This journal is©The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10, 1581--1589 | 1581

Cite this: Energy Environ. Sci.,

2017, 10, 1581

Soft, stretchable, high power density electronicskin-based biofuel cells for scavenging energyfrom human sweat†

Amay J. Bandodkar, ‡a Jung-Min You,‡a Nam-Heon Kim,‡a Yue Gu,‡a

Rajan Kumar,a A. M. Vinu Mohan,a Jonas Kurniawan, a Somayeh Imani,b

Tatsuo Nakagawa,a Brianna Parish,a Mukunth Parthasarathy,a Patrick P. Mercier,*b

Sheng Xu *a and Joseph Wang *a

This article describes the fabrication, characterization, and real-life

application of a soft, stretchable electronic-skin-based biofuel cell

(E-BFC) that exhibits an open circuit voltage of 0.5 V and a power

density of nearly 1.2 mW cm�2 at 0.2 V, representing the highest

power density recorded by a wearable biofuel cell to date. High

power density is achieved via a unique combination of lithographically-

patterned stretchable electronic framework together with screen-

printed, densely-packed three-dimensional carbon-nanotube-based

bioanode and cathode array arranged in a stretchable ‘‘island-bridge’’

configuration. The E-BFC maintains its performance even under

repeated strains of 50%, and is stable for two days. When applied

directly to the skin of human subjects, the E-BFC generates B1 mW

during exercise. The E-BFC is able to power conventional electronic

devices, such as a light emitting diode and a Bluetooth Low Energy

(BLE) radio. This is the first example of powering a BLE radio by a

wearable biofuel cell. Successful generation of high power density

under practical conditions and powering of conventional energy-

intense electronic devices represents a major step forward in the field

of soft, stretchable, wearable energy harvesting devices.

Wearable electronics has garnered enormous academic andindustrial interest owing to their immense biomedical,1–3

computing,4 entertainment,5 defense,6 and environment7–9 relatedapplications. Many examples of wearable devices that areembedded in textiles10,11 or directly mounted on the humanskin12 have been demonstrated in recent years. Ideally, wear-able electronics should be thin, compact, soft, and stretchable,in order to enable conformal integration with the human skinwithout causing a somatosensory backlash. Researchers havebeen successful in realizing such wearable devices owing to

advances in new manufacturing techniques,13–15 materialsinnovation,16,17 and nanotechnology.18 However, a key chal-lenge to enable adoption in practical use-cases is the lackof correspondingly thin wearable energy sources. Most priorefforts either rely on the integration of bulky batteries,19,20

which severely compromise wearability. Efforts have been madetowards developing thin, stretchable batteries21,22 and super-capacitors23,24 to address this issue. However, such systemshave limited energy storage capacity and require frequentrecharging. An alternate approach is to wirelessly power thewearable device, typically via near-field communication (NFC)chipsets;25–27 yet, this approach requires a large, proximalpower source that tethers the subject to a fixed location (albeitwirelessly), preventing use by freely-behaving subjects. A promisingalternative approach is to develop a wearable energy harvesterthat scavenges energy from body motion,28–31 the sun,32,33 bodyheat,34,35 and biofluids.36–43 Of these, utilizing biofluids, especiallyhuman sweat, for generation of electricity by wearable biofuel cellsis an appealing avenue since these systems rely on innocuousbiomolecules for energy conversion. Our group demonstrated thedevelopment of wearable biofuel cells that convert readily availablesweat lactate into electricity.44,45 However, progress in this field hasbeen limited due to low power density and lack of soft, stretchablenature for conformal skin contact of the wearable biofuel systems.

Recent efforts have attempted to address the issue of skinconformality by developing stretchable biofuel cells.46–48 However,these rely on random composite-based stretchable systems wherethe underlying electrodes are stretchable, but the overlaying activelayers are non-stretchable. The mismatch between the mechanicalproperties can strain the active layers, leading to gradual delami-nation and leaching, thus decreasing the performance. While workrelated to device conformity has been initiated, not much has beenaccomplished in improving the power density of wearable biofuelcells – one of the most crucial factors for the success of the field.Rapid generation of high levels of power by wearable biofuel cellsis essential considering the transient nature of lactate concen-tration in human sweat,49 where sweat lactate concentration

a Department of NanoEngineering, University of California, San Diego, La Jolla,

CA 92093-0448, USA. E-mail: josephwang@eng.ucsd.edu, shengxu@eng.ucsd.edub Department of Electrical & Computer Engineering, University of California,

San Diego, La Jolla, CA 92093-0407, USA. E-mail: pmercier@ucsd.edu

† Electronic supplementary information (ESI) available: Experimental details andvideos. See DOI: 10.1039/c7ee00865a‡ Equal contribution.

Received 29th March 2017,Accepted 15th June 2017

DOI: 10.1039/c7ee00865a

rsc.li/ees

Energy &EnvironmentalScience

COMMUNICATION

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article OnlineView Journal | View Issue

1582 | Energy Environ. Sci., 2017, 10, 1581--1589 This journal is©The Royal Society of Chemistry 2017

increases during the initial perspiration phase and then dropsdue to increased sweat rate.50,51 The low power density of prior-artwearable biofuel cells limits their ability to power useful electronicloads in practice. For example, most modern wearable devicesrequire far-field wireless transmission of sensed information to asmartphone or smartwatch. However, commercially available radiotransmitters that can accomplish this (e.g., Bluetooth) consumepower on the order of mW. Most prior-art wearable BFCs have onlybeen demonstrated to trivial power loads with current draws muchless than 100 mA.36,38,45 While some prior-art BFCs have beendemonstrated to power custom52 or commercially availableradios,53,54 these examples were optimized to operate either asan implant within the human body, or in fruit, both of whichoffer higher fuel concentrations that are not available for awearable device. Thus, increasing power density in wearableBFCs is essential, especially when considering the transientnature of lactate concentration in human sweat.

The present work addresses both the above-mentioned issues, bydeveloping a soft, deterministically stretchable electronic skin-type

biofuel system (E-BFC) that can generate a high-power density of4B1 mW cm�2 from scavenging lactate present in human sweat(Fig. 1). This represents nearly 10 times higher power density thanthe previously reported wearable biofuel cells.45 Table S1 in the ESI†shows the comparison of the present work with previously reportedwearable biofuel cells. The E-BFC sustains its high power underrepeated strains (e = 50%) and is able to harness the energy in thesweat to power an energy intense Bluetooth Low Energy (BLE) radioand a light emitting diode (LED).

The power generated by the E-BFC depends on its activesurface area. The E-BFC will thus generate the highest powerif it comprises of a single continuous anode and cathode.However, metallic films are rigid in nature and easily fractureunder a strain of B5–10%, less than what is commonlyencountered by wearable devices. Hence, the anode and cathodeof the E-BFC were designed to have an island-bridge architecturewith the central areas of the islands firmly bonded to the under-lying stretchable substrate while the serpentine bridges free tounwind and deform under stress. The firmly bonded islands

Fig. 1 Images and schematics of E-BFC. Actual photograph of (A) bare and (B) carbon-coated Au island-bridge architecture. (C) Zoomed in image ofcarbon-coated Au island-bridge architecture. Photographs of (D) complete and (E) a close-up section of the E-BFC. (F) Pictorial representation ofexploded view of the E-BFC, illustrating its different layers (Au, gold; Ca, carbon and Ac, active bio-anodic/cathodic components). (G) Imagesdemonstrating the complete and zoomed view of an E-BFC adorned by a human subject along with its potential for various wearable applications.(A, B–E and G) Scale bar, 5 mm.

Communication Energy & Environmental Science

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online

This journal is©The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10, 1581--1589 | 1583

form the active electrodes of the E-BFC that experience negligiblestrain, owing to the stretchable nature of the interconnectingserpentine bridges.

The Au bridge-island architecture was designed to possess ahexagonal close packed structure with anode and cathodeforming interdigitated electrode systems (Fig. 1F). The hexagonalclose-packed structure enables the system to acquire high fillfactor (B60%) while the interdigitated electrode pattern allowsthe anode and cathode to be closely located for reducing internalresistance. The islands possess a circular geometry to mitigatenon-uniform strain generation at the electrodes.

In the interdigitated island-bridge E-BFC design, only theislands forming the anode and cathode are electrically connectedvia the serpentine interconnects. However, such a design can leadto non-uniform stress generation around each island. To distributethe stress uniformly, each circular anodic (cathodic) electrode isradially connected to adjacent cathodic (anodic) electrodes vianon-conducting polyimide serpentine bridges and by conductivebridges to adjoining anodic (cathodic) electrodes (Fig. 1A). Such adesign enables the stress to be uniformly distributed around thecircular electrodes while preventing the anode and cathode frombeing short-circuited. The distance between adjoining islandswas maintained at 1 mm to ensure sufficient stretchability andconvenience of screen printing.

The functionalization steps of the Au islands with active BFCmaterial were designed to realize a system that could generatehigh levels of power at physiologically relevant lactate concen-trations. With this aim, the anode islands were functionalizedwith densely packed carbon nanotube-naphthoquinone (CNT-NQ)3D structures while the cathode comprised of compact 3D carbonnanotube-silver oxide (CNT-Ag2O) structures. The 3D porousnature of these structures enables high loadings of activecomponents, which in turn lead to manifold enhancement inpower generated by the BFC. It is critical that the active compo-nents and CNTs are firmly held in close contact to mitigateleaching and the concomitant decrease in power output. Thiswas achieved by dispersing the CNTs and the active materials inwater insoluble polymeric binders that hold the pellet compo-nents together while allowing facile transport of fuel through its3D structure. In the case of anodic pellets, a highly cross-linkedchitosan biopolymer was employed as a binder while the cathodicpellet was firmly held by the water-insoluble Nafion polymer.To firmly bond the 3D anodic and cathodic pellets to the under-lying Au islands, a layer of conductive carbon ink was screenprinted onto the Au islands (Fig. 1B and C) followed by affixing thepellets to the carbon-coated Au islands using conductive carbonglue (Fig. 1D and E). Such a hybrid printing-lithography fabrica-tion process resulted in strongly bonded 3D anodic and cathodicpellets arranged in an interdigitated format (Fig. 1F). The asfabricated soft, stretchable E-BFC permits its easy integrationwith the curvilinear epidermis for real-time generation of energyfrom human sweat to power electronics (Fig. 1G). Rapid catalyticoxidation of fuel is essential for high-performance BFCs. In thisregard, the enzyme must be able to easily interact with the fuel.To achieve this, the lactate oxidase (LOx) enzyme was drop-castedon top of the CNT-NQ pellet. Such a functionalization step

ensured minimal use of LOx enzyme to attain rapid and maxi-mum interaction of the enzyme with sweat lactate to generatehigh electricity levels. However, in the case of the cathode, theAg2O particles were well dispersed within the 3D CNT framework,since cathodic reduction reaction is independent of sweatconstituents; close contact between the Ag2O and CNT facili-tates electron transfer (Fig. 2A). Details of the device fabricationare discussed in the ESI.†

The ability of the E-BFC to generate power by scavenginglactate as fuel was tested by exposing it to buffer solution spikedwith various physiologically relevant lactate concentrations. Forthis study, a unit cell comprising a single anode and cathode wasemployed, as marked in Fig. 2A. In the presence of lactate,spontaneous oxidation of lactate at the anode and reduction ofAg2O to Ag at cathode occur as illustrated in Fig. 2A (scheme).SEM images of both anode (Fig. 2B) and cathode (Fig. 2C)pellets were captured to observe the dispersion of differentconstituents within the pellets. The uniform distribution ofCNTs within the anode and cathode pellets is obvious from theSEM images. Such homogeneity is crucial for efficient electronicpathways between the active components of the anode/cathodeand the underlying carbon current collector electrode. The SEMimages (Fig. S1, ESI†) also reveal a highly porous morphology thatprovides a large surface area, thus enabling high loading of theenzyme. The granular structures observed in the SEM image ofthe cathodic pellet can be attributed to Ag2O. EDX was furtheremployed to obtain better information on the distribution of Ag2Owithin the cathodic pellet. The color-coded EDX images, taken forcarbon (Fig. 2D) and silver (Fig. 2E), illustrate the homogeneity ofthe Ag2O distribution within the CNT matrix. The correspondingSEM image is shown in Fig. S1B (ESI†).

Fig. 2F illustrates a typical power density versus voltagecurve for a unit cell of E-BFC. As expected, the power densityincreases in a nearly linear fashion with increasing lactateconcentration. It is important to note the high power densitygenerated by the BFC. This represents the highest powerdensity reported to date for a lactate biofuel cell and nearly10 times higher than previously reported wearable lactateBFCs.44,45 It should be noted that, although the net surfacearea of the present E-BFC is higher than previous studies,44,45

the power density in the present and early studies is calculatedper unit active electrode area, providing a reliable parameter tocompare various BFCs. The amount of enzyme loaded in thepresent system is similar to previously reported BFCs, indicatingthat the E-BFC owes its high performance to the 3D pellet featuresand the use of a Ag2O based cathode. The 3D structure of pellets(Fig. S1, ESI†) allows high loading of the mediator in the anodicpellet and Ag2O in the cathodic pellet. The 3D porous structurealso enables much higher surface area where the immobilizedLOx enzyme can interact with the lactic acid fuel and mediator togenerate high currents. The collective effect of these features ofthe E-BFC enables it to generate significantly higher power densitythan previously reported wearable lactic acid based BFCs. Therehave been a handful of studies reporting high power densityBFCs.55–57 However, these relied on glucose as a fuel and requiredmuch higher fuel concentrations to generate power density

Energy & Environmental Science Communication

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online

1584 | Energy Environ. Sci., 2017, 10, 1581--1589 This journal is©The Royal Society of Chemistry 2017

comparable to the present E-BFC and are not compatible forinterface with the soft, curvilinear human epidermis. TheE-BFC’s ability to generate high levels of power, even at lowerfuel concentrations, can be attributed to the synergistic effect ofimmobilization of LOx directly on top of the 3D CNT-NQ pillarand the use of Ag2O based cathode. The enzyme immobiliza-tion procedure enables efficient fuel-enzyme interaction forfacile bioelectrocatalysis. On the other hand, the Ag2O basedcathode is shown to undergo faster catalytic reduction ascompared to sluggish oxygen reduction reaction occurring atnoble metal or enzyme-based cathodes.46,58 It was noted thatthe power generated by the E-BFC saturates for concentrationsabove 20 mM. This can be attributed to the low Michaelis–Menten constant for the LOx enzyme. However, this is not amajor concern since the average concentration of lactic acidin human sweat is around 14 mM.59 One could extend thesaturation limit by coating the anode with diffusion membranes,but this can lead to lower power density for lactic acid concentrations

usually present in human sweat. The main aim of an energyharvester is to extract maximum energy from commonlyencountered settings, and hence priority has been given hereto maximizing the power output of the E-BFC rather thanextending the saturation limit. The ability of E-BFC to generatestable power for extended periods is a critical parameter thatone needs to analyze when considering it for practical applications.In this regard, an E-BFC was exposed to 14 mM lactic acid solutionand the power output was recorded for two days (Fig. 2G). As clearlyillustrated in the figure, the E-BFC generates stable power with themaximum power density decreasing by only B15% at 0.2 V. Thestudy reveals that the E-BFC can produce stable, high levels ofpower for long durations. A slight increase in the open-circuitvoltage of the E-BFC is observed while recording the power curvesafter 24 h and 48 h. This could be attributed to the swelling of thechitosan biopolymer at the anode that leads to enhanced inter-action between lactate, enzyme and the mediator. The stability ofthe E-BFC ultimately will depend on stability of the enzyme and

Fig. 2 In vitro characterization of E-BFC. (A) Optical image showing the close-up view of anode and cathode of the E-BFC along with a schematicillustrating the different components of each electrode and the corresponding electrochemical reactions occurring at them. SEM images of (B) anodeand (C) cathode. EDX images for (D) carbon and (E) silver captured for cathode. Power density curves recorded for (F) increasing lactate concentrationand following (G) 48 h when the E-BFC is periodically exposed to 14 mM lactate. Scale bar; (A) 2 mm, (B and C) 1 mm and (D and E) 50 mm.

Communication Energy & Environmental Science

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online

This journal is©The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10, 1581--1589 | 1585

availability of Ag2O. Major efforts have been devoted to improve theenzyme stability at ambient conditions.40 Such routes can augmentthe stability of anode. Preliminary studies on loading of Ag2Owithin the cathodic pellets revealed no significant improvementin power output when the loading was increased beyond 70%.Thus, rather than increasing the loading of Ag2O within each pellet,increasing the areal density of cathodic pellets (with 70% Ag2O) byrelying on innovative device architectures will increase the totalamount of Ag2O available for cathodic reduction. Combininganodes with highly stabilized enzyme and higher density ofcathodic pellets while increasing the number of anodes and

cathodes will therefore help extend the life-span of the E-BFCfor extended use. Stretchability is a crucial requirement forwearable devices. Therefore, the ability of the E-BFC to endureextreme strains was studied in in-vitro settings. The E-BFC wasexposed to 14 mM lactic acid solution and power curves wererecorded after every 20 stretching iterations of 50% strain.As illustrated in Fig. 3A, the ability of the E-BFC to generatepower is negligibly hampered even after being subjected tosuch extreme strain cycles. It must be noted that the humanarm skin can endure a maximum strain of B27%.60 Therefore,the present E-BFC system is capable of performing desirably

Fig. 3 Mechanical resiliency studies and effect of fuel coverage on the E-BFC. (A) Power density versus voltage curves for E-BFC after repeatedstretching iterations (e = 50%). Image of a blue LED powered by E-BFC when subjected to (B) e = 0% and (C) e = 50% strain. (D) Effect of repeated e = 50%strains on the short-circuit current and open-circuit voltage. Images showing the effect of strain on the (E and F) short circuit current and (G and H)open-circuit voltage. (I) Effect of increasing the areal coverage of E-BFC by fuel on the power output. (J) Image of the E-BFC illustrating the extent ofspatial coverage by fuel during study conducted in (I). (A–J) 14 mM lactic acid was used as fuel. (B, C and E–H) Scale bar, 5 mm.

Energy & Environmental Science Communication

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online

1586 | Energy Environ. Sci., 2017, 10, 1581--1589 This journal is©The Royal Society of Chemistry 2017

under the circumstances commonly experienced by the epidermis.Subsequently, the ability of the E-BFC to power electronics undersimilar fatigue cycle was studied by integrating a custom-madeDC-DC converter to power a LED. Circuit diagram for the DC-DCconverter is shown in Fig. S2A (ESI†). As clearly demonstrated inVideo S1 (ESI†) and Fig. 3B, C, the LED continues to light evenwhen the E-BFC is stretched repeatedly by 50%. In addition,a separate study was also conducted to measure the effect ofstrain (e = 50%) on the short-circuit current and open circuitvoltage of the E-BFC (Fig. 3D–H). The data plotted in Fig. 3D,Videos S2, S3 (ESI†) and the images of Fig. 3E–H, clearlydemonstrate the negligible impact of such repeated strains onthe E-BFC performance. The initiation and the flow of sweat overthe epidermis are non-uniform. Thus, at a given moment theE-BFC will be non-uniformly covered with human sweat. The netpower generated by the E-BFC depends on the lactate-fuelconcentration as well as the spatial coverage of the E-BFC bythe sweat. Therefore, the effect of fractional coverage of E-BFC bythe lactate solution on the generated power was studied. A fixedconcentration of 14 mM lactate was utilized for this study sincethe median concentration of lactate in human sweat is aroundthis value.59 As illustrated in Fig. 3I, the net power generated bythe E-BFC increases linearly as the coverage increases from 20%to 100% (Fig. 3J). It is important to note that even with 20% arealcoverage the E-BFC is capable of generating B400 mW in thepresence of 14 mM lactate solution, due to the compact 3Dstructure of the CNT-based anodes and cathodes. Also, the netpower generated by the E-BFC will increase linearly with increasingthe number of anodes and cathodes present in the device. Thus,the power output can be modulated by varying the number of unitBFCs in the device.

The ability of the E-BFC to power a BLE device under in-vitroconditions was also demonstrated. Circuit diagrams for theDC-DC converter and the BLE device is shown in Fig. S2 (ESI†).As shown in Fig. 4, the BFC was able to charge the 2.2 mFcapacitor to 3.5 V in B53 s, indicating an average generatedpower of approximately 0.5 mW during this experiment. Due tothe large current draw of the BLE System-on-Chip (SoC) during

its initialization routine, the capacitor output voltage temporarilydropped until the initialization routine completed. After initialization,the BFC was demonstrated to deliver sufficient power to powerthe BLE SoC at a stabilized voltage of 3.5 V for B10 min. Thedrop in the voltage can be attributed to consumption of thefuel. Without the E-BFC input power, the 2.2 mF decouplingcapacitor would only be able to power the load for less than 45 sbefore its voltage dropped to below 2 V, thereby indicating thatthe E-BFC is sustaining the load throughout the experiment.This represents the first example of a wearable BFC capable ofpowering a BLE device.

Ultimately, the E-BFC system was tested in real-life scenariosby applying it to consenting human subjects while they per-spired on a stationary bike. During these tests, the E-BFC waseither applied to the upper arm or on upper trapezius muscle.The distribution of sweat around the E-BFC modulates theelectric field density between the anode and cathode arrays.At the initial stage of sweating, the amount of sweat is insuffi-cient to provide a uniform conductive path between all anodesand cathodes of the E-BFC. Thus, the electric field produced bythe half-cell reaction occurring at the anode or cathode is non-uniformly distributed, resulting in areas of high electric fieldsnear the interface between the electrodes and the underlyingskin. Such high electric field coupled with the high resistanceof the stratum corneum could lead to resistive loss and maylead to skin irritation. In order to mitigate this issue, the E-BFCwas coated with a thin layer of hydrogel which provided auniform conductive path between anode and cathode arrayswhile allowing facile sweat transport for rapid power generation.Fig. 5A and B represent real-time power extracted from the E-BFCwhen applied to perspiring human subjects. The profiles followedin these studies correlate well with earlier studies. With increasingstress level, higher levels of lactate are released by the body in thesweat. This leads to continuously increasing power generated bythe skin-worn E-BFC. However, after a certain period, the well-known dilution of sweat lactate occurs,49 leading to decrease ofthe power generated by the E-BFC. Although fresh E-BFCs wereused for every on-body tests, the stability study, as discussed in

Fig. 4 Powering of a BLE device by E-BFC. (A) Photograph illustrating data transmission from an E-BFC powered BLE to a laptop. (B) Plot showingvoltage variations of the energy-buffering capacitor during the experiment. (1) Charging of the capacitor; (2) connecting the BLE microcontroller.

Communication Energy & Environmental Science

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online

This journal is©The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10, 1581--1589 | 1587

Fig. 2G, and studies performed in our earlier work,44 indicate thatthe same E-BFC can be used multiple times.

The ability of the E-BFC to power common electronics wasalso demonstrated. Almost all commercially available electronicdevices require an operating voltage that is much higher than thevoltage at which BFCs produce the highest power.61 Thus, in orderto exploit BFCs to power commercially available electronics, acustom DC-DC converter circuit board was coupled to the E-BFC.This setup was mounted on a human subject and the sweatgenerated by the subject was harvested to power a blue LED(Fig. 5C–E). Initially, the LED remained off but started blinkingwithin two mins after the subject started sweating. The LEDcontinued to blink for B4 min (Video S4, ESI†) and then thefrequency of blinking gradually decreased. This could be due tothe consumption of lactate present in the sweat by the E-BFC anddecrease in lactate concentration with increased sweat rate.50 Theperformance of the E-BFC was several orders better than that ofpreviously reported wearable BFC, which could power a LEDfor only a few seconds.45 When the LED stopped blinking, the

E-BFC was disconnected while the subject continued cycling,leading to the generation of fresh lactate. After a couple ofminutes, the E-BFC was reconnected and the LED started blinkingagain. Such a demonstration reveals the ability of the E-BFC as apotential power source for wearable applications.

Conclusions

The present work demonstrates the first example of a determi-nistically stretchable, high power density wearable electronicskin-based biofuel cell for harvesting usable energy fromhuman sweat. The E-BFC can generate B1.2 mW cm�2 ofpower in the presence of physiologically relevant concentrationsof lactate. This represents the highest power density reported for alactate biofuel cell. The wearable E-BFC’s impressive performancecan be attributed to its reliance on 3D anodes and cathodes, theenzyme immobilization method, and use of Ag2O as the activecathode material. In vitro characterization of the E-BFC systemrevealed that it can maintain its power output for long durations

Fig. 5 On-body generation of power by E-BFC. (A and B) Real-time scavenging of electrical energy by E-BFC applied to two different human subjects.(C) Image of a perspiring human subject adorning an E-BFC, interfaced with a DC–DC converter, to power an LED by harvesting lactate present in sweatas a fuel. Close-up image of the LED in (D) ‘off’ state in absence of sweat and (E) ‘on’ state in presence of lactate in sweat.

Energy & Environmental Science Communication

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online

1588 | Energy Environ. Sci., 2017, 10, 1581--1589 This journal is©The Royal Society of Chemistry 2017

and when under repeated stretching iterations. The negligibleimpact on its performance when subjected to fatigue cyclesreflects its island-bridge architecture which enables the systemto endure repeated stretching up to biaxial e = 50%, sufficient forall wearable applications. The high power generated by the E-BFCwas leveraged to power a BLE device. This is the first example of awearable biofuel cell powered BLE radio. Ultimately, the E-BFCwas reduced to practice by mounting it on human subjects togenerate electrical energy in real-time. Thereafter, a skin-wornE-BFC was also interfaced with a DC–DC converter to power aLED with sweat lactate as a fuel. Although the present workdemonstrates practicality of the BFCs as a wearable power source,additional efforts must be undertaken to harness the truepotential of such systems. For example, although Ag2O basedcathodes are superior to commonly used platinum-based ones,they are light sensitive and can degrade over time. One could lookinto other active materials, such as manganese oxide, to addressthis issue. Similarly, integrating micro-fluidics with the E-BFC willsupply fresh sweat to the fuel cell while directing away used, oldsweat and help improve its performance. At the same time, effortsmust be made to develop soft, stretchable architecture for theDC–DC converter electronics based on recent advances in wearableelectronics.62 Additionally, suitable routes to integrate E-BFC withother forms of energy harvesters, such as, solar cells, tribo-electricand thermos-electric generators must be developed so that energycan be scavenged from multiple sources. Strategies involvingmesoporous electrode architectures for further enhancing thepower output, as demonstrated recently,63,64 should also be consi-dered. Additional efforts can be made to further improve the poweroutput of the E-BFC by developing multi-fuel biofuel cells andcomplete oxidation of fuel by involving cascading enzymaticreactions. Further research will focus on these issues as well asperforming detailed E-BFC stability studies and exploring the effectof pellet thickness and porosity and enzyme loading on the poweroutput. Nevertheless, the attractive performance of the new epidermalBFC, particularly its ability to maintain very high power outputsunder repeated mechanical strains, represents a major step forwardin the field of soft, stretchable, wearable energy harvesting devices.

Acknowledgements

This work was supported by the Defense Threat Reduction AgencyJoint Science and Technology Office for Chemical and BiologicalDefense (HDTRA 1-16-1-0013). A. J. B. and R. K. acknowledgesupport from the Siebel Scholars Foundation and NationalScience Foundation Graduate Research Fellowship (Award DGE-1144086), respectively. P. P. M. and S. I. acknowledge the supportof the Arnold and Mabel Beckman Foundation. Any opinion,findings, and conclusions or recommendations expressed in thismaterial are those of the author(s) and do not necessarily reflectthe views of the funding agencies.

Notes and references

1 K. I. Jang, S. Y. Han, S. Xu, K. E. Mathewson, Y. Zhang, J. W.Jeong, G. T. Kim, R. C. Webb, J. W. Lee, T. J. Dawidczyk,

R. H. Kim, Y. M. Song, W. H. Yeo, S. Kim, H. Y. Cheng, S. IlRhee, J. Chung, B. Kim, H. U. Chung, D. Lee, Y. Yang, M. Cho,J. G. Gaspar, R. Carbonari, M. Fabiani, G. Gratton, Y. Huang andJ. A. Rogers, Nat. Commun., 2014, 5, 5779–5788.

2 J. W. Zhong, Y. Zhang, Q. Z. Zhong, Q. Y. Hu, B. Hu,Z. L. Wang and J. Zhou, ACS Nano, 2014, 8, 6273–6280.

3 X. W. Wang, Y. Gu, Z. P. Xiong, Z. Cui and T. Zhang, Adv.Mater., 2014, 26, 1336–1342.

4 N. Al-Naffakh, N. Clarke, P. Dowland and F. Li, InternetTechnology and Secured Transactions (ICITST), 2016 11thInternational Conference for, IEEE, 2016, pp. 189–195.

5 J. Kim, A. Banks, H. Y. Cheng, Z. Q. Xie, S. Xu, K. I. Jang,J. W. Lee, Z. J. Liu, P. Gutruf, X. Huang, P. H. Wei, F. Liu,K. Li, M. Dalal, R. Ghaffari, X. Feng, Y. G. Huang, S. Gupta,U. Paik and J. A. Rogers, Small, 2015, 11, 906–912.

6 A. J. Bandodkar, A. M. O’Mahony, J. Ramirez, I. A. Samek,S. M. Anderson, J. R. Windmiller and J. Wang, Analyst, 2013,138, 5288–5295.

7 V. Misra, A. Bozkurt, B. Calhoun, T. N. Jackson, J. S. Jur,J. Lach, B. Lee, J. Muth, O. Oralkan, M. Ozturk, S. Trolier-McKinstry, D. Vashaee, D. Wentzloff and Y. Zhu, Proc. IEEE,2015, 103, 665–681.

8 Y. J. Yun, W. G. Hong, N. J. Choi, B. H. Kim, Y. Jun andH. K. Lee, Sci. Rep., 2015, 5, 10904.

9 H. T. Li, X. Y. Mu, Y. N. Yang and A. J. Mason, IEEE Sens. J.,2014, 14, 3391–3399.

10 J. Lee, H. Kwon, J. Seo, S. Shin, J. H. Koo, C. Pang, S. Son,J. H. Kim, Y. H. Jang, D. E. Kim and T. Lee, Adv. Mater., 2015,27, 2433–2439.

11 Y. N. Meng, Y. Zhao, C. G. Hu, H. H. Cheng, Y. Hu, Z. P. Zhang,G. Q. Shi and L. T. Qu, Adv. Mater., 2013, 25, 2326–2331.

12 W. H. Yeo, Y. S. Kim, J. Lee, A. Ameen, L. K. Shi, M. Li,S. D. Wang, R. Ma, S. H. Jin, Z. Kang, Y. G. Huang andJ. A. Rogers, Adv. Mater., 2013, 25, 2773–2778.

13 B. C. K. Tee, A. Chortos, R. R. Dunn, G. Schwartz, E. Easonand Z. A. Bao, Adv. Funct. Mater., 2014, 24, 5427–5434.

14 T. Yokota, P. Zalar, M. Kaltenbrunner, H. Jinno,N. Matsuhisa, H. Kitanosako, Y. Tachibana, W. Yukita,M. Koizumi and T. Someya, Sci. Adv., 2016, 2, e1501856.

15 A. Mohan, N. Kim, Y. Gu, A. J. Bandodkar, J. M. You,R. Kumar, J. F. Kurniawan, S. Xu and J. Wang, Adv. Mater.Technol., 2017, 2, 1600284.

16 J. Wang, X. H. Li, Y. L. Zi, S. H. Wang, Z. L. Li, L. Zheng, F. Yi,S. M. Li and Z. L. Wang, Adv. Mater., 2015, 27, 4830–4836.

17 Y. Ma, M. Pharr, L. Wang, J. Kim, Y. Liu, Y. Xue, R. Ning,X. Wang, H. U. Chung, X. Feng, J. A. Rogers and Y. Huang,Small, 2017, 13, 1602954.

18 D. H. Zhang, M. H. Miao, H. T. Niu and Z. X. Wei, ACS Nano,2014, 8, 4571–4579.

19 B. Schazmann, D. Morris, C. Slater, S. Beirne, C. Fay,R. Reuveny, N. Moyna and D. Diamond, Anal. Methods,2010, 2, 342–348.

20 A. J. Bandodkar, D. Molinnus, O. Mirza, T. Guinovart,J. R. Windmiller, G. Valdes-Ramirez, F. J. Andrade,M. J. Schoning and J. Wang, Biosens. Bioelectron., 2014, 54,603–609.

Communication Energy & Environmental Science

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online

This journal is©The Royal Society of Chemistry 2017 Energy Environ. Sci., 2017, 10, 1581--1589 | 1589

21 R. Kumar, J. Shin, L. Yin, J. M. You, Y. S. Meng and J. Wang,Adv. Energy Mater., 2017, 7, 1602096.

22 M. Koo, K. I. Park, S. H. Lee, M. Suh, D. Y. Jeon, J. W. Choi,K. Kang and K. J. Lee, Nano Lett., 2012, 12, 4810–4816.

23 B. H. Xie, C. Yang, Z. X. Zhang, P. C. Zou, Z. Y. Lin, G. Q. Shi,Q. H. Yang, F. Y. Kang and C. P. Wong, ACS Nano, 2015, 9,5636–5645.

24 C. J. Yu, C. Masarapu, J. P. Rong, B. Q. Wei and H. Q. Jiang,Adv. Mater., 2009, 21, 4793–4797.

25 J. Kim, A. Banks, H. Y. Cheng, Z. Q. Xie, S. Xu, K. I. Jang,J. W. Lee, Z. J. Liu, P. Gutruf, X. Huang, P. H. Wei, F. Liu,K. Li, M. Dalal, R. Ghaffari, X. Feng, Y. G. Huang, S. Gupta,U. Paik and J. A. Rogers, Small, 2015, 11, 906–912.

26 J. Kim, P. Gutruf, A. M. Chiarelli, S. Y. Heo, K. Cho, Z. Q. Xie,A. Banks, S. Han, K. I. Jang, J. W. Lee, K. T. Lee, X. Feng,Y. G. Huang, M. Fabiani, G. Gratton, U. Paik andJ. A. Rogers, Adv. Funct. Mater., 2017, 27, 1604373.

27 D. P. Rose, M. E. Ratterman, D. K. Griffin, L. L. Hou,N. Kelley-Loughnane, R. R. Naik, J. A. Hagen, I. Papautskyand J. C. Heikenfeld, IEEE Trans. Biomed. Eng., 2015, 62,1457–1465.

28 J. Wang, S. M. Li, F. Yi, Y. L. Zi, J. Lin, X. F. Wang, Y. L. Xuand Z. L. Wang, Nat. Commun., 2016, 7, 12744–12751.

29 X. L. Cheng, B. Meng, X. S. Zhang, M. D. Han, Z. M. Su andH. X. Zhang, Nano Energy, 2015, 12, 19–25.

30 R. Yang, Y. Qin, C. Li, G. Zhu and Z. L. Wang, Nano Lett.,2009, 9, 1201–1205.

31 Z. Li, G. A. Zhu, R. S. Yang, A. C. Wang and Z. L. Wang, Adv.Mater., 2010, 22, 2534–2537.

32 B. J. Kim, D. H. Kim, Y. Y. Lee, H. W. Shin, G. S. Han,J. S. Hong, K. Mahmood, T. K. Ahn, Y. C. Joo, K. S. Hong,N. G. Park, S. Lee and H. S. Jung, Energy Environ. Sci., 2015,8, 916–921.

33 Y. H. Lee, J. S. Kim, J. Noh, I. Lee, H. J. Kim, S. Choi, J. Seo,S. Jeon, T. S. Kim, J. Y. Lee and J. W. Choi, Nano Lett., 2013,13, 5753–5761.

34 F. Suarez, A. Nozariasbmarz, D. Vashaee and M. C. Ozturk,Energy Environ. Sci., 2016, 9, 2099–2113.

35 S. J. Kim, J. H. We and B. J. Cho, Energy Environ. Sci., 2014, 7,1959–1965.

36 P. Cinquin, C. Gondran, F. Giroud, S. Mazabrard, A. Pellissier,F. Boucher, J. P. Alcaraz, K. Gorgy, F. Lenouvel, S. Mathe,P. Porcu and S. Cosnier, PLoS One, 2010, 5, e10479–10482.

37 S. Cosnier, A. Le Goff and M. Holzinger, Electrochem. Com-mun., 2014, 38, 19–23.

38 K. MacVittie, J. Halamek, L. Halamkova, M. Southcott,W. D. Jemison, R. Lobeld and E. Katz, Energy Environ. Sci.,2013, 6, 81–86.

39 A. Szczupak, J. Halamek, L. Halamkova, V. Bocharova,L. Alfonta and E. Katz, Energy Environ. Sci., 2012, 5, 8891–8895.

40 A. J. Bandodkar, J. Electrochem. Soc., 2017, 164, H3007–H3014.41 A. J. Bandodkar and J. Wang, Energy Environ. Sci., 2016, 28,

1188–1200.

42 E. Katz and K. MacVittie, Energy Environ. Sci., 2013, 6,2791–2803.

43 S. Cosnier, A. Le Goff and M. Holzinger, Electrochem. Commun.,2014, 38, 19–23.

44 W. Jia, G. Valdes-Ramirez, A. J. Bandodkar, J. R. Windmillerand J. Wang, Angew. Chem., Int. Ed., 2013, 52, 7233–7236.

45 W. Jia, X. Wang, S. Imani, A. J. Bandodkar, J. Ramirez, P. P.Mercier and J. Wang, J. Mater. Chem. A, 2014, 2, 18184–18189.

46 I. Jeerapan, J. R. Sempionatto, A. Pavinatto, J. M. You andJ. Wang, J. Mater. Chem. A, 2016, 4, 18342–18353.

47 A. J. Bandodkar, I. Jeerapan, J. M. You, R. Nunez-Flores andJ. Wang, Nano Lett., 2016, 16, 721–727.

48 Y. Ogawa, Y. Takai, Y. Kato, H. Kai, T. Miyake andM. Nishizawa, Biosens. Bioelectron., 2015, 74, 947–952.

49 J. M. Green, R. C. Pritchett, T. R. Crews, J. R. McLester andD. C. Tucker, Eur. J. Appl. Physiol., 2004, 91, 1–6.

50 M. J. Buono, N. V. L. Lee and P. W. Miller, J. Physiol. Sci.,2010, 60, 103–107.

51 J. M. Green, R. C. Pritchett, D. C. Tucker, T. R. Crews andJ. R. McLester, J. Sport Sci., 2004, 22, 321–327.

52 T. Hanashi, T. Yamazaki, W. Tsugawa, K. Ikebukuro andK. Sode, J. Diabetes Sci. Technol., 2011, 5, 1030–1035.

53 K. MacVittie, T. Conlon and E. Katz, Bioelectrochemistry,2015, 106, 28–33.

54 M. Falk, V. Andoralov, Z. Blum, J. Sotres, D. B. Suyatin,T. Ruzgas, T. Arnebrant and S. Shleev, Biosens. Bioelectron.,2012, 37, 38–45.

55 C. H. Kwon, S. H. Lee, Y. B. Choi, J. A. Lee, S. H. Kim,H. H. Kim, G. M. Spinks, G. G. Wallace, M. D. Lima,M. E. Kozlov, R. H. Baughman and S. J. Kim, Nat. Commun.,2014, 5, 3928–3934.

56 A. Zebda, C. Gondran, A. Le Goff, M. Holzinger, P. Cinquinand S. Cosnier, Nat. Commun., 2011, 2, 1365–1370.

57 H. Sakai, T. Nakagawa, Y. Tokita, T. Hatazawa, T. Ikeda,S. Tsujimura and K. Kano, Energy Environ. Sci., 2009, 2,133–138.

58 Y. Yu, M. Xu, L. Bai, L. Han and S. Dong, Biosens. Bioelectron.,2016, 75, 23–27.

59 C. J. Harvey, R. F. LeBouf and A. B. Stefaniak, Toxicol.In Vitro, 2010, 24, 1790–1796.

60 C. Jacquemoud, K. Bruyere-Garnier and M. Coret, J. Biomech.,2007, 40, 468–475.

61 A. Zebda, S. Cosnier, J. P. Alcaraz, M. Holzinger, A. Le Goff,C. Gondran, F. Boucher, F. Giroud, K. Gorgy, H. Lamraouiand P. Cinquin, Sci. Rep., 2013, 3, 1516–1520.

62 S. Xu, Y. Zhang, L. Jia, K. E. Mathewson, K.-I. Jang, J. Kim,H. Fu, X. Huang, P. Chava, R. Wang, S. Bhole, L. Wang,Y. J. Na, Y. Guan, M. Flavin, Z. Han, Y. Huang and J. A.Rogers, Science, 2014, 344, 70–74.

63 L. Zou, Y. Qiao, Z. Y. Wu, X. S. Wu, J. L. Xie, S. H. Yu, J. Guoand C. M. Li, Adv. Energy Mater., 2016, 6, 1501535.

64 L. Zou, Y. Qiao, C. Zhong and C. M. Li, Electrochim. Acta,2017, 229, 31–38.

Energy & Environmental Science Communication

Publ

ishe

d on

15

June

201

7. D

ownl

oade

d by

Uni

vers

ity o

f C

alif

orni

a -

San

Die

go o

n 14

/07/

2017

18:

50:3

3.

View Article Online