4 removes the residual silver oxides and allows re-configuration of the surface gold atoms via repeated electro-oxidation/reduction.

- 113 -

Chapter 5

Figure 5.3 (A) Schematic diagram of the electrochemical dealloying process. (B-E) SEM images of the porous structure of NPG obtained at different conditions. B: anodisation in 0.5 M NaF at 1.5 V vs. SCE for 10 min, anodisation and cycling

potential in 1 M H2SO4 for 1 (C), 2 (D) and 15 (E) potential cycles.

Table 5.1 Summary of results obtained by cycling potential in 1 M H2SO4 of as-anodised NPG at 1.05 V.

- 114 -

Chapter 5

Figure 5.4 SEM images the porous structure of NPG obtained with different conditions. A: anodisation in 0.5 M NaF at 1.05 V vs. SCE for 10 min, anodisation

and cycling potential in 1 M H2SO4 for 1 (B), 2 (C) and 15 (D) potential cycles.

Table 5.2 Summary of results obtained by cycling potential in 1 M H2SO4 of as-anodised NPG at 1.5 V.

- 115 -

Chapter 5

Table 5.3 Summary of results obtained by cycling potential in 1 M H2SO4 of

Ag70/Au30 alloy without anodisation.

Figure 5.5 SEM images the porous structure of NPG obtained by cycling potential in 1 M H2SO4 for 2 (A), 15 (B) and 30 (C) potential cycles without anodisation.

Figure 5.6 Plots of (A) roughness factor; (B) residual silver content; (C) pore size; (D) crack width obtained after potential cycling of as-anodised NPG.

- 116 -

Chapter 5

Anodisation followed by cyclic voltammetry in in H2SO4 was performed to dealloy the less noble metal Ag from a PET supported Ag70/Au30 alloy. The morphology showed a continuous porous structure with an average pore size of 20.9±4.2 nm and a crack width of 28.4±4.6 nm. The obtained NPG could be bent (Figure 5.7A) with no obvious structural deformation after a 40-degree bend (Figure 5.7C). Streaks were observed when the electrode was bent by 60-degrees (Figure 5.7D) and the electrode was still conductive due to the presence of the underlying Au layer. No delamination occurred even after a 90-degree bend due to the Ti adhesive

-1 layer. The sheet resistance (Rs) of the flexible NPG on PET was 3.4± 0.1 Ω sq (Figure 5.8), with slight increases to 3.5± 0.1 and 3.6 Ω sq-1 after bending to 40- and

o 90- , respectively. The measured Rs was similar to that of a NPG leaf electrode (2.5 Ω sq-1) [56]. This type of PET supported NPG is expected to find applications such as supercapacitors [40], point-of-care diagnostics [55] and surface enhanced resonance Raman scattering (SERRS) sensors [57]. Here, we demonstrated the use of such electrodes for EBFCs.

- 117 -

Chapter 5

Figure 5.8 (A) Schematic drawing of the experimental setup for resistance measurements according to reference [56]. (B) The relationship between bending angles of the NPG film and sheet resistance.

5.3.2. Characterisation of the bioelectrodes FAD-dependent LOx was selected as it can be easily immobilised on an electrode surface with redox polymers [21], which simultaneously immobilise the enzyme and shuttle electrons from the enzyme to the electrode surface. Os poly(N- vinylimidazole) redox polymers have been shown to be promising mediators for use - 118 -

Chapter 5 with LOx [58, 59]. As shown in Scheme 5.2, the oxidation of L-lactate to pyruvate is catalyzed by LOx(FAD), with the redox center FAD converted to be the reduced form, FADH2. Oxidation of FADH2 by Os(III) in the redox polymer regenerates FAD and produces Os(II) which is subsequently re-oxidised to Os(III) at the electrode. Instead of drop-casting which leads to a film with relatively poor stability, electrodeposition can be used to co-immobilise enzymes and Os polymers containing weakly coordinated chloride ions [60]. At a cathodic potential, Os3+ was reduced to Os2+, accompanied by exchange of chloride ions with the more strongly coordinating pyridine or imidazole groups on the polymer. Such a crosslinking process resulted in irreversible polymer precipitation onto the electrode. A negative potential of -1.1 V vs. SCE was used for deposition in the presence of Os(bpy)2PVI and LOx. The electrocatalytic responses of the resulting bioelectrodes varied with deposition time (Figure 5.11A). Surface coverages of the Os polymer (Figure 5.11A, blue curve) increased with deposition time. In other words, longer deposition time resulted in an increase in the amounts of Os polymer and enzyme that were immobilised [35]. The highest response was obtained with a moderate deposition time of 360 s, reflecting a compromise between the loading of the enzyme and mass-transport resistance of lactate through the film [35]. As shown previously, transmission electron microscopic (TEM) images clearly showed that the pores of NPG were blocked by the polymer film when the deposition time was too long, leading to a decreased catalytic response

- 119 -

Chapter 5

[35]. A deposition time of 360 s resulted in a thin film growing along the ligaments without plugging the pores (Figure 5.9).

Scheme 5.2 Electron transfer route between the electrode surface and LOx

mediated by Os(bpy)2PVI.

Figure 5.9 SEM image of the deposition layer of NPG/Os(bpy)2PVI/LOx (360 s deposition).

Cyclic voltammograms (CVs) of the optimal NPG/Os(bpy)2PVI/LOx bioanode, performed in air-equilibrated PBS, exhibited a pair of well-defined redox peaks with an integrated anodic-to-cathodic peak area ratio of approximately one and a midpoint redox potential of 193±1 mV vs. SCE (Figure 5.11B), attributed to the rapid oxidation/reduction of the Os2+/3+. In the presence of 3 mM lactate, a typical sigmoidal catalytic wave appeared, with an onset potential of 47±30 mV vs. SCE and

- 120 -

Chapter 5 a net catalytic response of 62.2±1.9 μA cm-2. It’s noteworthy that the reduced form of

LOx (FADH2) can also be oxidised by O2 [58], decreasing the current and generating unwanted H2O2 that could deactivate the enzyme. Additionally, an oxygen depleting bioanode will reduce the substrate concentration for the oxygen-reducing biocathode [61]. Thus, oxygen competition was studied by comparing the catalytic response in either air-equilibrated or N2-saturated PBS containing 3 mM lactate. The ratio of measured current density from air-/N2 solution was 87.7%, implying a relatively small fraction (12.3%) was assigned to oxygen depletion. This is consistent with previous studies that indicated that competition with oxygen was significant at low lactate concentrations and became negligible at high substrate concentrations [58]. By varying the lactate concentration, it became clear that NPG/Os(bpy)2PVI/LOx displayed a saturated current density when the concentration was above 3 mM (Figure 5.11C). In other words, oxygen competition at the bioanode was not a critical concern for the EBFC operation in the presence of 3 mM lactate (Figure 5.11C). The electrode had a sensitivity of 19.7±1.4 µA cm-2 mM-1 with a linear range up to 3 mM. The apparent Michaelis-Menten constant of the enzyme modified electrode was 1.0±0.4 mM, which is lower than that obtained from FcMe2-LPEI/LOx modified buckypaper

(1.6±0.1 mM) [21]. NPG/Os(bpy)2PVI/LOx showed considerable stability at +250 mV vs. SCE in an air-equilibrated 3 mM lactate PBS solution retaining 63% of its original response after 2 hours’ continuous operation (Figure 5.10).

- 121 -

Chapter 5

Figure 5.10 Operational stability of NPG/Os(bpy)2PVI/LOx (360 s deposition) in air-equilibrated PBS.

To prepare a biocathode with a high onset potential, BOx was covalently attached to a diazonium layer modified NPG (Figure 5.1) in order to achieve direct electron transfer (DET). The porous structure is believed to accommodate BOx in a manner that provides a favourable orientation for DET [36]. The catalytic activity of the BOx-diazonium modified NPG was studied by cyclic voltammetry in either N2- bubbled or air-equilibrated PBS at a scan rate of 5 mV s-1 (Figure 5.11D). In the air- equilibrated PBS, a sigmoidal catalytic curve was obtained with an onset potential of 503±15 mV vs. SCE and a background-corrected current density of 19.5±0.1 μA cm-2. The lower net current density obtained in comparison to the LOx based bioanode (62.2±1.9 μA cm-2) demonstrated that the output of the EBFCs was limited by the biocathode. The same electrode registered a net current density of 57.3±1.9 μA cm-2 in an O2-bubbled solution (Figure 5.11D), indicating a substrate concentration dependent catalytic behaviour. The obtained current density compares well with a previous report [36], where the oxygen reduction response could be enhanced by increasing the thickness of NPG up to 500 nm.

- 122 -

Chapter 5

various deposition times. (B) CVs of NPG/Os(bpy)2PVI/LOx (360 s deposition) in air-equilibrated 0.1 M pH 7.0 PBS at a scan rate of 5 mV s-1. (C) Catalytic response of

5.3.3. Performance of EBFC The performance of the assembled EBFC was tested in air-equilibrated PBS and artificial tear solution containing 3 mM lactate, respectively. The maximum power density achieved was 2.4±0.2 μW cm-2 at ca. 237 mV and a maximum short- circuit current density of 15.1±3.0 μW cm-2 in a PBS solution (Figure 5.12C). The obtained open-circuit voltage (OCV) was 455±21 mV in PBS, which is consistent with the difference in the onset potentials for lactate oxidation and oxygen reduction

- 123 -

Chapter 5 occurring at the bioanode and biocathode, respectively. However, the maximum power density and current density decreased to 1.7±0.1 μW cm-2 and 11.6±1.5 μA cm- 2 when tested in the artificial tear solution (Figure 5.12C). This likely arose from the interference of species such as ascorbate and the increased solution viscosity leading to mass transport resistance together with biofouling on the electrode surface caused by proteins. The anti-biofouling effect of NPG has been described in previous reports [62]. Significant resistances to fouling caused by BSA and fibrinogen were observed, due to the nanoporous structure preventing entry of large proteins into the nanoporous network.

Figure 5.12 Photograph of the contact lens encapsulated EBFC (A) and testing setup (B). (C) Polarisation and power curves for the EBFC consisting of

NPG/Os(bpy)2PVI/LOx bioanode and NPG-BOx biocathode. (D) Operational stability of the EBFC at 150 mV in artificial tears.

The decrease in performance was mainly attributed to interference by ascorbate as it can be easily oxidised on the nanostructured gold electrode [17, 30].

- 124 -

Chapter 5

The oxidation of ascorbate should be taken into account as it greatly perturbs the biocathode performance, although it improves the current observed at the bioanode [31]. Detailed investigation of ascorbate interference was performed on both the anode and cathode in air-equilibrated PBS solutions (Figure 5.13). Bare NPG displayed a faradaic response in 0.18 mM ascorbate with a current density of 7.86 μA cm-2 at 174 mV vs. SCE (inset of Figure 5.13A). The catalytic response decreased

-2 to 3.15 μA cm on NPG/Os(bpy)2PVI/LOx (Figure 5.13A), 5% of the catalytic current of 62.2±2.0 μA cm-2 to 3 mM lactate. This implies that the Os polymer did not act as a mediator for the oxidation of ascorbate and the coating layer restricted the interference of ascorbate to some extent. Figure 5.13B shows the effect of ascorbate upon the BOx cathode. A decrease of current density by 36% and a shift of the onset potential by 90±5mV, consistent with a decrease in OCV from 455±21 to 380±28 mV, were observed. Thus, we can conclude that ascorbate does not change the bioanode response greatly, but diminishes the biocathode performance in terms of onset potential and current response. Table 5.4 compares the performance of the proposed EBFC with previously reported tear based EBFCs. The maximum power density obtained was of the same order of magnitude as other lactate/O2 EBFCs [31,

32]. The maximum power density is ca. two times higher than that of a glucose/O2 EBFC due to the low glucose concentration in tear fluid [30]. While the OCV was ca.

100 mV less than that of a DET based glucose/O2 EBFC [30], and similar to the values reported for lactate/O2 EBFCs [31, 32]. The theoretical OCV for a lactate/O2 EBFC is 1.0 V [23], indicating that a higher OCV is possible. A mediator with a lower redox potential (that is still greater than that of FAD (ca. -0.43 V vs. SCE) [63]) could be used to prepare a LOx modified bioanode with a low onset potential. As described earlier, interference from ascorbate results in a decrease of the onset potential. Suppression of ascorbate interference, e.g. using a coating of Nafion [64], would also improve the OCV.

- 125 -

Chapter 5

Figure 5.13 Effects of the presence of 0.18 mM ascorbate towards bioanode (A) and biocathode (B). Inset of (A) shows the response on a bare NPG.

A stable response is a prerequisite for an applicable EBFC. The operational stability of the EBFC was examined by placing the NPG electrodes between two contact lenses (Scheme 5.1 and Figure 5.12B) with drop-by-drop supply of artificial tears containing 3 mM lactate, etc. The EBFC survived over a period of 5.5 h working at 150 mV (Figure 5.12D). The power density observed decreased significantly by more than 50% during the initial 1 hour period. The EBFC retained ca. 20% of the original output after 5.5 h. The stability decay was mainly due to the deterioration of the response of the BOx based biocathode [35], as the response of the Os polymer modified bioanode was robust (Figure 5.10). This stability is acceptable for use in one-day disposable lenses.

- 126 -

Chapter 5

Table 5.4 List of properties of tear based EBFCs.

OCV Pmax Anode Cathode Stability Ref. (mV) (µWcm-2)

More than 20 h AuNPs/CtCDH AuNPs/BOx 570 1 operational [30] 0.05 mM glucose Air-saturated half-life in human tears

77% loss for the AuNPs/TTF-TCNQ AuNPs/BOx first 1 hour in 540 3.1 [17] 0.665 mM ascorbate Air-saturated human lachrymal tears

An-pyr- 80% BP/poly- MWCNT/TBAB-modified loss for 4 MG/LDH/NAD+ 410±60 8.01±1.4 [31] Nafion/BOx hours in 3 mM lactate Solution: n/a artificial tear

An-pyr-

LOx/FcMe2-LPEI MWCNT/TBAB-modified 440 ± 80 2.4 ± 0.9 n/a [32] 3 mM lactate Nafion/BOx Air-saturated

~20% of initial PBS: PBS: power NPG/Os(bpy)2PVI/LOx NPG-diazonium-BOx 455±21 2.4±0.2 This retained after 3 mM lactate Air-equilibrated Artificial Artificial work 5.5 h tear: 380±28 tear: 1.7±0.1 operation in artificial tear

Note: CtCDH: Corynascus thermophilus cellobiose dehydrogenase; AuNPs: gold nanoparticles; TTF-TCNQ: tetrathiafulvalene-tetracyanoquinodimethane; BP: buckypaper; poly-MG: polymerized methylene green; FcMe2-LPEI: dimethylferrocene-modified linear polyethyleneimine; An-pyr: anthracene-pyrene; MWCNT: multi-walled carbon nanotube; TBAB: tetrabutylammonium bromide.

- 127 -

Chapter 5

5.4. Conclusions

Flexible NPG were successfully fabricated using an electrochemical dealloying method. The electrodes were modified with lactate oxidase and bilirubin oxidase for use as a lactate/O2 biofuel cell, which was subsequently tested in a solution of artificial tears. The flexible EBFC holds potential as an autonomous power supply for wearable electronic devices. Ascorbate interference, especially at the biocathode, were responsible for the decrease in performance in tears in comparison to the performance in phosphate buffer solution. A coating film on the biocathode may alleviate such interference effects. The response of the assembled EBFC was limited by current density and operational stability of the biocathode. Improvements in the observed current density of the biocathode will enable the development of a self-powered lactate biosensor on a contact lens, where the power density of the EBFC could be correlated with the concentration of lactate.

5.5 References

[1] M. Rasmussen, S. Abdellaoui, S.D. Minteer, Enzymatic Biofuel Cells: 30 Years of Critical Advancements, Biosens. Bioelectron., 76 (2015) 91-102. [2] D. Leech, P. Kavanagh, W. Schuhmann, Enzymatic Fuel Cells: Recent Progress, Electrochim. Acta, 84 (2012) 223-234. [3] S. Shleev, Quo Vadis, Implanted Fuel Cell?, ChemPlusChem, 82 (2017) 522-539. [4] A.T. Yahiro, S.M. Lee, D.O. Kimble, Bioelectrochemistry: I. Enzyme Utilizing Bio-Fuel Cell Studies, Biochim. Biophys. Acta, 88 (1964) 375-383. [5] R. Drake, B. Kusserow, S. Messinger, S. Matsuda, A Tissue Implantable Fuel Cell Power Supply, Trans. Amer. Soc. Artif. Intern. Organs, 16 (1970) 199-205. [6] P. Cinquin, C. Gondran, F. Giroud, S. Mazabrard, A. Pellissier, F. Boucher, J.-P. Alcaraz, K. Gorgy, F. Lenouvel, S. Mathé, A Glucose Biofuel Cell Implanted in Rats, PLoS ONE, 5 (2010) e10476. [7] V. Andoralov, M. Falk, D.B. Suyatin, M. Granmo, J. Sotres, R. Ludwig, V.O. Popov, J. Schouenborg, Z. Blum, S. Shleev, Biofuel Cell Based on Microscale Nanostructured Electrodes with Inductive Coupling to Rat Brain Neurons, Sci. Rep., 3 (2013) 3270-3281.

- 128 -

Chapter 5

[8] J.A. Castorena-Gonzalez, C. Foote, K. MacVittie, J. Halámek, L. Halámková, L.A. Martinez-Lemus, E. Katz, Biofuel Cell Operating in Vivo in Rat, Electroanalysis, 25 (2013) 1579-1584. [9] A. Zebda, S. Cosnier, J.P. Alcaraz, M. Holzinger, A. Le Goff, C. Gondran, F. Boucher, F. Giroud, K. Gorgy, H. Lamraoui, P. Cinquin, Single Glucose Biofuel Cells Implanted in Rats Power Electronic Devices, Sci. Rep., 3 (2013) 1516. [10] L. Halámková, J. Halámek, V. Bocharova, A. Szczupak, L. Alfonta, E. Katz, Implanted Biofuel Cell Operating in a Living Snail, J. Am. Chem. Soc., 134 (2012) 5040-5043. [11] K. MacVittie, J. Halamek, L. Halamkova, M. Southcott, W.D. Jemison, R. Lobel, E. Katz, From "Cyborg" Lobsters to a Pacemaker Powered by Implantable Biofuel Cells, Energy Environ. Sci., 6 (2013) 81-86. [12] D. Pankratov, L. Ohlsson, P. Gudmundsson, S. Halak, L. Ljunggren, Z. Blum, S. Shleev, Ex vivo Electric Power Generation in Human Blood Using an Enzymatic Fuel Cell in a Vein Replica, RSC Adv., 6 (2016) 70215-70220. [13] M. Cadet, S. Gounel, C. Stines-Chaumeil, X. Brilland, J. Rouhana, F. Louerat, N. Mano, An Enzymatic Glucose/O2 Biofuel Cell Operating in Human Blood, Biosens. Bioelectron., 83 (2016) 60-67. [14] M. Falk, D. Pankratov, L. Lindh, T. Arnebrant, S. Shleev, Miniature Direct Electron Transfer Based Enzymatic Fuel Cell Operating in Human Sweat and Saliva, Fuel Cells, 14 (2014) 1050-1056. [15] G. Göbel, M.L. Beltran, J. Mundhenk, T. Heinlein, J. Schneider, F. Lisdat, Operation of a Carbon Nanotube-based Glucose/Oxygen Biofuel Cell in Human Body Liquids-Performance Factors and Characteristics, Electrochim. Acta, 218 (2016) 278-284. [16] W. Jia, G. Valdés-Ramírez, A.J. Bandodkar, J.R. Windmiller, J. Wang, Epidermal Biofuel Cells: Energy Harvesting from Human Perspiration, Angew. Chem. Int. Ed., 52 (2013) 7233-7236. [17] M. Falk, V. Andoralov, M. Silow, M.D. Toscano, S. Shleev, Miniature Biofuel Cell as a Potential Power Source for Glucose-Sensing Contact Lenses, Anal. Chem., 85 (2013) 6342-6348. [18] S. Anastasova, B. Crewther, P. Bembnowicz, V. Curto, H.M.D. Ip, B. Rosa, G.-Z. Yang, A Wearable Multisensing Patch for Continuous Sweat Monitoring, Biosens. Bioelectron., 93 (2017) 139-145. [19] E.R. Berman, Biochemistry of the Eye, Springer Science & Business Media, New York, 1991. [20] D.A. Sakharov, M.U. Shkurnikov, M.Y. Vagin, E.I. Yashina, A.A. Karyakin, A.G. Tonevitsky, Relationship between Lactate Concentrations in Active Muscle Sweat and Whole Blood, Bull. Exp. Biol. Med., 150 (2010) 83-85. [21] D.P. Hickey, R.C. Reid, R.D. Milton, S.D. Minteer, A Self-powered Amperometric Lactate Biosensor Based on Lactate Oxidase Immobilized in Dimethylferrocene-modified LPEI, Biosens. Bioelectron., 77 (2016) 26-31. [22] A.J. Bandodkar, W. Jia, J. Wang, Tattoo-Based Wearable Electrochemical Devices: A Review, Electroanalysis, 27 (2015) 562-572. - 129 -

Chapter 5

[23] D. Pankratov, E. González-Arribas, Z. Blum, S. Shleev, Tear Based Bioelectronics, Electroanalysis, 28 (2016) 1250-1266. [24] H. Yao, A.J. Shum, M. Cowan, I. Lähdesmäki, B.A. Parviz, A Contact Lens with Embedded Sensor for Monitoring Tear Glucose Level, Biosens. Bioelectron., 26 (2011) 3290-3296. [25] H. Yao, Y. Liao, A.R. Lingley, A. Afanasiev, I. Lähdesmäki, B.P. Otis, B.A. Parviz, A Contact Lens with Integrated Telecommunication Circuit and Sensors for Wireless and Continuous Tear Glucose Monitoring, J. Micromech. Microeng., 22 (2012) 075007. [26] M. Leonardi, E.M. Pitchon, A. Bertsch, P. Renaud, A. Mermoud, Wireless Contact Lens Sensor for Intraocular Pressure Monitoring: Assessment on Enucleated Pig Eyes, Acta Ophthalmol., 87 (2009) 433-437. [27] M.X. Chu, K. Miyajima, D. Takahashi, T. Arakawa, K. Sano, S.-i. Sawada, H. Kudo, Y. Iwasaki, K. Akiyoshi, M. Mochizuki, K. Mitsubayashi, Soft Contact Lens Biosensor for in situ Monitoring of Tear Glucose as Non-invasive Blood Sugar Assessment, Talanta, 83 (2011) 960-965. [28] A.R. Lingley, M. Ali, Y. Liao, R. Mirjalili, M. Klonner, M. Sopanen, S. Suihkonen, T. Shen, B.P. Otis, H. Lipsanen, B.A. Parviz, A Single-pixel Wireless Contact Lens Display, J. Micromech. Microeng., 21 (2011) 125014. [29] C.-C. Peng, M.T. Burke, B.E. Carbia, C. Plummer, A. Chauhan, Extended Drug Delivery by Contact Lenses for Glaucoma Therapy, J. Control. Release, 162 (2012) 152-158. [30] X. Wang, M. Falk, R. Ortiz, H. Matsumura, J. Bobacka, R. Ludwig, M. Bergelin, L. Gorton, S. Shleev, Mediatorless Sugar/Oxygen Enzymatic Fuel Cells Based on Gold Nanoparticle-Modified Electrodes, Biosens. Bioelectron., 31 (2012) 219-225. [31] R.C. Reid, S.D. Minteer, B.K. Gale, Contact Lens Biofuel Cell Tested in a Synthetic Tear Solution, Biosens. Bioelectron., 68 (2015) 142-148. [32] R.C. Reid, S.R. Jones, D.P. Hickey, S.D. Minteer, B.K. Gale, Modeling Carbon Nanotube Connectivity and Surface Activity in a Contact Lens Biofuel Cell, Electrochim. Acta, 203 (2016) 30-40. [33] Y. Ding, Y.J. Kim, J. Erlebacher, Nanoporous Gold Leaf: “Ancient Technology”/Advanced Material, Adv. Mater., 16 (2004) 1897–1900. [34] X. Xiao, P. Si, E. Magner, An Overview of Dealloyed Nanoporous Gold in Bioelectrochemistry, Bioelectrochem., 109 (2016) 117-126. [35] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, A Symmetric Supercapacitor/Biofuel Cell Hybrid Device Based on Enzyme-Modified Nanoporous Gold: An Autonomous Pulse Generator, Biosens. Bioelectron., 90 (2017) 96-102. [36] T. Siepenkoetter, U. Salaj-Kosla, X. Xiao, P.Ó. Conghaile, M. Pita, R. Ludwig, E. Magner, Immobilization of Redox Enzymes on Nanoporous Gold Electrodes: Applications in Biofuel Cells, ChemPlusChem, 82 (2017) 553-560. [37] X. Xiao, E. Magner, A Biofuel Cell in Non-Aqueous Solution, Chem. Commun., 51 (2015) 13478-13480. - 130 -

Chapter 5

[38] T. Siepenkoetter, U. Salaj‐Kosla, X. Xiao, S. Belochapkine, E. Magner, Nanoporous Gold Electrodes with Tuneable Pore Sizes for Bioelectrochemical Applications, Electroanalysis, 28 (2016) 2415-2423. [39] T. Siepenkoetter, U. Salaj-Kosla, E. Magner, The Immobilization of Fructose Dehydrogenase on Nanoporous Gold Electrodes for the Detection of Fructose, ChemElectroChem, 4 (2017) 905-912. [40] F. Meng, Y. Ding, Sub-Micrometer-Thick All-Solid-State Supercapacitors with High Power and Energy Densities, Adv. Mater., 23 (2011) 4098-4102. [41] X. Xiao, J. Ulstrup, H. Li, J. Zhang, P. Si, Nanoporous Gold Assembly of Glucose Oxidase for Electrochemical Biosensing, Electrochim. Acta, 130 (2014) 559-567. [42] J. Pozuelo, V. Compañ, J.M. González-Méijome, M. González, S. Mollá, Oxygen and Ionic Transport in Hydrogel and Silicone-Hydrogel Contact Lens Materials: An Experimental and Theoretical Study, J.Memb. Sci., 452 (2014) 62-72. [43] E.M. Kober, J.V. Caspar, B.P. Sullivan, T.J. Meyer, Synthetic Routes to New Polypyridyl Complexes of Osmium(Ii), Inorg. Chem., 27 (1988) 4587-4598. [44] R.J. Forster, J.G. Vos, Synthesis, Characterization, and Properties of a Series of Osmium- and Ruthenium-containing Metallopolymers, Macromolecules, 23 (1990) 4372-4377. [45] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, J. Schindelin, I. Arganda- Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, NIH Image to ImageJ: 25 Years of Image Analysis, Nat. Methods, 9 (2012) 671. [46] S. Trasatti, O. A. Petrii, Real Surface Area Measurements in Electrochemistry, Pure Appl. Chem., 63 (1991) 711-734. [47] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, An Oxygen- Independent and Membrane-Less Glucose Biobattery/Supercapacitor Hybrid Device, Biosens. Bioelectron., 98 (2017) 421-427. [48] X. Xiao, H. Li, K. Zhang, P. Si, Examining the Effects of Self- Assembled Monolayers on Nanoporous Gold Based Amperometric Glucose Biosensors, Analyst, 139 (2014) 488-494. [49] M.D. Scanlon, U. Salaj-Kosla, S. Belochapkine, D. MacAodha, D. Leech, Y. Ding, E. Magner, Characterization of Nanoporous Gold Electrodes for Bioelectrochemical Applications, Langmuir, 28 (2011) 2251-2261. [50] J. Snyder, K. Livi, J. Erlebacher, Dealloying Silver/Gold Alloys in Neutral Silver Nitrate Solution: Porosity Evolution, Surface Composition, and Surface Oxides, J. Electrochem. Soc., 155 (2008) C464-C473. [51] Q. Zhang, X. Wang, Z. Qi, Y. Wang, Z. Zhang, A Benign Route to Fabricate Nanoporous Gold through Electrochemical Dealloying of Al–Au Alloys in a Neutral Solution, Electrochim. Acta, 54 (2009) 6190-6198. [52] K. Shao, C. Fang, Y. Yao, C. Zhao, Z. Yang, J. Liu, Z. Zou, An Easily Modified Method Using FeCl3 to Synthesize Nanoporous Gold with a High Surface Area, RSC Adv., 7 (2017) 18327-18332.

- 131 -

Chapter 5

[53] P. Delahay, M. Pourbaix, P.V. Rysselberghe, Potential‐pH Diagram of Silver Construction of the Diagram—Its Applications to the Study of the Properties of the Metal, its Compounds, and its Corrosion, J. Electrochem. Soc., 98 (1951) 65-67. [54] L.D. Burke, P.F. Nugent, The Electrochemistry of Gold: I The Redox Behaviour of the Metal in Aqueous Media, Gold Bull., 30 (1997) 43-53. [55] Z. Matharu, P. Daggumati, L. Wang, T.S. Dorofeeva, Z. Li, E. Seker, Nanoporous-Gold-Based Electrode Morphology Libraries for Investigating Structure– Property Relationships in Nucleic Acid Based Electrochemical Biosensors, ACS Appl. Mater. Interfaces, 9 (2017) 12959-12966. [56] F. Meng, X. Yan, J. Liu, J. Gu, Z. Zou, Nanoporous Gold as Non- Enzymatic Sensor for Hydrogen Peroxide, Electrochim. Acta, 56 (2011) 4657-4662. [57] L. Zhang, H. Chang, A. Hirata, H. Wu, Q.-K. Xue, M. Chen, Nanoporous Gold Based Optical Sensor for Sub-ppt Detection of Mercury Ions, ACS Nano, 7 (2013) 4595-4600. [58] T.J. Ohara, R. Rajagopalan, A. Heller, "Wired" Enzyme Electrodes for Amperometric Determination of Glucose or Lactate in the Presence of Interfering Substances, Anal. Chem., 66 (1994) 2451-2457. [59] T.-M. Park, E.I. Iwuoha, M.R. Smyth, R. Freaney, A.J. McShane, Sol- gel Based Amperometric Biosensor Incorporating an Osmium Redox Polymer as Mediator for Detection Of L-Lactate, Talanta, 44 (1997) 973-978. [60] Z. Gao, G. Binyamin, H.H. Kim, S.C. Barton, Y. Zhang, A. Heller, Electrodeposition of Redox Polymers and Co‐Electrodeposition of Enzymes by Coordinative Crosslinking, Angew. Chem. Int. Ed., 41 (2002) 810-813. [61] R.D. Milton, K. Lim, D.P. Hickey, S.D. Minteer, Employing FAD- Dependent Glucose Dehydrogenase within a Glucose/Oxygen Enzymatic Fuel Cell Operating in Human Serum, Bioelectrochem., 106, Part A (2015) 56-63. [62] J. Patel, L. Radhakrishnan, B. Zhao, B. Uppalapati, R.C. Daniels, K.R. Ward, M.M. Collinson, Electrochemical Properties of Nanostructured Porous Gold Electrodes in Biofouling Solutions, Anal. Chem., 85 (2013) 11610-11618. [63] K. Rabaey, W. Verstraete, Microbial fuel cells: novel biotechnology for energy generation, Trends Biotechnol., 23 (2005) 291-298. [64] S.L. O'Riordan, K. Mc Laughlin, J.P. Lowry, In vitro physiological performance factors of a catalase-based biosensor for real-time electrochemical detection of brain hydrogen peroxide in freely-moving animals, Anal. Methods, 8 (2016) 7614-7622.

- 132 -

CHAPTER 6:

A continuous fluidic bioreactor utilising electrodeposited silica for lipase immobilisation onto nanoporous gold

Chapter 6

Chapter 6. A continuous fluidic bioreactor utilising electrodeposited silica for lipase immobilisation onto nanoporous gold

6.1. Introduction

Immobilised enzymes [1] have been successfully used in applications such as biocatalysis [2, 3], biosensors [4] and biofuel cells [5-7]. Silicate materials including controlled pore glass (CPG), sol-gel derived silicate and mesoporous silicate (MPS) are biocompatible and widely used as solid supports for the immobilisation of enzymes [8]. Sol-gel derived silicate materials possess features that include ease of preparation, chemical inertness, negligible swelling and optical transparency [9]. In addition, the sol-gel process enables enzymes to be immobilised without significant losses in activity [10]. Electrodeposition provides a controllable and rapid route to grow uniform silica layers onto a conductive substrate regardless of the roughness of the surface [11]. The process of electrodeposition is initiated by an increase in pH in the proximity of the cathode as a consequence of hydrogen evolution, which consumes protons. Hydrolysis and condensation of precursors such as tetraethoxysilane (TEOS) are triggered by this change in pH [12], resulting in three- dimensional Si-O-Si networks that encapsulates enzyme in solution on the electrode/electrolyte interface (Scheme 5.1).

- 133 -

Chapter 6

Scheme 6.1 Schematic drawing of the electrodeposition of silica for enzyme immobilisation at a constant negative potential.

A wide range of enzymes including glucose oxidases [13] and dehydrogenases [14, 15] etc. have been successfully immobilised in electrochemically generated silica matrices for use as biosensors. Various electrochemically derived sol-gel silica/nanomaterial hybrids, such as carbon nanotubes [16], gold nanoparticles [17], and macroporous gold electrodes [11], have been prepared. Nanoporous gold (NPG), fabricated by etching gold alloys, has been used as a support material for the immobilisation of enzymes [18]. NPG possesses characteristic pores and ligaments whose sizes can be tailored by adjusting the dealloying conditions [19]. NPG has been modified with electrodeposited thin films of thiophene polymers [20] and Os modified polymers [5, 21] for the immobilisation of enzymes. In contrast to Os polymer modified electrodes that are prepared by drop-casting, electrodeposited films are more physically stable [22]. Electrodeposition enables control of the degree of modification of the entire surface [18] and the thickness of the deposited film can be tuned to optimise the enzyme loading [5]. Immobilisation makes it possible for enzymes to be used in a continuous flow mode [23]. In comparison to the conventional batch approach, flow methods are more efficient due to the large surface-to-volume ratio, ease of collection and on-line analysis of products [24]. Micro-reactors can be generally subdivided into three types: packed-bed, monolith and wall-coated reactors [25]. The former two can suffer from possible blockage and pressure gradients along the microchannels [26], making it

- 134 -

Chapter 6 difficult to tune flow dynamics. Inversely, wall-coated channels enable smoother flow with negligible mass transport resistance and thus predictable fluidic conditions. However, the catalyst loading of a wall-coated reactor is inferior to those of packed- bed and monolithic approaches [25], arising from the low working surface area. Porous supports can be used to improve the catalyst loading [27]. In this chapter, high-surface-area NPG was used as a solid support for enzymes in a wall-coated channel. To immobilise enzymes, NPG electrodes with an average pore size of ca. 30 nm [5, 28] were functionalised by the electrodeposition of silica with simultaneous encapsulation of lipase [29] (Scheme 5.1). The bio-modified electrodes were placed into a bespoke flow channel device (Scheme 5.2). The hydrolysis of 4-nitrophenyl butyrate (4-NPB) by lipase was used as a model system (Scheme 5.3). The catalytic performance was affected by the thickness of the silicate layer and by the concentration of lipase. The conversion of substrate to product depended on the flow rate and full conversion was feasible upon recycling the solution.

Scheme 6.2 (A) CAD drawing of the flow cell consisting of top-plate and base. (B) Sectional view of the base (unit: mm). (C) Detailed view of part 3 from (B). (D)

- 135 -

Chapter 6

Photograph of the flow cell during operation; the arrows indicate the direction of flow.

Scheme 6.3 Hydrolysis reaction of 4-NPB catalysed by lipase.

6.2. Experimental section

6.2.1. Materials and apparatus Potassium phosphate monobasic (≥99 %) and dibasic (≥98 %), hydrochloric acid (HCl, 37%), nitric acid (HNO3, 70%), sulfuric acid (H2SO4, 97%), 4-nitrophenyl butyrate (4-NPB, ≥98 %), Bradford reagent, bovine serum albumin (BSA, lyophilized powder, ≥96%) were obtained from Sigma-Aldrich Ireland, Ltd. Lipase (Sigma- Aldrich Ireland, Ltd) from Thermomyces lanuginosus (≥100 kUg-1, EC no.: 3.1.1.3) was supplied as a solution containing 73% (w/v) water, 25% (w/v) propylene glycol, 2% (w/v) lipase, and 0.5% calcium chloride. 4-(2-Hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES, ≥99 %) and tetraethoxysilane (TEOS, 99.9%) were purchased from Fisher Scientific, Ireland. All solutions were prepared with deionised water (resistivity of 18.2 MΩ cm) from an Elgastat maxima-HPLC (Elga Purelab Ultra, UK). NPG leaves were fabricated by etching ca. 100 nm thick Au/Ag leaf alloy films (12-carat, Eytzinger, Germany) in concentrated HNO3 for 30 min at 30 °C. The NPG films were then attached onto a pre-polished glassy carbon electrode (GCE) with

- 136 -

Chapter 6 a diameter of 4 mm and were allowed to dry. Prior to use, NPG electrodes were activated by scanning the potential over the range of -0.2 to 1.65 V in 1 M H2SO4 at a scan rate of 100 mV s-1 for 15 cycles. Electrochemical experiments were performed with a CHI802 potentiostat (CH Instruments, Austin, Texas) consisting of a NPG working electrode, a Pt wire counter electrode and an Ag/AgCl reference electrode. Samples mounted on 300- mesh copper grids (S147-3, Agar Scientific, UK) were characterised by transmission electron microscopy (TEM, JEOL JEM-2100, operating voltage of 200 kV). The average pore size and deposition layer thickness were measured with ImageJ software (National Institutes of Health, Bethesda, Maryland) [30]. Absorbance was recorded on a Cary 60 UV-Vis spectrophotometer (Agilent, USA).

6.2.2. Enzyme immobilisation procedures and activity measurement A typical silica sol was prepared by dissolving 2.125 g tetraethoxysilane (TEOS), 2 mL of deionised water and 2.5 mL of 0.01 M aqueous HCl, which were mixed for 12 h using a magnetic stirrer and then diluted 3 times with water for further use. 900 μL of lipase solution in 0.1 M pH 7.0 phosphate buffer solution (PBS) was mixed with 100 μL of the above hydrolysed sol. A range of lipase concentrations was used. As shown in Scheme 6.1, sol-gel electro-assisted deposition was performed at an applied potential of -1.1 V vs. Ag/AgCl for different durations. The same method was applied to deposit silica/lipase onto planar gold electrodes (diameter: 3 mm) for comparison. Regeneration of the modified NPG electrodes was achieved by polishing and cleaning the glassy carbon electrode, followed by attachment of a new NPG film and subsequent deposition of silica. All experiments were performed at room temperature (20±2 °C). The activity of immobilised lipase was evaluated by the hydrolysis of 4-NPB (Scheme 6.3). Flow experiments were performed by pumping buffer solution (10 mM pH 7.0 HEPES) containing 75 μM 4-NPB at various flow rates (Scheme 6.2D). The system was washed (at a relatively fast flow rate of 0.12 mL min-1 for 10 min) to allow the detachment of loosely-bounded enzymes prior to measurements. The absorbance of

- 137 -

Chapter 6 the product was measured at a wavelength of 420 nm. The stability of an electrode soaked in a solution of 4-NPB (2 mL, 75 μM) at room temperature was checked once a day. Absorbance changes at 420 nm were recorded by immersing the electrode into a fresh solution (2 mL, 75 μM 4-NPB) for one hour. To achieve 100% conversion of 4-NPB, the absorbance of the effluent solution was determined prior to recycling into the channel.

6.2.3. Flow cell design Devices were designed using SolidWorks 3D CAD software (Dassault Systèmes, 2017). The flow cell consisted of two parts: a top-plate and a base (Scheme 6.2A). The top-plate was mounted above the base and fixed with 10 screws (Scheme 6.2D). Four NPG modified GCEs were inserted into the vertical holes of the device, sitting flush on top of a 500 μm high channel (Scheme 6.2B). As shown in Scheme 6.2C, O-rings between the top-plate and base mitigated against any leakage of solution. A secondary function is that the action of tightening the top-plate onto the O-rings also caused the O-rings to squeeze the GCEs, effectively securing them in place on the top of the channel and preventing movement. An Instech P720 peristaltic pump was used to pump the solution into the channel (Scheme 6.2D) via round push- fit adaptors. The transition from a round adaptor to the laminar flow channel in the base was internally rounded to mitigate turbulence (Scheme 6.2C). A Stratasys Objet Connex 500 3D printer featuring PolyJet Matrix™ Technology was used to print the flow cell parts with a laminar flow channel of a depth of 500 μm. Printing was performed by successively laying down an acrylate based material (VeroClear RGD810) in 16 micron thick layers followed by UV curing and hardening of each layer. Exogenous material in cavities and overhangs was removed using a jet wash system. VeroClear RGD810 is a transparent material that facilitates removal of the support material from the flow channel by allowing visual inspection of the jet washing step. This material also allows the visual detection of air bubbles forming inside the channel during operation.

- 138 -

Chapter 6

6.2.4. Determination of immobilised protein concentration The amount of immobilised lipase into the silica film onto the electrode was determined by the Bradford assay [31] of the initial and residual enzyme content in the electrodeposition solution. A range of concentrations of BSA from 0.15 to 1 mg mL-1 in water was used to obtain a standard curve. The absorbances (595 nm) of mixtures of solutions of protein (50 uL) and of Bradford’s reagent (1.5 mL) were measured.

6.3. Results and discussion

Application of a potential of -1.1 V vs. Ag/AgCl initiates the sol-gel process (Scheme 6.1) via the production of hydroxyl ions that trigger the condensation of TEOS. The deposition time was optimised using a solution containing 1 mg mL-1 lipase. The effects of deposition time on the catalytic response are summarised in Figure 6.1A. In agreement with previous studies [5, 20], the response increased with time for deposition times less than 180 s, and can be ascribed to the increased amount of enzyme immobilised. For longer deposition times, the response decreased which likely arise from limitations in the supply of 4-NPB supply to lipase. TEM images indicate that the thicker films block the pores (Figure 6.2) [5]. A deposition time of 180 s was therefore utilised as a compromise between higher enzyme loading and mass-transport through the film. Similar phenomena were reported for the immobilization of D-sorbitol dehydrogenase [16] and of haemoglobin [11] in electrodeposited silica layers.

- 139 -

Chapter 6

Figure 6.1 (A) The effect of deposition time on the catalytic performance of NPG/silica/lipase obtained in 1 mg mL-1 lipase (measured by immersion in 2 mL of 75 μM 4-NPB for 0.5 h). (B) The effect of lipase concentration on the catalytic performance of NPG/silica/lipase obtained by depositing for 180 s (measured by immersion in 2 mL of 75 μM 4-NPB for 1 h).

TEM confirmed the formation of a silica film on NPG (Figure 6.2). NPG (dark region of the micrographs) preserved its porous structure after electrodeposition with gold ligands growing thicker (Figure 6.3). The relatively bright outer layer along the pores can be distinguished as the silica film with a uniform thickness. The optimal deposition time of 180 s resulted in a layer thickness of ca. 9.3±1.1 nm (in contrast to ca. 30 nm pores). A 60-s deposition time resulted in a 2.8±0.6 nm thick film (Figure 6.2A), while the pores were filled after a deposition time of 360 s (Figure 6.2C).

Figure 6.2 TEM images showing the NPG/silica/lipase obtained in 1 mg mL- 1 lipase with various deposition durations: (A): 60 s, (B): 180 s, (C): 360 s; the arrow in (A) and (B) indicates the silica/lipase layers. The pores are filled with silica/lipase in (C); Scale bars at the right-bottom of (A), (B) and (C) indicate 30 nm.

- 140 -

Chapter 6

Figure 6.3 SEM image of a bare NPG and NPG/silica/lipase (180 s deposition). The effect of lipase concentration was optimised using a deposition time of 180 s. The catalytic response increased at lipase concentrations below 3 mg mL-1 (Figure 6.1B), at which point the response leveled off. Similar results were reported for the sol-gel deposition of D-sorbitol dehydrogenase on macroporous gold electrodes [14]. Silica/lipase was also deposited on a planar gold electrode using the optimised protocol (3 mg mL-1 lipase, 180 s deposition time). Using the Bradford assay, the enzyme loading on planar gold was determined to be 1.8 nmol cm-2, in comparison to 3.6 nmol cm-2 on NPG. Despite a roughness factor of ca. 8 [28], NPG only encapsulated double the amount of lipase than that on the planar gold. This is likely due to the accelerated sol-gel process arising from the increased rate of hydrogen evolution that is observed on NPG in comparison to planar gold [32]. Similarly, Mazurenko et al. found that sol-gel derived bioelectrode performance was sensitive to the quantity of carbon nanotubes (CNTs) on the GCE as a high quantity of CNTs led to facilitated sol-gel process and faster silica film deposited [16]. When the stability of modified NPG and planar gold electrodes was examined (Figure 6.4), NPG modified electrodes retained 49.3% of the original response after 5 days’ storage, in comparison to 20.2% on planar gold. The observed decrease in activity likely arises from loss of enzyme activity and/or leakage of the enzyme. Lipase can operate as a catalyst in aqueous and nonaqueous solutions in a stable manner [2]. For example, lipase that was physically adsorbed on NPG showed negligible activity loss after 10 successive uses, and maintained 74% of its initial activity after 20 cycles [33]. Lipase immobilised onto SBA-15 retained 95% activity after 7 cycles [2]. It is thus - 141 -

Chapter 6 likely that the stability arises from leakage of the enzyme, especially given that sol-gel derived silica is generally brittle and can become cracked during long-term manipulation and use. The curved surface of NPG could provide a more stable environment for the film. Similar examples of enhanced stability on NPG have been observed. For example, electrodeposited MnO2 on NPG preserved 64% of its capacitance after 50 cycles of charge-discharge, in comparison to 26% for planar

Au/MnO2 [21].

Figure 6.4 Storage stability of NPG/silica/lipase; Response was measured by immersion in 2 mL of 75 μM 4-NPB for 1 h.

- 142 -

Chapter 6

The dependence of the catalytic response of a single NPG/silica/lipase electrode on flow rates was examined over the range 0.01 - 0.12 mL min-1 (Figure 6.5). The catalytic response decreased with increasing flow rate due to the decreased residence time, in agreement with previous studies of lipase immobilized in flow reactors [34, 35].

Figure 6.5 The effect of flow rate at the catalytic behavior of NPG/silica/lipase towards 75 μM 4-NPB.

- 143 -

Chapter 6

Four NPG/silica/lipase electrodes were mounted in the flow cell (Scheme 6.2D). A linear increase in the amount of substrate conversion was observed in the first four cycles with the level of conversion leveling off in subsequent cycles. As can be seen from Figure 6.6, the reaction was completed after eight cycles.

Figure 6.6 Conversion ratio of 2 mL of 75 μM 4-NPB by cycling in a loop at a flow rate of 0.05 mL min-1.

To examine enzyme leaching under flow conditions, a control experiment was performed by flushing the channel at a rate of 0.05 mL min-1 with blank PBS for 20 min. The solution collected (1 mL) was mixed with 1 mL of 75 μM 4-NPB and incubated for 0.5 h. An absorbance change of 0.0054 (at 420 nm) was observed corresponding to leaching of 0.36% of the enzyme. This result is consistent with the data in Figure 6.4 that showed a decrease in activity with time. This leaching likely arises from removal of the silica from the electrode.

- 144 -

Chapter 6

The NPG was physically attached on to the GCE and could be removed. Freshly prepared NPG leaves can be re-attached onto the GCE with newly generated silica layers to immobilise lipase. Figure 6.7 shows the normalised activity of 10 regenerated NPG/silica/lipase electrodes. A RSD of 4.6% demonstrated the reproducibility of the method.

Figure 6.7 Regeneration of an NPG/silica/lipase electrode. The response was measured by immersion of the electrode in 2 mL of 75 μM 4-NPB for 1 h.

6.4. Conclusions

In this study, direct incorporation of lipase into sol-gel derived silica onto NPG was proposed. The effects of electrodeposition time and lipase concentration in the electrolyte on the catalytic response were systematically evaluated. The porous structure of NPG had a remarkable, positive influence in enhancing operational stability, although only a two-fold increase for the initial catalytic response over that on a planar Au electrode supported silica/lipase. The laminar flow cell allowed the study of the effect of flow rate. Operated in a loop mode at a flow rate of 0.05 mL min-1 enabled the complete hydrolysis of 4-NPB (2 mL of 75 μM solution) after eight

- 145 -

Chapter 6 cycles. The underlying study is to integrate NPG supported enzymatic biofuel cells (EBFCs) with the bespoke flow cell.

6.5 References

[1] U. Hanefeld, L. Gardossi, E. Magner, Understanding enzyme immobilisation, Chem. Soc. Rev., 38 (2009) 453-468. [2] N.H. Abdallah, M. Schlumpberger, D.A. Gaffney, J.P. Hanrahan, J.M. Tobin, E. Magner, Comparison of mesoporous silicate supports for the immobilisation and activity of cytochrome c and lipase, J. Mol. Catal. B: Enzym., 108 (2014) 82-88. [3] R. DiCosimo, J. McAuliffe, A.J. Poulose, G. Bohlmann, Industrial use of immobilized enzymes, Chemical Society Reviews, 42 (2013) 6437-6474. [4] A. Heller, B. Feldman, Electrochemical Glucose Sensors and Their Applications in Diabetes Management, Chem. Rev., 108 (2008) 2482-2505. [5] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, A symmetric supercapacitor/biofuel cell hybrid device based on enzyme-modified nanoporous gold: An autonomous pulse generator, Biosens. Bioelectron., 90 (2017) 96-102. [6] D. Leech, P. Kavanagh, W. Schuhmann, Enzymatic fuel cells: Recent progress, Electrochim. Acta, 84 (2012) 223-234. [7] S. Calabrese Barton, J. Gallaway, P. Atanassov, Enzymatic Biofuel Cells for Implantable and Microscale Devices, Chem. Rev., 104 (2004) 4867-4886. [8] E. Magner, Immobilisation of enzymes on mesoporous silicate materials, Chem. Soc. Rev., 42 (2013) 6213-6222. [9] W. Jin, J.D. Brennan, Properties and applications of proteins encapsulated within sol-gel derived materials, Anal. Chim. Acta, 461 (2002) 1-36. [10] R. Gupta, N.K. Chaudhury, Entrapment of biomolecules in sol-gel matrix for applications in biosensors: Problems and future prospects, Biosens. Bioelectron., 22 (2007) 2387-2399. [11] F. Qu, R. Nasraoui, M. Etienne, Y.B.S. Côme, A. Kuhn, J. Lenz, J. Gajdzik, R. Hempelmann, A. Walcarius, Electrogeneration of ultra-thin silica films for the functionalization of macroporous electrodes, Electrochem. Commun., 13 (2011) 138-142. [12] R. Shacham, D. Avnir, D. Mandler, Electrodeposition of Methylated Sol-Gel Films on Conducting Surfaces, Adv. Mater., 11 (1999) 384-388. [13] W.-Z. Jia, K. Wang, Z.-J. Zhu, H.-T. Song, X.-H. Xia, One-Step Immobilization of Glucose Oxidase in a Silica Matrix on a Pt Electrode by an Electrochemically Induced Sol-Gel Process, Langmuir, 23 (2007) 11896-11900. [14] Z. Wang, M. Etienne, G.-W. Kohring, Y. Bon-Saint-Côme, A. Kuhn, A. Walcarius, Electrochemically assisted deposition of sol–gel bio-composite with co- immobilized dehydrogenase and diaphorase, Electrochim. Acta, 56 (2011) 9032-9040.

- 146 -

Chapter 6

[15] Z. Wang, M. Etienne, F. Quilès, G.-W. Kohring, A. Walcarius, Durable cofactor immobilization in sol–gel bio-composite thin films for reagentless biosensors and bioreactors using dehydrogenases, Biosens. Bioelectron., 32 (2012) 111-117. [16] I. Mazurenko, M. Etienne, O. Tananaiko, V. Zaitsev, A. Walcarius, Electrophoretically deposited carbon nanotubes as a novel support for electrogenerated silica–dehydrogenase bioelectrodes, Electrochim. Acta, 83 (2012) 359-366. [17] R. Toledano, D. Mandler, Electrochemical Codeposition of Thin Gold Nanoparticles/Sol-Gel Nanocomposite Films, Chem. Mater., 22 (2010) 3943-3951. [18] X. Xiao, P. Si, E. Magner, An overview of dealloyed nanoporous gold in bioelectrochemistry, Bioelectrochem., 109 (2016) 117-126. [19] T. Siepenkoetter, U. Salaj‐Kosla, X. Xiao, S. Belochapkine, E. Magner, Nanoporous Gold Electrodes with Tuneable Pore Sizes for Bioelectrochemical Applications, Electroanalysis, 28 (2016) 2415-2423. [20] X. Xiao, M.e. Wang, H. Li, P. Si, One-step fabrication of biofunctionalized nanoporous gold/poly (3, 4-ethylenedioxythiophene) hybrid electrodes for amperometric glucose sensing, Talanta, 116 (2013) 1054-1059. [21] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, An oxygen- independent and membrane-less glucose biobattery/supercapacitor hybrid device, Biosens. Bioelectron., 98 (2017) 421-427. [22] Z. Gao, G. Binyamin, H.H. Kim, S.C. Barton, Y. Zhang, A. Heller, Electrodeposition of Redox Polymers and Co‐Electrodeposition of Enzymes by Coordinative Crosslinking, Angew. Chem. Int. Ed., 41 (2002) 810-813. [23] S. Kundu, A.S. Bhangale, W.E. Wallace, K.M. Flynn, C.M. Guttman, R.A. Gross, K.L. Beers, Continuous Flow Enzyme-Catalyzed Polymerization in a Microreactor, J. Am. Chem. Soc., 133 (2011) 6006-6011. [24] Y. Asano, S. Togashi, H. Tsudome, S. Murakami, Microreactor technology: innovations in production processes, Pharm. Eng., 30 (2010) 32. [25] R. Munirathinam, J. Huskens, W. Verboom, Supported Catalysis in Continuous-Flow Microreactors, Adv. Synth. Catal., 357 (2015) 1093-1123. [26] K. Szymańska, M. Pietrowska, J. Kocurek, K. Maresz, A. Koreniuk, J. Mrowiec-Białoń, P. Widłak, E. Magner, A. Jarzębski, Low back-pressure hierarchically structured multichannel microfluidic bioreactors for rapid protein digestion – Proof of concept, Chem. Eng. J., 287 (2016) 148-154. [27] E.V. Rebrov, A. Berenguer-Murcia, H.E. Skelton, B.F.G. Johnson, A.E.H. Wheatley, J.C. Schouten, Capillary microreactors wall-coated with mesoporous titania thin film catalyst supports, Lab Chip, 9 (2009) 503-506. [28] X. Xiao, J. Ulstrup, H. Li, J. Zhang, P. Si, Nanoporous gold assembly of glucose oxidase for electrochemical biosensing, Electrochim. Acta, 130 (2014) 559- 567. [29] I. Itabaiana, L.S. de Mariz e Miranda, R.O.M.A. de Souza, Towards a continuous flow environment for lipase-catalyzed reactions, J. Mol. Catal. B: Enzym., 85 (2013) 1-9.

- 147 -

Chapter 6

[30] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, J. Schindelin, I. Arganda- Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, NIH image to imageJ: 25 years of image analysis, Nat. Methods, 9 (2012) 671. [31] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem., 72 (1976) 248-254. [32] X. Xiao, C. Engelbrekt, Z. Li, P. Si, Hydrogen evolution at nanoporous gold/tungsten sulfide composite film and its optimization, Electrochim. Acta, 173 (2015) 393-398. [33] X. Wang, X. Y. Liu, X. L. Yan, P. Zhao, Y. Ding, P. Xu, Enzyme- Nanoporous Gold Biocomposite: Excellent Biocatalyst with Improved Biocatalytic Performance and Stability, PLoS ONE, 6 (2011) e24207. [34] I. Itabaiana, F.K. Sutili, S.G.F. Leite, K.M. Goncalves, Y. Cordeiro, I.C.R. Leal, L.S.M. Miranda, M. Ojeda, R. Luque, R.O.M.A. de Souza, Continuous flow valorization of fatty acid waste using silica-immobilized lipases, Green Chem., 15 (2013) 518-524. [35] I. Denčić, S. de Vaan, T. Noël, J. Meuldijk, M. de Croon, V. Hessel, Lipase-Based Biocatalytic Flow Process in a Packed-Bed Microreactor, Ind. Eng. Chem. Res., 52 (2013) 10951-10960.

- 148 -

CHAPTER 7:

Conclusions and recommendations

Chapter 7

Chapter 7. Conclusions and recommendations

7.1. Conclusions

The objective of this thesis is to exploit the applications of dealloyed nanoporous gold (NPG) in bioelectrochemistry and in particular for use in biofuel cells to harvest energy. The fabrication and characterisation of NPG and enzyme- modified NPG is described in detail.

A membrane-less glucose/O2 biofuel cell that coupled a NPG/GOx bioanode with a NPG/BOx biocathode was examined in organic solvents containing a very small amount of water (≤5%) (Chapter 2). The power output of the cell decreased from 3.65 µW cm-2 to 0.47 µW cm-2 in 95% acetonitrile, with full restoration of the power output when the the electrodes were placed in aqueous buffer. Maximum power densities decreased with increasing log P in straight-chain monohydric alcohols (MeOH, EtOH, PrOH, BuOH, PeOH). The use of this BFC is limited to a small range of solvents due to the low solubility of glucose in non-aqueous media (generally not higher than 5 mM glucose in 95 % organic solvent). Possible applications exist for the generation of electricity from samples of pharmaceuticals and petrochemicals with low water content. A supercapacitor/enzymatic biofuel cell hybrid device was prepared by a facile, one-step electrodeposition of poly(3,4-ethylenedioxythiophene) (PEDOT)/Os polymer/enzyme onto NPG electrodes (Chapter 3). The dual-function properties of this hybrid device allowed the energy yielded by the biofuel cell to be stored in the supercapacitor and delivered at a significantly high-power pulse. For instance, it permitted a pulse current density of 2 mA cm-2, with an instant maximum power density of 609 μW cm-2, 468 times higher than that of the BFC. The modification layer showed reasonable stability without visible leakage of the redox mediators after 50 cycles operation at 0.2 mA cm-2 for approximately 7 hours. In contrast to the

- 149 -

Chapter 7 planar Au based system displaying an open-circuit voltage (OCV) of 365 mV, a maximum current density of 1.5 µA cm-2 and a maximum power density of 0.08 µW cm-2, NPG electrodes improved the performance in terms of lower resistance, higher bioelectrochemical signal (an OCV of 459.6±9.5 mV, a maximum current density of 28.9 µA cm-2 and a maximum power density of 1.3 µW cm-2) and capacitance. A proof-of-concept pulse generator (0.2 Hz pulse at 10 μA for 0.5 ms) to mimic a pacemaker was demonstrated using electrodes connected in series. An oxygen independent and membrane-less glucose biobattery/supercapacitor hybrid device delivering high-power-density pulses was presented in Chapter 4. MnO2 could replace oxygen reducing enzymes as the cathode, due to its features including the capability to be discharged at a reasonable potential, inert to glucose and insoluble reduced state. Most importantly, when only a fraction was discharged in the pulse mode, the spontaneous potential recovery of MnO2 occurred due to the redistribution of the oxidation states. Coupled with a flavin adenine dinucleotide-dependent glucose dehydrogenase (FAD-GDH) based supercapacitive bioanode, the hybrid device function in the similar way of a biosupercapacitor. This biobattery based hybrid device holds promise as an intermittent power source, which can overcome the limited oxygen supply occurring on the conventional biofuel cells, for implanted medical devices. In Chapter 5, flexible NPG were successfully fabricated using an electrochemical dealloying method. The electrodes were modified with lactate oxidase and bilirubin oxidase for use as a lactate/O2 biofuel cell, which was subsequently tested in a solution of artificial tears. The flexible biofuel cell holds potential as an autonomous power supply for wearable electronic devices. Ascorbate interference, especially at the biocathode, were responsible for the decrease in performance in tears in comparison to the performance in phosphate buffer solution. A coating film on the biocathode may alleviate such interference effects. The response of the assembled biofuel cell was limited by current density and operational stability of the biocathode.

- 150 -

Chapter 7

Chapter 6 proposed a direct incorporation of lipase into sol-gel derived silica onto NPG. The effects of electrodeposition time and lipase concentration in the electrolyte on the catalytic response were systematically evaluated. The porous structure of NPG had a remarkable, positive influence in enhancing operational stability, although only a two-fold increase for the initial catalytic response over that on a planar Au electrode supported silica/lipase. The laminar flow cell allowed the study of the effect of flow rate. Operated in a loop mode at a flow rate of 0.05 mL min-1 enabled the complete hydrolysis of 4-nitrophenyl butyrate (2 mL of 75 μM solution) after eight cycles. The bespoke flow cell has the potential to integrate NPG supported enzymatic biofuel cells.

7.2. Recommendations

Based on the conclusions drawn above, following recommendations are proposed for further investigation: Regarding to Chapter 2, future work on the use of enzymes (e.g. pyrroquinoline quinone-dependent alcohol dehydrogenase, PQQ-ADH) utilising substrates (e.g. EtOH) with high solubility in organic media would be of interest. PQQ- ADH may be immobilised onto NPG in a manner that enables direct electron transfer, thus allowing for the development of high-energy-density biofuel cells. Enzymatic biofuel cells operating in other non-aqueous solutions such as ionic liquids (proposed in 2009 [1]) are also of interest. Ionic liquids can be promising candidates as electrolytes of biofuel cells due to properties such as negligible vapour pressure, non-flammability, chemical and thermal stability, high ionic conductivity and excellent electrochemical stability. In addition, the oxygen solubility in ionic liquids can be higher than that in aqueous solution. However, the selction of suitable ionic liquids is a challenge given the deactivation effects on enzymes [2]. Chapter 3 and 4 show that potentials of the delivered pulses are limited by the open-circuit voltage of the biofuel cell or biobattery. There is still room to improve the power output by selecting redox mediators with appropriate redox potentials that are - 151 -

Chapter 7 close to that of redox centres of the enzymes for bioanode and biocathode, as well as fabricating biofuel cells undergoing direct electron transfer at both electrodes. The results described in Chapter 5 indicate that a self-powered lactate biosensor on a contact lens could be developed, where the power density of the EBFC could be correlated with the concentration of lactate [3]. Polymer films can be used to alleviate ascorbate interference. Furthermore, the thickness of NPG can be increased to up to 300 nm without detachment under bending, which can enhance the obtained roughness factor, and thus, current response. The laminar flow cell developed in Chapter 6 can be used for the study of fluidic biofuel cells and sequential enzymatic reactions. The biofuel cell can be constructed following the methodology in Chapter 3, as the electrodeposited Os polymer is quite robust without significant peel-off. Sequential enzymatic reactions using several different enzymes can be separately immobilized on the four-electrode system. The enzyme loading can be tuned to optimise overall reaction efficiency considering different enzymatic activity for each type of enzyme. A wearable platform consisting of biofuel cells as power sources powering sensor arrays [4] to monitor various health-relative parameters is the ultimate aim of this work. The sensor array comprises amperometric biosensors, potentiometric ion- selective electrodes, a pH sensor and a thermal sensor. The preparation of such an array of modified electrodes and the successful development of stable sensors are significant challenges. Collaborative efforts from perspectives of enzyme engineering, material design and device fabrication remain a very high priority to realize this goal.

7.3 References

[1] M. Armand, F. Endres, D.R. MacFarlane, H. Ohno, B. Scrosati, Ionic- liquid materials for the electrochemical challenges of the future, Nat Mater, 8 (2009) 621-629. [2] M. Bihari, T.P. Russell, D.A. Hoagland, Dissolution and Dissolved State of Cytochrome c in a Neat, Hydrophilic Ionic Liquid, Biomacromolecules, 11 (2010) 2944-2948.

- 152 -

Chapter 7

[3] D.P. Hickey, R.C. Reid, R.D. Milton, S.D. Minteer, A Self-powered Amperometric Lactate Biosensor Based on Lactate Oxidase Immobilized in Dimethylferrocene-modified LPEI, Biosens. Bioelectron., 77 (2016) 26-31. [4] W. Gao, S. Emaminejad, H.Y.Y. Nyein, S. Challa, K. Chen, A. Peck, H.M. Fahad, H. Ota, H. Shiraki, D. Kiriya, D.-H. Lien, G.A. Brooks, R.W. Davis, A. Javey, Fully integrated wearable sensor arrays for multiplexed in situ perspiration analysis, Nature, 529 (2016) 509.

- 153 -

Appendix 1:

First-Authored Publications

Bioelectrochemistry 109 (2016) 117–126

Contents lists available at ScienceDirect

Bioelectrochemistry

journal homepage: www.elsevier.com/locate/bioelechem

An overview of dealloyed nanoporous gold in bioelectrochemistry☆

Xinxin Xiao a,b,PengchaoSia,⁎,EdmondMagnerb,⁎ a Key Laboratory for Liquid-Solid Structural Evolution and Processing of Materials, Ministry of Education, School of Materials Science and Engineering, Shandong University, Jinan 250061, People's Republic of China b Department of Chemical and Environmental Sciences, Materials and Surface Science Institute, University of Limerick, Limerick, Ireland article info abstract

Article history: Nanoporous gold (NPG) obtained via dealloying of Au alloys has potential applications in a range of ﬁelds, and in Received 2 November 2015 particular in bioelectrochemistry. NPG possesses a three dimensional bicontinuous network of interconnected Received in revised form 23 December 2015 pores with typical pore diameters of ca. 30-40 nm, features that are useful for the immobilisation of enzymes. Accepted 30 December 2015 This review describes the common routes of fabrication and characterization of NPG, the use of NPG as a support Available online 31 December 2015 for oxidoreductases for applications in biosensors and biofuel cells together with recent progress in the use of NPG electrodes for applications in bioelectrochemistry. Keywords: Nanoporous gold © 2015 Elsevier B.V. All rights reserved. Glucose biosensor Electrochemistry Electron transfer

1. Introduction 2. Dealloyed NPG fabrication and characterization

Dealloyed nanoporous gold (NPG) is a porous material containing 2.1. Fabrication three dimensional frameworks of bicontinuous pores and ligaments, that is prepared by electro−/chemically dissolving the less noble Corrosion processing of alloys with preferential removal of the less component from Au alloys (Figs. 1 and 2A) [1].Forsuchmaterials noble component (i.e. dealloying), originally known as “depletion gild- to act as a support for enzymes, it is essential that the pore diameter ing”, was established by the Indians of pre-Columbian Central America sufficiently large to accommodate the enzyme and to enable effec- in order to create a layer of pure gold by etching copper from Cu/Au al- tive transport of substrate to the enzyme. The pore sizes of NPG can loys [20]. Forty and Pickering showed that dealloying of a binary alloy be tailored in the range from ~5 to 700 nm [2] by tuning the compo- resulted in a “spongy” morphology [20,21]. In 2001, Erlebacher et al. sition of the alloy and the dealloying conditions [3]. NPG possesses demonstrated that the formation of pores was due to an intrinsic dy- advantages such as high surface-to-volume ratio, good electrical namic formation process [22,23]. In the fabrication of NPG from Ag/Au conductivity, chemical stability, biocompatibility and permeability alloys, Ag atoms are dissolved under corrosive conditions such as con- [3]. NPG has been the subject of much attention for a range of appli- centrated nitric acid. At the solid/electrolyte interface, Au atoms aggre- cations including catalysis [4–6], optical sensing [7,8] and analysis [9]. gate into clusters and form islands rather than spreading over the NPG electrodes have been used in the development of immunosensors surface, allowing the formation of pores. Subsequently, newly formed [10–12], DNA sensors [13,14], enzymatic biosensors [15,16] and as pores open up regions of virgin alloy for further etching of Ag, allowing enzyme-free sensors [17–19]. the dealloying process to continue (Fig. 2A). In this mini-review, we describe the use of dealloyed NPG for According to the morphology of the precursor, dealloyed NPG can bioelectrochemical applications in biosensors and enzymatic biofuel form nanoparticles [24], nanowires [25,26], microwires [18] and films cells. The review focuses on NPG fabricated via dealloying methods, [3]. Bulk forms of NPG, such as wires and films are more convenient rather than porous gold electrodes obtained with other methods (tem- for manipulation. In 2004, Erlebacher et al. dealloyed commercially plate methods etc.). available 12-carat Ag/Au leaves (thickness of approximately 100 nm) to prepare crack-free NPG [3], which could be transferred and physically stabilized onto glassy carbon electrodes (GCEs) for further use (Fig. 2B ☆ This is a part of Special Issue BES2015. ⁎ Corresponding authors. and C). NPG leaves mounted on GCE have been widely adopted for a E-mail addresses: [email protected] (P. Si), [email protected] (E. Magner). range of electrochemical sensing applications [13,17]. However, the

Fig. 1. Main view (A) and side view (B) scanning electron microscope (SEM) images of NPG leaf (reprinted with permission from [69]). (C) Transmission electron microscope (TEM) image of NPG (reprinted with permission from [70]). (D) Contact mode atomic force microscope (AFM) images of NPG. Scan area: 500 × 500 nm2.(E) Height profile of a cross section along the line as indicated by a and b in (D) (reprinted with permission from [69]). (F) TEM image of a [01] nanopore and the corresponding diffraction pattern (inset). (G) High-angle annular dark- field-scanning TEM (HAADF-STEM) image of the labeled square indicated as b in (F). The intensity profile (inset) along the dotted line represents a stepped surface. (H) HAADF-STEM image of the labeled square indicated as d in (F) (reprinted with permission from [51]).

electrodes are brittle and can become detached from the GCE over time subsequent anodic potential scan, Zn was removed from the alloy, [27]. A mechanically robust NPG can be prepared by sputtering a Ag/Au resulting in NPG (Fig. 2E). alloy onto a glass support with subsequent dealloying of Ag, which greatly facilitates their ease of use [27]. The multilayered system was 2.2. Characterization comprised of a glass support, a ~ 10 nm thick Ti adhesion layer, a thin Au substrate layer (~35 nm) that improves adhesion and prevents de- The morphology (pore/ligament structure) and roughness of NPG lamination of the Ag/Au alloy layer during dealloying, and a ~ 100 nm are critical parameters for their successful utilisation. Microscopy thick Ag67Au33 alloy layer (Fig. 2D). Magnetron sputtering enables var- techniques such as scanning or transmission electron microcopy ious substrates (either conductive or non-conductive), variable alloy (SEM/TEM) and atomic force microscopy (AFM) can be used to ex- composition and pre-patterned deposition layers to be used. NPG can amine the morphology of NPG (Fig. 1). For NPG with uniform pore- also be fabricated through reconstruction of Au electrode surfaces by size-distribution and bicontinuous structures, the characteristic applying alternative potential scans in a solution containing ZnCl2 in sol- length scales (i.e. gold ligament width and pore channel diameter) vents such as benzyl alcohol [28], dimethyl sulfoxide [29] and in ionic can be determined by rotationally averaging the fast Fourier transform liquids [30,31]. Zinc was ﬁrst electrodeposited on Au to form a Zn/Au (FFT) power spectrum of a micrograph obtained by SEM or TEM [32]. alloy at an elevated temperature during cathodic scanning. In the TEM is particularly useful in characterizing surface modiﬁcation of X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126 119

obtained by oxidation and reduction of the outermost layer of gold

atoms. The roughness factor Rf, can then be calculated, assuming that a speciﬁc charge of 390 μCcm−2 is required for gold oxide reduction [37]. The measurement of the capacitance of the double- layer can also be used to determine the surface area [27].Spectro- scopic methods such as energy dispersive X-ray (EDX) spectroscopy and X-ray photoelectron spectroscopy (XPS) have been used to ana- lyzethechemicalcompositiontheelectrodes.XPSisusefulforthe analysis of NPG functionalized with biomolecules at relatively low concentrations [38,39].

3. Properties of dealloyed NPG for bioelectrochemical applications

Porous gold electrodes can be fabricated using various routes that range from the hard template route [40–43], dynamic hydrogen bubble template method [44], to the anodization of gold electrodes [45] etc. The preparation of template-directed macroporous gold electrodes has been reviewed [46,47]. Generally, hard template

Fig. 2. Cyclic voltammogram for a dealloyed NPG electrode in 1 M H2SO4 (scan rate: routes involve several steps: assembly of monodisperse spheres 0.1 V s−1). (e.g. polystyrene or silica with diameters ranging from 100 nm to 2 μm), electrodeposition of the metal and subsequent removal of the hard template (Fig. 5A) [40]. This approach allows precise con- NPG (e.g. with nanoparticles or thin films) [33–35]. For example, mod- trol the pore diameter and thickness of the porous structure, ification of NPG by MnO2 particles, thin films of tungsten sulfide, silica which are important parameters for bioelectrochemical applications and conducting polymer can be clearly distinguished by TEM (Fig. 3). [40,48]. In the dynamic hydrogen bubble process, the hydrogen bub- The immobilisation of bovine serum albumin (BSA) and immunoglobu- bles arising from the electrochemical reduction of H+ act as tem- lin G (IgG) on NPG has been examined using AFM [36], with individual plates for Au electrodeposition (Fig. 5B) [44].AnodizationofAu features of proteins observed on the surface of the support. Electro- takes place in an oxalate containing solution at a high potential chemical techniques, such as cyclic voltammetry (CV), are primarily ap- (Fig. 5C) [45], with the nanopores generated by the formation of a plied to measure the electroactive surface areas (Fig. 4). During carbonaceous passivation film and its subsequent breakdown. Each potential cycling of NPG in sulfuric acid solution, anodic peaks between route results in a porous gold material with specific morphology 1.1–1.4 V (vs. SCE) and a well-defined cathodic peak at ~0.9 V are (Fig. 5 and Table 1).

Fig. 3. (A) Bright-field TEM image of NPG/MnO2 hybrid materials for electrochemical supercapacitors (reprinted with permission from [35]). (B) HRTEM image of the NPG/MnO2 hybrid showing nanocrystalline MnO2 with a grain size of ~5 nm. (C) HAADF-STEM image taken from a gold/MnO2 interface region. (D) HRTEM image of the NPG/amorphous tungsten sulfide for electrochemical hydrogen evolution (reprinted with permission from [34]). (E) TEM image of NPG decorated with a thin glucose oxidase doped silica film generated by electro-assisted sol–gel process (unpublished results). The enzyme was catalytically active. (F) TEM image of NPG/poly(3,4-ethylenedioxythiophene) biocomposite with glucose oxidase entrapped in the film (reprinted with permission from [33]). 120 X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126

Fig. 4. (A) Schematic diagram of the porosity evolution process during dealloying of a Ag/Au alloy (modified from [6]). (B) 12-carat Ag/Au (white gold) leaf. (C) NPG leaf attached onto a GCE with a diameter of 4 mm. (D) Diagram of the layered Ag/Au alloy film obtained by magnetron sputtering (not to scale), modified from [27]. (E) Schematic diagram of NPG fabrication using in-situ electrodeposition and dealloying, modified from [28].

NPG obtained via dealloying methods possess a number of interest- lipase-NPG biocomposite retained 33 and 38% of activity in ing features for applications in bioelectrochemistry. Four main points ethyl acetate and chloroform, respectively. are summarized below: Secondly, the high content of low index crystalline faces [51] on the Firstly, enzymes confined within NPG display higher stability on surface of NPG (Fig. 1G, H) can enhance the rate of electron transfer, exposure to high temperatures [49,50] and organic solvents [38] enabling significantly enhanced electrochemical responses. Exam- due to the nanopore providing a protective environment to the ples include glucose [52], hydrogen peroxide [17], nitrophenol enzyme. After incubation at 50 °C for 2 h, only 6% of the initial ac- [53], hydrazine [54], nitrite [55], etc. The electrocatalytic response tivity of free laccase remained in comparison to 60% for laccase to glucose and hydrogen peroxide is discussed in more detail in immobilised on NPG [49]. The activity of xylanase immobilised Section 4. on NPG was assayed at 50 °C and exhibited only a 25% loss Thirdly, thiol self-assembled monolayers (SAMs) are more stable on after 10 cycles [39]. After incubation for 30 min at 50, 60 and NPG than on planar Au electrodes [56,57], as the monolayers benefit 70 °C, 62, 59 and 54% of the initial activity was observed for from the presence of defective sites, lattice strain and residual Ag on free lipase, compared with 75, 74, and 76%, respectively, for li- the ligament surface. In alkaline media, the peak potential for SAM pase adsorbed on NPG (Scheme 1A) [38]. After treatment with desorption from dealloyed NPG occurred at a more negative poten- ethyl acetate (log P of 0.68), 27% of the initial activity of free li- tial (−1.15 V vs. SCE) versus that observed at planar gold (−0.74 V) pase remained, while 81% activity was observed for the lipase- [56], indicative of a stronger Au-S bond. NPG biocomposite [38]. On exposure to chloroform (log P of Fourthly, NPG can exclude large proteins from accessing the internal 2.0) free and immobilised lipase retained 42 and 85% of the ini- pores, reducing the effects of biofouling [58,59]. In the presence of − tial activity, respectively. After 20 cycles of treatment, the 2mgmL 1 of BSA [58], the time for the initial electrochemical X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126 121

Fig. 5. Schematic diagram of hard template route (A), dynamic hydrogen bubble template method (B) and anodization of gold electrodes (C). SEM images of assembly of ﬁve layers of 680 nm silica nanoparticles on electrode (D) and the ﬁnal macroporous gold structure obtained (E) (reprinted with permission from [40]). The SEM images of porous Au via dynamic hydrogen bubble template method (F) (Copyright 2015, reprinted from [44]) and anodization (G) (reprinted with permission from [45]).

3− response of [Fe(CN)6] to decrease by 50% was 3 min on planar available glucose biosensors utilise GOx due to its relatively low price, gold, 12 min on macroporous gold (1200 nm pore network, obtained high selectivity and stability [62,63].GOxisaflavoprotein that catalyzes β δ via hard templating of latex spheres), and 38 min on hierarchical gold the oxidation of -D-glucose to D-glucono- -lactone, which is then fl (1200/60 nm bimodal pore network). In contrast, at dealloyed NPG hydrolyzed to gluconic acid [64]. The redox active group in GOx, avine adenine dinucucleotide (FAD), is concomitantly reduced to FADH . (~5–50 nm pore, R of 15), the current decreased by ca. 12% after 60 2 f GOx(FAD) can be regenerated using molecular oxygen, electron media- mi, indicative of a significant resistance to fouling. On exposure to a so- tors or directly at the electrode surface. Electron transfer (ET) can occur fi −1 lution of brinogen (1 mg mL )fortwohours[60], the faradaic peak via mediated (MET) or direct ET (DET). The latter process has the advan- associated with the oxidation of ascorbic acid disappeared at planar tages that it avoids possible leakage and diffusional limitations associated gold electrodes but remained at NPG. Collectively, the voltammetric with the use of mediators. However, the redox active centers of GOx are 3− response to [Fe(CN)6] and ascorbic acid, displaying fast and slow deeply buried with a minimal distance of 13–18 Å between the periphery ET kinetics, respectively, was sensitive to surface contamination at pla- of GOx and the N7 nitrogen of the isoalloxazine rings of FAD [65].Achiev- nar gold but was significantly less affected at NPG. ing DET of GOx (third-generation glucose biosensors [63]) has been ham- pered by requirements including optimization of enzyme orientation and minimization of the distance between the active center of the enzyme 4. Glucose biosensors and the surface of the electrode. When oxygen is used as a mediator, hydrogen peroxide generated in 4.1. Enzyme based glucose biosensors the process can be detected via either oxidation or reduction. A potential of +0.7 V (vs. SCE) is required in order to observe a faradaic re-

The ﬁrst glucose biosensor utilised glucose oxidase (GOx) sponse to H2O2 at planar gold electrodes, a rather high detection immobilised on a platinum electrode [61]. The majority of commercially potential which may cause interference from other species. In contrast,

Table 1 Comparison of porous gold materials fabricated with different methods.

Method Substrate Morphology Pore size

Dealloying Au or Si, glass etc. Bicontinuous pores and ligaments; Randomly distributed and interconnected pores ~5–700 nm Hard template route Au or Pt Periodically arranged pores; Interconnected pores 100–2000 nm Dynamic hydrogen bubble template method Au or Pt Open structure with larger pores on the top of smaller ones; Interconnected pores ~1–30 μm Anodization Au Spongelike conﬁguration ~20 nm 122 X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126

Scheme 1. Methods of immobilising enzyme on NPG, including physical adsorption of lipase (A) (reprinted with permission from [38]), covalent bonding in the assistance of SAM (B) (reprinted with permission from [69]) and entrapment within conducting polymers (C) (reprinted with permission from [33]).

NPG can oxidize H2O2 at an onset potential of +0.2 V (vs. SCE) and a glucose monitoring technology [71,72]. By enveloping the redox en- peak potential of +0.4 V (vs. SCE) [66]. According to Erlebacher et al. zymes with redox polymers, electron communication of redox centers [67], the primary reason that NPG displays a higher catalytic activity of enzymes and polymers with electrode surfaces can be established than planar gold towards the reduction or oxidation of H2O2 is due to [73]. Encapsulation of GOx within Os-redox polymer onto NPG has the higher concentration of steps on NPG (Fig. 1G, H). A glucose biosen- been employed as a glucose biosensor (with a detection limit of 2.0 sor has been developed by physical adsorption of GOx within the NPG (±0.1) μM [74]), and for a glucose/O2 biofuel cell [75]. Drop casting of followed by covering the electrode with a Nafion film. The electrode redox hydrogel/enzyme mixture onto porous electrodes generally displayed a linear range up to 18 mM, a sensitivity of 0.7 μAcm−2 mM−1 leads to a “cap” like configuration [76], which hampers the full occupan- and a detection limit of 196 μM (applied potential +0.4 V vs. SCE) [66]. cy of deeper pores. Electrodeposition of a redox polymer is recommend-

The reduction of H2O2 at NPG occurs at an onset potential of −0.1 V (vs. ed to enable the generation of a stable and uniformly distributed film Ag/AgCl) [15,17]. Responses to H2O2 reduction at NPG with pore sizes of [77]. 18, 30, 40 and 50 nm have been compared, and the electrodes with The nicotinamide adenine dinucleotide (NAD) dependent enzyme smaller pores yielded higher sensitivity, arising from the larger surface glucose dehydrogenase (GDH) is also used for the detection of glucose area of these electrodes. After immobilisation of GOx via thiol linking s The oxidation of NADH requires a high overpotential [78]. NPG can en- and at an applied potential of −0.2 V (Ag/AgCl), the highest sensitivity hance the sluggish ET kinetics of NADH oxidation [66], with a peak po- (8.6 μAcm−2 mM−1) to glucose was obtained from NPG electrodes tential for the oxidation of NADH of +0.52 V (vs. SCE), while +0.72 V with pore sizes of 30 nm [15], suggesting that the small pores of on planar gold. Alcohol dehydrogenase (ADH) modified NPG electrodes 18 nm NPG were not sufficiently large to accommodate higher can operate at an applied potential of +0.5 V (vs. SCE) for the detection amounts of enzyme. Prussian Blue (PB) was electrodeposited onto of ethanol [66], though this potential is still high for many practical ap-

NPG for enhanced activity to H2O2 reduction [68].Atapotentialof plications. The flavocytochrome cellobiose dehydrogenase (CDH) can 0 V (vs. Ag/AgCl), a linear response up to 30 mM glucose with a sen- be utilised in the development of third-generation glucose biosensors sitivity of 50 μAcm−2 mM−1, as well low levels of interference from [79]. CDH processes two separate domains connected via a polypeptide lactate, uric acid and ascorbic acid was obtained. linker region. The flavodehydrogenase domain is catalytically active and Mediators such as p-benzoquinone (BQ) and ferrocenecarboxylic the cytochrome domain with a haem b group acts as an electron relay, acid (FCA) have been used to oxidise GOx immobilised on NPG elec- i.e. built-in mediator. This dual-domain feature enables efficient trodes (Scheme 1B) [69,70]. In the presence of mediators that under- DET between the active site of the enzyme and the electrode surface. go diffusion controlled and fast kinetic redox reactions, the enzymes NPG modified with SAMs of 1-thioglycerol can promote the DET of immobilised deeply inside the NPG electrode do not contribute to Corynascus thermophiles CDH (CtCDH) [80]. CtCDH immobilised the detection of glucose, as the redox reaction occurs at the external with an Os-redox polymer on NPG displayed a limit of detection of regions of the pores. Redox couples that display with fast heteroge- 16 (±0.1) μM for the determination of glucose [74]. The main con- 4−/3- 6 3−/2- neous kinetics, such as [Fe(CN)6] and [Ru(NH3) ] ,show straint of CDH for blood glucose sensing is its promiscuous characteris- similar responses on NPG [27]. Electrodes utilising such mediators do tic since it is capable of catalyzing several carbohyrates. However, it may not display significant increases in the current, similarly, diffusional be useful in the analysis of glucose in samples where glucose is present limitations arising from substrate diffusion to enzyme molecules within in much higher concentrations. the pores results in responses that are similar to those obtained on pla- Practical application of glucose biosensors requires testing in real nar supports. samples including blood, serum or other human body fluids (e.g. saliva, In the presence of excess levels of mediator, the sensitivity of a urine and sweat). However, operation in blood or serum results in sensor is determined by the amount of accessible immobilised en- fouling of the electrode arising from protein adsorption on the zymes [69]. Entrapment is one way to increase enzyme loading. electrode surface, leading to a decreased sensor response [81].As 3− One-step electropolymerization of a hybrid film of poly(3,4- already mentioned, the voltammetric response of [Fe(CN)6] in ethylenedioxythiophene) (PEDOT) and GOx on nanoporous gold phosphate buffer containing 2 mg mL−1 of BSA was stable on NPG for (NPG) has been described (Scheme 1C) [33]. Using BQ as a mediator, 22hbutnotonaplanargoldelectrode[58]. Glucose biosensors the biosensor, exhibited a sensitivity of 7.3 μAcm−2 mM−1 and a linear should be capable of operating in whole blood samples [82]. To date, response over the concentration range 0.1–15 mM. Redox hydrogel NPG based biosensors are mostly tested in 10–100 fold diluted serum “wired” glucose oxidase been very successful in the development of [12,70,83,84]. X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126 123

4.2. Nonenzymatic glucose biosensors occur [89]. To address this issue, a thin film (~5 nm) of CuO has been electrodeposited onto NPG with a sensitivity of 374 μAcm−2 mM−1 to- The development of numerous nonenzymatic glucose sensors, based wards glucose and a linear range up to 12 mM obtained operation in al- on direct electrooxidation of glucose, has been described [85]. NPG is kaline media [90]. promising as it can promote the relatively sluggish kinetics of glucose By decorating NPG wires with cobalt oxide, the resultant nonenzy- oxidation and its high specific surface area is less prone to interference matic glucose biosensor had a high sensitivity (2500 μAcm−2 mM−1) from ascorbic acid and uric acid. Au-OHads layers, formed by adsorption and a low detection limit of 5 nM [18]. In both metal oxide modification − of OH on gold electrodes in a neutral or alkaline solution act as “pre- cases (NPG/CuO and NPG/Co3O4), however, no response to glucose was oxidation precursors” and are active in the oxidation of glucose at low observed at neutral pH. Thus dilution of serum samples in alkaline potentials [86]. The reaction pathways are shown in Eqs. (1) to (4) media is generally required [90]. Although there have been a wide [52,87]: range of publications on the use of transition metal oxides as the basis for nonenzymatic glucose sensors, their applications for the determina- þ −→ − ðÞ1−λ − þ λ − ð Þ fi Au OH Au OHads e 1 tion of glucose concentration are limited due to their lack of speci city and the need to operate in alkaline media. Direct oxidation of glucose In neutral media: with inorganic catalysts, coupled with a Pt based cathode for oxygen reduction, are more applicable in alkaline glucose fuel cells or enzyme- − þ Au−OHads þ Glu cose→Gluconolactone þ Au þ e þ H2O þ H ð2Þ free (abiotic) glucose fuel cells [91,92].

In neutral media: 5. Bioelectrochemical applications of NPG

− − Au−OHads þ Glu cose þ OH →Gluconolactone þ Au þ e þ 2H2O ð3Þ 5.1. Heme-proteins

Gluconolactone→Gluconic acidðÞ hydrolysis ð4Þ Cytochrome c (cyt c) is a model protein that can undergo DET at SAM modified. The topography of gold electrodes were shown to play an im- The presence of residual amounts of Ag on NPG significantly en- portant role in the adsorption and electrochemical response of cyt c [93]. hances the electrooxidation of glucose in alkaline solutions, as Ag pro- Various substrates with different surface roughness including evaporat- motes the chemisorption of OH− at active Au atoms [52]. In 0.1 M ed, bulk, single crystal, and epitaxially grown gold on mica were studied. NaOH, planar gold displayed an onset potential of −0.6 V (vs. SCE) for On a Au electrode with smoother surface, SAMs exhibited an increased glucose oxidation, while −0.9 V (vs. SCE) for NPG with 12 nm pores ability to block a diffusing probe molecule, implying a lower level of de- [88]. The catalytic current density of NPG with 12 nm pores was over fectiveness. Moreover, the extent of adsorption and the electrochemical 1.5 and 17 times higher than that of NPG with 35 nm pores and planar response of the adsorbed cyt c decreased significantly on smooth sub- gold. Similar pore size effects can be seen when working in neutral strates. Thus, rough gold surfaces, resulting in SAMs with a high degree pH, with a sensitivity of 20.1 μAcm−2 mM−1 and linear range up to of defects, were more suitable for the optimal faradaic response [93].

18 mM. NPG with 18 nm pores gave rise to an improved response to glu- Well-deﬁned, nearly symmetric voltammograms (peak separation ΔEp cose in comparison to NPGs with pores of 30, 40 and 50 nm in size [89]. of 18 ± 1 mV) were obtained by covalently attaching cyt c on NPG mod- Poisoning of NPG by the irreversible adsorption of chloride ions can iﬁed with a mixed SAM [27]. The surface coverage of active cyt c on a

Scheme 2. Schematic diagram of a NPG based glucose/O2 BFC fabricated by drop-casting a solution of osmium redox polymer, enzyme and cross-liners. Reproduced from [75] with permission from The Royal Society of Chemistry. 124 X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126

NPG electrode (Rf of 28) was ~11 times higher than that achieved at a obtained on planar gold and 2½ sphere gold macroporous electrodes, planar Au surface. The adsorbed proteins were stable and no decreases respectively. in the voltammetric response with continuous potential cycling (30 Trametes hirsute laccase (ThLc) showed well-defined DET at unmod- scans) in aqueous buffer were observed. ified NPG electrodes, in contrast to the absence of a response at unmodified polycrystalline gold electrodes [99]. The significant response of laccase obtained on NPG was proposed to arise from preferential orien- 5.2. Multi-copper oxidases tation of enzymes onto the surface of NPG, decreasing the distance for ET to the T1 copper sites. On covering the electrode with an epoxy Multi-copper oxidases (MCOs, e.g. laccase and bilirubin oxidase cap, significantly higher currents (by a factor of 10) were obtained. (BOD)) modified electrodes have been extensively employed as Higher responses were obtained when the temperature was increased cathodes of enzymatic biofuel cells (BFC) [94], which are of significant from 20 to 37 °C. The current obtained from DET was 30% of that interest due to their potential applications as autonomous power observed for ThLc co-immobilised with an osmium redox polymer, suppliers [95,96]. Generally, MCOs contain four copper atoms: the T1 indicating that higher amounts of enzyme are electrochemically ad- copper site acts as an electron acceptor while O2 is reduced to H2Oat dressable in the presence of the redox polymer. Moreover, pore size the T2/T3 copper sites [97]. NPG based BFCs have been constructed dependent performance has also observed when laccase was physically based on GOx or CDH at the anodes and laccase or BOD at the cathode adsorbed on different NPG surfaces (10–20 nm, 40–50 nm and 90– [57,74,75]. The electrodes displayed improved power densities and 100 nm, a larger pore size typically refers to smaller surface area) [49]. stabilities in comparison to planar electrodes. For example, glucose/O2 Aporesizeof40–50 nm resulted in the largest amount of immobilised BFCs operating via mediated ET at a GOx anode and BOD cathode, laccase, as mass transport of the enzyme to the inner pores was lim- displayed maximum power densities of 35 vs. 11 μWcm−2 in 100 mM ited. A layer-by-layer assembly of AuNPs and laccase onto ordered glucose at osmium polymer-modified nanoporous and planar gold macroporous gold electrodes enabled the DET of laccase, with a for- (Scheme 2), respectively [74]. Using GOx and laccase with Os redox poly- mal potential of 0.25 V vs. Ag/AgCl, which was very close to the the- mers [98], maximum power densities of 17 and 38 μWcm−2 were oretical value for the T2 copper site (0.21 V) [100].

Fig. 6. Schematic diagram of a cross-section of NPG electrodes with adsorbed MvBOD without (A) and with P017-epoxy caps (C) and their electrochemical behaviors (B, D) towards Ar

(dotted line) and O2 (full and dashed line for the ﬁrst and second scan), respectively. Inset of (B): SEM image of the surface of a NPG electrode. Inset of (D): the proposed structure of P017-epoxy. Conditions: 0.1 M pH 7.0 citrate–phosphate buffer, scan rate of 5 mV s−1. Reproduced from [76] with permission. X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126 125

Previous reports have described DET of BOD on bare or negatively [11] Q. Wei, Y. Zhao, C. Xu, D. Wu, Y. Cai, J. He, H. Li, B. Du, M. Yang, Nanoporous gold fi fi lm based immunosensor for label-free detection of cancer biomarker, Biosens. charged SAMs modi ed single-crystal electrodes [101] and three- Bioelectron. 26 (2011) 3714–3718. dimensional AuNP electrodes [102]. BOD was physically embedded [12] M. Li, M. Zhang, S. Ge, M. Yan, J. Yu, J. Huang, S. Liu, Ultrasensitive into bare NPG and the resulted electrodes displayed bioelectrocatalytic electrochemiluminescence immunosensor based on nanoporous gold electrode fi and Ru-AuNPs/graphene as signal labels, Sensors Actuators B Chem. 181 (2013) reduction of molecular oxygen, undergoing ef cient DET (Fig. 6) [76]. 50–56. However, BOD was relatively loosely bound and easily removed from [13] K. Hu, D. Lan, X. Li, S. Zhang, Electrochemical DNA biosensor based on nanoporous the surface of the electrode as only 50% of potential scan was retained gold electrode and multifunctional encoded DNA-Au bio bar codes, Anal. Chem. 80 – in second scan. On coating the electrode with an epoxy film by creating (2008) 9124 9130. [14] P. Daggumati, Z. Matharu, L. Wang, E. Seker, Biofouling-resilient nanoporous gold covalent bonds with nucleophilic groups such as amine, thiol and electrodes for DNA sensing, Anal. Chem. 87 (2015) 8618–8622. hydroxyl groups on the surface of the enzyme, current densities were [15] L.Y. Chen, T. Fujita, M.W. Chen, Biofunctionalized nanoporous gold for electro- – stable on repeated potential scans. As a result, a minimal inhibition of chemical biosensors, Electrochim. Acta 67 (2012) 1 5. − [16] C. Wu, H. Sun, Y. Li, X. Liu, X. Du, X. Wang, P. Xu, Biosensor based on glucose F with ca. 3.5% activity loss was achieved on a biocathode stabilized oxidase-nanoporous gold co-catalysis for glucose detection, Biosens. Bioelectron. − with epoxy, indicating the exclusion of F from the pores by the poly- 66 (2015) 350–355. mer and also the confinement of the enzyme in the pores of the elec- [17] F. Meng, X. Yan, J. Liu, J. Gu, Z. Zou, Nanoporous gold as non-enzymatic sensor for − hydrogen peroxide, Electrochim. Acta 56 (2011) 4657–4662. trode in a manner which precludes binding of F to the T2/T3 site. [18] X.-Y. Lang, H.-Y. Fu, C. Hou, G.-F. Han, P. Yang, Y.-B. Liu, Q. Jiang, Nanoporous gold supported cobalt oxide microelectrodes as high-performance electrochemical bio- 6. Conclusions and prospects sensors, Nat. Commun. 4 (2013) 2169. [19] C. Xiao, Y.-L. Liu, J.-Q. Xu, S.-W. Lv, S. Guo, W.-H. Huang, Real-time monitoring of H2O2 release from single cells using nanoporous gold microelectrodes decorated Nanoporous gold has emerged as a promising support for the im- with platinum nanoparticles, Analyst 140 (2015) 3753–3758. mobilisation of enzymes for bioelectrochemical applications (i.e. [20] A.J. Forty, Corrosion micromorphology of noble metal alloys and depletion gilding, Nature 282 (1979) 597–598. biosensors and biofuel cells) with improved response and stability. [21] H.W. Pickering, C. Wagner, Electrolytic dissolution of binary alloys containing a The porous nature of the support enables the immobilisation of noble metal, J. Electrochem. Soc. 114 (1967) 698–706. high amounts of enzyme, though such loadings may not be of benefit [22] J. Erlebacher, M.J. Aziz, A. Karma, N. Dimitrov, K. Sieradzki, Evolution of nanoporosity in dealloying, Nature 410 (2001) 450–453. for very deep pores where diffusion of enzyme substrate into the [23] J. Erlebacher, An atomistic description of dealloying porosity evolution, the critical pores can become rate limiting. The use of flow systems can enhance potential, and rate-limiting behavior, J. Electrochem. Soc. 151 (2004) C614–C626. substrate transport and overcome such limitations. A micro-fluidic [24] X. Li, Q. Chen, I. McCue, J. Snyder, P. Crozier, J. Erlebacher, K. Sieradzki, Dealloying of noble-metal alloy nanoparticles, Nano Lett. 14 (2014) 2569–2577. device utilising NPG is promising due to its potential advantages of [25] Z. Liu, P.C. Searson, Single nanoporous gold nanowire sensors, J. Phys. Chem. B 110 miniaturisation, integration and high-performance. NPG can be (2006) 4318–4322. deposited on plastic supports for flexible devices, enabling the devel- [26] X. Gu, X. Cong, Y. Ding, Platinum-decorated Au porous nanotubes as highly effi- fl cient catalysts for formic acid electro-oxidation, ChemPhysChem 11 (2010) opment of thin exible polymer-based electronic devices, for exam- 841–846. ple, that are integrated with “smart” contact lenses for biomedical [27] M.D. Scanlon, U. Salaj-Kosla, S. Belochapkine, D. MacAodha, D. Leech, Y. Ding, E. applications [103]. Magner, Characterization of nanoporous gold electrodes for bioelectrochemical applications, Langmuir 28 (2011) 2251–2261. [28] F. Jia, C. Yu, Z. Ai, L. Zhang, Fabrication of nanoporous gold film electrodes with ul- Acknowledgements trahigh surface area and electrochemical activity, Chem. Mater. 19 (2007) 3648–3653. [29] H. Dong, X. Cao, Nanoporous gold thin film: fabrication, structure evolution, and The National Natural Science Foundation of China (NSFC No. electrocatalytic activity, J. Phys. Chem. C 113 (2009) 603–609. 51301096), Shandong Provincial Natural Science Foundation [30] L. Lu, Y. Dong, J. Wang, Q. Li, X. Wu, Direct electrochemistry and bioelectrocatalysis (ZR2013EMQ012) and the Fundamental Research Funds of Shandong of horseradish peroxidase entrapped in a self-supporting nanoporous gold electrode: a new strategy to improve the orientation of immobilized enzymes, Anal. University (No. 2014JC017) are appreciated by P. Si. The European Com- Methods 7 (2015) 6686–6694. mission (FP7-PEOPLE-2013-ITN 607793 “Bioenergy”) and Government [31] J.F. Huang, B.T. Lin, Application of a nanoporous gold electrode for the sensitive de- of Ireland Postgraduate Scholarship (GOIPG/2014/659) are greatly ac- tection of copper via mercury-free anodic stripping voltammetry, Analyst 134 – knowledged by X. Xiao. Assoc. Prof. Jingdong Zhang and Prof. Yi Ding (2009) 2306 2313. [32] F. Takeshi, C. Ming Wei, Characteristic length scale of bicontinuous nanoporous are appreciated for their kind support and meaningful discussions. We structure by fast Fourier transform, Jpn. J. Appl. Phys. 47 (2008) 1161. thank Dr. Urszula Salaj-Kosla for reading the manuscript. [33] X. Xiao, M.e. Wang, H. Li, P. Si, One-step fabrication of bio-functionalized nanoporous gold/poly (3, 4-ethylenedioxythiophene) hybrid electrodes for amperometric glucose sensing, Talanta 115 (2013) 1054–1059. References [34] X. Xiao, C. Engelbrekt, Z. Li, P. Si, Hydrogen evolution at nanoporous gold/tungsten sulfide composite film and its optimization, Electrochim. Acta 173 (2015) [1] A. Wittstock, J. Biener, J. Erlebacher, Nanoporous Gold: From an Ancient Technology 393–398. to a High-tech Material, RSC Publishing, 2012. [35] X. Lang, A. Hirata, T. Fujita, M. Chen, Nanoporous metal/oxide hybrid electrodes for [2] L. Qian, X. Yan, T. Fujita, A. Inoue, M. Chen, Surface enhanced Raman scattering of electrochemical supercapacitors, Nat. Nanotechnol. 6 (2011) 232–236. nanoporous gold: smaller pore sizes stronger enhancements, Appl. Phys. Lett. 90 [36] Y.H. Tan, J.R. Schallom, N.V. Ganesh, K. Fujikawa, A.V. Demchenko, K.J. Stine, Char- (2007) 153120. acterization of protein immobilization on nanoporous gold using atomic force mi- [3] Y.Ding,Y.J.Kim,J.Erlebacher,Nanoporousgoldleaf:“ancient technology”/ croscopy and scanning electron microscopy, Nanoscale 3 (2011) 3395–3407. advanced material, Adv. Mater. 16 (2004) 1897–1900. [37] S. Trasatti, O.A. Petrii, Real surface area measurements in electrochemistry, Pure [4] X. Zhang, Y. Ding, Unsupported nanoporous gold for heterogeneous catalysis, Catal. Appl. Chem. 63 (1991) 711–734. Sci. Technol. 3 (2013) 2862–2868. [38] X. Wang, X.Y. Liu, X.L. Yan, P. Zhao, Y. Ding, P. Xu, Enzyme-nanoporous gold [5] A.W. Arne Wittstock, Marcus Bäumer, nanoporous gold as a platform for a building biocomposite: excellent biocatalyst with improved biocatalytic performance and block catalyst, ACS Catal. 2 (2012) 2199–2215. stability, PLoS One 6 (2011) e24207. [6] J. Biener, M.M. Biener, R.J. Madix, C.M. Friend, Nanoporous gold: understanding the [39] X. Yan, X. Wang, P. Zhao, Y. Zhang, P. Xu, Y. Ding, Xylanase immobilized origin of the reactivity of a 21st century catalyst made by pre-Columbian technol- nanoporous gold as a highly active and stable biocatalyst, Microporous Mesopo- ogy, ACS Catal. (2015) 6263–6270. rous Mater. 161 (2012) 1–6. [7] Y. Ding, M. Chen, Nanoporous metals for catalytic and optical applications, MRS [40] R. Szamocki, S. Reculusa, S. Ravaine, P.N. Bartlett, A. Kuhn, R. Hempelmann, Tailored Bull. 34 (2009) 569–576. mesostructuring and biofunctionalization of gold for increased electroactivity, [8] L.Zhang,H.Chang,A.Hirata,H.Wu,Q.-K.Xue,M.Chen,Nanoporousgoldbased Angew. Chem. Int. Ed. 45 (2006) 1317–1321. optical sensor for sub-ppt detection of mercury ions, ACS Nano 7 (2013) [41] R. Szamocki, A. Velichko, C. Holzapfel, F. Mücklich, S. Ravaine, P. Garrigue, N. Sojic, 4595–4600. R. Hempelmann, A. Kuhn, Macroporous ultramicroelectrodes for improved electro- [9] M.M. Collinson, Nanoporous gold electrodes and their applications in analytical analytical measurements, Anal. Chem. 79 (2007) 533–539. chemistry, ISRN Anal. Chem. 2013 (2013) 1–21. [42] P.N. Bartlett, J.J. Baumberg, P.R. Birkin, M.A. Ghanem, M.C. Netti, Highly ordered [10] C. Ding, H. Li, K. Hu, J.-M. Lin, Electrochemical immunoassay of hepatitis B surface macroporous gold and platinum films formed by electrochemical deposition antigen by the amplification of gold nanoparticles based on the nanoporous gold through templates assembled from submicron diameter monodisperse polysty- electrode, Talanta 80 (2010) 1385–1391. rene spheres, Chem. Mater. 14 (2002) 2199–2208. 126 X. Xiao et al. / Bioelectrochemistry 109 (2016) 117–126

[43] M. Heim, S. Reculusa, S. Ravaine, A. Kuhn, Engineering of complex macroporous [75] X. Xiao, E. Magner, A biofuel cell in non-aqueous solution, Chem. Commun. 51 materials through controlled electrodeposition in colloidal superstructures, Adv. (2015) 13478–13480. Funct. Mater. 22 (2012) 538–545. [76] U. Salaj-Kosla, S. Pöller, Y. Beyl, M.D. Scanlon, S. Beloshapkin, S. Shleev, W. Schuhmann, [44] H. du Toit, M. Di Lorenzo, Continuous power generation from glucose with two dif- E. Magner, Direct electron transfer of bilirubin oxidase (Myrothecium verrucaria)at ferent miniature flow-through enzymatic biofuel cells, Biosens. Bioelectron. 69 an unmodified nanoporous gold biocathode, Electrochem. Commun. 16 (2012) (2015) 199–205. 92–95. [45] K. Nishio, H. Masuda, Anodization of gold in oxalate solution to form a nanoporous [77] Z. Gao, G. Binyamin, H.H. Kim, S.C. Barton, Y. Zhang, A. Heller, Electrodeposition of black film, Angew. Chem. Int. Ed. 50 (2011) 1603–1607. redox polymers and co-electrodeposition of enzymes by coordinative crosslinking, [46] A. Walcarius, A. Kuhn, Ordered porous thin films in electrochemical analysis, TrAC, Angew. Chem. Int. Ed. 41 (2002) 810–813. Trends Anal. Chem. 27 (2008) 593–603. [78] F.D. Munteanu, N. Mano, A. Kuhn, L. Gorton, Mediator-modified electrodes for cat- [47] A. Walcarius, Template-directed porous electrodes in electroanalysis, Anal. Bioanal. alytic NADH oxidation: high rate constants at interesting overpotentials, Chem. 396 (2010) 261–272. Bioelectrochemistry 56 (2002) 67–72. [48] R. Szamocki, A. Velichko, F. Mücklich, S. Reculusa, S. Ravaine, S. Neugebauer, W. [79] R. Ludwig, R. Ortiz, C. Schulz, W. Harreither, C. Sygmund, L. Gorton, Cellobiose de- Schuhmann, R. Hempelmann, A. Kuhn, Improved enzyme immobilization for en- hydrogenase modified electrodes: advances by materials science and biochemical hanced bioelectrocatalytic activity of porous electrodes, Electrochem. Commun. 9 engineering, Anal. Bioanal. Chem. 405 (2013) 3637–3658. (2007) 2121–2127. [80] U. Salaj-Kośla, Nanoporous Gold Electrodes as Matrices for Enzyme Immobilization [49] H. Qiu, C. Xu, X. Huang, Y. Ding, Y. Qu, P. Gao, Adsorption of laccase on the surface for Application in Biosensors and Biofuel Cells(PhD thesis) University of Limerick, of nanoporous gold and the direct electron transfer between them, J. Phys. Chem. C Limerick, Ireland, 2012. 112 (2008) 14781–14785. [81] A. Barfidokht, J.J. Gooding, Approaches toward allowing electroanalytical devices to [50] M. Hakamada, M. Takahashi, M. Mabuchi, Enhanced thermal stability of laccase be used in biological fluids, Electroanalysis 26 (2014) 1182–1196. immobilized on monolayer-modified nanoporous Au, Mater. Lett. 66 (2012) 4–6. [82] E. Magner, Detection of ferricyanide as a probe for the effect of hematocrit in whole [51] T. Fujita, P.F. Guan, K. McKenna, X.Y. Lang, A. Hirata, L. Zhang, T. Tokunaga, S. Arai, blood biosensors, Analyst 126 (2001) 861–865. Y. Yamamoto, N. Tanaka, Y. Ishikawa, N. Asao, Y. Yamamoto, J. Erlebacher, M.W. [83] S. Ge, X. Jiao, D. Chen, Ultrasensitive electrochemical immunosensor for CA 15-3 Chen, Atomic origins of the high catalytic activity of nanoporous gold, Nat. using thionine-nanoporous gold-graphene as a platform and horseradish Mater. 11 (2012) 775–780. peroxidase-encapsulated liposomes as signal amplification, Analyst 137 (2012) [52] Z. Liu, L. Huang, L. Zhang, H. Ma, Y. Ding, Electrocatalytic oxidation of D-glucose at 4440–4447. nanoporous Au and Au–Ag alloy electrodes in alkaline aqueous solutions, [84] C. Wu, X. Liu, Y. Li, X. Du, X. Wang, P. Xu, Lipase-nanoporous gold biocomposite Electrochim. Acta 54 (2009) 7286–7293. modified electrode for reliable detection of triglycerides, Biosens. Bioelectron. 53 [53] Z. Liu, J. Du, C. Qiu, L. Huang, H. Ma, D. Shen, Y. Ding, Electrochemical sensor for de- (2014) 26–30. tection of p-nitrophenol based on nanoporous gold, Electrochem. Commun. 11 [85] C. Chen, Q. Xie, D. Yang, H. Xiao, Y. Fu, Y. Tan, S. Yao, Recent advances in electro- (2009) 1365–1368. chemical glucose biosensors: a review, RSC Adv. 3 (2013) 4473–4491. [54] X. Yan, F. Meng, S. Cui, J. Liu, J. Gu, Z. Zou, Effective and rapid electrochemical de- [86] M.W. Hsiao, R.R. Adžić, E.B. Yeager, Electrochemical oxidation of glucose on single tection of hydrazine by nanoporous gold, J. Electroanal. Chem. 661 (2011) 44–48. crystal and polycrystalline gold surfaces in phosphate buffer, J. Electrochem. Soc. [55] X. Ge, L. Wang, Z. Liu, Y. Ding, Nanoporous gold leaf for amperometric determina- 143 (1996) 759–767. tion of nitrite, Electroanalysis 23 (2011) 381–386. [87] S.B. Aoun, Z. Dursun, T. Koga, G.S. Bang, T. Sotomura, I. Taniguchi, Effect of metal [56] M. Hakamada, M. Takahashi, T. Furukawa, K. Tajima, K. Yoshimura, Y. Chino, M. ad-layers on Au(1 1 1) electrodes on electrocatalytic oxidation of glucose in an al- Mabuchi, Electrochemical stability of self-assembled monolayers on nanoporous kaline solution, J. Electroanal. Chem. 567 (2004) 175–183. Au, Phys. Chem. Chem. Phys. 13 (2011) 12277–12284. [88] Q. Li, S. Cui, X. Yan, Electrocatalytic oxidation of glucose on nanoporous gold mem- [57] M. Hakamada, M. Takahashi, M. Mabuchi, Enzyme electrodes stabilized by branes, J. Solid State Electrochem. 16 (2012) 1099–1104. monolayer-modified nanoporous Au for biofuel cells, Gold Bull. 45 (2012) 9–15. [89] L.Y. Chen, X.Y. Lang, T. Fujita, M.W. Chen, Nanoporous gold for enzyme-free elec- [58] J. Patel, L. Radhakrishnan, B. Zhao, B. Uppalapati, R.C. Daniels, K.R. Ward, M.M. trochemical glucose sensors, Scr. Mater. 65 (2011) 17–20. Collinson, Electrochemical properties of nanostructured porous gold electrodes [90] X. Xiao, M.e. Wang, H. Li, Y. Pan, P. Si, Non-enzymatic glucose sensors based on in biofouling solutions, Anal. Chem. 85 (2013) 11610–11618. controllable nanoporous gold/copper oxide nanohybrids, Talanta 125 (2014) [59] S. Saraf, C.J. Neal, S. Park, S. Das, S. Barkam, H.J. Cho, S. Seal, Electrochemical study 366–371. of nanoporous gold revealing anti-biofouling properties, RSC Adv. 5 (2015) [91] S. Kerzenmacher, J. Ducrée, R. Zengerle, F. Von Stetten, Energy harvesting by im- 46501–46508. plantable abiotically catalyzed glucose fuel cells, J. Power Sources 182 (2008) 1–17. [60] A.A. Farghaly, M. Lam, C.J. Freeman, B. Uppalapati, M.M. Collinson, Potentiometric [92] C. Corti, R. Holliday, D. Thompson, Progress towards the commercial application of measurements in biofouling solutions: comparison of nanoporous gold to planar gold catalysts, Top. Catal. 44 (2007) 331–343. gold, J. Electrochem. Soc. 163 (2016) H3083–H3087. [93] M.C. Leopold, E.F. Bowden, Influence of gold substrate topography on the volt- [61] L.C. Clark, C. Lyons, Electrode systems for continuous monitoring in cardiovascular ammetry of cytochrome c adsorbed on carboxylic acid aterminated self-assembled surgery, Ann. N. Y. Acad. Sci. 102 (1962) 29–45. monolayers, Langmuir 18 (2002) 2239–2245. [62] R. Wilson, A.P.F. Turner, Glucose oxidase: an ideal enzyme, Biosens. Bioelectron. 7 [94] M. Falk, Z. Blum, S. Shleev, Direct electron transfer based enzymatic fuel cells, (1992) 165–185. Electrochim. Acta 82 (2012) 191–202. [63] J. Wang, Electrochemical glucose biosensors, Chem. Rev. 108 (2008) 814–825. [95] R.A.S. Luz, A.R. Pereira, J.C.P. de Souza, F.C.P.F. Sales, F.N. Crespilho, Enzyme biofuel [64] S.B. Bankar, M.V. Bule, R.S. Singhal, L. Ananthanarayan, Glucose oxidase — an over- cells: thermodynamics, kinetics and challenges in applicability, ChemElectroChem view, Biotechnol. Adv. 27 (2009) 489–501. 1 (2014) 1751–1777. [65] H.J. Hecht, D. Schomburg, H. Kalisz, R.D. Schmid, The 3D structure of glucose oxi- [96] M. Cooney, V. Svoboda, C. Lau, G. Martin, S. Minteer, Enzyme catalysed biofuel cells, dase from Aspergillus niger. Implications for the use of GOD as a biosensor enzyme, Energy Environ. Sci. 1 (2008) 320–337. Biosens. Bioelectron. 8 (1993) 197–203. [97] T. Sakurai, K. Kataoka, Basic and applied features of multicopper oxidases, CueO, [66] H. Qiu, L. Xue, G. Ji, G. Zhou, X. Huang, Y. Qu, P. Gao, Enzyme-modified nanoporous bilirubin oxidase, and laccase, Chem. Rec. 7 (2007) 220–229. gold-based electrochemical biosensors, Biosens. Bioelectron. 24 (2009) [98] S. Boland, D. Leech, A glucose/oxygen enzymatic fuel cell based on redox polymer 3014–3018. and enzyme immobilisation at highly-ordered macroporous gold electrodes, Ana- [67] R. Zeis, T. Lei, K. Sieradzki, J. Snyder, J. Erlebacher, Catalytic reduction of oxygen and lyst 137 (2012) 113–117. hydrogen peroxide by nanoporous gold, J. Catal. 253 (2008) 132–138. [99] U. Salaj-Kosla, S. Pöller, W. Schuhmann, S. Shleev, E. Magner, Direct electron trans- [68] T. Li, F. Jia, Y. Fan, Z. Ding, J. Yang, Fabrication of nanoporous thin-film working fer of Trametes hirsuta laccase adsorbed at unmodified nanoporous gold electrodes, electrodes and their biosensingapplications, Biosens. Bioelectron. 42 (2013) 5–11. Bioelectrochemistry 91 (2013) 15–20. [69] X. Xiao, J. Ulstrup, H. Li, J. Zhang, P. Si, Nanoporous gold assembly of glucose oxi- [100] L. Deng, F. Wang, H. Chen, L. Shang, L. Wang, T. Wang, S. Dong, A biofuel cell with dase for electrochemical biosensing, Electrochim. Acta 130 (2014) 559–567. enhanced performance by multilayer biocatalyst immobilized on highly ordered [70] X. Xiao, H. Li, K. Zhang, P. Si, Examining the effects of self-assembled monolayers macroporous electrode, Biosens. Bioelectron. 24 (2008) 329–333. on nanoporous gold based amperometric glucose biosensors, Analyst 139 (2014) [101] M. Tominaga, M. Ohtani, I. Taniguchi, Gold single-crystal electrode surface modi- 488–494. fied with self-assembled monolayers for electron tunneling with bilirubin oxidase, [71] A. Heller, Electrical connection of enzyme redox centers to electrodes, J. Phys. Phys. Chem. Chem. Phys. 10 (2008) 6928–6934. Chem. 96 (1992) 3579–3587. [102] K. Murata, K. Kajiya, N. Nakamura, H. Ohno, Direct electrochemistry of bilirubin ox- [72] A. Heller, B. Feldman, Electrochemical glucose sensors and their applications in di- idase on three-dimensional gold nanoparticle electrodes and its application in a abetes management, Chem. Rev. 108 (2008) 2482–2505. biofuel cell, Energy Environ. Sci. 2 (2009) 1280–1285. [73] A. Heller, Electron-conducting redox hydrogels: design, characteristics and synthe- [103] M. Falk, V. Andoralov, Z. Blum, J. Sotres, D.B. Suyatin, T. Ruzgas, T. Arnebrant, S. sis, Curr. Opin. Chem. Biol. 10 (2006) 664–672. Shleev, Biofuel cell as a power source for electronic contact lenses, Biosens. [74] U. Salaj-Kosla, M.D. Scanlon, T. Baumeister, K. Zahma, R. Ludwig, P.Ó. Conghaile, D. Bioelectron. 37 (2012) 38–45. MacAodha,D.Leech,E.Magner,Mediatedelectron transfer of cellobiose dehydrogenase and glucose oxidase at osmium polymer-modified nanoporous gold electrodes, Anal. Bioanal. Chem. 405 (2013) 3823–3830.

Due to copyright restrictions the full text of this article is not included with the electronic version of this thesis

Chemical Communication 2015, 51, 13478 A biofule cell in non-aqueous soltuion Xinxin Xiao and Edmond Magner http://dx.doi.org/10.1039/c5cc04888e http://hdl.handle.net/10344/5064

Biosensors and Bioelectronics 90 (2017) 96–102

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journal homepage: www.elsevier.com/locate/bios

A symmetric supercapacitor/biofuel cell hybrid device based on enzyme- MARK modiﬁed nanoporous gold: An autonomous pulse generator ⁎ Xinxin Xiaoa, Peter Ó Conghaileb, Dónal Leechb, Roland Ludwigc, Edmond Magnera, a Department of Chemical and Environmental Sciences, Bernal Institute, University of Limerick, Limerick, Ireland b School of Chemistry & Ryan Institute, National University of Ireland Galway, Galway, Ireland c Department of Food Science and Technology, BOKU-University of Natural Resources and Life Sciences, Muthgasse 18, 1190 Vienna, Austria

ARTICLE INFO ABSTRACT

Keywords: The integration of supercapacitors with enzymatic biofuel cells (BFCs) can be used to prepare hybrid devices in Biofuel cell order to harvest signiﬁcantly higher power output. In this study, a supercapacitor/biofuel cell hybrid device was Supercapacitor prepared by the immobilisation of redox enzymes with electrodeposited poly(3,4-ethylenedioxythiophene) Hybrid device +/2+ (PEDOT) and the redox polymer [Os(2,2′-bipyridine)2(polyvinylimidazole)10Cl] (Os(bpy)2PVI) on dealloyed Nanoporous gold nanoporous gold. The thickness of the deposition layer can be easily controlled by tuning the deposition Osmium redox polymer conditions. Once charged by the internal BFC, the device can be discharged as a supercapacitor at a current Pulse generator density of 2 mA cm−2 providing a maximum power density of 608.8 μWcm−2, an increase of a factor of 468 when compared to the power output from the BFC itself. The hybrid device exhibited good operational stability for 50 charge/discharge cycles and ca. 7 h at a discharge current density of 0.2 mA cm−2. The device could be used as a pulse generator, mimicking a cardiac pacemaker delivering pulses of 10 μA for 0.5 ms at a frequency of 0.2 Hz.

1. Introduction presence of conductive fluids within the body, implantable cell stacks suffer from the problem of short-circuits between individual cells Enzymatic biofuel cells (BFCs) utilizing oxidoreductases as electro- (Andoralov et al., 2013; MacVittie et al., 2013). In such systems, catalysts can be used to generate electricity from fuels such as sugars or isolation of the cells is essential. Another route is to couple BFCs with alcohols in combination with dioxygen (Calabrese Barton et al., 2004; external electronic devices to increase the voltage. For example, using a Cooney et al., 2008; Leech et al., 2012). BFCs are of interest as power charge pump and a DC-DC converter, a fluidic BFC utilizing PQQ- sources for biosensors (Pinyou et al., 2015; Zloczewska et al., 2014), dependent glucose dehydrogenase and laccase with an intrinsic OCV of medical implants (e.g. insulin pumps, cardiac pacemakers (MacVittie 0.47 V was sufficient to power a pacemaker (Southcott et al., 2013). et al., 2013)), and other devices (Falk et al., 2012; Ó Conghaile et al., Falk et al. (2014) presented a self-powered wireless lactose biosensing 2016). To be able to activate commonly used microelectronic devices system, consisting of an energy harvesting module including a voltage (such as commercial pacemakers), appropriate output voltages (mini- amplifier and capacitor to build a power source based on a BFC using mum of 1.4 V) are required (MacVittie et al., 2013). The open circuit bilirubin oxidase (BOx) and cellobiose dehydrogenase (CDH). voltage (OCV) of glucose and oxygen BFCs is limited by the thermo- In addition to low voltage outputs, BFCs are also limited by their dynamic value of 1.179 V (Pankratov et al., 2016), and in practice by low current/power densities, which can be improved through efficient the difference between the onset redox potentials of the bioanode and substrate diffusion (Murata et al., 2009), enhanced rates of electron biocathode (Cracknell et al., 2008). The observed OCV can be increased transfer between enzymes and electrodes, improving catalytic activity change "The observed OCV can be increased by..." to "The observed (Suraniti et al., 2013) and loading of enzymes (Flexer et al., 2011), as OCV can be increased closer to the theoretical value by..." by using well as utilizing enzyme cascades for deep and complete oxidation direct electron transfer (DET) or by the use of redox mediators with pathways (Kim et al., 2013; Shao et al., 2013; Xu and Minteer, 2012). redox potentials closer to those of the enzyme/cofactor (Rasmussen The introduction of capacitors into the BFC circuit enables the et al., 2015). The OCV can also be increased by using multiple cells accumulation of charge, resulting in output pulses of higher power. connected in series (MacVittie et al., 2013). However, due to the Sode et al. proposed the concept of a “BioCapacitor” with the integra-

⁎ Corresponding author. E-mail address: [email protected] (E. Magner). http://dx.doi.org/10.1016/j.bios.2016.11.012 Received 17 July 2016; Received in revised form 21 October 2016; Accepted 5 November 2016 Available online 09 November 2016 0956-5663/ © 2016 Elsevier B.V. All rights reserved. X. Xiao et al. Biosensors and Bioelectronics 90 (2017) 96–102 tion of a charge pump/capacitor and a BFC that resulted in higher were prepared with deionised water (resistivity of 18.2 MΩ cm) from voltages and currents (Hanashi et al., 2009; Sode et al., 2016). an Elgastat maxima-HPLC (Elga, UK). Os(bpy)2PVI was synthesised Electrochemical capacitors (known as supercapacitors) (Winter and using a published procedure (Jenkins et al., 2009). BOx from Brodd, 2004) take advantage of the electrical double layer capacitance Myrothecium verrucaria (EC 1.3.3.5, 2.63 U mg−1) was purchased attained via ion adsorption or pseudocapacitance achieved by fast and from Amano Enzyme Inc., Japan. Recombinant, (in Pichia pastoris) reversible faradaic reactions, offering high specific power density and expressed Glomerella cingulata GDH (EC 1.1.99.10) with a specific great durability. Supercapacitors externally connected to a laccase- activity of 572 U mg−1 was prepared according to a published route based cathode and zinc anode based biobattery, had higher power (Sygmund et al., 2011). stability than the battery itself (Skunik-Nuckowska et al., 2014). Recent NPG leaves were fabricated by dealloying ca. 100 nm thick Au/Ag progress has seen BFC assemblies with capacitive bioelectrodes (Agnes leaf alloy (12-carat, Eytzinger, Germany) in concentrated HNO3 et al., 2014; González-Arribas et al., 2016; Kizling et al., 2015a, 2015b; (Sigma-Aldrich) for 30 min at 30 °C. The NPG films were then attached Pankratov et al., 2014b). These supercapacitor/BFC hybrids, or self- onto pre-polished glassy carbon electrodes (GCEs) with a diameter of charging biocapacitors, are based on the fabrication of hybrid compo- 4 mm. Prior to use, cyclic voltammetry (CV) of NPG in 1 M H2SO4 were site modified electrodes with integration of enzymes and capacitive carried out to create clean surfaces and left to dry naturally. materials. The main feature is their ability to generate cyclic, high power pulses from the discharge of the supercapacitor, which is 2.2. Enzyme immobilisation procedures recharged towards to OCV via the internal BFC in the following open-circuit mode (Agnes et al., 2014). A supercapacitor/BFC hybrid The electrodeposition solutions contained 0.1 M pH 7.0 phosphate based on wired enzymes on carbon nanotubes (CNTs) was capable of buffer solution (PBS) with 2 mM polyethylene glycol 3400 (PEG3400), −1 −1 delivering discharge pulses for 5 days in the presence of glucose and O2 20 mM EDOT, 0.5 mg ml Os(bpy)2PVI and either 0.5 mg ml of (Agnes et al., 2014). The majority of supercapacitor/BFC systems have FAD-GDH or BOx. The presence of PEG enabled the dispersion of relied on the use of high-surface-area carbon nanomaterials (CNMs), EDOT in aqueous media and increased the hydrophilicity of the such as CNT (Agnes et al., 2014; Kizling et al., 2015a, 2015b; polymer (Reiter et al., 2001; Fabiano et al., 2002). A pulse sequence Pankratov et al., 2014b) and graphene (González-Arribas et al., of 0.9 V (2 s) and −0.4 V (3 s) was used for deposition. The electrodes 2016). However, the potential toxicity of CNMs (Magrez et al., 2006) were then gently rinsed with PBS. For comparison purposes, films were should be taken into account for in vivo applications and direct also deposited onto polycrystalline planar Au electrodes. exposure to CNMs should be avoided in implantable devices (Miyake et al., 2011). 2.3. Morphology characterisation Dealloyed nanoporous gold (NPG), a porous material in a self- supporting bulk form comprising three-dimensional frameworks of Scanning electron microscopy (SEM) images were recorded using a bicontinuous pores and ligaments (Ding et al., 2004), has been Hitachi SU-70 microscope operated at 15 kV. Transmission electron investigated as conductive and non-toxic supports for supercapacitors microscopy (TEM) images were obtained using a JEOL JEM-2100 (Lang et al., 2011; Meng and Ding, 2011) and enzyme immobilisation instrument at an acceleration voltage of 200 kV. The average pore size (Scanlon et al., 2012; Xiao and Magner, 2015; Xiao et al., 2016, 2014), and deposition layer thickness were obtained by performing at least 30 separately. Mediators are required to enable efficient electron transfer different measurements with ImageJ software (National Institutes of between the cofactor of the enzyme and the NPG surface (Xiao et al., Health, Bethesda, Maryland) (Schneider et al., 2012). 2013). In this context, alternate potential pulses could be applied to electrodeposit osmium redox polymers with the co-immobilisation of 2.4. Electrochemical measurements enzymes onto electrode surface (Gao et al., 2002; Habermüller et al., 2000; Schuhmann et al., 1997). Unlike other soluble mediators that are Generally, electrochemical studies were performed using a CHI802 prone to leakage (González-Arribas et al., 2016), the electrodeposited potentiostat (CH Instruments, Austin, Texas) in a standard three- redox polymer is robust, even in hydrodynamic conditions. For electrode electrochemical cell containing 0.1 M pH 7.0 PBS and 0.1 M example, a laccase/redox polymer composite film showed little loss in KCl. Enzyme-modified electrodes, a platinum wire and saturated response after rotation for 24 h at 2500 rpm (Shen et al., 2013). In this calomel electrode (SCE) were used as the working, counter and contribution, we electrodeposited poly(3,4-ethylenedioxythiophene) reference electrodes, respectively. The polarisation and power curves (PEDOT) and the redox polymer [Os(2,2′- of the assembled biofuel cells were measured using the bioanode as +/2+ bipyridine)2(polyvinylimidazole)10Cl] (Os(bpy)2PVI) onto NPG working electrode and the biocathode as a combined counter/reference − electrodes with the co-immobilisation of enzymes. Flavin adenine electrode. The potential was scanned at a scan rate of 1 mV s 1 in the dinucleotide-dependent glucose dehydrogenase (FAD-GDH, EC presence of O2-bubbled 20 mM glucose, while recording the current in 1.1.99.10, D-glucose: acceptor 1-oxidoreductase) was used as an the circuit. All experiments were carried out at room temperature (20 oxygen-insensitive enzyme at the anode (Zafar et al., 2012), in contrast ± 2 °C). The current densities or power densities were calculated using to glucose oxidase (GOx) which depletes dissolved oxygen and pro- the geometric surface area of the working electrode or bioanode unless duces unwanted hydrogen peroxide (Milton et al., 2015). BOx was stated otherwise. immobilised at the cathode and the properties of the cell were The specific capacitance of the individual electrode in three- characterised in detail. A proof-of-concept pulse generator for a pace- electrode system was calculated from the cyclic voltammograms in maker was demonstrated, which was able to deliver a 10 μA pulse at a the region where no faradaic processes were occurring (Eq. S1). The frequency of 0.2 Hz. specific capacitance of the assembled supercapacitor was obtained from the galvanostatic discharge curves (Eq. S2). 2. Experimental section The charge/discharge performance of the hybrid devices in air- equilibrated buffer solution containing 20 mM glucose was examined 2.1. Materials with an Autolab PGSTAT100 potentiostat (Eco Chimie, Netherlands) using the biocathode as working electrode and the bioanode as a Potassium phosphate monobasic (≥99%) and dibasic (≥98%), D- combined counter/reference electrode. Testing of the devices involved (+)-glucose (99.5%), 3,4-ethylenedioxythiophene (EDOT, 97%) were the test sequence: (i) charging at open-circuit mode using the BFC obtained from Sigma-Aldrich Ireland, Ltd. Potassium chloride (KCl, component and (ii) galvanostatic discharge of the capacitor at various ≥99%) was purchased from Fisher Scientific Ireland, Ltd. All solutions current densities (Fig. 4C).

97 X. Xiao et al. Biosensors and Bioelectronics 90 (2017) 96–102

polymer, the weakly coordinated chloride ions exchanged with more strongly coordinating pyridine or imidazole groups on proximal chains when Os3+ was reduced to Os2+ during the cathodic pulse (Gao et al., 2002). This crosslinking eﬀect led to irreversible polymer precipitation onto the electrode. The resting period at the anodic potential enabled the reestablishment of the bulk concentration of precursor at the electrode surface. Overall, the potential sequence led to the alternate

deposition of PEDOT and Os(bpy)2PVI, which was confirmed by electrochemical studies (Fig. 1A). Enzymes in the deposition solution were physically and/or coordinately entrapped into the resulting films. CVs of various modified electrodes (deposition time of 300 s) in PBS at 100 mV s−1 were compared to confirm the successful electrodeposition of the polymers (Fig. 1A). CVs of NPG/PEDOT/

Os(bpy)2PVI/FAD-GDH electrodes showed the faradaic redox reaction 2+/3+ of Os (ΔEp of 76 mV) superimposed on the charge/discharge capacitive currents. The Os polymer modified NPG electrode without PEDOT displayed a pair of reversible redox peaks with a peak separation of 20 mV. NPG/PEDOT exhibited a rectangular charge/ discharge curve without any redox peaks. Table S1 compares CV derived specific capacitances that are normalised with respect to the projected surface area. Bare NPG showed a 9.6-fold higher capacitance than that possible with the bare planar gold electrode, consistent with the surface roughness factor (the ratio between the electrochemically addressable and geometric surface areas) obtained from the outermost − layer of Au oxide stripping (a specific charge of 390 μCcm 2 is required for gold oxide reduction (Trasatti and Petrii, 1991)). NPG/PEDOT and

NPG/PEDOT/Os(bpy)2PVI had 3.2 and 4.4 times higher capacitance than that of bare NPG. The amount of deposited hybrid polymer, with the associated increase in the capacitance, and the enzyme loading increased with potential cycling before levelling off after a number of cycles (200 cycles for the case of (Gao et al., 2002)). On increasing the deposition time, the resulting film tended to block the pores (Fig. S3) of the NPG electrode (Fig. S1). For the FAD-GDH modified electrode, a deposition time of 300 s exhibited the optimal response to 10 mM glucose (Fig. S4A), attributed to a compromise between loading of biocatalyst and mass transport of substrate through the film. For the BOx modified electrode, a shorter pulse duration of 150 s afforded the highest electrocatalytic response to oxygen, with relatively low capacitance (Fig. S4B). A pulse of 300 s duration was chosen as a compromise between the electrochemical response and capacitance. Both bioelectrodes were separately studied in detail at a scan rate of 5mVs−1 (Fig. 1B and C). As can be seen from Fig. 1B, NPG/PEDOT/

Os(bpy)2PVI/FAD-GDH displayed a pair of redox peaks with a midpoint potential, Em, of +191 mV (vs. SCE), in agreement with the reduction-oxidation of the Os2+/3+ couple. The ratio of the integrated area of the anodic to cathodic peak was ca. 1.1. The variation of peak current with scan rate was linear, indicative of a surface controlled process (Fig. S5). In the presence of 10 mM glucose, a sigmoidal catalytic wave with an onset potential of −18 ± 9 mV vs. SCE and a −2 Fig. 1. (A) Cyclic voltammograms (CVs) of various electrodes (deposition time: 300 s). background-corrected limiting current density of 59.7 ± 2.4 μAcm

CVs of (B) NPG/PEDOT/Os(bpy)2PVI/FAD-GDH and (C) NPG/PEDOT/Os(bpy)2PVI/ was observed. These results were indicative of the successful immobi- −1 BOx electrodes at a scan rate of 5 mV s . lisation of FAD-GDH with high activity. The apparent Michaelis- app Menten constant, KM , of the enzyme modiﬁed electrode was 3. Results and discussion 7.9 mM (Fig. S6), lower than the value of 17.4 mM obtained from the same enzyme when chemically crosslinked onto graphite electrode 3.1. Electrochemical characterisation of the capacitive bioelectrodes (Zafar et al., 2012). This decrease may arise from improved substrate transport through the thin immobilizing layer (Fig. 2). BOx based Cyclic voltammograms (CVs) displayed an onset potential of 0.7 V cathodes also showed a pair of redox peaks in N2 bubbled PBS (vs. SCE) for the growth of PEDOT on the NPG electrode in an aqueous (Fig. 1C). The ratio of the integrated area of the anodic to cathodic solution (Xiao et al., 2013). The potentiostatic pulse comprised an peak was however less than 1, due to competition with residual O2 for anodic potential of 0.9 V (2 s) to generate the radical cation and a the oxidation of BOx. The Em was +202 mV, a slight increase in cathodic potential of −0.4 V (3 s) to enable the EDOT concentration in comparison to that of FAD-GDH modiﬁed electrodes. In O2 bubbled the proximity of the electrode surface to return to that in the bulk state, solution, electrocatalytic reduction commenced at 387 ± 13 mV and thus allowing polymer formation on the electrode surface (Schuhmann reached a maximum net catalytic current density of 65.2 ± 4.5 μAcm−2. et al., 1997). In the presence of Os poly(N-vinylimidazole) redox

98 X. Xiao et al. Biosensors and Bioelectronics 90 (2017) 96–102

galvanostatic charge/discharge at a given external current density of 10 µA cm−2 (Fig. S7). A specific capacitance of 391.9 ± 2.1 µF cm−2 was obtained (Eq. S2). The total capacitance of the supercapacitor is determined by the series connection of the two capacitive electrodes (Eq. S3)(Khomenko et al., 2005), leading to a lower overall capacitance compared with those of individual electrodes. Recent reports described the underlying mechanism of a hybrid supercapacitor/microbial fuel cell (Pankratov et al., 2014a; Santoro et al., 2016). The integration of a BFC with a capacitor enables the hybrid device to work as a self-powered capacitor, without the requirement for external input. In rest conditions, i.e. in open-circuit, the cell voltage tended to the equilibrium potential, i.e. OCV of the BFC. The existing potential difference between the two electrodes polarised the anode and cathode, leading the NPG backbones to be negatively or positively charged, respectively, and triggering the p-dopable PEDOT film to insert/deinsert anions (Fig. 4A). In other words, the capacitive cell was electrostatically charged at the thermodynamically induced potential difference, driving its voltage profile close to the value of OCV. Fig. 2. TEM image of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH (300 s deposition). As shown in Fig. 4C, the voltage increased with time initially rising rapidly before levelling off with time. 3.2. Morphology characterisation The energy stored in the biocapacitor could be subsequently discharged at desired currents by releasing ions (Fig. 4B). As can be NPG and NPG/PEDOT/Os(bpy)2PVI/FAD-GDH electrodes were seen in Fig. 4C and D, a galvanostatic discharging current density of examined by SEM (Fig. S2A and B). Typical porous structures with 0.2 mA cm−2, almost 7 times higher than the 28.9 µA cm−2 possible bicontinuous pores/ligaments of NPG were observed. The average pore with the BFC mode, resulted in a rapid release of power. In the size was 30.6 ± 4.7 nm for the bare NPG (Figs. S1 and S2). The following cycle, the rest step at open-circuit mode without any external deposited layer uniformly grew along the pore surfaces, making the load enabled the recovery of the cell potential to OCV (0.45 V) of the pores smaller and ligaments thicker, but not plugging the pores. The BFC. The following cycles almost overlapped, indicative of the excellent core-shell structure was clearly observed by TEM (Fig. 2), with the stability (Fig. 4C). On a closer examination of the first discharging fi fi contrast between the modi ed lm and the gold support clearly visible. segment (Fig. 4D), a capacitance of 357 µF cm−2 was calculated by The spatially homogeneous film was 7.4 ± 1.4 nm in thickness, a size dividing the given current density, jpulse, by the absolute value of the sufficient to encapsulate the enzyme. slope of the discharging curve (Eq. S2). A jpulse dependent voltage drop of 11 mV was observed due to the internal resistance (Eq. S4), probably 3.3. Hybrid device testing assigned to the ohmic resistance of electrode, mass/charge transfer resistance, and/or low intrinsic biocatalytic activity (Liang et al., 2007). NPG/PEDOT/Os(bpy)2PVI/FAD-GDH and NPG/PEDOT/ The resistance was predominantly attributed to the internal resistance Os(bpy)2PVI/BOx electrodes were subsequently assembled into a from the capacitor, instead of the biocatalytic processes, as the cell also dual-functioning device comprising a BFC and a capacitor. This type showed a voltage drop when used only as a capacitor (Fig. S7). of device can perform as a glucose/O2 BFC when connected to a load in The long-term operation (7 h, 50 cycles) of the hybrid device was an external circuit (Fig. 3A). The polarisation curve of the BFC was tested by recording the potential at the open-circuit with a cutoff at −1 obtained with linear sweep voltammetry at a scan rate of 1 mV s , with 0.4 V, followed by discharge at 0.2 mA cm−2 for 0.5 s (test sequence is the power curve calculated accordingly (Fig. 3B). The BFC registered an shown as the red line of Fig. 4C and D). For a period of 7 h (50 cycles) −2 OCV of 459.6 ± 9.5 mV, a maximum current density of 28.9 µA cm , (Fig. 5A), the discharge finishing potentials remained constant at −2 and a maximum power density of 1.3 µW cm at a potential of 0.09 V 0.07 V, demonstrating the stable capacitance of the supercapacitor in O2 bubbled PBS containing 20 mM glucose. The assembled cell can for each discharge cycle. The reset time did increase, e.g. ca. 300 and also act as a supercapacitor, whose performance was examined by

Fig. 3. (A) Schematic diagram of the BFC. (B) Polarisation and power curve for the BFC consisting of NPG/PEDOT/Os(bpy)2PVI/FAD-GDH bioanode and NPG/PEDOT/Os(bpy)2PVI/ BOx biocathode.

99 X. Xiao et al. Biosensors and Bioelectronics 90 (2017) 96–102

Fig. 4. Schematic diagrams of the hybrid device working at the self-charging (A) and galvanostatic discharging mode (B) (with simplified charge-discharge description on the capacitive NPG/PEDOT hybrid). (C) Charge/discharge curves of the as-assembled biocapacitor (black line); Experimental setup: reset at open-circuit and cutoff at 0.4 V, followed by discharging at 0.2 mA cm−2 for 0.5 s (red line). (D) Magnified image of the first discharge segment. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

800 s for the ﬁrst and ﬁnal cycles, respectively. After ca. 7 h of operation, the device exhibited a loss of 70% in maximum power density (0.38 µW cm−2)(Fig. S8A) when tested as a BFC. Cyclic

voltammograms of PEDOT/Os(bpy)2PVI showed little change, indicative of a stable modiﬁcation layer (Fig. S8B). Thus, decreased enzymatic activity, in particular of the FAD-GDH based bioanode (data not shown), was responsible for the extended self-charge time. The discharge capability of the biocapacitor at various current densities was examined (Fig. 5B), with a current density up to 2mAcm−2. Generally, a larger discharge current density provided a larger power density (Eq. S8), as well as a longer recovery time. Table S2 compares the instant maximum power densities that can be delivered. For example, power pulses of 352 and 609 μWcm−2 at 1 and 2 mA cm−2 were achieved, 271 and 468 times higher than that obtained from a traditional BFC conﬁguration (1.3 μWcm−2). The

maximum voltage output, Vmax, decreased with higher pulse current densities due to the potential loss caused by the equivalent series resistance (ESR) (Santoro et al., 2016) (Eq. s7, Table S2). As a result, doubling the current density did not result in the same increase in the Fig. 5. (A) Charge/discharge curves of the biocapacitor for 50 cycles; Experimental maximum power density. Decreasing the ESR could improve the setup: reset at open-circuit and cutoﬀ at 0.4 V, followed by discharging at 0.2 mA cm−2 maximum power density (Santoro et al., 2016). for 0.5 s (B) Charge/discharge curves of the biocapacitor upon various discharging To highlight the important role of the NPG substrate, a planar Au ﬀ current densities; Experimental setup: reset at open-circuit and cuto at 0.4 V, followed based hybrid electrode system was constructed using the same condi- by discharging at 0.005 (a), 0.01 (b), 0.02 (c), 0.05 (d), 0.1 (e), 0.2 (f), 0.5 (g), 1 (h), 2 (i) − tions. Au based BFC displayed a poor performance with an OCV of mA cm 2 for 0.2 s. 365 mV, a maximum current density of 1.5 µA cm−2, and a maximum −2 power density of 0.08 µW cm at a potential of 0.11 V in O2 bubbled PBS containing 20 mM glucose (Fig. S9A). The internal resistance was

100 X. Xiao et al. Biosensors and Bioelectronics 90 (2017) 96–102

reasonable stability without visible leakage of the redox mediators after 50 cycles operation at 0.2 mA cm−2 for approximately 7 h. In contrast to the planar Au based system, nanoporous gold electrodes improved the performance in terms of lower resistance, higher bioelectrochemical signal and capacitance. A proof-of-concept pulse generator (0.2 Hz pulse at 10 μA for 0.5 ms) to mimic a pacemaker was demonstrated using electrodes connected in series.

Acknowledgments

This project has received funding from the European Union's Seventh Framework Programme for research, technological development and demonstration under Grant agreement No 607793. X. Xiao acknowledges a Government of Ireland Postgraduate Scholarship (GOIPG/2014/659). The assistance of Dr. T. Kennedy is acknowledged.

Appendix A. Supplementary material Fig. 6. Charge/discharge curves of the series connection of three biofuel cells (see Fig. S9); Experimental setup: the connected cells were allowed to reset at open-circuit for 2 h, Supplementary data associated with this article can be found in the followed by discharging at 10 μA for 0.5 ms every 5 s reset. (A) is the first measurement online version at http://dx.doi.org/10.1016/j.bios.2016.11.012. of 1500 discharging pulses; (B) is for the measurement of 1300 discharging pulses upon refilling of fresh solutions; insets show zooms at the specific cycles. References larger, leading to a voltage drop of 84 mV (Fig. S9B). A specific Agnes, C., Holzinger, M., Le Goff, A., Reuillard, B., Elouarzaki, K., Tingry, S., Cosnier, S., −2 capacitance of 31.6 µF cm , 11 times lower than that reported on 2014. Energy Environ. Sci. 7 (6), 1884–1888. the NPG based device, was estimated. Andoralov, V., Falk, M., Suyatin, D.B., Granmo, M., Sotres, J., Ludwig, R., Popov, V.O., Schouenborg, J., Blum, Z., Shleev, S., 2013. Sci. Rep. 3. http://dx.doi.org/10.1038/ srep03270. 3.4. A proof-of-concept pulse generator Calabrese Barton, S., Gallaway, J., Atanassov, P., 2004. Chem. Rev. 104 (10), 4867–4886. Cooney, M., Svoboda, V., Lau, C., Martin, G., Minteer, S., 2008. Energy Environ. Sci. 1 A cardiac pacemaker possesses dual-function of sensing and pacing (3), 320–337. the heart (Sanders, 2008). To be able to pace the heart, an electric Cracknell, J.A., Vincent, K.A., Armstrong, F.A., 2008. Chem. Rev. 108 (7), 2439–2461. stimulus generated by the pulse generator with a fixed pulse potential Ding, Y., Kim, Y.J., Erlebacher, J., 2004. Adv. Mater. 16 (21), 1897–1900. Fabiano, S., Tran-Minh, C., Piro, B., Dang, L.A., Pham, M.C., Vittori, O., 2002. Mater. Sci. and width (i.e. threshold, the minimum voltage from the pacemaker to Eng.: C 21 (1–2), 61–67. initiate a heartbeat) is required. Previous attempts proposed the Falk, M., Andoralov, V., Blum, Z., Sotres, J., Suyatin, D.B., Ruzgas, T., Arnebrant, T., possibility of using BFCs as power sources to replace lithium based Shleev, S., 2012. Biosens. Bioelectron. 37 (1), 38–45. batteries currently used in pacemakers (MacVittie et al., 2013; Falk, M., Alcalde, M., Bartlett, P.N., De Lacey, A.L., Gorton, L., Gutierrez-Sanchez, C., Haddad, R., Kilburn, J., Leech, D., Ludwig, R., Magner, E., Mate, D.M., Conghaile, Southcott et al., 2013). To increase the voltage to the required value, P.Ó., Ortiz, R., Pita, M., Pöller, S., Ruzgas, T., Salaj-Kosla, U., Schuhmann, W., three individual cells were connected in series, giving a potential of ca. Sebelius, F., Shao, M., Stoica, L., Sygmund, C., Tilly, J., Toscano, M.D., 1.24 V for 2 h in the presence of 20 mM glucose. A test sequence of 5 s Vivekananthan, J., Wright, E., Shleev, S., 2014. PLoS One 9 (10), e109104. μ Flexer, V., Brun, N., Courjean, O., Backov, R., Mano, N., 2011. Energy Environ. Sci. 4 (6), reset (i.e. a frequency of 0.2 Hz) and 0.5 ms discharge at 10 A that 2097–2106. matched typical pacemaker working characteristic was applied (Mallela Gao, Z., Binyamin, G., Kim, H.H., Barton, S.C., Zhang, Y., Heller, A., 2002. Angew. Chem. et al., 2004). As shown in Fig. 6A, the voltage dropped steadily to ca. Int. Ed. 41 (5), 810–813. – González-Arribas, E., Pankratov, D., Gounel, S., Mano, N., Blum, Z., Shleev, S., 2016. 0.8 V in the initial 200 250 cycles, as 5 s was not long enough for Electroanalysis 28 (6), 1290–1297. voltage recovery (inset of Fig. 6A), and then maintained at ca. 0.74 V in Habermüller, K., Ramanavicius, A., Laurinavicius, V., Schuhmann, W., 2000. the following long-term testing cycles. Such an output voltage stabilised Electroanalysis 12 (17), 1383–1389. Hanashi, T., Yamazaki, T., Tsugawa, W., Ferri, S., Nakayama, D., Tomiyama, M., at 0.74 V is enough to exceed the mean pacing threshold (e.g. 0.51 ± Ikebukuro, K., Sode, K., 2009. Biosens. Bioelectron. 24 (7), 1837–1842. 0.22 V reported previously (Ritter et al., 2015)). Upon refilling the Jenkins, P.A., Boland, S., Kavanagh, P., Leech, D., 2009. Bioelectrochemistry 76 (1–2), solution, the series connected cells recovered to a potential of ca. 1.21 V 162–168. Khomenko, V., Frackowiak, E., Béguin, F., 2005. Electrochim. Acta 50 (12), 2499–2506. after a 2-h incubation period, followed by 1300 cycles of charge/ Kim, Y.H., Campbell, E., Yu, J., Minteer, S.D., Banta, S., 2013. Angew. Chem. Int. Ed. 52 discharge (Fig. 6B). The output voltage gradually decreased for the (5), 1437–1440. initial 200 cycles due to the relatively short reset time of 5 s and Kizling, M., Draminska, S., Stolarczyk, K., Tammela, P., Wang, Z., Nyholm, L., Bilewicz, – attained a stable value of 0.7 V after the 220th cycle. The results R., 2015a. Bioelectrochemistry 106 (Part A), 34 40. Kizling, M., Stolarczyk, K., Kiat, J.S.S., Tammela, P., Wang, Z., Nyholm, L., Bilewicz, R., demonstrate that three cells connected in series can mimic a pace- 2015b. Electrochem. Commun. 50, 55–59. maker generating 0.2 Hz pulses (10 μA, 0.5 ms) with a stable output Lang, X., Hirata, A., Fujita, T., Chen, M., 2011. Nat. Nanotechnol. 6, 232–236. – potential of 0.7 V. Leech, D., Kavanagh, P., Schuhmann, W., 2012. Electrochim. Acta 84 (0), 223 234. Liang, P., Huang, X., Fan, M.-Z., Cao, X.-X., Wang, C., 2007. Appl. Microbiol. Biotechnol. 77 (3), 551–558. 4. Conclusions MacVittie, K., Halamek, J., Halamkova, L., Southcott, M., Jemison, W.D., Lobel, R., Katz, E., 2013. Energy Environ. Sci. 6 (1), 81–86. Magrez, A., Kasas, S., Salicio, V., Pasquier, N., Seo, J.W., Celio, M., Catsicas, S., A supercapacitor/enzymatic biofuel cell hybrid device was prepared Schwaller, B., Forró, L., 2006. Nano Lett. 6 (6), 1121–1125. by a facile, one-step electrodeposition of PEDOT/Os polymer/enzyme Mallela, V.S., Ilankumaran, V., Rao, N., 2004. Ind. Pacing Electrophysiol. J. 4 (4), 201–212. onto dealloyed nanoporous gold electrodes. The dual-function proper- Meng, F., Ding, Y., 2011. Adv. Mater. 23 (35), 4098–4102. ties of this hybrid device allowed the energy yielded by the biofuel cell Milton, R.D., Lim, K., Hickey, D.P., Minteer, S.D., 2015. Bioelectrochemistry 106 (Part to be stored in the supercapacitor and delivered at a significantly high A), 56–63. Miyake, T., Haneda, K., Nagai, N., Yatagawa, Y., Onami, H., Yoshino, S., Abe, T., power pulse. For instance, it permitted a pulse current density of – −2 −2 Nishizawa, M., 2011. Energy Environ. Sci. 4 (12), 5008 5012. 2mAcm , with an instant maximum power density of 609 μWcm , Murata, K., Kajiya, K., Nakamura, N., Ohno, H., 2009. Energy Environ. Sci. 2 (12), 468 times higher than that of the BFC. The modification layer showed 1280–1285.

101 X. Xiao et al. Biosensors and Bioelectronics 90 (2017) 96–102

Ó Conghaile, P., Falk, M., MacAodha, D., Yakovleva, M.E., Gonaus, C., Peterbauer, C.K., 12 (12), 1157–1167. Gorton, L., Shleev, S., Leech, D., 2016. Anal. Chem. 88 (4), 2156–2163. Shao, M., Nadeem Zafar, M., Sygmund, C., Guschin, D.A., Ludwig, R., Peterbauer, C.K., Pankratov, D., Blum, Z., Shleev, S., 2014a. ChemElectroChem 1 (11), 1798–1807. Schuhmann, W., Gorton, L., 2013. Biosens. Bioelectron. 40 (1), 308–314. Pankratov, D., González-Arribas, E., Blum, Z., Shleev, S., 2016. Electroanalysis 28 (6), Shen, W., Deng, H., Teo, A.K.L., Gao, Z., 2013. J. Power Sources 226, 27–32. 1250–1266. Skunik-Nuckowska, M., Grzejszczyk, K., Stolarczyk, K., Bilewicz, R., Kulesza, P.J., 2014. Pankratov, D., Blum, Z., Suyatin, D.B., Popov, V.O., Shleev, S., 2014b. ChemElectroChem J. Appl. Electrochem. 44 (4), 497–507. 1 (2), 343–346. Sode, K., Yamazaki, T., Lee, I., Hanashi, T., Tsugawa, W., 2016. Biosens. Bioelectron. 76, Pinyou, P., Conzuelo, F., Sliozberg, K., Vivekananthan, J., Contin, A., Pöller, S., Plumeré, 20–28. N., Schuhmann, W., 2015. Bioelectrochemistry 106 (Part A), 22–27. Southcott, M., MacVittie, K., Halamek, J., Halamkova, L., Jemison, W.D., Lobel, R., Katz, Rasmussen, M., Abdellaoui, S., Minteer, S.D., 2015. Biosens. Bioelectron. 76, 91–102. E., 2013. Phys. Chem. Chem. Phys. 15 (17), 6278–6283. Reiter, S., Habermüller, K., Schuhmann, W., 2001. Sens. Actuat. B: Chem. 79 (2–3), Suraniti, E., Courjean, O., Gounel, S., Tremey, E., Mano, N., 2013. Electroanalysis 25 (3), 150–156. 606–611. Ritter, P., Duray, G.Z., Steinwender, C., Soejima, K., Omar, R., Mont, L., Boersma, L.V., Sygmund, C., Staudigl, P., Klausberger, M., Pinotsis, N., Djinović-Carugo, K., Gorton, L., Knops, R.E., Chinitz, L., Zhang, S., Narasimhan, C., Hummel, J., Lloyd, M., Simmers, Haltrich, D., Ludwig, R., 2011. Microb. Cell Fact. 10 (1), 1–9. T.A., Voigt, A., Laager, V., Stromberg, K., Bonner, M.D., Sheldon, T.J., Reynolds, D., Trasatti, S., Petrii, O.A., 1991. Pure Appl. Chem. 63 (5), 711–734. 2015. Eur. Heart J.. Winter, M., Brodd, R.J., 2004. Chem. Rev. 104 (10), 4245–4270. Sanders, R.S., 2008. The pulse generator. In: Kusumoto, F.M., Goldschlager, N.F. (Eds.), Xiao, X., Magner, E., 2015. Chem. Commun. 51 (70), 13478–13480. Cardiac Pacing for the ClinicianBoston, MA. Springer, US, 47–71. Xiao, X., Si, P., Magner, E., 2016. Bioelectrochemistry 109, 117–126. Santoro, C., Soavi, F., Serov, A., Arbizzani, C., Atanassov, P., 2016. Biosens. Bioelectron. Xiao, X., Wang, Me, Li, H., Si, P., 2013. Talanta 115 (15), 1054–1059. 78, 229–235. Xiao, X., Ulstrup, J., Li, H., Zhang, J., Si, P., 2014. Electrochim. Acta 130, 559–567. Scanlon, M.D., Salaj-Kosla, U., Belochapkine, S., MacAodha, D., Leech, D., Ding, Y., Xu, S., Minteer, S.D., 2012. ACS Catal. 2 (1), 91–94. Magner, E., 2012. Langmuir 28 (4), 2251–2261. Zafar, M.N., Beden, N., Leech, D., Sygmund, C., Ludwig, R., Gorton, L., 2012. Anal. Schneider, C.A., Rasband, W.S., Eliceiri, K.W., Schindelin, J., Arganda-Carreras, I., Frise, Bioanal. Chem. 402 (6), 2069–2077. E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., 2012. Nat. Methods 9 (7), 671. Zloczewska, A., Celebanska, A., Szot, K., Tomaszewska, D., Opallo, M., Jönsson- Schuhmann, W., Kranz, C., Wohlschläger, H., Strohmeier, J., 1997. Biosens. Bioelectron. Niedziolka, M., 2014. Biosens. Bioelectron. 54, 455–461.

102 Biosensors and Bioelectronics 98 (2017) 421–427

Contents lists available at ScienceDirect

Biosensors and Bioelectronics

journal homepage: www.elsevier.com/locate/bios

An oxygen-independent and membrane-less glucose biobattery/ MARK supercapacitor hybrid device ⁎ ⁎ Xinxin Xiaoa, , Peter Ó Conghaileb,1, Dónal Leechb, Roland Ludwigc, Edmond Magnera, a Department of Chemical Sciences and Bernal Institute, University of Limerick, Limerick, Ireland b School of Chemistry & Ryan Institute, National University of Ireland Galway, Galway, Ireland c Department of Food Science and Technology, BOKU-University of Natural Resources and Life Sciences, Muthgasse 18, 1190 Vienna, Austria

ARTICLE INFO ABSTRACT

Keywords: Enzymatic biofuel cells can generate electricity directly from the chemical energy of biofuels in physiological Biobattery ﬂuids, but their power density is signiﬁcantly limited by the performance of the cathode which is based on Enzymatic biofuel cell oxygen reduction for in vivo applications. An oxygen-independent and membrane-less glucose biobattery was Supercapacitor prepared that consists of a dealloyed nanoporous gold (NPG) supported glucose dehydrogenase (GDH) Hybrid device bioanode, immobilised with the assistance of conductive polymer/Os redox polymer composites, and a solid- Oxygen-independent state NPG/MnO cathode. In a solution containing 10 mM glucose, a maximum power density of 2.3 µW cm−2 Nanoporous gold 2 at 0.21 V and an open circuit voltage (OCV) of 0.49 V were registered as a biobattery. The potential of the

discharged MnO2 could be recovered, enabling a proof-of-concept biobattery/supercapacitor hybrid device. The resulting device exhibited a stable performance for 50 cycles of self-recovery and galvanostatic discharge as a supercapacitor at 0.1 mA cm−2 over a period of 25 h. The device could be discharged at current densities up to 2mAcm−2 supplying a maximum instantaneous power density of 676 μWcm−2, which is 294 times higher than that from the biobattery alone. A mechanism for the recovery of the potential of the cathode, analogous to that

of RuO2 (Electrochim. Acta 42(23), 3541–3552) is described.

1. Introduction respectively (Maggs et al., 1995)), together with possible mass transport limitation of oxygen, making oxygen reducing biocathode a The use of enzymatic biofuel cells (EBFCs) is of promise in significant limiting factor in the application of EBFCs. For example, generating electricity from fuels (Leech et al., 2012; Rasmussen the theoretical power output of an in vivo 1 cm long tubular glucose/ et al., 2015). EBFCs function at physiological temperature and pH, in oxygen EBFC is solely determined by the oxygen reduction reaction comparison to traditional fuel cells utilising abiotic catalysts which (ORR) at the cathode (Pankratov et al., 2016b). Moreover, the stability generally operate in harsh environments (e.g. strongly acidic or alka- of the enzymes used, predominantly multi-copper oxidases such as line media). Immobilisation of enzymes at the anode and cathode can laccase and bilirubin oxidase (BOx), needs to be considered. Laccase eliminate the requirement for membranes that are required in con- prefers a weakly acidic environment (ca. pH 4–5) and is inhibited by ventional fuel cells to separate the anode and cathode compartments. halide ions (Salaj-Kosla et al., 2013; Spira-Solomon et al., 1986; Vaz- In vivo EBFCs utilising oxygen and glucose are of significant interest Dominguez et al., 2008; Xu, 1996). In comparison to laccases, BOx is due to potential applications as miniaturised power sources for more stable under physiological conditions (pH 7.4, no inhibition in − implantable medical devices (Calabrese Barton et al., 2004) such as the presence of Cl ). However, the operational stability of BOx based cardiac pacemakers (MacVittie et al., 2013) and insulin pumps. electrodes is limited, for example, an osmium polymer “wired” However, the successful application of autonomous biomedical devices Trachyderma tsunodae BOx displayed a current loss of 78% after 2 h is a significant challenge due to the requirements for high power rotation at 100 rpm, a loss that was mainly ascribed to the irreversible density, biocompatibility and long lifetime (Shleev, 2017). The con- deactivation of BOx Cu-centers in the oxidised state (Kang et al., 2006). centration of oxygen in vivo is significantly lower (0.14 mM in arterial Air-breathing biocathodes can be employed to circumvent limita- blood and 0.08 mM in intestinal tissue (Carreau et al., 2011; Shleev, tions in the supply of oxygen, but can only be used in subcutaneous 2017)) than that of glucose (3.3 and 4.8 mM in muscle and plasma, devices (Miyake et al., 2011). Recently, molecular oxygen-independent

⁎ Corresponding authors. E-mail addresses: [email protected] (X. Xiao), [email protected] (E. Magner). 1 Current address: National Centre for Sensor Research, School of Chemical Sciences, Dublin City University, Dublin 9, Ireland. http://dx.doi.org/10.1016/j.bios.2017.07.023 Received 11 June 2017; Received in revised form 5 July 2017; Accepted 9 July 2017 Available online 11 July 2017 0956-5663/ © 2017 Elsevier B.V. All rights reserved. X. Xiao et al. Biosensors and Bioelectronics 98 (2017) 421–427

++ − hybrid EBFCs or biobatteries relying on a combination of enzymatic Mn(IV)O2 +xNa +yH + ( x + y ) e ↔ Mn ( III )(+xy) Mn ( IV ) 1−(+xy) anodes and solid-state cathodes have been proposed to address the OONa H underlying problems of enzymatic cathode based EBFCs. These abiotic xy (1) cathodes utilise cheap and abundantly available materials such as where 0 < (x+y)≤1. In this case, the discharged form is insoluble, Prussian Blue (PB) (Addo et al., 2011), Ag2O/Ag (Yu et al., 2016b) and avoiding issues with leakage. (ii) a moderate onset potential, resulting MnO2 (Yu et al., 2016a), which can be reduced/discharged via an in a biobattery with a considerable OCV (Yu et al., 2016a). (iii) external circuit, resulting in rechargeable biobatteries. For example, operation at neutral pH that is amenable to enzymes. (iv) inert to the the oxidation of PB to Berlin Green (BG) occurs at a high potential of oxidation of glucose, as confirmed by Dong et al. (Yu et al., 2016a), ff 0.87 V vs. SCE (Ne , 1985), exceeding the redox potentials of multi- resulting in a membrane-less biobattery. A spontaneous recovery of the copper oxidases. Minteer et al. developed a rechargeable ethanol potential of the discharged NPG/MnO2 was observed in open-circuit biobattery based on an alcohol dehydrogenase (ADH) modified bioa- mode, similar to that reported with a pseudo-capacitive RuO2 electrode node and a PB paste cathode that registered an open circuit voltage (Liu et al., 1997). The assembled NPG/PEDOT/Os(bpy)2PVI/FAD- (OCV) of up to 1.2 V (Addo et al., 2011). Dong et al. combined a GDH//NPG/MnO2 biobattery/supercapacitor hybrid device delivered glucose dehydrogenase (GDH) bioanode with an Ag2O/Ag (Yu et al., intermittent electric signals, with a power density much higher than 2016b) or MnO2 cathode (Yu et al., 2016a) to fabricate oxygen- that of the biobattery itself. independent recycled biobatteries with reported OCVs of 0.59 V and 0.43 V, respectively. Microbial biobatteries consisting of anodes colonized by microorganisms and reoxidisable solid-state cathodes such as 2. Experimental section Ag2O/Ag (Xie et al., 2013) and PB (Xie et al., 2015) were stable, showing no loss of capacity over 20 cycles of operation (Xie et al., 2.1. Materials and apparatus 2015). Biofuel cell (BFC)/supercapacitor hybrid devices, or self-charging Sodium phosphate (monobasic dehydrate ≥99% and dibasic ≥99%), biocapacitors, utilising capacitive bioelectrodes are of great interest due sodium sulfate (≥99.99%), manganese(II) acetate tetrahydrate to their ability to generate repeated electric pulses, with an instanta- (99.99%), D-(+)-glucose (99.5%), 3,4-ethylenedioxythiophene (EDOT, neous power density that is significantly higher than that from the BFC 97%) were obtained from Sigma-Aldrich Ireland, Ltd. All solutions itself (Agnes et al., 2014). Biocapacitors taking advantage of enzymes were prepared with deionised water (18.2 MΩ cm, Elga Purelab Ultra,

(Agnes et al., 2014; Kizling et al., 2015; Knoche et al., 2016; Pankratov UK). Os(bpy)2PVI was prepared according to an established procedure et al., 2014), microbes (Santoro et al., 2016) and thylakoids (Forster and Vos, 1990; Kober et al., 1988). Oxygen-insensitive, (Pankratova et al., 2017) have been presented. Recently, we described recombinant Glomerella cingulata FAD-GDH (EC 1.1.99.10, D-glu- a supercapacitive EBFC prepared by the immobilisation of flavin cose: acceptor 1-oxidoreductase) was expressed in Pichia pastoris and adenine dinucleotide-dependent GDH (FAD-GDH) and BOx with purified with a specific activity of 572 U mg−1 (Sygmund et al., 2011). electrodeposited poly(3,4-ethylenedioxythiophene) (PEDOT) and the Dealloyed NPG leaves were obtained by floating ca. 100 nm thick ′ +/2+ redox polymer [Os(2,2 -bipyridine)2(polyvinylimidazole)10Cl] Au/Ag leaves (12-carat, Eytzinger, Germany) on concentrated HNO3 (Os(bpy)2PVI) on dealloyed nanoporous gold (NPG) (Xiao et al., (Sigma-Aldrich) for 30 min at 30 °C (Xiao and Magner, 2015; Xiao 2017). The device could operate as a pulse generator to mimic that et al., 2014). And then placed on well-polished glassy carbon electrodes in a cardiac pacemaker, producing 10 μA pulses for 0.5 ms at a (GCEs, diameter: 4 mm). The NPG electrodes were cleaned by scanning frequency of 0.2 Hz. the potential over the range of −0.2–1.65 V in 1 M H2SO4 at a scan rate In this contribution, we substitute the BOx biocathode with a non- of 100 mV s−1 for 15 cycles. enzymatic MnO2 cathode to assemble an oxygen-independent glucose Scanning electron microscopy (SEM) images were collected using a biobattery/supercapacitor hybrid device (Scheme 1). At neutral pH Hitachi SU-70 microscope (operating at 15 kV), equipped with an MnO2 only shows catalytic activity towards oxygen at negative poten- energy dispersive X-ray spectroscopy (EDX). Transmission electron tials (Zhang et al., 2009), outside the potential window needed in this microscopy (TEM, JEOL JEM-2100, operating voltage of 200 kV) work and is thus used as a consumed cathode. MnO2 has been selected images of the electrodes were obtained on samples mounted on 300- based on several considerations: (i) a higher pseudo-capacitance in mesh copper grids (S147-3, Agar Scientific, UK). The average pore size comparison to carbon materials (Simon and Gogotsi, 2008). MnO2 is and layer thickness were measured with ImageJ software (National partially charged/discharged via the intercalation/deintercalation of Institutes of Health, Bethesda, Maryland) (Schneider et al., 2012) using + electrolyte cations (e.g. Na ) and protons according to the reaction: at least 30 measurement points. Raman spectra of MnO2 deposited on

Scheme 1. Schematic diagrams of the hybrid device working at the reset (left) and galvanostatic discharging mode (right). The scheme in the middle depicts the relevant potential diﬀerences, with potential shifts caused by galvanostatic discharging (blue arrows) and on the recovery of the potential during the quiescent step (red arrows).

422 X. Xiao et al. Biosensors and Bioelectronics 98 (2017) 421–427 gold foils (thickness: 0.1 mm, purity: 99.9%) were recorded with a times, e.g. 300 s, resulted in the formation of thick ﬁlms that blocked LabRAM 300 Raman spectrometer (Horiba Jobin Yvon) using an the pores (Fig. S2C), which were likely to be detached from the excitation source at 514 nm (Ar laser). electrode, leading to signiﬁcantly degraded operational stability. A deposition period of 180 s was chosen for further electrochemical

2.2. Preparation of the enzyme modiﬁed anode and NPG/MnO2 study. NPG/MnO2 (180 s) showed an initial speciﬁc capacitance of 1.5 −2 cathode ± 0.1 mF cm , which was almost four times higher than that of MnO2 on planar gold obtained using the same procedure and six-fold higher

Electrodeposition was performed in phosphate buffer solution than that of bare NPG (Xiao et al., 2017). NPG/MnO2 retained 64% of (PBS, 0.1 M pH 7.0) containing 2 mM polyethylene glycol 3400 its capacitance, while planar Au/MnO2 only retained 26% after 50 −1 −1 (PEG3400), 20 mM EDOT, 0.5 mg ml Os(bpy)2PVI and 0.5 mg ml charge-discharge cycles (Fig. S5)reflecting the role of the substrate of FAD-GDH using a pulse sequence of 0.9 V(2 s) and −0.4 V (3 s) for a NPG in stabilising the coating layer due to the confinement effects. total time of 300 s (Xiao et al., 2017). Fig. 1D shows a linear sweep voltammogram (LSV) of NPG/MnO2 NPG/MnO2 was fabricated via potentiostatic electrodeposition in in 0.1 M pH 7.0 PBS, exhibiting a cathodic reduction with an onset 0.1 M Na2SO4 and 0.1 M Mn(CH3COO)2 solution at 0.45 V vs. SCE and potential of ca. +433 mV and a net cathodic current density of −2 30 °C for certain durations. The as-prepared NPG/MnO2 electrodes 72 µA cm at 0.15 V. This reaction was oxygen independent (Eq. + + were subsequently immersed in solutions of 1 M H2SO4 and deionized (1)), undergoing insertion of H and Na (Yang et al., 2016). The water. observed discharge ability enables NPG/MnO2 to act as a consumed solid-state cathode (Yu et al., 2016a), a potential alternative to ORR 2.3. Electrochemical measurements active enzymes based biocathodes.

Electrochemical studies were performed with a CHI802 potentio- 3.2. Electrochemical performance of the bioanode and assembled stat (CH Instruments, Austin, Texas) in a three-electrode electroche- biobattery mical cell, with the NPG electrode, platinum wire and saturated calomel electrode (SCE) as the working, counter and reference A previously optimised NPG/PEDOT/Os(bpy)2PVI/FAD-GDH electrodes, respectively. To obtain the power density proﬁle of the bioanode was prepared for the oxidation of glucose (Xiao et al., assembled biobattery, the bioanode and NPG/MnO2 cathode were used 2017). Brieﬂy, a pulse sequence consisting of anodic (0.9 V for 2 s) as the working electrode and combined counter/reference electrode in and cathodic −0.4 V (3 s) potentials resulted in the successive deposi- a two-electrode system. The current was recorded over the potential tion of PEDOT and Os(bpy)2PVI with the co-immobilisation of enzyme range open circuit voltage of the BFC to 0 V at a scan rate of 1 mV s−1 into the polymer matrix. CVs of the bi-functional electrode displayed a in N2-bubbled 0.1 M pH 7.0 PBS containing 10 mM glucose. The power response corresponding to the charge/discharge currents from the density curve was calculated accordingly. All experiments were carried capacitive materials and the redox reaction of Os2+/3+ (Fig. 2A). The out at room temperature (20 ± 2 °C) unless stated otherwise. midpoint potential of the osmium redox couple was +210 mV vs. SCE, Testing of the charge/discharge properties of the biobattery was very close to its reported formal potential of +220 mV vs. Ag/AgCl performed in a PBS solution in (0.1 M pH 7.0) containing 10 mM (Jenkins et al., 2009). On addition of 10 mM glucose, a sigmoidal glucose using an Autolab PGSTAT100 potentiostat (Eco Chimie, response (Fig. 2A) arising from the catalytic oxidation of glucose was

Netherlands). The NPG/MnO2 and bioanode were used as working observed (vide infra, indicative of the immobilisation of FAD-GDH). An and combined counter/reference electrodes, respectively. The testing onset potential of −18 ± 9 mV vs. SCE was observed. sequence comprised (i) stand at open-circuit while recording the open The NPG/MnO2 cathode and FAD-GDH based bioanode were circuit potential (OCP) and (ii) galvanostatic discharge at defined assembled and tested without using a membrane. For the first test current densities. (blank line, Fig. 2B), the biobattery registered a maximum current density of 14 µA cm−2, a maximum power density of 2.3 µW cm−2 at 3. Results and discussion 0.21 V and an OCV of 0.49 V. This performance is an improvement over an equivalent EBFC with a BOx cathode that had a maximum −2 3.1. Electrochemical performance of NPG/MnO2 power density of 1.3 µW cm and an OCV of 0.46 V (Xiao et al., 2017). A subsequent test (red line, Fig. 2B) showed a decreased power density Anodic deposition is a widely-used method to oxidise Mn(II) (max. 1.9 µW cm−2) and OCV (0.39 V) due to partial discharge of dissolved in solution to MnO2 which is deposited as a film on an MnO2. To demonstrate the recovery behavior of the cathode, NPG/ electrode (Tench and Warren, 1983). The specific capacitance of NPG/ MnO2 was then transferred into a three-electrode cell containing PBS MnO2 electrodes increased linearly with deposition time (Fig. S1), in and oxidised at 0.5 V vs. SCE for 120 s. The OCV was restored to 0.49 V agreement with previous reports (Chen et al., 2013; Kang et al., 2013). (blue line, Fig. 2B), the same value of the initial test, with a maximum The formation of a coating layer was verified by SEM (Fig. S2 and power density of 2.1 µW cm−2, approaching the initial value. The Fig. 1A) and the presence of Mn was confirmed by EDX (Fig. 1B). A recovery of maximum power density and the OCV also implied that Raman band at 657 cm−1 was assigned to manganese oxide in the form the Mn(IV) was reduced to Mn(III) which is insoluble and retained in of Mn(III) and Mn(IV) (Fig. S3)(García et al., 2005). Unmodified NPG the film, unlike Mn(II) which is soluble and could diffuse into solution had a typical porous structure comprising interconnected pores and causing unwanted side reactions. ligaments (Xiao et al., 2016), with a uniform diameter of 30.6 ± 5 nm (Fig. S2A). The coating layer obtained after 30 s deposition was not 3.3. Electrochemical performance of the hybrid device clearly visible in the SEM image (Fig. S2B), but could be clearly identified after deposition for 180 (Fig. 1A) and 300 s (Fig. S2C). The cell was also tested as a hybrid device in N2-bubbled 10 mM Using TEM, the NPG/MnO2 composite material could be distin- glucose solution. It was reset at the open-circuit mode for 30 min (cut- guished by the contrast difference between the modified layer and the off at 0.4 V) and subsequently galvanostatic discharged at 0.1 mA cm−2 gold skeleton (Fig. 1C, Fig. S4). The electrodeposited layer along the (cut-off at 0 V), a level significantly higher than the discharge current of pore surfaces showed a relatively uniform thickness of 5.4 ± 1 nm. The the biobattery mode (14 µA cm−2). Once the potential of the built-in previous report, where the same methodology was used but with much asymmetric capacitor was discharged to a potential close to zero, thicker coating layers, showed that electrodeposited MnO2 nanocrys- interestingly, the potential recovered towards the OCV of the biobattery tals possess a spinel structure (Kang et al., 2013). Long deposition (Fig. 3A). The mechanism of this is described in the next section. The

423 X. Xiao et al. Biosensors and Bioelectronics 98 (2017) 421–427

Fig. 1. SEM (A) and TEM (C) image of NPG/MnO2 (deposition time: 180 s). (B) EDX spectra of bare NPG and NPG/MnO2 (deposition time: 180 s). (D) LSV of NPG/MnO2 (deposition time: 180 s) in 0.1 M pH 7.0 PBS at a scan rate of 2 mV s−1. device could be used for 50 cycles (25 h) of discharge with slight the observed voltage drop. The ability to generate high-power-density decreases in the onset potential for discharge (in the range of 0.35 and pules is promising in the development of a hybrid device as a 0.39 V). miniaturised power source to generate electric stimuli (e.g. cardiac The hybrid devices were discharged at various current densities up pacemakers). to 2 mA cm−2 (Fig. 3B). Current densities of 1 and 2 mA cm−2 led to Table 1 summarises the performance of representative enzyme maximum instantaneous power densities of 378 and 676 μWcm−2, based power sources consuming glucose as substrate. In comparison to respectively, 164 and 294 times higher than that from a biobattery EBFCs utilising gold nanomaterials including gold nanoparticles configuration (2.3 µW cm−2). The significantly improved instantaneous (AuNPs) (Wang et al., 2012), highly-ordered macroporous gold power density was attributed to the intrinsic nature of the super- (MPG) (Boland and Leech, 2012) and dealloyed NPG (Siepenkoetter capacitor. The specific capacitance of the asymmetric supercapacitor et al., 2017; Xiao and Magner, 2015) under similar testing conditions, was 320 μFcm−2 according to the galvanostatic discharge curve. The the biobattery displays reasonable output in terms of maximum power ohmic resistance during discharge was estimated to be 458 Ω based on density and OCV. Significantly higher power density was achieved with

Fig. 2. (A) CVs of the NPG/PEDOT/Os(bpy)2PVI/FAD-GDH bioanode. (B) The performance of the biobattery in the presence of 10 mM glucose.

424 X. Xiao et al. Biosensors and Bioelectronics 98 (2017) 421–427

exchange (Ardizzone et al., 1990). The presence of abundant oxidised Ru species in the bulk region enables re-oxidation of the surface region via an electron-hopping/charge transfer mechanism. In an analogous

manner, MnO2 may undergo a similar process. The OCP of the discharged NPG/MnO2 was examined in a solution that had been saturated with either N2 or O2 (Fig. 4A). The potential slowly recovered to ca. 0.37 V in both cases, indicating that the potential recovery was

not aﬀected by O2 over a period of 30 min. Assuming that the electrodeposition of MnO2 is 100% faradaically eﬃcient (Chen et al., 2013), the amount of deposited MnO2 can be calculated by integrating the i-t curve. Applying a potential of 0.45 V vs. SCE for 180 s resulted

in the deposition of 30 nmol MnO2 onto the NPG. Galvanostatic discharge at 0.1 mA cm−2 for a short period of 3 s reduced 0.39 nmol

MnO2 on the surface, assuming that each MnO2 accepted a single electron (i.e. conversion from Mn(IV) to Mn(III)). This data indicates that each discharge step consumed a very small fraction (1.3%) of the Fig. 3. (A) Potential profile of the device for 50 cycles. Solution: 0.1 M 7.0 PBS and 10 mM glucose. Experimental protocol: reset at open-circuit for 30 min and cutoff at bulk MnO2. In the open-circuit mode, redistribution of the concentra- 0.4 V, followed by discharging at 0.1 mA cm−2 and cutoff at 0 V. (B) Charge/discharge tions of Mn(IV) in the bulk and Mn(III) at the surface leads to the curves of the biocapacitor upon various discharging current densities; Experimental recovery of the potential, in a similar manner as described with RuO2. setup: reset at open-circuit for 30 min, followed by discharging at 0.005 (a), 0.01 (b), In the discharge step, the OCP of the assembled device decreased −2 ff 0.02 (c), 0.05 (d), 0.1 (e), 0.2 (f), 0.5 (g), 1 (h), 2 (i) mA cm for 0.2 s and cuto at 0 V. rapidly to approach 0 V (Fig. 3A), i.e. a low potential difference between the bioanode and cathode. In the reset step, the charge a carbon nanotube (CNT) based biobattery (Yu et al., 2016a, 2016b), transfer within the MnO film enabled redistribution of oxidation ff 2 however, this system su ers from the disadvantage that it requires the states that resulted in the return of the potential of the cathode to + use of NAD as a cofactor. The performance of the biobattery/super- 0.37 V vs. SCE (Fig. 4A), while the catalytic oxidation of glucose by the capacitor hybrid device compares well with that of a biosupercapacitor bioanode caused the potential to decrease with time to 0.01 V vs. SCE (Xiao et al., 2017) and with an Os polymer based EBFC/supercapacitor (Fig. 4B) (Pankratov et al., 2016a). Simultaneously, the potential hybrid (Pankratov et al., 2016a) but is lower than that of CNT based difference allowed the capacitive material on the bioanode to be hybrid devices in the presence of high glucose concentration (100 and recharged, whose charge would be released together with the MnO2 200 mM) (Agnes et al., 2014; Narvaez Villarrubia et al., 2016). cathode in the next discharge step (Scheme 1). To emphasise the role of the catalytic active bioanode, we tested a device comprising a NPG/PEDOT anode without FAD-GDH and a 3.4. Potential recovery of NPG/MnO2 NPG/MnO2 cathode (Fig. S6). The potential recovery was also observed, but with a maximum OCP no higher than 0.1 V, which is Conway et al. described the mechanism involved in the recovery of lower than that (0.4 V) in the presence of FAD-GDH. Therefore, we the electrode potential of RuO electrodes that had undergone dis- 2 confirm that a bioanode is essential to harness the potential charge (Liu et al., 1997). During discharge, the outer region of the difference, making the output potential and power density of the metal oxide layer is reduced first, with reduction occurring at a much hybrid device acceptable. slower rate in the bulk material due to the limited rate of proton

Table 1 List of properties of enzymatic power sources utilising glucose as substrate.

−2 Power source Anode Cathode [Glucose] OCV (V) PMax (µWcm ) Stability Ref. (mM)

Au electrode based Au/AuNPs/ CtCDH Au/AuNPs/ MvBOx 5 0.68 3.3 ∼20% drop in 12 h of (Wang et al., 2012) EBFC continuous operation

MPG/ Os(dmbpy)2PVI/ MPG/Os(bpy)2PVI/ 10 ∼0.52 38 N/A (Boland and Leech, AnGOx MaLc 2012)

NPG/Os(dmbpy)2PVI/ NPG/Os(bpy)2PVI/ 5 0.56 3.65 25% drop in 12 h of storage (Xiao and Magner, 2015) AnGOx MvBOx

NPG/Os(bpy)2PVI/FAD- NPG-MvBOx 5 0.45 17.5 ∼40% drop in 8 h of (Siepenkoetter et al., GDH continuous operation 2017) + Biobattery CFP/NAD -IL-SWCNTs/ GF/MnO2 30 0.43 40.5 N/A (Yu et al., 2016a) GDH/CS

GCE/MWCNTs/MDB/ Ag2O/Ag 30 0.59 275 Less than 50% drop in 6 h of (Yu et al., 2016b) GDH continuous operation EBFC/SC hybrid MWCNTs/ AnGOx/catalase MWCNTs/Lc 200 1 ± 0.1 EBFC: 16 mW Charge/discharge for 5 days (Agnes et al., 2014) Hybrid: N/A at 3 mA NPG/PEDOT/ NPG/PEDOT/Os 10 0.46 EBFC: 1.3 50 cycles charge/discharge (Xiao et al., 2017) −2 Os(bpy)2PVI/FAD-GDH (bpy)2PVI/MvBOx Hybrid: 608.8 for ∼7 h at 0.2 mA cm Graphite/Os polymer/ Graphite/Os 20 0.45 EBFC: ∼3 Charge/discharge for ∼50 h (Pankratov et al., 2016a) PQQ-GDH polymer/ MvBOx Hybrid: N/A BP/MWCNTs/pMG/GDH BP/MWCNTs/BOx 100 ∼0.56 EBFC: N/A Charge/discharge for 3 days (Narvaez Villarrubia Diﬀusing NAD+ Hybrid: 1070 at 0.4 mA cm−2 et al., 2016)

Biobattery/SC NPG/PEDOT/Os NPG/MnO2 10 0.49 Biobattery: 2.3 50 cycles charge/discharge This work −2 hybrid (bpy)2PVI/FAD-GDH Hybrid: 676 for ∼25 h at 0.1 mA cm

+/2+ CDH: cellobiose dehydrogenase; Os(dmbpy)2PVI: [Os(4,4′-dimethyl-2,2′-bipyridine)2(polyvinylimidazole)10Cl] ; GOx: glucose oxidase; Lc: laccase; N/A: not available; SC: supercapacitor; MWCNTs: multi-walled carbon nanotubes; BP: buckypaper; pMG: polymerized methylene green; MDB: Meldola's blue; CFP: carbonﬁber paper; SWCNTs: single- walled carbon nanotubes; IL: ionic liquid; CS: chitosan; GF: graphite ﬂake.

425 X. Xiao et al. Biosensors and Bioelectronics 98 (2017) 421–427

−1 Fig. 4. (A) OCP of NPG/MnO2 in the presence of N2 or O2. The electrode was discharged by scanning potential from 0.5 to 0 V vs. SCE at a scan rate of 1 mV s . (B) OCP of the bioanode upon the addition of 10 mM glucose.

4. Conclusions Chem. A 1 (32), 9202–9207. Forster, R.J., Vos, J.G., 1990. Macromolecules 23 (20), 4372–4377. García, M.A., Ruiz-González, M.L., Quesada, A., Costa-Krämer, J.L., Fernández, J.F., An oxygen independent and membrane-less glucose biobattery/ Khatib, S.J., Wennberg, A., Caballero, A.C., Martín-González, M.S., Villegas, M., supercapacitor hybrid device delivering high-power-density pulses was Briones, F., González-Calbet, J.M., Hernando, A., 2005. Phys. Rev. Lett. 94 (21), 217206. could replace oxygen reducing enzymes as the presented. MnO2 Jenkins, P.A., Boland, S., Kavanagh, P., Leech, D., 2009. Bioelectrochem 76 (1–2), cathode, due to its features including the capability to be discharged 162–168. at a reasonable potential, inert to glucose and insoluble reduced state. Kang, C., Shin, H., Heller, A., 2006. Bioelectrochem 68 (1), 22–26. Most importantly, when only a fraction was discharged in the pulse Kang, J., Hirata, A., Kang, L., Zhang, X., Hou, Y., Chen, L., Li, C., Fujita, T., Akagi, K., Chen, M., 2013. Angew. Chem. Int. Ed. 52 (6), 1664–1667. mode, the spontaneous potential recovery of MnO2 occurred due to the Kizling, M., Draminska, S., Stolarczyk, K., Tammela, P., Wang, Z., Nyholm, L., Bilewicz, redistribution of the oxidation states. Coupled with a FAD-GDH based R., 2015. Bioelectrochemistry. 106 (Part A), 34–40. supercapacitive bioanode, the hybrid device function in the similar way Knoche, K.L., Hickey, D.P., Milton, R.D., Curchoe, C.L., Minteer, S.D., 2016. ACS Energy Lett., 380–385. of a biosupercapacitor. This biobattery based hybrid device holds Kober, E.M., Caspar, J.V., Sullivan, B.P., Meyer, T.J., 1988. Inorg. Chem. 27 (25), promise as an intermittent power source, which can overcome the 4587–4598. limited oxygen supply occurring on the conventional biofuel cells, for Leech, D., Kavanagh, P., Schuhmann, W., 2012. Electrochim. Acta 84 (0), 223–234. Liu, T., Pell, W.G., Conway, B.E., 1997. Electrochim. Acta 42 (23), 3541–3552. implanted medical devices. MacVittie, K., Halamek, J., Halamkova, L., Southcott, M., Jemison, W.D., Lobel, R., Katz, E., 2013. Energy Environ. Sci. 6 (1), 81–86. Acknowledgments Maggs, D.G., Jacob, R., Rife, F., Lange, R., Leone, P., During, M.J., Tamborlane, W.V., Sherwin, R.S., 1995. J. Clin. Invest. 96 (1), 370–377. Miyake, T., Haneda, K., Nagai, N., Yatagawa, Y., Onami, H., Yoshino, S., Abe, T., This project has received funding from the European Union's Nishizawa, M., 2011. Energy Environ. Sci. 4 (12), 5008–5012. Seventh Framework Programme for research, technological develop- Narvaez Villarrubia, C.W., Soavi, F., Santoro, C., Arbizzani, C., Serov, A., Rojas- Carbonell, S., Gupta, G., Atanassov, P., 2016. Biosens. Bioelectron. 86, 459–465. ment and demonstration under grant agreement no 607793. X. Xiao Neﬀ, V.D., 1985. J. Electrochem. Soc. 132 (6), 1382–1384. acknowledges a Government of Ireland Postgraduate Scholarship Pankratov, D., Blum, Z., Suyatin, D.B., Popov, V.O., Shleev, S., 2014. ChemElectroChem (GOIPG/2014/659). P. Ó Conghaile acknowledges a Technology 1 (2), 343–346. Innovation Development Award through Science Foundation Ireland Pankratov, D., Conzuelo, F., Pinyou, P., Alsaoub, S., Schuhmann, W., Shleev, S., 2016a. Angew. Chem. Int. Ed. 55 (49), 15434–15438. (15/TIDA/2887). Pankratov, D., Ohlsson, L., Gudmundsson, P., Halak, S., Ljunggren, L., Blum, Z., Shleev, S., 2016b. RSC Adv. 6 (74), 70215–70220. Appendix A. Supporting information Pankratova, G., Pankratov, D., Hasan, K., Åkerlund, H.-E., Albertsson, P.-Å., Leech, D., Shleev, S., Gorton, L., 2017. Adv. Energy Mater.. http://dx.doi.org/10.1002/ aenm.201602285. Supplementary data associated with this article can be found in the Rasmussen, M., Abdellaoui, S., Minteer, S.D., 2015. Biosens. Bioelectron. 76, 91–102. online version at doi:10.1016/j.bios.2017.07.023. Salaj-Kosla, U., Pöller, S., Schuhmann, W., Shleev, S., Magner, E., 2013. Bioelectrochemistry. 91, 15–20. Santoro, C., Soavi, F., Serov, A., Arbizzani, C., Atanassov, P., 2016. Biosens. Bioelectron. References 78, 229–235. Schneider, C.A., Rasband, W.S., Eliceiri, K.W., Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., 2012. Nat. Methods 9 (7), 671. Addo, P.K., Arechederra, R.L., Minteer, S.D., 2011. J. Power Sources 196 (7), Shleev, S., 2017. ChemPlusChem 82 (4), 522–539. http://dx.doi.org/10.1002/ – 3448 3451. cplu.201600536. ﬀ Agnes, C., Holzinger, M., Le Go , A., Reuillard, B., Elouarzaki, K., Tingry, S., Cosnier, S., Siepenkoetter, T., Salaj-Kosla, U., Xiao, X., Conghaile, P.Ó., Pita, M., Ludwig, R., Magner, – 2014. Energy Environ. Sci. 7 (6), 1884 1888. E., 2017. ChemPlusChem 82 (4), 553–560. – Ardizzone, S., Fregonara, G., Trasatti, S., 1990. Electrochim. Acta 35 (1), 263 267. Simon, P., Gogotsi, Y., 2008. Nat. Mater. 7 (11), 845–854. – Boland, S., Leech, D., 2012. Analyst 137 (1), 113 117. Spira-Solomon, D.J., Allendorf, M.D., Solomon, E.I., 1986. J. Am. Chem. Soc. 108 (17), Calabrese Barton, S., Gallaway, J., Atanassov, P., 2004. Chem. Rev. 104 (10), 5318–5328. – 4867 4886. Sygmund, C., Staudigl, P., Klausberger, M., Pinotsis, N., Djinović-Carugo, K., Gorton, L., ‐ Carreau, A., Hafny Rahbi, B.E., Matejuk, A., Grillon, C., Kieda, C., 2011. J. Cell. Mol. Haltrich, D., Ludwig, R., 2011. Microb. Cell Fact. 10 (1), 1–9. – Med. 15 (6), 1239 1253. Tench, D., Warren, L.F., 1983. J. Electrochem. Soc. 130 (4), 869–872. Chen, L.Y., Kang, J.L., Hou, Y., Liu, P., Fujita, T., Hirata, A., Chen, M.W., 2013. J. Mater. Vaz-Dominguez, C., Campuzano, S., Rüdiger, O., Pita, M., Gorbacheva, M., Shleev, S.,

426 X. Xiao et al. Biosensors and Bioelectronics 98 (2017) 421–427

Fernandez, V.M., De Lacey, A.L., 2008. Biosens. Bioelectron. 24 (4), 531–537. Xie, X., Ye, M., Liu, C., Hsu, P.-C., Criddle, C.S., Cui, Y., 2015. Energy Environ. Sci. 8 (2), Wang, X., Falk, M., Ortiz, R., Matsumura, H., Bobacka, J., Ludwig, R., Bergelin, M., 546–551. Gorton, L., Shleev, S., 2012. Biosens. Bioelectron. 31 (1), 219–225. Xu, F., 1996. Biochemistry 35 (23), 7608–7614. Xiao, X., Magner, E., 2015. Chem. Commun. 51 (70), 13478–13480. Yang, L., Cheng, S., Wang, J., Ji, X., Jiang, Y., Yao, M., Wu, P., Wang, M., Zhou, J., Liu, Xiao, X., Ulstrup, J., Li, H., Zhang, J., Si, P., 2014. Electrochim. Acta 130, 559–567. M., 2016. Nano Energy 30, 293–302. Xiao, X., Si, P., Magner, E., 2016. Bioelectrochem 109, 117–126. Yu, Y., Han, Y., Lou, B., Zhang, L., Han, L., Dong, S., 2016a. Chem. Commun. 52 (92), Xiao, X., Conghaile, P.Ó., Leech, D., Ludwig, R., Magner, E., 2017. Biosens. Bioelectron. 13499–13502. 90, 96–102. Yu, Y., Xu, M., Bai, L., Han, L., Dong, S., 2016b. Biosens. Bioelectron. 75, 23–27. Xie, X., Ye, M., Hsu, P.-C., Liu, N., Criddle, C.S., Cui, Y., 2013. Proc. Natl. Acad. Sci. 110 Zhang, L., Liu, C., Zhuang, L., Li, W., Zhou, S., Zhang, J., 2009. Biosens. Bioelectron. 24 (40), 15925–15930. (9), 2825–2829.

Research Article

Cite This: ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 www.acsami.org

Nanoporous Gold-Based Biofuel Cells on Contact Lenses † † ‡ § ‡ † Xinxin Xiao, Till Siepenkoetter, Peter ÓConghaile, , Donaĺ Leech, and Edmond Magner*, † Department of Chemical Sciences and Bernal Institute, University of Limerick, Limerick V94 T9PX, Ireland ‡ School of Chemistry & Ryan Institute, National University of Ireland Galway, Galway H91 TK33, Ireland

*S Supporting Information

ABSTRACT: A lactate/O2 enzymatic biofuel cell (EBFC) was prepared as a potential power source for wearable microelectronic devices. Mechanically stable and flexible nanoporous gold (NPG) electrodes were prepared using an electrochemical dealloying method consisting of a pre-anodization process and a subsequent electrochemical cleaning step. Bioanodes were prepared by the electrodeposition of an Os polymer and Pediococcus sp. lactate oxidase onto the NPG electrode. The electrocatalytic response to lactate could be tuned by adjusting the deposition time. Bilirubin oxidase from Myrothecium verrucaria was covalently attached to a diazonium-modified NPG surface. A flexible EBFC was prepared by placing the electrodes between two commercially available contact lenses to avoid direct contact with the eye. When tested in air-equilibrated artificial tear solutions (3 mM lactate), a maximum power density of 1.7 ± 0.1 μWcm−2 and an open-circuit voltage of 380 ± 28 mV were obtained, values slightly lower than those obtained in phosphate buffer solution (2.4 ± 0.2 μWcm−2 and 455 ± 21 mV, respectively). The decrease was mainly attributed to interference from ascorbate. After 5.5 h of operation, the EBFC retained 20% of the initial power output. KEYWORDS: electrochemical dealloying, nanoporous gold, enzymatic biofuel cell, contact lens, lactate oxidase, bilirubin oxidase

1. INTRODUCTION for continuous health monitoring and applications in sports 18 Enzymatic biofuel cells (EBFCs) have been extensively science. In comparison to implantable glucose/O2 EBFCs − investigated,1 3 since the first prototype consisting of a glucose that utilize glucose and O2 in blood, lactate/O2 biofuel cells oxidase (GOx)-modified anode and a Pt cathode was have more potential for use with wearable electronics due to introduced by Yahiro et al. in 1964.4 EBFCs enjoy advantages the higher concentration of lactate in tears and in sweat. For such as ease of miniaturization and the ability to operate at example, the normal concentrations of glucose and lactate in human blood range from 3.3 to 6.5 mM and 0.5−0.8 mM, physiological conditions. EBFCs that mimic metabolic path- − − ways (most interestingly, glucose oxidation) in the body have respectively, in comparison to 0.1 0.6 mM and 2 5 mM, respectively in human tears.19 Lactate is an important been envisioned as autonomous power suppliers for implant- ffi able medical devices since the 1970s.5 Significant efforts have biomarker of metabolic e ciency during physical exercise. A been directed toward designing implantable EBFCs, with correlation between the concentration of lactate in sweat and in − EBFCs examined in vivo in rats,6 9 snails,10 lobsters,11 etc. blood has been described by Sakharovet al., indicating that 12,13 sweat lactate levels can be measured to evaluate changes in Recently, EBFCs have been tested ex vivo in human blood. 20 However, practical applications of EBFCs operating in the blood lactate concentrations. Self-powered lactate biosensors human body have been hindered by (i) the relatively large size based on the construction of lactate EBFCs have been 9 demonstrated showing a linear increase in power density for of the devices, (ii) the limited lifetime due to the deactivation 21 and/or leakage of enzyme,13 and (iii) low power density due to lactate concentrations between 0 and 5 mM. ffi A promising type of wearable device is the temporary tattoo ine cient rates of electron transfer between the enzymes and 22 the electrode surface, accompanied by the limited mass EBFC. Wang et al. successfully fabricated an EBFC that can be attached on the skin to harvest energy from lactate present transport of substrates. The supply of O2 in particular is a 16 significant constraint for in vivo operation, and thus, oxygen- in human sweat during physical exercise. The transferable reducing biocathodes are a significant limiting factor.3,12 tattoo-based EBFC employed a mediated lactate oxidase (LOx) Noninvasive EBFCs utilizing fuels in saliva,14,15 sweat,16 bioanode and a platinum black cathode. Another interesting 23 tears,17 etc. offer an alternative that can circumvent issues type of wearable EBFC is that incorporated on a contact lens. caused by implantation. This type of EBFC avoids direct contact with the immune system and removes the necessity of a Received: December 8, 2017 surgical procedure. Such devices can be easily discarded and Accepted: February 6, 2018 replaced and can be used to activate wearable medical devices Published: February 6, 2018

Scheme 1. Schematic Diagram of the Assembly of the Modiﬁed Contact Lens (A) and the Conﬁguration of the EBFC (B)

Basal tears containing a range of species such as lactate, glucose, solution (PBS) and artificial tears, exhibiting a maximum power − ascorbate, saturated air, etc. keep the cornea moist, making the density of 2.4 ± 0.2 and 1.7 ± 0.1 μWcm 2, respectively. preparation of continuous and self-sustained EBFCs possible during physical movement and under quiescent conditions. 2. EXPERIMENTAL SECTION Contact lenses floating on the cornea for myopia correction are 2.1. Materials and Apparatus. Sodium phosphate (monobasic commercially available. Recently, new roles for such systems, − dehydrate ≥99% and dibasic ≥99%), sodium fluoride (NaF, 99.99%), including sensors,24 27 digital displays,28 and drug release29 − hydrochloric acid (HCl, 37%), sulfuric acid (H2SO4,95 98%), D- fi ≥ have been described. Falk et al. rst reported experimental (+)-glucose (99.5%), sodium nitrite (NaNO2, 99.999%), 6-amino-2- ≥ proof of a 3D nanostructured gold-wire-supported glucose/O2 naphthoic acid (NA, 90%), 3-mercaptopropionic acid (MPA, 99%), EBFC that could be used on contact lenses.30 Follow-up work L-ascorbic acid (≥99%), urea (≥99.5%), sodium L-lactate (≥99%), N- cyclohexyl-N′-(2-morpholinoethyl) carbodiimide metho-p-toluenesul- from the same group described a hybrid EBFC relying on an ≥ abiotic anode to oxidize ascorbate and a bilirubin oxidase fonate (CMC, 99%), lysozyme human (EC 3.2.1.17), bovine serum albumin (BSA), mucin from porcine stomach, and Pediococcus sp. LOx (BOx)-based biocathode to reduce oxygen in real human ≥ −1 17 (EC 1.13.12.4, 20 U mg ) were purchased from Sigma-Aldrich tears. Minteer et al. described the preparation of a Ireland, Ltd. Anhydrous acetonitrile (>99.8%) was obtained from buckypaper-supported EBFC assembled with a lactate dehy- Fisher Scientific, Ireland. Myrothecium verrucaria BOx (EC 1.3.3.5, 2.63 drogenase (LDH) bioanode and a BOx biocathode that was Umg−1) was obtained as a gift from Amano Enzyme Inc., Japan. 31 43,44 deposited on a curved elastomeric substrate, registering a Os(bpy)2PVI was synthesized using an established procedure. maximum power density of 2.4 ± 0.9 μWcm−2.32 Silicon-hydrogel contact lenses (−0.5 and −9.0 diopter) were obtained Ω Electrode materials for tear-based EBFCs should (i) be locally. Deionized water (18.2 M cm, Elga Purelab Ultra, UK) was fl used for all preparations. exible enough to be seamlessly attached onto the eyeball, (ii) Morphology studies were performed with a Hitachi SU-70 scanning exhibit high surface areas to ensure high enzyme loadings for electron microscope (SEM, 10 kV), equipped with an energy larger current densities, and (iii) be biocompatible for use on dispersive X-ray spectrometer (EDX) for residual Ag determination. the eye.23 Dealloyed nanoporous gold (NPG)33 possesses a ImageJ software (National Institutes of Health, Bethesda, Maryland)45 three-dimensional porous structure fabricated via the etching of was used to measure the average pore size and crack width of NPG by − gold alloys and is a promising substrate for EBFCs.34 37 The analyzing at least 30 measurement points. pore sizes can be finely tuned to accommodate enzymes in a 2.2. Electrochemical Dealloying. Magnetron sputtered Ag/Au 36,38,39 alloy was prepared in an ultrahigh vacuum chamber at room manner that optimizes the electrocatalytic response. Thin 38 temperature according to a previous report. Briefly, Ar-plasma- (100 nm) NPG leaves have been used as the electrode material treated microscope glass slides or 100 μm thin PET substrates were fl 40 − of exible supercapacitors. However, Au Ag NPG prepared coated with a 10 nm Ti adhesive layer, ca. 35 nm Au protective layer 34 from these leaves are brittle, while the preparation conditions and 100 nm Ag70/Au30 (atomic %) alloy layer, subsequently. The (concentrated nitric acid) used for chemical etching are highly glass-supported alloy sheets were cut using a circular saw and painted corrosive,37,41 with concomitant safety and environmental with dielectric paste (Gwent Group, UK) to define an electroactive surface area of ca. 0.3 cm2. Cleanroom tape composed of a polyamide concerns. To overcome these issues, we electrochemically fi fi * dealloyed Au−Ag alloys at neutral pH to fabricate mechanically lm (VWR, Ireland) was used to de ne the electrode area (0.35 0.35 cm2) of the PET-supported alloy sheets. robust NPG electrodes. A polyethylene terephthalate (PET) fi fl To prepare NPG, the alloy was anodized at +1.05 and +1.5 V vs lm was used to provide a exible substrate. LOx and BOx were SCE in 0.5 M NaF at room temperature (20 ± 2 °C) for 10 min and immobilized onto the electrode with the assistance of [Os(2,2′- subsequently cleaned by scanning the potential from −0.2 to 1.65 V in +/2+ 35 −1 bipyridine)2(polyvinylimidazole)10Cl] (Os(bpy)2PVI) 1MH2SO4 at a scan rate of 100 mV s for a range of potential cycles 36 and diazonium grafting, separately, for the bioanode and (1 to 15). The electrochemically addressable surface area (Areal)of biocathode. The EBFC was enclosed between two commer- NPG and the roughness factor (Rf), i.e., the ratio of Areal to the geometric area (Ageo), was obtained by cyclic voltammetry using a cially available contact lenses (Scheme 1) to avoid direct −2 31,32 value of 390 μCcm for the reduction of a single layer of gold contact with the eye. Hydrophilic silicon-hydrogel contact oxide.46 lenses contain microchannels to enable the transport of 2.3. Enzyme Immobilization. NPG-based bioanodes were 42 solutions and oxygen to the EBFC. The performance of the prepared by electrodeposition at −1.1 V for different durations EBFC was examined in solutions containing phosphate buffer (60−600 s) in 0.1 M pH 7.0 PBS containing 1 mg mL−1 of

7108 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article

Figure 1. (A) Linear sweep voltammogram of Ag70/Au30 alloy sputtered on glass in 0.5 M NaF. (B) Cyclic voltammograms of the as-anodized NPG (1.5 V) in 1 M H2SO4.

Figure 2. (A) Schematic diagram of the electrochemical dealloying process. (B−E) SEM images of the porous structure of NPG obtained at diﬀerent conditions. Anodization in 0.5 M NaF at 1.5 V vs SCE for 10 min (B), anodization and cycling potential in 1 M H2SO4 for 1 (C), 2 (D), and 15 (E) potential cycles.

Os(bpy) PVI and 1 mg mL−1 of LOx. The surface coverage (nmol constant (96 485 C mol−1), and A (cm−2) is the geometric area of the − 2 cm 2) of the Os polymer on the electrode was calculated according to electrode. the equation BOx was covalently attached to NPG via a 2-carboxy-6-naphtoyl diazonium salt (NA-DS) modiﬁcation layer, which was synthesized Q surface coverage = according to a previous report.36 Brieﬂy, a fresh NA-DS solution was (1) nFA obtained by dropwise addition of 2 mL of 2 M HCl containing 2 mM where Q (nC) is the charge regarding to the oxidation/reduction of Os NaNO2 into a 2 mL solution of 20 mM NA in acetonitrile, in an ice polymer determined based on the cyclic voltammograms (CVs) in a bath. Electrografting was achieved by electrochemically reducing NA- blank PBS, n is the number of electrons involved, F is the Faraday DS at the electrode surface with a single potential scan over the

7109 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article

Figure 3. Plots of (A) roughness factor; (B) residual silver content; (C) pore size; (D) crack width obtained after potential cycling of as-anodized NPG.

− − potential range 0.6 to −0.6 V at a scan rate of 200 mV s 1 (Figure S1). 15.7 M) at temperatures of 30 °C or higher.38,47 49 In contrast, The modified electrodes were immersed into a 1 mM MPA aqueous there has only been a few reports describing the dealloying of fi solution overnight to block any unmodi ed gold surface, followed by gold alloys at neutral pH. Such methods are based on the carefully rinsing with deionized water and drying in vacuum. A 20 μL −1 electrochemical oxidation of the less noble element and aliquot of BOx (0.5 mg mL ) was drop-cast onto the surface of the 50,51 electrode, incubated in a vacuum chamber for 5 min, and then subsequent removal of the oxidized product. Expensive transferred to a fridge at 4 °C for 1 h. The modified electrodes were salt solutions (such as AgNO3) were used to dealloy Ag65/Au35 then immersed in a solution of CMC (5 mM) at 4 °C for 2 h to cross- (atomic %) at an applied potential between 1.4 and 2 V vs link the enzyme molecules. NHE.50 Al/Au alloys were electrochemically dealloyed in 2.4. Electrochemical Measurements. Electrochemical experi- solutions of NaCl.51 However, in dealloying Ag/Au alloys, the ments were carried out with a CHI802 potentiostat (CH Instruments, use of NaCl as an electrolyte may result in the precipitation of Austin, Texas) in a three-electrode electrochemical cell consisting of insoluble AgCl.52 Therefore, NaF was selected as a low cost the Au alloy or NPG-based working electrodes, a platinum counter alternative electrolyte for electrochemical dealloying. electrode, and saturated calomel electrode (SCE) as the reference electrode. The linear sweep voltammogram (LSV) of the Ag70/Au30 The assembled EBFC was tested in a two-electrode system by using alloy (sputtered on glass) in 0.5 M NaF (Figure 1A) displayed a LOx-based bioanode as the working electrode and a BOx-based peaks corresponding to the oxidation of Ag (0.77 to 1.22 V vs biocathode as the combined counter/reference electrode, recording SCE) and the oxygen evolution reaction (OER, onset potential the current in the potential range between the open-circuit voltage of of 1.25 V vs SCE), consistent with data from the Pourbaix − the EBFC and 0 V at 1 mV s 1. The power density curve was diagram for Ag.53 To illustrate the effect of oxidation potential, calculated using the geometric area of the limiting electrode. All two representative potentials, 1.05 and 1.5 V in the region of Ag ± ° experiments were performed at room temperature (20 2 C) unless oxidation and OER, respectively, were chosen (Figure 1A). stated otherwise. After 10 min at room temperature, both potentials led to NPG To test the contact-lens-supported EBFC, a contact lens (diopter − fi with similar morphology (Figures 2B and S2A), showing 9.0) was rst mounted onto the polycarbonate packing material used ± ∼ to package the contact lens. The material had the same curved surface pinholes with very small average pore sizes (8.2 2 and 5nm fi shape of the lens. It was rst treated with O2 plasma (30 s, Solarus 950 at 1.05 and 1.5 V, respectively, Figure 3 and Tables S1 and S2) Advanced Plasma System, Gatan, U.S.A.), and the NPG-based EBFC and wide cracks due to stress release on volume contraction electrodes were then placed on the lens followed by a thinner contact during the dealloying process49 (20 ± 4.2 and 16.8 ± 4.1 nm at 32 lens (diopter −0.5) (Scheme 1). Artificial tear solutions (a mixture 1.05 and 1.5 V, respectively Figure 3 and Tables S1 and S2). of 50 μM glucose, 3 mM lactate, 180 μM ascorbate, 5.4 mM urea, 2.47 fi fi ± −1 −1 −1 EDX analysis con rmed that a signi cant amount of Ag (12.2 mg mL lysozyme, 0.2 mg mL BSA, and 0.15 mg mL mucin in 0.1 0.3% at 1.05 V and 13.6 ± 0.4% at 1.5 V, Figure 3 and Tables M pH 7.0 PBS) were maintained at 35 °C and continuously dropped onto the contact lenses with a peristaltic pump (P720, Instech, U.S.A.). S1 and S2) remained, due to the presence of residual silver oxide passivating further silver dissolution.50 This was consistent with the observation that longer anodization times 3. RESULTS AND DISCUSSION (>10 min) resulted in no significant changes in terms of average 3.1. Electrochemical Dealloying. Dealloying of gold pore size and Ag content. To obtain NPG electrodes with alloys is normally performed in concentrated nitric acid (ca. suitable pore sizes for enzyme immobilization, the residual

7110 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article oxide was removed by cycling the applied potential in 1 M 50 H2SO4, which is an established protocol to create clean gold electrode surfaces.30,41,54 As potential cycling continued (Figure 1B), the small peaks corresponding to removal of Ag at ca. + 0.2 V started to disappear after the second potential cycle. Meanwhile, the reduction peak of gold oxide at ca. 0.85 V (Figure 1B) decreased due to coarsening of NPG55 that was associated with an increase in the average pore size and a decrease in the specific surface area.41,48 SEM (Figures 2B−D and S2) and EDX (Figure 3, Table S1 and S2) results reinforced these observations. NPG anodized at 1.05 and 1.5 V (Figure 3 and Tables S1 and S2) both showed decreases in the amounts of silver remaining, with increases in pore sizes and decreases in the surface roughness. The observed cracks grew in size with continuous potential cycling, until they were in the same size range of the pores and no longer distinguishable from them after the 15th scan.38 The observed pore evolution process during potential scanning is quite similar to that Figure 4. (A) Digital photo of the PET-supported NPG obtained via reported on glass-supported Ag/Au alloys during etching by electrochemical dealloying. (B) SEM images of the electrochemically 38 concentrated nitric acid. NPG prepared at potentials of 1.05 dealloyed PET-NPG, and the corresponding microstructure under a and 1.5 V both generated satisfactory nanoporous structures in 40° (C) and 60° bend (D). terms of acceptable Rf and pore sizes. For further studies, 1.5 V was used as the optimal potential as the resulting pore sizes deformation after a 40° bend (Figure 4C). Streaks were ± ± (20.2 5.1 nm) and large Rf values (17.0 0.8) were suitable observed when the electrode was bent by 60° (Figure 4D), and for enzyme immobilization. the electrode was still conductive due to the presence of the Control experiments performed by cycling the potential in underlying Au layer. No delamination occurred even after a 90° H2SO4 without anodization also resulted in the appearance of bend due to the Ti adhesive layer. The sheet resistance (Rs)of nanoporous structures (Figure S3 and Table S3). After two the flexible NPG on PET was 3.4 ± 0.1 Ω sq−1 (Figure S4), cycles, only 4.7 ± 0.4% Ag remained in the alloy. The with slight increases to 3.5 ± 0.1 and 3.6 Ω sq−1 after bending ° electrodes had an average pore size of ca. 5 nm and also to 40 and 90 , respectively. The measured Rs was similar to that displayed large cracks (40.3 ± 8.2 nm, Figure S3A, Table S3). of a NPG leaf electrode (2.5 Ω sq−1).56 This type of PET- In contrast, anodization resulted in enriched Ag content (12.2 supported NPG is expected to find use as supercapacitors40 as ± 0.3% at 1.05 V and 13.6 ± 0.4% at 1.5 V, Figure 3, Table S1 well as in point-of-care diagnostics55 and surface-enhanced and S2) on the surface, and etching of Ag was relatively slower resonance Raman scattering (SERRS) sensors.57 Here, we as it was impeded by the presence of silver oxide, thus resulting demonstrated the use of such electrodes for EBFCs. in smaller crack sizes (20 ± 4.2 nm for 1.05 V and 16.8 ± 4.1 3.2. Characterization of the Bioelectrodes. FAD- nm for 1.5 V, Figure 3 and Table S1 and S2). Cyclic dependent LOx was selected as it can be easily immobilized 21 voltammograms of the electrodes in H2SO4 resulted in the on an electrode surface with redox polymers, which extraction of the surface-enriched Ag and reconfiguration of simultaneously immobilize the enzyme and shuttle electrons surface Au atoms (i.e., surface rearrangement).41,50,55 Without from the enzyme to the electrode surface. Os poly(N- anodization, direct etching in H2SO4 stripped Ag from the bulk, vinylimidazole) redox polymers have been shown to be immediately leading to rapid volume changes and resulting in promising mediators for use with LOx.58,59 As shown in large cracks. Additional potential scans (up to 15 cycles (Figure Scheme S1, the oxidation of L-lactate to pyruvate is catalyzed by ± S3B, Table S3)) resulted in NPG with low Rf (4.0 0.2) and LOx(FAD), with the redox center FAD converted to be the ± expanded crack sizes (44.6 5.7 nm). The obtained reduced form, FADH2. Oxidation of FADH2 by Os(III) in the microstructure showed unconnected ligaments when scanned redox polymer regenerates FAD and produces Os(II), which is to the 30th cycle (Figure S3B). In conclusion, potential cycling subsequently reoxidized to Os(III) at the electrode. Instead of in H2SO4 without anodization resulted in NPG with pores that drop-casting, which leads to a film with relatively poor stability, were too small for enzymes to enter (2 cycles) or with electrodeposition can be used to coimmobilize enzymes and Os unsatisfactory roughness (15 and 30 cycles) (Table S3). Thus, polymers containing weakly coordinated chloride ions.60 At a it is essential to first anodize the precursor to obtain the optimal cathodic potential, Os3+ was reduced to Os2+, accompanied by NPG structure. Figure 2A illustrates a possible mechanism of an exchange of chloride ions with the more strongly the two-step electrochemical dealloying process. Anodization coordinating pyridine or imidazole groups on the polymer. generates silver oxide passivated NPG with pinholes. Such a cross-linking process resulted in irreversible polymer − Subsequent coarsening of the surface in H2SO4 removes the precipitation onto the electrode. A negative potential of 1.1 V fi residual silver oxides and allows recon guration of the surface vs SCE was used for deposition in the presence of Os(bpy)2PVI gold atoms via repeated electro-oxidation/reduction. and LOx. The electrocatalytic responses of the resulting Anodization followed by cyclic voltammetry in H2SO4 was bioelectrodes varied with deposition time (Figure 5A). Surface performed to dealloy the less noble metal Ag from a PET- coverages of the Os polymer (Figure 5A, blue curve) increased supported Ag70/Au30 alloy. The morphology showed a with deposition time. In other words, longer deposition time continuous porous structure with an average pore size of 20.9 resulted in an increase in the amounts of Os polymer and ± 4.2 nm and a crack width of 28.4 ± 4.6 nm. The obtained enzyme that were immobilized.35 The highest response was NPG could be bent (Figure 4A) with no obvious structural obtained with a moderate deposition time of 360 s, reflecting a

7111 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article

ff Figure 5. (A) E ect of the deposition time for NPG/Os(bpy)2PVI/LOx on the catalytic response toward 3 mM lactate in air-equilibrated 0.1 M pH 7.0 PBS at 250 mV vs SCE. Blue line indicates the surface coverages of the Os polymer obtained by various deposition times. (B) CVs of NPG/ −1 Os(bpy)2PVI/LOx (360 s deposition) in air-equilibrated 0.1 M pH 7.0 PBS at a scan rate of 5 mV s . (C) Catalytic response of NPG/ Os(bpy)2PVI/LOx (360 s deposition) toward various concentrations of lactate in air-equilibrated 0.1 M pH 7.0 PBS at 250 mV vs SCE. (D) CVs of the BOx-modified electrode in 0.1 M pH 7.0 PBS at a scan rate of 5 mV s−1. compromise between the loading of the enzyme and the mass- negligible at high substrate concentrations.58 By varying the fi 35 transport resistance of lactate through the lm. As shown lactate concentration, it became clear that NPG/Os(bpy)2PVI/ previously, transmission electron microscopic (TEM) images LOx displayed a saturated current density when the clearly showed that the pores of NPG were blocked by the concentration was above 3 mM (Figure 5C). In other words, polymer film when the deposition time was too long, leading to oxygen competition at the bioanode was not a critical concern a decreased catalytic response.35 A deposition time of 360 s for the EBFC operation in the presence of 3 mM lactate resulted in a thin film growing along the ligaments without (Figure 5C). The electrode had a sensitivity of 19.7 ± 1.4 μA plugging the pores (Figure S5). cm−2 mM−1 with a linear range up to 3 mM. The apparent Cyclic voltammograms (CVs) of the optimal NPG/Os- Michaelis−Menten constant of the enzyme-modified electrode ± (bpy)2PVI/LOx bioanode, performed in air-equilibrated PBS, was 1.0 0.4 mM, which is lower than that obtained from fi fi ± 21 exhibited a pair of well-de ned redox peaks with an integrated FcMe2-LPEI/LOx-modi ed buckypaper (1.6 0.1 mM). anodic-to-cathodic peak area ratio of approximately one and a NPG/Os(bpy)2PVI/LOx showed considerable stability at +250 midpoint redox potential of 193 ± 1 mV vs SCE (Figure 5B), mV vs SCE in an air-equilibrated 3 mM lactate PBS solution, attributed to the rapid oxidation/reduction of the Os2+/3+.In retaining 63% of its original response after 2 h of continuous the presence of 3 mM lactate, a typical sigmoidal catalytic wave operation (Figure S7). appeared, with an onset potential of 47 ± 30 mV vs SCE and a To prepare a biocathode with a high onset potential, BOx net catalytic response of 62.2 ± 1.9 μAcm−2. It is noteworthy was covalently attached to a diazonium-layer-modified NPG that the reduced form of LOx (FADH2) can also be oxidized by (Figure S1) in order to achieve direct electron transfer (DET). 58 O2, decreasing the current and generating unwanted H2O2 The porous structure is believed to accommodate BOx in a that could deactivate the enzyme. Additionally, an oxygen manner that provides a favorable orientation for DET.36 The depleting bioanode will reduce the substrate concentration for catalytic activity of the BOx-diazonium-modified NPG was 61 the oxygen-reducing biocathode. Thus, oxygen competition studied by cyclic voltammetry in either N2-bubbled or air- was studied by comparing the catalytic response in either air- equilibrated PBS at a scan rate of 5 mV s−1 (Figure 5D). In the equilibrated or N2-saturated PBS containing 3 mM lactate. The air-equilibrated PBS, a sigmoidal catalytic curve was obtained ± ratio of measured current density from air/N2 solution was with an onset potential of 503 15 mV vs SCE and a 87.7%, implying a relatively small fraction (12.3%) was assigned background-corrected current density of 19.5 ± 0.1 μAcm−2. to oxygen depletion (Figure S6). This is consistent with The lower net current density obtained in comparison to the previous studies that indicated that competition with oxygen LOx-based bioanode (62.2 ± 1.9 μAcm−2) demonstrated that was significant at low lactate concentrations and became the output of the EBFCs was limited by the biocathode. The

7112 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article

Figure 6. Photograph of the contact lens encapsulated EBFC (A) and testing setup (B). (C) Polarization and power curves for the EBFC consisting ﬁ of NPG/Os(bpy)2PVI/LOx bioanode and NPG-BOD biocathode. (D) Operational stability of the EBFC at 150 mV in arti cial tears.

Figure 7. Effects of the presence of 0.18 mM ascorbate toward bioanode (A) and biocathode (B). Inset of (A) shows the response on a bare NPG. same electrode registered a net current density of 57.3 ± 1.9 μA power density and current density decreased to 1.7 ± 0.1 μW −2 −2 ± μ −2 fi cm in an O2-bubbled solution (Figure S8), indicating a cm and 11.6 1.5 Acm when tested in the arti cial tear substrate-concentration-dependent catalytic behavior. The solution (Figure 6C). This likely arose from the interference of obtained current density compares well with a previous species such as ascorbate and the increased solution viscosity report,36 where the oxygen reduction response could be leading to mass transport resistance together with biofouling on enhanced by increasing the thickness of NPG up to 500 nm. the electrode surface caused by proteins. The antibiofouling 3.3. Performance of EBFC. The performance of the effect of NPG has been described in previous reports.62 assembled EBFC was tested in air-equilibrated PBS and Significant resistances to fouling caused by BSA and fibrinogen artificial tear solution containing 3 mM lactate, respectively. were observed, due to the nanoporous structure preventing − The maximum power density achieved was 2.4 ± 0.2 μWcm 2 entry of large proteins into the nanoporous network. at ca. 237 mV and a maximum short-circuit current density of The decrease in performance was mainly attributed to 15.1 ± 3.0 μWcm−2 in a PBS solution (Figure 6C). The interference by ascorbate, as it can be easily oxidized on the obtained open-circuit voltage (OCV) was 455 ± 21 mV in PBS, nanostructured gold electrode.17,30 The oxidation of ascorbate which is consistent with the difference in the onset potentials should be taken into account as it greatly perturbs the for lactate oxidation and oxygen reduction occurring at the biocathode performance, although it improves the current bioanode and biocathode, respectively. However, the maximum observed at the bioanode.31 Detailed investigation of ascorbate

7113 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article

Table 1. List of Properties of Tear-Based EBFCs

μ ‑2 anode cathode OCV (mV) Pmax ( Wcm ) stability ref. AuNPs/CtCDH, 0.05 AuNPs/BOx, air-saturated 570 1 more than 20 h operational half-life in 30 mM glucose human tears AuNPs/TTF-TCNQ, AuNPs/BOx, air-saturated 540 3.1 77% loss for the first 1 hour in human 17 0.665 mM ascorbate lachrymal tears BP/poly-MG/LDH/ An-pyr-MWCNT/TBAB-modified 410 ± 60 8.01 ± 1.4 80% loss for 4 hours in artificial tears 31 NAD+, 3 mM lactate Nafion/BOx, solution: n/a fi ± ± LOx/FcMe2-LPEI, 3 An-pyr-MWCNT/TBAB-modi ed 440 80 2.4 0.9 n/a 32 mM lactate Nafion/BOx, air-saturated ± ± ∼ NPG/Os(bpy)2PVI/ NPG-diazonium-BOx, air- PBS: 455 21, PBS: 2.4 0.2, 20% of initial power retained after 5.5 this LOx, 3 mM lactate equilibrated artificial tear: artificial tear: h operation in artificial tears work 380 ± 28 1.7 ± 0.1 aNote: CtCDH: Corynascus thermophilus cellobiose dehydrogenase; AuNPs: gold nanoparticles; TTF-TCNQ: tetrathiafulvalene-tetracyanoquinodi- fi methane; BP: buckypaper; poly-MG: polymerized methylene green; FcMe2-LPEI: dimethylferrocene-modi ed linear polyethylenimine; An-pyr: anthracene-pyrene; MWCNT: multiwalled carbon nanotube; TBAB: tetrabutylammonium bromide. interference was performed on both the anode and cathode in 4. CONCLUSIONS air-equilibrated PBS solutions (Figure 7). Bare NPG displayed Flexible NPG were successfully fabricated using an electro- a faradaic response in 0.18 mM ascorbate with a current density chemical dealloying method. The electrodes were modified of 7.86 μAcm−2 at 174 mV vs SCE (inset of Figure 7A). The − with lactate oxidase and bilirubin oxidase for use as a lactate/O2 catalytic response decreased to 3.15 μAcm2 on NPG/ biofuel cell, which was subsequently tested in a solution of Os(bpy)2PVI/LOx (Figure 7A), 5% of the catalytic current of artificial tears. The flexible EBFC holds potential as an − 62.2 ± 2.0 μAcm 2 to 3 mM lactate. This implies that the Os autonomous power supply for wearable electronic devices. polymer did not act as a mediator for the oxidation of Ascorbate interference, especially at the biocathode, was ascorbate, and the coating layer restricted the interference of responsible for the decrease in performance in tears in ascorbate to some extent. Figure 7B shows the effect of comparison to the performance in phosphate buffer solution. ascorbate upon the BOx cathode. A decrease of current density Acoatingfilm on the biocathode may alleviate such ff by 36% and a shift of the onset potential by 90 ± 5 mV, interference e ects. The response of the assembled EBFC consistent with a decrease in OCV from 455 ± 21 to 380 ± 28 was limited by current density and operational stability of the mV, were observed. Thus, we can conclude that ascorbate does biocathode. Improvements in the observed current density of not change the bioanode response greatly but diminishes the the biocathode will enable the development of a self-powered biocathode performance in terms of onset potential and current lactate biosensor on a contact lens, where the power density of the EBFC could be correlated with the concentration of lactate. response. Table 1 compares the performance of the proposed EBFC with previously reported tear-based EBFCs. The ■ ASSOCIATED CONTENT maximum power density obtained was of the same order of *S Supporting Information magnitude as other lactate/O EBFCs.31,32 The maximum 2 The Supporting Information is available free of charge on the power density is ca. two times higher than that of a glucose/O 2 ACS Publications website at DOI: 10.1021/acsami.7b18708. EBFC due to the low glucose concentration in tear fluid.30 The Supplementary figures, tables, and plots including OCV was ca. 100 mV less than that of a DET-based glucose/O2 30 electrodeposition of diazonium salt, scanning electron EBFC and similar to the values reported for lactate/O2 31,32 microscopy (SEM), resistance measurement of flexible EBFCs. The theoretical OCV for a lactate/O2 EBFC is 23 nanoporous gold, bioanode operational stability (PDF) 1.0 V, indicating that a higher OCV is possible. A mediator with a lower redox potential (that is still greater than that of FAD (ca. −0.43 V vs SCE)63) could be used to prepare a LOx- ■ AUTHOR INFORMATION modified bioanode with a low onset potential. As described Corresponding Author * earlier, interference from ascorbate results in a decrease of the E-mail: [email protected]; Fax: +353 61 213529; Tel: onset potential. Suppression of ascorbate interference, e.g., +353 61 234390. using a coating of Nafion,64 would also improve the OCV. ORCID A stable response is a prerequisite for an applicable EBFC. Edmond Magner: 0000-0003-2042-556X The operational stability of the EBFC was examined by placing Present Address § the NPG electrodes between two contact lenses (Scheme 1 and National Centre for Sensor Research, School of Chemical Figure 6B) with drop-by-drop supply of artificial tears Sciences, Dublin City University, Dublin 9, Ireland containing 3 mM lactate, etc. The EBFC survived over a Notes period of 5.5 h working at 150 mV (Figure 6D). The power The authors declare no competing financial interest. density observed decreased significantly by more than 50% during the initial 1 h period. The EBFC retained ca. 20% of the ■ ACKNOWLEDGMENTS original output after 5.5 h. The stability decay was mainly due This project has received funding from the European Union’s to the deterioration of the response of the BOx-based Seventh Framework Programme for research, technological biocathode,35 as the response of the Os-polymer-modified development, and demonstration under grant agreement no. bioanode was robust (Figure S5). This stability is acceptable for 607793. X.X. acknowledges a Government of Ireland use in 1-day disposable lenses. Postgraduate Scholarship (GOIPG/2014/659).

7114 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article ■ REFERENCES Immobilized in Dimethylferrocene-modified LPEI. Biosens. Bioelectron. 2016, 77,26−31. (1) Rasmussen, M.; Abdellaoui, S.; Minteer, S. D. Enzymatic Biofuel (22) Bandodkar, A. J.; Jia, W.; Wang, J. Tattoo-Based Wearable Cells: 30 Years of Critical Advancements. Biosens. Bioelectron. 2016, 76, − − Electrochemical Devices: A Review. Electroanalysis 2015, 27, 562 572. 91 102. (23) Pankratov, D.; Gonzalez-Arribas,́ E.; Blum, Z.; Shleev, S. Tear (2) Leech, D.; Kavanagh, P.; Schuhmann, W. Enzymatic Fuel Cells: − − Based Bioelectronics. Electroanalysis 2016, 28, 1250 1266. Recent Progress. Electrochim. Acta 2012, 84, 223 234. (24) Yao, H.; Shum, A. J.; Cowan, M.; Lahdesmä ki,̈ I.; Parviz, B. A. A (3) Shleev, S. Quo Vadis, Implanted Fuel Cell? ChemPlusChem 2017, Contact Lens with Embedded Sensor for Monitoring Tear Glucose 82, 522−539. Level. Biosens. Bioelectron. 2011, 26, 3290−3296. (4) Yahiro, A. T.; Lee, S. M.; Kimble, D. O. Bioelectrochemistry: I. (25) Yao, H.; Liao, Y.; Lingley, A. R.; Afanasiev, A.; Lahdesmä ki,̈ I.; Enzyme Utilizing Bio-fuel Cell Studies. Biochim. Biophys. Acta, Spec. Otis,B.P.;Parviz,B.A.AContactLenswithIntegrated Sect. Biophys. Subj. 1964, 88, 375−383. Telecommunication Circuit and Sensors for Wireless and Continuous (5) Drake, R. F.; Kusserow, B. K.; Messinger, S.; Matsuda, S. A Tear Glucose Monitoring. J. Micromech. Microeng. 2012, 22, 075007. Tissue Implantable Fuel Cell Power Supply. Trans. Am. Soc. Artif. Int. (26) Leonardi, M.; Pitchon, E. M.; Bertsch, A.; Renaud, P.; Organs 1970, 16, 199−205. Mermoud, A. Wireless Contact Lens Sensor for Intraocular Pressure (6) Cinquin, P.; Gondran, C.; Giroud, F.; Mazabrard, S.; Pellissier, ́ Monitoring: Assessment on Enucleated Pig Eyes. Acta Ophthalmol. A.; Boucher, F.; Alcaraz, J.-P.; Gorgy, K.; Lenouvel, F.; Mathe,S.A − Glucose Biofuel Cell Implanted in Rats. PLoS One 2010, 5, e10476. 2009, 87, 433 437. (7) Andoralov, V.; Falk, M.; Suyatin, D. B.; Granmo, M.; Sotres, J.; (27) Chu, M. X.; Miyajima, K.; Takahashi, D.; Arakawa, T.; Sano, K.; Ludwig, R.; Popov, V. O.; Schouenborg, J.; Blum, Z.; Shleev, S. Biofuel Sawada, S.-i.; Kudo, H.; Iwasaki, Y.; Akiyoshi, K.; Mochizuki, M.; Cell Based on Microscale Nanostructured Electrodes with Inductive Mitsubayashi, K. Soft Contact Lens Biosensor for in situ Monitoring of Tear Glucose as Non-invasive Blood Sugar Assessment. Talanta 2011, Coupling to Rat Brain Neurons. Sci. Rep. 2013, 3, 3270. − (8) Castorena-Gonzalez, J. A.; Foote, C.; MacVittie, K.; Halamek,́ J.; 83, 960 965. Halamková ,́ L.; Martinez-Lemus, L. A.; Katz, E. Biofuel Cell Operating (28) Lingley, A. R.; Ali, M.; Liao, Y.; Mirjalili, R.; Klonner, M.; in Vivo in Rat. Electroanalysis 2013, 25, 1579−1584. Sopanen, M.; Suihkonen, S.; Shen, T.; Otis, B. P.; Lipsanen, H.; Parviz, (9) Zebda, A.; Cosnier, S.; Alcaraz, J. P.; Holzinger, M.; Le Goff, A.; B. A. A Single-pixel Wireless Contact Lens Display. J. Micromech. Gondran, C.; Boucher, F.; Giroud, F.; Gorgy, K.; Lamraoui, H.; Microeng. 2011, 21, 125014. Cinquin, P. Single Glucose Biofuel Cells Implanted in Rats Power (29) Peng, C.-C.; Burke, M. T.; Carbia, B. E.; Plummer, C.; Chauhan, Electronic Devices. Sci. Rep. 2013, 3, 1516. A. Extended Drug Delivery by Contact Lenses for Glaucoma Therapy. − (10) Halamková ,́ L.; Halamek,́ J.; Bocharova, V.; Szczupak, A.; J. Controlled Release 2012, 162, 152 158. Alfonta, L.; Katz, E. Implanted Biofuel Cell Operating in a Living Snail. (30) Falk, M.; Andoralov, V.; Blum, Z.; Sotres, J.; Suyatin, D. B.; J. Am. Chem. Soc. 2012, 134, 5040−5043. Ruzgas, T.; Arnebrant, T.; Shleev, S. Biofuel Cell as a Power Source for − (11) MacVittie, K.; Halamek, J.; Halamkova, L.; Southcott, M.; Electronic Contact Lenses. Biosens. Bioelectron. 2012, 37,38 45. Jemison, W. D.; Lobel, R.; Katz, E. From ″Cyborg″ Lobsters to a (31) Reid, R. C.; Minteer, S. D.; Gale, B. K. Contact Lens Biofuel Pacemaker Powered by Implantable Biofuel Cells. Energy Environ. Sci. Cell Tested in a Synthetic Tear Solution. Biosens. Bioelectron. 2015, 68, − 2013, 6,81−86. 142 148. (12) Pankratov, D.; Ohlsson, L.; Gudmundsson, P.; Halak, S.; (32) Reid, R. C.; Jones, S. R.; Hickey, D. P.; Minteer, S. D.; Gale, B. Ljunggren, L.; Blum, Z.; Shleev, S. Ex vivo Electric Power Generation K. Modeling Carbon Nanotube Connectivity and Surface Activity in a − in Human Blood Using an Enzymatic Fuel Cell in a Vein Replica. RSC Contact Lens Biofuel Cell. Electrochim. Acta 2016, 203,30 40. Adv. 2016, 6, 70215−70220. (33) Ding, Y.; Kim, Y. J.; Erlebacher, J. Nanoporous Gold Leaf: “ ” (13) Cadet, M.; Gounel, S.; Stines-Chaumeil, C.; Brilland, X.; Ancient Technology /Advanced Material. Adv. Mater. 2004, 16, − Rouhana, J.; Louerat, F.; Mano, N. An Enzymatic Glucose/O2 Biofuel 1897 1900. Cell Operating in Human Blood. Biosens. Bioelectron. 2016, 83,60−67. (34) Xiao, X.; Si, P.; Magner, E. An Overview of Dealloyed (14) Falk, M.; Pankratov, D.; Lindh, L.; Arnebrant, T.; Shleev, S. Nanoporous Gold in Bioelectrochemistry. Bioelectrochemistry 2016, 109, 117−126. Miniature Direct Electron Transfer Based Enzymatic Fuel Cell ́ Operating in Human Sweat and Saliva. Fuel Cells 2014, 14, 1050− (35) Xiao, X.; Conghaile, P. O.; Leech, D.; Ludwig, R.; Magner, E. A 1056. Symmetric Supercapacitor/Biofuel Cell Hybrid Device Based on (15) Göbel, G.; Beltran, M. L.; Mundhenk, J.; Heinlein, T.; Enzyme-modified Nanoporous Gold: An Autonomous Pulse Gen- erator. Biosens. Bioelectron. 2017, 90,96−102. Schneider, J.; Lisdat, F. Operation of a Carbon Nanotube-based ́ Glucose/Oxygen Biofuel Cell in Human Body Liquids-Performance (36) Siepenkoetter, T.; Salaj-Kosla, U.; Xiao, X.; Conghaile, P. O.; Factors and Characteristics. Electrochim. Acta 2016, 218, 278−284. Pita, M.; Ludwig, R.; Magner, E. Immobilization of Redox Enzymes on (16) Jia, W.; Valdes-Ramírez,́ G.; Bandodkar, A. J.; Windmiller, J. R.; Nanoporous Gold Electrodes: Applications in Biofuel Cells. Wang, J. Epidermal Biofuel Cells: Energy Harvesting from Human ChemPlusChem 2017, 82, 553−560. Perspiration. Angew. Chem., Int. Ed. 2013, 52, 7233−7236. (37) Xiao, X.; Magner, E. A Biofuel Cell in Non-aqueous Solution. (17) Falk, M.; Andoralov, V.; Silow, M.; Toscano, M. D.; Shleev, S. Chem. Commun. 2015, 51, 13478−13480. Miniature Biofuel Cell as a Potential Power Source for Glucose- (38) Siepenkoetter, T.; Salaj-Kosla, U.; Xiao, X.; Belochapkine, S.; Sensing Contact Lenses. Anal. Chem. 2013, 85, 6342−6348. Magner, E. Nanoporous Gold Electrodes with Tuneable Pore Sizes for (18) Anastasova, S.; Crewther, B.; Bembnowicz, P.; Curto, V.; Ip, H. Bioelectrochemical Applications. Electroanalysis 2016, 28, 2415−2423. M. D.; Rosa, B.; Yang, G.-Z. A Wearable Multisensing Patch for (39) Siepenkoetter, T.; Salaj-Kosla, U.; Magner, E. The Immobiliza- Continuous Sweat Monitoring. Biosens. Bioelectron. 2017, 93, 139− tion of Fructose Dehydrogenase on Nanoporous Gold Electrodes for 145. the Detection of Fructose. ChemElectroChem 2017, 4, 905−912. (19) Berman, E. R. Biochemistry of the Eye; Springer Science & (40) Meng, F.; Ding, Y. Sub-micrometer-thick All-solid-state Business Media: New York, 1991. Supercapacitors with High Power and Energy Densities. Adv. Mater. (20) Sakharov, D. A.; Shkurnikov, M. U.; Vagin, M. Y.; Yashina, E. I.; 2011, 23, 4098−4102. Karyakin, A. A.; Tonevitsky, A. G. Relationship between Lactate (41) Xiao, X.; Ulstrup, J.; Li, H.; Wang, M.; Zhang, J.; Si, P. Concentrations in Active Muscle Sweat and Whole Blood. Bull. Exp. Nanoporous Gold Assembly of Glucose Oxidase for Electrochemical Biol. Med. 2010, 150,83−85. Biosensing. Electrochim. Acta 2014, 130, 559−567. (21) Hickey, D. P.; Reid, R. C.; Milton, R. D.; Minteer, S. D. A Self- (42) Pozuelo, J.; Compañ, V.; Gonzalez-Mé ijome,́ J. M.; Gonzalez,́ powered Amperometric Lactate Biosensor Based on Lactate Oxidase M.; Molla,́ S. Oxygen and Ionic Transport in Hydrogel and Silicone-

7115 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 ACS Applied Materials & Interfaces Research Article

Hydrogel Contact Lens Materials: An Experimental and Theoretical Nanostructured Porous Gold Electrodes in Biofouling Solutions. Anal. Study. J. Membr. Sci. 2014, 452,62−72. Chem. 2013, 85, 11610−11618. (43) Kober, E. M.; Caspar, J. V.; Sullivan, B. P.; Meyer, T. J. Synthetic (63) Rabaey, K.; Verstraete, W. Microbial fuel cells: novel Routes to New Polypyridyl Complexes of Osmium(Ii). Inorg. Chem. biotechnology for energy generation. Trends Biotechnol. 2005, 23, 1988, 27, 4587−4598. 291−298. (44) Forster, R. J.; Vos, J. G. Synthesis, Characterization, and (64) O’Riordan, S. L.; Mc Laughlin, K.; Lowry, J. P. In vitro Properties of a Series of Osmium- and Ruthenium-containing physiological performance factors of a catalase-based biosensor for Metallopolymers. Macromolecules 1990, 23, 4372−4377. real-time electrochemical detection of brain hydrogen peroxide in (45) Schneider, C. A.; Rasband, W. S.; Eliceiri, K. W. NIH Image to freely-moving animals. Anal. Methods 2016, 8, 7614−7622. Imagej: 25 Years of Image Analysis. Nat. Methods 2012, 9, 671. (46) Trasatti, S.; Petrii, O. A. Real Surface Area Measurements in Electrochemistry. Pure Appl. Chem. 1991, 63, 711−734. (47) Xiao, X.; Conghaile, P. Ó.; Leech, D.; Ludwig, R.; Magner, E. An Oxygen-independent and Membrane-less Glucose Biobattery/Super- capacitor Hybrid Device. Biosens. Bioelectron. 2017, 98, 421−427. (48) Xiao, X.; Li, H.; Wang, M.; Zhang, K.; Si, P. Examining the Effects of Self-Assembled Monolayers on Nanoporous Gold Based Amperometric Glucose Biosensors. Analyst 2014, 139, 488−494. (49) Scanlon, M. D.; Salaj-Kosla, U.; Belochapkine, S.; MacAodha, D.; Leech, D.; Ding, Y.; Magner, E. Characterization of Nanoporous Gold Electrodes for Bioelectrochemical Applications. Langmuir 2012, 28, 2251−2261. (50) Snyder, J.; Livi, K.; Erlebacher, J. Dealloying Silver/Gold Alloys in Neutral Silver Nitrate Solution: Porosity Evolution, Surface Composition, and Surface Oxides. J. Electrochem. Soc. 2008, 155, C464−C473. (51) Zhang, Q.; Wang, X.; Qi, Z.; Wang, Y.; Zhang, Z. A Benign Route to Fabricate Nanoporous Gold through Electrochemical Dealloying of Al−Au Alloys in a Neutral Solution. Electrochim. Acta 2009, 54, 6190−6198. (52) Shao, K.; Fang, C.; Yao, Y.; Zhao, C.; Yang, Z.; Liu, J.; Zou, Z. An Easily Modified Method Using FeCl3 to Synthesize Nanoporous Gold with a High Surface Area. RSC Adv. 2017, 7, 18327−18332. (53) Delahay, P.; Pourbaix, M.; Rysselberghe, P. V. Potential-pH Diagram of Silver Construction of the Diagram−Its Applications to the Study of the Properties of the Metal, its Compounds, and its Corrosion. J. Electrochem. Soc. 1951, 98,65−67. (54) Burke, L. D.; Nugent, P. F. The Electrochemistry of Gold: I The Redox Behaviour of the Metal in Aqueous Media. Gold Bull. 1997, 30, 43−53. (55) Matharu, Z.; Daggumati, P.; Wang, L.; Dorofeeva, T. S.; Li, Z.; Seker, E. Nanoporous-Gold-Based Electrode Morphology Libraries for Investigating Structure−Property Relationships in Nucleic Acid Based Electrochemical Biosensors. ACS Appl. Mater. Interfaces 2017, 9, 12959−12966. (56) Meng, F.; Yan, X.; Liu, J.; Gu, J.; Zou, Z. Nanoporous Gold as Non-enzymatic Sensor for Hydrogen Peroxide. Electrochim. Acta 2011, 56, 4657−4662. (57) Zhang, L.; Chang, H.; Hirata, A.; Wu, H.; Xue, Q.-K.; Chen, M. Nanoporous Gold Based Optical Sensor for Sub-ppt Detection of Mercury Ions. ACS Nano 2013, 7, 4595−4600. (58) Ohara, T. J.; Rajagopalan, R.; Heller, A. ″Wired″ Enzyme Electrodes for Amperometric Determination of Glucose or Lactate in the Presence of Interfering Substances. Anal. Chem. 1994, 66, 2451− 2457. (59) Park, T.-M.; Iwuoha, E. I.; Smyth, M. R.; Freaney, R.; McShane, A. J. Sol-gel Based Amperometric Biosensor Incorporating an Osmium Redox Polymer as Mediator for Detection Of L-Lactate. Talanta 1997, 44, 973−978. (60) Gao, Z.; Binyamin, G.; Kim, H. H.; Barton, S. C.; Zhang, Y.; Heller, A. Electrodeposition of Redox Polymers and Co-electrodeposition of Enzymes by Coordinative Crosslinking. Angew. Chem., Int. Ed. 2002, 41, 810−813. (61) Milton, R. D.; Lim, K.; Hickey, D. P.; Minteer, S. D. Employing FAD-dependent Glucose Dehydrogenase within a Glucose/Oxygen Enzymatic Fuel Cell Operating in Human Serum. Bioelectrochemistry 2015, 106 (Part A), 56−63. (62) Patel, J.; Radhakrishnan, L.; Zhao, B.; Uppalapati, B.; Daniels, R. C.; Ward, K. R.; Collinson, M. M. Electrochemical Properties of

7116 DOI: 10.1021/acsami.7b18708 ACS Appl. Mater. Interfaces 2018, 10, 7107−7116 Journal of Electroanalytical Chemistry 812 (2018) 180–185

Contents lists available at ScienceDirect

Journal of Electroanalytical Chemistry

journal homepage: www.elsevier.com/locate/jelechem

A continuous ﬂuidic bioreactor utilising electrodeposited silica for lipase ☆ T immobilisation onto nanoporous gold ⁎ Xinxin Xiaoa, Till Siepenkoettera, Robert Whelanb, Urszula Salaj-Koslaa, Edmond Magnera, a Department of Chemical Sciences and Bernal Institute, University of Limerick, Limerick, Ireland b School of Design, University of Limerick, Limerick, Ireland

ARTICLE INFO ABSTRACT

Keywords: An electrochemically triggered sol-gel process was used to generate a thin silica layer for the immobilisation of Lipase lipase from Thermomyces lanuginosus onto dealloyed nanoporous gold (NPG). The catalytic response of the en- Electrodeposition trapped lipase was examined using the hydrolysis of 4-nitrophenyl butyrate (4-NPB) as a model reaction. For the Nanoporous gold electrodeposition process, parameters including the deposition time and the concentration of lipase affected the Enzyme immobilisation − observed catalytic activity. A deposition time of 180 s and a lipase concentration of 3 mg mL 1 were used to Flow cell prepare the optimised electrode. The operational stability of the silica immobilised enzyme was enhanced on Microfluidic enzymatic reactor NPG in comparison to that on planar gold, which may arise from confinement of the enzyme in the porous structure. The modified electrodes were incorporated into a 3D printed flow cell with conversion efficiencies of up to 100% after 8 cycles.

1. Introduction electrodes [11], have been prepared. Nanoporous gold (NPG), fabricated by etching gold alloys, has been used as a support material for the Immobilised enzymes [1] have been successfully used in applica- immobilisation of enzymes [18]. NPG possesses characteristic pores and tions such as biocatalysis [2,3], biosensors [4] and biofuel cells [5–7]. ligaments whose sizes can be tailored by adjusting the dealloying Silicate materials including controlled pore glass (CPG), sol-gel derived conditions [19]. NPG has been modified with electrodeposited thin silicate and mesoporous silicate (MPS) are biocompatible and widely films of thiophene polymers [20] and Os modified polymers [5,21] for used as solid supports for the immobilisation of enzymes [8]. Sol-gel the immobilisation of enzymes. In contrast to Os polymer modified derived silicate materials possess features that include ease of pre- electrodes that are prepared by drop-casting, electrodeposited films are paration, chemical inertness, negligible swelling and optical transpar- more physically stable [22]. Electrodeposition enables control of the ency [9]. In addition, the sol-gel process enables enzymes to be im- degree of modification of the entire surface [18] and the thickness of mobilised without significant losses in activity [10]. Electrodeposition the deposited film can be tuned to optimise the enzyme loading [5]. provides a controllable and rapid route to grow uniform silica layers Immobilisation makes it possible for enzymes to be used in a con- onto a conductive substrate regardless of the roughness of the surface tinuous flow mode [23]. In comparison to the conventional batch ap- [11]. The process of electrodeposition is initiated by an increase in pH proach, flow methods are more efficient due to the large surface-to- in the proximity of the cathode as a consequence of hydrogen evolution, volume ratio, ease of collection and on-line analysis of products [24]. which consumes protons. Hydrolysis and condensation of precursors Micro-reactors can be generally subdivided into three types: packed- such as tetraethoxysilane (TEOS) are triggered by this change in pH bed, monolith and wall-coated reactors [25]. The former two can suffer [12], resulting in three-dimensional Si-O-Si networks that encapsulate from possible blockage and pressure gradients along the microchannels enzyme in solution on the electrode/electrolyte interface (Scheme 1). [26], making it difficult to tune flow dynamics. Inversely, wall-coated A wide range of enzymes including glucose oxidases [13] and de- channels enable smoother flow with negligible mass transport re- hydrogenases [14,15] etc. have been successfully immobilised in elec- sistance and thus predictable fluidic conditions. However, the catalyst trochemically generated silica matrices for use as biosensors. Various loading of a wall-coated reactor is inferior to those of packed-bed and electrochemically derived sol-gel silica/nanomaterial hybrids, such as monolithic approaches [25], arising from the low working surface area. carbon nanotubes [16], gold nanoparticles [17], and macroporous gold Porous supports can be used to improve the catalyst loading [27].

☆ Dedicated to Professor Renata Bilewicz on the occasion of her 65th birthday. ⁎ Corresponding author at: Department of Chemical Sciences, University of Limerick, Limerick, Ireland. E-mail address: [email protected] (E. Magner). https://doi.org/10.1016/j.jelechem.2017.11.059 Received 26 September 2017; Received in revised form 15 November 2017; Accepted 22 November 2017 Available online 23 November 2017 1572-6657/ © 2017 Published by Elsevier B.V. X. Xiao et al. Journal of Electroanalytical Chemistry 812 (2018) 180–185

Scheme 1. Schematic drawing of the electrodeposition of silica for enzyme immobilisation at a constant negative potential.

Scheme 2. (A) CAD drawing of the flow cell consisting of top-plate and base. (B) Sectional view of the base (unit: mm). (C) Detailed view of part 3 from (B). (D) Photograph of the flow cell during operation; the arrows indicate the direction of flow.

Scheme 3. Hydrolysis reaction of 4-NPB catalysed by lipase.

NPG electrodes with an average pore size of ca. 30 nm [5,28] were (Scheme 3). The catalytic performance was affected by the thickness of functionalised by the electrodeposition of silica with simultaneous en- the silicate layer and by the concentration of lipase during electro- capsulation of lipase [29] (Scheme 1). The bio-modified electrodes were deposition. The conversion of substrate to product depended on the placed into a bespoke flow channel device (Scheme 2). The hydrolysis flow rate and full conversion was feasible upon recycling the solution. of 4-nitrophenyl butyrate (4-NPB) by lipase was used as a model system

181 X. Xiao et al. Journal of Electroanalytical Chemistry 812 (2018) 180–185

2. Experimental section 2.3. Flow cell design

2.1. Materials and apparatus Devices were designed using SolidWorks 3D CAD software (Dassault Systèmes, 2017). The flow cell consisted of two parts: a top-plate and a Potassium phosphate monobasic (≥99%) and dibasic (≥98%), base (Scheme 2A). The top-plate was mounted above the base and fixed hydrochloric acid (HCl, 37%), nitric acid (HNO3, 70%), sulfuric acid with 10 screws (Scheme 2D). Four NPG modified GCEs were inserted (H2SO4, 97%), 4-nitrophenyl butyrate (4-NPB, ≥98%), Bradford re- into the vertical holes of the device, sitting flush on top of a 500 μm agent, bovine serum albumin (BSA, lyophilized powder, ≥96%) were high channel (Scheme 2B). As shown in Scheme 2C, O-rings between obtained from Sigma-Aldrich Ireland, Ltd. Lipase (Sigma-Aldrich the top-plate and base mitigated against any leakage of solution. A − Ireland, Ltd) from Thermomyces lanuginosus (≥100 kU g 1, EC no.: secondary function is that the action of tightening the top-plate onto the 3.1.1.3) was supplied as a solution containing 73% (w/v) water, 25% O-rings also caused the O-rings to squeeze the GCEs, effectively se- (w/v) propylene glycol, 2% (w/v) lipase, and 0.5% calcium chloride. 4- curing them in place on the top of the channel and preventing move- (2-Hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, ≥99%) and ment. An Instech P720 peristaltic pump was used to pump the solution tetraethoxysilane (TEOS, 99.9%) were purchased from Fisher Scientific, into the channel (Scheme 2D) via round push-fit adaptors. The transi- Ireland. All solutions were prepared with deionised water (resistivity of tion from a round adaptor to the laminar flow channel in the base was 18.2 MΩ cm) from an Elgastat maxima-HPLC (Elga Purelab Ultra, UK). internally rounded to mitigate turbulence (Scheme 2C). NPG leaves were fabricated by etching ca. 100 nm thick Au/Ag leaf A Stratasys Objet Connex 500 3D printer featuring PolyJet Matrix™ alloy films (12-carat, Eytzinger, Germany) in concentrated HNO3 for Technology was used to print the flow cell parts with a laminar flow 30 min at 30 °C. The NPG films were then attached onto a pre-polished channel of a depth of 500 μm. Printing was performed by successively glassy carbon electrode (GCE) with a diameter of 4 mm and were al- laying down an acrylate based material (VeroClear RGD810) in 16 μm lowed to dry. Before future utilisation, the NPG electrodes were acti- thick layers followed by UV curing and hardening of each layer. vated by scanning the potential over the range of −0.2 to 1.65 V in 1 M Exogenous material in cavities and overhangs was removed using a jet −1 H2SO4 at a scan rate of 100 mV s for 15 cycles. wash system. VeroClear RGD810 is a transparent material, that facil- Electrochemical experiments were performed with a CHI802 po- itates removal of the support material from the flow channel by al- tentiostat (CH Instruments, Austin, Texas) consisting of a NPG working lowing visual inspection of the jet washing step. This material also al- electrode, a Pt wire counter electrode and a Ag/AgCl reference elec- lows the visual detection of air bubbles forming inside the channel trode. Samples mounted on 300-mesh copper grids (S147-3, Agar during operation. Scientific, UK) were characterised by transmission electron microscopy (TEM, JEOL JEM-2100, operating voltage of 200 kV). The average pore 2.4. Determination of immobilised protein concentration size and deposition layer thickness were measured with ImageJ software (National Institutes of Health, Bethesda, Maryland) [30]. Absor- The amount of immobilised lipase into the silica film onto the bance was recorded on a Cary 60 UV–Vis spectrophotometer (Agilent, electrode was determined by the Bradford assay [31] of the initial and USA). residual enzyme content in the electrodeposition solution. A range of − concentrations of BSA from 0.15 to 1 mg mL 1 in water were used to obtain the standard curve. The absorbances (595 nm) of mixtures of 2.2. Enzyme immobilisation procedures and activity measurement solutions of protein (50 μL) and Bradford reagent (1.5 mL) were measured. A typical silica sol was prepared by dissolving 2.125 g tetraethoxysilane (TEOS) with 2 mL of deionised water and 2.5 mL of 3. Results and discussion 0.01 M aqueous HCl, which were mixed for 12 h using a magnetic stirrer and then diluted 3 times with water for further use. 900 μLof Application of a potential of −1.1 V vs. Ag/AgCl initiates the sol-gel lipase solution in 0.1 M pH 7.0 phosphate buffer solution (PBS) was process (Scheme 1) via the production of hydroxyl ions that trigger the mixed with 100 μL of the above hydrolysed sol. A range of lipase condensation of TEOS. The deposition time was optimised using a so- − concentrations was used. As shown in Scheme 1, sol-gel electro-assisted lution containing 1 mg mL 1 lipase. The effects of deposition time on deposition was performed at an applied potential of −1.1 V vs. Ag/ the catalytic response are summarised in Fig. 1A. In agreement with AgCl for different durations. The same route was applied to deposit previous studies [5,20], the response increased with time for deposition silica/lipase onto planar gold electrodes (diameter: 3 mm) for com- times < 180 s, and can be ascribed to the increased amount of enzyme parison. Regeneration of the modified NPG electrodes was achieved by immobilised. For longer deposition times, the response decreased which polishing and cleaning the glassy carbon electrode, followed by at- likely arise from limitations in the supply of 4-NPB supply to lipase. tachment of a new NPG film and subsequent deposition of silica. TEM images indicate that the thicker films , e.g. obtained for 360 s All experiments were performed at room temperature (20 ± 2 °C). deposition, block the pores (Fig. 2C) [5]. A deposition time of 180 s was The activity of the immobilised lipase was evaluated by the hydrolysis therefore utilised as a compromise between higher enzyme loading and of 4-NPB (Scheme 3). Flow experiments were performed by pumping mass-transport through the film. Similar phenomena were reported for buffer solution (10 mM pH 7.0 HEPES) containing 75 μM 4-NPB at the immobilisation of D-sorbitol dehydrogenase [16] and of hae- various flow rates (Scheme 2D). The system was washed (at a relatively moglobin [11] in electro-deposited silica layers. − fast flow rate of 0.12 mL min 1 for 10 min) to allow the detachment of TEM confirmed the formation of a silica film on NPG (Fig. S1 and loosely-bounded enzymes prior to measurements. The absorbance of Fig. 2). NPG (dark region of the micrographs) preserved its porous the product was measured at a wavelength of 420 nm. The stability of structure after electrodeposition with gold ligands growing thicker (Fig. an electrode soaked in a solution of 4-NPB (2 mL, 75 μM) at room S2). The relatively bright outer layer along the pores can be dis- temperature was checked once a day. Absorbance changes at 420 nm tinguished as the silica film with a uniform thickness. The optimal were recorded by immersing the electrode into a fresh solution (2 mL, deposition time of 180 s resulted in a layer thickness of ca. 75 μM 4-NPB) for 1 h. To achieve 100% conversion of 4-NPB, the ab- 9.3 ± 1.1 nm (in contrast to ca. 30 nm pores). A 60-s deposition time sorbance of the effluent solution was determined prior to recycling into resulted in a 2.8 ± 0.6 nm thick film (Fig. 2A), while the pores were the channel. filled after a deposition time of 360 s (Fig. 2C). The effect of lipase concentration was optimised using a deposition time of 180 s. The catalytic response increased at lipase concentrations

182 X. Xiao et al. Journal of Electroanalytical Chemistry 812 (2018) 180–185

− Fig. 1. (A) The eﬀect of deposition time on the catalytic performance of NPG/silica/lipase obtained in 1 mg mL 1 lipase (measured by immersion in 2 mL of 75 μM 4-NPB for 0.5 h). (B) The eﬀect of lipase concentration on the catalytic performance of NPG/silica/lipase obtained by depositing for 180 s (measured by immersion in 2 mL of 75 μM 4-NPB for 1 h).

− Fig. 2. TEM images showing the NPG/silica/lipase obtained in 1 mg mL 1 lipase with various deposition durations: (A): 60 s, (B): 180 s, (C): 360 s; arrows in (A) and (B) indicate the silica/lipase coating layers, while pores are ﬁlled with silica/lipase in (C); scale bars at the right-bottom of (A), (B) and (C) indicate 30 nm.

− below 3 mg mL 1 (Fig. 1B), at which point the response leveled off. Similar results were reported for the sol-gel deposition of D-sorbitol dehydrogenase on macroporous gold electrodes [14]. Silica/lipase was also deposited on a planar gold electrode using the − optimised protocol (3 mg mL 1 lipase, 180 s deposition time). Using the Bradford assay, the enzyme loading on planar gold was determined − − to be 1.8 nmol cm 2, in comparison to 3.6 nmol cm 2 on NPG. Despite a roughness factor of ca. 8 [28], NPG only encapsulated double the amount of lipase than that on the planar gold. This is likely due to the accelerated sol-gel process arising from the increased rate of hydrogen evolution that is observed on NPG in comparison to planar gold [32]. Similarly, Mazurenko et al. found that sol-gel derived bioelectrode performance was sensitive to the quantity of carbon nanotubes (CNTs) on the GCE as a high quantity of CNTs led to facilitated sol-gel process and faster silica film deposited [16]. When the stability of modified NPG and planar gold electrodes was examined (Fig. 3), NPG modified electrodes retained 49.3% of the original response after 5 days' storage, in comparison to 20.2% on planar gold. The observed decrease in activity likely arises from loss of enzyme activity and/or leakage of the enzyme. Lipase can operate as a catalyst in aqueous and nonaqueous Fig. 3. Storage stability of NPG/silica/lipase; response was measured by immersion in solutions in a stable manner [2]. For example, lipase that was physically 2mLof75μM 4-NPB for 1 h. adsorbed on NPG showed negligible activity loss after 10 successive

183 X. Xiao et al. Journal of Electroanalytical Chemistry 812 (2018) 180–185

Fig. 4. The eﬀect of ﬂow rate at the catalytic behavior of NPG/silica/lipase towards Fig. 6. Regeneration of a NPG/silica/lipase electrode. The response was measured by 75 μM 4-NPB. immersion of the electrode in 2 mL of 75 μM 4-NPB for 1 h.

(1 mL) was mixed with 1 mL of 75 μM 4-NPB and incubated for 0.5 h. An absorbance change of 0.0054 (at 420 nm) was observed corresponding to leaching of 0.36% of the enzyme. This result is consistent with the data in Fig. 3 that showed a decrease in activity with time. This leaching likely arises from removal of the silica from the electrode. The NPG was physically attached on to the GCE and could be removed. Freshly prepared NPG leaves can be re-attached onto the GCE with newly generated silica layers to immobilise lipase. Fig. 6 shows the normalised activity of 10 regenerated NPG/silica/lipase electrodes. A RSD of 4.6% demonstrated the reproducibility of the method.

4. Conclusions

In this study, direct incorporation of lipase into sol-gel derived silica onto NPG was proposed. The effects of electrodeposition time and lipase concentration in the electrolyte on the catalytic response were systematically evaluated. The porous structure of NPG had a remarkable, positive influence in enhancing operational stability, although Fig. 5. Conversion ratio of 2 mL of 75 μM 4-NPB by cycling in a loop at a flow rate of only a two-fold increase for the initial catalytic response over that on a − 0.05 mL min 1. planar Au electrode supported silica/lipase. The laminar flow cell allowed the study of the effect of flow rate. Operated in a loop mode at a fl −1 uses, and maintained 74% of its initial activity after 20 cycles [33]. ow rate of 0.05 mL min enabled the complete hydrolysis of 4-NPB μ Lipase immobilised onto SBA-15 retained 95% activity after 7 cycles (2 mL of 75 M solution) after eight cycles. The underlying study in the [2]. It is thus likely that the stability arises from leakage of the enzyme, group is to integrate NPG supported enzymatic biofuel cells (EBFCs) fl especially given that sol-gel derived silica is generally brittle and can with the bespoke ow cell. become cracked during long-term manipulation and use. The curved surface of NPG could provide a more stable environment for the film. Acknowledgments Similar examples of enhanced stability on NPG have been observed. For This project has received funding from the European Union's example, electrodeposited MnO2 on NPG preserved 64% of its capacitance after 50 cycles of charge-discharge, in comparison to 26% for Seventh Framework Programme for research, technological development and demonstration under grant agreement no 607793. X. Xiao planar Au/MnO2 [21]. The dependence of the catalytic response of a single NPG/silica/ acknowledges a Government of Ireland Postgraduate Scholarship lipase electrode on flow rates was examined over the range (GOIPG/2014/659). − 0.01–0.12 mL min 1 (Fig. 4). The catalytic response decreased with increasing flow rate due to the decreased residence time, in agreement Appendix A. Supplementary data with previous studies of lipase immobilised in flow reactors [34,35]. Four NPG/silica/lipase electrodes were mounted in the flow cell Supplementary data to this article can be found online at https:// (Scheme 2D). A linear increase in the amount of substrate conversion doi.org/10.1016/j.jelechem.2017.11.059. was observed in the first four cycles with the level of conversion leveling off in subsequent cycles. As can be seen from Fig. 5, the reaction References was completed after eight cycles. fl [1] U. Hanefeld, L. Gardossi, E. Magner, Understanding enzyme immobilisation, Chem. To examine enzyme leaching under ow conditions, a control ex- – fl Soc. Rev. 38 (2) (2009) 453 468. periment was performed by ushing the channel at a rate of [2] N.H. Abdallah, M. Schlumpberger, D.A. Gaffney, J.P. Hanrahan, J.M. Tobin, − 0.05 mL min 1 with blank PBS for 20 min. The solution collected E. Magner, Comparison of mesoporous silicate supports for the immobilisation and

184 X. Xiao et al. Journal of Electroanalytical Chemistry 812 (2018) 180–185

activity of cytochrome c and lipase, J. Mol. Catal. B Enzym. 108 (2014) 82–88. nanoporous gold/poly(3,4-ethylenedioxythiophene) hybrid electrodes for ampero- [3] R. DiCosimo, J. McAuliffe, A.J. Poulose, G. Bohlmann, Industrial use of immobilized metric glucose sensing, Talanta 116 (15) (2013) 1054–1059. enzymes, Chem. Soc. Rev. 42 (15) (2013) 6437–6474. [21] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, An oxygen-independent [4] A. Heller, B. Feldman, Electrochemical glucose sensors and their applications in and membrane-less glucose biobattery/supercapacitor hybrid device, Biosens. diabetes management, Chem. Rev. 108 (7) (2008) 2482–2505. Bioelectron. 98 (2017) 421–427. [5] X. Xiao, P.Ó. Conghaile, D. Leech, R. Ludwig, E. Magner, A symmetric super- [22] Z. Gao, G. Binyamin, H.H. Kim, S.C. Barton, Y. Zhang, A. Heller, Electrodeposition capacitor/biofuel cell hybrid device based on enzyme-modified nanoporous gold: of redox polymers and co-electrodeposition of enzymes by coordinative cross- an autonomous pulse generator, Biosens. Bioelectron. 90 (2017) 96–102. linking, Angew. Chem. Int. Ed. 41 (5) (2002) 810–813. [6] D. Leech, P. Kavanagh, W. Schuhmann, Enzymatic fuel cells: recent progress, [23] S. Kundu, A.S. Bhangale, W.E. Wallace, K.M. Flynn, C.M. Guttman, R.A. Gross, Electrochim. Acta 84 (0) (2012) 223–234. K.L. Beers, Continuous flow enzyme-catalyzed polymerization in a microreactor, J. [7] S. Calabrese Barton, J. Gallaway, P. Atanassov, Enzymatic biofuel cells for im- Am. Chem. Soc. 133 (15) (2011) 6006–6011. plantable and microscale devices, Chem. Rev. 104 (10) (2004) 4867–4886. [24] Y. Asano, S. Togashi, H. Tsudome, S. Murakami, Microreactor technology: in- [8] E. Magner, Immobilisation of enzymes on mesoporous silicate materials, Chem. Soc. novations in production processes, Pharm. Eng. 30 (1) (2010) 32. Rev. 42 (15) (2013) 6213–6222. [25] R. Munirathinam, J. Huskens, W. Verboom, Supported catalysis in continuous-flow [9] W. Jin, J.D. Brennan, Properties and applications of proteins encapsulated within microreactors, Adv. Synth. Catal. 357 (6) (2015) 1093–1123. sol-gel derived materials, Anal. Chim. Acta 461 (1) (2002) 1–36. [26] K. Szymańska, M. Pietrowska, J. Kocurek, K. Maresz, A. Koreniuk, J. Mrowiec- [10] R. Gupta, N.K. Chaudhury, Entrapment of biomolecules in sol-gel matrix for ap- Białoń, P. Widłak, E. Magner, A. Jarzębski, Low back-pressure hierarchically plications in biosensors: problems and future prospects, Biosens. Bioelectron. 22 structured multichannel microfluidic bioreactors for rapid protein digestion – proof (11) (2007) 2387–2399. of concept, Chem. Eng. J. 287 (2016) 148–154. [11] F. Qu, R. Nasraoui, M. Etienne, Y.B.S. Côme, A. Kuhn, J. Lenz, J. Gajdzik, [27] E.V. Rebrov, A. Berenguer-Murcia, H.E. Skelton, B.F.G. Johnson, A.E.H. Wheatley, R. Hempelmann, A. Walcarius, Electrogeneration of ultra-thin silica films for the J.C. Schouten, Capillary microreactors wall-coated with mesoporous titania thin functionalization of macroporous electrodes, Electrochem. Commun. 13 (2) (2011) film catalyst supports, Lab Chip 9 (4) (2009) 503–506. 138–142. [28] X. Xiao, J. Ulstrup, H. Li, J. Zhang, P. Si, Nanoporous gold assembly of glucose [12] R. Shacham, D. Avnir, D. Mandler, Electrodeposition of methylated sol-gel films on oxidase for electrochemical biosensing, Electrochim. Acta 130 (2014) 559–567. conducting surfaces, Adv. Mater. 11 (5) (1999) 384–388. [29] I. Itabaiana, L.S. de Mariz e Miranda, R.O.M.A. de Souza, Towards a continuous [13] W.-Z. Jia, K. Wang, Z.-J. Zhu, H.-T. Song, X.-H. Xia, One-step immobilization of flow environment for lipase-catalyzed reactions, J. Mol. Catal. B Enzym. 85 glucose oxidase in a silica matrix on a Pt electrode by an electrochemically induced (2013) 1–9. sol-gel process, Langmuir 23 (23) (2007) 11896–11900. [30] C.A. Schneider, W.S. Rasband, K.W. Eliceiri, J. Schindelin, I. Arganda-Carreras, [14] Z. Wang, M. Etienne, G.-W. Kohring, Y. Bon-Saint-Côme, A. Kuhn, A. Walcarius, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, NIH image to imageJ: Electrochemically assisted deposition of sol–gel bio-composite with co-immobilized 25 years of image analysis, Nat. Methods 9 (7) (2012) 671. dehydrogenase and diaphorase, Electrochim. Acta 56 (25) (2011) 9032–9040. [31] M.M. Bradford, A rapid and sensitive method for the quantitation of microgram [15] Z. Wang, M. Etienne, F. Quilès, G.-W. Kohring, A. Walcarius, Durable cofactor quantities of protein utilizing the principle of protein-dye binding, Anal. Biochem. immobilization in sol–gel bio-composite thin films for reagentless biosensors and 72 (1) (1976) 248–254. bioreactors using dehydrogenases, Biosens. Bioelectron. 32 (1) (2012) 111–117. [32] X. Xiao, C. Engelbrekt, Z. Li, P. Si, Hydrogen evolution at nanoporous gold/tungsten [16] I. Mazurenko, M. Etienne, O. Tananaiko, V. Zaitsev, A. Walcarius, sulfide composite film and its optimization, Electrochim. Acta 173 (2015) 393–398. Electrophoretically deposited carbon nanotubes as a novel support for electro- [33] X. Wang, X.Y. Liu, X.L. Yan, P. Zhao, Y. Ding, P. Xu, Enzyme-nanoporous gold generated silica–dehydrogenase bioelectrodes, Electrochim. Acta 83 (2012) biocomposite: excellent biocatalyst with improved biocatalytic performance and 359–366. stability, PLoS ONE 6 (9) (2011) e24207. [17] R. Toledano, D. Mandler, Electrochemical codeposition of thin gold nanoparticles/ [34] I. Itabaiana, F.K. Sutili, S.G.F. Leite, K.M. Goncalves, Y. Cordeiro, I.C.R. Leal, sol-gel nanocomposite films, Chem. Mater. 22 (13) (2010) 3943–3951. L.S.M. Miranda, M. Ojeda, R. Luque, R.O.M.A. de Souza, Continuous flow valor- [18] X. Xiao, P. Si, E. Magner, An overview of dealloyed nanoporous gold in bioelec- ization of fatty acid waste using silica-immobilized lipases, Green Chem. 15 (2) trochemistry, Bioelectrochemistry 109 (2016) 117–126. (2013) 518–524. [19] T. Siepenkoetter, U. Salaj-Kosla, X. Xiao, S. Belochapkine, E. Magner, Nanoporous [35] I. Denčić, S. de Vaan, T. Noël, J. Meuldijk, M. de Croon, V. Hessel, Lipase-based gold electrodes with tuneable pore sizes for bioelectrochemical applications, biocatalytic flow process in a packed-bed microreactor, Ind. Eng. Chem. Res. 52 Electroanalysis 28 (2016) 2415–2423. (32) (2013) 10951–10960. [20] X. Xiao, M.e. Wang, H. Li, P. Si, One-step fabrication of bio-functionalized

185

Due to copyright restrictions the full text of this article is not available in the electronic version of this thesis

Chemical Communications 2018, 54, 5823 A quasi-solid-state and self-powered biosupercapacitor based on flexible nanoporous gold electrodes Xinxin Xiao and Edmond Magner http://dx.doi.org/10.1039/c8cc02555j http://hdl.handle.net/10344/6817