1 The Bleeding Edge – Part II: Nano Energy TechnologiesToday we'll discuss the possible use of nano in power production and storage: Starting with a short visit to James Clerk Maxwell, then proceeding to . . . A careful look at Photovoltaics (a.k.a. "solar cells"), including: - The critical difference between photovoltaics and photoconductors - The impact (present and future) of nano-quantum structures Then on to ways of storing electrical energy: - Nano Capacitors - Nano Batteries - Nano Fuel Cells A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm

2 For even more information on today's topics see:Lectures from my "Introduction to Sustainable Energy Systems" class including: Solar Power Next Generation Solar Power Batteries and Fuel Cells Power Cycles & Energy Storage A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm

3 MUCH attention is now focused on use of nano for powerFor PRODUCTION of power via "Quantum Dot" solar cells Using light to move electrons Or STORAGE of power via nano capacitors and batteries Into which excess electrons can be driven and temporarily stored That is as far as most news reports (and much research) goes But if this is all you consider (and engineer): Weak photo-current will briefly flow from quantum dot cell, and then stop Or nano capacitors will only ever store nano electrical power What then are the problems?

4 A short visit with James Clerk Maxwell:Maxwell's 1st Equation: Electric Field builds in proportion to net charge "Net charge" = Positive charge density – Negative charge density Electric force is then proportional to the strength of that electric field So just a TINY ACCUMULATION of net charge => HUGE FORCES For a second or two: Then there is a loud snap as charge build-up dissipates On scales much greater than molecular dimensions Nature will not LET you remove or add significant net charge!

5 So "electricity" is almost all about pumping chargePUMP charge in one end of something and out the other end: "Something" = Battery, solar cell, generator, That's WHY it's called electrical current: An analogy to incompressible water: Can pump water THROUGH pipes, but if try to increase water IN pipe => Explosion! Battery, solar cell, generator are all CHARGE PUMPS And pumps are judged on basis of the flow and pressure they can generate: Water Power = Flow x Pressure which is analogous to: Electrical Power = "Current" x "Voltage" A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm

6 Photovoltaics: What happens when light strikes a material?Case 1) Photon energy < Material's bond energy: Photons can't shake anything loose, most just proceed on through That is, material is ~ transparent to these too low energy photons Case 2) Photon Energy ~ Material's bond energy Photon IS now absorbed and its energy used to kick an electron out of a bond Case 3) Photon Energy > Material's bond energy Photon is absorbed: Part of its energy kicks an electron out of a bond Rest of its energy also goes to that electron in the form of kinetic energy That is, photon kicks electron out of the bond, then kicks it in the butt!

7 So when we shine light having bond energy onto a material:(Remembering that I want is power = current x voltage out of a solar cell) This light is (at least eventually) going to be absorbed by a bond in our material: Before: Atom cores (positive nuclei + inner electrons) + bonding electrons In 1D: ~ After: One negative electron is liberated, leaving behind a positive region: / / Electron is drawn BACK to positive region => falling back into bond (or a FEW might wander out the left or right end)

8 But this only gives me a photoconductor:Most electrons just wander until pulled back into bonds Or ones that DO exit are equally likely to exit right or left Nothing is pumping (pushing) electrons to flow in one direction! Application? ADD external battery/power supply and use as a light detector: No light: All electrons in bonds, no current through sample (despite battery) Light: Freed electrons. Battery can now suck them out one end and push back into other But where did the light's energy go? Into the atoms Freed electrons fell back into atoms' clutches Giving the atoms a kick => atomic vibrations (a.k.a. heat)

9 To produce power we've ALSO got to drive (PUMP) electrons somewhere!Classic Technique: START with fully-bonded electrically neutral material, most commonly silicon It sets the bonding rules with its crystal structure: Rule with Si = four bonds ADD atom of almost same size but with one less bonding electron (e.g. boron) Fits into crystal, steals electron from elsewhere, making it an Acceptor (thief?) Bond where electron stolen from now becomes a positive Hole Add neutral Acceptor atoms to Si => Negative ions + Liberated holes: Silicon atoms = Grey (fixed neutral atoms) Acceptor ions also FIXED in position Holes = MOBILE Why? ANSWER: Hole grabs electron from neighbor, leaving hole in a NEW place . . . And holes don't fill with electrons from outside because that would add net charge

10 Can also add things that will shed electronsDonor = Similar to Si in size, but with one additional bonding electron (e.g. P, As) Fits into crystal but final electron has nothing to pair with and bond Add room temperature heat, and this atom ionizes => Donor: By liberating its last, weakly attached, electron: Silicon with added Donor impurity atoms => Positive ions + Liberated electrons: Here only liberated electrons are MOBILE And, as in other material, net charge is still zero! So James Clerk Maxwell is still happy And if mobile electrons return home, heat will eventually kick them back out! NOTE: Acceptor and Donor impurities are called "DOPANTS"

11 Payoff comes when you put two such "doped" regions side by side:Acceptor ions + Mobile Holes: Donor ions + Mobile Electrons: At intersection ("junction”) mobile electrons are going to rush across to FILL mobile holes!! (Because holes ARE just bonds that have lost one of the normal paired electrons!) Mobile electrons filling the holes (in the bonds) is called "recombination"

12 Central junction thus becomes depleted of ALL mobile charges (liberated electrons or holes):But this leaves uncompensated FIXED acceptor ions (-) / donor ions (+) at the junction Which produces a growing Electric field at that junction Migration / recombination continues UNTIL field is strong enough to block further migration Because Electric field pushes positive charges left and negative charges right Electric field thus locks remaining mobile holes and electrons on their respective sides

13 NOW add light to knock electrons out of background silicon:Light photon knocks an electron out of a bond, creating a wandering electron + hole (traveling together ="exciton") New electron and hole can both wander, but if reach "junction:" "Built-in" electric field traps new electron on right, but propels hole to left If instead created on left, hole trapped on left, but electron swept to right = A CHARGE PUMP (BTW this is also a DIODE: Can only force current through it in ONE direction)

14 NET RESULT (again) = Build up of electric field at boundaryMore general way of creating boundary charge-separating electric field: ABOVE: ONE MATERIAL but divided it in TWO DIFFERENTLY BEHAVING REGIONS Made two regions different by adding acceptor OR donor impurity atoms ALTERNATIVE: Just put two DIFFERENT MATERIALS side by side Electrons at higher energies on one side will try to cross over to other side NET RESULT (again) = Build up of electric field at boundary

15 Leading to common rules for almost all photovoltaics (solar cells):Must have at least one set of paired materials: Be it two distinctly different materials OR One basic material (e.g., silicon) modified into two differently acting layers In that pair, one layer/material must cling onto electrons more tightly So that electrons will flow into it from second material Until shift of charge across boundary builds up ELECTRIC FIELD at interface Which will tend to counter further shifting of charge That "interfacial" electric field will then provide the critical push Light energy => breaking electron bonds But ELECTRIC FIELD then pushes freed electrons all in one direction

16 Solar cell efficiency = Power produced / Power receivedBut solar energy is all about "efficiency" Solar cell efficiency = Power produced / Power received Power received = Solar power: Above the earth's atmosphere, peaks at about Watts / square meter This value is referred to as "AM0" (air mass zero) Atmosphere will absorb ~ 25% of this power => ~ 1000 Watts / square meter Referred to as "AM1.5" (air mass 1.5) But this = MAX surface intensity Because this is value for sun DIRECTLY overhead Which happens only in certain locations, in certain seasons, once a day AND clouds / haze / fog will further reduce intensities!

17 But what is the amount of POWER (= current x voltage) PRODUCED?Current comes from the number of electrons liberated by light / second - Function of how strongly that material absorbs photons of that color - AND of how much material is doing the absorbing (e.g. its layer thickness) Voltage comes from charge driving/separating junction ELECTRIC FIELD Which was created by process of bond filling/liberating. Leading to fact that: Photo-electrons/holes are driven out of cell by ~ 60-70% of liberation energy => Solar cell voltage ~ (0.65) (liberation energy) / (electron charge = "e") For Si solar cell, "Voc" ~ (0.65) (Si electron liberation energy = 1.1 eV) / e ~ 0.7 Volt Larger liberation-energy ("bandgap") solar cell materials => More VOLTAGE But Less current: Why? (Answer => motivation for using quantum dots)

18 For answer, we must go back to solar spectrumEnergy Wavelength If this strikes a solar cell made of a material having Small liberation energy => MOST colors liberate electrons, but are driven out of cell by small voltages If this strikes a solar cell made of a material having Large liberation energy => Only HIGH ENERGY light liberates electrons But fewer electrons that ARE liberated will be driven by higher voltages! So now try to find optimum combination (based on choice of optimum material):

19 Energy absorbed by cells made of different materials:Cell with small "bandgap" 0 eV 5 eV Light's Energy VIS Energy absorbed by solar cell Electron Liberation Butt Kicking Cell with medium "bandgap" Electron Liberation Butt Kicking 0 eV 5 eV Light's Energy VIS Energy absorbed by solar cell Cell with large "bandgap" Electron Liberation Butt Kicking 0 eV 5 eV Light's Energy VIS Energy absorbed by solar cell Blue Triangles => Energy passing right thru the cell! Yellow Triangles => Energy lost to heating of the cell Rectangles (only!) => Electricity out of the cell

20 Plotting backward, as percentages (preparing to blend in solar intensities):5 eV 2 1.5 1.25 1 eV 0.75 eV 0.5 eV 250 500 750 1000 1500 2000 2500 Sun's wavelengths (in nm) 100% 50% 0% Electron Liberation Electron Butt Kicking For solar cell using material with 0.5 eV electron liberation energy (bandgap): 5 eV 2 1.5 1.25 1 eV 0.75 eV 0.5 eV 250 500 750 1000 1500 2000 2500 Sun's wavelengths (in nm) 100% 50% 0% Electron Liberation Electron Butt Kicking For solar cell using material with 0.75 eV electron liberation energy (bandgap):

21 Now for solar cells with a LOT of different bandgaps:5 eV 2 1.5 1.25 1 eV 0.75 eV 0.5 eV 250 500 750 1000 1500 2000 2500 Wavelength in nm (or above, equivalent energy in eV) 100% 50% 0% Sun's AM0 Power Spectrum PERCENTAGE of light's energy captured as electricity by different bandgap materials: Sun's Energy Spectrum: Energy captured by different bandgap cells: (I really should've used Sun's AM1.5 spectrum)

22 Larger the area under a bottom curve => More solar energy capturedBIGGEST area comes between 1 eV and 1.5 eV curves, for ~ 1.3 eV material Material of this bandgap could capture & convert ~ 35% of Sun's energy It's called the Shockley-Queisser Limit after William Shockley & Hans Queisser So fact that single-material solar cells efficiencies top out at 35% is NOT because we are doing a poor job of engineering! It is instead because: We ONLY CAPTURE part of light energy liberating electron from bond, REST of light energy is wasted giving liberated electron kick in the butt All because photons refuse to divide their energy between multiple electrons! (Sure would be nice if we could weasel our way around that!)

23 Si has ALMOST perfect bandgap to reach S-Q Limit!So can approach ~ 30% efficiency. But silicon is also fragile and expensive "TIME OUT! Elsewhere in class you say Si is tough and cheap!" Fragility: It's tougher than OTHER semiconductors, but it is still brittle/breakable Expense: In solar cell want light-liberated electrons/holes to wander a long time So have good chance of wandering into electric field at junction (essential!) But wandering electrons/holes tend to STOP at impurities If both stop there, likely that electrons will fill holes (effectively vanishing) So solar cell grade Si must be about 1000X more pure than electronic grade About 1 part in 1012 pure! => Much more expensive than normal Si So would really like some sort of breakthrough!

24 How? By stacking solar cells of DIFFERENT materials atop one anotherTo beat the Shockley-Quiesser Limit, minimize the electron butt kicking: How? By stacking solar cells of DIFFERENT materials atop one another Materials with different bond energies/"bandgaps" Top: Material with large bond energies ~ purple/blue light: High energy photons use ALMOST ALL of their energy liberating electrons While less energetic photons pass right through! Middle: Material with medium bond energies ~ green light: Medium energy photons use ALMOST ALL of their energy liberating electrons Bottom: Material with low bond energies ~ red light: Low energy photons use ALMOST ALL of their energy liberating electrons ~ ALL photon energy => electron liberation => ~ 100% energy capture!!

25 Called "Multi-Junction" or "Tandem" Solar Cells"BIG REMAINING PROBLEM: Shockley-Quiesser Limit was ~ 33% energy capture efficiency Multi-junction cell could approach ~ 3X (= 100%) energy capture efficiency but requires the combination of ~ 3 different cells to get there So it likely costs (at least) 3X times as much!!!! To beat this must produce ~ three layers of cells for cost of about one layer Very difficult with semiconductors which don't like to grow on one another: Different crystal structures OR different atom spacings OR dope one another Here is where quantum dots (or other tunable nanostructures) might come in!

26 Remember the "Quantum Size Effect" of Lecture 3 and Labs?Happened when trapped waves in box Only waves that survived had to "fit the box" Quantum Dots act as such electron wave containing boxes Smaller dots => Smaller electron standing waves => Higher electron energies Can TUNE quantum dot size to TUNE electron energy And, at least in principle, grow layers of different size dots atop one another Yielding a "multi-junction" Shockley-Quiesser beating Quantum Dot solar cell

27 Let's explore that possibility in smaller steps:Building from earlier example of conventional semiconductor photovoltaic cell: ELECTRIC FIELD from acceptor / donor ions provides sorting & direction of flow OUT Electrical Current Pump IN An analagous one Quantum Dot photovoltaic cell: For instance, when semiconductor dot or drop of organic polymer meets metal: OUT IN But nano dots put out nano power, so want LOTS of them to work together:

28 A multi quantum dot (but not yet "multi-junction") solar cell:Sticking, for the moment, with dots still of the same size: Transparent front conductor Quantum dots Charge separating Electric fields Back Conductor What is going on here? Have specially selected quantum dot and back conductor materials So that interface between them swaps charge, setting up electric field Which propels ONLY photo-generated holes into back conductor Leaving photo-generated electrons to be collected by front conductor Result is MANY nano electron pumps working together ("in parallel"): But this still allows for only ONE very thin light-absorbing QD layer => Very little current

29 An improved multi quantum dot solar cell:THIS would be much better = MORE DOTS! With materials chosen so that: - Blue metal collects only charge from green - Green sucks positive charge from dots - Yellow sucks negative charge from dots - Gray metal collects only current from yellow - Green and yellow self-segregate into such a pillared structure - Dots go to interface Sounds incredibly complex, doesn't it? But such designs ARE being researched Green and Yellow = Immiscible conducting polymers (e.g. "Block co-polymers") But additional quantum dots STILL achieve light to electrical energy conversion of only ~ 9% Compared to polycrystalline Si solar cell efficiencies of ~ 20% Or to single crystal Si or GaAs efficiencies of % !!

30 A REALLY improved quantum dot solar cell:Using Quantum Size Effect + flexibility of quantum dots to give multi-junction design: First capture Blue Light with small quantum dots => few higher voltage electrons Then capture Red Light with deeper large quantum dots => many lower voltage electrons I1, V1 I2, V2 Done right (probably with more layer/sizes), might efficiently capture ALL colors But would require incredible control of internal arrangement and electrical current paths Making the creation of quantum dots the almost trivial part of task REAL CHALLENGE here is required complex 3D SELF-ASSEMBLY (Why I discussed possibility of QD's tethered to DNA scaffolds in last lecture!!)

31 It'd be LOT easier if quantum dots could be randomly distributed in layers!More like this: Yes, but we still need electric fields to separate light-liberated electrons from holes! Here we would NOT want fields between the dots and their surrounding layer: For instance, this field would drive positive holes out of dot, trapping electrons Until dot got so negative that it started pulling holes back Then NOTHING (electrons nor holes) would escape to deliver power! - + - - - - - -

32 You'd instead want electric fields at layer boundariesWhich COULD be accomplished by: 1) Choosing dot and layer materials with similar energy levels So they don't naturally transfer charge / build up interface electric fields 2) Choosing layer materials with differing energy levels To promote charge transfer across their interfaces Thereby engineering properly directed charge-pumping electric fields To produce this alternate QD tandem solar cell: I1, V I2, V2 3D self-assembly => A LOT easier Layer material selection => A LOT more difficult

33 HAVE "multi-junction" and/or Quantum Dot solar cells broken through?Following figure: Compilation by U.S. National Renewable Energy Lab (NREL) of latest, greatest, one of a kind, possibly never reproduced (or horrendously expensive), solar cell efficiency records:

34 Best RESEARCH solar cells (1976 – 2015):

35 At lower resolution but with some guidance as to cell types:Multi-junction / Tandem Single crystal GaAs Single crystal Si Polycrystal thin film Si Other thin films

36 "Hero" (best in lab / single shot) efficiencies, top to bottom:Multi-junction solar cells: Highest at almost 45% So have beat, but not shattered, Shockley-Quiesser Limit Crystalline GaAs solar cells (more exotic/$ crystal than Si): Hair over 34% Crystalline silicon solar cells: Highest at 27.6% Thin-film cells (e.g. polycrystalline/amorphous Si and CdTe): Highest at 23% Perovskite cells: Highest just over 20% Dye-sensitized, organic cells: Highest at 12% Quantum Dot solar cells: Highest at 9.2% NOTE: In minority of cases where affordable commercial versions even exist, they have efficiencies ½ or less than above hero numbers!

37 What happened to Nano's would-be impact upon photovoltaics?Quantum Dot cells - As illustrated by my earlier figures: Difficult to get multiple dots in single layer wired together and working together EXTREMELY difficult to get different layers of different dots to work together So the real (still in the future) breakthrough must be in SELF-ASSEMBLY GRAPHENE – Where did this "wonder material" disappear to? As a career specialist in photon capture I see four themes in graphene research (very often confused and hugely over sold): 1) Graphene's exceptional conductivity could be used for input/output wiring 2) Graphene's exceptional conductivity could be used for transparent top contact 3) Graphene could substitute for top contact AND top semiconductor layer (changing semiconductor "P-N"diode to "Schottky Barrier" diode) 4) Might get single photon to liberate MORE than one graphene bonding electron (Some reports of this, including by former Bell Labs colleague Art Hebard)

38 What exactly MAKES something a good "conductor" or "top contact?""Good conductor" must pass a lot of electrical current easily Electrical current = (#number of free electrons) x (average electron speed) Graphene's electrons travel at exceptionally high speeds So graphene is a good conductor despite having fewer electrons than metals "Good top contact" has to be good conductor AND let most light pass through Light = Oscillating electric (and magnetic) fields When light's electric field strikes conductor, field shifts electrons: Electron shift => polarization => counter electric field If there are enough electrons, counter electric field cancels light field And such free-electron rich materials (metals) end up acting as mirrors But graphene has fewer electrons => poor mirror (i.e. is more transparent!)

39 Input/output wiring (small impact)Combining this with pictorial depiction => Graphene's potential impact: Input/output wiring (small impact) Transparent front contact (medium impact) Top contact AND top semiconductor layer replaced with graphene (medium impact) N (donor doped) semiconductor P (acceptor doped) semiconductor Top layers replaced with graphene AND Get 1 photon to break MULTIPLE bonds (huge impact) The final alternative (only) would completely sidestep Shockley-Quiesser limit: ~ ALL photon energy would go into freeing electrons (= recoverable energy) ~ No photon energy would go into electron kinetic energy (=unrecoverable energy)

40 2) Nano Capacitors: Key to tomorrow's energy storage?What IS a capacitor? A clever way of side-stepping Maxwell's 1st equation: Trying to push more charge into an object doesn’t really work So instead build two closely spaced parallel conducting plates: Overcomes need for local charge balance Excess +’s on top plate don’t like one another But repulsion balanced by attraction to -’s below Becomes more effective the closer the plates are: Capacitance = Area x (dielectric constant in gap) / (gap thickness) = A e / d So where might we come up with some really small gaps? "A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm

41 Possible nano capacitor materials?What about graphite? No, in normal graphite electrons are able to jump between layers But MIGHT prevent by alternating graphene layers with insulating spacer layers What about nested carbon nanotubes? Would again need non-electron conducting spacer layer to hold conducting tubes apart From: Carbon Nanotubes: Potential Benefits and Risks of Nanotechnology in Nuclear Medicine, Journal of Nuclear Medicine 48, 1039 (2007)

42 With some success, such nano capacitors ARE being builtMuch of the effort is going into creating those thin insulating spacer layers Rather than trying to slide insulating layers into place Better strategy is to form them IN PLACE. For instance, by oxidizing surfaces Quantum mechanical tunneling sets limit on minimum spacer thickness: Because (lecture 3): If insulator is too thin, electrons just "tunnel" right through it How thin is "too thin?" Spacer must be > 0.5 – 1 nanometer But capacitance = A e / d So nano capacitors could still have d's ~ 1 nm While conventional macro capacitors use d's of microns or even millimeters So, per unit area, nano capacitors might store 103 to 106 more charge !

43 Important ramifications:When (in previous section) we reduced solar cells to nano size (i.e. to single Q-dots) We only got nano-power out (i.e. power out shrunk with cell area) Forcing us to try to wire huge numbers (1 / nano) of Q-dot solar cells together! But situation is better for nano capacitors: Stored charge shrinks with capacitor area / layer separation Smaller numerator hurts. But smaller denominator helps So, done right, a single nano capacitor could store a LOT of charge per volume ! Making them great candidates for powering nano things But to power MACRO things, would STILL have to wire huge numbers together Bringing us right back to 3D SELF-ASSEMBLY challenge of last section

44 3) Nano Batteries: Chemistry class meets realityThe most important nano-engineered battery is the lithium ion battery Batteries maximize power by using Group I alkali metals: Li, Na, K, Rb, Cs, Fr Alkali metals hold onto their electrons VERY loosely (= least electronegative) Highest voltage batteries then pair alkali metals with highly electronegative atoms or compounds

45 But that is a double edged sword:At right of the periodic table, oxygen is one of the most electronegative elements Alkali metal reacting with oxygen (or water) maximizes energy of electron transfer RESULT: Alkali metals can burn spontaneously in contact with air Some even explode violently if dropped in water Thus Li in lithium batteries is generally provided as a compound, e.g.: LiCoO2 Which is used as the battery's cathode: Ionization/decomposition: n LiCoO2 = (n-m) Li + n CoO2 + m Li+ + m e- For the anode, carbon or silicon are the first choices (for reasons I'll soon explain): Ionization/decomposition: LimC6 (solid) = 6 C (solid) + m Li+ + m e- "A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm

46 Yielding Li ion battery structure and behavior:During CHARGING, Li is actually transferred from inside cathode to inside anode DISCHARGE reverses this: Li transferred from inside the anode to inside cathode: - - - Anode: Li absorbing and deionizing + + Cathode: Li dissolving and ionizing + - - - Anode: Li desorbing and ionizing + + Cathode: Li absorbing and deionizing +

47 Look more closely at action during charging of the Li ion battery:Cathode starts out as a naturally layered chemical compound LiCoO2 : During charging Li must diffuse out: Anode instead starts out as pure solid carbon or silicon: But then, during charging, how the heck does Li get inside! It's made possible by uniquely accommodating crystal structures of C / Si + - + - An Introduction to Sustainable Energy Systems:

48 Carbon and silicon are column IV neighbors in the periodic table:They thus share the tetrahedrally bonded "diamond" crystal structure: Diamond Carbon (bond length = nm ) Silicon (bond length = nm) "Diamond" type of crystal structure has a lot of open space Li is small enough to squeeze into spaces between between Si atoms But Li is less likely to squeeze into spaces between closely spaced C atoms And such true diamond electrodes would be hopelessly expensive anyway! Figures from my "UVA Virtual Lab" website: https://WeCanFigureThisOut.org/VL/Semiconductor_crystals.htm

49 But we know that carbon has a second possible crystal structure:Carbon in its alternate (also cheaper) "graphite" structure, with stacked planes: Li CAN slide between these "graphitic" carbon planes But in some ways, Si anodes are still more attractive Because silicon crystal growth was perfected by the microelectronics industry And crystals are now available in huge sizes (30 cm dia. x meters long) With precut and fully polished wafers costing only ten's of dollars Figures from my "UVA Virtual Lab" website: https://WeCanFigureThisOut.org/VL/Nanocarbon.htm

50 But there is still a problem (or challenge) for Si anodes:To increase Li battery capacity we cram as much Li into the anode as possible But when a LOT of Li slithers into the spaces, the Si crystal actually expands With enough added Li, silicon expands by 2-3 times, actually changing its structure: Top from my "UVA Virtual Lab" website: https://WeCanFigureThisOut.org/VL/Semiconductor_crystals.htm Bottom:

51 When Li ion battery discharges, silicon anode shrinks and reorders:Or at least it may do that the first few (or few hundred) times But during charging it's likely that Li is not added uniformly to the Si And during discharging it's likely that Li is not removed uniformly Resulting in non-uniform expansion and contraction of the silicon => Huge non-uniform stress across the crystal => Eventual development of cracks and fractures With these cracks/fractures, as silicon shrinks upon battery discharge: Si pieces separate => Electrical contact between pieces is lost => Shrinking effective anode size & capacity

52 A solution is provided by nanoscale self-assembly:On Si wafer, lay down nanopattern of metal, heat to melt, then expose to SiH4 vapor: <= SiH4 approaching one of a vast array of now molten metal dots SiH4 decomposes, releasing Si to dissolve into the molten metal dot Si diffuses down to wafer where it solidifies creating a growing column of new Si: "A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm

53 With results such as these:Lorelle Mansfield -NIST: U. Helsinki: Small size + accessibility => Uniform Li absorption, even stress, minimal Si cracking: New nano-structured Li ion battery anodes Designing nanostructured Si anodes for high energy lithium batteries, Wu & Cui, Nano Today 7, pp (2012)

54 Nanostructures are also being investigated for Fuel Cells:For example, ones such as this to power automobiles Fueled by hydrogen gas + oxygen gas Producing only water vapor as exhaust: Incoming H2 gas High surface area (or nano-porous) catalyst Incoming O2 gas (in air) H H + - O Electrolyte capable of passing H+ ions: Aqueous OR Solid Solution OR Proton permeable membrane Outgoing water

55 Batteries and Fuel Cells are "kissing cousins"Battery: Pair of "redox" materials comes from the electrodes themselves Fuel Cell: Redox materials come from external fuels And electrodes themselves are not consumed however: They often act as catalysts => Need for exotic / $$$ materials E.G. Platinum = 36,667 $/kg (27 March 2015) Good News: Quantity of catalyst can be vastly reduced w/ high surface area nanostructures Bad News: Fuel cell energy return ~ 40% vs. batteries energy return ~ 80% For more on fuel cells / nanostructured fuel cells see my other class's lecture: Batteries and Fuel Cells

56 But BOTH batteries and fuel cells have BIG weight + size problemsWith critical ramification that for transportation (especially air transportation!): Fossil fuels are going to be REALLY HARD to replace: Comparison of various energy storage technologies (energy / mass vs. gasoline): Conventional Battery = x Gasoline's energy density PC Battery = x Gasoline's energy density TNT (0.65 Cal/gm) = 1/15 x Gasoline's energy density Butyl alcohol = 0.9 x Gasoline's energy density Kerosene / JP-A / Jet Fuel = x Gasoline's energy density Gasoline (9.75 Cal /gm) = 1 x Gasoline's energy density Natural Gas (liquid) = 1.3 x Gasoline's energy density Hydrogen (liquid) = 2.6 x Gasoline's energy density (data mostly from Richard A. Muller's book "Physics for Future Presidents")

57 Conclusions Discussed: Nanostructured Photovoltaics Nano CapacitorsNano batteries (and very similar nano electrochemical cells) These represent only a small sample of the “bleeding edge” of nano energy But, as in Bleeding Edge: Part I, range of potential applications is already stunning Nevertheless, this lecture also highlighted a recurring nanotech issue: Making the nano things was once again often the "easy" part But getting them organized often ends up limiting technological application "A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm

58 Credits / AcknowledgementsFunding for this class was obtained from the National Science Foundation (under their Nanoscience Undergraduate Education program). This set of notes was authored by John C. Bean who also created all figures not explicitly credited above. Copyright John C. Bean (2017) (However, permission is granted for use by individual instructors in non-profit academic institutions) "A Hands-on Introduction to Nanoscience: WeCanFigureThisOut.org/NANO/Nano_home.htm