Sublattice-Based Defect Engineering

• Sublattice-based defect engineering is a strategy that uses inequivalent lattice sites to direct defect formation, coupling, and phase selection in functional materials.
• It enables precise tuning of electronic, optical, magnetic, and topological properties through controlled defect placement in systems like chalcogenides, perovskites, and 2D materials.
• The approach integrates advanced synthesis and diagnostic techniques to manipulate defect landscapes, offering pathways for enhanced catalytic, quantum, and photonic device performance.

Searching arXiv for recent and relevant papers on sublattice-based defect engineering across materials platforms. arxiv_search(query="sublattice defect engineering materials defects sublattice", max_results=10, sort_by="relevance") arxiv_search(query="sublattice-based defect engineering", max_results=10, sort_by="relevance") Sublattice-based defect engineering is the deliberate use of crystallographically inequivalent site networks—cation and anion sublattices, bipartite A/B lattices, or majority and minority site sets—as separate control variables for defect formation, defect coupling, phase selection, and functional response. In this view, the decisive question is not only which defect is present, but which sublattice it occupies, depletes, couples to, or reconstructs. Across layered chalcogenides, honeycomb materials, perovskites, ceramics, topological magnets, and artificial lattices, this strategy has been used to tune band gaps, orbital hybridization, charge separation, magnetic exchange, octahedral rotations, helium trapping, and topological defect states (Ushkov et al., 9 Dec 2025, Wang et al., 2020, Shin et al., 2021, Daghbouj et al., 6 Jan 2026).

1. Sublattices as the primary degrees of freedom

The central structural premise is that many functional solids are not chemically uniform graphs of equivalent sites. CdPS$_3$, for example, contains a Cd sublattice of Cd$^{2+}$ ions in octahedral CdS$_6$ coordination, a P sublattice in covalently bonded $\mathrm{P_2}$ units within $\mathrm{P_2S_6^{4-}}$, and an S sublattice shared between Cd–S and P–S bonding. Monolayer hBN is a two-sublattice honeycomb with inequivalent B and N sites. In perovskites such as EuTiO$_3$ and LaAlO$_3$, the A-site, B-site, and oxygen sublattices are distinct chemical actors rather than interchangeable lattice points. In $\alpha$-SiC, the Si and C sublattices have markedly different vacancy formation energies, so irradiation does not populate them symmetrically (Ushkov et al., 9 Dec 2025, Wang et al., 2020, Shin et al., 2021, Jia et al., 2022, Daghbouj et al., 6 Jan 2026).

This distinction matters because defects inherit the symmetry, coordination, and chemistry of the sublattice on which they are created. A carbon vacancy in 3C-SiC is not equivalent to a silicon vacancy; a C$_\text{B}$ defect in hBN is not equivalent to a V$_\text{N}$; an Eu vacancy in EuTiO$^{2+}$0 perturbs exchange pathways differently from an oxygen vacancy; and a line defect that connects A–A sites in a buckled honeycomb lattice does not behave like one that connects A–B sites. In artificial bipartite systems, even the imbalance between majority and minority sublattices within a unit cell can be used to isolate defect zero modes on only one sublattice by selectively breaking chiral symmetry on the other (Poli et al., 2015).

A recurrent consequence is that the electronic states associated with a defect may live on a different sublattice from the structural defect itself. In ReS$^{2+}$1, sulfur vacancies are defects of the S sublattice, yet the corresponding in-gap states are localized mainly on neighboring Re atoms. In the proposed Hf- and Zr-vacancy qubits in 4H-SiC and w-AlN, the vacancy is on the anion sublattice, while the active spin density resides primarily on cation dangling bonds with $^{2+}$2 symmetry. This separation between structural location and electronic support is one of the defining features of the field (Jiang et al., 2018, Seo et al., 2017).

2. Microscopic mechanisms of sublattice selectivity

Several distinct mechanisms recur across the literature. One is non-equilibrium phase selection. In femtosecond pulsed laser ablation in liquid of CdPS$^{2+}$3, the solvent is explicitly described as “not a passive heat sink”: it sets redox potential, solvation of fragments, and quenching rate. Water preserves near-ideal Cd:P:S $^{2+}$4, acetonitrile yields mixed CdPS$^{2+}$5/CdS populations, and isopropanol drives strong P-sublattice depletion with CdS formation and Cd$^{2+}$6 defect sites. The thermodynamic bias is also sublattice-specific: the formation energy of CdS is $^{2+}$7, more negative than $^{2+}$8 for CdPS$^{2+}$9, so once Cd, P, and S are liberated into the plasma, the Cd–S sublattice preferentially reforms as binary CdS (Ushkov et al., 9 Dec 2025).

A second mechanism is defect-orbital hybridization constrained by lattice geometry. In hBN, pairs of in-gap defect orbitals hybridize into bonding and antibonding combinations with splittings that depend not only on distance but on sublattice, orientation, and out-of-plane configuration. For nearest-neighbor pairs, the reported splittings are $_6$0 meV for cis CH$_6$1–CH$_6$2@1, $_6$3 meV for trans CH$_6$4–CH$_6$5@1, and $_6$6 meV for in-plane C$_6$7–C$_6$8@1. The paper emphasizes that the host lattice is “not merely a passive dielectric background”: local strain and relaxation reshape orbital overlap and binding energies, and even produce non-monotonic overlap as the separation vector changes on the honeycomb lattice (Wang et al., 2020).

A third mechanism is electron-mediated sublattice ordering. In dilute graphene with on-site adsorbates or vacancies, defects act as Ising variables $_6$9 depending on whether they occupy A or B sites. The effective interaction is long-ranged, ferromagnetic in sublattice space, and decays as $\mathrm{P_2}$0, favoring same-sublattice occupation. The resulting order parameter $\mathrm{P_2}$1 generates a Dirac mass $\mathrm{P_2}$2, opens a gap, and produces domain walls with mid-gap channels (Cheainov et al., 2010).

A fourth mechanism is defect–lattice coupling through local dipoles and tilts. In rhombohedral LaAlO$\mathrm{P_2}$3, the A-site antisite Al$\mathrm{P_2}$4 is much smaller than La$\mathrm{P_2}$5, off-centers by about $\mathrm{P_2}$6 Å, and creates a large local dipole. DFT and the effective lattice model show that this A-sublattice defect couples directly to the oxygen sublattice and locally changes the octahedral rotation pattern from the bulk $\mathrm{P_2}$7 state to $\mathrm{P_2}$8 near the defect, with rotation angles increasing from about $\mathrm{P_2}$9 in pristine rhombohedral LaAlO$\mathrm{P_2S_6^{4-}}$0 to about $\mathrm{P_2S_6^{4-}}$1 near the defect (Jia et al., 2022).

Taken together, these results show that “sublattice-based” does not denote a single mechanism. It can mean solvent-directed reoccupation of distinct chemical networks, distance- and orientation-dependent orbital hybridization, electron-mediated ordering of bipartite occupations, or coupling of a defect dipole on one sublattice to collective distortions on another. This suggests that sublattice selectivity is best understood as a symmetry-and-chemistry filter applied to whichever microscopic process dominates in a given material.

3. Representative material systems and outcomes

Representative realizations span photocatalytic chalcogenides, quantum defects, photoconductors, oxides, radiation-tolerant ceramics, interfacial metals, and magnetic topological compounds (Ushkov et al., 9 Dec 2025, Schöler et al., 2019, Shin et al., 2021, Daghbouj et al., 6 Jan 2026, Jain et al., 10 Nov 2025, Chen et al., 25 Dec 2025, Jiang et al., 2018).

Platform	Sublattice or defect handle	Reported consequence
CdPS$\mathrm{P_2S_6^{4-}}$2/CdS	P-sublattice depletion; Cd–S reallocation; Cd$\mathrm{P_2S_6^{4-}}$3 sites	$\mathrm{P_2S_6^{4-}}$4 Methylene Blue degradation in 30 min under 532 nm
hBN	B-sublattice substitutions and N-sublattice vacancies	Splittings from $\mathrm{P_2S_6^{4-}}$5 meV to nearly $\mathrm{P_2S_6^{4-}}$6 eV; visible-to-near-IR optical tuning
Graphene	A/B sublattice ordering of dilute defects	Gap opening and domain-wall channels
3C-SiC	Carbon vacancies and Si-sublattice dopant complexes	ZPLs at 1.278, 1.233, 1.171, and 1.138 eV
EuTiO$\mathrm{P_2S_6^{4-}}$7	Eu–O vacancies on A-site and oxygen sublattices	AFM-to-FM transition with $\mathrm{P_2S_6^{4-}}$8
$\mathrm{P_2S_6^{4-}}$9-SiC	Pre-created vacancy clusters, dominated by C-sublattice vacancies	Stable nanobubbles instead of platelets and nanocracks
EG/SiC + Ag	Vacancy/line defects in EG versus sp$_3$0-rich ZLG	Phase-selective Ag(1) or Ag(2) monolayers
ReS$_3$1	S-sublattice vacancies decorated by H$_3$2PP	$_3$3 Jones

Several of these platforms illustrate how the same conceptual vocabulary translates into very different physical regimes. In 3C-SiC, Si-rich sublimation growth favors carbon vacancies on the C sublattice, and those vacancies become the common structural element of isolated $_3$4, $_3$5, $_3$6, and the proposed $_3$7 complex. The observed near-infrared zero-phonon lines at 1.278, 1.233, 1.171, and 1.138 eV are therefore a direct record of how the C-sublattice vacancy couples to different defects on the Si sublattice (Schöler et al., 2019).

In EuTiO$_3$8, the engineered defects sit on the Eu and O sublattices, not on Ti. XPS and XRD correlate decreasing Eu/Ti ratio with increased unit-cell volume, and magnetic measurements show that sufficiently defect-rich films display a ferromagnetic transition at $_3$9 K, with $_3$0 approaching $_3$1. Optical spectroscopy and DFT connect this to the suppression of Eu–Ti–Eu antiferromagnetic superexchange and the emergence of ferromagnetic Eu–O–Eu exchange (Shin et al., 2021).

In $_3$2-SiC under helium implantation, the relevant engineering target is not a single point defect but an entire defect landscape. DFT gives $_3$3 eV and $_3$4 eV, so irradiation preferentially generates C-sublattice vacancies. When vacancy clusters are pre-created before helium exposure, helium is trapped in stable, uniformly dispersed nanobubbles instead of forming extended platelets and nanocracks. The authors describe this as a shift “from a passive property to an actively tunable design parameter” (Daghbouj et al., 6 Jan 2026).

At the graphene/SiC interface, defect class controls which 2D Ag phase forms under confinement heteroepitaxy. Vacancy and line defects in epitaxial graphene expose the SiC surface and favor the nearly commensurate Ag(1) phase; sp$_3$5-rich zero-layer graphene instead stabilizes the denser Ag(2) phase. The phase choice propagates into different graphene doping levels and a three-order magnitude difference in second-order susceptibility (Jain et al., 10 Nov 2025).

In ReS$_3$6, the relevant functional problem is not phase stability but trap-state distribution. Sulfur vacancies generate shallow and deep traps; Protoporphyrin molecules selectively decorate those vacancies, largely eliminate the deep traps that produce minute-scale decay, and leave carrier recombination plus shallow traps to dominate the transient response (Jiang et al., 2018).

4. Electronic, optical, magnetic, and topological consequences

One major branch of the field concerns defect complexes as engineered quantum states. In hBN, the C$_3$7–V$_3$8 complex is reinterpreted as an artificial molecule formed from a B-sublattice substitution and an N-sublattice vacancy. As the defect separation changes from $_3$9 to $\alpha$0, the lowest excitation energy decreases from about $\alpha$1 eV to about $\alpha$2 eV, shifting the predicted absorption from the red part of the visible spectrum to the near-infrared. The same work reports bonding–antibonding splittings ranging from $\alpha$3 meV to nearly $\alpha$4 eV, establishing a direct link between sublattice geometry and in-gap molecular spectroscopy (Wang et al., 2020).

A second branch concerns topological and symmetry-protected defect states. In a dimerized Lieb lattice, partial chiral symmetry breaking is implemented by adding couplings only on the majority sublattice, while leaving the minority-sublattice block chiral. This lifts the flat-band compactons away from zero energy yet leaves a defect zero mode fully supported on the minority sublattice spectrally isolated. Sublattice-staggered absorption then suppresses all modes with majority-sublattice weight while leaving the minority-sublattice zero mode essentially unchanged (Poli et al., 2015).

Related line-defect physics appears in graphene and silicene. In graphene with a staggered A/B potential, a mirror-symmetric line defect supports a gapless state that traverses the bulk gap and is independent of ribbon width, defect position, and edge potentials; conductance calculations yield a single-channel plateau at $\alpha$5, so the defect acts as a one-dimensional quantum channel controlled by the defect on-site energy $\alpha$6 (Song et al., 2013). In zigzag silicene nanoribbons with a 5–5–8 line defect, the decisive distinction is whether the defect connects opposite sublattices or same-type sublattices. For the A–A configuration, changing the sublattice potential, the defect on-site energy, and the spin–orbit coupling can switch between a band gap and a gapless state and can even produce a spin–valley-polarized metal (Wan et al., 2021).

A third branch concerns interacting topological spin centers. In the interacting many-body SSH model, the bipartite A/B structure and the placement of bond defects determine where zero modes appear. Once on-site Hubbard repulsion is included, these zero modes become localized spin centers. By careful addition of defects, the paper shows that one can engineer any number of localized spin centers, grouped into singlet or triplet channels within blocks separated by the defects, thereby constructing a chain of spin qubits (Wang et al., 24 Jul 2025).

Magnetic topological materials add another layer. In Mn(Bi$\alpha$7Sb$\alpha$8)$\alpha$9Te$_\text{B}$0, the crucial defects are Mn$_\text{B}$1 and Bi/Sb$_\text{B}$2 antisites. First-principles calculations show that increasing antisite density progressively destroys the field-forced magnetic Weyl state: a clean type-II Weyl semimetal at low defect density becomes a gapped state and then a trivial magnetic insulator as antisite density increases. This is a literal case in which sublattice occupancy controls whether Weyl nodes exist at all (Chen et al., 25 Dec 2025).

5. Engineering routes and diagnostic toolkits

The engineering routes reported in this literature are unusually diverse. In CdPS$_\text{B}$3, femtosecond PLAL is performed with a Yb:KGW source at $_\text{B}$4, pulse width $_\text{B}$5, repetition rate $_\text{B}$6, pulse energy $_\text{B}$7, focused to $_\text{B}$8 for a fluence of $_\text{B}$9. A bulk crystal is ablated for 10 min in 5 mL of water, acetonitrile, or isopropanol, with the focus rastered over a $_\text{N}$0 area. The method is explicitly described as surfactant-free and scalable, and the structural outcome is read out with SAED, EDX, Raman, and photoluminescence (Ushkov et al., 9 Dec 2025).

In 3C-SiC, the route is growth-embedded rather than post-growth. Sublimation epitaxy under Si-rich conditions and growth rates between $_\text{N}$1 and $_\text{N}$2 control whether carbon-vacancy-related NIR centers appear at all; the sharp $_\text{N}$3-related band is strongest in the $_\text{N}$4–$_\text{N}$5 regime and decreases again at the highest rates because of concentration quenching. The emphasis is not on creating defects after synthesis but on embedding them into the crystal during growth (Schöler et al., 2019).

In radiation-tolerant ceramics, the route is pre-damage before service. In $_\text{N}$6-SiC, C self-ions create vacancy clusters before helium implantation. The resulting defect landscape is quantified by advanced microscopy, nano-precession electron diffraction strain mapping, ERDA helium depth profiling, positron annihilation spectroscopy, and atomistic simulations. Here the important diagnostic is not a single spectroscopic line but the conversion of the helium defect population from large platelets and nanocracks to uniformly dispersed nanobubbles (Daghbouj et al., 6 Jan 2026).

At the EG/SiC interface, phase-selective growth of Ag is controlled by preparing either low-defect EG, plasma-damaged EG, or sp$_\text{N}$7-rich ZLG before intercalation. The structural fingerprint is then extracted by LEED, cross-sectional TEM, HAADF-STEM, STM, AES, Raman, and ARPES. The combined dataset allows the Ag(1)/Ag(2) phase boundary to be correlated with the original defect texture of the graphene cap (Jain et al., 10 Nov 2025).

In Mn(Bi$_\text{N}$8Sb$_\text{N}$9)$^{2+}$00Te$^{2+}$01, optimized chemical vapor transport rather than local microscopy is the decisive processing advance. The route uses MnTe:(Bi,Sb)$^{2+}$02Te$^{2+}$03 starting ratios $^{2+}$04 with $^{2+}$05–3, iodine transport, and a composition-dependent two-zone temperature schedule to suppress antisites on the magnetic and heavy-element sublattices. The resulting sublattice occupancies are not read out directly by atom counting alone, but by combining EDX with magnetization through the relations $^{2+}$06 and $^{2+}$07 (Chen et al., 25 Dec 2025).

A broader methodological point emerges from these cases. The same concept can be implemented by laser-driven non-equilibrium synthesis, defect-specific chemical growth windows, pre-irradiation, molecule decoration, patterned line defects, or topological bond engineering. This suggests that sublattice-based defect engineering is less a single technique than a common design language that can be translated into whichever synthesis modality best addresses a given sublattice.

6. Misconceptions, limitations, and open problems

Several studies directly challenge recurring misconceptions. One is the idea that the surrounding medium or host lattice is passive. In fs-PLAL of CdPS$^{2+}$08, the solvent is “not a passive heat sink”; in hBN artificial molecules, the host lattice is “not merely a passive dielectric background”; and in ceramic helium tolerance, the defect landscape is redefined from a fixed material property to an actively tunable design parameter (Ushkov et al., 9 Dec 2025, Wang et al., 2020, Daghbouj et al., 6 Jan 2026).

Another misconception is that defect minimization is always the correct objective. The ReS$^{2+}$09 photoconductor shows the opposite: deep traps are detrimental because they generate $^{2+}$10 s decay tails, but a small residual density of shallow traps is functionally useful because it sustains photogating and high detectivity. Likewise, in $^{2+}$11-SiC, deliberately pre-creating vacancy clusters before helium exposure is the protective strategy rather than a source of additional damage (Jiang et al., 2018, Daghbouj et al., 6 Jan 2026).

The literature also makes clear that chemical substitution alone is often insufficient if it changes defect chemistry in the wrong direction. In Mn(Bi$^{2+}$12Sb$^{2+}$13)$^{2+}$14Te$^{2+}$15, Sb substitution is used to tune the Fermi level toward charge neutrality, but it also exacerbates Mn–Sb antisite formation; reducing antisite density by optimized synthesis is therefore a prerequisite for recovering the type-II Weyl state (Chen et al., 25 Dec 2025).

A number of open questions remain material-specific. In 3C-SiC, the $^{2+}$16 assignment of the 1.171 eV line is explicitly tentative, and the work does not yet provide ODMR, $^{2+}$17, $^{2+}$18, or $^{2+}$19 for the grown defects (Schöler et al., 2019). In hBN, transition energies and geometry trends are robust, but the paper notes that quantitative emission energies are shifted by PBE band-gap underestimation and missing many-body corrections (Wang et al., 2020). In Ag(2)/SiC, pure-interface DFT yields self-doped metallic behavior, while the experiment with graphene capping is semiconducting; the proposed reconciliation is charge transfer to graphene plus trapping in point defects within the Ag layer (Jain et al., 10 Nov 2025).

Taken together, these studies suggest a practical set of design rules. First, identify the inequivalent sublattices that actually control the target property. Second, determine whether the active electronic state lives on the structural defect site or on neighboring sites of another sublattice. Third, treat kinetics, charge state, and local relaxation as co-equal design variables, not afterthoughts. Fourth, engineer defect density rather than simply defect presence or absence. Under those conditions, “defect engineering” becomes a precise form of sublattice engineering: a way to redistribute occupation, symmetry, and coupling across a crystal so that a chosen function—catalytic, magnetic, optical, topological, or quantum—emerges by design.