Materials, Nano & Molecular Engineering¶
-
Exit capability
Move from quantum/chemical mechanisms to computational screening, characterization, processing, interfaces, reliability, and manufacturable materials.
-
Frontier targets
Superconductors; batteries; catalysts; smart/self-healing materials; nanotech; high-temperature structures.
-
Choose this track if
Best if you want to create new physical capabilities rather than assemble known ones.
Electronic structure¶
Week 1: Solid-state/quantum refresh¶
Reading/source: David Tong solid-state notes + MIT 3.091 electronic materials
Know: Connect crystal symmetry and electronic structure to observable material properties.
Reconstruct: Derive reciprocal-lattice/band-filling intuition and density-of-states role.
Do: Compute/plot tight-binding bands for 1D/2D toy lattices.
Context: Materials design starts with structure/electrons but must end in measurement.
Defend: Which predicted property is most sensitive to model approximation?
Gate: Pass if qualitative band behavior is predicted before computation.
Source: source
Thermodynamics¶
Week 2: Phase stability and chemical potentials¶
Reading/source: MIT 3.091 + Materials Project thermodynamic methodology
Know: Use free energies, convex hulls and chemical potentials to reason about phase stability.
Reconstruct: Derive binary convex-hull stability and lever/chemical-potential intuition.
Do: Build hull from toy formation energies; identify metastability margin.
Context: Materials Project exposes computed thermodynamic stability at scale.
Defend: Why can a metastable phase still be manufacturable/useful?
Gate: Pass if equilibrium prediction is separated from kinetic accessibility.
Source: source
Defects¶
Week 3: Defects, doping and nonstoichiometry¶
Reading/source: MIT 3.091 defects/diffusion
Know: Understand vacancies/interstitials/substitution, charge compensation and defect-controlled properties.
Reconstruct: Derive Arrhenius/Boltzmann defect concentration and diffusion length.
Do: Simulate defect concentration/diffusion versus T and process time.
Context: Many functional materials are engineered through controlled imperfection.
Defend: When does a 'defect' become the desired feature?
Gate: Pass if process conditions map to defect/property target.
Source: source
Transport¶
Week 4: Electronic, ionic and thermal transport¶
Reading/source: Solid-state + transport methods
Know: Connect carriers/scattering/phonons/ions to conductivity and heat flow.
Reconstruct: Derive Drude-like conductivity scaling and diffusion-mobility relationship.
Do: Compare electronic/ionic/thermal transport tradeoffs for battery/thermoelectric toy materials.
Context: High performance often requires contradictory transport properties.
Defend: Which transport mechanism is rate limiting at device scale?
Gate: Pass if material metric is translated into device consequence.
Source: source
Computation¶
Week 5: Density-functional-theory workflow literacy¶
Reading/source: Materials Project methodology/workflows
Know: Understand what DFT predicts, approximations, relaxation, energy/band/elastic workflows and systematic error.
Reconstruct: Reconstruct Kohn-Sham workflow conceptually without pretending to derive DFT from scratch.
Do: Use Materials Project data to compare computed properties across a chemical family.
Context: MP publishes standardized high-throughput calculation workflows and methodologies.
Defend: What error comes from functional/model versus database/data processing?
Gate: Pass if every downloaded property includes method/version caveat.
Source: source
Week 6: Materials Project API and reproducible data¶
Reading/source: Materials Project API getting started
Know: Programmatically query structures/properties and preserve provenance/database version.
Reconstruct: Write schema for material_id, structure, property, uncertainty/method metadata.
Do: Build reproducible search pipeline for candidate battery/catalyst/semiconductor materials.
Context: MP warns users to consider benchmarking/systematic errors and database versions.
Defend: How could database selection bias distort your candidate ranking?
Gate: Pass if query/version/source are reproducible from clean environment.
Source: source
Week 7: Descriptors and surrogate models¶
Reading/source: Materials informatics methods + Wave-1 ML
Know: Build property models from composition/structure descriptors and quantify extrapolation.
Reconstruct: Derive train/validation split strategy by chemical family rather than random rows.
Do: Train baseline property predictor; compare random split vs composition-held-out split.
Context: Materials ML is especially vulnerable to leakage from chemically similar samples.
Defend: Does your model interpolate chemistry or discover new chemistry?
Gate: Pass if OOD split is materially harder and reported.
Source: source
Week 8: Active learning and Bayesian optimization¶
Reading/source: Wave-1 probability/optimization + materials search
Know: Select next computation/experiment based on uncertainty and expected improvement.
Reconstruct: Derive Gaussian-process/expected-improvement intuition or equivalent uncertainty-guided search.
Do: Search toy composition space with active learning versus random screening.
Context: Closed-loop materials discovery depends on informative experiment selection.
Defend: What if your uncertainty model is confidently wrong?
Gate: Pass if acquisition is stress-tested on misspecified surrogate.
Source: source
Characterization¶
Week 9: Diffraction and structure determination¶
Reading/source: MIT 3.091 crystal/XRD materials
Know: Infer structure from diffraction and understand ambiguity/resolution.
Reconstruct: Derive Bragg law and reciprocal-space peak relation.
Do: Generate/index synthetic XRD patterns with noise/mixtures.
Context: Predicted crystal structure means little without characterization.
Defend: What distinct structures could produce similar patterns?
Gate: Pass if uncertainty/multiphase alternatives are considered.
Source: source
Week 10: Electrical/magnetic/thermal measurements¶
Reading/source: Materials characterization principles + MP property methodology
Know: Design measurements that discriminate mechanisms, not merely produce one number.
Reconstruct: Regenerate four-probe/contact-resistance distinction and temperature/field sweep logic.
Do: Design superconductivity/semiconductor conductivity experiment in simulation with contact artifacts.
Context: Frontier material claims frequently fail at measurement interpretation.
Defend: What measurement would falsify your favored mechanism?
Gate: Pass if at least two independent observables are proposed.
Source: source
Week 11: Microscopy/spectroscopy and multiscale evidence¶
Reading/source: MIT 3.091 + characterization survey
Know: Connect spatial/chemical/structural probes to microstructure/property claims.
Reconstruct: Map measurement resolution/volume to defect length scales.
Do: Create characterization plan for grain-boundary-controlled material.
Context: No single instrument sees 'the material'; each samples scales/contrasts.
Defend: What feature could be invisible to your chosen measurement?
Gate: Pass if cross-validation across modalities is designed.
Source: source
Replication¶
Week 12: Reproduce a Materials Project trend¶
Reading/source: MP API + methodology
Know: Verify a computed materials trend and its methodological caveats.
Reconstruct: Rebuild candidate-selection criteria and one computed descriptor.
Do: Replicate published/database trend across a material family and compare to available experiment.
Context: MP exists to accelerate materials discovery through public computed property data.
Defend: Where is computation systematically biased?
Gate: Replication Gate: notebook + provenance + experimental cross-check where available.
Source: source
Electrochemistry¶
Week 13: Battery materials and interfaces¶
Reading/source: Chem/materials core + MP battery-relevant properties
Know: Connect voltage, capacity, diffusion, phase stability and interfaces to cell behavior.
Reconstruct: Derive intercalation voltage from free-energy difference intuition.
Do: Screen cathode/electrolyte candidates; build simple cell-level energy/degradation model.
Context: Battery gains require interfaces, safety and manufacturing beyond bulk energy density.
Defend: Which bulk-property improvement is erased at cell/pack level?
Gate: Pass if candidate is evaluated at material and device levels.
Source: source
Catalysis¶
Week 14: Catalysts and surface energetics¶
Reading/source: MP surface/adsorption methodology + physical chemistry
Know: Understand adsorption, activation barriers, selectivity and catalyst stability.
Reconstruct: Derive Sabatier-volcano qualitative tradeoff.
Do: Create toy catalyst screening with activity/selectivity/stability multiobjective score.
Context: Catalyst discovery is constrained optimization over competing mechanisms.
Defend: Why is strongest binding usually not best?
Gate: Pass if selectivity and degradation are not omitted.
Source: source
Functional materials¶
Week 15: Semiconductors, dielectrics, piezoelectrics, magnetics¶
Reading/source: Materials Project property methodology
Know: Map band gaps, dielectric/piezo/magnetic properties to devices.
Reconstruct: Derive simple device figure of merit from material constants.
Do: Query MP for a functional-material family and construct Pareto frontier.
Context: Materials databases expose many computed device-relevant properties.
Defend: Which figure of merit is incomplete at device/manufacturing scale?
Gate: Pass if screening includes stability/abundance constraints.
Source: source
Frontier¶
Week 16: Superconductivity and extraordinary claims¶
Reading/source: Solid-state theory + rigorous characterization
Know: Understand signatures/constraints without pretending current theory predicts high-Tc materials reliably.
Reconstruct: Regenerate zero-resistance vs Meissner distinction and critical field/current concepts.
Do: Write preregistered verification protocol for claimed ambient superconductor.
Context: The target is scientifically open; evidence standards must be unusually strong.
Defend: What mundane artifact could mimic each claimed signature?
Gate: Pass if protocol includes independent replication and sample provenance.
Source: source
Processing¶
Week 17: Processing-structure-property¶
Reading/source: MIT 3.091 phase/kinetics + Wave-2 manufacturing
Know: Treat synthesis/annealing/deposition as state-control of microstructure.
Reconstruct: Derive diffusion/process-time scaling and nucleation/growth intuition.
Do: Optimize a virtual heat-treatment/process schedule for target microstructure/property.
Context: A computational candidate that cannot be synthesized reproducibly is not a technology.
Defend: Which process variable controls variance rather than mean property?
Gate: Pass if process window and measurement plan are specified.
Source: source
Scale¶
Week 18: Manufacturability, supply and reliability¶
Reading/source: Wave-2 manufacturing + materials constraints
Know: Integrate yield, purity, critical minerals, recycling and lifetime.
Reconstruct: Compute material intensity per GWh/million devices and sensitivity to yield.
Do: Scale candidate material to 10k/1M units; identify precursor/purity/thermal/process bottleneck.
Context: Materials innovation can shift bottlenecks into supply chains.
Defend: What happens if the rarest element becomes 10× more expensive?
Gate: Pass if substitute/recycling path is evaluated.
Source: source
Research¶
Week 19: Failure and uncertainty map¶
Reading/source: All track work
Know: Rank computational, synthesis, characterization and lifetime uncertainties.
Reconstruct: Construct Bayesian/causal dependency graph from composition/process -> structure -> property -> device.
Do: Run sensitivity/value-of-information analysis to choose next measurement.
Context: High leverage often comes from resolving uncertainty, not searching more candidates.
Defend: What experiment most changes your decision?
Gate: Pass if next step is chosen quantitatively.
Source: source
Week 20: Reproduce a current materials paper/workflow¶
Reading/source: Choose open computational/experimental paper + MP data
Know: Practice reproducible materials science.
Reconstruct: Reconstruct structure, method, characterization and claimed mechanism.
Do: Replicate computed trend or public data analysis; challenge with alternate functional/split/metric.
Context: Current materials research is increasingly data/workflow intensive.
Defend: Which conclusion survives method changes?
Gate: Extension Gate: reproduction + one countermodel/alternative explanation.
Source: source
Week 21: Closed-loop discovery prototype¶
Reading/source: MP API + surrogate/active-learning stack
Know: Integrate candidate generation, prediction, uncertainty and acquisition.
Reconstruct: Regenerate decision loop and stopping rule.
Do: Build autonomous computational materials-discovery loop on bounded search space.
Context: This is a safe analog of an autonomous lab before real synthesis.
Defend: How do you detect that the loop is exploiting model/database artifacts?
Gate: Pass if hidden 'ground truth' benchmark exposes and measures exploitation.
Source: source
Week 22: Mechanism-focused extension¶
Reading/source: Selected material system
Know: Move beyond leaderboard property prediction to mechanistic hypothesis.
Reconstruct: Write mechanism + discriminating observation.
Do: Add descriptor/physics constraint/measurement simulation testing mechanism.
Context: Industry-leading materials work couples prediction to explanatory measurement.
Defend: What observation would force a different mechanism?
Gate: Pass if hypothesis makes risky prediction.
Source: source
Capstone¶
Week 23: New-material design dossier¶
Reading/source: All track sources
Know: Produce candidate -> computation -> synthesis concept -> characterization -> device -> scale chain.
Reconstruct: Reconstruct full processing-structure-property-performance argument.
Do: Capstone material family for battery/catalyst/semiconductor/smart material with reproducible screening.
Context: No 'miracle material' claim without evidence ladder.
Defend: Which link in the chain is least verified?
Gate: Systems Gate: computational + experimental + manufacturing reviewers.
Source: source
Week 24: Research program and kill criteria¶
Reading/source: All sources
Know: Define experiments and milestones that could rapidly falsify the program.
Reconstruct: Write top hypotheses, predicted signatures, null models and value-of-information ordering.
Do: Paper/notebook + candidate database + characterization plan + scale/cost/supply model + 12-month roadmap.
Context: Good materials programs kill bad candidates early.
Defend: What result makes you stop working on this material?
Gate: Capstone Gate: defend novelty, evidence, manufacturability and falsifiability.
Source: source
Research gates¶
G1 Replication¶
Required performance: Reproduce computed/experimental material trend with provenance and method sensitivity.
Minimum artifacts: Notebook; database version; structures; method; uncertainty; experimental cross-check.
Pass criterion: Random row split alone is not adequate ML evidence.
G2 Extension¶
Required performance: Test a mechanism or improve candidate selection under realistic OOD/process constraints.
Minimum artifacts: Competing mechanism predictions; OOD split; active-learning/measurement plan.
Pass criterion: Must include a risky discriminating prediction.
G3 System Closure¶
Required performance: Close composition->process->structure->property->device->manufacturing/supply chain.
Minimum artifacts: Characterization ladder; process window; device model; yield/material-intensity model.
Pass criterion: A computed property without synthesis/measurement path fails.
G4 Research Defense¶
Required performance: Propose fastest sequence of experiments that kills bad candidates and validates good ones.
Minimum artifacts: Candidate dossier; characterization plan; scale/cost/supply; 12-month roadmap.
Pass criterion: Must state exact observation that ends the program.
Frontier technologies primarily routed here¶
- 5. Grid-scale batteries: class A
- 12. AI materials scientist: class B
- 15. Additive manufacturing / primitive matter printers: class A
- 35. Self-healing materials: class B
- 36. Smart materials: class B
- 43. Next-generation batteries: class B
- 46. Nanomedicine: class B
- 52. Programmable matter-primitive forms: class B
- 54. Room-temperature/ambient-pressure superconductivity: class B
- 85. Molecular assemblers: class D
- 86. Medical nanorobots: class D