Inkjet-Printed Graphene for Flexible Micro-Supercapacitors L.T. Le1, M.H. Ervin2, H. Qiu1, B.E. Fuchs3, J. Zunino3, and W.Y. Lee1 1 Chemical Engineering and Materials Science, Stevens Institute of Technology, Hoboken, NJ 07030, USA 2 U.S. Army Research Laboratory, RDRL-SER-L, 2800 Powder Mill Road, Adelphi, MD 20783, USA 3 U.S. Army Armament Research, Development and Engineering Center, Picatinny Arsenal, NJ, 07806, USA E-mail: firstname.lastname@example.orgAbstract — Here we report our multi-institutional effort in advances in mW-scale energy harvesting from mechanicalexploring inkjet printing, as a scalable manufacturing vibration and other sources [2-4], we envision the possibilitypathway of fabricating graphene electrodes for flexible of inkjet printing a micro-supercapacitor and integrating itmicro-supercapacitors. This effort is founded on our recent with a printable energy harvester on an implantablediscovery that graphene oxide nanosheets can be easily biomedical device. Such a self-powered implant does notinkjet-printed and thermally reduced to produce and pattern have to be surgically removed from the patient’s body due tographene electrodes on flexible substrates with a lateral the cycle life limitation associated with a rechargeablespatial resolution of ∼50 µm. The highest specific energy battery.and specific power were measured to be 6.74 Wh/kg and2.19 kW/kg, respectively. The electrochemical performance However, to a large extent, integrated flexible micro-of the graphene electrodes compared favorably to that of supercapacitors do not exist in the marketplace today due toother graphene-based electrodes fabricated by traditional miniaturization challenges associated with conventionalpowder consolidation methods. This paper also outlines our fabrication methods such as screen printing and spraycurrent activities aimed at increasing the capacitance of the deposition of electrode materials. In contrast to theseprinted graphene electrodes and integrating and packaging techniques, inkjet printing offers (1) the ability to preciselywith other supercapacitor materials. pattern inter-digitized electrodes with a lateral spatial resolution of ∼50 µm; (2) direct phase transformation fromIndex Terms – Graphene, Graphene oxide, Inkjet Printing, liquid inks to heterogeneous nanoscale structures in anSupercapacitor, Flexible Electronics additive, net-shape manner with minimum nanomaterial use, handling and waste generation; and (3) rapid translation of I. INKJET-PRINTING FOR MICRO-SUPERCAPACTIORS new discoveries into integration with flexible electronics using commercially available inkjet printers ranging fromThere is a tremendous need for rechargeable power sources desktop to roll-to-roll. Some of these transformativethat have long cycle life and can be rapidly charged and attributes are captured in our concept device design (Fig. 1).discharged beyond what is possible with rechargeablebatteries. Electric double layer capacitors, commonlyreferred to as “supercapacitors,” are promising in terms ofproviding fast charge/discharge rates in seconds while beingable to withstand millions of charge/discharge cycles incomparison to thousands of cycles for batteries .Supercapacitors utilize nanoscale electrostatic chargeseparation at electrode-electrolyte interfaces as an energystorage mechanism. This mechanism avoids faradicchemical reactions, dimensional changes, and solid-statediffusion between electrodes and electrolytes, andconsequently provides long-term cycle stability and highspecific power. For high capacitance, electrodes aretypically fabricated of electrically conductive materials such Fig. 1. Flexible micro-supercapacitor concept.as activated carbon with high surface area. II. GRAPHENE AS IDEAL ELECTRODE MATERIALWhile many supercapacitor research efforts are currentlyaimed at developing supercapacitors for electric vehicle In order to increase capacitance, significant efforts are beingapplications, there is also another exciting opportunity to made to explore carbon nanotubes (CNT) and graphene asdevelop micro-supercapacitors for the rapidly emerging ideal electrode materials with their theoretical surface areasflexible electronics market. For example, with recent of 1315 m2/g and 2630 m2/g, respectively [5,6]. Also, their
chemical stability, high electrical and thermal conductivity, spherical ink droplets without clogging nozzles at a lateraland mechanical strength and flexibility are attractive as spatial resolution of ∼50 µm. For example, the dot structureelectrode materials. However, for inkjet printing, these in Fig. 2c was produced with 20 printing passes to show thatnanomaterials as well as activated carbon nanoparticles are drop-to-drop placement and alignment could be repeated tohydrophobic, and thus segregate in water even at very low increase thickness. Also, the average distance between theconcentrations (e.g., 5 ppm for single-walled CNT) unless center locations of two neighbouring droplets could besurfactants are added or their surfaces are functionalized. adjusted to form continuous films. The overlap spacing ofHowever, the use of surfactants and surface modification 15 µm was used for the film shown in Fig. 2d.during supercapacitor electrode fabrication is generally notdesired, since they can significantly decrease capacitance.In contrast to CNT and graphene, the recent “re-discovery”and commercial availability of hydrophilic graphene oxide(GO) at a reasonable price presents a unique opportunity todevelop and use GO as an ideal ink with stable dispersion inpure water (up to 1 wt %) . Although GO itself is notelectrically conductive, it can be thermally, chemically, andphotothermally reduced to graphene . As shown in Fig. 2,we have recently found  that GO, stably dispersed inwater at 0.2 wt %, can be inkjet-printed using a bench-scaleinkjet printer (Fujifilm Dimatix DMP2800) andsubsequently reduced at a moderate temperature of 200°C inflowing N2 as a new means of producing andmicropatterning electrically conductive graphene electrodes.Fig. 2. Inkjet Printing: (a) ink formulation based on stable GOdispersion in water, (b) ink droplets jetted by piezoelectric nozzles,(c) SEM image of a graphene dot printed on titanium substrate, and(d) SEM image of continuous graphene film on titanium. FromReference .At room temperature, the viscosity and surface tension of thewater-based GO ink at 0.2 wt% were measured to be 1.06mPa•s and 68 mN/m, respectively, and were similar to thoseof de-ionized water (0.99 mPa•s and 72 mN/m). Thephysical properties of the GO ink were outside of the rangesrecommended for normal inkjet printing (e.g., 10-12 mPa•s Fig. 3. Initial electrochemical performance: (a) cyclic voltammetryand 28-32 mN/m). Nevertheless, as shown in Fig. 2b, we measured at different scan rates (b) specific capacitance retainedfound that manipulating the firing voltage of piezoelectric over 1000 charge/discharge cycles at a constant scan rate of 50 mV/s and (c) Ragone plot. From Reference .nozzles as a function of time was effective in generating
Titanium foils from Sigma Aldrich (100 µm thick, 99.99%purity) was used as an example of flexible substrate andcurrent collector for our initial electrochemicalcharacterization. Electrochemical performance wasevaluated by cyclic voltammetry (Fig. 3a) and galvanostaticcharge/discharge. Two identical electrodes were clampedwith a Celgard separator. 1 M H2SO4 was used as theelectrolyte. The specific capacitance of the grapheneelectrodes was measured to be 48-132 F/g in the scan rangeof 0.5 to 0.01 V/s. As shown in Fig. 3b, 96.8 % capacitancewas retained over 1000 cycles. The specific power andenergy density of the graphene electrodes are plotted in Fig.3c.Table 1. Comparison of electrochemical performancesAs compared in Table 1, the capacitance of the grapheneelectrodes was similar to that reported for other grapheneelectrodes prepared by conventional powder-based methodsin the absence of any pseudocapacitance materials added tothe electrodes [5,10,11]. However, the power density ofIPGEs was considerably lower than that of CNT-basedelectrodes which has been reported as high as 100 kW/kg[12,13]. The lower power density of the graphene electrodesmay be partly explained by the lack of: (1) interconnectivityamong 2D graphene nanosheets for electron conduction and(2) 3D mesoscale porosity for ion conduction. Nevertheless,the initial performance of the inkjet-printed is promising,and is expected to be further improved by optimizingprinting and reduction conditions and by optimizing its 3Dmorphology. III. CHALLENGES AND CURRENT ACTIVITIESThe fundamental scientific challenge for this research stemsfrom the lack of understanding of and experience withgraphene and GO as new nanoscale building blocks for 3Dassembly. For example, our initial results show that we arecurrently utilizing less than 12% of the theoreticalcapacitance possible with graphene (i.e., 132 out of 1104 F/gfor H2SO4 electrolyte). We are currently exploring a conceptof adding nanospacers to control the stacking behavior ofconformal graphene nanosheets and therefore to increasespecific surface area and capacitance. Also, as illustrated inFig. 4, we are focusing on droplet coalescing as an importantprinting parameter that: (1) will determine optimum printingspeed and (2) can be used to create disordered 3D assembly Fig. 4. Overlapped droplet spacing of: (a) 5 µm (b) 25 µm and (c)of graphene nanosheets as another means of controlling the 15 µm. (d) illustration of nozzle and substrate movements duringconformal stacking behavior of the nanosheets. inkjet printing.
We observed the significant effect of droplet overlap spacing the specific electrolyte development and packaging issueson the formation of continuous boundaries which appear as and challenges associated with realizing micro-“white” lines in the SEM images (Figs. 4a-c). As evident supercapacitors that can be integrated with flexiblefrom these SEM images, the average distance between the electronics.boundaries corresponded well to the overlap spacing ofneighboring droplets used to prepare these graphene thinfilms. At a high magnification (Fig. 4c), graphene sheetsappeared more wrinkly and less uniform at the boundariesthan in areas between the boundaries. The results suggestthat we may be able to control and use these boundaries as amechanism to produce more disordered 3D assembly of thenanosheets.Fig. 4d illustrates the 3D operation of multi-nozzle printing.d1 and d2 are the overlap spacings between two neighboringdroplets, which can be controlled as low as 5 µm in the x-and y-directions, respectively. During typical operation, theprinthead moves in the x-direction to place the first row ofdroplets for a specified distance. When the printhead comesback to its original x location, the substrate stage moves inthe y-direction so that the printhead can place the secondrow of droplets. In addition to the spacing parameters, thereare several key time variables to consider from a scalingperspective. t1 is the time between placing two neighboringdroplets in the x-axis direction with the controllable range of∼0.5 ms, t2 is the time it takes for the printhead to be readyto print the next row droplets in the y-direction (e.g., ∼10 sfor 1 cm x-direction motion). t3 is the time between placingthe two layers of droplets in the z-direction (e.g., ∼4 min for1 cm2).The effects of these variables on the development ofboundaries with GO ink are being evaluated. Once we areable to understand and control the formation of continuousboundaries, the new processing/structure knowledge may beused to: (1) assess surface area and capacitanceenhancements associated with morphology tailoring and (2) Fig. 5. Islands formation as a function of substrate hydrophobicity:scale fabrication using bench- and industrial scale printers (a) hydrophobic surface of as-received Kapton and (b) hydrophilicwhile controlling electrode morphology. surface of treated Kapton.On the concept device fabrication and demonstration fronts, IV. CONCLUSIONSwe have undertaken several activities. Kapton (DuPont) is Hydrophilic GO dispersed in water was found to be a stableinitially chosen as a flexible substrate material. Inkjet ink for inkjet printing of GO with the lateral spatialprinting of the GO ink on as-received Kapton substrate resolution of 50 µm. Subsequent thermal reduction of thesurface resulted in the formation of islands of about 1 to 2 printed GO produced electrically conductive graphenemm (Fig. 5a). After the substrate surface was treated with electrodes with promising initial electrochemicalpotassium hydroxide for 3 h, the island formation was performance for flexible micro-supercapacitor applications.considerably reduced (Fig. 5b). This change was attributedto the spreading of hydrophilic ink droplets on the Kaptonsurface becoming hydrophilic with the treatment. For ACKNOWLEDGMENTcurrent collector, a commercially available silver The authors thank the U.S. Army - ARDEC for funding thisnanoparticles CCi-300 ink (Cabot Inc.) is selected. This ink project under the contract of W15QKN-05-D-0011.contains 20 nm silver nanoparticles suspended in a mixtureof ethanol and ethylene glycol. We are evaluating several REFERENCESelectrolytes for electrochemical compatibility with inkjet-  Conway, B.E., Electrochemical Supercapacitors: Scientificprinted silver. For packaging, we are exploring a heat- Fundamentals and Technological Applications, 2nd edition. 1999:sealing approach using heat-sealable Kapton FN as a Springer.  X. Chen, et al., 1.6 V Nanogenerator for Mechanical Energyprimary method and soft-lithography as an alternative Harvesting Using PZT Nanofibers, Nano Letters, 2010, 10(6), p.option. These initial investigations are expected to uncover 2133-2137.
 R. Yang, et al., Power Generation with Laterally Packaged Piezoelectric Fine Wires, Nature Nanotechnology, 2009, 4, p. 34-39. R. Yang, et al., Converting Biomechanical Energy into Electricity by a Muscle-Movement-Driven Nanogenerator, Nano Letters, 2009, 9(3), p. 1201-1205. Stoller, M.D., et al., Graphene-Based Ultracapacitors. Nano Letters, 2008. 8(10): p. 3498-3502. Geim, A.K. et al., The rise of graphene. Nature Materials, 2007. 6(3): p. 183-191. Paredes, J.I., et al., Graphene Oxide Dispersions in Organic Solvents. Langmuir, 2008. 24(19): p. 10560-10564. Zangmeister, C.D., Preparation and Evaluation of Graphite Oxide Reduced at 220 °C. Chemistry of Materials, 2010. 22(19): p. 5625- 5629. Le, L.T., et al., Graphene supercapacitor electrodes fabricated by inkjet printing and thermal reduction of graphene oxide. Electrochemistry Communications, 2011. 13(4): p. 355-358. Vivekchand, S., et al., Graphene-based electrochemical supercapacitors. Journal of Chemical Sciences, 2008. 120(1): p. 9-13. Liu, C., et al., Graphene-Based Supercapacitor with an Ultrahigh Energy Density. Nano Letters, 2010. 10(12): p. 4863-4868. Kaempgen, M., et al., Printable Thin Film Supercapacitors Using Single-Walled Carbon Nanotubes. Nano Letters, 2009. 9(5): p. 1872- 1876. Honda, Y., et al., Aligned MWCNT Sheet Electrodes Prepared by Transfer Methodology Providing High-Power Capacitor Performance. Electrochemical and Solid-State Letters, 2007. 10(4): p. A106-A110. Z.S. Wu,et al., Anchoring Hydrous RuO2 on Graphene Sheets for High-Performance Electrochemical Capacitors, Advanced Functional Materials, 2010, 20(20), p. 3595-3602. H. Gómez, et al., Graphene-Conducting Polymer Nanocomposite as Novel Electrode for Supercapacitors, Journal of Power Sources, 2011, 196(8), p. 4102-4108