Biosensors and Nanotheranostics

1. Introduction

Anisotropic metal nanostructures, a family that includes nanorods, nanoprisms, nanostars, bimetallic core–shell architectures and three-dimensional (3D) chiral nanohelicoids, now sit at the centre of nanophotonics, catalysis and sensing (Langille et al., 2013; Tan et al., 2017). Much of their appeal traces back to a single phenomenon, the localized surface plasmon resonance (LSPR), in which free conduction electrons oscillate collectively and coherently in response to incident light (Baffou & Quidant, 2013; Besteiro et al., 2021). A spherical particle typically shows one dipolar absorption band. Break that symmetry, however, and the optical response fragments into several shape-dependent modes that concentrate the electromagnetic field into "hotspots" at tips, edges and interparticle gaps, allowing the response to be tuned across the visible and near-infrared (NIR) (Besteiro et al., 2021; Brongersma et al., 2015).

How such shapes are made has, for a long time, been a question for wet chemistry. Seed-mediated growth, direct chemical reduction and surfactant-assisted capping remain the workhorses of the field, and they are very good at what they do: they deliver crystalline particles with narrow size distributions in respectable yields (Nikoobakht & El-Sayed, 2003; Xia et al., 2017). Yet these routes are bounded by the face-centred cubic (fcc) symmetry of gold and silver and by the thermodynamics of their surface facets (Langille et al., 2013; Xia et al., 2017). Producing a genuinely asymmetric or chiral particle usually means recruiting chiral organic ligands or biomolecular templates such as peptides and DNA (Kuzyk et al., 2017; Lv et al., 2022). Thermal solution chemistry also offers little in the way of spatial addressability, wavelength selectivity or dynamic reconfiguration, which makes it difficult to place anisotropic particles precisely on solid substrates (Golze et al., 2019; Vinnacombe-Willson et al., 2023).

Light-directed, or plasmon-driven, synthesis offers one way around these constraints. The idea is deceptively simple. When noble-metal seeds are illuminated, plasmon excitation converts photon energy into non-equilibrium hot electrons and hot holes, or into localized heat, and these species drive redox chemistry, regioselective reduction, site-specific etching and symmetry breaking at sub-wavelength length scales (Brongersma et al., 2015; Long et al., 2020; Tan et al., 2017). The first clear demonstration came when R. Jin et al. (2001) watched silver nanospheres turn into triangular nanoprisms under ordinary room light. Since then, light has gradually been reframed: no longer just a probe, it can act as a spatially resolved reagent that sets particle shape, aspect ratio and even 3D handedness (Besteiro et al., 2021; X. Jin et al., 2026).

The field has moved quickly, and not always tidily. Several questions that matter for anyone trying to reproduce or extend this work are still open. It remains surprisingly hard to separate genuinely non-thermal carrier chemistry from local heating, because the two dissipate energy on overlapping timescales (Kamarudheen et al., 2018; Zhan et al., 2019). In colloidal circularly polarized light (CPL)-driven galvanic replacement, chirality appears only when illumination is synchronized with the earliest redox steps, before unpolarized dark reactions take over (Feferman et al., 2026). And background light can itself become a confounding variable: in N-methyl-D-glucamine (NMGA)-mediated nanostar synthesis, it is the dark-grown particles, not the light-grown ones, that show the stronger NIR extinction and higher photothermal efficiency (Samaranayake et al., 2026).

This review attempts to draw these threads together. Our aims are fourfold. First, we examine how hot electrons, hot holes, plasmon-induced charge separation (PICS) and photothermal fields direct nucleation, anisotropic growth, etching and particle coalescence (Besteiro et al., 2021; Dahi et al., 2024; Huang et al., 2022; Tan et al., 2017). Second, we compare light-directed strategies across monometallic, bimetallic and hybrid systems, in solution and on substrates (Feferman et al., 2026; Golze, 2022; Habib & Sadtler, 2020; Langille et al., 2013). Third, we ask how optical parameters (wavelength, intensity, linear or circular polarization, optical vortices) interact with surface chemistry, including surfactants, halides, pH and exposure timing (Feferman et al., 2026; X. Jin et al., 2026; Samaranayake et al., 2026). Finally, we identify the technical bottlenecks and outline directions for chiral sensing, metamaterials and plasmon-enhanced photocatalysis (Dahi et al., 2024; X. Jin et al., 2026).

2. Plasmon-Driven Light-Directed Synthesis of Anisotropic and Chiral Metal Nanostructures

2.1 Historical Evolution: From Seed-Mediated Chemistry to Photoinduced Synthesis

2.1.1 The seed-mediated paradigm and its limits

The rational synthesis of anisotropic metal nanostructures has long been a central concern of nanomaterials research, largely because their optical, catalytic and surface-enhanced spectroscopic properties depend so strongly on geometry (Langille et al., 2013; Zhai et al., 2016). For decades, seed-mediated growth was the method of choice for controlling size, aspect ratio and exposed facets (Xia et al., 2017). In a typical protocol, small isotropic seeds are introduced into a growth solution containing a metal precursor, a mild reductant and a shape-directing surfactant such as cetyltrimethylammonium bromide (CTAB) or polyvinylpyrrolidone (PVP) (Nikoobakht & El-Sayed, 2003; Zhai et al., 2016). The approach gives excellent yields of rods and cubes, but it works within the bounds of thermodynamic control or surfactant-guided kinetic stabilization (Langille et al., 2013). Lower-symmetry shapes, and particles grown site-specifically on solid supports, have proved much harder to obtain by thermal routes alone (Golze et al., 2019; Vinnacombe-Willson et al., 2023).

2.1.2 Light enters the flask

The turning point, as most accounts agree, was the report by R. Jin et al. (2001) that spherical silver particles convert into uniform triangular nanoprisms under room-light illumination. The finding mattered less for the prisms themselves than for what it implied: light could drive an anisotropic transformation rather than merely report on one. Later studies extended photoinduced growth to other metals and shapes, including silver nanorods and cubes with unusual twinning, gold nanoprisms and branched nanostars (Langille et al., 2013; Personick et al., 2013; Zhai et al., 2016). Linearly polarized and phase-structured light was also shown to imprint its own symmetry on photoelectrodeposited films and on laser-reorganized nanoparticle layers (Carim et al., 2016; Li et al., 2021).

2.1.3 From field-shaping to carrier chemistry

Early explanations leaned on the idea that anisotropic near fields simply created favoured growth zones. Current thinking is more specific: the transformations are driven by non-radiative plasmon decay, which produces energetic hot carriers and localized heat (Besteiro et al., 2021; Dahi et al., 2024; Tan et al., 2017). By loosening the grip of crystallographic symmetry and thermal equilibrium, light-directed synthesis offers spatial addressability, spectral selectivity and reconfigurability that conventional chemistry struggles to match (X. Jin et al., 2026; Long et al., 2020). The milestones of this conceptual shift are summarized in Figure 1.

2.2 Microscopic Mechanisms: Hot Charge Carriers versus Photothermal Effects

2.2.1 The plasmon decay cascade

Everything in plasmon-directed synthesis begins with the LSPR (Baffou & Quidant, 2013; Besteiro et al., 2021). After excitation, plasmons decay non-radiatively by Landau damping within about 1–10 fs, generating strongly non-equilibrium electron–hole pairs, the so-called hot carriers (Brongersma et al., 2015; Wang et al., 2018). These relax through carrier–carrier scattering on the 100 fs–1 ps scale and through electron–phonon coupling over roughly 1–10 ps, eventually handing their energy to the surrounding medium as heat (Baffou & Quidant, 2013; Dahi et al., 2024). The full cascade, and the chemistry that branches off at each stage, is sketched in Figure 2. Telling the non-thermal and thermal branches apart is, arguably, still the most stubborn conceptual problem in plasmon-driven chemistry (Kamarudheen et al., 2018; Zhan et al., 2019).

2.2.2 Hot electrons, hot holes and spatial selectivity

Hot electrons with enough energy to cross interfacial Schottky barriers, or to reach unoccupied molecular orbitals, can transfer into solution and reduce adsorbed ions such as AuCl4− or Ag+ (Besteiro et al., 2021; Long et al., 2020). Because the local generation rate scales with the square of the local field, |E(r)|2, reduction concentrates at sharp corners, apexes and gaps (Besteiro et al., 2021; Wang et al., 2018). Hot holes, meanwhile, can oxidize and dissolve surface atoms, so deposition and dissolution compete with one another on the same particle (Dahi et al., 2024; Feferman et al., 2026). The main pathways are compared in Table 2.

2.2.3 Direct spatial evidence

Some of the most persuasive evidence for carrier-driven chemistry comes, perhaps unexpectedly, from polymers. Wang et al. (2018) showed that low-power laser excitation of single gold particles initiates radical polymerization of acrylate monomers along the incident polarization axis, without any molecular photoinitiator. Since the temperature rise was too small to drive thermal polymerization, the site-selective polymer lobes were taken as a map of hot-electron injection that matched the calculated near field.

2.2.4 The non-local character of heat

Heat behaves differently. The high thermal conductivity of noble metals equilibrates temperature across a particle within picoseconds (Baffou & Quidant, 2013; Kamarudheen et al., 2018). Local heating can anneal, melt or strip ligands, but it is poorly suited to producing directional growth; regioselective modification seems to require the directionality of carrier transfer or of near-field gradients (Besteiro et al., 2021; Dahi et al., 2024; X. Jin et al., 2026).

2.3 Non-Classical Growth Modes and Reaction Kinetics

2.3.1 Beyond monomer addition

Classical models assume that anisotropic crystals grow atom by atom or through Ostwald ripening, in which small, high-energy particles dissolve to feed larger ones (Xia et al., 2017). Single-particle studies now suggest that plasmon-driven synthesis frequently follows other routes,

Figure 1. Conceptual evolution of anisotropic metal nanocrystal synthesis, from classical seed-mediated chemistry to light-directed, chirality-encoded growth. The timeline groups milestone studies into three eras. Era I (blue) represents thermodynamically and surfactant-controlled wet chemistry; Era II (orange) begins with the photoinduced conversion of silver nanospheres into nanoprisms and includes the mechanistic recognition of hot-carrier and photothermal contributions; Era III (violet) covers substrate-bound growth, circularly polarized light (CPL)-imprinted chirality and kinetic control of nucleation and galvanic replacement. Citations beneath each milestone identify the primary sources discussed in Section 2.1.

Figure 2. Energy-flow cascade following plasmon excitation and the chemical pathways that branch from it. Photon absorption excites the localized surface plasmon resonance (LSPR), which decays by Landau damping within about 1–10 fs into hot electrons (green) and hot holes (orange). Hot electrons reduce adsorbed precursors at field hotspots, with a generation rate proportional to |E(r)|2, whereas hot holes etch, strip ligands, oxidize Pb2+ or inject into quantum dots. Carriers that are not captured relax by carrier–carrier and electron–phonon scattering into localized heat (yellow), which acts non-selectively. The figure summarizes Section 2.2 and the mechanisms compiled in Table 2.

including hot-hole-regulated coalescence and competing light and dark kinetics (Dahi et al., 2024; Huang et al., 2022; Samaranayake et al., 2026).

2.3.2 Hot-hole-regulated coalescence

Huang et al. (2022) provided a clear example while studying the light-driven conversion of gold nanospheres into triangular nanoplates in the presence of iodide. Combining nanopore sensing with time-resolved electron microscopy, they identified an edge-selective coalescence pathway: hot holes destabilize capping ligands at particle edges, and small spheres attach in an oriented fashion to form planar, twinned plates. Controlled coalescence, in other words, can still deliver crystalline and uniform anisotropic products.

2.3.3 Wavelength as a switch between growth and etching

Dahi et al. (2024) examined the interplay between wavelength and redox chemistry during femtosecond-laser reshaping of gold nanorods. Irradiation at 400 nm, within the interband absorption of gold, generated carriers fairly uniformly and drove overgrowth along both sides and tips. Tuning to 800 nm, in resonance with the longitudinal LSPR, instead concentrated hot holes at the tips, where free AuCl4− etched them and steadily lowered the aspect ratio.

2.3.4 Ambient light as a hidden variable

Illumination during early nucleation can matter just as much. In the seedless, surfactant-free synthesis of gold nanostars with NMGA as a green reductant, room light during mixing accelerated Au3+ reduction, producing many small nuclei that grew into stars with short branches and small cores (Samaranayake et al., 2026). Running the same reaction in darkness suppressed this nucleation burst and allowed the remaining precursor to feed tip elongation, giving larger, sharper stars with markedly better NIR photothermal performance. The lesson, one suspects, extends well beyond this particular system.

2.4 Light-Imprinted Chirality and Structural Symmetry Breaking

2.4.1 Why chirality, and why light

Chiral nanostructures lack mirror symmetry in three dimensions and can show intense circular dichroism (CD) and large dissymmetry factors (g) (Besteiro et al., 2021; X. Jin et al., 2026). Conventional routes depend heavily on chiral templates, such as amino acids, peptides or DNA origami (Kuzyk et al., 2017; Lv et al., 2022). Light-directed synthesis offers a ligand-free alternative in which handedness is written directly onto achiral precursors using CPL or structured vector beams (Besteiro et al., 2021; Feferman et al., 2026; X. Jin et al., 2026).

2.4.2 Physical origin of CPL-imprinted chirality

The mechanism rests on symmetry breaking in the near field (Besteiro et al., 2021). When a low-symmetry achiral particle, such as a cuboid, nanobar or triangular prism, is illuminated with CPL, orthogonal field components interfere to produce a phase-shifted, asymmetric intensity at opposite corners (Besteiro et al., 2021; Feferman et al., 2026). That asymmetry drives localized PICS and carrier transfer, so that deposition or galvanic dissolution reproduces the handedness of the light in the 3D structure (Besteiro et al., 2021; X. Jin et al., 2026). Representative chiral architectures are compiled in Table 3, and the general sequence is illustrated in Figure 3.

2.4.3 A two-step kinetic pathway

Feferman et al. (2026) investigated CPL-controlled galvanic replacement in colloidal Au@Ag nanobars exposed to Au3+, Pt4+ and Pd4+ complexes. Chirality developed in two stages. A fast step, lasting only seconds, combined regioselective hot-electron reduction at specific apexes with hot-hole dissolution of silver from the lateral faces and accounted for roughly half of the final CD. A slower step over several minutes, plasmon-guided redeposition of silver onto the newly asymmetric template, supplied the rest. Timing proved critical: delaying CPL exposure by only a few seconds after precursor addition caused a sharp loss of CD, and the choice of halide counterion and pH strongly modulated both reaction rate and final dissymmetry.

2.4.4 Structured light and direct writing

Uniform CPL is not the only option. Beams carrying spatial polarization gradients or orbital angular momentum (OAM), such as optical vortices, add further spatial control (X. Jin et al., 2026). Kim et al. (2024) used CPL-driven direct writing to print plasmonic silver nanohelicoids, and related approaches have produced spiral metasurfaces and chiral needle arrays over large areas without chiral reagents (X. Jin et al., 2026).

2.5 Substrate-Bound Synthesis versus Solution-Phase Photogrowth

Table 1. Light-directed synthetic methodologies, irradiation parameters and morphological outcomes for representative anisotropic gold nanostructures. The table compares five representative light-directed routes, listing for each the precursor chemistry (including capping, reducing and hole-scavenging agents), the irradiation conditions (source, wavelength, power or colour temperature), the dominant growth or reshaping mode and the resulting morphology and optical signature. Dimensions are reported as in the original studies (mean ± SD where given). The comparison illustrates that wavelength, ambient light and seed immobilization can each change the product as decisively as chemical composition. Abbreviations: AuNS, gold nanostar; NR, nanorod; NMGA, N-methyl-D-glucamine; CTAB, cetyltrimethylammonium bromide; PVP, polyvinylpyrrolidone; DVB, divinylbenzene; MMA, methyl methacrylate; LSPR, localized surface plasmon resonance; NIR, near-infrared.

Table 2. Fundamental physical and chemical mechanisms underlying plasmon-driven transformations of metal nanocrystals. Five mechanisms are distinguished by their excitation and environmental requirements, the energetic species involved, the spatial pattern of the resulting reaction and a representative material system. Hot-electron and hot-hole pathways are spatially selective because carrier generation follows the local near field, whereas photothermal heating is distributed across the particle by rapid thermal conduction. Interband hot-hole injection into quantum dots is included as a non-classical, non-synthetic counterpart that illustrates the same carrier physics. Abbreviations: QD, quantum dot; PEL, plasmon-enhanced photoluminescence; LSPR, localized surface plasmon resonance.

2.5.1 The problem of tumbling

A basic distinction runs between colloidal photogrowth and substrate-bound, on-chip fabrication (Besteiro et al., 2021; Golze et al., 2019). In suspension, particles rotate randomly, which averages out the directional information carried by linear polarization or CPL; structural anisotropy is preserved only when local reaction rates outpace rotational relaxation (Besteiro et al., 2021; Feferman et al., 2026). Figure 4 contrasts the two platforms.

2.5.2 Orientation-locked seeds

Immobilizing seeds on a substrate fixes their orientation relative to the light field and breaks planar symmetry (Golze et al., 2019; Neal et al., 2022). Heteroepitaxial strain, planar defects and dielectric boundaries then stabilize two-dimensional or vertically oriented growth (Golze, 2022; Golze et al., 2019). Golze et al. (2019) used seeds lined with planar stacking faults to grow periodic arrays of gold nanoplates on sapphire from aqueous HAuCl4, PVP and methanol over square-centimetre areas. Adding a photoactive carbon nitride ligand introduced bending strain during growth, lifting one edge of each plate off the support and producing spiral nanohelicoid arrays (Golze, 2022; Golze et al., 2021). Neal et al. (2022) similarly reported large-area arrays of atomically flat single-crystal gold nanotriangles formed directly on substrates.

2.5.3 Flow, scale and integration

Vinnacombe-Willson et al. (2023) placed substrate-bound seeds inside microfluidic reactors, continuously refreshing precursors to control aspect ratio and facet development while suppressing homogeneous nucleation. Substrate-bound photogrowth thus links colloidal chemistry with device integration and points toward lithography-free optical chips, chiral sensors and photocatalysts (Golze, 2022; X. Jin et al., 2026; Neal et al., 2022). Hybrid compositions extend the palette further; plasmon-mediated growth of silver–platinum structures, for example, couples optical activity with catalytic function (Habib & Sadtler, 2020).

2.6 Synthesis of the Literature and Remaining Gaps

2.6.1 What the literature agrees on

Read together, these studies converge on a few points. Light acts through carriers whose spatial distribution follows the near field; wavelength determines whether those carriers are spread evenly or concentrated; and kinetics, not thermodynamics, increasingly decide the outcome (Besteiro et al., 2021; Dahi et al., 2024; Tan et al., 2017). The synthetic methods and their outcomes are collected in Table 1.

2.6.2 Where it remains unsettled

Less agreement exists on how to quantify the relative weight of thermal and non-thermal contributions in any given synthesis (Kamarudheen et al., 2018; Zhan et al., 2019), on how general the light–dark kinetic effects are (Samaranayake et al., 2026), and on how far colloidal chirality can be pushed before rotational averaging wins (Feferman et al., 2026). These gaps shaped the questions addressed in the remainder of this review.

3. Methods

3.1 Review Design and Reporting Framework

3.1.1 Design

This work was conducted as a structured narrative and integrative review. A fully quantitative meta-analysis was not attempted, because the included studies differ substantially in material system, light source and outcome metric. To keep the process transparent and repeatable, we nonetheless followed the staged logic of systematic reviews (identification, screening, eligibility assessment, data extraction, appraisal and synthesis), as summarized in Figure 5.

3.1.2 Review questions

Four questions, aligned with the objectives stated in Section 1, guided the search and extraction. RQ1: Which photophysical and photochemical mechanisms (hot electrons, hot holes, PICS, photothermal heating) direct nucleation, anisotropic growth, etching and coalescence? RQ2: How do light-directed strategies compare across monometallic, bimetallic and hybrid systems, in solution and on substrates? RQ3: How do optical parameters interact with chemical parameters to set facet selectivity, aspect ratio and chirality? RQ4: What technical bottlenecks limit reproducibility and application?

3.2 Information Sources and Search Strategy

3.2.1 Databases and time window

Web of Science Core Collection, Scopus, PubMed and Google Scholar were searched for records published from January 2001, the year of the first photoinduced nanoprism report (R. Jin et al., 2001), to March 2026. Searches were limited to English-language records.

3.2.2 Search string

The core Boolean string, adapted to the syntax of each database, was: ("plasmon-mediated" OR "plasmon-driven" OR "light-directed" OR "photoinduced" OR "photochemical") AND ("synthesis" OR "growth" OR "etching" OR "reshaping" OR "galvanic replacement") AND ("gold" OR "silver" OR "Au" OR "Ag" OR "bimetallic" OR "core–shell") AND ("nanorod" OR "nanoprism" OR "nanoplate" OR "nanostar" OR "anisotropic" OR "chiral" OR "nanohelicoid"). Supplementary strings combined "hot carrier" OR "hot electron" OR "hot hole*" with "photothermal" to capture mechanistic studies, and "circularly polarized light" OR "optical vortex" with "chirality" to capture symmetry-breaking work.

3.2.3 Citation chasing

Backward and forward citation searching was seeded from three anchor reviews (Langille et al., 2013; Tan et al., 2017; X. Jin et al., 2026) and from the reference lists of all included primary studies.

3.3 Eligibility Criteria

3.3.1 Inclusion criteria

Records were eligible if they (a) reported the synthesis, growth, etching or reshaping of metal or metal-containing nanostructures in which light excitation was a controlled variable; (b) provided mechanistic, theoretical or quantitative evidence relevant to hot-carrier, photothermal or polarization-dependent pathways; (c) described the optical conditions (at minimum the wavelength or spectral range) and the chemical system; or (d) were authoritative reviews or theses that framed the mechanisms or methods under discussion. Foundational works on conventional seed-mediated growth and on chiral templating were included as comparators (Kuzyk et al., 2017; Lv et al., 2022; Nikoobakht & El-Sayed, 2003; Xia et al., 2017).

3.3.2 Exclusion criteria

We excluded studies in which light served only for characterization; purely photothermal therapy studies without a synthetic component; conference abstracts without retrievable data; and reports lacking sufficient methodological detail to identify the light source and precursor chemistry.

3.4 Study Selection

Titles and abstracts were screened against the eligibility criteria, followed by full-text assessment. Disagreements were resolved by discussion. The final evidence base comprised 31 sources: primary experimental studies, theoretical and computational studies, review articles, one manuscript under review (Trotsiuk et al., 2026) and one doctoral dissertation (Golze, 2022).

3.5 Data Items and Extraction

3.5.1 Extraction form

A standardized extraction form was piloted on five studies and then applied to all included primary reports. Extracted fields were: nanostructure type and composition; seed type and whether seeds were colloidal or substrate-bound; precursor identity and concentration; capping, reducing and hole-scavenging agents; halide ions and pH; light source (LED, continuous-wave or pulsed laser, broadband lamp); wavelength or spectral range; power or power density; polarization state (unpolarized, linear, left- or right-circular, vortex); timing of illumination relative to precursor addition; dominant growth or reshaping mode; morphological outcome (dimensions reported as mean ± standard deviation where available); optical signature (LSPR position, extinction, CD ellipticity, g-factor, enantiomeric excess); and functional metrics (photothermal conversion efficiency, photoluminescence enhancement).

3.5.2 Harmonization of units

Values were recorded as reported. Wavelengths were standardized to nanometres, power densities to mW cm−2 or W cm−2, and times to seconds or minutes. Dissymmetry factors were treated as dimensionless. No values were imputed or recalculated, apart from simple ratios explicitly stated in the source text.

3.6 Appraisal of Mechanistic Evidence

Because many conclusions in this field rest on mechanistic inference, each primary study was appraised against four questions. First, did the study include a control capable of separating hot-carrier from photothermal contributions, for instance a temperature-matched dark control, a wavelength series or an intensity dependence (Kamarudheen et al., 2018; Zhan et al., 2019)? Second, was the spatial pattern of reaction correlated with a calculated or measured field distribution (Besteiro et al., 2021; Wang et al., 2018)? Third, were outcomes quantified with statistics rather than described qualitatively? Fourth, were optical and chemical conditions reported in sufficient detail for replication? These appraisals informed the weight given to each study in the synthesis but were not used to exclude studies.

3.7 Synthesis Strategy

Extracted data were grouped thematically into four domains corresponding to Tables 1–4: synthetic methodologies and morphological outcomes (Table 1); fundamental mechanisms (Table 2); chiral synthesis with CPL and vector fields (Table 3); and functional performance (Table 4). Within each domain, findings were compared narratively, with attention to agreement, contradiction and the conditions under which each result was obtained. Conceptual flow diagrams (Figures 1–4) were constructed to summarize the literature review, and Figure 6 was drawn from numerical values reported by Samaranayake et al. (2026) and the qualitative outcomes reported by Dahi et al. (2024). Figures were produced in Python (Matplotlib); quantitative panels reproduce published means and standard deviations without re-analysis.

3.8 Minimum Reporting Items Used to Judge Reproducibility

To make the appraisal criteria explicit, and to offer a template for future primary studies, we defined a minimum set of reporting items for light-directed syntheses: (i) light source type and model; (ii) emission spectrum or central wavelength and bandwidth; (iii) power density at the sample, and how it was measured; (iv) polarization state and how it was verified; (v) beam geometry, spot size and vessel material; (vi) illumination onset and duration relative to precursor addition; (vii) solution temperature, monitored during irradiation; (viii) control of ambient light during preparation; (ix) concentrations of precursor, reductant, surfactant, halide and hole scavenger, together with pH; (x) seed preparation and, for substrate-bound systems, substrate identity and seed density; and (xi) the characterization methods and number of particles counted. Studies reporting at least eight of these eleven items were considered highly reproducible for the purposes of this review.

4. Synthesis of Findings: Light as a Kinetic, Spatial and Chiral Handle in Nanocrystal Growth

4.1 Overview of the Evidence Base

4.1.1 What the studies have in common

Across colloidal, substrate-supported and hybrid systems, one theme recurs with some insistence: light does more in nanocrystal growth than the older wet-chemical picture allowed. It behaves less like an external stimulus and more like a handle on both kinetics and spatial selectivity. Absorbed photons are converted into hot electrons, hot holes or heat, and these, in turn, appear to govern interfacial redox chemistry, regioselective deposition, local etching and symmetry breaking at the nanometre scale (Besteiro et al., 2021; Tan et al., 2017).

4.1.2 How the findings are organized

The synthetic conditions and morphological outcomes are compiled in Table 1, the underlying mechanisms in Table 2, the chiral architectures in Table 3 and the functional metrics in Table 4. The sections below follow that order, though the categories overlap more than a tidy table might suggest.

4.2 Light versus Dark: Illumination as a Gate on Nucleation

4.2.1 Ambient light accelerates nucleation

Perhaps the most counter-intuitive result in the corpus is that, in at least one system, less light gives better particles. In seedless, surfactant-free nanostar growth mediated by NMGA, ambient fluorescent or LED illumination during mixing accelerated the reduction of HAuCl4 and produced a burst of small nuclei (Samaranayake et al., 2026). The precursor was consumed early. Little remained to build branches, and the resulting stars had cores of 61.0 ± 11.6 nm and blunted tips of only 23.2 ± 5.6 nm (Figure 6a; Table 1).

4.2.2 Darkness favours elongation

When light was excluded, nucleation was held back, and the unreacted gold fed sustained, directional tip growth. Dark-grown stars were larger, with 94.1 ± 17.4 nm cores and 54.2 ± 10.9 nm tips, and their extinction broadened into the NIR near 900 nm (Samaranayake et al., 2026). Whether this light-accelerated nucleation is general, or a peculiarity of NMGA chemistry, is not yet clear. It seems prudent, at any rate, to treat ambient light as a variable to be reported rather than ignored.

4.3 Wavelength Selects between Growth and Etching

Table 3. Chiral nanostructures synthesized with circularly polarized light (CPL) and structured vector light fields. The table lists five chiral architectures obtained from achiral seeds or precursors, together with the seed platform (colloidal or substrate-bound), the polarization state and wavelength of the light, the proposed symmetry-breaking mechanism and the reported chiroptical response. Values for circular dichroism (CD), dissymmetry factor (g), luminescence dissymmetry (glum) and enantiomeric excess are given as reported. Comparison across rows shows that orientation-locked substrates and structured beams generally yield stronger and more controllable chirality than freely rotating colloids. Abbreviations: GRR, galvanic replacement reaction; PICS, plasmon-induced charge separation; OAM, orbital angular momentum; DLW, direct laser writing.

Table 4. Functional performance of light-engineered plasmonic nanostructures: classical versus non-classical enhancement pathways. Four systems are compared by emitter or functional component, excitation wavelength and resonance condition, the measured performance outcome and the dominant mechanism. The table contrasts classical electromagnetic (Purcell or near-field) enhancement with non-classical pathways such as hot-hole charge rebalancing and kinetically tuned morphology. Performance metrics are reported as in the original studies and are not directly comparable across rows because they measure different properties. Abbreviations: QD, quantum dot; PL, photoluminescence; AuNS, gold nanostar; η, photothermal conversion efficiency.

4.3.1 Interband excitation: isotropic overgrowth

Photon energy can do more than speed up a reaction; it can reverse its sign. When CTAB-capped gold nanorods were irradiated with femtosecond pulses in the presence of free AuCl4−, excitation at 400 nm engaged interband d→sp transitions and produced carriers distributed fairly evenly across the rod (Dahi et al., 2024). These reduced nearby AuCl4−, and the rods grew in both length and width (Figure 6b; Table 1).

4.3.2 Resonant excitation: tip etching

At 800 nm, matched to the longitudinal LSPR, the picture changed. Field enhancement, and with it two-photon absorption, concentrated at the tips; hot holes accumulated there and drove oxidative dissolution by AuCl4− (Dahi et al., 2024). The tips were etched back, the aspect ratio fell and the LSPR shifted continuously to the blue by more than 130 nm (Table 2; Table 4). The same particle and the same precursor gave opposite outcomes, depending only on wavelength.

4.4 Non-Classical Assembly and Substrate-Guided Growth

4.4.1 Coalescence instead of monomer addition

Not all light-directed growth proceeds atom by atom. Colloidal gold spheres can become single-crystalline triangular plates through iodide-regulated, hot-hole-driven coalescence (Huang et al., 2022). Under 500 nm illumination, hot holes appear to strip ligands along particle perimeters and twin boundaries, encouraging oriented attachment and fusion into flat plates (Table 1; Table 2).

4.4.2 Substrates stabilize two-dimensional growth

When seeds are fixed on solid supports, planar boundaries and heteroepitaxial strain stabilize the same two-dimensional mode, giving periodic arrays of single-crystal plates and, with suitable ligands, 3D spiral nanohelicoids without homogeneous precipitation in solution (Golze, 2022; Golze et al., 2019; Golze et al., 2021). Comparable large-area arrays of atomically flat triangles (Neal et al., 2022) and flow-controlled growth from substrate-bound seeds (Vinnacombe-Willson et al., 2023) suggest that the approach is reasonably transferable between laboratories (Figure 4).

4.5 Separating Hot Carriers from Localized Heat

4.5.1 Polymer maps of hot-electron injection

Whether a given transformation is carrier-driven or heat-driven is, in many papers, asserted rather than demonstrated. A few studies, however, offer cleaner tests. Plasmon-directed polymerization of divinylbenzene, styrene and methyl methacrylate on single gold particles under 0.5–2.0 mW excitation at 640 nm produced polymer protrusions aligned with the incident polarization, in agreement with finite-difference time-domain maps of |E|2 (Wang et al., 2018). The calculated temperature rise, below about 5 K, was too small to initiate thermal polymerization, which strongly implies a non-thermal, hot-electron mechanism (Table 2).

4.5.2 Quantitative separation

Kamarudheen et al. (2018) and Zhan et al. (2019) took a more quantitative route, designing experiments to apportion rate enhancements between photothermal and carrier contributions. Their results suggest that both channels often operate together and that their balance depends on intensity, particle geometry and the thermal environment, which is a useful caution against single-mechanism explanations (Figure 2).

4.5.3 Hole injection across a metal–emitter interface

Hybrid metal–semiconductor systems add a further twist. When cysteamine-capped CdSe/ZnS quantum dots (QDs) were electrostatically bound to gold spheres or rods at gaps below 1.5 nm, non-resonant excitation at 400 nm produced an approximately four-fold increase in photoluminescence and longer excited-state lifetimes (Trotsiuk et al., 2026). Classical theory would predict quenching at such distances. Because 400 nm lies far from the LSPR, Purcell enhancement can be largely ruled out; calculations instead pointed to d-band hot holes crossing the interface and neutralizing trapped charges, returning "off"-state QDs to an emissive "on" state (Trotsiuk et al., 2026). This result is still under peer review and should be read with that in mind.

4.6 Emergence of Chiroptical Activity under CPL and Vector Fields

4.6.1 Mechanistic basis

Using CPL to induce 3D chirality in achiral seeds offers a ligand-free route to chiroptical materials. The mechanism relies on PICS and on the asymmetric near fields created when orthogonal polarization components interfere at seed boundaries (Besteiro et al., 2021; Feferman et al., 2026; X. Jin et al., 2026).

4.6.2 Fast and slow stages of chiral galvanic replacement

In colloidal Au@Ag nanobars undergoing galvanic replacement with Au3+, Pt4+ or Pd4+ under 565 nm CPL, optical activity emerged in two steps (Feferman et al., 2026). During the first 0–5 s, CPL excited overlapping longitudinal and transverse modes with a 90° phase shift; gold deposited at particular corners while silver dissolved from lateral faces, generating about half of the final CD. Between roughly 5 s and 5 min, plasmon-guided redeposition of dissolved silver onto the asymmetric template supplied the remaining half, reaching g ≈ 4.3 × 10−3 (Figure 3; Table 3).

4.6.3 A race against the dark reaction

The kinetics impose a strict rule: the light-driven rate must clearly outpace the unpolarized dark reaction. Delaying CPL by just 20 s after precursor addition cut CD intensity by more than half (Feferman et al., 2026). Replacing chloride with bromide or iodide raised the peak ellipticity to 93 mdeg (g ≈ 5.0 × 10−3), presumably by facilitating interfacial electron transfer (Table 3).

4.6.4 Substrates and structured beams amplify chirality

Removing rotational tumbling helps considerably. Achiral gold seeds fixed on TiO2/ITO and subjected to CPL-driven oxidation of Pb2+ to PbO2 reached enantiomeric excesses of up to 43%, with clear CD sign reversal on switching handedness (X. Jin et al., 2026). Direct writing with focused CPL produced helical silver nanostructures (Kim et al., 2024), and vortex beams carrying OAM yielded spiral relief patterns with luminescence dissymmetry factors (glum) of up to 0.5 over macroscopic areas (X. Jin et al., 2026). Theory had anticipated this trend, predicting substantially larger g-factors for substrate-fixed than for freely rotating particles (Besteiro et al., 2021).

4.7 Functional Performance

4.7.1 Photothermal conversion

Structural control translates, with some caveats, into functional gains (Table 4). Under 808 nm irradiation at 1.7 W cm−2, dark-grown NMGA nanostars consistently outperformed light-grown ones, reaching photothermal conversion efficiencies (η) of about 34% at [Au]f = 0.50 mM and RNMGA = 14, an average improvement of roughly 45% (Samaranayake et al., 2026). The longer tips concentrate the local field and raise extinction at the laser wavelength.

4.7.2 Emission recovery and spectral tuning

In CdSe/ZnS QD–gold complexes, interband hole injection recovered photoluminescence by up to 400% without spectral matching or Purcell enhancement, which may prove useful for charge-balanced emitters and biosensors (Trotsiuk et al., 2026). Femtosecond reshaping of nanorods allowed continuous post-synthetic LSPR tuning over more than 130 nm, potentially valuable for optical filters, plasmonic colour printing and tailored photocatalysts (Dahi et al., 2024). Substrate-bound nanohelicoid arrays, finally, offer dense chiral metamaterials for enantioselective sensing and circular-polarization photodetection (Golze, 2022; X. Jin et al., 2026).

5. Discussion: Toward Predictive, Photon-Programmed Nanocrystal Synthesis

5.1 Principal Findings in Context

5.1.1 A shift in what controls shape

Taken as a whole, the evidence reviewed here suggests a quiet but real shift in how nanocrystal shape is controlled. In classical seed-mediated chemistry, morphology follows from facet energies and ligand binding (Nikoobakht & El-Sayed, 2003; Xia et al., 2017). In the studies summarized in Table 1, the decisive variables are increasingly optical and kinetic: photon energy, polarization and the timing of illumination. This does not make the older picture wrong. It does, however, make it incomplete.

5.1.2 Three recurring principles

Three principles seem to hold across systems. Carrier generation follows the near field, so geometry and wavelength jointly decide where chemistry happens (Besteiro et al., 2021; Wang et al., 2018). Electrons tend to build and holes tend to remove, and the balance between them can be tilted by wavelength or by scavengers (Dahi et al., 2024; Huang et al., 2022). And kinetics frequently matter more than equilibrium, whether in nucleation (Samaranayake et al., 2026) or in chirality transfer (Feferman et al., 2026). Figure 2 and Table 2 summarize how these strands connect.

5.2 Hot Carriers versus Heat: Unresolved, but Tractable

5.2.1 Why the question persists

The persistence of this debate is, in part, a measurement problem. Carrier relaxation and heating occur within

Figure 3. Mechanism of circularly polarized light (CPL)-imprinted chirality in achiral plasmonic seeds. A low-symmetry achiral seed, such as an Au@Ag nanobar, is exposed to left- or right-handed CPL (or an optical vortex carrying orbital angular momentum). Interference of longitudinal and transverse plasmon modes with a 90° phase shift produces an asymmetric near field at handedness-specific corners. A fast step (≈0–5 s) of regioselective reduction and silver dissolution generates about half of the final circular dichroism (CD), and a slower redeposition step (≈5 s–5 min) supplies the rest. Kinetic constraints and halide effects are noted at the bottom; data are from Feferman et al. (2026).

Figure 4. Comparison of colloidal and substrate-bound platforms for plasmon-driven photogrowth. In colloidal suspension (left, blue), random Brownian rotation averages the directional information carried by linear or circular polarization, so anisotropy and chirality survive only when redox kinetics outpace rotational diffusion. Substrate-bound seeds (right, violet) are fixed relative to the optical field; planar symmetry breaking, planar defects, heteroepitaxy and dielectric boundaries stabilize two-dimensional or vertical growth. This yields periodic nanoplate arrays, nanohelicoids, larger dissymmetry factors and enantiomeric excesses of up to 43%. The figure summarizes Section 2.5.

picoseconds of each other, and most syntheses are monitored over minutes, so the two contributions are inevitably convolved (Baffou & Quidant, 2013; Kamarudheen et al., 2018). It would be surprising if one mechanism dominated everywhere.

5.2.2 What the stronger studies show

The more convincing studies share a design feature: they include an internal control that heat alone cannot explain. Polarization-aligned polymer growth under negligible heating (Wang et al., 2018), wavelength-switched growth and etching on the same rods (Dahi et al., 2024) and emission recovery under non-resonant excitation (Trotsiuk et al., 2026) are all difficult to reconcile with a purely thermal account. Systematic intensity- and temperature-dependent protocols (Kamarudheen et al., 2018; Zhan et al., 2019) offer a template that, in our view, deserves to become standard.

5.3 Kinetics as the Hidden Variable

5.3.1 Light can harm as well as help

The light-versus-dark nanostar result (Figure 6a) is instructive precisely because it runs against intuition. Light is often assumed to be beneficial in plasmon-mediated synthesis; here, uncontrolled ambient light degraded the product (Samaranayake et al., 2026). A plausible reading is that photons accelerated reduction indiscriminately, before any anisotropic template existed, so the additional nucleation simply diluted the precursor. If so, the effect may be widespread but unreported, hidden in the batch-to-batch variability that many groups quietly accept.

5.3.2 Synchronization in chiral synthesis

CPL-driven galvanic replacement poses the same issue in sharper form. Chirality is encoded only while the light-driven rate exceeds the dark rate, and the window closes within seconds (Feferman et al., 2026). The strong dependence on halide and pH suggests that chemists have more levers here than is sometimes assumed; tuning the dark reaction down may be as effective as tuning the light reaction up (Table 3).

5.4 Platform Choice: Colloids versus Substrates

5.4.1 Trade-offs

Figure 4 captures what is, essentially, a trade-off. Colloids are scalable and solution-processable but lose directional information through rotational averaging (Besteiro et al., 2021; Feferman et al., 2026). Substrates lock orientation, stabilize two-dimensional growth and deliver much larger chiroptical responses, but at the cost of throughput and with the substrate itself becoming part of the chemistry (Golze et al., 2019; X. Jin et al., 2026; Neal et al., 2022).

5.4.2 Toward hybrid strategies

Microfluidic delivery to substrate-bound seeds (Vinnacombe-Willson et al., 2023) hints at a middle path, retaining orientation control while improving precursor supply and reproducibility. Whether such platforms can be scaled to areas relevant for devices remains, for now, an open question.

5.5 Translational Implications

5.5.1 Biomedical and catalytic applications

The functional data in Table 4 are encouraging, if still fragmentary. A 45% gain in photothermal efficiency from a change in ambient lighting is cheap to implement and may be relevant to photothermal therapy (Samaranayake et al., 2026). Light-guided growth of bimetallic Ag–Pt and related hybrids couples plasmonic absorption with catalytic surfaces, a combination that seems well suited to plasmon-enhanced catalysis (Habib & Sadtler, 2020; Langille et al., 2013).

5.5.2 Photonic and sensing applications

On the photonic side, wavelength-programmable reshaping (Dahi et al., 2024), polarization-sensitive laser-written metasurfaces (Li et al., 2021) and printable chiral nanohelicoids (Kim et al., 2024) all point toward lithography-free fabrication of optical components. Chiral nanostructures produced without templates could reduce reliance on biological scaffolds (Kuzyk et al., 2017; Lv et al., 2022), although their dissymmetry factors, at 10−3 in colloids, remain modest compared with the best templated assemblies.

5.6 Limitations

5.6.1 Limitations of the evidence

Several limitations should temper these conclusions. The primary literature is heterogeneous in light sources, intensities and outcome metrics, which makes quantitative comparison hazardous. Many key results come from single laboratories and have not yet been independently replicated, and one of the studies discussed is still under peer review (Trotsiuk et al., 2026). Reporting of power

Figure 5. Workflow used to identify, screen, extract, appraise and synthesize the literature included in this review. The six stages proceed from definition of the review questions (1) through database searching and citation chasing (2), eligibility screening against predefined criteria (3), standardized data extraction of optical, chemical and outcome variables (4), appraisal of mechanistic evidence and reporting completeness (5), and thematic synthesis into Tables 1–4 and Figures 1–4 and 6 (6). Blue, orange and violet boxes denote identification, selection/extraction and analysis phases, respectively. The final evidence base comprised 31 sources published between 2001 and 2026.

density, polarization purity and ambient light is often incomplete, which limited our appraisal.

5.6.2 Limitations of this review

The review was restricted to English-language sources and to the evidence base described in Section 3, and a meta-analysis was not feasible. Publication bias, favouring striking morphologies over failed or ambiguous attempts, is likely but could not be assessed formally.

5.7 Future Directions

5.7.1 Mechanistic and kinetic tools

Progress will probably depend on better tools for watching growth as it happens. Single-particle methods such as nanopore sensing (Huang et al., 2022), combined with in situ spectroscopy and temperature-resolved controls (Zhan et al., 2019), could make mechanistic claims more falsifiable. Kinetic models that treat light-driven and dark rates explicitly would help predict when chirality or anisotropy will survive (Feferman et al., 2026).

5.7.2 Standards and integration

The field would also benefit from shared reporting standards along the lines of the checklist in Section 3.8. On the application side, combining substrate-bound growth with structured light (X. Jin et al., 2026; Kim et al., 2024) and extending light-directed growth to semiconductor and oxide hybrids (Tan et al., 2017; Trotsiuk et al., 2026) seem among the most promising routes toward functional chiral sensors, metamaterials and photocatalysts.

6. Conclusion

Light, it now seems clear, can be more than a probe in nanocrystal synthesis; it can serve as a programmable reagent. The studies reviewed here show that hot electrons build material at field hotspots, hot holes etch, strip ligands or drive coalescence, and heat acts broadly and without much spatial selectivity. Wavelength can switch gold nanorods from growth to etching. Ambient light can tip nucleation kinetics enough to change photothermal efficiency by about 45%. Circularly polarized light can write handedness onto achiral seeds, provided the light-driven chemistry outruns the dark reaction, and substrates amplify these effects by removing rotational averaging. Important gaps remain, especially in separating carrier from thermal contributions, synchronizing kinetics and reporting conditions consistently. Addressing them through in situ measurements, explicit kinetic models and shared reporting standards should help move light-directed growth from striking demonstrations toward predictable, scalable fabrication of chiral and anisotropic nanomaterials for sensing, catalysis and photonics.

References