Direct-write 3D printing of plasmonic nanohelicoids by circularly polarized light

Significance

Chiroplasmonic nanostructures show giant polarization rotation that can advance biosensing, catalysis, photonics, and information technologies. However, complex fabrication processes and poor scalability of three-dimensional (3D) lithography hinder their fundamental and translational studies. Light-to-mater chirality transfer enables simple and versatile direct-write 3D printing of chiroplasmonic nanostructures on various substrates. Macroscale patterns with a wide range of chiroptical properties can be continuously printed at a rate of several square centimeters per minute. The single-step, mask-free, direct-write printing from aqueous solutions of silver salt allows for rapid production of chiroplasmonic devices for biosensing and optoelectronics.

Abstract

Chiral plasmonic surfaces with 3D “forests” from nanohelicoids should provide strong optical rotation due to alignment of helical axis with propagation vector of photons. However, such three-dimensional nanostructures also demand multi-step nanofabrication, which is incompatible with many substrates. Large-scale photonic patterns on polymeric and flexible substrates remain unattainable. Here, we demonstrate the substrate-tolerant direct-write printing and patterning of silver nanohelicoids with out-of-plane 3D orientation using circularly polarized light. Centimeter-scale chiral plasmonic surfaces can be produced within minutes using inexpensive medium-power lasers. The growth of nanohelicoids is driven by the symmetry-broken site-selective deposition and self-assembly of the silver nanoparticles (NPs). The ellipticity and wavelength of the incident photons control the local handedness and size of the printed nanohelicoids, which enables on-the-fly modulation of nanohelicoid chirality during direct writing and simple pathways to complex multifunctional metasurfaces. Processing simplicity, high polarization rotation, and fine spatial resolution of the light-driven printing of stand-up helicoids provide a rapid pathway to chiral plasmonic surfaces, accelerating the development of chiral photonics for health and information technologies.

The strong polarization rotation of transmitted and reflected light by chiral inorganic nanostructures with subwavelength dimensions (1–5) catalyzed their rapid development in the last decade (6–8). The plasmonic surfaces with chirality on the nano- and micrometer scale show particularly high sensitivity to changes in their dielectric environment making them promising materials platforms for biosensing, catalysis, photonics, and information technologies (9–14). While their synthesis in solutions is simple and versatile, the translation of these processes to surfaces is not. Currently, subwavelength metal geometries on substrates with three-dimensional (3D) chirality are created primarily by two-photon 3D lithography followed by electroplating of metals and etching by plasma (10, 11, 15–17). Other patterning methods include ion/electron beam induced deposition (18) and glancing angle deposition (19–23). Transfer of pre-synthesized 3D plasmonic elements on substrates had been also recently suggested as an alternative method (13, 14). All of these methods are accurate but complicated and expensive; they also involve time-consuming and multi-step processing at low pressure and/or high temperature condition. The specialized hardware for these techniques is rare even for high-end facilities, which hinders access to chiral nanostructured surfaces for researchers in multiple scientific disciplines and geographic locations.

Direct creation of subwavelength helical metal geometries at the interfaces through spontaneous self-assembly of nanoparticles (NPs) is the holy grail of 3D printing and chiral plasmonic surface production. However, realization of this concept encounters several fundamental barriers especially for devices with strong polarization rotation. First, as shown in the typical design of lithographic 3D chiral metasurfaces, the rotational axis of the self-assembled geometries must be aligned to the surface normal vector to maximize their optical rotatory power for photon propagating through the substrate (24). The spontaneous out-of-plane alignment is, however, difficult even for the closely packed interfacial self-assembly of chiral molecules (7, 25–30), producing frequent orientational defects (29, 30). It becomes nearly impossible for the assemblies of NPs because such alignment is associated with order of magnitude higher thermodynamic penalties. The stand-up placement of helical geometries on surfaces (like trees in the forests) inevitably leads to high potential energy and mechanical instability. Unlike closely packed monolayers from chiral molecules (7, 25–30), the empty space between plasmonic elements is a critical requirement for strong polarization effects from in-gap plasmonic resonances. However, the empty space between the nanostructures in self-assembly process is associated with the loss of close-range van der Waals interactions, which make such complex geometries even more thermodynamically unfavorable.

Photon-to-matter chirality transfer offers both simplicity and universality to the chiral synthesis (31, 32) in the course of 3D printing, suggesting a promising research direction to overcome these fundamental constraints. In this study, we show that the direct printing of 3D chiral metallic nanostructures is possible by illumination of different substrates with circularly polarized light (CPL). It enables the fabrication of centimeter-scale chiral plasmonic surfaces, surpassing spatial and geometrical limitations, all in a single-step, mask-free, and additive manufacturing process.

Results

The Light-driven Formation of Silver Nanohelicoids.

The “forests” of conical metallic spirals with nanoscale diameter (nanohelicoids) quickly emerge when solid substrates are illuminated with left- and right-handed circularly polarized light (L-CPL and R-CPL, respectively). Surfaces coated with nanohelicoids showed distinctive handedness with enantiomeric excess above 80% benefiting from efficient directional assembly starting from the fixed nucleation spot at the liquid–solid interface. The normal incidence of CPL on various solid substrates, for instance, glass, indium titanium oxide (ITO), and polydimethylsiloxane (PDMS), immersed in an aqueous mixture solution of sodium citrate (2.5 mM to 12.5 mM) and silver nitrate (2.5 mM), results in heterogeneous silver nucleation at the liquid–solid interface and the subsequent growth into nanohelicoids in the direction normal to the substrate. See the methods for details.

The time required for nanohelicoids to form is dependent on light intensity and growth conditions. In most cases, we used focused laser beams to expedite the direct-write printing. In some cases, the circularly polarized laser beams were broadened to elucidate growth mechanisms (Fig. 1A and SI Appendix, Fig. S1A). For a broadened 532 nm laser beam with a photon flux of 3.07 × 10²¹ photons s⁻¹ m⁻², the nanohelicoid arrays covering an area of ca 3.14 cm² were produced within 30 min, displaying angle-dependent reflective color typical for plasmonic metasurfaces (Fig. 1B and SI Appendix, Fig. S1B). By changing the ellipticity of polarized light (i.e., by changing angle of quarter-wave plate against the linear polarizer) from 5 to 45 degree, the resulting nanostructures showed different circular dichroism (CD) and optical dissymmetry g-factor (Fig. 1 C and D). The L-CPL and R-CPL illumination of solid substrates results in the growth of right- and left-handed nanohelicoids denoted as Δ- and Λ-enantiomers, respectively (Fig. 1 E and F). In contrast, SEM images of silver particles formed under linearly polarized light (LP) provides clear evidence that CPL is essential for the formation of nanohelicoids. Even with an extended exposure time of 60 min under LP, we could not find any particles or their assemblies comparable to the size or morphology to nanohelicoids formed under CPL (SI Appendix, Fig. S2). The Δ- and Λ-nanohelicoid stand-up arrays displayed strong mirror-symmetrical CD spectra with bands at 350, 430, 545, and 780 nm, whereas the silver particles formed under LP showed no chiroptical activities (Fig. 1G). Besides glass, similar arrays with identical polarity were grown on ITO, silicon wafers and PDMS (SI Appendix, Figs. S1 and S3).

Fig. 1.

CPL–driven synthesis of silver nanohelicoids oriented along the normal direction to the substrate surface. (A) Schematic of the experimental design for the CPL-printing of the silver nanohelicoid forests. (B) Photograph of silver nanohelicoids metasurfaces printed on glass substrate. (C) CD and (D) g-factor of silver nanohelicoids synthesized by different ellipticity (changing angle between linear polarized light and quarter-wave plates from 5 to 45 degree) of the polarized light. (E and F) Scanning electron microscope (SEM) image of Δ-nanohelicoids and Λ-nanohelicoid formed on ITO/glass by L-CPL and R-CPL, respectively. (G) CD spectra of Δ- nanohelicoids Λ-nanohelicoids, and linear NPs formed on glass substrates after 30-min exposure to a 532 nm L-CPL, R-CPL, and linearly polarized light respectively. (H–J) Experimental observation of different growth stages of Δ-nanohelicoids. (H) SEM images of Δ-nanohelicoids at different illumination times up to 30 min. The scale bar is 50 nm. (I) STEM images of a Δ-nanohelicoid after 10 min illumination. Secondary electron (SE), high-angle annular dark-field (HAADF), and bright field (BF) detectors are simultaneously used to analyze the internal structure of the rotational assembly. Note that the structure is transferred from the original substrate to the transmission electron microscopy (TEM) grid and thus flipped over on the grid. The scale bar is 100 nm. (J) CD spectra of Δ-nanohelicoids at varying times of L-CPL illumination up to 30 min.

Different growth stages of nanohelicoids were characterized spectroscopically and imaged by electron microscopy (Fig. 1 H–J and SI Appendix, Note S1). After a 5 min exposure, asymmetric dimers of small silver NPs (15 to 20 nm in diameter) were observed (Fig. 1H). The corresponding dimers showed two CD bands in the ultra-violet (UV) and violet parts of the spectrum at 330 and 420 nm with positive and negative signs, respectively (Fig. 1J). After a 10 min exposure, vivid helicity of the nanostructures emerged. The scanning transmission electron microscopy (STEM) images of nanostructure at intermediate growth stage (at 10 min, Fig. 1I) revealed that the final complex nanostructures (i.e., Fig. 1H at 30 min) are formed by twisted horseshoe shaped segments consisting of 5 to 7 NPs with a diameter of 15 to 65 nm. CPL exposure past this 10 min growth time resulted in the formation of vertical stacks of the horseshoe segments (SI Appendix, Fig. S4). The growth of the nanohelicoid is accompanied by the gradual red-shift of the chiroptical bands and emergence of two additional bands for green and near infrared-red wavelengths at 545 and 780 nm, respectively (Fig. 1j).

The geometry and optical properties of nanohelicoids could depend on the incident angle of the CPL. Thus, we examined the morphology of the nanostructures and CD spectra of the resulting chiral plasmonic surface when the laser beam formed four different angles, θ = 0, 5, 10, and 15 degrees, to the surface normal. The Λ-nanohelicoids formed under R-CPL impinging at the substrate surface for larger θ showed a progressively stronger redshift in their CD spectra (SI Appendix, Figs. S5). The anisotropy of individual nanohelicoids and their alignment on substrates, introduced by off-normal incidence of CPL, translated into anisotropic optical activity, including a change in angle-dependent CD for different sample directions (SI Appendix, Figs. S5–S7 and Notes S2 and S3).

Computational Analysis of Different Growth Stages of Silver Nanohelicoids.

To better understand the photon-to-matter chirality transfer process leading to the stand-up forests of nanohelicoids, we used electromagnetic simulations of chirality-dependent plasmonic resonances on objects with complex geometries (Fig. 2A). The observed shape transitions are explained by dynamically changing “hot-spots” whose positions are determined by the handedness of the incident CPL light. Taking NP dimers as a starting point, exposure to CPL results in an intense electrical field concentrated in different parts of the nanostructure (Fig. 2B, Movies S1 and S2, and SI Appendix, Note S4), facilitating the site-selective deposition of NPs (SI Appendix, Fig. S8A and Movie S3). The introduction of metallic NPs at these points with high intensity electromagnetic filed (SI Appendix, Fig. S8B) will induce further deposition of the metal and changes in the position of the hot spots in 3D space. The trajectory of the hot spots and the newly formed metal will produce a croissant-shaped segment curved around the rotational axis of the growing helicoid. The handedness of each newly made segment and, thus, helicoid will be determined by the helicity of the incident photons (Fig. 2C and SI Appendix, Movie S4). Modeling this growth process computationally, we achieved the twisted horseshoe self-assembled geometry of the intermediate staged nanohelicoids (V2 in Fig. 2A) that matches the experimental geometry (Fig. 1I). After this stage, we assumed that the twisted horseshoe-shaped chirality motifs are stacked into multi-layers as we observed in the SEM images of intermediate stacking stage (SI Appendix, Fig. S4). The simulated CD spectra of the models representing the three different intermediate growth stages (V1-3, Fig. 2A) are nearly identical to the experimental ones, with respect to the polarity of the CD peaks and their spectral position (Fig. 2D and SI Appendix, Note S6), albeit some broadening of the CD peaks is inevitable due to some degree of stochasticity in the shape and size of the nanohelicoids.

Fig. 2.

Computational analysis of different growth stages of Δ-nanohelicoids. (A) 3D geometrical models at different growth stages of nanohelicoids. The models were built based on the prediction of dynamic hot spots under illumination with L-CPL, and from left to right, the initial to the final growth stage of nanohelicoid models are presented. (B) The initial dimer model on the substrate (Left) was used as an input to predict asymmetric cross-sectional electric field distribution (Right) under R-(Top) and L-(Bottom) CPL, respectively. (C) The electric field volume plot (Left) under R- (Top) and L- (Bottom) CPL extracted from (B) is used to predict the anchor points of further deposition of silver, altering geometry (SI Appendix, Fig. S8A and Movie S3). The subsequent change in the cross-sectional electric field is observed upon importing such asymmetric nanostructure but still shows rotatory electric field in the same handedness (Right). (D) Calculated differential extinction cross-section spectra under L-CPL to R-CPL (Δ extinction cross-section) of the three different stages (V1-3). (E–I) Electromagnetic models for multi-layered structure in Δ-nanohelicoids. (E) Twisted horseshoe models that represent two different growth stages in (A) (single layer: V2; double layers: V3). (F) Calculated extinction cross sections of the single and stacked twisted horseshoe model under L-CPL and R-CPL. (G) Δ Extinction cross section for the twisted horseshoe model. Calculated current density vector map of (H) single and (I) twisted horseshoe models at different resonance modes.

Notably, from stage V2 to V3, two additional peaks appeared in Fig. 2D (marked by stars). To prove that these new peaks originate from the stacking of the twisted horseshoe (TH) segments on top of each other, we investigate their electromagnetic resonance modes (Fig. 2E) by comparing calculated extinction spectra of single- and double-layered structures (SI Appendix, Note S7). The evolution of the calculated difference in extinction cross section under L- and R-CPL from single- to double-layer clearly shows the splitting of resonance peak and matched experimental observation (Fig. 2 F and G). Each of the two resonance modes at 400 and 525 nm observed for single-stratum assemblies split into two new ones, generating the four resonance modes in total that are characteristic of the double-strata structures. To visualize different resonance modes, the volume arrow plots of current density of both single and double-strata TH were analyzed at their resonance wavelengths (Fig. 2 H and I). While a single-stratum TH shows two modes (i.e., bonding and anti-bonding type of plasmon coupling), additional resonance modes were found in double-strata TH due to interlayer coupling. The multipolar decomposition of scattering calculation of this complex hierarchical structure showed that the modes from interlayer coupling appeared in all four electronic and magnetic dipole and quadrupole scattering components (SI Appendix, Fig. S9).

Biomimetic Aspect of Nanohelicoid Arrays.

The polarization-dependent scattering from the nanohelicoids results in strong polarization-dependent reflection of CPL. Note that that this optical effect is conceptually similar to CPL-dependent colors previously observed for cuticles of Chrysina gloriosa beetles (33) (SI Appendix, Fig. S10A) but made from man-made materials and, in fact much stronger, especially compared to the total thickness of the chiral material. The surface exposed to L-CPL for 15 min (Δ-nanohelicoids) starts to have visible color/brightness variations under L-CPL and R-CPL filters (SI Appendix, Fig. S10B).

Circularly Polarized Light Emission (CPLE) from Single Silver Nanohelicoids.

To further investigate the scattering dissymmetry from nanohelicoids, we analyzed single-particle circularly polarized light emission (CPLE) of nanohelicoidfs with different shapes. We excited Δ- and Λ-nanohelicoid samples prepared on microscope coverslips specialized for single-particle spectroscopy (SI Appendix, Fig. S11). The samples were illuminated with a depolarized laser at a wavelength of 488 nm (details in SI Appendix, Note S8 and Fig. S12). Single nanohelicoids with different handedness displayed different circular polarizations (Fig. 3A). The average optical dissymmetry g-factor of −0.33 indicates that the Δ-nanohelicoids emit R-CPL stronger than L-CPL, and vice versa for Λ-nanohelicoids: g = +0.15. As a benchmark we also measured achiral gold nanospheres that showed nearly zero dissymmetry factor as expected (Fig. 3B and SI Appendix, Table S2). The repeated observation of CPLE intensities from single Δ-nanohelicoids compared to achiral nanosphere control over a hundred imaging frames demonstrates polarization-selective light emission as well as the particles’ robustness and photostability (Fig. 3 C and D).

Fig. 3.

Single-particle analysis of chiroptical activity of silver nanohelicoids on substrate. (A) Single-particle optical microscopy images of Δ-nanohelicoids, Λ-nanohelicoid, and nanosphere during collection of L-CPL (Left) or R-CPL (Right) luminescence upon excitation with randomly polarized 488-nm continuous wave laser light (Scale bars: 250 nm.); (B) Dissymmetry g-factor for Δ- and Λ-nanohelicoids and achiral nanosphere. Each point is the measurement of g for a single particle. The control experiment with achiral nanosphere shows nearly zero asymmetry factor. (C, D) Repeated observations of the max intensity/median intensity of emitted light from a single Δ-nanohelicoid (C) and achiral nanosphere (D) (Odd frames: R-CPL detection; Even frames: L-CPL detection).

Direct 3D Printing of Chiral Plasmonic Surfaces.

The simplicity of CPL-induced formation of the chiroplasmonic nanostructures is a promising pathway for direct-write 3D printing of metasurfaces. Employing a programmable moving stage for the 532 nm laser beam with 0.36 mm spot size (Fig. 4A), we created patterns of discrete small dots alternating translation and illumination (Fig. 4B and SI Appendix, Fig. S13A and Note S9). The focused beam (photon flux: 9.48 × 10²⁴ photons s⁻¹ m⁻²) was parked for 30 s before translation to create a patterned surface with 1 mm diameter. Constant movement of the stage at 1.1 mm/s during illumination further allows the fabrication of more complex patterned surfaces. This continuous printing was demonstrated via the printing of two rows of mirrored letter patterns with R-CPL and L-CPL for the upper and lower row, respectively (Fig. 4C and SI Appendix, Fig. S13B, printing demo recording found in Movie S5). The CD spectra of these two rows display opposing signs for the major chiroptical band at 500 nm (Fig. 4D). Analysis of these two rows by microscale Mueller matrix polarimetry (MMP) (SI Appendix, Figs. S14–S17) also confirmed a complete set of polarization parameters and high-contrast polarization from the patterns. To show application of the printed chiral plasmonic surface patterns in polarization-encryption, we printed the Greek letter χ (as in χειρ, i.e. “hand”) and its mirror image on the same substrate using L-CPL and R-CPL, respectively (Fig. 4E and SI Appendix, Fig. S13C). The circular birefringence satisfied the Kramers–Kronig relation with the CD spectrum and confirmed high rotational power of the nanohelicoid arrays (Fig. 4F). The maps of the Mueller matrix elements acquired for the prominent 500 nm band in the CD spectrum were obtained for the sample area of 5.2 cm by 9.2 cm (Fig. 4 G–J and SI Appendix, Note S9). The images obtained for lasers with different polarization rotation vividly showed negative and positive CD and g-factors in the MMP maps.

Fig. 4.

Direct-write 3D printing of chiral plasmonic surfaces of nanohelicoids. (A) Schematic illustration of the CPL-driven printing system with a motorized stage. (B) A photograph of the grid pattern on a glass substrate. (C) Photographs of continuous writing patterns formed by R- (Upper row) and L (Lower row) CPL using 532 nm lasers, along with their (D) CD spectra. (E–J) Two mirrored letter, “χ”, printed by 532 nm L-CPL (Left) and R-CPL (Right), respectively. (E) Photograph of the sample. (F) CD and CB spectra of Λ-nanohelicoid pattern measured by Mueller matrix polarimetry. (G) Transmittance, (H) absorbance, (I) CD, and (J) g-factor map of the whole pattern obtained by Mueller matrix polarimetry with spatial resolution of 1 mm.

Silver Nanohelicoids Formed under CPL with Different Wavelengths.

Consistent with the chirality transfer determined by the spin angular momentum of the photons, the maximum of the chiroptical response of the silver nanohelicoids is dependent on the wavelength of light source (Fig. 5A). The required photon flux to form silver nanohelicoid patterns can be as low as of 9.74 × 10²⁰ photons s⁻¹ m⁻² when the broadened 405 nm laser source was used. Although 660 nm is far off from the resonance of small silver NPs, the CPL illumination can still successfully produce the superstructures but at a slower rate (SI Appendix, Note S10). The SEM images of the resulting structures under different wavelengths of light source showed that longer wavelengths form bigger nanohelicoids (Fig. 5C). The electromagnetic simulation results from the simplified nanohelicoid model based on the experimental structural parameters of these three superstructures also show spectral range changes in their chiroptical bands (Fig. 5 B and D and SI Appendix, Fig. S19 and S20 and Notes S10).

Fig. 5.

Nanohelicoids formed with various CPL sources with different wavelengths and proof-of-concept sensing. (A–D) Δ-nanohelicoids printed using three different wavelengths: 405, 532, and 660 nm. (A) Experimental CD spectra and (B) calculated differential extinction cross section of (D) 3D Δ-nanohelicoid models; the model built based on the experimental geometry parameters observed in (C) SEM images of Δ-nanohelicoids synthesized by three different wavelength. (E) A photograph depicts a miniaturized 24-variation library grid pattern, comprising bi-layered silver nanohelicoids. These structures were fabricated using different combinations of light handedness (L-CPL and R-CPL) and wavelengths (405, 532, and 660 nm). (F and G) Color and intensity maps illustrate the highest positive (F) and negative (G) CD peaks obtained from the bi-layered library samples shown in Fig. 5E. The legend indicates the wavelength of CPL used for manufacturing the 1st and 2nd layers, represented as + and − for L- and R-CPL illumination, respectively. The diameter of the bubble plot is directly proportional to intensity; full CD and ORD spectra and the CD peak readout data of each sample numbered with gray boxed label is available in SI Appendix, Figs. S22 and S23 and Table S3). (H–J) Detection of Pepsin using CD change in Λ-nanohelicoids plasmonic surface. (H) CD spectra change in increase of concentration of Pepsin. CD peak change plot with increase of concentration of Pepsin in the range of (I) 0 to 2 µM and (J) their low concentration region (0 to 300 pM). All error bars shown in this figure are SD based on the measurements of three sets of samples.

The handedness, wavelength, and power density of the laser beam can be easily varied while we print the patterns in different parts of the substrates. Such flexibility gives the possibility of fast and versatile printing of integrated multi-grid or compartment patterns with various local geometries, which lithographic methods can hardly provide. To show on-the-fly tunability of the chiroptical activity of the deposited silver forests, we printed 24 different grid spots, each with a 1 mm diameter, using a two-step exposure method with various combinations of light source handedness (L-CPL or R-CPL) and wavelength (405 nm, 532 nm, or 660 nm) as illustrated in Fig. 5E (detailed in SI Appendix, Note S11). The resulting CD spectra (SI Appendix, Fig. S22) and optical rotatory dispersion spectra (ORD, SI Appendix, Fig. S23) indicate that the chiroptical spectra can be modulated across the entire UV and visible wavelength range, as demonstrated in Fig. 5 F and G (Additional details and numerical data are provided in SI Appendix, Tables S3 and S4). The optical rotatory power from the nanohelicoid model was also demonstrated using finite-difference time-domain (FDTD) simulations (SI Appendix, Fig. S24 and Note S12).

Proof-of-concept Biosensing with the CPL-printed Chiral Plasmonic Surfaces.

Biomimetic aspects of chiral nanostructures and strong chiroptical activity prompted us to explore the applicability of CPL-printed nanohelicoids for spectroscopic detection of chiral analytes with biological relevance. As a model analyte, we used a protein pepsin whose detection is essential both for medicine and biotechnology. Its sensing can be based on wavelength shift of the CD peak of the nanohelicoids in the technologically convenient visible part of the spectrum (Fig. 5 H–J) as opposed to the far UV spectroscopic window typical for chiroptical activity of amino acids and proteins. The CD spectra gradually red-shifted with increase in pepsin concentration and the limit of detection can be as low as 100 pM (Fig. 5J). We also confirmed the enantiomer selectivity of the system through the different absolute shift observed upon absorption of L-and D-lysine on Λ-nanohelicoids (SI Appendix, Fig. S25).

Discussion

The photoreduction process of silver nitride and citrate has been well studied (34, 35), but a light-induced dynamic process starting from heterogeneous nucleation at the interface has not been yet observed. The possibility of nanoscale dynamic processes on the substrate should still be considered, adding complexity compared to processes in a solution. Information indicating that dynamic assembly here is unlikely as colloids includes SEM imaging of different growth stages (Fig. 1H) and microscopy data that align with our step-by-step simulation results at initial to intermediate stages (Fig. 2). Computational data for the model of the nanohelicoid at the final stage of formation (SI Appendix, Fig. S21) also support our arguments regarding the suggested dynamic process.

The cumulative overview of the data obtained for CPL-printing point to the technological potential of the method for optoelectronics, biosensing, biomedicine and information technologies. Our computer-driven direct-write system can print the patterns with local variations of optical activity. With the spot diameter in Fig. 1B being 2 cm, the patterned features can span four orders of magnitude in x–y dimensions. The simplicity of CPL-driven formation of nanohelicoids under focused laser beam will allow customizing chiral nanostructures for different biomedical communities, exemplified by multi-well microplates plates for high throughput chiroptical analysis (SI Appendix, Fig. S26A). The successful printing of miniaturized multi-grid patterns with various combinations of local patterns (SI Appendix, Fig. S26B) suggests potential development of the chiroptical diagnostics with the higher selectivity and sensitivity than currently ones due to “giant” polarization rotation in helicoids with normal orientation to the substrates. For instance, the combinational chiroptical activity changes of the 24 different nanohelicoids patterns (Fig. 5 E–G) will provide a unique footprint for the specific analyte as each of 24 chiral elements will react differently (SI Appendix, as exemplified in the insets of SI Appendix, Fig. S26B). The continuous patterning ability of our technique shown in Fig. 4C could also allow direct-write printing of the chiral metallic channel in microfluidic channel and sped up applications of metasurfaces as an in-line analysis tool for continuous flow microreactors (SI Appendix, Fig. S26C), which can be used for the study of various chemical reactions that lead to a change of the CD spectrum of the sample.

Finally, it should be also emphasized that the direct-write printing of helicoids does not involve any chiral organic ligands. In fact, it does not involve any chiral or achiral organics, which is utterly essential for various technologies because it avoids chiral contaminants. The chiral purity of photons also avoids uncertainty about unexpected low-occurrence stereomers and chiral additives. This light-specific distinction can provide better sensitivity to its dielectric environment and eliminates charge transport dilemma of solution-processed nanomaterials that often hinders their practical application in electrochemical, photocatalytic and optoelectronic devices (36).

In conclusion, the chirality of incident photons can be transferred to the chirality of nanostructures assembled at liquid-solid interfaces. The rotational axis of produce nanohelicoids is aligned to the surface normal of the substrate, which enables high chiroptical activity and g-factors. The growth of nanohelicoids on a variety of substrates occurs under ambient condition, which enables rapid large-area patterning. The chiroplasmonic surfaces can be used for photonic, optoelectronic, and electromechanical devices. Particular benefits are expected for biosensing and multiparameter health monitoring based on detection of chiral metabolites that require patterns with variable photonic activity. The direct-write CPL patterning may also be suitable for non-metallic optically active nanostructures (32, 37). Altogether, it provides an alternative to nanolithography with advantages of simplicity, versatility, additivity, and scalability.

Materials and Methods

Growth of Silver Nanohelicoids on Substrates.

To synthesize the silver nanohelicoids, silver nitrate (AgNO₃, Cat.# 209139), and trisodium citrate dihydrate (Cat.# S4641) were purchased from Sigma-Aldrich (Milwaukee, WI). Ultrapure water from a Direct-Q3 system (18.2 MΩ cm, Millipore; Billerica, MA) was used in this work., the desired substrate [either glass, indium tin oxide (ITO)/glass, polydimethylsiloxane (PDMS), or silicon] was submerged in a solution of silver precursor (AgNO₃, 2.5 mM) and citrate (12.5 mM), while the height from substrate to surface of the solution was fixed at 3 cm. The thin PDMS substrate (SI Appendix, Fig. S3) was prepared using a PDMS mixture [PDMS curing agent mix (1:10), Dow Corning, US] spun on a silicon mold at 1,000 rpm for 30 s, followed by incubation at 70 °C for 2 h. Three different lasers (wavelength of 405, 532, and 660 nm), modulated to emit either left- or right-circularly polarized light (L-CPL and R-CPL, respectively), were directed perpendicular to the substrate with varying power density for different time up to 30 min depending on purpose of experiments. The patterned substrates were immersed in clear DI water solution and washed three times before optical or imaging analysis.

Optical Setup for CPL-induced Centimeter-scale Patterning of Surfaces with Nanohelicoids.

The diode-pumped continuous wave lasers with three different wavelengths of 405, 532, and 660 nm (CrystaLaser, NV, USA) were used as light sources. The laser emissions were broadened into spots with a diameter of 1 to 2 cm using a combination of plano-concave and -convex lenses (Fig. 1A). The broadened beams were transformed to CPL by directing it through a linear polarizer and a quarter-wave plate. Since the quarter-wave plate is made with a birefringent material, the linearly polarized light turned to CPL by passing through the quarter-wave plate with the 45° transmission angle. By rotating the quarter-wave plate 90° relative to the previous angle, the handedness of the CPL can be changed. We used achromatic waveplates for printing with different wavelengths (see the details in SI Appendix, Note S11).

CD Spectroscopy.

The nanohelicoid arrays for optical analysis were created on cover glass with a size of 22 × 22 mm and thickness of 1.5 mm. CD spectra of the samples were obtained using a Jasco J-815 or J-1700 CD spectrometer equipped with one PMT detector in 200 to 800 nm range and two InGaAs near infrared (NIR) detectors in 800 to 1,600 nm and 1,600 to 2,500 nm range. The CD spectra were measured with the typical scanning parameters: scanning speed, 500 nm/min; data pitch, 1 nm; bandwidth, 5 nm (NIR bandwidth, 10 nm); digital integration time, 0.25 s; and one accumulation. The angle-dependent CD measurements were performed with the sample aligned differently to the incident direction of light source in the J-815, which ranges from 0 to 15 degree from surface normal of substrate.

Mueller Matrix Polarimetry.

Mueller matrix polarimetry (Hinds Inc) was used to obtain a 4 × 4 Mueller Matrix (M). Briefly, a total of four photoelastic modulators (PEMs) were employed to extract M from the samples; 2 of them were for a polarization state generator and the rest of them were for a polarization state analyzer. As described elsewhere (38, 39), four input Stokes vectors generated by PEMs of the polarization state generator and four output vectors analyzed by PEMs of analyzer resulting in a matrix of 16 measured light intensities that can be directly used to determine 16 elements of the M of the sample. The optical parameters [horizontal linear dichroism (LD), 45° linear dichroism (LD′), horizontal linear birefringence (LB), 45° linear birefringence (LB′), circular dichroism (CD), circular birefringence (CB)] were consequently extracted from the M.

Electron Microscopy.

Scanning electron microscopy (SEM) samples were prepared as described above by using ITO-coated glass slides (Nanocs Inc.) as substrates, followed by drying at room temperature. After checking their CD spectra, SEM imaging on samples were performed in FEI Nova NanoLab Dual Beam SEM and FEI Helios Nanolab at 5 kV accelerating voltage and 0.4 nA beam current under secondary electron detection mode. The tilting series of SEM images were obtained by FEI Helios Nanolab to determine handedness of the nanohelicoids (SI Appendix, Note S5 and Fig. S4). Scanning transmission electron microscopy (STEM) samples were prepared on PDMS and transferred to a copper grid coated with holey carbon supported on a continuous carbon film (TedPella 01824) by stamping technique. Simultaneous imaging with three different detectors were performed with FEI Helios Nanolab at 29 kV accelerating voltage.

Single-Particle Circularly Polarized Luminescence Measurements.

Silver nanohelicoids were prepared on glass coverslips and imaged by single-particle spectroscopy under 488-nm continuous wave (CW) randomly polarized light (excitation intensity: 3.87 × 10⁵ mW/cm²). The redshifted emission was selected through a 505-nm long pass filter, and the emission polarization was analyzed with a liquid crystal variable retarder (Thorlabs LCC1223T-A). See SI Appendix, Note S8 for detailed methods.

Electromagnetic Computation.

To elucidate the growth mechanism and chiroptical evolution of the silver nanohelicoids, the electromagnetic computations on initial stages and simplified horseshoe model in Fig. 2 were performed using finite-element-analysis (FEA) software, COMSOL Multiphysics 5.5. See SI Appendix, Notes S4, S6, and S7 for detailed methods. The chiroptical change in nanohelicoid structures with different sizes (Fig. 5 A–D) were evaluated by constructing simplified 3D models based on SI Appendix, Table S1 and electromagnetic simulation by finite-difference time-domain method (FDTD) utilizing the commercial software Lumerical (See SI Appendix, Notes S7 and S10 and Figs. S19 and S20 for details). The validity of simplified models were confirmed with a more realistic model of Δ-nanohelicoid formed by 532 nm laser, which consist of the individual nanoparticle assemblies and their multi-stacks (SI Appendix, Fig. S21). A similar FDTD analysis were performed to evaluate the optical rotatory power of the nanohelicoid structures for linearly polarized light as shown in SI Appendix, Fig. S24. (See SI Appendix, Note S12 for detailed method and discussion).

Direct-write and Two-step Grid Patterning of Chiral Plasmonic Surfaces.

The CPL-induced chiral plasmonic structure printing was demonstrated with a prototype printing set up equipped with a motorized sample stage (Zaber Technology, Inc.). Note that the optical set up for the CPL source in the printing system was configured without broadening optics (neither plano-concave nor -convex lenses were used). The customized Matlab code was used for pixelating imported images and programming the sample stage movement corresponding to x–y coordinates of pixelated images shown in SI Appendix, Fig. S13 (See SI Appendix, Note S9 for more details). Two-step grid printing with combination of CPLs with three-different wavelengths (Fig. 5 E–G) was performed with various combinations of the CPL sources while quartz slide were mounted on motorized sample stage. First and second layer of each dot was formed by exposing the spot under the selected CPL for 60 s for each layer (see SI Appendix, Note S11 for more details).

Biosensing Experiments with Chiral Plasmonic Surface Patterns.

The circular Λ-nanohelicoids surface pattern with 1 cm diameter was printed on cover slides. Lyophilized (salt-free) Pepsin (Cat.# 10108057001) and L- and D-lysine (Cat.# L8021) were purchased from Sigma-Aldrich and dissolved in DI water with various concentrations. The CD spectra of printed pattern were measured after addition of 10 µL of analyte solutions with coverslip.

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Electric field plot of dimer model under L-CPL.

Movie S2.

Electric field plot of dimer model under R-CPL.

Movie S3.

Chiral dimer model reconstruction process.

Movie S4.

Electric field plot of Δ-dimer model under L-CPL.

Movie S5.

CPL- printing of chiral metasurfaces.

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.