Hydrogen Storage

Hydrogen Storage Domain

The hydrogen domain provides evaluation models for hydrogen storage materials, including interstitial metal hydrides, complex hydrides, metal-organic frameworks (MOFs), and chemical hydrogen carriers.

Physics Model

• Gravimetric capacity is calculated from the stoichiometry of the hydrogen-bearing phase, accounting for molecular weights and the maximum hydrogen-to-metal ratio.
• Desorption temperature is derived from the van't Hoff equation using formation enthalpies estimated via the Miedema model with DFT-calibrated corrections.
• Kinetics are modeled using Johnson-Mehl-Avrami-Kolmogorov (JMAK) nucleation-and-growth kinetics, with rate constants dependent on particle size, catalyst doping, and temperature.

Default Parameters

| Parameter | Type | Bounds | Unit | Description | |—————-|———|————|———|——————-| | host_metal_a | categorical | [Ti, Zr, V, La, Mg] | — | Primary metal | | host_metal_b | categorical | [Ni, Fe, Mn, Co, Cr] | — | Secondary metal | | a_b_ratio | continuous | [0.5, 3.0] | — | A:B stoichiometric ratio | | catalyst_dopant | categorical | [none, Pd, Pt, Nb, V] | — | Catalytic additive | | dopant_wt_pct | continuous | [0.0, 5.0] | wt% | Dopant concentration | | ball_mill_time | continuous | [0.5, 48.0] | hours | Mechanical activation | | particle_size | continuous | [0.1, 100.0] | um | Mean particle diameter |

Default Objectives

| Objective | Direction | Unit | |—————-|—————-|———| | gravimetric_capacity | maximize | wt% H2 | | desorption_temp | minimize | C | | absorption_rate | maximize | wt%/min |

Key Trade-Offs

• Capacity vs. desorption temperature: Materials with stronger metal-hydrogen bonds have higher capacities but require more heat to release hydrogen. The DOE target is >5.5 wt% at <85 C.
• Kinetics vs. capacity: Nanostructuring and doping improve kinetics but can reduce reversible capacity through surface oxidation.
• Gravimetric vs. volumetric capacity: Light metals (Mg, Li) offer high gravimetric capacity but low density limits volumetric performance.

Example: MgH2-Based System

name: mg-hydride-optimization
domain: hydrogen

parameters:
  - name: host_metal_b
    type: categorical
    choices: [Ni, Fe, Ti, V]
  - name: a_b_ratio
    type: continuous
    bounds: [1.5, 2.5]
  - name: catalyst_dopant
    type: categorical
    choices: [none, Nb, V, Pd]
  - name: dopant_wt_pct
    type: continuous
    bounds: [0.0, 3.0]
  - name: ball_mill_time
    type: continuous
    bounds: [1.0, 24.0]

objectives:
  - name: gravimetric_capacity
    direction: maximize
    unit: wt%
  - name: desorption_temp
    direction: minimize
    unit: C

optimizer:
  method: cma-es
  budget: 300
  batch_size: 15

MOF Storage Mode

For metal-organic frameworks, the parameter space shifts to linker chemistry and pore geometry:

parameters:
  - name: pore_diameter
    type: continuous
    bounds: [5.0, 30.0]
    unit: angstrom
  - name: surface_area
    type: continuous
    bounds: [500, 7000]
    unit: m2/g
  - name: framework_density
    type: continuous
    bounds: [0.1, 1.5]
    unit: g/cm3

The MOF evaluation model uses grand canonical Monte Carlo (GCMC)-derived isotherms parameterized by pore geometry.

Typical Results

MgH2-based campaigns typically discover:

• Pure MgH2: 7.6 wt% capacity, desorption above 300 C, slow kinetics
• Nb-catalyzed MgH2: 6.5 wt%, desorption at ~250 C, fast absorption
• MgNi alloys: 3.6 wt%, desorption below 200 C, moderate kinetics

The Pareto front reveals the fundamental trade-off between capacity and practical operating temperatures.