Resolving Band Structure and DOS Mismatches: A Comprehensive Troubleshooting Guide for Computational Researchers

Abstract

This article provides a systematic framework for identifying, diagnosing, and resolving discrepancies between band structure and density of states (DOS) calculations in computational materials science. Covering fundamental principles through advanced validation techniques, we address common pitfalls including k-point sampling inadequacies, smearing parameter selection, and methodological inconsistencies. Through practical troubleshooting protocols and comparative analysis methods, this guide enables researchers to achieve consistent electronic structure characterization essential for reliable materials design and drug development applications.

Understanding the Roots of Band Structure-DOS Discrepancies

Fundamental Differences Between Band Structure and DOS Calculation Methods

Frequently Asked Questions (FAQs)

Q1: Why do my band structure and density of states (DOS) plots show different band gaps?

The band gap derived from a band structure calculation can differ from the one obtained from the DOS due to fundamental differences in how these properties are computed [1] [2].

• Different k-space Sampling: The DOS calculation typically uses a uniform k-point grid that samples the entire Brillouin Zone (BZ). In contrast, a band structure calculation is performed along specific high-symmetry lines between high-symmetry points [3] [2]. The conduction band minimum (CBM) and valence band maximum (VBM) might lie at a k-point not included in the high-symmetry path but captured by the uniform k-grid used for the DOS [2].
• Interpolation vs. Direct Calculation: The band gap printed in output files is often from an "interpolation method" that quadratically interpolates bands over the whole BZ. The band structure plot uses a "from band structure method" with a dense sampling along a specific path. These methods can yield different results for the gap [1].

Q2: My DOS shows states in an energy region where the band structure has a gap. What is wrong?

This common problem, known as "missing DOS," is almost always caused by insufficient k-point sampling during the self-consistent field (SCF) calculation that generates the charge density [4] [1]. A coarse k-grid fails to accurately capture the electronic states across the entire BZ. The solution is to restart the calculation with a finer k-point grid (improved KSpace%Quality or a denser Monkhorst-Pack grid) [4] [1]. Additionally, ensure the energy grid for the DOS is sufficiently fine by decreasing the DOS%DeltaE parameter [1].

Q3: What should I do if my calculation fails due to a "dependent basis" error?

A "dependent basis" error indicates that the set of Bloch functions constructed from your basis set is nearly linearly dependent, threatening numerical accuracy [1]. Do not simply adjust the dependency criterion to bypass the error [1]. Instead, address the root cause by adjusting the basis set. The most common solution is to use confinement to reduce the range of diffuse basis functions, which are usually the culprits, especially in highly coordinated systems [1].

Troubleshooting Guide: Band Structure and DOS Mismatch

Problem Definition

A mismatch occurs when the electronic information from a band structure calculation does not align with the information from the Density of States (DOS). This includes discrepancies in the band gap, the presence or absence of states at specific energies, or the general shape of spectral features [1] [2].

Diagnosis and Resolution Protocol

Follow this workflow to diagnose and resolve the mismatch.

Step 1: Verify K-point Grid Convergence

The most common cause of mismatch is an inadequately converged k-point grid during the initial self-consistent charge calculation [3] [1].

• Action: Perform a convergence test. Run a series of SCF calculations with progressively denser k-point grids (e.g., SupercellFolding with 4x4x4, 8x8x8, etc.) and monitor the total energy. A well-converged calculation shows minimal energy change with a denser grid [3].
• Protocol:
  • Use your primary input with varying KPointsAndWeights blocks.
  • Set Scc = Yes and SccTolerance = 1e-5 (or tighter) [3].
  • Copy the resulting charges.bin from the converged calculation for subsequent band structure and DOS runs.

Step 2: Address Band Gap Discrepancies

If the band gap from the DOS and band structure differ, it's often because the CBM/VBM are not on the high-symmetry path [1] [2].

• Action: Recompute the band gap directly from the DOS data, which samples the entire BZ and is often more reliable for this specific property [2].
• Protocol (using pymatgen with Materials Project API):

Step 3: Resolve Feature Mismatches and Artifacts

• Missing DOS Peaks: If the band structure shows bands in an energy range where the DOS is zero, the k-grid for the DOS is likely too coarse. Restart the DOS calculation with a significantly denser k-point grid [4] [1].
• Energy Shifts: Ensure the Fermi level is consistent. Use the same reference energy when plotting both band structure and DOS.

Experimental Protocols for Robust Calculations

Protocol 1: Standard Workflow for Band Structure and DOS

This protocol ensures consistent results by using a well-converged charge density as the starting point for both non-self-consistent field (NSCF) calculations [3] [2].

Detailed Steps:

• Self-Consistent Field (SCF) Calculation

  • Purpose: Obtain a converged charge density for the system [3].
  • Input Settings:
    • Hamiltonian = DFTB { Scc = Yes } (or equivalent in your code).
    • SccTolerance = 1e-5 or lower [3].
    • Use a dense, uniform k-point grid (e.g., a SupercellFolding or Monkhorst-Pack scheme equivalent to at least 4x4x4 for a simple solid, but test for convergence) [3].
  • Output: The crucial charges.bin file.

• Band Structure Calculation (NSCF)

  • Purpose: Calculate eigenvalues along a high-symmetry path.
  • Input Settings:
    • ReadInitialCharges = Yes (Crucial: reads charges.bin).
    • MaxSCCIterations = 1 (since no new SCC is needed).
    • KPointsAndWeights = Klines { ... } with a list of high-symmetry points and the number of k-points between them [3].

• DOS Calculation (NSCF)

  • Purpose: Calculate eigenvalues on a dense, uniform k-grid for DOS.
  • Input Settings:
    • ReadInitialCharges = Yes.
    • MaxSCCIterations = 1.
    • A dense, uniform k-point grid (often denser than the SCF grid).
    • Enable DOS and PDOS output (e.g., Analysis = { ProjectStates { ... } } in DFTB+) [3].

Protocol 2: K-point Convergence Testing

This protocol establishes a reliable k-point grid for the SCF calculation.

• Identify a key observable: Total energy is a sensitive probe [3].
• Run a series of calculations: Using identical inputs except for the k-point grid. Systematically increase the grid density (e.g., 2x2x2, 4x4x4, 6x6x6, 8x8x8).
• Analyze results: Plot the total energy versus the inverse of the k-grid density. The grid is considered converged when the energy change is smaller than your desired accuracy (e.g., 1e-3 eV) [3].

The Scientist's Toolkit: Research Reagent Solutions

Table 1: Essential Computational "Reagents" for Band Structure and DOS Calculations

Item / Software Tool	Function / Purpose	Key Parameters & Notes
DFTB+	An efficient software for electronic structure calculations using Density Functional based Tight Binding (DFTB). Used for calculating band structures, DOS, and PDOS [3].	SccTolerance, KPointsAndWeights, ReadInitialCharges. Requires Slater-Koster parameter files (e.g., mio, tiorg) [3].
dptools Package	A set of utilities distributed with DFTB+. Contains scripts for post-processing results [3].	dp_dos tool converts band.out to plottable DOS files. Use -w flag for PDOS files [3].
VASP	A widely used software for performing ab initio quantum mechanical calculations using Density Functional Theory (DFT).	Key for computing ground-state densities, band structures, and DOS with PAW pseudopotentials.
pymatgen	A robust Python library for materials analysis. Provides powerful tools for analyzing DOS and band structure objects [2].	get_gap(), get_cbm_vbm() methods. Can interface with databases like the Materials Project [2].
K-point Grid	The sampling mesh in reciprocal space. The most critical "reagent" for convergence [3] [2].	Use SupercellFolding or MonkhorstPack. Must be tested for convergence for both SCF and DOS [3].
Slater-Koster Files	Parameterized files containing integrals for the DFTB Hamiltonian. Act as the "basis set" for DFTB+ calculations [3].	Examples: mio, tiorg. Must be specified in the input via SlaterKosterFiles [3].
1-Octyn-3-one, 1-phenyl-	1-Octyn-3-one, 1-phenyl-, CAS:62059-61-4, MF:C14H16O, MW:200.28 g/mol	Chemical Reagent
Octyl galactofuranoside	Octyl galactofuranoside, MF:C14H28O6, MW:292.37 g/mol	Chemical Reagent

Frequently Asked Questions

1. Why does the sum of my projected density of states (PDOS) not match the total DOS? This is often due to an incomplete basis set for the projections. Specifically, the pseudopotential file used may only contain atomic wavefunctions for a limited set of orbitals (e.g., 2s and 2p for carbon). If the calculation includes bands with higher orbital character (e.g., d-bands) that are not present in the pseudopotential, those states will not be captured in the PDOS, causing a mismatch with the total DOS [5]. The solution is to ensure your pseudopotential includes the necessary orbital channels for the energy range you are investigating.

2. Why is there a discrepancy between the band gap reported in my DOS calculation and my band structure calculation? The DOS and band structure are typically calculated using different k-point grids. The uniform k-point grid used for DOS might not include the specific k-points where the valence band maximum (VBM) or conduction band minimum (CBM) occurs, which are precisely mapped in a line-mode band structure calculation. Therefore, the band gap from the DOS can differ from the fundamental gap determined from the band structure [2]. It is recommended to use the band structure for determining the fundamental gap.

3. My calculation shows a 0 eV band gap for a material known to be an insulator. Is this a physical result or an error? A reported 0 eV band gap can have several causes. It could be a physical result if the material is actually a metal or semimetal. However, it can also be a limitation of the DFT functional (like GGA-PBE, which is known to underestimate band gaps), a parsing artifact in the database, or an issue with the automatic detection of band edges in a complex DOS [2]. The band gap should be recomputed manually from the DOS or band structure data to verify.

4. My SCF calculation does not converge for a metallic system. What can I do? Systems with metallic character can be more challenging to converge. You can adopt more conservative settings, such as decreasing the SCF mixing parameter or the DIIS dimension (DIIS%Dimix) [1]. Alternatively, using a finite electronic temperature at the beginning of a geometry optimization can help achieve initial convergence, with the temperature being reduced as the geometry optimizes [1].

5. I cannot see core-level bands or DOS peaks in my visualization. What is wrong? By default, the energy window for saving bands is often limited. To see deep core levels, you need to increase the EnergyBelowFermi parameter in the band structure or DOS input to a value large enough to encompass the core states (e.g., 10000 eV) [1]. Additionally, you must ensure that no frozen core approximation is active by setting the frozen core to None [1].

Troubleshooting Guide

Symptom 1: Missing Peaks in Projected DOS (PDOS)

When the summed PDOS across all atoms or orbital types does not equal the total DOS, particularly at higher energies, follow this diagnostic workflow:

Diagnosis and Solution: The most common cause is that the pseudopotential (PSP) used for the projection lacks the higher atomic orbital channels needed to describe the conduction bands. For example, a carbon pseudopotential might only contain 2s and 2p orbitals. If the conduction bands have significant 3d character, these states will be absent from the PDOS [5].

Protocol:

• Inspect Pseudopotential: Check the header of your pseudopotential file to see which orbital channels (e.g., 3s, 3p, 3d) are included and which are neglected.
• Identify Missing Orbitals: Based on the atomic species and energy range of interest, identify the missing orbital channels. Consultation with electronic structure literature for the element is helpful.
• Generate New Pseudopotential: Use the ld1.x code (part of Quantum ESPRESSO) or a similar tool to generate a new pseudopotential that includes the previously neglected orbital channels [5].
• Re-run Calculation: Perform a new self-consistent field (SCF) calculation and subsequent PDOS calculation using the updated pseudopotential.

Symptom 2: Band Gap Inconsistencies between DOS and Band Structure

Discrepancies in the reported band gap value when comparing DOS and band structure outputs are often a sampling issue.

Diagnosis and Solution: The DOS is calculated on a uniform k-point grid that samples the entire Brillouin zone. The band structure is calculated along a specific high-symmetry path. It is possible that the CBM or VBM lies at a k-point not on this path or not sampled by the uniform grid. The band gap from the band structure is generally considered the fundamental gap, while the DOS gap can be different if the k-grid misses the extrema [2].

Protocol for Verification:

• Converge K-point Grid: Systematically increase the density of the k-point grid (KSpace%Quality) for the DOS calculation until the band gap value stabilizes [1].
• Use Band Structure for Gap: Rely on the band structure calculation for reporting the fundamental band gap, as it can interpolate between k-points to find the true VBM and CBM [1] [6].
• Manual Inspection: Manually inspect the band structure plot and the DOS near the Fermi level to identify the locations of the VBM and CBM.

Symptom 3: Unexpected Zero Band Gap

When a material is calculated or reported to have a 0 eV band gap unexpectedly.

Diagnosis and Solution: This can be a true physical result, a known DFT error (band gap underestimation), or a parsing/analysis artifact [2].

Verification Protocol (using the Materials Project API and pymatgen):

• Recompute from DOS:
  If this gives a non-zero value, the initial 0 eV gap was likely a parsing error [2].
• Recompute from Band Structure (using VBM from DOS): If the Fermi level in the band structure data is misplaced, you can correct it using the VBM from the more robust DOS calculation [2].

The Scientist's Toolkit: Research Reagent Solutions

The following table lists key computational parameters and their functions, which are essential for diagnosing and resolving DOS and band gap issues.

Item/Parameter	Primary Function	Troubleshooting Role
K-point Grid Quality (KSpace%Quality)	Determines the sampling density of the Brillouin Zone.	A coarse grid can cause DOS peaks to be missing or smeared and lead to band gap inconsistencies. Improving k-point quality is a primary convergence step [1] [2].
Pseudopotential (PSP) File	Defines the interaction between ions and valence electrons, including the atomic orbitals available for projection.	An incomplete PSP lacking higher orbitals (e.g., 3d for C) is the primary cause of missing PDOS peaks in the conduction band [5].
Energy Window (EnergyBelowFermi, EnergyAboveFermi)	Sets the energy range for which band structure and DOS data are saved and plotted.	If set too small, it can cause core-level bands and DOS peaks to be missing from the output and visualization [1].
SCF Convergence Parameters (SCF%Mixing, DIIS%Dimix)	Controls the algorithm for achieving self-consistency in the electronic density.	Poor SCF convergence can lead to incorrect total DOS and spurious results. Tuning these is crucial for difficult systems like metals [1].
Band Structure DeltaK	Sets the interpolation step between high-symmetry k-points for band structure plots.	A large DeltaK can result in a non-smooth band structure that might misrepresent the true band dispersion and gap [6].
Furan, 2-butyl-5-iodo-	Furan, 2-butyl-5-iodo-, CAS:558472-48-3, MF:C8H11IO, MW:250.08 g/mol	Chemical Reagent
4-Methyl-2-vinylthiophene	4-Methyl-2-vinylthiophene|RUO	4-Methyl-2-vinylthiophene (C7H8S) for research. A versatile thiophene building block for material science and bioactive compound synthesis. For Research Use Only.

The table below provides a quick reference for key parameters discussed in this guide, their typical symptoms when misconfigured, and recommended actions.

Symptom Pattern	Critical Parameters to Check	Recommended Action / Code Command
Missing PDOS Peaks	Pseudopotential orbital channels	Generate a new PSP with ld1.x including all relevant channels (e.g., 3s, 3p, 3d) [5].
Band Gap Inconsistencies	KSpace%Quality, DOS%DeltaE	Increase k-grid density for DOS; use band structure for fundamental gap [1] [2].
Unexpected 0 eV Gap	- (Often a DFT/parsing issue)	Recompute gap from DOS: dos.get_gap() in pymatgen [2].
Invisible Core Levels	BandStructure%EnergyBelowFermi, Frozen Core	Set EnergyBelowFermi to a large value (e.g., 10000) and FrozenCore to None [1].
Non-Smooth Bands	BandStructure%DeltaK	Decrease DeltaK (e.g., to 0.03) for a smoother band structure plot [6].

The Critical Role of k-point Sampling in BZ Integration

Frequently Asked Questions

• What is k-point sampling and why is it critical? K-points are discrete points used to sample the Brillouin Zone (BZ), the unit cell of the reciprocal lattice. Accurate integration over the BZ is essential because all electronic properties of a crystal, like total energy and charge density, are determined by integrating over this zone [7]. Insufficient sampling leads to errors in total energies, inaccuracies near the Fermi level, and a poor representation of the Density of States (DOS) [7].

• Why might my band structure and Density of States (DOS) show mismatches? This is a common issue rooted in how these two properties are calculated [1].

  • DOS Calculation (Interpolation Method): The DOS is derived from a k-space integration that samples the entire Brillouin zone, typically using interpolation between a grid of k-points [1]. It can therefore capture the true band gap if the k-grid is sufficiently dense.
  • Band Structure Calculation (Band Structure Method): The band structure is calculated by plotting energies along a specific, high-symmetry path in the BZ [1]. While the k-point sampling along this path can be very dense, it might miss the actual points where the valence band maximum (VBM) or conduction band minimum (CBM) occur if they are not located on this path [1]. A converged DOS is often the more reliable method for determining the fundamental band gap [1].

• My calculation is for a metal and won't converge. What can I do? Metals are challenging because of the discontinuous Fermi surface. You can try:

  • Use the tetrahedron method: This method, especially with Blöchl corrections, is often the best choice for small metallic systems as it provides excellent convergence for the single-particle sum [8].
  • Increase k-point density: Metallic systems require a much higher density of k-points to converge the total energy because the Fermi surface creates a discontinuity [7] [9].
  • Employ smearing: Use Methfessel-Paxton or Fermi-Dirac smearing to artificially smear occupational states around the Fermi level. This makes the integrand smoother and accelerates convergence, though it introduces a small, controllable error [8].

• How do I systematically test for k-point convergence?

  • Start with a coarse k-point grid.
  • Gradually increase the sampling density (e.g., from 4×4×4 to 8×8×8, etc.).
  • Monitor the change in the total energy of your system. The calculation is considered converged when the energy change between successive grids falls below a desired threshold (e.g., 1 meV/atom) [7] [10].
  • Create a table to track your progress [7]:

  k-grid	Number of k-points	Total Energy (eV)	ΔE (meV)
  4x4x4	...	...	...
  6x6x6	...	...	...
  8x8x8	...	...	...

Troubleshooting Guides

Problem: Band Structure and DOS Do Not Agree on Band Gap Value

Issue Description A user finds a significant discrepancy (e.g., 0.27 eV for silicon) between the band gap reported by the band structure object and the DOS object in their analysis code [11].

Diagnostic Workflow The following diagram outlines the logical steps to diagnose and resolve a band structure-DOS mismatch.

Resolution Steps

• Verify the k-point grid for DOS: The DOS is calculated using the same k-grid as your self-consistent field (SCF) calculation. If this grid is too coarse, the DOS will be inaccurate. Systematically increase the k-point density for the SCF/DOS calculation and monitor the convergence of the total energy and the band gap [7] [12].
• Check the band structure path: Confirm that the high-symmetry path used for the band structure plot passes through the points in the Brillouin zone where the valence band maximum (VBM) and conduction band minimum (CBM) are actually located. The band structure can only show the gap along the specific path it calculates [1].
• Understand the methods: Recognize that the "interpolation method" used for DOS integration over the entire BZ is often more reliable for finding the true fundamental gap than the "band structure method," which is restricted to a specific path [1]. The gap printed in the main output file (often from the interpolation method) is typically more trustworthy than one estimated from a band structure plot.

Problem: Total Energy Fails to Converge with k-points

Issue Description The total energy of the system does not stabilize as the k-point grid is densified, a problem particularly acute in metallic systems [9].

Resolution Steps

• Switch to a generalized k-point grid: Standard Monkhorst-Pack (MP) grids may not provide the most efficient symmetry reduction. Using generalized regular (GR) grids or grids from modern algorithms (e.g., from the Mueller Group's K-Point Grid Server) can provide better convergence with fewer irreducible k-points, accelerating calculations by a factor of about 2 on average [13] [9].
• Use the tetrahedron method for metals: For metallic systems, the Blöchl-corrected tetrahedron integration method is highly recommended. It can converge the single-particle sum to within a few μRy by improving the error scaling from ( O(n^{-2}) ) to ( O(n^{-4}) ) or better for simple bands [8].
• Employ smearing techniques: Gaussian (Methfessel-Paxton) or Fermi-Dirac smearing can be used to smooth the Fermi-level discontinuity. While this introduces a small finite temperature, it significantly improves convergence behavior with the k-point grid [8].
• Exploit symmetry for specific materials: In some systems, like graphene, including a specific high-symmetry k-point (e.g., the K-point) in your sampling can be more important than using a generally dense grid. This can perfectly pin the Fermi level at the Dirac cone, even with a relatively coarse grid that includes that point [7].

Experimental Protocols

Protocol 1: Systematic k-point Convergence Test

Objective To determine the minimally sufficient k-point grid that yields a total energy converged to within a target accuracy (e.g., 1 meV/atom).

Materials and Reagents

• Software: A DFT code such as SIESTA [7], VASP [10], or Questaal [8].
• System: The crystal structure of interest (e.g., Silicon in a diamond structure) [10].

Methodology

• Initial Setup: Begin with a structurally relaxed system and a well-converged basis set or plane-wave cutoff.
• Grid Generation: Start with a coarse Monkhorst-Pack k-point grid (e.g., 2×2×2). The grid can be defined manually or using an automated parameter like kgridcutoff [7].
• Calculation Series: Perform a single-point energy calculation. Repeat the calculation, progressively increasing the k-point density (e.g., to 4×4×4, 6×6×6, 8×8×8, etc.) [7] [10].
• Data Collection: For each calculation, record the k-grid, the total number of k-points, and the total energy.
• Analysis: Plot the total energy as a function of k-point density or the inverse of the grid size. The energy is considered converged when the change (ΔE) between successive calculations is less than your target threshold [10].

Protocol 2: Resolving Band Gap Mismatch

Objective To obtain a consistent and accurate band gap value from both DOS and band structure calculations.

Materials and Reagents

• Software: A DFT code with separate options for SCF k-grid and band structure k-path.
• System: A semiconductor or insulator (e.g., Silicon).

Methodology

• Converge the SCF k-grid: First, follow Protocol 1 to find a k-grid that converges the total energy. This grid will be used for the DOS calculation.
• Perform a DOS Calculation: Using the converged k-grid, calculate the DOS. Use the tetrahedron method or a small smearing for insulators. Note the band gap from the DOS.
• Calculate the Band Structure: Generate a band structure along a high-symmetry path. Ensure the path is dense (many points between high-symmetry points) and includes all critical points where the VBM and CBM are suspected to be.
• Cross-Reference Results: Compare the gap from the DOS with the gap observed on the band structure plot. The DOS gap is typically more reliable. If they disagree, it is likely the band path does not intersect the true CBM and/or VBM. Consult literature to confirm the location of the CBM/VBM and adjust your band path accordingly [1].

The Scientist's Toolkit

Tool / Reagent	Function in k-point Sampling
Monkhorst-Pack Grids	A systematic method to generate uniform k-point meshes for sampling the Brillouin zone. It is the most common and traditional approach [7].
Generalized Regular (GR) Grids	Advanced k-point grids that can provide better symmetry reduction than standard MP grids, leading to fewer irreducible k-points and faster calculations for the same accuracy [9].
Tetrahedron Method	An integration technique that divides the Brillouin zone into tetrahedra. It is particularly accurate for metals, especially when enhanced with Blöchl corrections [8].
Smearing Methods	Techniques (e.g., Methfessel-Paxton, Fermi-Dirac) that broaden occupational states around the Fermi level. This smoothens integrands, accelerating SCF convergence in metals at the cost of a small, controlled error [8].
K-point Grid Servers	Online tools (e.g., Mueller Group's Server) that generate highly efficient, optimized k-point grids for a given crystal structure, streamlining the setup process [13].
Carbocisteine-13C3	Carbocisteine-13C3, MF:C5H9NO4S, MW:182.17 g/mol
5-(Bromomethyl)isoindoline	5-(Bromomethyl)isoindoline, MF:C9H10BrN, MW:212.09 g/mol

How Smearing Methods and Energy Grids Affect Spectral Resolution

Core Concepts & Definitions

What are smearing methods and why are they used?

Smearing methods, also known as broadening techniques, are computational algorithms used in electronic structure calculations to determine the fractional occupation of electronic states near the Fermi level. In first-principles calculations of metallic systems, the binary occupation of states (fully occupied or completely empty) can lead to numerical instabilities and slow convergence. Smearing techniques introduce a small amount of "smearing" around the Fermi energy, which allows for fractional occupation numbers and significantly improves the convergence behavior of self-consistent field (SCF) cycles, particularly for metals [14].

What is the energy grid and how does it affect spectral resolution?

The energy grid refers to the discrete set of energy points at which the Density of States (DOS) is calculated. The parameter DOS%DeltaE in many computational codes controls the spacing between these energy points [1]. A finer energy grid (smaller DeltaE) provides higher resolution in the DOS, allowing for better identification of sharp spectral features, van Hove singularities, and narrow band gaps. Conversely, a coarser grid can miss these fine details, leading to inaccurate representations of the electronic structure.

Troubleshooting Guides

Band structure does not match the DOS

Problem Description: Researchers often encounter a discrepancy where the electronic band structure plot does not align with the features observed in the Density of States (DOS). Key features, such as band edges or peaks, appear at different energies in the two representations [1].

Diagnosis Flowchart:

Step-by-Step Resolution Protocol:

• Verify K-Space Integration Quality: The DOS is derived from a k-space integration method that interpolates bands across the entire Brillouin Zone (BZ), while the band structure is plotted along a specific high-symmetry path. A mismatch is often caused by an unconverged KSpace%Quality parameter.

  • Action: Systematically increase the KSpace%Quality setting and rerun the DOS calculation. A sufficiently high quality should make the DOS features consistent with the band structure, provided the path covers all relevant features [1].

• Refine the DOS Energy Grid: A coarse energy grid can blur sharp features in the DOS.

  • Action: Decrease the value of DOS%DeltaE to obtain a finer energy grid for the DOS calculation. This improves the resolution and may reveal features that were previously smeared out [1].

• Check Band Path Coverage: The band structure plot is only as good as the chosen path in the BZ. It is possible that the top of the valence band (TOVB) or bottom of the conduction band (BOCB)—and thus the band gap—does not lie on this path.

  • Action: Consult literature or use symmetry analysis to ensure your band path includes all critical points. The true band gap from the "interpolation method" (printed in the output file) is more reliable as it samples the entire BZ [1].

Choosing the correct smearing method

Problem Description: An inappropriate choice of smearing technique and width (SIGMA) can lead to incorrect total energies, unphysical occupation of band gaps, and inaccurate forces, which is particularly detrimental for geometry relaxations and phonon calculations [14].

Diagnosis Flowchart:

Step-by-Step Resolution Protocol:

• Classify Your System: The optimal smearing method depends critically on whether the system is a metal, semiconductor, or insulator.

  • Metals (for relaxations): Use the Methfessel-Paxton method (ISMEAR=1 in VASP). Choose a SIGMA value as large as possible while keeping the entropy term T*S (reported in the output file) negligible (e.g., < 1 meV/atom). A default of SIGMA=0.2 is often a reasonable starting point [14].
  • Semiconductors/Insulators: Use Gaussian smearing (ISMEAR=0) or the tetrahedron method (Blöchl corrections, ISMEAR=-5). Crucially, avoid ISMEAR > 0 as it can lead to unphysical occupation of the band gap and errors in forces exceeding 20% [14].
  • Unknown Systems/High-Throughput: Gaussian smearing (ISMEAR=0) with a small SIGMA (0.03 to 0.1) is the safest and most recommended default [14].

• Converge SIGMA for Gaussian Smearing: When using Gaussian smearing, the total energy must be extrapolated to SIGMA=0. The output file typically reports an extrapolated value. Ensure that both the energy and the forces are converged with respect to a systematic reduction of SIGMA [14].

• Use Tetrahedron for Final DOS: For the calculation of highly accurate total energies or the DOS on a converged structure, use the tetrahedron method (ISMEAR=-5) with a dense k-point mesh. This method provides a sharper representation of band edges compared to smearing methods [14].

Frequently Asked Questions (FAQs)

Q1: I see two different band gaps reported in my output. Which one should I trust? A1: This discrepancy arises from two different evaluation methods. The "interpolation method" (used for k-space integration and the gap printed in the main output) samples the entire Brillouin Zone and is generally more reliable. The "band structure method" only evaluates energies along a specific path. While the latter can use a denser k-point sampling along the path, it relies on the assumption that the band edges lie on that path. For a definitive answer, the gap from the interpolation method is preferable, but the band structure method can confirm the location of the extrema [1].

Q2: My SCF calculation will not converge, especially for a metallic system. How can smearing help? A2: Smearing is specifically designed to address SCF convergence in metals. By allowing fractional occupations near the Fermi level, it prevents large, discontinuous changes in orbital occupations between SCF cycles, which is a primary source of charge sloshing and divergence. Switching from the default ISMEAR=-5 (tetrahedron) to ISMEAR=1 (Methfessel-Paxton) or ISMEAR=0 (Gaussian) with an appropriate SIGMA is often the key to achieving convergence in metallic systems [14]. Other stabilizing measures include decreasing the mixing parameter and using the MultiSecant or LIST DIIS methods [1].

Q3: Why are my phonon frequencies imaginary (negative)? Could this be related to smearing? A3: Yes, the choice of smearing can indirectly cause imaginary frequencies. The two most common causes are: 1) The geometry is not fully optimized to a minimum, and 2) The numerical accuracy of the forces used for the phonon calculation is insufficient. Using an overly large SIGMA or an inappropriate smearing method (e.g., ISMEAR>0 for an insulator) can lead to inaccurate forces and, consequently, unphysical phonon spectra. Always ensure your geometry is fully converged and that your smearing method is appropriate for your system to obtain reliable forces [1] [14].

Q4: For a system that is difficult to converge, should I use a finite electronic temperature? A4: Yes, applying a finite electronic temperature (i.e., Fermi-Dirac smearing, ISMEAR=-1) can significantly improve SCF convergence, much like other smearing techniques. This is often employed in automated workflows during the initial stages of a geometry optimization when forces are still large. The electronic temperature can be set high initially and then automatically reduced as the geometry converges, ensuring accuracy in the final energy [1].

Table 1: Smearing Method Comparison and Guidelines

Smearing Method (VASP ISMEAR)	Best For System Type	Recommended SIGMA	Key Advantages	Key Disadvantages/Cautions
Gaussian (ISMEAR=0)	Unknown, Semiconductors, Insulators	0.03 - 0.1	Safe default; provides energy(SIGMA→0) extrapolation [14].	Forces/stress are consistent with free energy, not extrapolated energy [14].
Methfessel-Paxton (ISMEAR=1)	Metals (for relaxations, forces, phonons)	Set so that T*S < 1 meV/atom [14].	Very accurate total energies for metals; corrects for entropy term [14].	Avoid for semiconductors/insulators; can cause severe errors [14].
Fermi-Dirac (ISMEAR=-1)	Finite-temperature properties	Corresponds to electronic temperature	Physically meaningful for real temperature effects [14].	Other methods are often preferred for ground-state calculations [14].
Tetrahedron + Blöchl (ISMEAR=-5)	Semiconductors, Insulators; Metals (for accurate DOS/final energy)	Not Applicable	Most accurate for DOS and total energies in bulk materials; sharp band edges [14].	Forces can be wrong (5-10%) for metals; not variational [14].

Table 2: Key Parameters for Spectral Resolution Control

Parameter	Typical Function	Effect on Spectral Resolution	Recommended Starting Value
SIGMA	Smearing width (eV)	Larger values smear out spectral features, lower resolution but improve metal SCF convergence [14].	0.1 (Gaussian), 0.2 (Methfessel-Paxton for metals) [14].
DOS%DeltaE	Energy grid spacing (eV)	Smaller values give higher resolution DOS, revealing sharp features [1].	System-dependent; must be converged.
KSpace%Quality	k-point mesh density	Finer mesh improves BZ sampling, essential for matching DOS and band structure [1].	System-dependent; must be converged.

The Scientist's Toolkit: Essential Research Reagents & Computational Parameters

Table 3: Key Computational "Reagents" for Band Structure and DOS Calculations

Item / Parameter	Function / Role	Example "Solution" / Value
Smearing Function	Determines fractional occupancy of states near Fermi level; critical for SCF convergence in metals.	Gaussian (ISMEAR=0), Methfessel-Paxton (ISMEAR=1), Tetrahedron (ISMEAR=-5) [14].
Smearing Width (SIGMA)	Controls the energy width over which states are smeared; a key convergence parameter.	0.1 eV (Gaussian default), 0.2 eV (Methfessel-Paxton for metals) [14].
k-point Mesh (KSpace%Quality)	Defines the sampling of the Brillouin Zone; affects accuracy of integration for DOS and charge density.	"Good" or "High" quality setting; must be converged for the system [1].
Energy Grid (DOS%DeltaE)	Sets the energy resolution for the DOS calculation; finer grid captures sharp features.	A small value (e.g., 0.01 eV); must be tested for convergence [1].
Fermi Energy Setting (EFERMI)	Determines the reference energy (0 eV) for band structures and DOS.	MIDGAP (for gapped systems), LEGACY (default, can be unstable) [14].
2-Thiazolamine, 5-ethoxy-	2-Thiazolamine, 5-ethoxy-, MF:C5H8N2OS, MW:144.20 g/mol	Chemical Reagent
Dextrounifiram	Dextrounifiram, CAS:865717-10-8, MF:C13H15FN2O3S, MW:298.34 g/mol	Chemical Reagent

Troubleshooting Guide: Band Structure and DOS Mismatch

This guide addresses common computational and experimental challenges researchers face when investigating the band structure and density of states (DOS) in materials like anatase TiOâ‚‚ and doped MgFâ‚‚.

FAQ: Resolving Discrepancies Between Calculated and Experimental Band Gaps

Q: My DFT-calculated band gap for anatase TiOâ‚‚ is significantly smaller than the experimental value (3.2 eV). What is the cause and how can I resolve this?

• A: This is a known limitation of standard DFT functionals (LDA, GGA), which underestimate band gaps. To address this:
  • Use advanced exchange-correlation functionals: Employ the Hubbard +U model (DFT+U) to account for strong electron correlations. For TiOâ‚‚, a U value of 8.2 eV applied to Ti 3d orbitals has been used successfully to yield a band gap of 3.14 eV, close to the experimental value [15] [16].
  • Consider hybrid functionals: Functionals like HSE06 incorporate a portion of exact Hartree-Fock exchange, which typically opens up the band gap to more experimental values.
  • Ensure convergence: Verify that your plane-wave cutoff energy and k-point sampling are well-converged. For anatase TiOâ‚‚, a 4×4×4 Monkhorst-Pack k-point grid has been shown to provide good convergence for the total energy [17].

Q: The projected DOS (PDOS) for my doped system shows unexpected peaks deep in the band gap. Are these physical or a sign of an error?

• A: Deep gap states can be physical, originating from certain types of defects or dopants, but they can also indicate problematic computational settings.
  • Identify the dopant's role: Metallic dopants like Cr or Fe in TiOâ‚‚ can introduce deep levels, while non-metallic dopants like F often create shallower states [16]. In MgFâ‚‚, Co²⁺ doping introduces a ground state level about 2 eV above the valence band maximum [18].
  • Check for spurious interactions: If using a supercell model for doping, ensure the defect concentration is not artificially high, which can lead to interactions between periodic images of the dopant. Increase your supercell size to dilute the dopant concentration and check if the gap states persist.
  • Verify dopant stability: Confirm the most stable configuration for your dopant. For F in TiOâ‚‚, substitutional doping (replacing an oxygen atom) is often more energetically favorable than interstitial doping [15].

Q: My band structure plot shows unphysical spikes or discontinuities. What could be wrong?

• A: This is often related to the k-point path used in the non-self-consistent field (NSCF) calculation.
  • Ensure a connected path: The k-points for the band structure calculation must be defined along a continuous, high-symmetry path in the Brillouin zone. A common path for anatase TiOâ‚‚ is Z → Γ → X → P [17] [19].
  • Use sufficient points: Between high-symmetry points, use an adequate number of k-points (e.g., 20-50) to ensure a smooth band dispersion [17].

FAQ: Challenges in Doping and Defect Engineering

Q: I am synthesizing F-doped TiOâ‚‚, but my material does not show the expected red-shift in absorption or improved photocatalytic activity. What might have gone wrong?

• A: The effectiveness of F-doping depends heavily on the successful incorporation of F atoms into the TiOâ‚‚ structure and their specific coordination.
  • Confirm F incorporation and configuration: Use XPS to check the chemical state of F. Two XPS peaks in the F1s region (~684 eV for Ti-F bonds and ~688 eV for surface-adsorbed F) indicate successful doping [20]. The most active configuration for visible-light response is often the surface Ti₃-F species [20].
  • Optimize synthesis parameters: The doping concentration is critical. For TiOâ‚‚ nanorods, a precursor concentration of 0.05 mol/L NHâ‚„F was optimal, yielding a photocurrent 4.61 times higher than pure TiOâ‚‚. Higher concentrations can be detrimental [21].
  • Check for charge compensation: Doping can introduce charge imbalances that are compensated by native defects (e.g., oxygen vacancies), which may themselves alter the electronic structure.

Q: When I computationally model doped MgFâ‚‚, how do I achieve a reliable band gap for the host material?

• A: Pure MgFâ‚‚ is a wide-band-gap insulator, and standard DFT severely underestimates its gap.
  • Incorporate exact exchange: As demonstrated in studies of Co-doped MgFâ‚‚, including a portion of non-local exact exchange is crucial for properly reproducing the band gap and other bulk properties [18]. Hybrid functionals (e.g., PBE0, HSE) are recommended for this class of materials.

Table 1: Band Gap Modification in Doped Anatase TiOâ‚‚

Material System	Dopant Type/Concentration	Calculation Method	Band Gap (eV)	Key Change vs. Pure TiOâ‚‚	Experimental Validation
Pure Anatase TiOâ‚‚	-	DFT+U (U=8.2)	3.14 [15]	Reference	~3.2 eV [15]
F:Hi:TiOâ‚‚	F, H Interstitial	DFT+U	3.00 [15]	-0.14 eV	~3.0 eV [15]
Fo:Hi:TiOâ‚‚	F (Substitutional), H	DFT+U	2.60 [15]	-0.54 eV	Not Specified
Cl-doped TiOâ‚‚	Cl	DFT+U	Significant Reduction [15]	Introduction of gap states	Enhanced Hâ‚‚ production rate [15]
F-doped TiOâ‚‚ Nanorods	F (0.05M NHâ‚„F)	Experimental (UV-Vis)	Optimized [21]	Red shift & improved absorption	6.58x higher Hâ‚‚ production [21]

Table 2: Electronic Structure of Doped MgFâ‚‚ and Other Modifications

Material System	Dopant/Modification	Calculation Method	Key Electronic Finding	Experimental Correlation
Pure MgFâ‚‚	-	Ab initio (with non-local exchange)	Properly reproduced band gap [18]	Reliable baseline for calculations [18]
Co:MgF₂	Co²⁺ (3d⁷)	Ab initio	Ground state level ~2 eV above valence band top [18]	Good agreement with experimental data [18]
Anatase under Strain	8 GPa Biaxial Tensile Strain	DFT+U	Band gap reduction to 2.96 eV [16]	Epitaxial growth on lattice-mismatched substrates [16]
Cr-doped Anatase under Strain	Cr + 8 GPa Strain	DFT+U	Band gap reduction to 2.4 eV [16]	Not Specified

Experimental Protocols

Objective: To synthesize fluorine-doped TiOâ‚‚ nanorod arrays (F-T) on FTO glass for enhanced photoelectrochemical (PEC) water splitting.

Materials:

• Substrate: Fluorine-doped tin oxide (FTO) conductive glass.
• Precursors: Tetrabutyl titanate (TBOT), concentrated hydrochloric acid (HCl, 37%).
• Dopant Source: Ammonium fluoride (NHâ‚„F) solutions (0.01, 0.05, 0.1 mol/L).
• Solvents: Acetone, ethanol, deionized water.

Procedure:

• Substrate Cleaning: Clean FTO glass via ultrasonication in acetone, ethanol, and deionized water for 30 minutes each. Dry under a nitrogen stream.
• Hydrothermal Growth:
  • Mix deionized water and concentrated HCl in a 1:1 volume ratio.
  • Slowly add 0.6 ml of TBOT to the acid solution under vigorous stirring until the solution becomes clear.
  • Transfer the solution to a Teflon-lined autoclave and place the cleaned FTO glass inside, conductive side facing up.
  • Heat the autoclave at 170 °C for 6 hours in an oven.
  • After cooling, remove the FTO substrate, now coated with a white TiOâ‚‚ film, and rinse with ultrapure water.
• Fluorine Doping:
  • Soak the as-grown, unannealed TiOâ‚‚ nanorod array in an NHâ‚„F solution (e.g., 0.05 mol/L) for 5 minutes.
  • Remove the sample, discard excess solution, and dry with a nitrogen stream.
• Annealing: Place the sample in a tubular furnace and anneal at 450 °C for 1 hour under an argon atmosphere.

Characterization & Validation:

• PEC Performance: Use a standard 3-electrode system (e.g., CHI 760E electrochemical workstation) to measure photocurrent density. Optimally doped samples (0.05F-T) show a photocurrent of 7.34 mA/cm² at 1.8 V vs. RHE [21].
• Hydrogen Production: Measure Hâ‚‚ evolution under illumination. A 6.58-fold increase over pure TiOâ‚‚ is indicative of successful doping [21].
• Structural Analysis: Use XRD and TEM to confirm the anatase phase and successful F incorporation, which may manifest as changes in lattice spacing [21].

Objective: To calculate the electronic band structure, total DOS, and projected DOS (PDOS) of a periodic system like anatase TiOâ‚‚ using DFTB+.

Procedure:

• Ground-State Calculation (Self-Consistent Charge):
  • Geometry: Provide the crystal structure in GenFormat, specifying fractional coordinates.
  • Hamiltonian: Set SCC = Yes with a tight tolerance (e.g., SccTolerance = 1e-5). Use appropriate Slater-Koster files (e.g., mio set).
  • k-Points: Use a dense Monkhorst-Pack grid (e.g., 4x4x4 generated via SupercellFolding) to obtain converged charges.
  • PDOS Setup: In the Analysis block, use ProjectStates to define regions (e.g., Atoms = Ti and Atoms = O) with ShellResolved = Yes to output PDOS for individual atomic shells.
  • Execution: Run DFTB+ to generate band.out and PDOS files (e.g., dos_ti.1.dat, dos_o.1.dat).

• Band Structure Calculation (Non-SCC):
  • Input Charges: Copy the charges.bin file from the previous step and set ReadInitialCharges = Yes.
  • k-Points Path: Use the Klines method to define a path through high-symmetry points (e.g., Z, Γ, X, P for anatase), specifying the number of points between each.
  • Execution: Run DFTB+ with MaxSCCIterations = 1 to calculate eigenvalues along the specified path.

Post-Processing:

• Total DOS: Use dp_dos band.out dos_total.dat to generate the total DOS file.
• PDOS: Use dp_dos -w dos_ti.1.out dos_ti.s.dat (with the -w flag for weighting) to generate the PDOS for each atomic shell and species.

Conceptual Diagrams

F-TiO2 Synthesis and Effect Diagram

Diagram 1: F-TiO2 synthesis workflow and effects.

DOS Mismatch Troubleshooting Logic

Diagram 2: DOS mismatch troubleshooting logic.

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential Materials for TiOâ‚‚ and MgFâ‚‚ Doping Studies

Reagent / Material	Function / Role	Application Example	Critical Parameters
Ammonium Fluoride (NHâ‚„F)	Fluorine dopant source for TiOâ‚‚. Introduces F ions that can substitute for O or adsorb on the surface.	Synthesis of F-doped TiOâ‚‚ nanorods for enhanced PEC water splitting [21].	Concentration is critical (e.g., 0.05M optimal). Soaking time (5 min).
Tetrabutyl Titanate (TBOT)	Titanium precursor for the sol-gel and hydrothermal synthesis of TiOâ‚‚ nanostructures.	Hydrothermal growth of pristine TiOâ‚‚ nanorod arrays on FTO glass [21].	Purity, controlled hydrolysis in acidic conditions (HCl).
Slater-Koster Files (mio, tiorg)	Parameter sets containing pre-computed integrals for DFTB calculations. Essential for electronic structure simulations.	DFTB+ calculation of band structure and PDOS for anatase TiOâ‚‚ [17].	File path must be correctly specified in input. Must match element pairs (e.g., Ti-Ti, Ti-O, O-O).
FTO Conducting Glass	Transparent conducting oxide substrate. Serves as both the growth substrate and working electrode.	Growth of TiOâ‚‚ nanorod arrays for photoanode fabrication [21].	Surface cleanliness prior to synthesis is paramount.
Cobalt Fluoride (CoF₂)	Source of Co²⁺ ions for doping wide-band-gap fluorides like MgF₂.	Ab initio studies of Co-doped MgF₂ crystals for defect energy level analysis [18].	Purity, controlled incorporation during crystal growth.
Tolinapant dihydrochloride	Tolinapant dihydrochloride, CAS:1799328-50-9, MF:C30H44Cl2FN5O3, MW:612.6 g/mol	Chemical Reagent	Bench Chemicals
Gnetinc	Gnetinc, MF:C28H22O6, MW:454.5 g/mol	Chemical Reagent	Bench Chemicals

Robust Protocols for Converged Band Structure and DOS Calculations

Optimal k-point Grid Strategy for SCC Convergence and Band Accuracy

Frequently Asked Questions

What are the two primary convergence parameters I must check in DFT calculations? For all first-principles calculations, you must pay attention to two key convergence issues: the planewave energy cutoff (ecutwfc), which limits the wave-function expansion, and the number of k-points, which determines how well your discrete grid approximates the continuous integral over the Brillouin zone [22].

Why is a k-point convergence study necessary? The convergence of properties like total energy and band gap with respect to k-point density is "neither variational nor necessarily monotonous" [23]. Therefore, systematically testing different k-grids is the only reliable method to ensure your results are converged for your specific system and property of interest.

I found two different band gap values in my output. Which one is correct? The band gap can be determined by two main methods [1]:

• The "interpolation method": This method uses the analytical k-space integration scheme that determines the Fermi level and occupations. The gap printed in the main output file (e.g., the .kf file in BAND) typically comes from this method.
• The "band structure method": This is a post-SCF method that calculates bands along a high-symmetry path. It often uses a denser k-point sampling along that path and is generally considered the better way to find the gap, provided the path contains both the valence band maximum and conduction band minimum.

Why does my band structure plot not match my Density of States (DOS)? This common problem can have several causes [1]:

• Different k-space sampling: The DOS is derived from an interpolation method that samples the entire Brillouin zone, while the band structure is plotted along a single line. They are fundamentally different representations.
• Unconverged DOS: The DOS may not be converged with respect to the KSpace%Quality parameter. Try increasing this value.
• Insufficient energy grid: The energy grid for the DOS might be too coarse. You can make it finer using the DOS%DeltaE parameter.

My SCF calculation will not converge. What can I do? SCF convergence problems, especially in metallic systems or slabs, can often be resolved with more conservative settings [1]:

• Decrease the SCF%Mixing parameter and/or the DIIS%Dimix parameter.
• Try alternative SCF methods like the MultiSecant method or the LISTi variant of the DIIS method.
• For geometry optimizations, use "automations" to start with a higher electronic temperature and looser SCF criteria, which are then tightened as the geometry converges.

Troubleshooting Guides

Problem: Band Structure and DOS Mismatch

Issue: The features in your calculated density of states (DOS) do not align with the expected energy levels from your band structure plot.

Diagnosis and Solution:

• Verify k-space quality: The most common cause is that the DOS is not converged with respect to the k-point grid. The DOS requires a dense, uniform grid to accurately integrate over the entire Brillouin Zone, whereas the band structure uses a dense sampling along a specific path [1].
  • Action: Systematically increase the KSpace%Quality parameter (or equivalent in your code) and rerun the DOS calculation. The results are converged when the DOS no longer changes significantly with a finer k-grid.

• Check the band path: The band structure plot is only as good as the path chosen through the Brillouin Zone. It is possible that the chosen path misses the specific k-points where the valence band maximum or conduction band minimum occur [1].

  • Action: Ensure your band structure path includes all high-symmetry points. You may need to consult literature for your specific crystal structure to identify the correct path.

• Refine the DOS energy grid: A coarse energy grid can smear out sharp features in the DOS.

  • Action: Decrease the value of DOS%DeltaE to use a finer energy grid for calculating the DOS [1].

Problem: Significant Mismatch in Band Gap Values

Issue: Different methods or classes within your computational code report significantly different band gap values (e.g., a difference of 0.27 eV for Silicon) [11].

Diagnosis and Solution:

• Identify the source: Understand which method each reported gap comes from. As outlined in the FAQ, the gap from the Bandstructure object and the Dos object are computed differently [11] [1].
• Trust the band structure method: For determining the fundamental band gap, the "band structure method" is often more reliable as it can use a much denser sampling along a path to pinpoint critical points [1].
• Ensure consistency: For a valid comparison, both the band structure and DOS calculations must be based on the same self-consistent density and potential. Always use a well-converged SCF calculation as the starting point for both.

K-Point Convergence: Data and Protocols

The following table exemplifies the results of a k-point convergence study for a solid-state system, showing how total energy and band gap evolve with an increasingly dense k-grid [23].

Table 1: Example K-Point Convergence Data

k-grid	k-point Density	Total Energy (eV)	Band Gap (eV)	Converged?
2x2x2	1.01	-7900.073544	0.67	True
4x4x4	2.01	-7901.237159	0.77	True
6x6x6	3.02	-7901.317293	0.78	True
8x8x8	4.02	-7901.327057	0.67	True
10x10x10	5.03	-7901.328599	0.64	True
12x12x12	6.03	-7901.328883	0.64	True (1.0E-04 eV/atom)
16x16x16	8.04	-7901.328954	0.64	True (1.0E-05 eV/atom)
18x18x18	9.05	-7901.328957	0.65	True (1.0E-06 eV/atom)

This data shows that while the total energy converges monotonically, the band gap can oscillate (e.g., between 0.78 eV and 0.67 eV) before stabilizing, underscoring the need for thorough testing [23].

Experimental Protocol: K-Point Convergence Workflow

The following diagram illustrates the recommended workflow for performing a k-point convergence study.

Step-by-Step Methodology:

• Preparation: Start with a fully optimized crystal structure and a converged planewave energy cutoff.
• Grid Selection: Define a series of increasingly dense k-point grids (e.g., 2x2x2, 4x4x4, 6x6x6, ...). Automated workflows like the KPointConvergence class in aimstools can facilitate this task [23].
• Execution: Perform self-consistent field (SCF) calculations for each defined k-point grid.
• Data Extraction: For each calculation, extract the property of interest—most commonly the total energy per atom and the fundamental band gap.
• Convergence Analysis: Plot the property values against the k-point density or grid size. The property is considered converged when its change with a denser grid falls below a predefined threshold. Common thresholds are:
  • 1.0E-04 eV/atom: For sparse k-grids, suitable for initial geometry optimizations.
  • 1.0E-05 eV/atom: For relatively dense k-grids, used in production calculations.
  • 1.0E-06 eV/atom: For very dense k-grids, typically needed for metallic systems and optical spectra [23].

The Scientist's Toolkit: Essential Research Reagents

Table 2: Key Computational Tools and Inputs

Item / Software Code	Function / Purpose
Quantum ESPRESSO (PWSCF)	A full ab initio package for electronic structure, energy calculations, and linear response methods using plane waves and pseudopotentials [22].
VASP	A widely used DFT code implementing the projector augmented-wave (PAW) method and ultrasoft pseudopotentials [22].
KPointConvergence Workflow (aimstools)	An automated utility to set up, run, and evaluate k-point convergence studies [23].
ecutwfc	The key input parameter in PWSCF that sets the kinetic energy cutoff (in Rydberg) for the plane-wave basis set, controlling the quality of the wavefunction expansion [22].
K_POINTS automatic	The input parameter in PWSCF to define the Monkhorst-Pack k-point grid (e.g., 4 4 4 0 0 0 for a 4x4x4 grid) [22].
KSpace%Quality	A parameter in the BAND code that controls the density of the k-grid used for Brillouin Zone integration; crucial for converging the DOS [1].
Longipedlactone E	Longipedlactone E, MF:C30H38O6, MW:494.6 g/mol
4-Methylisoquinolin-8-amine	4-Methylisoquinolin-8-amine, MF:C10H10N2, MW:158.20 g/mol

Why is there a mismatch between my DOS and band structure results?

A mismatch between your Density of States (DOS) and band structure can occur because they are typically calculated using different methods and k-space sampling. The DOS is often computed via a method that interpolates over the entire Brillouin Zone (BZ), while the band structure is calculated along a high-symmetry path using a much denser k-point sampling [1]. If the k-grid for the DOS is not sufficiently converged, it might inaccurately determine the band edges (Valence Band Maximum and Conduction Band Minimum), leading to a different band gap value compared to the band structure plot [11].

How can I restart a calculation to improve the DOS?

You can recalculate the DOS with improved parameters, such as a denser k-grid, without restarting the full Self-Consistent Field (SCF) calculation. This is an efficient way to enhance the quality of your results [24].

Required Materials and Software

Item	Function
AMS/BAND Software	Software suite used for performing DFT calculations, including SCF, DOS, and band structure tasks.
Previous Calculation Results (.rkf file)	The output file from a prior SCF calculation; serves as the restart point for the new DOS calculation.
Input File Script	A text file specifying the new calculation parameters, such as a denser k-grid.
Linux/Unix Terminal	Environment for executing the AMS/BAND commands.

Step-by-Step Protocol

• Locate Your Previous Results: Ensure you have the .rkf files from a previously converged SCF calculation.
• Create a New Input File: Prepare an input file that loads the original atomic system and restarts from the previous SCF calculation. The key is to specify Dos true and BandStructure true within the Restart block, and to define a new, denser k-grid [24].
• Execute the Calculation: Run the new calculation using the command $AMSBIN/ams --delete-old-results < your_new_input_file.run [24].

The following diagram illustrates the workflow for restarting a DOS calculation:

What specific input settings should I use?

Below is a sample input script that demonstrates the restart procedure. The critical section is the Restart block, which instructs the code to use the potential from a previous calculation and only recalculate the DOS and band structure.

Example Input Script

Key Parameter Comparison

Parameter	Common Default	Improved Setting for Convergence	Explanation
KSpace%Quality	Normal	Good or VeryGood	Increases the number of k-points in the grid for better BZ sampling [1].
KSpace%Regular%DoubleCount	Not set	1	A simple way to double the density of the k-grid used in the previous calculation [24].
DOS%DeltaE	Default value	A smaller value (e.g., 0.005)	Makes the energy grid for the DOS finer, which can reveal sharper features [1].
BandStructure%EnergyBelowFermi	~10 Hartree	A larger value (e.g., 10000)	Ensures that deep-lying core bands are included in the band structure plot [1].

How do I verify that my new DOS is converged?

After running the new calculation, you should check the consistency of your results.

• Compare Band Gaps: Systematically compare the band gap value printed in the output file (from the k-space integration method) with the gap observed from the band structure plot [11]. They should be in much closer agreement.
• Visual Inspection: Plot the new DOS and the band structure on the same energy scale. The peaks in the DOS should align with the flat regions in the band structure.
• Quantitative Check: Use scripting tools, like the comparekf.py script mentioned in the documentation, to perform a dot-product comparison between the total DOS of your old and new calculations. A high similarity score indicates a well-converged DOS with respect to the k-grid [24].

Basis Set Selection and Confinement to Address Linear Dependency

FAQ: Understanding Linear Dependency

What is linear dependency in a basis set? In computational chemistry, a basis set is a set of functions used to represent molecular orbitals. Linear dependency occurs when one basis function can be represented as a linear combination of other functions in the set [25]. This makes the basis set over-complete and leads to numerical instability, causing slow or erratic Self-Consistent Field (SCF) convergence and inaccurate results [1] [26].

How is linear dependency detected? Programs detect linear dependency by analyzing the overlap matrix (a matrix of integrals representing the overlap between basis functions). The presence of very small eigenvalues in this matrix indicates near-linear-dependency [1] [26]. Most software will automatically project out these near-degeneracies if the eigenvalues fall below a default threshold, typically around 1×10⁻⁶ [26].

Why does my calculation have linear dependencies? Linear dependencies often arise from two main scenarios:

• Large, diffuse basis sets: Using very large basis sets, especially those with many diffuse functions (e.g., aug-cc-pV9Z), increases the chance of functions being numerically similar [27] [26]. This is common in calculations on anions or excited states.
• Systems with high coordination: In periodic systems like slabs or bulk materials, the basis functions of atoms in the inner layers can become overly diffuse, leading to dependencies with functions from neighboring atoms [1].

Troubleshooting Guide: Resolving Linear Dependency

Issue: Calculation fails with a "dependent basis" error message. This error indicates the program has identified a linear dependency it cannot safely ignore. Do not immediately relax the default tolerance; instead, adjust your basis set [1].

• Solution 1: Apply Confinement Confinement reduces the spatial extent of basis functions, which is especially useful for atoms in the interior of a system (like a slab) where diffuse functions are not needed.

  • Methodology: In the SCM BAND code, use the Confinement key to apply a radial potential that curtails the tail of the basis functions. The documentation suggests applying confinement to inner-layer atoms while using normal basis functions on surface atoms to properly describe electron density decay into vacuum [1].

• Solution 2: Manually Remove Problematic Functions If confinement is not sufficient or applicable, you can manually remove specific basis functions that cause dependencies.

  • Methodology: A practical approach is to identify and remove basis functions with exponents that are very similar in value. As demonstrated in a case study, the pairs of exponents with the smallest percentage difference (e.g., 94.8087090 and 92.4574853342) are often the primary culprits [27]. Removing one function from each of the N most similar pairs can cure N overly low eigenvalues in the overlap matrix [27].

• Solution 3: Use an Automated A Priori Method For a more robust and general solution, use a method based on the pivoted Cholesky decomposition of the overlap matrix.

  • Methodology: This algorithm processes the basis functions to identify and remove the minimal set required to eliminate linear dependencies before any expensive integrals are calculated [27]. Implementations are available in codes like ERKALE, Psi4, and PySCF [27].

• Solution 4: Adjust the Dependency Threshold (Use with Caution) As a last resort, you can tighten the threshold for identifying linear dependencies.

  • Methodology: In Q-Chem, this is controlled by the BASIS_LIN_DEP_THRESH $rem variable. A lower value (e.g., 5 for a threshold of 1×10⁻⁵) removes fewer functions but risks numerical instability. The default value of 6 (1×10⁻⁶) is generally reliable [26]. In ADF, the DEPENDENCY block and tolbas parameter can be used for similar control [28].

Issue: SCF convergence is slow or unstable, and I suspect linear dependency.

• Solution: First, try more conservative SCF convergence settings, such as decreasing the mixing parameter [1]. If problems persist, check your basis set for linear dependency using the methods above. Starting the SCF calculation with a smaller basis set (e.g., SZ) and then restarting with a larger one can also help [1].

Advanced Context: Connection to Band Structure and DOS Mismatch

Linear dependency in the basis set is a potential root cause of discrepancies between different electronic structure analyses, such as a mismatch between the band gap calculated from a band structure plot and that derived from the Density of States (DOS) [11].

The DOS is typically computed by sampling the entire Brillouin Zone (BZ) using an interpolation method, while a band structure plot is calculated along a specific high-symmetry path [1]. An inadequate or unstable basis set, potentially suffering from linear dependencies, can lead to an inaccurate representation of the electronic potential. This inaccuracy can cause inconsistencies between the two methods, as they probe the electronic structure differently. Therefore, ensuring a robust, non-linearly-dependent basis set is a critical step in troubleshooting band structure and DOS mismatches.

Research Reagent Solutions

The table below lists key computational tools and parameters relevant to managing basis set linear dependency.

Item	Function & Application	Key Parameters & Notes
Confinement Potential	Applies a radial potential to reduce the spatial extent of basis functions, curing dependencies in periodic systems [1].	Defined by a radius and potential shape. Particularly useful for inner atoms in slabs/bulk.
Pivoted Cholesky Decomposition	An automated algorithm to identify and remove linearly dependent functions from a basis set before integral calculation [27].	Available in ERKALE, Psi4, PySCF. More robust than manual removal.
Dependency Threshold (tolbas, BASIS_LIN_DEP_THRESH)	Numerical tolerance for identifying linear dependencies via the overlap matrix eigenvalues [28] [26].	Default: ~1e-6. Warning: Increasing tolerance (e.g., to 1e-5) can help SCF converge but may affect accuracy [26].
Overlap Matrix	The fundamental matrix used to diagnose linear dependence; its eigenvalues indicate the degree of dependency [1] [26].	Smallest eigenvalues are checked against the threshold. Cheap to compute.

Experimental Protocol: Basis Set Troubleshooting Workflow

The following diagram outlines a systematic workflow for diagnosing and resolving basis set linear dependency issues.

Workflow for Addressing Basis Set Linear Dependency

Partial DOS Projection for Orbital-Resolved Analysis

Frequently Asked Questions (FAQs)

Q1: What is the difference between total DOS and projected DOS (PDOS)? The total Density of States (DOS) describes the total number of available electronic states per energy level in a material, summed over all atoms and orbitals [29]. The projected DOS (PDOS) provides a more detailed breakdown, showing the contribution to the total DOS from specific atoms, specific orbitals (e.g., s, p, d), or specific spins [29] [3]. This allows you to understand which atomic species and orbitals are responsible for specific features in the electronic structure, such as the valence or conduction band edges [3].

Q2: Why are my PDOS results inconsistent with my total DOS? Inconsistencies between PDOS and total DOS can arise from several common setup errors:

• Insufficient k-point sampling: Using a sparse k-point mesh during the self-consistent field (SCF) calculation or the non-SCF DOS calculation can lead to poor quality and non-reproducible PDOS [30]. It is standard practice to use a denser k-point grid for DOS/PDOS calculations than for the initial SCF convergence [30].
• Incorrect orbital projection settings: Ensure that the parameters for projecting onto atoms or orbitals (e.g., natsph, iatsph in ABINIT, or the ProjectStates block in DFTB+) are correctly declared if you are targeting specific species [30].
• File handling error: When performing a multi-step calculation, you must correctly transfer the self-consistent charges from the ground-state calculation to the band structure/DOS calculation. For instance, in DFTB+, you need to copy the charges.bin file and set ReadInitialCharges = Yes [3].

Q3: How can I check the available orbital projections for my system? Most software packages provide ways to list all possible projection selections. For example, in VASP's py4vasp interface, you can use the selections() method to get a list of all available atoms and orbitals for projection [29]:

Q4: Can I perform spin-resolved PDOS analysis for magnetic materials? Yes, most modern DFT codes support spin-polarized calculations and can output spin-resolved PDOS. The selection syntax typically includes options for up, down, or total spin [29]. For instance, in ABINIT, you can use prtdos 3 in a spin calculation to obtain the projected DOS for each spin channel [30].

Q5: What does a "negative" Local Partial Density of States mean? Recent research has revealed that in mesoscopic systems, certain objects in the DOS hierarchy, like the local partial density of states, can become negative in the presence of a Fano resonance [31]. This negativity can be interpreted as a loss of coherent electrons in reverse time and may have implications for the thermodynamic properties of these systems. It has been demonstrated that this phenomenon is correlated with a Fano resonance featuring a π phase drop [31].

Troubleshooting Guides

Issue 1: Incorrect or No Orbital Projections in PDOS

Problem: The calculated PDOS is zero, does not appear, or does not match the expected contributions from specific atoms/orbitals.

Solution:

• Verify Projection Settings: Double-check the input file for the analysis block that enables PDOS.
  • In DFTB+, this is done in the Analysis block using ProjectStates and Region to specify the atoms and whether the projection should be shell-resolved [3].
  • In VASP, you must set LORBIT in the INCAR file to generate the projected DOS [29].
  • In QuantumATK, you use the ProjectedDensityOfStates block and choose the type of projection (e.g., on Elements and Shells) [32].
• Confirm Selection Syntax: When querying the results, ensure you use the correct syntax. For example, in py4vasp for VASP, a valid selection for the d-orbitals of Mn, Co, and Fe is: "d(Mn, Co, Fe)" [29].
• Check Basis Set: Ensure that the basis set used in the calculation includes the orbitals you are trying to project onto. For example, if you want to project onto d-orbitals, the basis set for the relevant atoms must have d-orbitals defined [3].

Issue 2: Discontinuities and Poor Resolution in DOS/PDOS Spectra

Problem: The DOS plot looks jagged, non-smooth, or has unexpected spikes, making it difficult to interpret.

Solution:

• Increase k-points for DOS step: The most common solution is to perform a non-SCF calculation on a much denser k-point mesh than the one used for the SCF convergence.
  • In ABINIT, use a higher ngkpt value in the dataset dedicated to the DOS calculation [30].
  • In DFTB+, for the band structure and DOS calculation, you define a custom Klines path or a dense mesh instead of reusing the SCF k-grid [3].
• Use Gaussian Smearing: Apply a small Gaussian broadening (smearing) to the discrete energy levels to create a smooth DOS. Tools like dp_dos in DFTB+ do this by default [3]. Adjust the smearing width carefully—a value that is too large will obscure important features, while one that is too small will not smooth the curve effectively.
• Ensure Proper Energy Range: When plotting, specify a sufficiently wide energy range around the Fermi level to capture all relevant features. For materials with a large bandgap like SiOâ‚‚, you may need to inspect a range of -15 eV to +10 eV [32].

Issue 3: Fermi Level Mismatch Between Total DOS and PDOS

Problem: The Fermi energy reported in the total DOS file is different from that in the PDOS file, leading to misaligned plots.

Solution:

• Single Reference Energy: Ensure that both the total DOS and PDOS are calculated in the same non-SCF run and that the Fermi energy is calculated only once from that same calculation and used as a common reference for all outputs.
• Consistent Post-Processing: When processing results, align all DOS data (total and partial) to the same Fermi energy value. Most codes output energies relative to the Fermi level by default. If you are using a script to plot, force it to use one consistent Fermi level for all data sets.
• Manual Alignment: If the above fails, you can manually shift the energy axes of your PDOS data to align with the total DOS, using the Fermi level as the zero point. This is often a last resort and indicates a potential issue with the calculation setup.

Experimental Protocols & Data Presentation

Protocol 1: Standard Workflow for PDOS Calculation in DFTB+

This protocol outlines the steps for obtaining PDOS using DFTB+ [3].

• Perform a Converged SCF Calculation:
  • Objective: Obtain self-consistent charges.
  • Method: Use a sufficiently dense k-point grid (e.g., a 4x4x4 Monkhorst-Pack set) and a tight SCC tolerance (e.g., SccTolerance = 1e-5).
  • Input Setup: In the Analysis block, define the projected DOS regions using ProjectStates. Specify the atoms (by element or index) and set ShellResolved = Yes to get orbital-level contributions.
• Calculate Band Structure and DOS on a Dense k-Path:
  • Objective: Get accurate band dispersion and DOS on a high-symmetry path.
  • Method: Copy the charges.bin from the previous step. In the new input file, set ReadInitialCharges = Yes and MaxSCCIterations = 1. Change the KPointsAndWeights to a Klines block that defines the path through high-symmetry points in the Brillouin zone.
• Post-Process and Plot:
  • Objective: Generate plottable DOS and PDOS files.
  • Method: Use the dp_dos tool. For the total DOS: dp_dos band.out dos_total.dat. For each PDOS file (e.g., dos_ti.1.dat), use the weighting option: dp_dos -w dos_ti.1.out dos_ti.s.dat.

Protocol 2: Accessing and Plotting PDOS in VASP via py4vasp

This protocol describes how to extract and visualize pre-calculated PDOS from a VASP calculation [29].

• Prerequisite Calculation: Run a VASP calculation with LORBIT set in the INCAR file to generate the required projection data.
• Access the DOS Object: In a Python script or Jupyter notebook, use py4vasp to access the calculation's Dos object.
• Select and Plot: Use the to_graph() or plot() method of the Dos object. Specify the desired orbital projections using the selection argument. The syntax allows for flexible combinations:
  • selection="1(p)" selects p-orbitals of the first atom.
  • selection="d(Mn, Co, Fe)" selects d-orbitals of Mn, Co, and Fe atoms.
  • selection="Ti(d) - O(p)" calculates the difference between the d-orbitals of Ti and the p-orbitals of O.

Quantitative Data from PDOS Analysis

The table below summarizes typical information that can be extracted from a PDOS analysis, using the example of anatase TiOâ‚‚ and silicon [3] [32].

Table 1: Key electronic properties derived from PDOS analysis for selected materials.

Material	Property	Value	Contributing Orbitals (from PDOS)
Anatase (TiOâ‚‚)	Valence Band Edge	Dominated by O p-orbitals	Oxygen p [3]
	Conduction Band Edge	Dominated by Ti d-orbitals	Titanium d [3]
Silicon (Si)	Indirect Band Gap	~1.1 eV (theoretical/experimental)	- [33]
	Conduction Band Minimum	Located at ~85% to X-point (0.425, 0, 0.425)	- [32]
SiOâ‚‚ (Quartz)	Indirect Band Gap (HSE06-DDH)	9.62 eV	- [32]
	Valence Bands	Dominated by O p shell	Oxygen p [32]
	Conduction Bands	Dominated by Si p shell	Silicon p [32]

The Scientist's Toolkit

Table 2: Essential software tools and functions for PDOS analysis.

Tool Name	Primary Function	Key Feature for PDOS
VASP	Planewave DFT code	PDOS via LORBIT; analyzed via py4vasp with flexible atomic/orbital selection syntax [29].
DFTB+	Density-functional tight-binding code	ProjectStates block for site- and shell-resolved PDOS; uses dp_dos for smearing and output [3].
QuantumATK	Multiscale platform	ProjectedDensityOfStates analyzer with projections on Elements, Shells, or Sites for local analysis [32].
ABINIT	Planewave DFT code	prtdos 3 outputs PDOS; variables natsph and iatsph for projections on specific atoms [30].
BAND (SCM)	DFT code for periodic systems	FatBands feature: the periodic equivalent of Mulliken population analysis, visualized as fatbands [6].
DDPC	Python data processing library	Aims to provide a unified interface for reading and manipulating DOS data from various DFT codes [34].
4-Methyl-1H-indol-3-amine	4-Methyl-1H-indol-3-amine	4-Methyl-1H-indol-3-amine (C9H10N2) is a research chemical for lab use. This RUO product is for professional research applications, not for human or veterinary use.
2-(2-Ethylhexyl)furan	2-(2-Ethylhexyl)furan|For Research Use	2-(2-Ethylhexyl)furan for research applications. This compound is For Research Use Only (RUO). Not for diagnostic, therapeutic, or personal use.

Workflow and Signaling Diagrams

PDOS Calculation and Troubleshooting Workflow

K-Point Sampling (Delta-K) Guidelines

Table 1: Recommended K-point sampling settings for different system types and calculation purposes.

System Type	Calculation Purpose	Recommended K-point Grid / Density	Key Considerations
Bulk Materials	Total Energy / DOS [14]	Tetrahedron method (ISMEAR = -5)	Requires at least 4 k-points to form a tetrahedron.
Metals	Geometry Relaxation [14]	Methfessel-Paxton (ISMEAR=1, SIGMA=0.2); ensure entropy term < 1 meV/atom.	Too sparse k-points can cause SCF convergence problems [1].
Semiconductors/Insulators	General Calculation [14]	Gaussian smearing (ISMEAR=0) with SIGMA=0.03-0.1.	Avoid ISMEAR > 0 (Methfessel-Paxton) as it can cause severe errors [14].
Unknown Systems / High-Throughput	General Screening [14]	Gaussian smearing (ISMEAR=0) with SIGMA=0.03-0.1.	A safe and generally applicable starting point.

Energy Grid & Smearing Parameters (Delta-E, SIGMA)

Table 2: Parameters controlling energy resolution and electronic smearing.

Parameter	Typical Value Range	Function & Effect	Convergence Check
DOS%DeltaE or NEDOS	Not specified in results	Controls the energy grid for DOS output; smaller values give higher resolution [1].	Decrease value until DOS profile does not change.
Gaussian Smearing (SIGMA)	0.03 - 0.1 eV (General) [14]	Smearing width for Gaussian (ISMEAR=0) or electronic temperature for Fermi-Dirac (ISMEAR=-1) [14].	Ensure free energy - total energy (T*S term) is negligible (~1 meV/atom for metals) [14].
SIGMA for Metals (Methfessel-Paxton)	~0.2 eV [14]	Smearing width for Methfessel-Paxton method (ISMEAR=1,2).	Monitor the T*S entropy term in the output file [14].

Troubleshooting Guides & FAQs

Frequently Asked Questions (FAQs)

Q1: Why does my calculated band structure not match my calculated Density of States (DOS)?

This is a common problem with several potential causes [1]:

• Different k-space sampling: The DOS is typically computed by sampling the entire Brillouin Zone (BZ), while the band structure is plotted along a specific high-symmetry path. The mismatch can occur if the k-point grid for DOS is not converged. Solution: Improve the KSpace%Quality parameter for the DOS calculation [1].
• Insufficient energy resolution: A coarse energy grid for the DOS can blur sharp features. Solution: Decrease the value of DOS%DeltaE to get a finer energy grid [1].
• Path selection: The chosen band structure path might miss key features (e.g., the actual valence band maximum or conduction band minimum) that are present elsewhere in the BZ [1].

Q2: How do I choose the correct smearing method (ISMEAR) and width (SIGMA) in VASP?

The choice is critical for accuracy and efficiency [14]:

• For semiconductors and insulators: Use Gaussian smearing (ISMEAR = 0) with a small SIGMA (0.03 to 0.1 eV) or the tetrahedron method (ISMEAR = -5) [14].
• For metals: Use the Methfessel-Paxton method (ISMEAR = 1) with a SIGMA that keeps the entropy term (T*S) below 1 meV per atom. A default of SIGMA = 0.2 is often a reasonable starting point [14].
• For unknown systems or high-throughput work: Always default to ISMEAR = 0 (Gaussian) with SIGMA = 0.1 [14].
• A crucial warning: Never use ISMEAR > 0 (Methfessel-Paxton) for semiconductors or insulators, as it can lead to incorrect results and large errors (e.g., >20% in phonon frequencies) [14].

Q3: What does the "dependent basis" error mean, and how can I resolve it?

This error indicates that the basis set used in the calculation is nearly linearly dependent, threatening numerical accuracy [1]. Do not simply loosen the convergence criterion. Instead:

• Use confinement: Apply spatial confinement to diffuse basis functions, which are often the cause, especially in highly coordinated atoms [1].
• Remove basis functions: Manually remove the most diffuse basis functions from your basis set [1].

Advanced Troubleshooting: Band Structure and DOS Mismatch

Problem: A researcher consistently observes a discrepancy between the bandgap measured from the band structure plot and the bandgap inferred from the DOS plot, within their thesis on band structure DOS mismatch.

Diagnosis Methodology:

• Verify the Fundamental Data Source: First, identify which method was used to determine the bandgap in each case. The gap printed in the main output file is typically from the "interpolation method" used for k-space integration and Fermi level determination. In contrast, the band structure plot uses a "band structure method" along a specific path [1].
• Isolate the Cause: The workflow below outlines the systematic diagnostic procedure.

Diagram 1: Diagnostic workflow for Band Structure-DOS mismatch.

Experimental Protocol for Resolution:

• Step 1 - Converge K-point Grid: Systematically increase the k-point density (e.g., from 6x6x6 to 12x12x12) and recalculate the DOS. The true bandgap from the DOS should converge to a stable value [1]. Use a SIGMA value appropriate for your system (see Table 2).
• Step 2 - Refine DOS Energy Grid: Set the energy grid parameter DOS%DeltaE (or NEDOS in VASP) to a smaller value (e.g., 0.01 eV or lower) to ensure sharp features are not artificially broadened by a coarse energy grid [1].
• Step 3 - Validate Band Path Coverage: Be aware that the band structure plot is only valid along the drawn path. If the top of the valence band or bottom of the conduction band lies at a k-point not on this path, the band structure plot will show an incorrect, smaller bandgap. The DOS, which samples the entire BZ, will show the correct, larger gap [1]. Cross-reference with literature or use specialized tools to find the true band extrema.

The Scientist's Toolkit: Research Reagent Solutions

Table 3: Essential computational parameters and their functions in band structure/DOS calculations.

Computational 'Reagent'	Function & Purpose	Technical Notes
K-point Grid	Samples the Brillouin Zone to compute integrals over k-space.	A denser grid is needed for metals and accurate DOS. Sparse grids can cause SCF convergence failure [1].
Smearing Function (ISMEAR)	Assigns fractional orbital occupations to improve convergence in metallic systems.	Critical choice: Gaussian for insulators/semiconductors; Methfessel-Paxton for metals [14].
Smearing Width (SIGMA)	Controls the width of the fractional occupation distribution.	Too large: incorrect energies. Too small: requires more k-points. Must be converged [14].
Tetrahedron Method (ISMEAR=-5)	A k-space integration method that interpolates bands between k-points.	Recommended for highly accurate DOS and total energy calculations in bulk materials [14].
Energy Grid (DeltaE)	Defines the resolution (bin width) for the DOS output.	A smaller DeltaE results in a higher-resolution DOS, revealing finer features [1].
Spatial Confinement	Reduces the range of diffuse basis functions.	A key remedy for "dependent basis" errors caused by highly coordinated atoms [1].

Diagram 2: Logical relationship between key tuning parameters and output spectra.

Diagnostic Framework for Band Structure-DOS Alignment

Systematic k-point Convergence Testing Protocol

FAQs: Addressing Common k-point Convergence Challenges

Why do I need to perform k-point convergence testing?

K-point convergence testing is essential because in practice, we must solve the Kohn-Sham equations for a finite number of k-wavevector values rather than across the entire continuous Brillouin zone [22]. Summing over a finite k-point grid approximates a continuous integral over the Brillouin zone, and an insufficient number of points leads to inaccurate total energies and derived properties [22]. For highly accurate calculations, such as thermodynamic studies for phase diagrams, extremely dense k-point sets are required to achieve total energy convergence better than 1 meV per atom [35].

What is the relationship between k-point sampling and my system's geometry?

The size and shape of your crystal's unit cell directly determine the size and shape of its Brillouin Zone (BZ) in reciprocal space [35]. A larger real-space unit cell results in a smaller reciprocal-space Brillouin zone. Consequently, systems with large unit cells (e.g., complex crystals or amorphous supercells) require fewer k-points for adequate sampling, sometimes only the Γ-point [35].

I'm getting inconsistent band gaps from my DOS and band structure calculations. Why?

This common discrepancy arises from the two different methods used to determine band edges [1] [36].

• Density of States (DOS): Typically uses an "interpolation method" that samples the entire Brillouin zone via a uniform k-point grid. The gap is the difference between the highest occupied and lowest unoccupied states found anywhere in this grid [1].
• Band Structure Plot: Calculated along a specific high-symmetry path in the Brillouin zone using a much denser linear sampling (DeltaK). This method can miss the true band edges if they do not lie on the chosen path [1].

To resolve this, ensure your DOS calculation uses a k-point grid of sufficient density, controlled by parameters like KSpace%Quality [1].

Troubleshooting Guides

SCF Convergence Failure with Dense k-point Grids

Problem: The Self-Consistent Field (SCF cycle fails to converge when using a high-density k-point grid.

Solutions:

• Conservative Mixing: Decrease the SCF mixing parameters to stabilize convergence [1]:

• Alternative SCF Methods: Switch from the default DIIS method to the MultiSecant method, which has a similar computational cost per cycle but might be more robust [1]:

• Two-Step Convergence: First, converge the SCF cycle using a smaller basis set (e.g., SZ). Then, restart the calculation with a larger basis set from this converged density [1].
• Automated Settings: Use engine automations to start with looser SCF convergence criteria and a higher electronic temperature, tightening them as the geometry optimization progresses [1].

Mismatch Between DOS and Band Structure

Problem: The calculated band gap differs between the Density of States (DOS) and the band structure plot.

Diagnosis and Resolution Workflow:

Specific Actions:

• Converge DOS k-grid: Systematically increase the k-point density for your DOS calculation (e.g., using KSpace%Quality). A recent study suggests that for meV-level accuracy, a k-point density as high as 5,000 k-points/Å⁻³ might be necessary [35].
• Inspect Band Structure Path: Ensure your band structure path passes through the likely locations of the valence band maximum and conduction band minimum. The band structure method assumes these critical points lie on the chosen path, which may not always be true [1].
• Refine DOS Energy Grid: Make the energy grid for the DOS calculation finer by decreasing the DOS%DeltaE parameter [1].

Essential Materials and Computational Reagents

Table: Key Components for k-point Convergence Studies

Item Name	Type	Function/Purpose	Example/Default Value
Monkhorst-Pack Grid	Algorithm	Generates a regular, balanced grid of k-points in the Brillouin Zone for efficient integration [35].	4 4 4 0 0 0 (for a cubic system)
KSpace%Quality	Parameter	Controls the fineness of the k-space mesh; higher values lead to denser sampling [1].	System-dependent; must be converged.
ecutwfc	Parameter	Plane-wave kinetic energy cutoff for the wavefunctions. Must be converged before k-points [22].	~30-100 Ry (system-dependent)
conv_thr	Parameter	The convergence threshold for the SCF cycle; tighter thresholds require better k-point sampling [22].	e.g., 1.0d-8
PWSCF (Quantum ESPRESSO)	Software Code	Performs DFT calculations using plane waves and pseudopotentials, requiring k-point input [22].	K_POINTS {automatic}
VASP	Software Code	A widely used DFT code employing the projector augmented-wave (PAW) method [22].	ISMEAR = 0; SIGMA = 0.XX for semiconductors
BiVO4 Pseudopotentials	Research Reagent	Specific pseudopotentials for elements; their choice influences the necessary ecutwfc and transferability [36].	Bi.pz-hgh.UPF, O.pz-hgh.UPF, V.pz-n-nc.UPF [36]

Step-by-Step Experimental Protocol

Systematic Protocol for k-point Convergence Testing

The following workflow outlines the complete procedure for establishing a converged k-point set, from initial setup to integration into a production calculation.

Step 1: Converge the Plane-Wave Cutoff

• Before starting k-point convergence, you must first converge the plane-wave energy cutoff (ecutwfc) [22]. The two are interdependent, but the standard practice is to fix a well-converged ecutwfc before testing k-points.

Step 2: Establish a Baseline Grid

• Start with a coarse k-point grid. For a cubic system, a reasonable starting point might be 4 4 4 [22]. The grid should be commensurate with your system's symmetry (e.g., 4 4 1 for a slab).

Step 3: Execute Convergence Calculations

• Run a series of total energy calculations (e.g., calculation = 'scf' in PWSCF), systematically increasing the number of k-points in each direction (e.g., 6 6 6, 8 8 8, 10 10 10, etc.) [10] [22].
• Keep all other parameters (cell geometry, atomic positions, ecutwfc) identical across all runs.

Step 4: Analyze Results and Determine Convergence

• For each calculation, extract the final total energy.
• Plot the total energy against the k-grid density (or the number of k-points).
• The energy will typically drop sharply initially and then plateau. The grid is considered converged when increasing the number of k-points changes the total energy by less than your desired precision (e.g., 1 meV/atom) [10] [35].

Step 5: Implement Converged Parameters

• Use the converged k-point grid for all subsequent property calculations (e.g., density of states, band structure, optical properties).
• For band structure calculations, perform a non-self-consistent field (NSCF) calculation on a densely spaced path of k-points using the previously converged charge density [36].

Addressing SCF Convergence Issues with Mixing and DIIS Parameters

How do SCF convergence problems relate to band structure and Density of States (DOS) mismatches?

In the context of band structure calculations, SCF convergence problems can directly lead to inconsistencies between the calculated band structure and the Density of States (DOS). A poorly converged SCF results in an inaccurate electron density and Fock matrix, which in turn affects the computed orbital energies (ε). Since the band structure is a k-space representation of these orbital energies along high-symmetry paths, and the DOS represents their density across the entire Brillouin zone, any error in the ε will cause a mismatch between the two. For instance, a calculation might show a band gap in the band structure but a continuous DOS at the Fermi level, indicating a failure to properly describe the system's electronic structure [37]. Furthermore, an insufficient k-point sampling grid for the DOS calculation, compared to the band structure, can also cause this discrepancy [38].

What are the fundamental parameters for controlling SCF convergence with DIIS and mixing?

The DIIS (Direct Inversion in the Iterative Subspace) algorithm and electron density mixing are primary tools for accelerating and stabilizing SCF convergence. Their behavior is controlled by several key parameters, summarized in the table below.

Parameter	Default (Typical) Value	Function	Effect of Increasing the Value
Mixing	0.1 - 0.3 [39] [40]	Fraction of the new Fock/Density matrix used in the linear combination.	More aggressive convergence, but less stable [39].
DIIS History (N)	5 - 10 [39]	Number of previous Fock/Density matrices used for extrapolation.	Increased stability, at the cost of higher memory usage [39].
DIIS Start Cycle (Cyc)	2 - 5 [39] [40]	SCF iteration at which DIIS begins.	More initial equilibration, leading to a more stable start [39].

What is a recommended DIIS and mixing parameter set for a difficult-to-converge system?

For systems with convergence challenges, such as those with small HOMO-LUMO gaps (e.g., transition metal complexes or metallic systems), a more stable and conservative parameter set is recommended. The following setup can serve as a starting point [39]:

• Mixing: 0.015 - 0.09
• DIIS History (N): 25
• DIIS Start Cycle (Cyc): 30

This configuration uses a smaller mixing parameter and a larger DIIS subspace, which slows down the convergence but significantly improves its stability by preventing large, erratic updates to the Fock or density matrix [39].

What is the step-by-step protocol for troubleshooting SCF convergence?

The following workflow provides a systematic approach to diagnosing and resolving persistent SCF convergence issues.

What are the functions of key research reagents and computational tools in SCF convergence?

In computational chemistry, "research reagents" are the algorithms and numerical techniques used to achieve a converged SCF solution. The table below details essential tools for this process.

Research Reagent (Algorithm)	Primary Function	Key Considerations
DIIS (Pulay Mixing)	Extrapolates a new Fock/Density matrix from a history of previous iterations to minimize the SCF error [41].	Increasing the history size (N) improves stability but uses more memory [39].
Damping	Mixes a large fraction of the previous density with the new one to prevent large oscillations in early cycles [40].	Often applied only for the first few iterations (NDamp) [42].
Level Shifting	Artificially increases the energy of virtual orbitals to widen the HOMO-LUMO gap, suppressing instability in the orbital optimization [39] [40].	Can affect properties dependent on virtual orbitals (e.g., excitation energies) [39].
Electron Smearing	Assigns fractional occupations to orbitals near the Fermi level, helping to converge metallic systems or those with near-degenerate states [39].	Introduces a finite electronic temperature; the smearing width should be as small as possible [39].
Quadratic Converger (QC)	Uses second-order methods (Newton-Raphson) to achieve robust convergence, often at a higher computational cost per iteration [42] [40].	A reliable last resort for difficult cases like open-shell transition metal complexes [42] [43].

What advanced algorithms exist if standard DIIS and mixing approaches fail?

When the parameter adjustments in the standard workflow fail, consider switching to more advanced SCF algorithms:

• Quadratically Convergent SCF (QC-SCF): This method uses a second-order algorithm to minimize the energy with respect to the orbital rotations. It is more robust and guaranteed to converge if started close enough to the minimum, but is computationally more expensive per iteration than DIIS [42] [40].
• Trust-Region Augmented Hessian (TRAH): Used in ORCA, this method is particularly effective for open-shell systems and ensures that the solution is a true local minimum on the orbital rotation surface [43].
• Anderson Acceleration: A generalization of DIIS that can be more effective for certain ill-conditioned nonlinear problems, including those found in computational chemistry [41] [44].

How can I ensure my final SCF solution is physically meaningful and stable?

Convergence of the SCF procedure does not guarantee that the solution is the true ground state. It is crucial to perform a stability analysis after convergence [40]. This analysis checks if the wavefunction is stable against small perturbations (e.g., breaking spin or spatial symmetry). If an instability is found, the calculation should be re-started from the unstable solution with the relevant constraints lifted (e.g., switching from RHF to UHF), allowing the SCF to converge to a lower-energy, stable solution [40].

Troubleshooting Guide: Frequently Asked Questions

Why is there a mismatch between my band structure and Density of States (DOS) calculations?

Discrepancies between band structure and DOS results are a common challenge in computational materials science, particularly for low-dimensional systems. Several factors can cause this:

• Different k-space sampling methods: The DOS is typically derived from a uniform sampling of the entire Brillouin Zone (BZ) using an interpolation method. In contrast, band structure calculations plot energies along specific high-symmetry paths in the BZ using a much denser k-point sampling along these lines [1]. If the chosen path misses the actual valence band maximum (VBM) or conduction band minimum (CBM), the band structure will report an incorrect band gap.

• Insufficient k-point convergence: A DOS calculation requires a well-converged k-point grid over the entire Brillouin zone. If the KSpace%Quality parameter is too low, the DOS will be inaccurate. Always test the convergence of your DOS with respect to this parameter [1].

• Coarse energy grid for DOS: The energy resolution of the DOS can be too low. Reducing the DOS%DeltaE parameter creates a finer energy grid, which can reveal sharper features and more accurate band edges [1].

• Fundamental methodological differences: The gap printed in output files is often from the BZ integration method, while the band structure provides a visual plot along a path. These can legitimately differ if the band extrema are not located on the high-symmetry path used for the band plot [1]. This is a common source of confusion, as highlighted by user reports on forums where band gaps from DOS and band structure differed by 0.27 eV for silicon [11] and 0.7 eV for BiVOâ‚„ [36].

How can I improve SCF convergence for metallic slabs and nanotubes?

Self-Consistent Field (SCF) convergence can be particularly difficult for low-dimensional systems like metal slabs or nanotubes due to their delocalized electrons and metallic character.

• Use more conservative mixing parameters: Reduce the mixing parameter to stabilize the convergence process [1].

• Employ alternative SCF algorithms: The DIIS method can be tuned, or more advanced methods like MultiSecant can be used, which comes at no extra cost per SCF cycle [1].

• Implement finite electronic temperature: For geometry optimizations, using a finite electronic temperature can help initial convergence. This can be automated to start with a higher temperature and gradually reduce it as the geometry converges [1].

• Start with a smaller basis set: If convergence is problematic, first run the calculation with a minimal basis set (e.g., SZ). Once converged, use the resulting density or wavefunctions as a starting point for a calculation with a larger basis set [1].

My geometry optimization for a slab/nanotube does not converge. What can I do?

If your SCF is converging but the geometry optimization is not, the issue likely lies with the forces or the optimization algorithm.

• Ensure accurate gradients: The forces (gradients of the energy) must be calculated accurately. Increase the numerical quality and the number of radial points in the basis functions to improve gradient accuracy [1].

• Verify the initial structure: For nanotubes, ensure your initial atomic coordinates and unit cell are physically reasonable. For slabs, confirm that the vacuum spacing is large enough to prevent interactions between periodic images.

• Check for soft vibrational modes: Low-dimensional systems often have soft modes. Negative frequencies in a phonon calculation can indicate that the geometry is not at a minimum or that the step size in the phonon run is too large [1].

What are the key differences in electronic structure between armchair and zigzag nanotubes?

The electronic properties of single-wall carbon nanotubes (SWCNTs) are critically dependent on their chirality, which is defined by the chiral vector (n, m).

The table below summarizes the fundamental distinctions:

Nanotube Type	Chirality	Electronic Behavior	Band Gap Origin
Armchair	(n, n)	Always metallic [45] [46]	Bands cross the Fermi level at ( k = \frac{2\pi}{\sqrt{3}a_0} ) [45].
Zigzag	(n, 0)	Metallic if n is divisible by 3; otherwise semiconducting [45] [46]	For semiconducting tubes, the gap is ( Eg \approx \frac{2\gamma a{\text{CC}}}{\d_{\text{tube}}} ) [45].
Chiral	(n, m), n≠m	Metallic if (n-m) is divisible by 3; otherwise semiconducting [46]	Behavior derived from the cutting lines in graphene's Brillouin zone [45].

This chirality dependence arises from the quantization of the wave vector in the circumferential direction and how this quantized set of lines intersects the graphene Brillouin zone, specifically the high-symmetry K-points where the valence and conduction bands meet [45].

Experimental Protocols & Workflows

Protocol 1: Standard Workflow for Band Structure and DOS Calculation

This protocol, adapted from DFTB+ and QuantumATK documentation [32] [3], ensures consistent and reliable results.

• Geometry Optimization

  • Fully relax the atomic coordinates and, if applicable, the lattice vectors of your nanotube or slab system using a converged k-point grid and accurate force thresholds.

• Self-Consistent Field (SCF) Calculation with Converged K-Points

  • Using the optimized geometry, perform an SCF calculation to obtain the ground-state electron density. Use a dense, uniform k-point grid (e.g., a Monkhorst-Pack grid) spanning the entire Brillouin zone.
  • Convergence Check: Systematically increase the k-point density until the total energy and band gap (for semiconductors) are stable (e.g., within 1-3 meV).

• Non-SCF Band Structure Calculation

  • Using the converged electron density from step 2, perform a non-SCF calculation where the k-points are defined along a high-symmetry path in the Brillouin zone (e.g., Γ-X-M-Γ). Keep the potential fixed (ReadInitialCharges = Yes in DFTB+ [3] or calculation='bands' in Quantum ESPRESSO [36]).

• DOS/PDOS Calculation

  • Using the same converged electron density from step 2, run a DOS calculation. This ensures the DOS and band structure are derived from the same electronic ground state. For Projected DOS (PDOS), specify the atoms and orbitals for projection.

The following workflow diagram illustrates this critical protocol and the relationship between different calculation types:

Protocol 2: Troubleshooting SCF Convergence in Metallic Slabs

This protocol addresses the specific challenge of SCF convergence in metallic low-dimensional systems, based on expert documentation [1].

• Initial Attempt with Defaults: Run with standard DIIS and mixing settings.
• Adjust Mixing Parameters: If the SCF oscillates, reduce the SCF%Mixing parameter to 0.05 or lower.
• Tweak DIIS Settings: Reduce DIIS%DiMix (e.g., to 0.1) and set Adaptable=false to disable automatic adjustments.
• Change SCF Algorithm: Switch from DIIS to the MultiSecant method (SCF%Method MultiSecant).
• Two-Stage Basis Set Strategy: If the above fails, converge the system using a minimal SZ basis set, then use the resulting density as a starting point for a calculation with the desired larger basis set.
• Finite Temperature & Automation: For geometry optimizations, implement an automation that starts with a higher electronic temperature (e.g., kT=0.01 Ha) and tighter SCF convergence criteria, which are automatically ramped down as the geometry converges.

The logical flow for diagnosing and resolving SCF convergence issues is outlined below:

The Scientist's Toolkit: Research Reagent Solutions

In computational materials science, "research reagents" refer to the fundamental numerical parameters and basis sets that define the accuracy of a simulation. The following table details essential "reagents" for simulating low-dimensional systems.

Tool Category	Specific Item / Parameter	Function & Application
Basis Sets	SZ (Single-Zeta), DZP (Double-Zeta plus Polarization)	Minimal SZ basis for pre-convergence; larger DZP for final production accuracy [1].
K-Point Grids	Monkhorst-Pack Grid, K-line Path	Uniform grid for DOS/charge density; high-symmetry path for band structure plotting [3].
SCF Mixing	Mixing parameter, DiMix parameter	Controls how the electron density is updated between cycles. Critical for stabilizing metallic systems [1].
Pseudopotentials	NC (Norm-Conserving), PAW (Projector Augmented-Wave)	Replace core electrons to reduce computational cost. Essential for heavy elements [36].
Numerical Grids	Becke Grid (Angular points), Radial Grid	For numerical integration. Heavy elements and slabs may require higher grid quality [1].
Exchange-Correlation	GGA (PBE), Hybrid (HSE06), LDA	Functional choice. Hybrids (HSE06) often give better band gaps [32].

A Quick Reference Guide

The table below summarizes the primary use cases and specific failure conditions for the Tetrahedron method and Gaussian smearing.

Method	Recommended For	Common Failure Conditions	Key Parameter(s)
Tetrahedron Method (ISMEAR = -5)	- Accurate total energy & DOS in bulk materials/semiconductors [14] [47]- Systems with sharp DOS features (e.g., Van Hove singularities) [48]	- Inaccurate forces/stress in metals (non-variational occupancies) [14]- Fails with coarse k-point meshes (less than 4 k-points per direction) [14] [47]- Can be fragile with specific k-grids or rounded cell parameters [49] [50]	KPOINT_MESH (Γ-centered)
Gaussian Smearing (ISMEAR = 0)	- Initial system screening & high-throughput calculations [14]- Semiconductors/insulators (safe choice) [14] [47]- Metallic relaxations (when using Fermi-Dirac) [14]	- Obscures sharp DOS features, appears converged to wrong DOS [48]- Incorrect total energy if SIGMA is too large [14]- Forces consistent with free energy, not SIGMA→0 energy [14]	SIGMA (smearing width)

The performance and accuracy of these methods are highly dependent on the k-point mesh density, as shown in the convergence data for a crystalline aluminum system below [51].

K-point Mesh	Total Energy (Hartree) with Smearing (MP)	Total Energy (Hartree) with Tetrahedron
4x4x4	-2.06181805	-2.07385026
8x8x8	-2.07161451	-2.07220663
12x12x12	-2.07212484	-2.07202169
16x16x16	-2.07195589	-2.07184510
24x24x24	-2.07182100	-2.07180814

---

Troubleshooting Guides

Troubleshooting Tetrahedron Method Failures

The tetrahedron method is prized for accuracy but has specific vulnerabilities.

• Problem: Inaccurate Forces During Metallic Relaxation

  • Cause: The tetrahedron method is not variational with respect to its partial occupancies. For metals, where states cross the Fermi level, this can lead to errors in the calculated forces and stress tensor, sometimes as high as 5-10% [14].
  • Solution: Do not use ISMEAR = -5 for ionic relaxations or molecular dynamics in metals. Switch to Methfessel-Paxton smearing (ISMEAR = 1) with a SIGMA value that keeps the entropy term (T*S in the OUTCAR file) below 1 meV/atom for the duration of the relaxation [14] [47].

• Problem: Calculation Crashes or Unphysical Results

  • Cause 1: Using a k-point mesh that is too coarse. The tetrahedron method requires at least 4 k-points in each direction to form a tetrahedron [14].
  • Solution 1: Use a denser k-point mesh.
  • Cause 2: A known issue where subtle rounding of cell parameters (e.g., to the 6th decimal place instead of full precision) can cause severe convergence problems in the tetrahedron method [50].
  • Solution 2: Ensure all atomic positions and lattice vectors are written to the input file with high precision (e.g., all 16 digits).

Troubleshooting Gaussian Smearing Failures

Gaussian smearing is a robust general-purpose method but can lead to silent errors.

• Problem: Incorrect or Underestimated Band Gap in DOS

  • Cause: A SIGMA value that is too large can artificially smear out the density of states, obscuring sharp features like band edges and Van Hove singularities. The DOS may appear visually converged with k-points but not be the correct DOS [48].
  • Solution: Systematically reduce SIGMA (e.g., to 0.03-0.1 eV) and reconverge your k-point mesh. For final, high-accuracy DOS calculations, the tetrahedron method (ISMEAR = -5) is strongly recommended [14] [48].

• Problem: Total Energy is Not Converged or Too High

  • Cause: The total energy computed with Gaussian smearing is sensitive to the SIGMA value. Large smearing introduces errors in the energy [14] [51].
  • Solution: Use the energy(SIGMA→0) value reported in the OUTCAR file, but note that this requires a systematic reduction of SIGMA to be accurate. Ensure the entropy term T*S is minimal. For metals, Methfessel-Paxton smearing (ISMEAR = 1) is often less sensitive to the smearing width and is easier to use for energy convergence [14] [47].

Diagram 1: A workflow for selecting between the tetrahedron and smearing methods, based on system type and accuracy requirements [14] [47].

---

Experimental Protocols

Protocol for DOS Convergence Study

This protocol is designed to diagnose and resolve discrepancies between band structure and DOS plots, a core issue in band structure DOS mismatch troubleshooting research.

• Initial SCF Calculation: Perform a self-consistent field (SCF) calculation on a fully relaxed structure using a relatively dense, uniform k-point grid (e.g., determined by a KSPACING of 0.04 or less in VASP) to obtain a ground-state charge density [2].
• Tetrahedron Method DOS: Run a non-self-consistent field (NSCF) calculation using the tetrahedron method with Blöchl corrections (ISMEAR = -5). Use a Γ-centered k-point mesh and the same dense k-grid from the SCF calculation [14] [52].
• Gaussian Smearing DOS: Using the same charge density and k-point mesh, run a second NSCF calculation with Gaussian smearing (ISMEAR = 0) and a small SIGMA value (e.g., 0.05 eV).
• Comparison and Analysis: Overlay the resulting DOS plots. The tetrahedron method should reveal sharper features. A significant discrepancy, especially at band edges, indicates the Gaussian SIGMA is too large or the k-mesh is insufficient [48].
• Validation with Band Structure: Compute the band gap from the DOS and compare it to the gap obtained from a band structure calculation along a high-symmetry path. For insulators, the band structure method is often more accurate for locating the CBM and VBM [1] [2].

Protocol for Force Accuracy in Metallic Systems

This protocol ensures accurate ionic forces during the relaxation of metallic systems, where the tetrahedron method fails.

• Preliminary Energy Calculation: Perform a static calculation using the tetrahedron method (ISMEAR = -5) on your initial metallic structure to establish a accurate baseline total energy [14].
• Switch to Smearing for Relaxation: In your relaxation (IBRION > 0) or molecular dynamics calculation, switch to Methfessel-Paxton smearing (ISMEAR = 1) [14].
• Converge SIGMA: Test a range of SIGMA values (e.g., 0.1 to 0.3 eV). For each, run a single-point calculation and check the OUTCAR file for the entropy term T*S.
• Select Optimal SIGMA: Choose the largest SIGMA value for which the entropy term T*S is negligible (typically < 1 meV per atom). This ensures computational efficiency without sacrificing accuracy [14] [47].
• Final Verification: For the final relaxed structure, a single static calculation with the tetrahedron method can be performed to obtain the most accurate total energy and DOS [14].

Diagram 2: A protocol for comparing DOS outputs from two methods to identify convergence issues [14] [48].

---

Frequently Asked Questions (FAQs)

Q1: Why does my semiconductor's total energy change drastically when I use ISMEAR=1 (Methfessel-Paxton)?

A1: Methfessel-Paxton smearing is designed for metals and can lead to unphysical partial occupancies in gapped systems. This introduces a large, incorrect contribution to the energy (entropy term). Always use ISMEAR = 0 (Gaussian) or ISMEAR = -5 (tetrahedron) for semiconductors and insulators [14] [47].

Q2: I see a band gap in my band structure plot, but my DOS calculation shows a zero gap. Why?

A2: This is a classic DOS mismatch problem. The most common cause is an insufficient k-point mesh in the DOS calculation. The band structure uses a dense path but the DOS uses interpolation over the entire Brillouin zone. A coarse k-grid can miss the exact location of the valence band maximum (VBM) and conduction band minimum (CBM). Solution: Converge your DOS calculation with a denser k-point mesh. Also, ensure you are using the tetrahedron method for the DOS to avoid smearing artifacts that can artificially close the gap [1] [48].

Q3: For a metal, should I always avoid the tetrahedron method?

A3: No. The tetrahedron method is excellent for calculating highly accurate total energies and the density of states (DOS) for metals in a single, static calculation. You should only avoid it when you need to compute accurate ionic forces or stress, such as during a geometry relaxation or molecular dynamics simulation [14].

Q4: What is the single most important setting to check for high-throughput calculations?

A4: Set ISMEAR = 0 and SIGMA = 0.05 (or similar small value). This Gaussian smearing setup is safe for insulators, semiconductors, and metals. It prevents catastrophic failures that can occur from using ISMEAR > 0 on gapped systems while providing reasonable results for metals, making it the most robust choice for automated workflows where the electronic nature of the material is not known in advance [14].

---

The Scientist's Toolkit: Key Computational Parameters

Item	Function/Description	Typical Value / Example
ISMEAR	Determines the method for setting partial occupancies [52].	-5 (Tetrahedron), 0 (Gaussian), 1 (Methfessel-Paxton)
SIGMA	Smearing width (eV) for broadening the electron occupancy [14].	0.05 (Semiconductors), 0.2 (Metals with MP)
KPOINTS	Defines the mesh of k-points for Brillouin Zone sampling.	8 8 8 0 0 0 (Monkhorst-Pack)
EFERMI	Controls the algorithm for determining the Fermi energy.	MIDGAP (Places Fermi level in middle of gap for insulators) [14]
IBRION	Type of ion relaxation algorithm.	2 (Conjugate Gradient)

Basis Set Dependency and Numerical Accuracy Optimization

Troubleshooting Guides

SCF Convergence Failure

Problem: The Self-Consistent Field (SCF procedure fails to converge, particularly challenging for systems like Fe slabs compared to Pd slabs [53] [1]

Solution: Implement more conservative SCF settings and improve numerical precision:

Alternative SCF methods can be employed [1]:

or

For precision-related convergence issues (indicated by many iterations after HALFWAY message) [53] [1]:

Experimental Protocol:

• Begin with a minimal SZ basis set for initial convergence
• Restart SCF with larger basis set from converged result
• For heavy elements, avoid small or no frozen core
• Implement finite electronic temperature during geometry optimization with automation scripts [1]

Basis Set Dependency Error

Problem: Calculation aborts with "dependent basis" message indicating linear dependency in Bloch functions [53] [1]

Solution: The program identifies problematic basis functions through dependency coefficients. Two primary resolution approaches:

Confinement Method (ideal for slabs) [53]:

Basis Function Removal:

• Examine dependency coefficients in output
• Remove basis functions with largest coefficients
• For two large coefficients, replace corresponding functions with an averaged function
• Prefer removing STOs over numerical Dirac orbitals [53]

Diagnostic Protocol:

• Check overlap matrix eigenvalues for each k-point
• Identify functions with largest dependency coefficients
• Remove one function per small eigenvalue
• Iterate until all k-points pass dependency check [53]

Band Structure and DOS Mismatch

Problem: Discrepancy between band structure plots and Density of States (DOS) calculations [1] [11]

Solution: Address the fundamental methodological differences:

k-Space Sampling:

Experimental Protocol:

• Converge DOS with improved KSpace%Quality
• Ensure band structure path covers high-symmetry points
• Verify consistent energy references between calculations
• Cross-validate with multiple k-point meshes [1]

Frequently Asked Questions

How do I resolve negative frequencies in phonon calculations?

Answer: Negative frequencies typically indicate either non-minimum geometry or excessive phonon step size [53] [1]:

What causes frozen core errors and how are they resolved?

Answer: Frozen core overlap criterion violations occur when frozen core approximation inadequately represents core functions [53]:

Safe Approach:

Performance-Optimized Approach:

Always validate with smaller test systems comparing to smaller core calculations [53]

Why does lattice optimization fail for GGA calculations?

Answer: Use analytical stress instead of numerical stress [1]:

Research Reagent Solutions

Table: Essential Computational Parameters for Numerical Accuracy

Component	Function	Recommended Settings
SCF Mixing	Controls charge density mixing between iterations	0.05 for problematic systems [53]
DIIS Dimension	Size of DIIS subspace for convergence acceleration	Reduced Dimix (0.1) with Adaptable false [53]
k-Space Quality	Brillouin zone sampling density	Normal or Good (avoid Basic) [53] [1]
Density Fit Quality	Accuracy of density fit approximations	Normal or Good ZlmFit [53]
Becke Grid Quality	Numerical integration grid for exchange-correlation	Normal or Good for heavy elements [53]
Frozen Core Criterion	Threshold for frozen core approximation	Default 0.98 or relaxed 0.8 with validation [53]

Experimental Workflows

Troubleshooting Logic Diagram

Basis Set Optimization Workflow

Cross-Verification and Benchmarking Strategies

A technical support guide for researchers encountering discrepancies in band gap analysis.

Why Do My Band Gap Values Disagree?

A mismatch between band gap values calculated by different methods is a common issue in computational materials science. This discrepancy often arises because Density of States (DOS) and band structure calculations use different k-point sampling methods [1] [2]. The DOS is typically computed on a uniform k-point grid covering the entire Brillouin zone, while the band structure is calculated along a high-symmetry path between specific points. It is possible for the uniform grid to miss the precise k-points where the valence band maximum (VBM) or conduction band minimum (CBM) occurs, leading to different gap values [2]. For example, a silicon calculation showed a 0.27 eV difference (0.61 eV from band structure vs. 0.88 eV from DOS) due to this sampling difference [11].

A Quick Diagnostic Table

Observation	Possible Cause	Next Steps
Small difference (e.g., < 0.3 eV) between DOS and band structure gap.	Different k-point sampling missing the exact VBM/CBM [2] [11].	Recompute the gap from the DOS, which is often more robust for the fundamental gap [2].
A previously non-zero gap now reads as 0 eV.	Database/parsing update, Fermi level placement issue, or the material is truly metallic/semi-metallic [2].	Use the Materials Project API to recompute the gap from the DOS data [2].
Severe SCF convergence issues, leading to unreliable results.	Bad initial precision, insufficient k-points, or problematic mixing parameters [1].	Use a smaller basis set (e.g., SZ) for an initial calculation, then restart with a larger one [1].
GGA/LDA functional calculates a band gap much smaller than experiment.	Known limitation of standard DFT functionals, which underestimate band gaps by ~40-50% on average [2] [54].	Use a more advanced method like Gâ‚€Wâ‚€ or hybrid functionals, applying a "scissor" correction if needed [2].
Band structure plot does not match DOS peaks.	The chosen high-symmetry path for the band structure may miss key features present in the full Brillouin zone sampled by the DOS [1].	Improve k-space quality for DOS convergence and ensure the energy grid (DOS%DeltaE) is fine enough [1].

Troubleshooting and Resolution Guide

Follow this workflow to diagnose and resolve band gap mismatches in your calculations.

Step 1: Ensure SCF Convergence

A foundational requirement for any reliable electronic structure calculation is that the Self-Consistent Field (SCF) procedure is fully converged. If the SCF is not converged, all subsequent analysis (including band structure and DOS) will be unreliable [1].

Troubleshooting SCF Convergence:

• Use Conservative Settings: Decrease the mixing parameters to stabilize the convergence process [1].

• Change the SCF Algorithm: Try the MultiSecant method, which comes at no extra cost, or the LISTi method, which may reduce the number of SCF cycles [1].
• Start Simple: For problematic systems, first run the calculation with a minimal basis set (e.g., SZ). Once converged, use the results as a restart point for a calculation with your target larger basis set [1].

Step 2: Recompute the Band Gap from the DOS

The most robust way to verify a band gap is to recalculate it directly from the Density of States data. This avoids potential artifacts from the band structure interpolation or Fermi level placement.

Protocol using the Materials Project API and pymatgen:

This method directly queries the DOS, which samples the entire Brillouin zone, providing a more reliable value for the fundamental band gap than the band structure might in cases of parsing errors [2].

Step 3: Verify Against Band Structure with Corrected Fermi Level

If you need to use the band structure, ensure the Fermi level is correctly aligned using the VBM from the DOS.

Protocol for Band Structure Correction:

This approach helps identify if a reported 0 eV gap is a physical property (the material is a metal or semimetal) or a computational artifact [2].

Key Reagents & Computational Solutions

The following tools and parameters are essential for robust band gap calculations.

Item / Software	Function / Purpose	Key Consideration
SCF Convergence Parameters	Stabilize the self-consistent field calculation.	Use more conservative Mixing and DiMix for difficult systems [1].
K-point Grid Quality	Controls sampling of the Brillouin zone for DOS/charge density.	A denser grid is needed for accurate DOS and to find the true CBM/VBM [1] [2].
Band Structure DeltaK	Controls the interpolation step along the high-symmetry path.	A smaller DeltaK (e.g., 0.03) yields smoother bands but increases cost [6].
Materials Project API	Accesses computed data to validate and benchmark your results.	Use to recompute band gaps from DOS data, bypassing potential parsing errors [2].
Gâ‚€Wâ‚€ Method	A more advanced, post-DFT method for accurate quasiparticle band gaps.	Reduces DFT's inherent band gap underestimation; different codes can show 0.1-0.3 eV variations [54].
Tauc Plot Analysis	Standard experimental method to determine the optical band gap from absorption data [55].	Fitting requires the absorption coefficient α; the Tauc relation is (αhν)¹/² ∝ (hν - E₉) for direct gaps [55].

Experimental Protocol: Band Gap via Absorption Spectrum Fitting (ASF)

This protocol allows for determining the optical band gap of thin films using only absorbance data, without needing film thickness or reflectance spectra [55].

1. Sample Preparation (CdSe Nanostructured Film Example)

• Substrate Cleaning: Clean glass slides with detergent, rinse with acetone, and perform ultrasonic cleaning followed by a final rinse with a mixture of double-distilled water and methanol [55].
• Chemical Bath Deposition (CBD): Prepare a solution from 0.25 M cadmium acetate and 0.25 M sodium selenosulfate (Naâ‚‚SeSO₃). Use ammonia to complex Cd²⁺ ions and adjust the pH. Immerse substrates vertically in the bath and maintain a constant temperature (e.g., room temperature) for a set duration (e.g., 4-24 hours) [55].
• Post-processing: Remove substrates, wash with deionized water and methanol to remove loosely adhered particles, and dry in air [55].

2. Data Collection

• Use a UV-Vis spectrometer to measure the absorbance spectrum Abs(λ) of the film across the relevant wavelength range [55].

3. Data Analysis and Fitting

• The optical band gap E_gap is determined by fitting the absorbance data to the Tauc model. For a direct band gap semiconductor, the relation is [55]: Abs(λ) ∝ (1/λ) * [ (1/λ) - (1/λ_g) ]^{1/2} where λ_g is the wavelength corresponding to the band gap.
• Fitting Procedure:
  • Plot [Abs(λ)]² as a function of photon energy hν (or 1/λ).
  • Identify the linear region of the plot.
  • Extrapolate the linear region to the x-axis ([Abs(λ)]² = 0). The intercept gives the direct optical band gap energy E_gap [55]. E_gap (eV) = 1239.83 / λ_g

Key Concepts: Direct vs. Indirect Band Gaps

Understanding the nature of the band gap is crucial for interpreting results and applications.

• Direct Band Gap: The maximum of the valence band and the minimum of the conduction band occur at the same k-vector in the Brillouin zone. Optical transitions can occur with just a photon, making these materials efficient for light emission and absorption (e.g., GaAs, InAs) [56] [57].
• Indirect Band Gap: The valence band maximum and conduction band minimum occur at different k-vectors. Optical transitions require both a photon and a phonon (lattice vibration) to conserve momentum, making these processes less probable. This results in weaker light emission/absorption (e.g., Si, Ge) [56] [57].

Interpolation vs Band Structure Method Gap Determination

Frequently Asked Questions (FAQs)

• Why does the band gap value from my density of states (DOS) calculation differ from the value in my band structure plot? This is a common discrepancy. The two methods use fundamentally different approaches to sample the Brillouin Zone (BZ). The DOS is typically derived from an interpolation method that samples the entire BZ using a uniform k-point grid, but with a coarser spacing. In contrast, the band structure is calculated using a band structure method that samples a specific high-symmetry path with a much denser k-point spacing (DeltaK). The gap can differ if the band structure path does not contain the actual k-points where the valence band maximum (VBM) or conduction band minimum (CBM) occur [1].

• Which band gap value should I trust for my calculations? The "band structure" method is often more reliable for determining the precise fundamental gap if you are confident that your chosen high-symmetry path contains both the VBM and CBM. This method allows for a very dense sampling along the path, which can accurately resolve the band edges [1]. However, the interpolation method is more rigorous for ensuring the VBM and CBM are found somewhere in the entire BZ, not just on a predefined path. For conclusive results, you should verify that the k-points identified as the VBM and CBM in the DOS are indeed on your band structure path.

• How can I improve the agreement between my DOS and band structure? The key is to improve the convergence of the DOS with respect to k-point sampling. You should try increasing the KSpace%Quality parameter to use a finer, more accurate k-point grid for the DOS calculation [1]. Additionally, you can make the energy grid for the DOS finer by decreasing the DOS%DeltaE parameter [1].

• My material is listed with a 0 eV band gap in a database, but I expect it to be a semiconductor. What should I do? A reported 0 eV gap can stem from either a physical reality (e.g., the material is a semimetal) or a computational artifact. The most robust action is to recompute the band gap directly from the DOS data, as automated detection of band edges can sometimes fail, especially with complex DOS profiles near the Fermi level [2].

---

Troubleshooting Guides

Problem 1: Mismatch Between DOS and Band Structure Gaps

This guide addresses the scenario where the band gap computed from the Density of States (DOS) does not match the value from the band structure plot.

Diagnosis Questions:

• Is the k-point grid used for the DOS calculation too coarse?
• Does the high-symmetry path used for the band structure miss the true k-point locations of the VBM or CBM?

Resolution Protocol:

• Verify Methodology: Confirm that you understand the two distinct methods for gap determination, as detailed in the FAQ section [1].
• Converge DOS K-Points: Systematically increase the k-point density for your DOS calculation and observe if the gap value converges. This is the most critical step.
• Inspect Band Edges: Examine the DOS data to identify the specific k-points and eigenvalues for the VBM and CBM. Cross-reference these k-points with the path used in your band structure plot to ensure they are included [1] [2].
• Refine DOS Energy Grid: If the DOS peaks appear broad, decrease the energy grid spacing (DOS%DeltaE) for a sharper, more accurate representation [1].

Workflow for Diagnosis and Resolution: The following diagram illustrates the logical process for diagnosing and resolving a DOS-band structure gap mismatch.

Problem 2: Unexpected Zero Band Gap

This guide helps when your calculation or a database reports a metallic (0 eV gap) result for a material expected to be semiconducting.

Diagnosis Questions:

• Is the reported 0 eV gap a physical property or a parsing/calculation artifact?
• Was the calculation converged, and did it use an appropriate functional?
• Is the Fermi level placement correct?

Resolution Protocol:

• Recompute from DOS: Manually recalculate the gap from the DOS data. Automated parsing can sometimes fail, especially if the DOS at the Fermi level is very small but non-zero [2].
• Check Task Hierarchy: In database analysis, verify which calculation type (DOS, band structure, etc.) was used to determine the gap. The hierarchy can affect the reported value [2].
• Verify Fermi Level: Ensure the Fermi level is correctly identified and placed between the VBM and CBM. In some cases, you may need to reconstruct the band structure object using the VBM from the DOS for an accurate gap [2].
• Consider Functional Limitations: Standard GGA functionals are known to severely underestimate band gaps and can incorrectly predict metals for known small-gap semiconductors. If steps 1-3 do not resolve the issue, consider repeating the calculation with a more advanced method like a hybrid functional (e.g., HSE) [58] [59].

Workflow for Diagnosing a Zero Gap:

---

Quantitative Data Comparison

The table below summarizes the key characteristics of the two primary band gap determination methods.

Feature	Interpolation Method (e.g., for DOS)	Band Structure Method
Primary Use Case	K-space integration for Fermi level & occupations; calculating total DOS [1]	Visualizing band dispersion; post-SCF analysis along a specific path [1]
BZ Sampling	Whole Brillouin zone with a uniform k-point grid [1]	Dense sampling along a chosen high-symmetry path [1]
Key Advantage	Systematically searches the entire BZ for the true VBM and CBM [1]	Can use very dense k-point spacing (DeltaK) along the path for high resolution [1]
Key Limitation	Coarser k-spacing may miss sharp band edges [1]	Relies on the path containing the actual VBM and CBM locations [1]
Reported Gap	The gap printed in output files is typically from this method [1]	The gap must be read directly from the band structure plot or data

---

The Scientist's Toolkit: Essential Computational Reagents

Item / "Reagent"	Function / Explanation
Hybrid Functional (HSE)	A more advanced exchange-correlation functional that mixes a portion of exact Hartree-Fock exchange with DFT, significantly improving band gap accuracy compared to standard GGA [58] [59].
K-Point Grid	A set of points in the reciprocal space used to numerically integrate over the Brillouin Zone. A denser grid is required for accurate DOS and gap convergence [1] [2].
High-Symmetry Path	A predefined trajectory through the Brillouin Zone connecting points of high symmetry. It is used for band structure plots to visualize dispersion relations [2].
Density of States (DOS)	A function that gives the number of electronic states per unit volume per unit energy. The band gap is identified as the energy range between the VBM and CBM where the DOS is zero [1] [2].
Post-Processing Code (e.g., pymatgen)	A software library used to programmatically analyze computational outputs, such as recomputing band gaps from DOS data to validate results and troubleshoot discrepancies [2].

Validation Through PDOS and Band Structure Correspondence

Troubleshooting Guides

FAQ: Why is there a mismatch between my band structure and my Density of States (DOS)?

Q: I have calculated both the band structure and the DOS for my system, but certain features, like a band gap or a prominent peak, appear in one but not the other. Why does this happen?

A: This is a known challenge in electronic structure calculations. The discrepancy often arises because the two properties are calculated using different sampling methods in the Brillouin Zone (BZ) [1].

• DOS Calculation: The DOS is typically derived from a method that interpolates eigenvalues over a uniform grid of k-points spanning the entire Brillouin Zone. This provides a good average picture of the available states at each energy level [17].
• Band Structure Calculation: The band structure is calculated by tracing the eigenvalues along a specific high-symmetry path in the Brillouin Zone. It shows the energy levels only for the k-points on this line [17].

A mismatch can occur if the chosen high-symmetry path for the band structure does not pass through the k-points where the valence band maximum (VBM) or conduction band minimum (CBM) are located. The DOS, sampling the entire zone, will correctly reflect the true band gap, while the band structure plot might not show it if the path misses these critical points [1].

Q: How can I resolve this discrepancy and ensure my results are valid?

A: Follow this structured troubleshooting protocol to diagnose and resolve the issue.

1. Verify k-Space Convergence for the DOS: The quality of the DOS is highly dependent on the density of the k-point grid used in the self-consistent field (SCF) calculation [17]. An insufficiently dense grid can lead to a poorly converged DOS that misses key features.

• Action: Systematically increase the density of your k-point grid (for example, from 4x4x4 to 8x8x8) and rerun the SCF and DOS calculations [17]. Compare the resulting DOS. A converged DOS should not change significantly with a further increase in k-points.

2. Refine the Band Structure Path: The standard high-symmetry path might not be sufficient for your specific material.

• Action: Consult the literature or crystallographic databases to identify all potential high-symmetry points in the Brillouin Zone where the VBM and CBM might occur. Consider calculating the band structure along multiple different paths to ensure you are not missing critical features.

3. Check Calculation Parameters for Consistency: Ensure that all underlying settings between the two calculations are consistent, except for the k-point sampling method.

• Action: Confirm that you are using the same:
  • Lattice parameters and atomic positions.
  • Exchange-correlation functional.
  • Basis set or plane-wave cutoff energy.
  • SCC (Self-Consistent Charge) tolerance for convergence [17].

The workflow below outlines the logical steps for resolving a band structure and DOS mismatch.

Experimental Protocols for Validation

Protocol 1: Achieving a Converged Density of States

A well-converged DOS is the foundation for valid comparison. This protocol outlines the steps for the DFTB+ code [17].

• Geometry Optimization: Fully optimize the crystal structure (lattice parameters and atomic positions) before any electronic property calculation.
• Initial SCF Calculation: Perform a self-consistent calculation with a moderately dense k-point grid to obtain converged charges. In DFTB+, this involves:
  • Setting Scc = Yes and a strict SccTolerance (e.g., 1e-5) [17].
  • Using a k-point grid generated via SupercellFolding or an equivalent Monkhorst-Pack scheme.
• DOS Calculation: Using the converged charges from the previous step, calculate the DOS. In DFTB+, the dp_dos tool is used to process the band.out file and generate a plottable DOS file (e.g., dos_total.dat) [17].
• Convergence Test: Repeat steps 2 and 3 with progressively denser k-point grids (e.g., 4x4x4, 6x6x6, 8x8x8). The DOS and the total energy are considered converged when they do not change significantly between iterations.

Protocol 2: Calculating a Corresponding Band Structure

Once a converged DOS is obtained, the band structure can be calculated.

• Read Converged Charges: Use the charges.bin file from the converged SCF calculation as the starting point. In the input file, set ReadInitialCharges = Yes and MaxSCCIterations = 1 to use the fixed, converged charges [17].
• Define High-Symmetry Path: Replace the uniform k-point grid with a path along high-symmetry points in the Brillouin Zone. In DFTB+, this is done using the Klines method, specifying the points and the number of k-points between them [17]. Example:

• Run Band Structure Calculation: Execute the calculation. The output will contain the eigenvalues for all bands along the specified path.

Key Parameters for DOS and Band Structure Calculations

The following table summarizes critical parameters that influence the results of DOS and band structure calculations, based on documentation from DFTB+ and BAND [17] [4] [1].

Parameter	Description	Role in Validation
k-Grid Quality	Density of the k-point mesh for SCF/DOS.	A coarse grid is a primary cause of DOS inaccuracies. Must be converged. [17] [1]
Band Structure Path (Klines)	The sequence of high-symmetry k-points for the band plot.	An ill-chosen path might miss the true VBM/CBM, causing a mismatch with the DOS. [17]
DOS Energy Grid (DeltaE)	The energy resolution (step size) for the DOS output.	A too-large DeltaE can smear out sharp peaks, obscuring features visible in the band structure. [4] [1]
SCC Tolerance (SccTolerance)	The convergence criterion for self-consistent charges.	Ensures the electronic charge density is stable before properties are calculated. [17]

The Scientist's Toolkit: Essential Computational Reagents

This table details key "reagents" or components used in performing and validating PDOS and band structure calculations.

Item	Function in the Calculation
Slater-Koster Files	Parameterized files containing pre-computed integrals for atomic interactions. They are essential for semi-empirical methods like DFTB to function [17].
k-Point Grid	A set of points in the Brillouin Zone used to numerically integrate periodic functions. It is crucial for achieving convergence in SCF calculations and the DOS [17].
High-Symmetry Path	A predefined route through the Brillouin Zone connecting points of high symmetry. It allows for the intuitive visualization of electronic bands in the band structure plot [17].
Projected DOS (PDOS)	A decomposition of the total DOS into contributions from specific atoms, orbitals, or groups. It is vital for understanding the chemical nature of electronic states [17] [4].
Hubbard U Correction	An empirical parameter added to DFT to better describe strongly correlated electrons (e.g., in d or f orbitals), which can significantly impact band gaps and orbital projections [60].

Benchmarking Against Established Systems and Reference Data

Frequently Asked Questions

Why is there a difference between the band gap reported by my band structure calculation and my Density of States (DOS) calculation?

This is a common issue that arises from the fundamental differences in how these two properties are computed [1]. The band gap can be determined by two primary methods:

• The "Interpolation Method": This method uses the results from the k-space integration scheme that determines the Fermi level and orbital occupations during the self-consistent field (SCF) cycle. It interpolates bands across the entire Brillouin zone (BZ) [1]. The total DOS is also derived from this method [1].
• The "Band Structure Method": This is a post-processing step where the one-electron levels are calculated for a dense set of k-points along a specific, high-symmetry path in the BZ, using a fixed potential from a prior SCF calculation [17] [1]. It does not determine the Fermi level or occupations [1].

The "band structure" method allows for a very dense sampling along a chosen path, which can sometimes more accurately locate the valence band maximum (VBM) and conduction band minimum (CBM) if they lie on that path. However, the "interpolation" method samples the entire BZ, which is a more complete approach, but typically with a coarser k-point grid. Therefore, a discrepancy often indicates that the true CBM or VBM is located at a k-point that is not on the high-symmetry path used for your band structure plot [1] [2].

My DOS plot shows a band gap, but my band structure plot appears metallic. What is wrong?

This is a specific manifestation of the issue described above. It is highly likely that the k-path you selected for the band structure calculation does not pass through the specific k-points in the Brillouin zone where the valence band maximum and the conduction band minimum are located. Consequently, the bands plotted along your path never show the highest point of the valence band or the lowest point of the conduction band, making the material appear to have no gap [1]. The DOS, being a integral over the entire Brillouin zone, correctly identifies the gap.

I have verified my k-path, but my DOS is still not converged and does not match the band structure. What should I do?

The most likely cause is that the k-point grid used for the SCF calculation, which generates the charge density used for both the DOS and band structure calculations, is too sparse [17] [4]. A common problem is "missing DOS," where there are electronic bands in a certain energy range, but the DOS shows no states. This is "caused by an insufficient k-space sampling. Try to Restart the DOS with a better k-grid" [4]. You should test the convergence of your DOS and total energy with respect to the k-point grid density.

What does it mean if my DFT-calculated band gap is significantly smaller than the experimental value?

This is a well-known limitation of standard Density Functional Theory (DFT) when using local (LDA) or semi-local (GGA) exchange-correlation functionals. The band gap error originates from approximations in the functional and a fundamental "derivative discontinuity." [2] Typically, band gaps computed with GGA are underestimated by about 40-50% [2]. For example, internal testing by the Materials Project found that computed gaps were underestimated by an average factor of 1.6, with a mean absolute error of 0.6 eV [2]. This is a systematic error, and more advanced (and computationally expensive) methods like GW approximation or hybrid functionals are required for more accurate gap predictions [2].

Troubleshooting Guides

Guide 1: Resolving Band Structure and DOS Mismatch

Problem: A significant discrepancy exists between the electronic band gap or features observed in the calculated band structure and the Density of States (DOS).

Scope: This guide applies to researchers using plane-wave or DFTB+ codes for electronic structure analysis of periodic systems. Resolving this is critical for accurate prediction of electronic, optical, and transport properties.

Diagnosis and Resolution:

Step	Action	Expected Outcome & Rationale
1	Verify k-grid convergence. Re-run the initial SCF calculation with a denser k-point mesh (e.g., increase KSpace%Quality or use a finer Monkhorst-Pack grid) [1] [4].	Total energy and DOS features become stable. A coarse k-grid is the most common cause of an unconverged DOS that fails to match the band structure.
2	Confirm the band structure path. Ensure the high-symmetry k-path in your band structure calculation passes through all potential locations of the VBM and CBM. Use robust k-path generation tools like those in pymatgen [2].	The band structure plot may reveal a gap at a different k-point. The true band gap (the fundamental gap) might not be on the default path.
3	Recompute the gap from DOS. As a robust check, calculate the band gap directly from the DOS object. In pymatgen, this is done with dos.get_gap() [2].	Provides a gap value based on integration over the entire Brillouin zone, which is often more reliable than the band structure plot for finding the fundamental gap [2].
4	Check calculation parameters. Ensure consistency. The band structure calculation must use the fixed charge density (ReadInitialCharges = Yes in DFTB+) [17] from the converged SCF calculation.	Prevents inconsistencies that arise from using different potentials or charges for the two types of calculations.

The following workflow summarizes the diagnostic process:

Guide 2: Addressing SCF Convergence Failure during Benchmarking

Problem: The self-consistent field (SCF cycle fails to converge when calculating the initial charge density for a new material system, preventing subsequent band structure and DOS analysis.

Scope: This issue is frequently encountered when benchmarking structurally complex systems, metallic systems, or slabs with large surface areas. Convergence is a prerequisite for any reliable electronic structure benchmark.

Diagnosis and Resolution:

Step	Action	Expected Outcome & Rationale
1	Employ more conservative mixing. Decrease the SCF mixing parameter (e.g., SCF%Mixing 0.05) and/or the DIIS parameter (DIIS%Dimix 0.1) [1].	Reduces large charge oscillations between cycles, stabilizing convergence for difficult systems.
2	Switch SCF algorithms. Try alternative algorithms like the MultiSecant method (SCF%Method MultiSecant) or LISTi (DIIS%Variant LISTi) [1].	Different algorithms can escape persistent cycles that cause DIIS to fail, potentially reducing the number of SCF cycles.
3	Apply a finite electronic temperature. Use a small electronic smearing (Convergence%ElectronicTemperature) to partially occupy states around the Fermi level [1].	Helps resolve degeneracies at the Fermi level that can prevent charge density convergence, especially in metals.
4	Increase numerical accuracy. Improve the integration grid quality (NumericalQuality Good) and, for all-electron codes, check the frozen core settings [1].	Ensures that precision issues, such as an insufficient density fit quality, are not the root cause of the convergence failure.

The following workflow outlines the escalation path for resolving SCF convergence failures:

Quantitative Data for Band Gap Analysis

Table 1: Systematic Error in DFT-Calculated Band Gaps

The following table summarizes the expected error in band gaps calculated using standard GGA functionals (like PBE), based on internal benchmarking by the Materials Project [2]. This data is crucial for setting expectations when benchmarking your own calculations.

Metric	Value	Context & Implications
Average Underestimation Factor	1.6x	Computed GGA gaps are, on average, 1.6 times smaller than the experimental value.
Mean Absolute Error (MAE)	0.6 eV	Even after accounting for the systematic shift, a significant residual error remains.
Typical Literature Reported Error	~50%	A common rule of thumb is that LDA and GGA gaps are underestimated by about half.

Experimental Protocols

Protocol 1: Computing a Band Structure and DOS with DFTB+

This protocol provides a detailed methodology for obtaining a band structure and DOS using the DFTB+ code, as described in its official documentation [17].

1. Compute the Self-Consistent Charge (SCC): The first step is to obtain the ground-state charge density.

• Input: Create a dftb_in.hsd file.
• Geometry: Specify the crystal structure using fractional coordinates [17].
• Hamiltonian: Set SCC = Yes and define a tight SccTolerance (e.g., 1e-5). Provide the path to the necessary Slater-Koster files and define the maximum angular momenta for each species [17].
• K-Points: Use a well-converged, dense k-point grid to sample the Brillouin zone. For example, use the SupercellFolding method to generate a 4x4x4 Monkhorst-Pack grid [17].
• Analysis: In the Analysis block, use ProjectStates to request the Partial DOS (PDOS) for specific atoms or shells (e.g., Ti d-orbitals and O p-orbitals) [17].
• Execution: Run DFTB+ with this input file.

2. Calculate the Band Structure:

• Input: Create a new dftb_in.hsd file in a different directory.
• Initial Charges: Copy the charges.bin file from the previous SCC calculation and set ReadInitialCharges = Yes. Limit the SCC cycles to 1 (MaxSCCIterations = 1) as the charges are now fixed [17].
• K-Points: Replace the uniform k-grid with a Klines block. This block defines the specific high-symmetry points (e.g., Z, Gamma, X, P) and the number of k-points to plot between them [17].
• Execution: Run DFTB+ with this new input.

3. Post-Processing and Plotting:

• Total DOS: Use the dp_dos tool from the dptools package on the band.out file from the first calculation to generate the total DOS: dp_dos band.out dos_total.dat [17].
• PDOS: Use the -w (weighting) option with dp_dos to convert the PDOS output files (e.g., dos_ti.1.out): dp_dos -w dos_ti.1.out dos_ti.s.dat [17].
• Visualization: Plot the generated .dat files using a tool like xmgrace or matplotlib.

Protocol 2: Recomputing a Band Gap from Materials Project Data

This protocol uses the Materials Project API and pymatgen to programmatically verify and recompute a material's band gap, which is essential for benchmarking and troubleshooting unexpected results (e.g., a reported 0 eV gap) [2].

1. Retrieve the Material's Data:

• Use the MPRester client from pymatgen to access the Materials Project database.
• Query the summary document for your material of interest (e.g., mp-1211100) to identify the task IDs for its band structure and DOS calculations [2].

2. Recompute the Gap from the DOS (Recommended):

• This is often the most reliable method. Use the task ID for the DOS calculation to fetch the CompleteDOS object.
• Call the get_gap() method on this object to obtain the band gap. This method computes the gap by integrating over the entire Brillouin zone [2].

3. (Optional) Recompute the Gap from the Band Structure:

• If the band structure's Fermi level is suspected to be incorrect, it can be corrected using the VBM from the DOS.
• Fetch the BandStructure object using its task ID.
• Fetch the CompleteDOS object and use its get_cbm_vbm() method to get the correct VBM and CBM energies.
• Create a new, corrected BandStructure object by supplying the original band structure data and the corrected VBM, then call get_gap() on this new object [2].

The Scientist's Toolkit: Research Reagent Solutions

Table 2: Essential Computational Tools for Electronic Structure Benchmarking

This table lists key software tools and data resources that function as the essential "reagents" for conducting and troubleshooting electronic structure calculations.

Item Name	Function / Purpose	Resource Link
DFTB+	A software package for fast quantum mechanical simulations using Density Functional based Tight Binding (DFTB). Used for calculating band structures, DOS, and PDOS.	DFTB+ Recipes
Pymatgen	A robust Python library for materials analysis. It provides powerful tools to parse, analyze, and validate VASP output, generate high-symmetry k-paths, and interface with the Materials Project API.	Pymatgen
Materials Project API	A programmatic interface to a vast database of computed materials properties. It is indispensable for retrieving reference data (band structures, DOS) to benchmark new calculations against.	Materials Project API
dptools	A set of utility scripts distributed with DFTB+ for post-processing results, including the dp_dos tool for generating plottable DOS files.	Bundled with DFTB+

Identifying True Physical Effects vs Computational Artifacts

Frequently Asked Questions (FAQs)

What does a "Band structure does not match the DOS" error mean? This common discrepancy occurs when the band structure plot suggests a semiconductor with a band gap, while the Density of States (DOS) plot shows no gap (metallic behavior), or when the sizes of the band gaps disagree [1] [37]. This can stem from different k-space sampling methods between the two calculations or other computational settings.

Why might my calculation show a metallic DOS but a semiconducting band structure? This inconsistency can have several causes [37]:

• Different k-point grids: The band structure is typically calculated along a high-symmetry path in the Brillouin zone, which might coincidentally show a gap. The DOS, however, uses a uniform grid over the entire Brillouin zone; if this grid is too sparse, it might miss the true band gap [1] [2].
• Excessive smearing: Applying a large smearing width to occupy electronic states can artificially smear out the band gap in the DOS, making an insulator appear metallic [37].
• Magnetic state convergence: For magnetic systems, the band structure and DOS calculations might have converged to different magnetic states or spin configurations, leading to inconsistent electronic pictures [37].

What is the fundamental reason DFT (GGA/PBE) underestimates band gaps? Density Functional Theory (DFT) with common functionals like LDA or GGA (e.g., PBE) is a ground-state theory. The underestimation of band gaps arises primarily from two sources [2]:

• Approximations in the exchange-correlation functional.
• A derivative discontinuity in the true exchange-correlation potential, which is not captured by standard functionals. This error can be significant; for instance, the Materials Project found computed gaps are underestimated by an average factor of 1.6 (~40%) compared to experiment [2].

My system has negative frequencies in its phonon spectrum. Is this a physical effect? Not typically for a stable, optimized structure. Negative frequencies (imaginary phonon modes) are most often a computational artifact indicating that the geometry was not fully optimized to a minimum or that the step size used in the phonon calculation was too large [1]. General numerical inaccuracies in integration can also be the cause [1].

Troubleshooting Guides

Band Structure and DOS Mismatch

Symptoms: The electronic band structure plot shows a band gap (semiconducting/insulating behavior), but the total DOS plot shows no gap at the Fermi level (metallic behavior) [37]. Alternatively, the band gap value extracted from the DOS differs from the value found in the band structure [11] [2].

Diagnosis and Solutions:

Table: Troubleshooting Band Structure and DOS Mismatch

Possible Cause	Diagnostic Check	Solution
Insufficient k-points for DOS [1] [2]	Check if the DOS converges with a higher KSpace%Quality or a denser k-mesh.	Increase the k-point sampling density for the DOS calculation. The DOS typically requires a denser uniform grid than a single SCF calculation.
Large Smearing Value [37]	Check the smearing width (sigma in VASP, smearing in Quantum ESPRESSO) used in the DOS calculation.	Reduce the smearing width, especially for semiconductors and insulators. Use the minimal value needed for convergence.
Different Fermi Level Placement	Verify the Fermi level is consistent between the band structure and DOS plots.	Manually align the Fermi level to zero in both plots during post-processing, or recompute the band structure's Fermi level using the VBM from the DOS [2].
Inconsistent Magnetic States [37]	Confirm the final magnetic moments on each ion are identical in both calculations.	Ensure both the SCF and non-SCF calculations are restarted from the same charge density and wavefunctions to maintain consistency.
Inaccurate Band Gap Parsing [2]	Use code to recompute the gap directly from the DOS or band structure data.	Recompute the band gap programmatically from the density of states object (e.g., dos.get_gap() in pymatgen) for a more robust value [2].

SCF Convergence Failure

Symptoms: The Self-Consistent Field (SCF) cycle oscillates and fails to converge within the set iteration limit.

Diagnosis and Solutions: This is common in metallic systems, slabs, and systems with heavy elements [1].

• Use more conservative mixing parameters. Decrease the mixing parameter to stabilize the convergence [1].

• Change the SCF algorithm. Switch from the default DIIS method to the MultiSecant method, which has a similar computational cost [1].

• Employ finite electronic temperature. Using a small electronic smearing (finite temperature) can help initial convergence. For geometry optimizations, use EngineAutomations to start with a higher temperature and reduce it as the geometry converges [1].
• Start from a smaller basis set. First, converge the calculation with a minimal basis set (e.g., SZ), then use the resulting density as a starting point for a calculation with the larger desired basis set [1].

Core-Level Shift Artifacts in Periodic Calculations

Symptoms: Calculated core-level binding energies (using the ΔSCF or transition-state method) show unphysical shifts of over 1 eV when changing the supercell size or for atoms far from the substrate [61].

Diagnosis and Solutions: This is a known artifact in periodic boundary condition calculations when a core hole is created in every unit cell, forming an artificial dipole layer that affects the electrostatic potential [61].

• Diagnosis: Monitor the calculated work function change due to the core-level excitation. A significant change can indicate this artifact [61].
• Solutions:
  • Use larger supercells to reduce the interaction between periodic images of the core hole [61].
  • Consider cluster-based calculations with electrostatic embedding to avoid the periodic replication of the core hole entirely [61].

Experimental Protocols & Methodologies

Protocol for a Robust Band Gap Calculation

This protocol, based on the methodology of the Materials Project, ensures consistent electronic structure analysis [2].

• Geometry Optimization: Fully optimize the crystal structure until all forces on atoms are below a tight threshold (e.g., 0.01 eV/Ã…) to ensure the geometry is in its ground state.
• Self-Consistent Field (SCF) Calculation: Perform a high-quality SCF calculation on the optimized structure using a dense, uniform k-point grid (e.g., a Monkhorst-Pack grid) to obtain the converged charge density.
• Non-SCF Density of States (DOS) Calculation: Using the converged charge density from step 2, perform a non-SCF calculation on an even denser uniform k-point grid to compute the DOS. The quality of this grid is critical for an accurate gap.
• Non-SCF Band Structure Calculation: Using the same converged charge density, perform a separate non-SCF calculation with k-points along high-symmetry lines in the Brillouin zone (e.g., following a standard path like Γ-K-M-Γ).
• Validation and Analysis:
  • Extract the band gap from the DOS using a robust method (e.g., dos.get_gap() in pymatgen) [2].
  • Compare this DOS gap with the gap observed in the band structure plot.
  • Ensure the Fermi level is consistently aligned.

The following workflow visualizes this protocol:

Logical Troubleshooting Pathway for DOS/Band Structure Mismatch

Follow this decision tree to systematically identify the cause of a discrepancy.

The Scientist's Toolkit: Research Reagent Solutions

Table: Essential Computational Parameters and Their Functions

Computational Parameter / 'Reagent'	Function / Role	Troubleshooting Application
k-point Grid Density	Determines the sampling density of the Brillouin zone. A denser grid leads to more accurate integration.	Primary fix for inaccurate DOS and mismatches with band structure. Use a denser grid for DOS than for the initial SCF [1] [2].
Smearing Width	Applies a finite electronic temperature to help SCF convergence in metals by occupying states near the Fermi level.	Can artificially smear out a band gap in the DOS. Reduce or turn off for insulators [37].
SCF Mixing Parameter	Controls how much of the new electron density is mixed with the old in each SCF cycle.	Decrease to fix SCF convergence oscillations [1].
MultiSecant / DIIS Method	Algorithms to find the self-consistent solution. They extrapolate the solution to speed up convergence.	Switch from DIIS to MultiSecant as an alternative, cost-effective收敛 algorithm [1].
Electronic Temperature (kT)	A form of smearing; a finite value helps initial convergence.	Use automations to start with a high kT during geometry optimization and reduce it as the structure converges [1].
Confinement Radius	Reduces the diffuseness of atomic basis functions.	Can resolve linear dependency errors in the basis set, which can cause SCF failures [1].

Conclusion

Resolving band structure-DOS mismatches requires meticulous attention to computational parameters and methodological consistency. Key takeaways include the necessity of k-point convergence testing, appropriate smearing selection for different dimensionalities, and systematic cross-validation between calculation methods. For biomedical and clinical research, these resolution strategies ensure reliable electronic structure predictions crucial for understanding drug-material interactions, designing biomedical devices, and developing novel therapeutic materials. Future directions should focus on automated convergence protocols, machine learning-assisted parameter optimization, and standardized benchmarking across computational platforms to enhance reproducibility in computational materials design.

References

• 1. — Troubleshooting .1 documentation BAND 2025 [https://www.scm.com/doc/BAND/Troubleshooting/Trouble_Shooting.html]
• 2. Electronic Structure | Materials Project Documentation [https://docs.materialsproject.org/methodology/materials-methodology/electronic-structure]
• 3. , DOS and PDOS — DFTB+ Recipes Band structure [https://dftbplus-recipes.readthedocs.io/en/latest/basics/bandstruct.html]
• 4. (DOS) — Density .1 documentation of States BAND 2025 [https://www.scm.com/doc/BAND/Analysis/DOS.html]
• 5. [Pw_forum] projwfc problem: sum of pdos doesn't match total dos [https://lists.quantum-espresso.org/pipermail/users/2008-August/009784.html]
• 6. Band Structure — BAND 2025.1 documentation [https://www.scm.com/doc/BAND/Analysis/Band_Structure.html]
• 7. of the Sampling with BZ - k — Siesta Documentation points [https://docs.siesta-project.org/projects/siesta/en/latest/tutorials/basic/kpoint-convergence/index.html]
• 8. Brillouin zone integration [https://www.questaal.org/docs/numerics/bzintegration/]
• 9. Efficiency of Generalized Regular k-point grids - ScienceDirect [https://www.sciencedirect.com/science/article/abs/pii/S0927025618304105]
• 10. k-point convergence - Mat3ra Documentation [https://docs.mat3ra.com/tutorials/dft/addons/kpt-convergence/]
• 11. Significant mismatch between the bandgap values computed by... [https://github.com/materialsproject/pymatgen/issues/449]
• 12. density functional theory - Increasing k-point grid to take DOS calculation - Matter Modeling Stack Exchange [https://mattermodeling.stackexchange.com/questions/5045/increasing-k-point-grid-to-take-dos-calculation]
• 13. Mueller Research Group - K-Point Grid Server [https://muellergroup.jhu.edu/K-Points.html]
• 14. technique - VASP Wiki Smearing [https://www.vasp.at/wiki/Smearing_technique]
• 15. Preparation of hydrogen, fluorine and chlorine doped and co-doped titanium dioxide photocatalysts: a theoretical and experimental approach | Scientific Reports [https://www.nature.com/articles/s41598-021-81979-x]
• 16. A roadmap of strain in doped anatase | Scientific Reports TiO 2 [https://www.nature.com/articles/s41598-018-30747-5?error=cookies_not_supported&code=e582c59c-eec7-41f0-817a-d83710e625bd]
• 17. , Band and PDOS — DFTB+ Recipes... structure DOS [https://dftbplus-recipes.readthedocs.io/en/stable/basics/bandstruct.html]
• 18. Электронный научный архив УрФУ: Ab initio calculations of pure and... [https://elar.urfu.ru/handle/10995/101553?mode=full]
• 19. Quantum Espressoで TiO アナターゼの電子のバンド構造と 2 を描く DOS [https://gist.github.com/t-nissie/ad99810ddf1bceb602f6]
• 20. Frontiersin [https://www.frontiersin.org/journals/chemistry/articles/10.3389/fchem.2025.1621188/xml/nlm]
• 21. Fluorine-doped titanium dioxide nanorod arrays for efficient photoelectrochemical water splitting [https://www.oaepublish.com/articles/microstructures.2024.165]
• 22. Wolverton Research Group - PWSCF Guide [https://www.wolverton.northwestern.edu/classes/458/pwscf-guide]
• 23. the Converging - k density — aimstools 0.6.4 documentation point [https://aims-tools.readthedocs.io/en/master/workflows/kpoint_convergence.html]
• 24. Example: DOS and BandStructure from a previous calculation ... [https://www.scm.com/doc/BAND/Examples/RestartDosAndBandStructure.html]
• 25. Properties and applications of basis with sets ... linearly dependent [https://link.springer.com/article/10.1007/s00214-025-03175-8]
• 26. 8.3.1 Customization‣ 8.3 Basis Symbolic Representation ‣ Chapter... Set [https://manual.q-chem.com/5.3/Ch8.S3.SS1.html]
• 27. Predicting linear dependencies in a basis set in advance of calculating any integrals - Matter Modeling Stack Exchange [https://mattermodeling.stackexchange.com/questions/1106/predicting-linear-dependencies-in-a-basis-set-in-advance-of-calculating-any-inte]
• 28. Dependency (basis set, fit set) — ADF 2025.1 documentation [https://www.scm.com/doc/ADF/Input/Dependency.html]
• 29. dos [https://www.vasp.at/py4vasp/latest/calculation/dos/]
• 30. calculation - Ground state - ABINIT Dos [https://discourse.abinit.org/t/dos-calculation/3701]
• 31. Multiple Manifestations of Negative Local Partial Density of States [https://arxiv.org/abs/2503.07053]
• 32. Band Structure, Projected Density of States and Effective Mass Calculations — | QuantumATKX-2025.06 Documentation [https://docs.quantumatk.com/intro/intro_tutorials/bandstructure-dos-effective-masses-Si-SiO2/bandstructure-dos-effective-masses.html]
• 33. Understanding Density in Quantum Mechanics and... of States [https://etcnmachining.com/blog/density-of-states/]
• 34. GitHub - DeeliN221/ddpc: DFT Data Processing Core [https://github.com/DeeliN221/ddpc]
• 35. Frontiers | The Basics of Electronic Structure Theory for Periodic Systems [https://www.frontiersin.org/journals/chemistry/articles/10.3389/fchem.2019.00106/full]
• 36. [Pw_forum] Inconsistency between DOS and Band structure [https://lists.quantum-espresso.org/pipermail/users/2016-September/036313.html]
• 37. inconsistent behavior between band and density of states structure [https://mattermodeling.stackexchange.com/questions/4096/inconsistent-behavior-between-band-structure-and-density-of-states]
• 38. Restart the DOS or band with structure — Tutorials... BAND [https://www.scm.com/doc/Tutorials/Analysis/RestartDosBandStructure.html]
• 39. Guidelines for ADF — ADF SCF .1 documentation Convergence 2025 [https://www.scm.com/doc/ADF/Rec_problems_questions/SCF.html]
• 40. Self-consistent field ( SCF ) methods — PySCF [https://pyscf.org/user/scf.html]
• 41. Anderson acceleration. Convergence analysis and applications to equilibrium chemistry - ScienceDirect [https://www.sciencedirect.com/science/article/abs/pii/S0168927424000229]
• 42. SCF | Gaussian.com [https://gaussian.com/scf/]
• 43. 7.7. SCF - ORCA 6.0 Manual Convergence [https://www.faccts.de/docs/orca/6.0/manual/contents/detailed/scfconv.html]
• 44. Convergence acceleration of iterative sequences for equilibrium chemistry computations | Computational Geosciences [https://link.springer.com/article/10.1007/s10596-021-10054-1]
• 45. 2.4.2 Electronic Band Structure of SW-CNTs [https://www.iue.tuwien.ac.at/phd/pourfath/node19.html]
• 46. Single Wall Carbon Nanotubes in the Presence of Vacancies and Related Energy Gaps | IntechOpen [https://www.intechopen.com/chapters/16863]
• 47. ISMEAR, SIGMA, FERWE, FERDO tag [https://wanglab.hosted.uark.edu/vasp/vasp/cms.mpi.univie.ac.at/vasp/vasp/node133.html]
• 48. [2103.03469] Comparison of the Tetrahedron Method to Smearing Methods for the Electronic Density of States [https://arxiv.org/abs/2103.03469]
• 49. [QE-users] smearing . vs - weird behavior tetrahedron [https://lists.quantum-espresso.org/pipermail/users/2019-March/042445.html]
• 50. density functional theory - The "Tetrahedron method does not include variations of the Fermi occupations" Problem - Matter Modeling Stack Exchange [https://mattermodeling.stackexchange.com/questions/11091/the-tetrahedron-method-does-not-include-variations-of-the-fermi-occupations-pr]
• 51. Crystalline Aluminum — STATE 0.0.1 documentation [https://state-doc.readthedocs.io/en/5.6.13/Tutorial/al.html]
• 52. ISMEAR - VASP Wiki [https://www.vasp.at/wiki/index.php/ISMEAR]
• 53. Documentation - SCM [https://www.scm.com/doc.2016/BAND/Rec_problems_questions/Trouble_Shooting.html]
• 54. Understanding Band in Materials Science - Simple Science Gaps [https://scisimple.com/en/articles/2025-04-24-understanding-band-gaps-in-materials-science--a3oxlyn]
• 55. Band gap determination using absorption spectrum fitting procedure | International Nano Letters [https://link.springer.com/article/10.1186/2228-5326-3-2]
• 56. Direct and indirect band - Wikipedia gaps [https://en.wikipedia.org/wiki/Direct_and_indirect_band_gaps]
• 57. Band gap - Wikipedia [https://en.wikipedia.org/wiki/Band_gap]
• 58. A band-gap database for semiconducting inorganic materials calculated with hybrid functional | Scientific Data [https://www.nature.com/articles/s41597-020-00723-8]
• 59. Electronic band , HSE (VASP) - Mat3ra Documentation structure [https://docs.mat3ra.com/tutorials/dft/electronic/hse-vasp-bg/]
• 60. A framework for quantifying uncertainty in DFT energy corrections - PMC [https://pmc.ncbi.nlm.nih.gov/articles/PMC8322326/]
• 61. Final-State Simulations of Core-Level Binding Energies at... [https://pmc.ncbi.nlm.nih.gov/articles/PMC7557941/]