Computational techniques for characterisation of electrically conductive MOFs: quantum calculations and machine learning approaches

7review sections

6material families

15review claims

15secondary benchmarks

28cited studies

6research gaps

Review scope

Summarise quantum-mechanical and machine-learning approaches for characterising and discovering electrically conductive MOFs, especially through band-structure, density-of-states and band-gap analysis.

Coverage: 1995–2021
Category: Review Theory Transport
Material scope: electrically conductive and semiconductive MOFs · 2D conjugated MOFs · UiO, MOF-74, MIL-53, ZIF and HKUST-type examples · ML-screened low-band-gap framework materials
Transport scope: band gap · band structure · density of states · charge-transport pathways · metallicity and semiconductivity screening

Application scope: electronics · sensors · transistors · resistive switching · energy storage · electrocatalysis
Explicit exclusions: primary extraction of synthesis recipes · exhaustive transcription of all band-gap rows · direct conductivity leaderboards
Source: 13584 · Abstract
Evidence role: Context, taxonomy and secondary benchmarking

Section map

The review’s argument is preserved as a navigable set of section summaries.

Band gap calculations

13587-13592

Synthesises band-structure, DoS, PBE/HSE/GW/DFT+U caveats, and representative conductive-MOF band-gap comparisons.

Relevance: Core · 13587 · Band gap calculations

Computational approaches overview

13585

Uses Fig. 1 to organise experimental methods, computational methods, DFT, ML and database-enabled discovery workflows.

Relevance: Core · 13585 · Introduction · Fig. 1

Conclusions

13595-13596

Concludes that periodic DFT remains central, method choice is context dependent, and ML is promising for triaging candidates but needs better data and uncertainty treatment.

Relevance: Core · 13595 · Conclusions

Geometry optimisation

13587

Explains when low-level DFT, dispersion corrections, fixed cell parameters or hybrid functionals are used for MOF geometry preparation.

Relevance: Supporting · 13587 · Geometry optimisation

Introduction

13584-13585

Frames conductive MOFs as uncommon but valuable porous electronic materials and introduces through-bond and through-space charge-transport routes.

Relevance: Core · 13584 · Introduction

Machine learning for characterisation of conductive MOFs

13592-13595

Reviews ML charge assignment, metallicity screening, QMOF/CGCNN/SOAP band-gap prediction and low-band-gap candidate discovery.

Relevance: Core · 13592 · Machine learning for characterisation of conductive MOFs

Quantum mechanical methods for characterisation of conductive MOFs

13585-13587

Reviews WFT, DFT and GW, then focuses on practical DFT choices including functionals, basis sets, pseudopotentials and k-point sampling.

Relevance: Core · 13586 · Quantum mechanical methods for characterisation of conductive MOFs

Taxonomies

Classification systems are attributed to this review and are not treated as a global material registry.

Band-Gap Magnitude

Band-gap conductivity classes

The review uses broad band-gap ranges to interpret whether conductive behaviour is expected from a material.

Categories: zero band gap conductors · 0-3 eV semiconductors · above about 4 eV insulators

13585 · Introduction

Electronic Coupling Mechanism

Charge-transport pathway types

Conductive MOFs are organised by whether charge transport is installed through extended bonding networks or by spatial orbital overlap.

Categories: through-bond extended pi-pi or pi-d conjugation · through-space overlapping orbitals

13584 · Introduction

Study RouteAuthor-proposed

Conductive MOF characterisation workflow

Fig. 1 maps characterisation from experimental/computational methods through electronic-structure study and screening outputs.

Categories: experimental methods · computational methods · comparison with experiment · discovery of new MOFs · use of existing databases

13585 · Introduction · Fig. 1

Screening Target

ML screening routes

ML methods are grouped by whether they estimate charges, classify metallicity, regress band gaps, or reduce candidates for detailed calculations.

Categories: partial-charge prediction · metallic/non-metallic classification · band-gap regression · candidate triage for high-level DFT

13593 · Machine learning for characterisation of conductive MOFs · Fig. 8

Model Representation

Periodic versus non-periodic DFT

The review distinguishes full periodic MOF calculations from molecular fragments used to probe conformations and HOMO-LUMO effects.

Categories: periodic crystal structure · non-periodic molecular building block

13585 · Introduction

Material families

Review-defined families retain their representative materials and conduction descriptions.

HKUST-1 guest-doped systems

3D Framework With Guest-Mediated Pathway

HKUST-1 and redox-active guest variants such as TCNQ@HKUST-1.

Conduction: TCNQ insertion creates donor-acceptor coupling and new states in the MOF HOMO-LUMO gap.

Representative materials: HKUST-1 · TCNQ@HKUST-1

Nodes / linkers: Cu paddlewheel · BTC · TCNQ guest

13592 · Band gap calculations

M-MIL-53 flexible frameworks

3D Flexible Framework

Isostructural trivalent-metal MIL-53 frameworks with large-pore and narrow-pore conformations.

Conduction: Band gaps change with pore conformation because narrower pores increase linker overlap.

Representative materials: Ti-MIL-53 · Fe-MIL-53 · V-MIL-53 · Sc-MIL-53 · Cr-MIL-53 · In-MIL-53 · Ga-MIL-53 · Al-MIL-53

Nodes / linkers: Ti · Fe · V · Sc · Cr · In · Ga · Al · MIL-53 dicarboxylate framework

13591 · Band gap calculations · Fig. 6

ML-screened low-band-gap framework candidates

Mixed Framework Dimensionalities

Framework materials from CSD/QMOF screening predicted to have low DFT+PBE band gaps and selected for follow-up.

Conduction: Low predicted band gap is treated as a triage signal, not proof of high conductivity.

Representative materials: RAXNEK · OTARUX · WAQMEJ · FAFJAZ · HIVPOU

Nodes / linkers: Fe · Ag · Rh · Cu · Ni/Cu · TCNQ derivatives · polyoxovanadate units · fluorinated or metal-fluoride species · pdt-type ligands

13595 · Notable low band gap materials found via machine learning · Table 2

MOF-74 and redox-modified analogues

3D Framework

MOF-74-type frameworks including Fe/Co/Mg variants and NDI/TTF-modified DSNDI-MOF-74 analogues.

Conduction: Band-gap tuning is discussed through DFT+U, redox-active ligands and donor-acceptor guest interactions.

Representative materials: Fe-MOF-74 · Co-MOF-74 · Mg-MOF-74 · DSNDI-MOF-74 · TTF-DSNDI-MOF-74

Nodes / linkers: Fe · Co · Mg · DOBDC-type · naphthalenediimide · tetrathiafulvalene guest/dopant

13592 · Band gap calculations

2D conjugated triphenylene-type MOFs

2D Layered

Layered conductive MOFs based on HITP, HHTP, HHTT or related linkers with transition-metal or lanthanide nodes.

Conduction: High conductivity is linked to in-plane pi/d conjugation and/or interlayer pi-pi stacking.

Representative materials: Ni-HITP · Cu-HITP · M-HHTT · La-HHTP · Cu-BHT

Nodes / linkers: Ni · Cu · Co · Mg · lanthanides · hexaiminotriphenylene · hexahydroxytriphenylene · hexahydroxy tetraazanaphthotetraphene · benzenehexathiol

13591 · Band gap calculations · Fig. 5

UiO-family MOFs

3D Frameworks

Zr- or Ce-based UiO frameworks with BDC/BPDC/TPDC linkers and functionalised analogues.

Conduction: Band gaps are sensitive to linker length, hydroxylation state, metal substitution and functional groups.

Representative materials: UiO-66 · UiO-67 · UiO-68 · UiO-66(Ce)

Nodes / linkers: Zr6O4(OH)4 · Ce · BDC · BPDC · TPDC · functionalised BDC

13591 · Band gap calculations

Synthesis strategies

Review-level synthesis principles remain separate from primary-study recipes.

Conformation and cell-volume tuning

Exploit pore breathing, interlayer distance or cell-volume changes to alter orbital overlap.

Claimed effects: Narrower pores or shorter stacking distances can reduce band gaps by increasing orbital overlap.

Controlling variables: pore conformation · pressure · temperature · gas adsorption · interlayer distance

Representative materials: M-MIL-53 · Ln-HHTP

Caveat: Applicability depends on flexible or layered structures and should be separated from irreversible chemical modification.

13591 · Band gap calculations

Guest-molecule or redox doping

Introduce redox-active guests or dopants to create donor-acceptor interactions and new electronic states inside the MOF gap.

Claimed effects: Can create new charge-transport pathways and significantly increase conductivity.

Controlling variables: guest identity · host-guest electronic coupling · solvation state · doping level

Representative materials: TCNQ@HKUST-1 · TTF-DSNDI-MOF-74

Caveat: Guest or solvent state can make the resulting conductivity extrinsic and condition-sensitive.

13592 · Band gap calculations

Organic linker functionalisation

Modify linker substituents to tune band gaps while preserving a common parent framework for comparison.

Claimed effects: Can narrow band gaps and make systematic structure-property comparisons possible.

Controlling variables: functional group identity · linker electronic character · parent framework consistency

Representative materials: UiO-66(Ce) · IRMOF-1 · IRMOF-20

Caveat: Presented as case-study evidence rather than a universal rule.

13592 · Band gap calculations · Fig. 7

Metal substitution in isoreticular families

Change metal nodes in related frameworks to alter d-orbital contributions, spin states and metal-ligand overlap.

Claimed effects: Can identify additional conductive analogues and change band gaps across a family.

Controlling variables: metal identity · open-shell character · metal-ligand orbital overlap

Representative materials: M-HITP · M-HHTT · M-MIL-53

Caveat: Open-shell transition metals can require hybrid or corrected methods for credible band gaps.

13591 · Band gap calculations

ML candidate triage followed by high-level calculations

Use ML models to down-select large MOF/material spaces before expensive DFT or experimental validation.

Claimed effects: Greatly reduces screening cost and can recover previously known conductive or semiconductive candidates.

Controlling variables: training-set chemistry · descriptors or graph representation · band-gap threshold · validation by HSE/experiment

Representative materials: CSD-42362 · QMOF · RAXNEK · WAQMEJ · OTARUX

Caveat: Positive low-band-gap or metallic predictions do not guarantee conductivity.

13593 · Machine learning for characterisation of conductive MOFs · Fig. 8

Review claims

These are the review authors’ synthesis, not newly measured results.

Author InterpretationHigh supportMeasurement Interpretation

Band gap helps screen candidates, but band structure and DoS are also needed to understand electronic states and likely carrier behaviour.

Evidence basis: review_reasoning

Caveat: This does not replace direct conductivity or mobility measurements.

13587 · Band gap calculations

Consensus SummaryHigh supportDefinition Scope

Conductive MOFs are uncommon because porosity, redox-inactive linkers and hard metal ions often impede charge delocalisation.

Evidence basis: review_reasoning

Caveat: Broad field-level tendency, not a quantitative prevalence estimate.

13584 · Introduction

Consensus SummaryHigh supportConsensus

DFT is the main practical method for periodic electronic-structure calculations in conductive MOFs because it balances cost and accuracy.

Evidence basis: review_reasoning

Caveat: GW and wave-function methods are discussed but are usually too expensive for broad MOF studies.

13586 · Quantum mechanical methods for characterisation of conductive MOFs

Author InterpretationHigh supportCaveat

Geometry-optimisation method choice depends on material size, computational resources and the number of structures; low-level DFT may be justified for first-step screening.

Evidence basis: multi_reference

Caveat: High-level calculations remain appropriate for selected outstanding materials or flexible structures.

13587 · Geometry optimisation

Consensus SummaryHigh supportTransport Mechanism

Guest molecules or dopants can create donor-acceptor pathways and new gap states, changing apparent conductivity and band gaps.

Evidence basis: multi_reference

Caveat: Such behaviour may be extrinsic and sensitive to solvent or guest state.

13592 · Band gap calculations

Author InterpretationMedium supportStructure Property Link

For HHTT-based 2D MOFs, zero band gap alone is not enough; band dispersion and orbital contributions explain why Cu-HHTT is interpreted as more conductive than Co-HHTT.

Evidence basis: single_reference

Caveat: The review summarises one family as an interpretive example.

13588 · Band gap calculations · Fig. 2

Consensus SummaryHigh supportTransport Mechanism

HITP/HHTP/HHTT-type 2D MOFs are interpreted through in-plane metal-ligand conjugation and interlayer pi-pi stacking.

Evidence basis: multi_reference

Caveat: Individual materials still require primary transport and structural evidence.

13591 · Band gap calculations

Consensus SummaryHigh supportCaveat

LDA/GGA functionals such as PBE tend to underestimate band gaps, while HSE-type hybrid functionals often give values closer to experiment.

Evidence basis: multi_reference

Caveat: HSE is more expensive and can require fewer k-points or Gamma-only treatment in large MOFs.

13588 · Band gap calculations

Author InterpretationHigh supportCaveat

Low predicted band gap is not sufficient for good electrical conductivity because defects, disorder and thermal effects can also control resistance.

Evidence basis: multi_reference

Caveat: Candidate lists should be treated as triage, not evidence of device-level transport.

13594 · Notable low band gap materials found via machine learning

Author InterpretationHigh supportStructure Property Link

M-MIL-53 pore conformation changes can tune band gaps because narrow pores increase neighbouring-linker orbital overlap.

Evidence basis: single_reference

Caveat: The review notes Ti-MIL-53 as an exception to the general lp versus np pattern.

13591 · Band gap calculations · Fig. 6

Author InterpretationHigh supportCaveat

Binary metallic/non-metallic ML screens may miss semiconducting MOFs and a positive metallic prediction does not guarantee conductivity.

Evidence basis: single_reference

Caveat: Useful as a screening caveat for Chapter 1 ML discussion.

13593 · Machine learning for characterisation of conductive MOFs

Author InterpretationHigh supportApplication Relevance

ML can reduce very large MOF search spaces to manageable candidate sets for high-level DFT and experimental follow-up.

Evidence basis: multi_reference

Caveat: Requires relevant, diverse training data and later validation.

13592 · Machine learning for characterisation of conductive MOFs

Consensus SummaryHigh supportCaveat

MOFs containing open-shell transition metals are difficult for most DFT functionals because of spin-state and electronic-structure complications.

Evidence basis: multi_reference

Caveat: The review suggests hybrid DFT for more realistic band-gap descriptions in such cases.

13589 · Band gap calculations

Author InterpretationHigh supportConsensus

The review identifies a lack of systematic structure-conductivity investigation across MOF chemistries, limiting rational correlations.

Evidence basis: review_reasoning

Caveat: This is the review authors outlook rather than a meta-analytic result.

13595 · Conclusions

Consensus SummaryHigh supportTransport Mechanism

The review interprets conductive MOF design mainly through through-bond conjugation and through-space orbital-overlap pathways.

Evidence basis: review_reasoning

Caveat: A simplifying taxonomy across heterogeneous materials.

13584 · Introduction

Secondary benchmarks

Every row remains visibly secondary and links to a primary dossier only where the mapping is verified.

Material	Property	Reported value	Context and quality	Primary evidence	Review source
SecondaryQMOF/CSD-42362 MOF subset	ML versus DFT compute time	7 minutes for 13,058 MOFs versus 1.5 million hours via DFT+PBE	Review summary of CGCNN screening versus DFT+PBE computation Text · Exact Reported	No verified corpus mapping	13594 · Machine learning for characterisation of conductive MOFs
SecondaryQMOF/CSD-42362 MOF subset	CGCNN band-gap prediction performance	MAE 0.27 eV; R2 0.89	ML prediction of DFT-computed band gaps for optimised MOF structures Text · Exact Reported	No verified corpus mapping	13593 · Machine learning for characterisation of conductive MOFs · Fig. 9
SecondaryCu(TCNQCl2) / FAFJAZ	ML-screened band gap	DFT+PBE predicted 0.009 eV; experimental 0.032 eV	Table 2 notable conductive materials identified using ML techniques Table · Exact Reported	No verified corpus mapping	13595 · Notable low band gap materials found via machine learning · Table 2
SecondaryDSNDI-MOF-74	band gap	PBE 1.6 eV; HSE06 2.5 eV; experimental 2.1 eV	Table 1 and text summary before TTF guest doping Text · Exact Reported	research_0182	13592 · Band gap calculations
SecondaryFe-MOF-74	band gap comparison	PBE/DFT+U 0.3-1.75 eV; HSE06 1.38-2.44 eV; experimental 2.1-1.3 eV	Table 1 secondary summary; includes Hubbard-corrected values in footnote Table · Range	No verified corpus mapping	13590 · Band gap calculations · Table 1
SecondaryLa-HHTP	directional band gap	1.5 eV in-plane; no band gap out-of-plane	Review text interpreting band structure and DoS Text · Range	research_0047	13591 · Band gap calculations · Fig. 5
SecondaryM-HHTT	band gap	0 eV PBE	Table 1 secondary summary and Fig. 2 discussion for layered HHTT MOFs Table · Exact Reported	No verified corpus mapping	13590 · Band gap calculations · Table 1
SecondaryM-MIL-53	lp-np band-gap difference	0.35 eV for V3+ to 1.39 eV for In3+	HSE06 comparison of large-pore and narrow-pore conformations Text · Range	No verified corpus mapping	13591 · Band gap calculations · Fig. 6
SecondaryNi-HITP	band gap	PBE 0.12 eV; HSE06 0.2/0 eV	Table 1 secondary summary; low-level and high-level DFT values Table · Range	No verified corpus mapping	13590 · Band gap calculations · Table 1
SecondaryFe(squarate)(bpee)(H2O)2 / RAXNEK	ML-screened band gap	DFT+PBE predicted 0.382 eV; reported 1.06 eV by HSE06-D3(BJ)	Table 2 notable conductive materials identified using ML techniques Table · Exact Reported	No verified corpus mapping	13595 · Notable low band gap materials found via machine learning · Table 2
SecondaryTCNQ@HKUST-1	electrical conductivity increase	six orders of magnitude to 7 x 10^-2 S cm^-1	TCNQ infiltration of HKUST-1; secondary review summary of original work Text · Approximate	research_0088	13592 · Band gap calculations
SecondaryTTF-DSNDI-MOF-74	band gap	PBE 0.9 eV; HSE06 1.5 eV; experimental 1 eV	Table 1 and review text after TTF guest doping Text · Exact Reported	research_0182	13592 · Band gap calculations
SecondaryUiO-66(Ce)	functionalisation band-gap narrowing	from 2.66 eV to 1 eV with SH and NH functional groups	HSE06 calculated DoS of functionalised UiO-66(Ce) Text · Approximate	No verified corpus mapping	13592 · Band gap calculations · Fig. 7
SecondaryUiO-66	band gap comparison	PBE 2.92 eV; HSE06 4.03 eV; experimental ca. 3.94-4.07 eV	Review text/Table 1 comparison of low-level, high-level and experimental band gaps Text · Range	No verified corpus mapping	13591 · Band gap calculations · Table 1
Secondary(TTF)[Rh2(CH3CO2)4]2TCNQ / WAQMEJ	ML-screened band gap	DFT+PBE predicted 0.151 eV; reported 0.71 eV by HSE06-D3(BJ)	Table 2 and Fig. 10 low-band-gap candidate context Table · Exact Reported	No verified corpus mapping	13595 · Notable low band gap materials found via machine learning · Table 2

Research gaps

Open questions are presented as review-author priorities, not conclusions from the primary database.

Mechanistic understanding

High

The review states that the nature of conductivity in MOFs is not yet explicitly understood at the atomistic level.

Proposed direction: Develop atomistic models that identify low-energy charge-transport pathways and compare them against transport measurements.

13584 · Abstract

Band-gap measurement comparability

High

Measured band gaps can vary with technique, synthesis, environment and post-treatment conditions.

Proposed direction: Pair computational benchmarks with careful experimental metadata and uncertainty reporting.

13589 · Band gap calculations

Uncertainty in ML predictions

High

The review calls for better reference band-gap calculations and methods to quantify prediction uncertainty.

Proposed direction: Report uncertainty estimates alongside ML candidate rankings and validate promising cases with high-level DFT/experiment.

13596 · Conclusions

ML training-set chemistry

High

ML studies may struggle with rare elements and hypothetical chemistry that are absent from training data.

Proposed direction: Expand training sets to chemically diverse MOFs and include more exact MOF-specific reference calculations.

13593 · Machine learning for characterisation of conductive MOFs

Open-shell transition-metal MOFs

Medium

Electronic structures of MOFs containing open-shell metals are difficult for most DFT functionals.

Proposed direction: Use hybrid or otherwise corrected methods and benchmark against experiments for Ti, V, Fe and related systems.

13589 · Band gap calculations

Structure-conductivity correlation

High

The review identifies a lack of systematic investigation over wide MOF chemistries, limiting rational structure-conductivity correlations.

Proposed direction: Build comparative datasets that vary metal nodes, linkers, guests and defects while controlling computational and experimental protocols.

13595 · Conclusions

Cited-study map

Mappings show which printed review references have a verified counterpart in the frozen primary corpus.

Show 28 cited-study records

Reference	Study	Role and context	Corpus mapping
Ref. 1042019	Title unavailable	solvation_caveat · ml_candidateSelected for the review caveat that conductivity can be sensitive to solvated state.	research_0052
Ref. 782014	Title unavailable	open_shell_caveat · band_gap_benchmarkCited for transition-metal/open-shell and DFT band-gap limitations in MOFs.	Unmapped
Ref. 922015	Title unavailable	band_gap_benchmark · hitp_familyOriginal Ni-HITP/Cu-HITP band-gap benchmark used in Table 1 and review discussion.	Unmapped
Ref. 1362019	Title unavailable	database_contextDatabase context for ML screening and the scale of available computational MOF structures.	Unmapped
Ref. 1532014	Title unavailable	ml_candidate · band_gap_benchmarkSource for Table 2 reported band gaps of selected Fe-containing ML-screened candidates.	Unmapped
Ref. 762013	Title unavailable	open_shell_caveatCited in the review for difficulty of describing transition metals with unpaired electrons.	Unmapped
Ref. 512021	Title unavailable	band_structure · transport_benchmarkOriginal HHTT study used by the review for band-structure/DoS comparison and zero-band-gap family benchmark.	Unmapped
Ref. 912016	Title unavailable	metal_substitution · hitp_familyCited for further M-HITP metal substitution after Ni-HITP calculations.	Unmapped
Ref. 1002017	Title unavailable	guest_doping · band_gap_benchmarkProvides DSNDI-MOF-74 and TTF-doped DSNDI-MOF-74 band-gap benchmarks and donor-acceptor interpretation.	research_0182
Ref. 1412018	Title unavailable	ml_screening · metallicity_classificationFirst ML application described by the review for conductive MOF metallicity prediction using descriptors and multi-voting.	research_0682
Ref. 772013	Title unavailable	open_shell_caveat · transition_metalsCited in the review discussion of electronic-structure and open-shell/transition-metal DFT limitations.	Unmapped
Ref. 882015	Title unavailable	guest_doping · transport_mechanismUsed for computational mechanism of TCNQ-induced conductance in HKUST-1.	Unmapped
Ref. 532015	Title unavailable	band_gap_benchmark · uio_familySelected because Table 1 and text use UiO-66 HSE/PBE/experiment comparison as a key band-gap benchmark.	Unmapped
Ref. 141964	Title unavailable	method_foundationFoundational density-based DFT reference used in the review explanation of electronic structure calculations.	Unmapped
Ref. 462006	Title unavailable	method_foundation · hybrid_functionalHybrid-functional reference cited for high-level band-gap and geometry calculations.	Unmapped
Ref. 131965	Title unavailable	method_foundationFoundational DFT reference cited in the review quantum-mechanical methods discussion.	Unmapped
Ref. 82020	Title unavailable	transport_pathway_reviewCited by the review for through-bond and through-space charge-transport pathway framing.	Unmapped
Ref. 402015	Title unavailable	band_gap_benchmark · flexible_frameworksOriginal M-MIL-53 conformation/band-gap benchmark used for flexibility-driven electronic tuning.	Unmapped
Ref. 1542010	Title unavailable	ml_candidate · band_gap_benchmarkSource for Table 2 Cu(TCNQCl2) experimental low-band-gap benchmark.	Unmapped
Ref. 602014	Title unavailable	linker_functionalisation · homo_lumoUsed for halogen/linker modification HOMO-LUMO and band-gap trends.	Unmapped
Ref. 1052021	Title unavailable	ml_screening · qmof · band_gap_predictionMain review source for QMOF, CGCNN/SOAP performance, ML-screened candidates and Table 2 context.	Unmapped
Ref. 1502017	Title unavailable	ml_candidate · band_gap_benchmarkSource for the WAQMEJ low-band-gap candidate reported in Table 2 and Fig. 10 context.	Unmapped
Ref. 452014	Title unavailable	geometry_optimisation · vdw_correctionCited as an example where PBE-D2 accounted for strong van der Waals interactions in layered Ni-HITP.	Unmapped
Ref. 932020	Title unavailable	band_structure · dimensionalityUsed for lanthanide-HHTP anisotropic band-gap and stacking-distance interpretation.	research_0047
Ref. 1292014	Title unavailable	guest_doping · transport_benchmarkOriginal TCNQ infiltration work summarised for conductivity increase and charge-transfer mechanism.	research_0088
Ref. 1282018	Title unavailable	linker_functionalisation · band_gap_benchmarkOriginal functionalised UiO-66(Ce) study used for band-gap narrowing benchmark.	Unmapped
Ref. 62020	Title unavailable	conductive_mof_context · transport_caveatCited by the review for broader conductive-MOF context and caveats around defects and transport.	Unmapped
Ref. 542013	Title unavailable	band_gap_benchmark · dft_uOriginal MOF-74 DFT+U band-gap example used to discuss system-specific correction schemes.	Unmapped