1 Next Generation Solar PowerIn earlier lectures I covered present day power technologies Including recent developments and new directions However, two of those technologies may actually undergo discontinuous change: Solar might be transformed by different materials and designs Nuclear might become much safer & less expensive Either development could be an absolute game changer! Thus: Next Gen Solar is the topic of this lecture, with Next Gen Nuclear to follow An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

2 A caution about today's Next Gen Solar discussionExperience tells me that many of you are particularly interested in solar power But I know that understanding its many forms can be REALLY FRUSTRATING (!) However, my field IS "Optoelectronics" (= light  electronics conversion) And I have taught that field for all of my academic half-career So today I will attempt to explain almost ALL leading-edge photovoltaics research But via words and drawings I think you CAN absorb Which will hopefully supply you with a GATEWAY into this field That field is the subject of the following "Best Research Cell Efficiencies" chart Released annually by the U.S. National Renewable Energy Lab (NREL) Around which I will organize this lecture Exploring each of the solar cells types it reports

3 The full (high resolution version) of the NREL chart:This file is HUGE, starting out more than four times the size of a PowerPoint slide So I will only use the hi-res version here (and you may have come back for details)

4 Its efficiency vs. year data fall into groups by cell type:Multi-junction / Tandem Single crystal GaAs Single crystal Si Polycrystal thin films Amorphous Si thin film "Emerging PV" I'll explore all of these groups in the course of this lecture

5 But first we need to recall, and extend, our photovoltaic basics:(link to my earlier Solar Power lecture)1 Photovoltaic's ambitious goal is to efficiently convert light of ALL different colors 1 eV 0.5 eV 2 eV 5 eV Energy Wavelength UV Infrared Vis 1)

6 But different colors interact very differently with matter:Infrared Visible UV Liberates electrons from bonds IFF the material has an electron liberation energy ≤ light energy Liberates electrons from bonds but gives them so much excess energy that they ricochet all around Excess energy is lost as struck atoms start to vibrate (=heat) Vibrates a few atoms But most light just passes right through

7 But liberation of electrons & creation of holes is not enough!Liberated & created, they have no reason to move in ANY particular direction Term for what they will do is diffuse = Randomly wander Diffusion is analogous to "the drunkard's walk" Drunk randomly bouncing off light posts: But we've got two types of wandering drunks: electrons and holes And these unique types of drunks can "annihilate" one another That is, the electron can just fill the hole (in the bond) => zip! To create a net flow (=> electricity) a photovoltaic cell MUST include a cliff Over which the "electron drunks" will fall (and the "hole drunks" ascend) Or the way I was first taught it: Electrons ~ Ball bearings (which fall DOWN) Holes ~ bubbles (which FLOAT up) Figure:

8 Solar cell's selective "cliff" is provided by an electric fieldWhich WE do NOT have to create by applying an external battery or power supply Instead, electric fields are naturally generated at the junction between two materials But only if electrons choose to shift from one material to the another OCCURS ONLY WHEN: Material at right has electrons in energy levels higher than empty levels in material at left => Electron finds new lower energy home! This negative charge displacement ALSO creates an electric field

9 Which, via careful materials selection, gets us to this point:Sunlight in Electric Field Light-liberated electron (and bond hole) created Electric field sorts things out, pushing holes left and electrons right => "electricity" But only if drunkard's walk of light-liberated holes & electrons GETS them to the electric field BEFORE they have recombined with one another! In they recombine first, their energy is instead lost as heat (a.k.a. vibrating atoms) So why use a thick layer that puts the field (cliff) so far away? (!) Because light is not all absorbed (and converted to electrons + holes) at the surface Strong light absorption requires tenths to tens of microns of material

10 So with necessarily thick layer, situation is more like this :Sunlight in Electric Field Light-liberated electrons (and bond holes) created It is then REALLY IMPORTANT that freed electrons & holes do NOT recombine before reaching the sorting electric field at the junction between the layers Their average random walking survival distance is called their diffusion length Diffusion length depends on the structure and purity of the material because: Electrons & holes are drawn to impurities and crystalline flaws Once together at such TRAPS, electrons & holes can easily recombine So the purity and perfection of PV crystals can be extremely important!

11 Returning to the NREL figure and its charter members: Si and GaAsSi single crystal ~ 28% efficiency GaAs single crystal ~ 35% efficiency Multi-junction / Tandem Single crystal GaAs Single crystal Si Polycrystal thin films Amorphous Si thin film "Emerging PV"

12 The crystallography of these most basic semiconductors:Silicon has same diamond structure as carbon (also column IV in periodic table): GaAs has a zincblende crystal structure which is really almost identical, except: Which is based on every atom having 4 symmetrically oriented bonding electrons (and neighbors): Here column III atoms bond to column V atoms Each PAIR has 8 bond electrons = Average of 4 per atom So it forms the SAME BONDING STRUCTURE as Si Trick ALSO works with paired column II and VI atoms! From my own UVA Virtual Lab website, here:

13 That makes these close cousin, plain vanilla, solar cellsThey are the oldest solar cells types, with strong further improvement unlikely GaAs is slightly closer to the ideal Shockley-Quiesser single material bandgap Which is why GaAs single crystal solar cells are ~5% more efficient Si raw material is more common/cheaper, and non-toxic Nevertheless, GaAs is a strongly bonded (very stable) chemical compound and thus significantly less toxic then pure arsenic Both require energy intensive, high temperature single crystal growth Further, the best Si solar cells require particularly low concentrations of "traps" Si raw material thus requires extra, energy intensive, purification These ARE the current photovoltaic gold standards! (with silicon's lower cost outweighing its slightly lower efficiencies)

14 But solar cells can also be made out of non-single crystalsPolycrystalline cells of various materials (including Si): ~ 20% efficiency Amorphous (non-crystalline) Si cells ~ 13% Multi-junction / Tandem Single crystal GaAs Single crystal Si Polycrystal thin films Amorphous Si thin film "Emerging PV"

15 Why do thin film efficiencies (of same materials) drop to 15-20%?Multi-crystalline: Poly/microcrystalline: Amorphous: Nature will generally find SOME way to complete depicted incomplete bonds But the result will often be distorted or incomplete bonds Excellent for attracting, or even binding, electrons or holes Producing a dense gauntlet of traps for our wandering electrons/holes to get past! Sunlight in Electric Field Light-liberated electrons (and bond holes) created

16 The irony of strong vs. mushy semiconductors:Silicon is a VERY strong, stable, high melting point, semiconductor Surpassed only its Column IV cousin diamond carbon, and their hybrid SiC However, this means that its atoms are NOT going to easily move around And it's very unlikely that faults in silicon crystals will self-heal So, from the outset, Si PV crystals must be extremely pure and perfect! In less stable, low melting point semiconductors, atoms CAN move around Thus (at least to some extent) these materials CAN self-heal And much poorer starting crystals (or polycrystals) often work well for PV But there is also a dark side: Low T semiconductors ARE less stable So their PV cells often readily degrade (accelerated by UV sunlight) Further, breaking down, they can release their often toxic components

17 Examples of mushy, surprisingly successful, thin film materials:For which energy intensive crystal growth can often be replaced by: Just spraying them onto metal backing sheets Or even printing them on as liquid or inkjet inks - Polycrystalline silicon (the old hand) - CdTe = column II + column VI compound Mimicking GaAs zincblende crystal, mimicking Si/C diamond bonding - CIGS = Copper Indium Gallium Selenide = CuInxGa1-xSe2 Cu takes one site, either In or Ga take the next, and Se the last:

18 CIGS crystal retains tetrahedra of diamond & zincblende crystalsThat simple strong geometry likely improves its general quality But especially useful: Purple sites can be either In or Ga Like earlier example of AlGaAs, it is an alloy of two compounds The two (fixed stoichiometry) compounds are: CuInSe2 with an electron liberation energy (bandgap) of 1.0 eV, and CuGaSe2 with an electron liberation energy (bandgap) of 1.7 eV Thus for CuInxGa(1-x)Se2, as x changes from 0 to 1, CIGS bandgap goes 1 to 1.7 eV So can tune one CIGS cell to maximum Shockley-Quiesser single material efficiency OR build multi-junction cell out of of different composition CIGS layers! An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

19 There is also a variant on CIGS called CZTSWhich stands for copper zinc tin sulfide = Cu2ZnSnS4 (vs. CIGS: CuInxGa1-xSe2) Having crystal structure: Versus CIGS crystal structure: With color coding of Cu, Zn, Sn, S Versus Cu, In or Ga, Se While not identical, they are very, very similar, as are their solar cell properties The big difference is that CZTS uses more abundant, less toxic elements And thus should be more environmentally friendly And could end up being cheaper But at this point CIGS cells are still ~ 2X more efficient (21.7% vs. 11.9%) CZTS figure:

20 But a particularly enigmatic thin film is amorphous Si:I depicted its amorphous structure as this: Those many screwed up or incomplete bonds can act as electron/hole traps: So how would ANY electron/holes EVER make it to the junction E-field? Solution (co-invented by RCA Labs friend) = It's not really amorphous Si! It's amorphous silicon hydride = Amorphous Si stuffed with hydrogen Tiny hydrogen atoms easily migrate into the structure And THEY end up attaching to any of the broken bonds above Effectively nullifying site's effectiveness as a trap Although what goes in can come back out => Possible lifetime/stability problems An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

21 Which prepares us to take a look at the top of the chart:Multi-junction / Tandem cells of various materials with efficiencies up to ~ 45% Multi-junction / Tandem Single crystal GaAs Single crystal Si Polycrystal thin films Amorphous Si thin film "Emerging PV"

22 These "multi-junction/tandem" cells look like the champs:They have the highest efficiencies! And their efficiencies are still climbing rapidly! But caution is in order: This chart does NOT deal with practicality This chart does NOT deal with cost It reports best, one of a kind, not yet reproduced, laboratory efficiencies ONLY AND, at present, it only extends up to 50% efficiency With multi-Junction / tandem cells still only producing 35-45% efficiencies AND these are THE most complex cells making practicality & cost BIG issues!

23 Recall that they are designed to overcome the Shockley-Quiesser Limit:Multi-Junction / Tandem cells = Stacking solar cells of different materials 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!!

24 But as noted in my first lecture on solar lecture energy:Multi-junction solar cells have not yet had any real impact. Why? Shortcoming #1) Efficiency Single-material / single-junction designs have 20-35% efficiency Multi-junction cells ought to be able to double or triple that efficiency But to date, at best, they increase efficiency ~ 1.5X (even if they ARE still heading upward!) Shortcoming #2) Complexity - which will have large impact on cost: Efficiency MUST increase MORE than the number of junctions: 3X efficiency + 3 junctions likely => MORE cost per watt out Instead, we need a way of making 2-3 junctions at ≤ 1.5X the cost An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

25 Ways of easily making MANY layers of cells/junctions?Create all cell layers using ~ one process and/or one apparatus Method #1) Use cells having almost identical crystal structures Materials above one another in this plot => Tightly combined cells of lattice constant vs. bandgap: Created in same or similar apparatus Silicon Based Semiconductor Heterostructures – Column IV Bandgap Engineering, John C. Bean, Proceedings of the IEEE, pp (April 1992)

26 Classic example: AlAs and GaAs: Each is a chemical compound= Having a strict stochiometery (here 1:1) But you can make alloy mixture of the two: AlxGa1-xAs = (fraction x of AlAs) + (fraction 1-x of GaAs) Atomistically: As atoms go on yellow sites, either Al or Ga atoms go on blue sites You can then build a tandem cell with say: Top cell of Al0.6Ga0.4As (targeting blue / UV light) Bottom cell of GaAs (targeting red / IR light) Or other alloys, or even alloys of alloys! From my own UVA Virtual Lab website, here:

27 Non or partially crystalline thin films might also be usedBut a newer possibility is Quantum Dot multi-junction/tandem solar cells Quantum Mechanics concluded that electrons act like waves But all types of wave act ~ the same Imagine making big water waves in a bowl: Only certain size waves persist Because wave of a given wavelength will bounce back and forth And to ADD to one another, the wave AND its reflections must be "in phase" For that to occur: The wave must be sized so that it "fits the box" To fit, distance out and back across bowl must = multiple of wavelength (l) Or, for a given size of "bowl" (L): Strong wavelengths = 2L / integer Acceptable wavelengths => Acceptable energies With energy of waves changing with SIZE of enclosure!

28 Because electrons ARE waves, same holds for them:Smaller "bowl" (quantum dot) => Smaller fitting wavelengths Smaller wavelength = More frequent wave oscillation => Higher Energy Hence smaller QUANTUM DOTs => Higher electron energy levels => Higher differences between those energy levels Difference between two of those levels = electron liberation energy (bandgap) So different quantum dot sizes absorb light at different energies (for energies we are interested in, dots must be nanometers in size) Different sized dots in different layers => Wavelength selective layers Offering a new type of Shockley-Quiesser beating multi-junction solar cell What are "quantum dots" made of? All sorts of different atoms & compounds!

29 Let's explore quantum dot possibility in smaller steps:Building from example of conventional semiconductor photovoltaic cell: Swap of charge across junction => ELECTRIC FIELD => Sorting of electrons/holes OUT Electrical Current Pump IN An analogous one Quantum Dot ("QD") photovoltaic cell: For instance, when semiconductor or polymer quantum dot meets metal: OUT IN But nano dots put out nano power, so we need LOTS of them to work together:

30 A quantum dot (but not yet "multi-junction") solar cell:Sticking, for the moment, with dots of the same size: Transparent front conductor Quantum dots Charge separating Electric fields Back Conductor What is going on here? You choose special 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 uses only ONE thin light-absorbing QD layer => Very little current

31 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% Putting it in low "Emerging PV" corner of the NREL chart:

32 REAL CHALLENGE here is the required complex 3D SELF-ASSEMBLYTo REALLY improve things we need multi-junction quantum dot solar cells: Using Quantum Size Effect AND 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), you might efficiently capture light of ALL colors But this 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 the required complex 3D SELF-ASSEMBLY

33 I'd be LOT easier if quantum dots could be randomly distributed in layers!More like this: But we'd still need electric fields to separate light-liberated electrons from holes! And HERE we would NOT want fields between the dots and their surrounding layer: For instance, field below 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! - +

34 You'd instead want electric fields at layer boundariesWhich COULD be accomplished by: 1) Choosing dot and layer materials that have similar energy levels So they don't naturally swap charge / build up interface electric fields 2) But choosing layer materials with differing energy levels So that they do transfer charge across their interfaces Thereby adding properly directed charge-pumping electric fields Producing this Quantum dot multi-junction / tandem solar cell: I1, V I2, V2 But full multi-junction quantum dot cells have NOT yet been realized! 3D self-assembly => A LOT easier Layer material selection => More difficult

35 But before leaving multi-junctions, what's NREL talking about here?Higher efficiencies for "concentrator"cells - what's this all about? It refers to concentrating sunlight on SMALL cells via LARGE lenses or mirrors:

36 Why would concentration be desirable?Reason #1) Photovoltaic cells get more efficient as light gets more intense Revising earlier figure to explicitly depict "traps" created by impurities/flaws: Even the most crystalline / highest purity PV materials have some "traps" ( ) These reduce the "diffusion length" of wandering holes and electrons Reducing number reaching to and sorted/pumped by electric field But there are limited number of "traps" interfering with wandering electrons/holes So overwhelm them by sending in more intense light => More electrons + holes => These quickly saturate (fill up) the traps => Allowing other electrons/holes to pass by! Sunlight in Electric Field Light-liberated electrons (and bond holes) created

37 Reason #2 for concentration is easier to understand:If the cost of a square meter of solar cell >> Cost of same size lens/mirror: (which is particularly likely for complex multi-junction solar cells) It's cheaper to buy large lens/mirror and combine with small solar cells! You still use ALL of the light captured by the full size lens/mirror AND You get further benefit of Reason #1 improvement in basic cell operation Thus "concentrator cells" at right can deliver same or higher power out! Lens Mirror vs. or

38 This finally gets us to NREL's "Emerging PV" category:These are new, long-shot possibilities, some of which are improving rapidly: Multi-junction / Tandem Single crystal GaAs Single crystal Si Polycrystal thin films Amorphous Si thin film "Emerging PV"

39 To make this busy corner clearer, let me enlarge it (and its key):I have already described: - Basic Quantum Dot cells - CZTS variant of CZTSSe Leaving us only:

40 THESE "emerging PVs" which can be sorted two different ways:By years of research (more mature 1st): Or efficiency (most efficient 1st): - Dye sensitized cells - Perovskite cells - Organic cells - Dye sensitized cells - Perovskite cells - Organic cells To those I want to add cells using Carbon Nanotubes (CNTs) and Graphene Because these are getting so much recent press attention! Despite the fact that they are "no-shows" on the NREL chart Onward: Newer designs/materials tend to build on the older And research papers assume you know the earlier research (and are very hard to comprehend if you don't!) So let's tackle these final types in ~ historical order:

41 Dye Sensitized Solar Cells (DSSC's):All of the preceding cells come from the physics & electrical engineering world Known as the "Device Physics Community" Of which, yes, I am a card-carrying member: DSSC's were instead developed by chemists (or even the occasional biologist) The subject matter is fundamentally the same But the terminology is fiercely different, as is the approach: - Physicists focus on atomic scale mechanisms (getting vague at larger scales) - Chemists focus on macroscopic properties (getting vague at the atomic scale) So what follows is my attempt at both bridge-building AND translation An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

42 Chem speak (vs. Physics speak):Papers on DSSC's (and the organic cells to follow) talk about three things: 1) Dyes or photo-sensitizers where light produces unbonded electrons & holes Using my 1st solar lecture key of = atomic cores, = electrons: So this part is easy to translate: It's a semiconductor 2) Hole transport materials (or layers) where only holes can move An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

43 Only holes? But weren't electrons really moving?Yes, but the subtle distinction is as follows: A single hole move = valence electron slithering from one bond to an adjacent one: And it really doesn't have to jump as it's actually a quantum mechanical cloud Which can sort of just ooze from one bond to another: This requires very little additional energy - Physics speak: This electron remains in the "valence band" - Chem speak: It's in the "HOMO" = Highest (normally) Occupied Molecular Orbital An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

44 As opposed to the movement of a truly unbonded electron:This electron cannot just ooze from one side of an atomic core to the other: Because it's outside the atomic core (which is already packed w/ electrons) So it's at a much higher energy - Physicist: it's in the "conduction band" Chemist: It's in the "LUMO" = Lowest (normally) Unoccupied Molecular Orbital Which, also in Chem-speak, makes the material above an: 3) Electron Transport Material (or layer) In their preferred liquid environment, Chemists can simplify things further: Electron Transport Material (wet) => Solution with negative ions (anions) Hole Transport Material (wet) => Solution with positive ions (cations)

45 All three things happen in the classic solar cell semiconductorWhy wouldn't all three things happen in a single material? 1) If bonds are too strong (the "bandgap" is too large), light won't be absorbed, => Electrons will not be liberated from bonds 2) If bonds are too strong, any loose electrons won't stay liberated, => Killing off (unbonded) electron transport 3) If material's bonds are oriented or separated differently, result might be that Quantum mechanical "oozing" of electrons between bonds is far less likely => Killing off valence band hole transport Chemists like to divide the three functions between different materials An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

46 Chemist's DSSC (or organic) solar cell:Which divides three functions between different layers Actual geometry is more complex: With light coming in through either Electron or Hole transport layer And instead of flat planes, they use much more convoluted 3D geometry This seems to do the required job of sorting electrons from holes Sending them in opposite directions to form an "electrical current" Sunlight in Thin Dye layer Thick Electron Transport layer Thick Hole Transport layer An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

47 But hold it: There is no "push" => No power!Remember? Electrical Power = Flow x Pressure = Current (I) x Voltage (V) Above scheme shows no way of producing the pressure/voltage: It's not yet a pump! It's just bucket leaking electrons to left / holes to right: To be a bit more technical: While it's "short circuit current" (unopposed flow) might be finite It's "open circuit voltage" (electrical force) would be ~ zero Thus, while papers on DSSC's and Organic Solar Cells seldom mention it, There has got to be something more! An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

48 There must be at least one pushing Electric Field:Sunlight in Thin Dye layer Thick Electron Transport layer Thick Hole Transport layer Interfacial Electric Fields THUS material of "hole transport layer" CANNOT ONLY transport holes Line up of its energy levels vs. those of the the Dye layer must promote electron transfer across that interface (=> E field) AND/OR Material of "hole transport layer" CANNOT ONLY transport electrons Line up of its energy levels vs. those of the Dye layer An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

49 Now (finally) moving on to actual DSSC configurations:Geometry: Corresponding energy levels: Dye/Sensitizer coats surfaces of TiO2 particles Electrons photo-generated in the dye move into the TiO2 electron transport layer Holes in dye layer are then filled by electrons from I- ions arriving in the electrolyte By absorbing holes, this serves the role of a hole transport layer (with I- ions then regenerated at far right Pt electrode) Energy levels are indeed carefully chosen to promote interfacial electron transfer Thereby supplying (even if not mentioned) the pushing electric fields Highly Efficient Dye-Sensitized Solar Cells: Progress and Future Challenges, Y. Zhang et al., Energy and Environmental Science 6, pp (2013)

50 Or getting rid of liquids and moving to all solid state:Perovskite Dye on Al2O3 particles (taking electron transport place of previous TiO2) (HTM = hole transport material, TCO = top contact transparent oxide) THICK Perovskite Dye (= better absorbing) on TiO2 particles (electron transporter) The Light and Shade of Perovskite Solar Cells, Michael Grätzel, Nature Materials 13, pp (November 2014)

51 Where I just snuck in perovskite solar cellsWhich have been rocketing up the NREL chart: Attributed to BOTH electrons and holes moving REALLY well in perovskites: LONG electron/hole diffusion lengths = very few and/or ineffective traps Perovskite = "Mushy" less stable but (remarkably) self-healing semiconductor

52 Perovskite family structure and formula:Take as the basic repeating crystal building block ("unit cell"): The CUBE with fat gray B's in its corners A = Organic or inorganic cations (+) = green: 1 per cell B = Metal cations (+) = gray: 1 per cell 8 corner atoms x (1/8 of each inside our cube) = 1 inside each unit cell X = Anion (-), frequently halide = purple: 3 per cell Don't count any atoms outside of the gray atom bounded cube Every gray cube edge (12 of them) has a purple X atom at its center But only ¼ of each of those purple X atoms is inside our cube (i.e. ) Thus X count = 12 edge atoms x (1/4 of each inside or cube) = 1 So general formula of repeating perovskite unit cell is: ABX3 Figure: The Light and Shade of Perovskite Solar Cells, Michael Grätzel, Nature Materials 13, pp (November 2014)

53 For perovskites used in solar cells:ABX3 is made up of: A organic cation = Methylammonium (CH3NH3+) B metal cation = Lead (Pb+) X anion = Iodine (I-) and/or Bromine (Br-) and/or Chlorine (Cl-) Synthesis DOES NOT REQUIRE energy intensive high temperature crystal growth! Perovskites can instead be evaporated onto surfaces OR Deposited from liquid using PbX2 and CH3NH3X dissolved in solvents Offering: 1) Simple (almost trivial) fabrication technology 2) Excellent performance due to long free electron / hole diffusion lengths 3) Combined or separated absorption + hole & electron transport by layer

54 Which makes perovskites sound like the PV silver bulletBut the "fly in the ointment" is: - It is built around toxic lead - Which, because things go together so easily, also come back apart! Specifically: In contact with water, lead iodide perovskite releases PbI Which is a known carcinogen Use of which is banned in many countries 1 This has fueled intensive research on alternate perovskites Including those which replace Pb with Sn But these have not yet produced comparable performance Nor do they achieve comparable cell stability/lifetimes So perovskite PV is a "Stay tuned for further developments!" sort of story 1The Light and Shade of Perovskite Solar Cells, Michael Grätzel, Nature Materials 13, pp (November 2014)

55 Bringing us to organic solar cellsSimilar to cells above but instead made from layers of organic chemicals They work, they can be very cheap to build, they can literally be flexible But you know what happens if organics are left in sunlight (e.g. rubbers & plastics) They fade, crack, and eventually crumble - Why? - Answer: UV not only liberates electrons from C-C bonds, breaks those bonds But some TOUGH ORGANICS have been getting a lot of press attention: Graphene: Carbon Nanotube (CNT) = Rolled up Graphene But they are NOWHERE on NREL's Best Research Cell Efficiencies chart So why all of the hype? What is really going on?

56 To explain I need to discuss the complete structure of a solar cell:Which would have to include at least these parts: But likely with these final proportions and arrangement: Sun Electric Field Light-liberated electrons & holes Metallic Contact Layer Transparent Metallic Wire An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm

57 In the full structure, where could Graphene/Nanotubes be used ?Well, what are they really good at / good for? 1) They are really, really strong – but that's largely irrelevant to solar cells 2) Electrons can move through them really, really easily and quickly Which is to say that they are superb conductors So they COULD be used in: Top wire, bottom metallic contact, bottom wire And this IS (frequently) suggested in the press and by researchers But in those places standard metals already work pretty well And substitution of graphene/CNTs => Minimal impact on efficiency Top contact offers a somewhat bigger opportunity for improvement: IT must be both a good conductor and largely transparent THAT is a rare and scientifically challenging combination:

58 A "good conductor" must pass a lot of electrical current easilyElectrical current = (#number of free electrons) x (average electron speed) Graphene/CNT electrons can travel at exceptionally high speeds So graphene is a good conductor despite having fewer electrons than metals A "transparent material" has to allow light (here sunlight) to 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. it's more transparent!)

59 So graphene/CNT's could help out in transparent front conductorThis could trim the cost a little Because alternatives such as Indium Tin Oxide (ITO) are costly And/or it might slightly goose up efficiencies (say, a few percent) Which would certainly be nice, but which would not yet be a huge deal To be a huge deal, graphene/CNT's would have get out of the boonies of the cell! AND instead enable big improvement in the function of the: Semiconductor/dye, electron or hole transport layers But graphene is NOT a natural semiconductor, it's effectively a metal meaning that: Electron liberation energy ~ 0 => No pushing force (Voltage) from layer(s) However, SOME forms of CNT are semiconductors so there is a possibility here

60 But ONLY possibility seeming to justify current level of excitement is:If, in the heart of the cell, CNT's (or modified form of graphene) could use excess photon energy to liberate additional electrons Instead, now, one photon gives all of its energy to only one electron IF that photon's energy is MORE than enough to liberate that electron excess energy goes into electron's kinetic energy and it just ricochets around, causing atoms to vibrate, sucking up that energy as waste heat = Fundamental reason ALL solar cell efficiencies are << 100% In graphene, some believe they've seen 1 energetic photon liberating 2 electrons HOWEVER, reports have been rare and that result very hard to achieve Leading to my personal conclusion that: Current press on graphene/CNT solar PV seems very premature and/or naive

61 Credits / AcknowledgementsSome materials used in this class were developed under a National Science Foundation "Research Initiation Grant in Engineering Education" (RIGEE). Other materials, including the "Virtual Lab" science education website, were developed under even earlier NSF "Course, Curriculum and Laboratory Improvement" (CCLI) and "Nanoscience Undergraduate Education" (NUE) awards. This set of notes was authored by John C. Bean who also created all figures not explicitly credited above. Copyright John C. Bean (However, permission is granted for use by individual instructors in non-profit academic institutions) An Introduction to Sustainable Energy Systems: WeCanFigureThisOut.org/ENERGY/Energy_home.htm