Notes from the 2015 ECM4 Conference, Aberdeen, Scotlandd
Notes from the ECM4 conference
Met with Mathew Halls from Schrodinger. He was curious about what I did, came up, gave me his card. Apparently he knows Chin Tang. He’s giving a talk on the last day.
Opening Remarks – Bob Collier – Aberdeen Chamber of Commerce
European Capital of Oil and Gas. 25% of Scotland GDP. Two major challenges, going to green energy, and low oil and gas prices mean efficiency is king. Wants solutions that can be as impactful as cell phone batteries.
Half of the conference is from USA, China, Japan, UK. ½ academics, ¼ industry, ¼ research societies. Really well credentialed attendees.
Moderators are reporting on the content of the presentations.
Plenary Speaker – Alexei Kornyshev – Imperial College of London
“The essential physics of supercapacitors and electroactuators at the nanoscale. Electricity vs mechanics”
Fellow of IOP, RSC, RSB, ISE
Room temp Ionic Liquids (RTIL) are large molecular ions, have irregular shapes, low melting points, miscible with oil and water. There are so many of them! Applications are numerous for catalysis, lubricants, electowetting, liquid-liquid interfacing.
EDLCs in supercapacitors -> need high area electrodes, separator prevents shorts. If the EDL is on the same scale as the electrode architecture, it doesn’t scale linearly. Gouy-Chapman-stern theory describes the classic electric double layer, or poisson-boltzmann theory. The last layer of condensed ions can have variable voltage dependent capacitance. ‘Double layer to ionic-Liquids: Paradigm change’. It is possible to actually possible to saturate the double layer, such that the double layer increases in thickness, reducing capacitance. Therefore, EDLC capacitance can be maximized! Asymmetric ion sizes can affect the precise maximization point. However, this is all theoretical, practically, electrochemistry occurs before approaching these scales. Overscreening leads to capacitative oscillations, where condensed charges bring out more screening ions than they should, which is compensated by the material. Applying a voltage kills overscreening.
3d EDLs are the new hot thing (last 3 years).
Water is bad compared to ionic liquids (it will electrolyze). Residual water in supercapacitors is a bad thing (it can be electroabsorbed to the surface, taking the potential place of an ion, reducing capacitance). Therefore, electrode design is just as imprortant as the RTIL. Space charges will dominate if it is smaller than the EDL capacitance. Graphene has much higher space charge.
Nanoporous supercapacitors. Carbon derived carbon is fabricated using RIE, gets 50% porous like activated charcoal. Maximized capacitance at 0.7nm when the ion size is about the pore size. Coulombic interactions are screened out at this scale (0.3 nm and lower). Basically, there is an optimal ion size for each pore size. Cations and Anions can have different adsorption affinities. More energy can be stored in ionophobic pores (analogous to a stiffer spring storing more energy than a softer spring). Dynamics are difficult (classic transmission line model breaks down at very thin layers, what’s the resistance of a single layer of ions?). Diffusive process dominates for ions charging and discharging in ionophilic pores. Simulation predicts a self diffusion coefficient increasing 1000x. Charging is greatly accelerated in ionophobic pores!
Electroactuators work by charge separation inducing mechanical stress on an electrode. The reverse process allows you to harvest energy!
During the questioning session, he mentioned in passing that nanoconfined water likes to adsorb!!!
Session A
Vibrating Floor for Power Generation on HEMS http://nscj.co.uk/ecm4/sessions/054_TakashiYoshikawa.pdf
Takashi Yoshikawa
They have energy problems in japan because there are poor fossil fuel resources. They want to create personally created energy through energy harvesting. Energy use in japan has actually plateaued per household. HEMS (home energy management system). Wireless energy through HEMS will have high fixed costs but low conitinuing costs. It’s turned into a consumer device talk. He’s describing a thermostat or a more integrated NEST, measuring the power consumption. When he actually gets to interesting things, he talks using a peltier element on human skin, harvesting energy from human action. You can actually harvest magnetic leakage from fluorescent light (up to 2.4 mW). Or from hall effect sensors on ventilation (40 uW). Mini hydroelectric power in drain pipes (5 mW). Temp gradient from a bathroom (1.6 mW). On a piezo embedded into the floor (60 uw). Rangone plot of all these different energy harvesting shows that ambient light is the most efficient. NIST requires wireless transmission above 0.13% at 0.55 meters.
Nice to get some information/hard numbers, but not a great talk.
Optimized vibration harvesting system based on piezoelectric transducer and impedance matching voltage converter
http://nscj.co.uk/ecm4/sessions/126_PeterSpies.pdf
Energy harvesting replaces or extends battery lifetime. Piezomaterials can be much easily be miniaturized compared to other technology. Goal: Supply a tracking system with piezos on storage boxes (think RFID + other sensors powered by vibrations, used to store fruit). 10uW wakeup power, 270 uWs needed for data transmission. Self oscillating system works by bending the ceramic piezo to generate charge. Maximum power output only happens if the stress is evenly distributed. Dual counterlevel structures harvest from all directions. All the piezo materials are fed into capacitors (AC-DC conversion + storage). 1 harvester produces 200 uWs. 4 wave bridge rectifier used to smooth out power, however diodes are capacitive, and thus phase shift the signals, reducing power generation. Impedance matching with an inductor puts things back into phase. A control circuit with transistors and an ASIC modifies the inductor. Efficiency is doubled compared to not having a control circuit. Control circuit uses 35 uW.
Variable capacitor energy harvesting based on polymer dielectric and composite electrode
http://nscj.co.uk/ecm4/sessions/341_RobertHahn.pdf
Fraunhofer corp is a IC packaging company working on micro fuel cells and microbatteries. Human based schemes rely on low frequency random motion. Electrostatic harvester works by stretching and overlapping electrodes, changing the capacitance. These kinds of harvesters require precharging. 0.4 mW can be generated in theory, but power conversion cuts this efficiency in half. Liquid metals were originally used, but now metallized elastomers and high-k dielectrics can be put into an insole. CV curves show storage. Parasitic resistances dominate at in the regime where low frequency random motion occurs. Thin layers of dielectric near the nano conductors degrade the capacitance (1/c smallest capacitances dominate). 10 Gohms needed to prevent parasitic materials.
Piezoelectric Soft MEMS for Energy Harvesting
http://nscj.co.uk/ecm4/sessions/234_FrancescoGuido.pdf
Converting mechanical strength into electrical energy. Basically just a metallized aluminum nitride film on kapton tape. Uses molybedenum for an electrode. Mo thin film can be grown epitaxically, and AlN is used as an adhesion layer to the kapton. The whole film structure is about 1 um thick, sealed with parylene. You need finite element modeling to predict the resonant frequency of the flexible substrate that imparts the energy to the piezo. 10.5 m/s wind speed generated 1.4 Vpp (257 uW/cm^3). Doped piezo materials appear to work very well.
Storing Energy on strained ZnO micro pillars for a Mechanical Battery
http://nscj.co.uk/ecm4/sessions/127_OscarSuchil.pdf
We are trying to get beyond 10 mW/cm^2. Conversion losses really prevent efficient transmission and energy generation. He proposes to keep mechanical energy mechanical. (si+au substrate, nanowire array with a compressor electrode on top). Nanowire acts as a spring. Think of calcium crystals in vestibular system. Many different parameters can effect the formation of the nanopillars (think seed crystals). The pillars are about 1 um diameters, 6 um tall. 300-500 mN loading forces per pillar. Electrical characterization through IV curves give you the energy in each compression setting. Looks like a schottky diode?!?!? He’s redesigning the setup because he isn’t confident in the results.
What good is a nanospring? Any strain on a nanopillar will be unevenly loaded. We aren’t winding nanoclocks!
Session B
Different measures for the capillarity-driven deformation of a nanoporous metal
http://nscj.co.uk/ecm4/sessions/136_LiHuaShao.pdf
Functional metallic nanomaterials need many different interfaces. Nanoporous gold (like a polymer membrane (100 nm rods)). Electroplating is used to form these structures (1mmx2mmx1mm). Creating textured nanoporous films can shift their mechanical properties (mechanical actuator deflects 3mm, frequency pole at 1.5 hz). They’re trying to get electro-absorption/desorption of materials on the gold films. Capillarity induced strain of porous materials rely on the stress within the bulk to be related to the surface tension. Work against surface stress and work against surface tension are not the same thing. Pressure in the fluid itself can be as strong as whatever surface tension is causing stress.
Carbon aerogels have ridiculous surface areas (3200 m^2/ cm^3).
Tailoring the surface states of dealloyed nanoporous metals for energy technologies
http://nscj.co.uk/ecm4/sessions/103_YiDing.pdf
Ranges of 0D-4D nanostructured metals. Fuel Cell commercialization relies on catalysts that are substantially better than the current Pt/C catalysts. Low-Pt electrocatalysts basically roughen a surface and deposit platinum upon the surface. Having a high activity isn’t necessarily the most desirable thing compared to stability. Three-phase boundary dilemma. We need good electron conductivity, but low proton conductivity, High density, High porosity. Nanoporous gold can provide an adhesion substrate for Pt, maximizing the amount of Pt exposed. Alloying Pt onto the surface occurs through the use of formic acid. This formic acid can be used as a catalyst itself to help Pt form its own reactions (150-400x better). Layering Pd on top of the nanoporous gold makes it even better. Reducing Pt -> reduce fuel cell costs. 30 ug/cm^2 is the metric to beat. Platinum I would deposit in a 50nm thick film would be 110 ug/cm^2. NAFION can be used as an attachment layer for gold!!!
Ordered mesoporous WO3 photocatalyst for the selective oxidation of methane to methanol
http://nscj.co.uk/ecm4/sessions/176_KatherineVilla.pdf
Methanol production at scale requires high temperatures and pressures. Photocatalysis can reduce this energy cost. Tungsten trioxide has a crystalline structure, is mesoporous, provides substrate surface for reaction to occur. Surface area is 151 m^2/g. Its only about 2x better than direct photolysis of water. When modified with fluorine on the surface it actually makes more hydroxyl groups which facilitates methanol production (10% increase). Doping additional metals into the W03 also promotes hydroxyl formation, resulting an additional doubling in the conversion efficiency.
Question: Would increasing the surface area to volume ratio improve the catalyst? Mesoporous moves to nanoporous.
Modelling of turbulent flow over porous media using a volume averaging approach and a Reynolds stress model
http://nscj.co.uk/ecm4/sessions/182_MichaelMoessner.pdf
The trailing edge of a wing, if porous, reduces the noise of the plane! How does that work? Turbulence and porous structures? Averaged out navier stokes equations. How turbulent is flow at different locations? Difficult question to answer. Boundary condition effects can effect the turbulence. Really not a great talk.
Session C
Molecular insights into electrical energy storage in supercapacitors with ionic liquid mixtures as electrolytes
http://nscj.co.uk/ecm4/sessions/011_GuangFeng.pdf
Supercapacitors do not have enough energy density to keep up with batteries. Complicated ionic liquids have a lot of charge -> higher energy, but water can get locked in and reduce the density. Removing water from electrified interfaces will be very important to improve this problem. Positive charges at the surface condense more water (function of water molecule polarity). This impacts the capacitance significantly (10% losses). Hydrogen bonds can form between water and the ionic liquids! When charged, the ionic liquids don’t bond so strongly, and the water is free to bond with itself!
Magnetic Supercapacitors based on Graphene-FeNi3 Hybrids
Dr. Gonzallo Abellan
Layered double hydroxides. Organic molecules can be shoved into these LDH. Combining LDH with graphene improves the conductivity of the LDH so they can be used for energy applications. Nanotubes can be grown on LDH pads, making fingers for an electrode in a supercapacitor. Carbon nanoonions form at low temperatures, delaminate and laydown different monolayers of graphene at higher temperatures. Spintronics applications too. He makes the point that everyone tries to optimize the internal components, and not the external environment in which the supercapacitor exists. Supercapacitance is affected by magnetic fields. It increases the effective capacitance (0.5 T 10x these hybrid’s energy storage). Magnetic Resistance decreases, the surface is oxidized, graphene is mostly unaffected. FeNi3 turns to NiO.
This talk was a big deal, the moderator basically said he was submitting to Nature on these results. The author didn’t seem to take questions very well, when people in the audience were trying to better understand the mechanism.
Growth and plasma functionalization of carbon nanotubes for supercapacitor applications
http://nscj.co.uk/ecm4/sessions/050_RogerAmadeRovira.pdf
Goal is to optimize the carbon nanotubes for supercapacitor applications. Nanoislands are seeded then CNTS are grown. Raman is used to differentiate between oxidized and non oxidized states.
Plenary Talk: Energy Harvesting using ZnO-Based Thin Films and nanostructures. Dr. David Rogers, Nanovation, France
ZnO is a wide bandgap semiconductor, much like GaN. Zn is easier to process but not easy to p-dope for energy applications. ZnO is biocompatible (GRAS). Applications in optoelectronics, gas sensors, varistors. 1 Billion dollars market now. New applications in photovoltaics and LEDs. Pulsed Laser Deposition has better plasma kinetics than just sputtering (10-100eV), putting a lot of oxygen in oxides. PLD can also be used to from nanostructures (rods, cones, pillars). These structures rely solely on the grain size and temperature of deposition. Using e-beam litho you can create patterned arrays. Energy Harvesting using ZnO relies on the fact that ZnO is transparent in solar cells (think like ITO). We are trying to replace ITO because it is a rare element (10% of photovoltaic price). Doping ZnO with Al helps lower the resistivity so that you can use them in solar cells. Using ZnO pillars as an antireflective coating will help improve the efficiency of solar cells (0.5% reflected vs 10% reflected). Not the strongest piezo material, but super flexible, meaning it has an excellent frequency response. Larger crystals produce larger voltages, but the triboelectric effect has 5x the power, so these researches have gone by the wayside. ZnO as a thermoelectric material (need a figure of merit better than 4). Current thermoelectrics rely on Te which is rare, expensive, and toxic (highest figure of merit = 2). Need high electrical conductivity, but low thermal conductivity. Reducing the thickness of the thermoelectrical materials reduces the thermal conductivity.
(note maybe we can dope pnc-Si). All the voids should reduce thermal conductivity, but 50nm thin film could be enough to maintain a good electrical conductivity.
Session D
Micro-supercapacitor electrodes based on hydrous ruthenium oxide and highly porous Au current collectors: the recipe for High Areal Capacitance
http://nscj.co.uk/ecm4/sessions/141_DanielGuay.pdf
Devices need to harvest, store, and run off their own energy. (Let this be known as the day that the machines took over ala the matrix). Microbatteries have power density and lifetime limitations. Each capacitance mechanism has a different shape on an IV curve. EDLCs have square shapes, redox reactions have definite peaks, pseudocapacitance looks like exponential superimposed on squares. Pseudocapacitance has a lot of capacitance/mass (RuO2 = 720 F/g). This is due to the wide multitude of redox states. Of course, these materials’ capacitances are the result of surface interface with the electrolyte; thick films are no good. If you need a certain thickness for stability, structure the current collector to interface with the thin film. E=0.5CV^2, usually V is fixed. Glancing angle deposition allows a sputtered material to nucleate columns, which shadow out other angles. This roughened structure increases the capacitance 4x. Mesoporous gold can be electrodeposited in sulfuric acid (hydrogen bubbles create voids in the film). Supercapacitor is on the same order as lithium microbattery. Cost is about 10 cents /cm^2, 3 F/cm^2.
Identification of parameters of the fractional model of supercapacitors
http://nscj.co.uk/ecm4/sessions/172_JoseQuintana.pdf
Not a great speaker. Couldn’t really understand him.
This talk focuses on the modeling behind EDLC in combination with battery circuits.
Layered Transition Metal Dichalcogenides and Reduced Graphene Oxide Hybrids for Supercapacitor Applications
http://nscj.co.uk/ecm4/sessions/015_ChandraSekharRout.pdf
I think I reviewed one of this group’s paper previously, using a vanadium battery. But this time they’ve integrated some of their vanadium into graphene sheets. They’ve improved upon tungsten selenide, increasing their results to 845 F/g.
Ionic liquid–added Biopolymer Electrolytes for Electric Double Layer Capacitors (EDLCs) Application
http://nscj.co.uk/ecm4/sessions/030_RameshTSubramaniam.pdf
Solid polymers have better electrochemical stability than liquids, however they have lower ionic conductivity and aren’t as mechanically strong. So, they blend ionic liquids into polymers to create new solid polymer electrolytes. PMMA-PVC is held as an example where two polymers act to provide mechanical and electrical strength. Plasticizers are used to promote the migration of charges by turning polymers into gels, but they aren’t particularly healthy, and act only over a narrow electrochemical window. By adding nanoparticles instead of plasticizers, we can create composite polymer electrolytes that are healthier and act over a wider range. Preparing electrodes was accomplished by forming activated carbon on starch.
Session E
In-situ Thermal Stability Analysis of Au Nanoparticles in a Metal Oxide Framework Based on a Sacrificial Fiber Template
http://nscj.co.uk/ecm4/sessions/231_JesseSJur.pdf
Inorganic modification of textiles. Uses ALD to put aluminum oxide, ZnO onto textiles. ALD on Au nanoparticles affects the SPR. Applying nanoparticles to clothing and then ALD. ALD precursors can infiltrate textile fibers (up to a micron deep), this can create like isotropic etches, increasing the surface area. You can burn out the scaffold and leave nanoparticles embedded into the infiltrated zone. (Padbury et al (Langmuir 2015)). This talk was very interesting, incorporating MEMS structures into clothing might be a future development.
Experimental analysis of the influence of contact conduction on the effective thermal conductivity of steel bar bundles
http://nscj.co.uk/ecm4/sessions/186_RafalWyczolkowski.pdf
Nice Polish Fellow.
Steel bars are heated in bundles to ensure the most efficient use of space, but they aren’t necessarily heated evenly. Surfaces are rough, so many microcontacts form, and the distribution depends on the local flatness, the elastic properties, and the mechanical load (grease works as a lubricant to evenly spread out the force on the microcontacts). Staggering the bars (highest packing density) produced the best thermal distribution.
High velocity interfacial collision: Is it a potential method to produce a nanoporous surface?
http://nscj.co.uk/ecm4/sessions/327_ThaneshanSapanathan.pdf
Not a great speaker.
Metal foams are used in a lot of places. Magnetic fields are used to drive ions into metallic films to make nanoporous structures. He used a cold welding process (magnetic) to run the ion collisions.
My talk
About half of the room (8 people) entered for my talk! It went pretty well. There were some questions about how my work is different than the porous silicon work done back in the 1980’s which I demurred other than the fabrication. I had some questions about how Karl’s NFT worked/was fabricated by an ALD expert who presented (Jesse Jur).
Session G
Analysis of Plasmonic Subwavelength Crossed Metallic Grating in Ag/SiO2/Ag sandwich nanostructures
http://nscj.co.uk/ecm4/sessions/364_HoangYanLin.pdf
Optoelectronic integrated circuits using quantum dots as a laser source. Smaller Quantum Dots have bigger bandgaps (smaller = bluer light). Metal insulator Metal nanocavities form the basis for lasing on ICs. Different gratings were tested. Excessive Si without oxygen when crystallized forms quantum dots (our pnc-Si potentially?)
Piezoelectric Biomedical microelectromechanical systems (BioMEMS) for implantable energy harvesting (EH) applications
http://nscj.co.uk/ecm4/sessions/390_MartinZalazar.pdf
BioMEMS applied to energy devices. Aluminum nitride plus ultra nanocrystalline diamond (a substrate) as a piezoelectric material. UNCD layers are seeded with nanodiamonds and microwave plasma CVD is used to grow the film. They sputter on AlN. He uses thin membranes (1 um film stack) much like we do (backside silicon etch with KOH). He plans to make integrated energy harvesters with rectifying diode bridges.
Improved electro-conjugate fluid jet generator with bypass micro-channel
http://nscj.co.uk/ecm4/sessions/372_RuzhaoZhang.pdf
Neat ECF generation with high voltages. (think like Blanca’s dielectrophoreseis structures, except organized to pump instead of concentrate). Nickel electrodes were formed in parallel to join the fluid flows in parallel. The pump can put out 40 kPa at 3000V (1000V gives you 10 kPa). The high voltages actually deform the microchannels causing delamination from the electrodes. The bypass microchannel evens out flow, but at a cost of output.