Energy is stored in chemistry
Nearly all the energy people use passes through a chemical reaction. Fossil fuels, which still supply about four-fifths of the world's primary energy, are sunlight stored as chemical bonds by photosynthesis millions of years ago and released by combustion. Food is the same, on a shorter timescale. Even the electricity from a wind turbine or a solar panel must be stored somewhere when the wind drops or the sun sets, and the best stores we have are chemical: ions shuttling inside a battery, or hydrogen made from water.
That is why the energy transition is, to a large extent, a chemistry problem. A fuel is valuable for how much energy it releases per kilogram and per litre, and for what it leaves behind; a battery for how many electrons each gram of electrode can store and at what voltage; a solar cell for which photons its semiconductor can use. This lesson works out those numbers from the thermochemistry of Thermochemistry, the free energies of Entropy and free energy and the cells and potentials of Electrochemistry, and then asks what limits them in practice.
Fuels and combustion
A fuel and the oxygen of the air together store chemical energy, released when weak bonds (such as the O=O bond of O₂ and C–H bonds) are replaced by the very strong bonds of CO₂ and water. Thermochemistry computed energies per gram; for an engineer, three numbers matter: energy per kilogram (for anything that moves), energy per litre (for the size of the tank) and CO₂ released per unit of energy (for the climate).
# Fuels compared: energy per kilogram and per litre, and CO2 per unit of energy.
# Standard enthalpies of combustion at 298 K, water as liquid ("higher heating value"),
# from the NIST WebBook (for hydrogen, methane and ammonia, the values for the gas).
fuels = {
# name ΔcH° kJ/mol M g/mol C atoms density kg/L (as stored)
"hydrogen, 700 bar": (-285.8, 2.016, 0, 0.040),
"hydrogen, liquid (20 K)": (-285.8, 2.016, 0, 0.071),
"methane, liquefied (LNG)": (-890.6, 16.04, 1, 0.42),
"octane (petrol)": (-5470.0, 114.23, 8, 0.703),
"methanol": (-726.1, 32.04, 1, 0.792),
"ammonia, liquid": (-382.8, 17.03, 0, 0.682),
"graphite (stand-in for coal)": (-393.5, 12.011, 1, None),
}
print(f"{'fuel':29s} {'MJ/kg':>6s} {'MJ/L':>6s} {'g CO2/MJ':>9s}")
for name, (dH, M, nC, rho) in fuels.items():
mj_per_kg = -dH / M # kJ/g = MJ/kg
per_litre = f"{mj_per_kg * rho:6.1f}" if rho else " -"
co2 = nC * 44.01 / (-dH / 1000) # g CO2 per MJ released
print(f"{name:29s} {mj_per_kg:6.1f} {per_litre} {co2:9.1f}")
# For comparison: a good lithium-ion cell stores about 250 Wh/kg and 650 Wh/L
print(f"{'lithium-ion cell':29s} {250 * 3600 / 1e6:6.2f} {650 * 3600 / 1e6:6.2f} {'none at use':>11s}")Hydrogen has by far the most energy per kilogram but very little per litre, even compressed to 700 bar or liquefied at 20 K; liquid hydrocarbons such as petrol pack the most into a tank. Per unit of energy, burning methane releases about 49 g of CO₂ per MJ, petrol 64 g and carbon (a stand-in for coal) 112 g, so per unit of heat, gas releases less than half the CO₂ of coal, and gas power stations are also more efficient. Hydrogen and ammonia release none when burned, but they are only as clean as the way they were made.
Combustion is a radical chain reaction (Chemical kinetics), and its details decide what comes out of the exhaust. In a petrol engine, the unburned mixture can ignite by itself before the flame front arrives, a destructive "knock"; branched alkanes resist this better than straight chains, and a fuel's octane number compares it with 2,2,4-trimethylpentane (100) and heptane (0). At flame temperatures, nitrogen and oxygen from the air combine to nitrogen oxides; incomplete combustion makes carbon monoxide and soot; sulfur in the fuel becomes SO₂ (Environmental chemistry). Lower-carbon liquid fuels include ethanol fermented from sugar cane (widely used in Brazil) and other crops, biodiesel made from vegetable oils and waste fats, and synthetic "e-fuels" made from hydrogen and captured CO₂, which are chemically sound but currently expensive in energy. Converting heat from combustion into work is limited by thermodynamics, as Lumenary ME's power plants and energy conversion explains.
Batteries: lithium-ion, sodium-ion and solid-state
A lithium-ion cell stores energy as lithium in two different solids. In the charged cell, lithium sits between the carbon layers of graphite at the negative electrode. On discharge, each lithium atom gives up an electron, which flows through the external circuit, and travels as Li⁺ through the electrolyte to the positive electrode, a layered metal oxide or a phosphate, where it slips back into the crystal and a metal ion is reduced (Co⁴⁺ to Co³⁺, for example). Charging pushes it back. Because lithium ions rock between two host structures without either being dissolved or rebuilt, the cell can be cycled thousands of times. M. Stanley Whittingham, John Goodenough and Akira Yoshino shared the 2019 Nobel Prize in Chemistry for the steps that made it work, and Sony sold the first commercial cells in 1991.
How much charge can an electrode hold? If each formula unit of a material of molar mass (g/mol) can take up electrons, its theoretical specific capacity is in mAh/g, where is the Faraday constant and 3.6 converts coulombs per gram to milliampere-hours per gram. Multiplying capacity by the cell's average voltage gives energy. The block computes capacities for the main electrode materials, then the specific energy of a few electrode pairs, counting only the active materials.
F = 96485.0 # C/mol
def capacity(electrons, molar_mass):
"""Theoretical specific capacity in mAh/g: n F / (3.6 M)."""
return electrons * F / (3.6 * molar_mass)
materials = {
# name electrons per formula unit, g/mol of host
"LiCoO2 (all Li removed)": (1.0, 97.87),
"LiCoO2 (half the Li, as used)": (0.5, 97.87),
"LiNi0.8Mn0.1Co0.1O2 (NMC811)": (1.0, 97.28),
"LiFePO4": (1.0, 157.76),
"Na2Fe[Fe(CN)6] (Prussian white)": (2.0, 313.77),
"graphite, LiC6 (per 6 C)": (1.0, 72.06),
"silicon, Li15Si4 (per 4 Si)": (15.0, 112.34),
"lithium metal": (1.0, 6.94),
}
print(f"{'electrode material':36s} mAh/g")
for name, (n, M) in materials.items():
print(f"{name:36s} {capacity(n, M):6.0f}")
def pair_energy(q_pos, q_neg, volts):
"""Specific energy (Wh/kg) of the two electrode materials alone, balanced: 1/q = 1/q+ + 1/q-."""
return volts / (1 / q_pos + 1 / q_neg)
print("\nactive materials only (practical capacities, mAh/g; average cell voltage):")
cells = [("LFP | graphite", 160, 360, 3.25), ("NMC811 | graphite", 200, 360, 3.65),
("NMC811 | silicon-graphite (500)", 200, 500, 3.6), ("NMC811 | lithium metal", 200, 3860, 3.8),
("Prussian white | hard carbon", 150, 300, 3.1)]
for name, qp, qn, v in cells:
e = pair_energy(qp, qn, v)
print(f" {name:32s} {e:4.0f} Wh/kg -> a real cell reaches roughly half: ~{0.5 * e:.0f}")Positive electrodes. Goodenough's lithium cobalt oxide, LiCoO₂, has a theoretical capacity of 274 mAh/g, but only about half its lithium can be removed before the layered structure becomes unstable, so it delivers about 140; it is still used in phones. Replacing most of the cobalt with nickel (as in NMC811, 80 per cent nickel) raises the usable capacity to about 200 mAh/g and lowers the use of cobalt, much of which is mined in the Democratic Republic of the Congo under widely documented social and environmental concerns. Lithium iron phosphate, LiFePO₄ (Padhi, Nanjundaswamy and Goodenough, 1997), works at a lower voltage (about 3.3 V) and stores less, but its strong P–O bonds hold the oxygen even when the cell is abused, so it is much less prone to thermal runaway, and it needs neither nickel nor cobalt. According to the IEA, LFP cells are almost 30 per cent cheaper per kWh than NMC, while LFP packs store about a fifth less energy per kilogram; LFP dominates in China and in stationary storage.
Negative electrodes. Graphite stores one lithium per six carbons, 372 mAh/g. Silicon, which alloys with lithium, holds almost ten times as much per gram, but it swells by about 300 per cent when fully lithiated and cracks its protective SEI, so today it is blended into graphite at a few to tens of per cent. Lithium metal is the ultimate negative electrode, but in liquid electrolytes it tends to grow needle-like dendrites that can short the cell.
Sodium-ion cells replace lithium by sodium, which is about a thousand times more abundant in the Earth's crust and is extracted from salt. Sodium ions are larger and do not fit well into graphite, so the negative electrode is hard carbon, a disordered carbon made from biomass or resins, and the positive electrode a layered oxide, a phosphate or a Prussian-blue analogue. The specific energy is lower than lithium-ion's, as the block shows, but the materials are cheaper and the cells perform better in the cold. Commercial production began in China in 2023, for small vehicles and grid storage.
Solid-state cells replace the flammable liquid electrolyte (a lithium salt, usually LiPF₆, in organic carbonates) with a solid ion conductor: a ceramic oxide such as the garnet Li₇La₃Zr₂O₁₂, a sulfide such as Li₁₀GeP₂S₁₂, whose conductivity rivals liquids (Kamaya et al., 2011), or a polymer. The prize is safety and, if a lithium-metal negative electrode can be used, roughly half as much energy again; the block's last lithium line shows why. The obstacles are chemical and mechanical: keeping solid–solid interfaces in contact as electrodes swell and shrink, side reactions at those interfaces, dendrites that can still grow along grain boundaries, and manufacturing at scale. As of 2026, cells with a partly solid electrolyte have appeared in a few vehicles, while fully solid-state cells are on pilot lines. The solid-state chemistry of these electrode crystals is the subject of Lumenary Materials' battery materials.
Hydrogen: electrolysis and fuel cells
In 2023 the world used about 97 million tonnes of hydrogen, mostly in oil refineries and to make ammonia and methanol, and almost all of it was made from natural gas or coal; low-emissions hydrogen was under a million tonnes (IEA Global Hydrogen Review 2024). Making it by electrolysis of water with low-carbon electricity, and using it in industry or turning it back into electricity in a fuel cell, are two halves of one chemical reaction run in opposite directions:
Two electrons pass per molecule of H₂, so the Gibbs energy fixes the reversible voltage, , and the enthalpy a second, the thermoneutral voltage , at which an electrolyser supplies all the energy, including the part, as electricity. Real cells need more voltage than that to drive the reactions at a useful rate (the overpotentials of Electrochemistry), and fuel cells deliver less; every extra volt in, or volt lost out, is energy turned into heat.
F, M_H2 = 96485.0, 2.016e-3 # C/mol, kg/mol
dH, dG = 285.83e3, 237.13e3 # J/mol for H2 + 1/2 O2 -> H2O(l) at 298 K
dH_lhv = 241.83e3 # J/mol if the water leaves as vapour (lower heating value)
E_rev, E_tn = dG / (2 * F), dH / (2 * F) # reversible and thermoneutral voltages
print(f"reversible voltage {E_rev:.3f} V, thermoneutral voltage {E_tn:.3f} V")
print(f"best possible fuel cell (ΔG/ΔH): {dG / dH:.1%} of the higher heating value")
def kwh_per_kg(volts):
"""Electrical energy per kg of H2 at a given cell voltage: 2 F V per mole of H2."""
return 2 * F * volts / M_H2 / 3.6e6
hhv = dH / M_H2 / 3.6e6
print(f"\nenergy content of hydrogen: {hhv:.1f} kWh/kg (HHV), {dH_lhv / M_H2 / 3.6e6:.1f} kWh/kg (LHV)")
print(" electrolyser cell voltage kWh per kg H2 efficiency (HHV)")
for V in (1.6, 1.8, 2.0, 2.2):
print(f" {V:4.2f} V {kwh_per_kg(V):5.1f} {E_tn / V:6.1%}")
print("\n fuel-cell voltage efficiency (LHV) round trip from a 1.9 V electrolyser")
for V in (1.0, 0.8, 0.7, 0.6):
print(f" {V:3.1f} V {V / (dH_lhv / (2 * F)):6.1%} {V / 1.9:6.1%}")An electrolyser cell running at 1.8 to 2.0 V needs about 48 to 53 kWh of electricity per kilogram of hydrogen, against an energy content of 39.4 kWh/kg; its efficiency is 74 to 82 per cent. A fuel cell at a typical working voltage of 0.7 V delivers 56 per cent of hydrogen's lower heating value; the rest leaves as heat, which is useful in buildings but not in a car. Put the two together and only about 37 per cent of the electricity comes back, before counting compression or liquefaction (which costs roughly a third of the energy content), against about 90 per cent for a lithium-ion battery. Hydrogen therefore makes most sense where electricity cannot easily go: steel made by reducing iron ore with hydrogen instead of coke, ammonia for fertiliser, long-distance shipping fuels, and seasonal storage.
The catalysts decide the cost. Alkaline electrolysers, a century-old technology, use nickel electrodes in potassium hydroxide solution. Proton-exchange-membrane (PEM) electrolysers and fuel cells use a thin acidic polymer membrane; in the acidic environment the oxygen-evolving anode needs iridium oxide and the hydrogen side platinum, and iridium, mined at only a few tonnes a year, is a real constraint. Solid-oxide cells run at 700 to 850 °C with ceramic electrolytes, where heat supplies part of the energy and no precious metals are needed. Finding cheap, durable catalysts for oxygen evolution, the slow four-electron half-reaction, is one of the central problems of electrocatalysis (Catalysis and green chemistry).
Energy from sunlight: leaves, artificial leaves and solar cells
Photosynthesis
Plants, algae and cyanobacteria run the largest chemical process on Earth. In the light reactions, two photosystems in series (the "Z-scheme") absorb red photons with chlorophyll: photosystem II uses their energy to oxidise water to O₂ at a cluster of four manganese ions and one calcium ion, the only catalyst in nature that does so, and photosystem I uses a second photon to raise the electrons high enough to reduce NADP⁺ to NADPH. The light reactions also pump protons across a membrane, and ATP synthase uses that gradient to make ATP (Biochemistry). In the Calvin cycle, NADPH and ATP drive the fixation of CO₂ into sugars, catalysed by RuBisCO, which is slow and sometimes grabs O₂ instead of CO₂, wasting energy in photorespiration.
Each CO₂ fixed needs at least eight photons. As the block below shows, eight red photons carry about three times the Gibbs energy stored, so even a perfect leaf using only absorbed red light would store 34 per cent; once the part of sunlight that chlorophyll cannot use, reflection, respiration and photorespiration are counted, the theoretical maximum is about 4.6 per cent for most plants and 6 per cent for maize and sugar cane, whose C₄ pathway suppresses photorespiration (Zhu, Long and Ort, 2008). Real crops average around 1 per cent over a year, which is why a field of solar panels produces far more usable energy than a field of energy crops.
Artificial photosynthesis
Artificial photosynthesis aims to do the plant's job, making fuel from sunlight, water and CO₂, more efficiently. In 1972 Akira Fujishima and Kenichi Honda showed that a titanium dioxide electrode lit with ultraviolet light could split water (1972). The semiconductor must absorb visible light, have its band edges placed to drive both half-reactions, and survive in water, and those demands conflict. In 2021 Kazunari Domen's group in Japan ran a 100 m² array of panels coated with an aluminium-doped strontium titanate photocatalyst, which made hydrogen safely for months at a solar-to-hydrogen efficiency below 1 per cent (Nishiyama et al., 2021): a demonstration of scale rather than of efficiency. Other approaches couple solar cells to electrolysers, or reduce CO₂ electrochemically on copper to make ethylene and ethanol. As of 2026, none competes on cost with solar electricity plus a separate electrolyser, and the research question is whether integration can make it so.
Solar cells
A solar cell uses the band gap of a semiconductor (The solid state and main-group chemistry). A photon with energy above the gap lifts an electron from the valence band to the conduction band; a junction between differently doped regions separates the electron and the hole it left, and they flow out as current. The band gap sets a trade-off.
A small band gap absorbs most of the sunlight but wastes most of each photon's energy as heat; a large one keeps more per photon but misses the red and infrared. In 1961 William Shockley and Hans Queisser worked out the best a single junction can do by detailed balance: in an ideal cell, the only loss besides those two is the light the cell itself must emit, because a good absorber is also a good emitter. The block reproduces their calculation, treating the Sun as a black body at 5778 K, and also does the photosynthesis arithmetic above.
import numpy as np
h, c, k, q, N_A = 6.62607e-34, 2.99792e8, 1.380649e-23, 1.602177e-19, 6.02214e23
# 1. Photosynthesis: at least 8 red photons (680 nm) per CO2 fixed into sugar
E_photon = h * c / 680e-9 * N_A / 1000 # kJ per mole of photons
dG_per_CO2 = 2870 / 6 # kJ/mol: glucose + 6 O2 -> 6 CO2 + 6 H2O, per carbon
print(f"680 nm photons: {E_photon:.0f} kJ/mol; 8 of them = {8 * E_photon:.0f} kJ per CO2; stored {dG_per_CO2:.0f} kJ")
print(f"best case for absorbed red light: {dG_per_CO2 / (8 * E_photon):.0%}")
# 2. The Shockley-Queisser limit: the Sun as a 5778 K black body, the cell at 300 K
def trapezoid(y, x):
return float(np.sum((y[1:] + y[:-1]) * np.diff(x)) / 2)
T_sun, T_cell, f_sun = 5778.0, 300.0, (6.957e8 / 1.496e11) ** 2 # f: Sun's angular size factor
pref = 2 * np.pi / (h**3 * c**2) # photons per m^2 s per J^3
E = np.linspace(0.05, 6.0, 60000) * q # photon energies, J
sun_flux = f_sun * pref * E**2 / np.expm1(E / (k * T_sun)) # photons per m^2 s per J
P_in = trapezoid(E * sun_flux, E) # W/m^2 arriving
def efficiency(Eg_eV):
Eg = Eg_eV * q
above = E >= Eg
J_sc = q * trapezoid(sun_flux[above], E[above]) # every photon above the gap gives one electron
Ec = np.linspace(Eg, Eg + 40 * k * T_cell, 4000) # the cell's own thermal emission above the gap
J_0 = q * trapezoid(pref * Ec**2 / np.expm1(Ec / (k * T_cell)), Ec)
V = np.linspace(0, Eg / q, 20000)
P = V * (J_sc - J_0 * np.expm1(q * V / (k * T_cell))) # ideal diode
return P.max() / P_in
print(f"\nsunlight above the atmosphere: {P_in:.0f} W/m^2")
gaps = {"silicon": 1.12, "GaAs": 1.42, "CdTe": 1.45, "MAPbI3 perovskite": 1.60, "wide-gap perovskite": 1.75}
for name, Eg in gaps.items():
print(f" {name:20s} Eg = {Eg:.2f} eV: limit {efficiency(Eg):.1%}")
grid = np.arange(0.9, 1.81, 0.01)
effs = [efficiency(g) for g in grid]
best = int(np.argmax(effs))
print(f" best single gap: {grid[best]:.2f} eV, limit {effs[best]:.1%}")With a black-body Sun, the best single gap is near 1.26 eV, with a limit of about 30 per cent; silicon's 1.12 eV is close to the optimum. With the real spectrum at the ground, which the atmosphere has filtered (the standard "AM1.5" spectrum), the same calculation gives the often-quoted 33.7 per cent near 1.34 eV. Two junctions stacked, a wide gap on top to use blue light and a narrow one below for the red, can in principle exceed 40 per cent.
Silicon makes the great majority of solar panels. It begins as quartz, reduced with carbon in electric arc furnaces, then purified through volatile silicon compounds that can be distilled to "nine nines" (99.9999999 per cent) or better, grown into crystals, sliced into wafers and doped with boron and phosphorus. Because its band gap is indirect, silicon absorbs weakly and wafers must be about 150 µm thick, and laboratory silicon cells have passed 27 per cent, close to silicon's practical limit of about 29 per cent. Metal-halide perovskites, with formula ABX₃ (for example methylammonium lead iodide, CH₃NH₃PbI₃), crystallise from solution at low temperature, absorb strongly in a film a few hundred nanometres thick, and have a band gap that can be tuned by mixing iodide and bromide. From 3.8 per cent in 2009 (Kojima et al.), certified efficiencies on the Best Research-Cell Efficiency Chart rose above 26 per cent for single perovskite cells and above 34 per cent for perovskite-on-silicon tandems. Their chemistry is also their weakness: the ionic lattice decomposes with moisture, heat and light, ions migrate in operation, and most contain lead, which must be contained and recycled. Long-term stability, not efficiency, decides their future. The band theory behind all this is in Lumenary Materials' electronic structure and bands and semiconductor materials, and the device physics in Lumenary EE's semiconductors and diodes.
The professional view, 2026
Systems, not molecules. A battery chemistry succeeds through cost per kWh, cycle life, safety, charging speed, supply chain and recyclability together. Grid operators now install large lithium-ion (mostly LFP) systems to shift solar power into the evening, and are testing longer-duration chemistries such as flow batteries (vanadium, or iron-based) and iron–air; Lumenary EE's renewables and storage treats the grid side.
Materials supply and recycling. Lithium comes mainly from Australian hard rock and from brines in Chile and Argentina, refining is concentrated in China, nickel increasingly comes from Indonesia, and graphite mostly from China, so supply, prices and the environmental and social conditions of mining are debated in every region. Recycling plants shred spent cells into "black mass" and recover cobalt, nickel, lithium and copper by hydrometallurgy (leaching and selective precipitation or solvent extraction) or by smelting; the IEA expects recycling to make a significant dent in demand for new minerals only after about a decade, when large numbers of electric-car batteries retire.
How research is done. Battery researchers combine operando X-ray diffraction and spectroscopy, which watch electrodes while they cycle, with DFT and machine-learned potentials to screen new crystal structures and electrolytes, and automated laboratories that formulate and test electrolytes in parallel (Self-driving laboratories). Life-cycle assessment (Catalysis and green chemistry) decides whether a new energy technology is an improvement: hydrogen, e-fuels and electric vehicles are only as clean as the electricity behind them. These threads continue in Sustainable chemistry.
Something safe to try at home
Watch leaves make oxygen. With an adult, dissolve a quarter of a teaspoon of baking soda (sodium hydrogencarbonate, a source of CO₂) and one drop of washing-up liquid in a glass of water. Punch about ten discs from a fresh spinach leaf with a drinking straw. Put the discs in a clean plastic syringe without a needle, draw in some of the solution, cover the tip with a finger and pull the plunger gently a few times to pull the air out of the leaves, until the discs sink. Tip them into a clear cup of the same solution under a bright lamp (an LED lamp stays cool). Over 10 to 20 minutes, the discs rise one by one, lifted by oxygen made by photosynthesis. Try it in dim light, or with plain water and no baking soda, and compare how long the discs take. Wash everything afterwards, and don't eat the discs.
Common misconceptions
Misconception
"Hydrogen is a source of energy." On Earth hydrogen is not found free in useful amounts; it is an energy carrier, made using more energy than it later returns. Whether it helps depends on where that energy comes from and what the hydrogen replaces.
Misconception
"A battery stores electricity." It stores chemical energy: lithium in one crystal rather than another, with a difference in Gibbs energy. Electricity is the flow of electrons that the reaction drives through the external circuit, which is why batteries age chemically and dislike heat.
Misconception
"Solar cells will soon be 100 per cent efficient." A single-junction cell in unconcentrated sunlight cannot exceed about a third, for the reasons the Shockley–Queisser calculation shows. Progress now comes from stacking junctions, lowering cost and improving lifetime, and records are measured against those limits.
Misconception
"Plants are efficient solar collectors." They are wonderful chemists, but crops store only about 1 per cent of the sunlight that falls on them over a year. Their advantage is that they repair and reproduce themselves.
Exercises
1. The chemistry of energy
Warm-upHumanButane, C₄H₁₀, the fuel in camping-stove cartridges, has a standard enthalpy of combustion of −2878 kJ/mol. CO₂ has a molar mass of 44.01 g/mol. How many grams of CO₂ are released per megajoule of heat? Give the nearest whole number.
2. The chemistry of energy
CoreHumanThe spinel cathode material LiMn₂O₄ (molar mass 180.81 g/mol) can release one lithium ion, and so one electron, per formula unit. Using C/mol, what is its theoretical specific capacity in mAh/g? Give the nearest whole number.
3. The chemistry of energy
CoreHumanAn electrolyser runs at a cell voltage of 1.9 V. Every molecule of H₂ needs two electrons ( C/mol; H₂ is 2.016 g/mol). How many kilograms of hydrogen does one megawatt-hour of electricity produce? Give the nearest whole number.
4. The chemistry of energy
CoreHumanLithium iron phosphate (LFP) cells are much less prone to thermal runaway than cells with layered nickel-rich oxide cathodes. What is the main chemical reason?
5. The chemistry of energy
CoreHumanWhy do sodium-ion cells use hard carbon rather than graphite as the negative electrode?
6. The chemistry of energy
Warm-upHumanWhich enzyme fixes CO₂ in the Calvin cycle, working slowly and sometimes reacting with O₂ instead of CO₂? Give its usual short name.
7. The chemistry of energy
CoreComputationalWrite capacity(n, molar_mass), the theoretical specific capacity in mAh/g of an electrode material that stores n electrons per formula unit of molar mass molar_mass (g/mol), with C/mol. Then write pair_energy(q_pos, q_neg, volts), the specific energy in Wh/kg of a balanced pair of electrodes with capacities q_pos and q_neg (mAh/g) at average voltage volts, counting the mass of both electrode materials: the pair's capacity is .
8. The chemistry of energy
StretchComputationalShockley and Queisser's first step was the ultimate efficiency: suppose every photon with energy above the band gap delivers exactly of energy, and photons below the gap deliver nothing. Treating the Sun as a black body at temperature , the photon flux per unit energy is proportional to . Write ultimate_efficiency(Eg_eV, T=5778.0) that returns
as a fraction, accurate to 0.001. Work in electronvolts: eV/K. Any constant prefactor cancels.
Further reading and resources
- The Nobel Prize in Chemistry 2019: John B. Goodenough, M. Stanley Whittingham and Akira Yoshino · Nobel Prize OutreachFor the development of lithium-ion batteries, with an illustrated popular summary.
- LixCoO2 (0 < x ≤ 1): a new cathode material for batteries of high energy density · Koichi Mizushima, Philip C. Jones, Philip J. Wiseman and John B. Goodenough · 1980
- Phospho-olivines as positive-electrode materials for rechargeable lithium batteries · Akshaya K. Padhi, K. S. Nanjundaswamy and John B. Goodenough · 1997
- A lithium superionic conductor · Noriaki Kamaya et al. · 2011
- Global EV Outlook 2025 · International Energy Agency
- Global Hydrogen Review 2024 · International Energy Agency
- What is the maximum efficiency with which photosynthesis can convert solar energy into biomass? · Xin-Guang Zhu, Stephen P. Long and Donald R. Ort · 2008
- Electrochemical photolysis of water at a semiconductor electrode · Akira Fujishima and Kenichi Honda · 1972
- Photocatalytic solar hydrogen production from water on a 100-m² scale · Hiroshi Nishiyama et al. · 2021
- Detailed balance limit of efficiency of p-n junction solar cells · William Shockley and Hans J. Queisser · 1961
- Organometal halide perovskites as visible-light sensitizers for photovoltaic cells · Akihiro Kojima, Kenjiro Teshima, Yasuo Shirai and Tsutomu Miyasaka · 2009
- Best Research-Cell Efficiency Chart · National Laboratory of the Rockies (formerly NREL)Certified record efficiencies for every kind of solar cell since 1976.
- NIST Chemistry WebBookThe combustion and formation enthalpies used in this lesson.