Diameter-dependent nanojoint formation and grain refinement in femtosecond laser nanojoining of AgNWs

Abstract

Femtosecond (fs) laser has emerged as an effective technique for nanojoining. However, the role of nanowire diameter and the associated electric field distribution under fs laser irradiation in nanojoint formation and polycrystalline evolution remains unclear. In this study, AgNW nanojoints were fabricated via fs laser irradiation, and the electric field intensity distribution was simulated using the three-dimensional finite-difference time-domain (FDTD) method. Nanojoint formation was examined across nanowires of different diameters under consistent laser processing conditions. High-resolution transmission electron microscopy (TEM) revealed that smaller-diameter nanowires (e.g., 60 nm) facilitate nanojoint formation with a relative joint cross-sectional area reaching 544%, significantly exceeding those of 200 nm (73%) and 300 nm (19%) nanowires. Moreover, regions with higher electric field intensity exhibited refined polycrystalline grains, with an average size of 4.42 nm at the nanojoint interface—substantially smaller than that at non-bonded regions (15.69 nm and 26.27 nm). These findings provide new physical insights into the role of localized electromagnetic enhancement in joint morphology and grain structure regulation during fs laser nanojoining.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-54617-7.

Introduction

Owing to their excellent optical, electrical, and mechanical properties^1–5, silver nanowire (AgNW) networks hold significant promise for numerous applications including flexible lighting^6–8, artificial skin^9,10, wearable devices^11–14, and flexible energy storage devices^12,15,16. As an important component of AgNW networks, Nanojoints have a significant impact on network performance¹⁷. Recently, methods for preparing AgNW nanojoints have primarily included solution processing¹⁸, mechanical pressing^19,20, high-energy beam irradiation^21–23, and intense light irradiation^24–26, etc. Among these, femtosecond (fs) laser irradiation is considered an ideal method, offering high peak power, low thermal effects, and minimal collateral damage^27,28.

The welding of AgNWs using fs laser is generally attributed to laser-induced photothermal effects. For instance, Liu et al. demonstrated that fs irradiation initiates nanojoint formation and localized melting, highlighting the role of complex photothermal processes²⁹. Further studies by Liang et al. established that precise control of laser energy enables low-damage ablation and welding³⁰. While Chen et al. revealed that the macroscopic stacking arrangement of nanowires dictates the mode and extent of the melting behavior³¹. However, these prior works have primarily focused on optimizing external laser parameters or macro-scale network geometries. In contrast, the governing role of intrinsic nanowire properties—specifically the diameter—in modulating the localized surface plasmon resonance (LSPR) and optical energy deposition remains underexplored. This study addresses this gap by elucidating the diameter-dependent plasmonic mechanisms and the concomitant evolution of grain structure with the nanojoints.

In this study, the nanojoints were fabricated using fs laser irradiation under controlled and reproducible conditions, with critical factors such as polyvinylpyrrolidone (PVP) removal being rigorously accounted for. Three-dimensional finite-difference time-domain (FDTD) method was employed to simulate the distribution of electric field intensity during the irradiation by fs laser. Nanojoint formation was observed in nanowires of different diameters under identical fs laser irradiation. Furthermore, polycrystalline grain size was analyzed across different regions of the same nanojoint. A high-resolution transmission electron microscopy (TEM) was utilized to analyze the change in crystalline structure of silver nanomaterials. It was found that under the identical fs laser irradiation, the smaller the diameter of the nanowires, the more readily nanojoints can be fabricated. Additionally, increased electric field intensity correlated with a decrease in polycrystalline grain size.

Experiments

Materials

AgNWs suspension (20 mg/ml) is purchased from XFNANO Materials (Jiangsu XFNANO Materials Tech Co., Ltd. Nanjing, China). The diameter of AgNWs ranges from 60 to 300 nm, and the length of AgNWs ranges from 7 to 13 μm. Silicon wafers are obtained from Shenshi materials (Shenshi Chemical Corporation Co.,LTD. Wuhan, China).

Sample preparation

To remove the PVP layers covering the AgNWs surfaces, the suspension was sonicated in 40 ℃ warm water for 20 s. Subsequently, the AgNWs suspension was concentrated via centrifugation at 4000 rpm. The clear supernatant was carefully decanted and discarded, leaving behind a high-concentration AgNWs paste. This paste was washed with deionized (DI) water to remove residual organics, followed by repeating the sonication and centrifugation steps. SEM characterization confirmed that this cleaning protocol effectively removed the PVP insulation layer without causing surface etching, breakage, or morphological deformation to the AgNWs, ensuring that the subsequent welding was purely metallic. The purified AgNWs paste was then diluted with DI water, deposited dropwise onto the substrate, and air-dried.

Laser irradiation

In this work, the fs laser with a central wavelength 1030 nm, a pulse duration 800 fs and a repetition rate 200 kHz was employed. The laser scanning speed was set at 1000 mm/s, the spot diameter was 10 μm, the scanning step size (pulse-to-pulse pitch) was 5 μm, and the laser scanning path spacing was 10 μm. Samples were welded using fs laser irradiation at fluences ranging from approximately 6 to 10 mJ/cm². To ensure statistical reliability and reproducibility, welding experiments were performed on over 50 distinct nanowire junctions for each diameter group. The results demonstrated high consistency, with a successful welding rate exceeding 95% under optimal fluence conditions. Schematic diagram of fs laser microprocessing system is shown in Fig. S1.

Characterizations

Scanning electron microscopy (SEM, MIRA 3 LMH, Czech Republic) was used to measure the surface morphology of AgNWs during fs laser irradiation. A focused ion beam (FIB, FEI Super X, USA) was used to prepare Transmission electron microscope (TEM) specimens. A high-resolution TEM (HR-TEM, Titan G2 60–300, USA) was employed to observe crystalline transformations within AgNWs nanojoints.

Results and discussion

The AgNW samples irradiated by fs laser are labelled as AgNW-P, where P stands for diameter of AgNW in nm (Table 1).

Table 1.

The diameter and label of silver materials. Sample diameter and length are also listed.

Category	a	b	c
Diameter (nm)	60	200	300
Label/AgNW-P	AgNW-60	AgNW-200	AgNW-300

The electromagnetic field distribution was simulated using the FDTD method. The AgNWs were modeled as cylinders with a fixed length of 2 μm (Fig. 1) and varying diameters (60, 200, and 300 nm) (Fig. 1(b)), using dielectric functions from the CRC Handbook³². A Gaussian source (λ = 1030 nm) was placed directly above the junction of the nanowire terminals. The source propagated along the − z direction with a polarization parallel to the x-axis, and the beam waist radius was set to 0.5 μm. The simulation volume was defined as 2 × 3 × 4.8µm3 (x×y×z). To ensure high resolution and convergence, a mesh size of 2 nm and a simulation time of 600 fs were employed, with Perfectly Matched Layers (10 layers) applied to all boundaries. Note that this model solves linear Maxwell’s equations for the steady-state field distribution; nonlinear absorption and fs pulse dynamics were not included.

Fig. 1.

(a1) Schematic illustration of fs laser irradiation at the end-to-end nanojoint of two AgNWs. (a2) Local enlargement of the purple dashed box in (a1). Distribution of electric vector in two vertical AgNWs with diameters of 60 nm (b1), 200 nm (b2), and 300 nm (b3).

Figure 1(a1) illustrates the schematic of fs laser irradiation, while Fig. 1(a2) details the simulation model where two AgNWs are arranged orthogonally with adjacent ends. The corresponding simulation results are presented in Fig. 1(b). The top-view field maps indicate that AgNW-60 exhibits the strongest surface electric field enhancement compared to AgNW-200 and AgNW-300. As shown in the insets of Fig. 1(b), the maximum electric field intensity is concentrated precisely at the junction of the nanowire ends.

The anisotropic field distribution observed in Fig. 1(b) is primarily governed by the laser polarization. Since the incident beam is linearly polarized along the x-axis, the horizontal nanowire (aligned with the polarization) satisfies the condition for strong longitudinal LSPR excitation. In contrast, the vertical nanowire, being perpendicular to the polarization direction, shows minimal field response.

Figure 2 shows the specific values of the electric field intensity on the surface of nanowires with different diameters (60 nm, 200 nm, 300 nm). Owing to the local geometric discontinuity in the nanowire regions³³, the electric field strengths of AgNW-60, AgNW-200 and AgNW-300 are all at their own maximum at the X-coordinate of 0 (where the nanowire ends junction). The maximum electric field strength at the end junction of AgNW-60 is 12.91 V/m, which is 1.56 times the maximum electric field strength at the end junction of AgNW-200 (8.29 V/m), and 1.71 times of that at the end junction of AgNW-300 (7.54 V/m), respectively.

Fig. 2.

Electric vector intensity of AgNWs with diameters of 60 nm, 200 nm, and 300 nm.

The transition from localized field enhancement to morphology evolution is governed by a synergy of non-thermal and thermal mechanisms. In the initial sub-picosecond stage, the intense electric field triggers high-density electronic excitation, potentially inducing “electronic softening” that facilitates atomic mobility through transient bond weakening³⁴. Subsequently, the energy is transferred to the lattice via electron-phonon coupling³⁵, where the deposited volumetric energy density (ω) scales with the local field intensity (ω ∝ Im(ε)|E|²)³⁶. Given that AgNW-60 exhibits a significantly higher |E|² than AgNW-300, the energy density in thinner wires readily exceeds the threshold for volumetric melting, whereas thicker wires are restricted to surface-limited melting. This theoretical framework explains the diameter-dependent joint areas observed in the following SEM analysis.

To verify this diameter-dependent mechanism, surface morphology characterization was conducted using SEM. Figure 3(a-c) shows the morphology of nanojoints of AgNW-60, AgNW-200, and AgNW-300, after irradiation at the same fs laser power, respectively, indicated by the light blue dashed rectangle. To rigorously quantify the welding efficacy and address the limitations of single-point measurements, we performed a statistical analysis on the Projected Joint Area Ratio. This ratio is defined as the maximum 2D projected area of the nanojoint (indicated by the yellow ellipse in Fig. 3a) divided by the projected area of the pristine nanowire (red ellipse). It should be acknowledged that while this metric provides a consistent quantitative index for comparing welding efficacy, it reflects the two-dimensional (2D) projected morphology rather than the full three-dimensional (3D) geometry of the junction. Nonetheless, it serves as a robust proxy for the overall material redistribution during the fs laser nanojoining process.

Fig. 3.

Morphology of nanojoints connected to the ends of the vertical nanowires with diameters of (a) 300 nm, (b) 200 nm, and (c) 60 nm. (d) Statistical box plots of the Projected Joint Area Ratio as a function of nanowire diameter. The values (544%, 73%, 19%) denote the average ratio for each group. Note that the Y-axis is presented on a logarithmic scale. Statistical data in (d) were derived from randomized junction populations (n > 23 per group) to ensure results are independent of local geometric variations.

As shown in the statistical box plots in Fig. 3(d), data were collected from independent samples with sizes of n = 32 (AgNW-60), n = 23 (AgNW-200), and n = 32 (AgNW-300). The results reveal a significant size-dependent trend. For the thickest wires (AgNW-300), the average ratio is relatively low at 19%, indicating limited surface melting. This value increases to 73% for AgNW-200. Most remarkably, for the AgNW-60 group, the average ratio surges to 544%. As revealed by our Molecular Dynamics (MD) simulations (Fig. S2), this is attributed to a ‘spherical collapse’ mechanism. In 60-nm wires, the concentrated energy density (ω) leads to total volumetric melting, causing the junction to coalesce into a near-spherical droplet to minimize surface energy. It should be noted that while the MD simulations were performed on smaller diameters (5–10 nm) to maintain computational efficiency, they effectively capture the qualitative transition from volumetric melting in thin wires to surface-limited melting in thicker counterparts. This transition provides a robust physical framework that explains the drastic increase in the projected joint area observed experimentally in the 60 nm AgNWs. The use of a logarithmic scale in Fig. 3(d) highlights the orders-of-magnitude difference in material redistribution between the 60-nm wires and the thicker counterparts. Despite the orientation-dependent nature of LSPR, the consistent diameter-driven trends observed across large-scale random networks underscore the robustness of the reported welding mechanism.

Unlike the end-to-end configuration discussed in Figs. 1 and 4(a1) presents the simulation model for a crossed-junction geometry, where the fs laser irradiates the overlapping intersection of two orthogonal AgNWs. A partial enlargement of the purple dashed box in Fig. 4(a1) is shown in Fig. 4(a2), revealing two vertically oriented silver nanowires intersecting at a junction. FDTD simulations (Fig. 4(b1)) clearly indicate a localized maximum in electric field strength at this nanowire intersection. Figure 4(b2) provides an enlarged view of the marked region in Fig. 4(b1). Three reference lines are defined in Fig. 4(b2): the red dashed line (“Gap”) corresponds to the intersection between AgNW A and AgNW B, the gray dashed line (“Top”) lies within nanowire A, and the cyan dashed line (“Bottom”) is positioned within nanowire B. As evident from the inset, the electric field strength along the “Gap” line significantly exceeds that along both the “Top” and “Bottom” lines.

Fig. 4.

(a1) Schematic illustration of fs laser irradiation at the crossed-nanojoint of two AgNWs. (a2) A partial enlargement of the purple dashed box in (a1). (b1) Cross-sectional electric field distribution at nanowires nanojoint calculated by FDTD. (b2) A partial enlargement of the purple dashed box in (b1). The results confirm that LSPR enhancement is highly localized at the contact points, where energy deposition is dominated by the junction geometry rather than the proximity of adjacent nanowires.

Figure 5 illustrates the electric field intensity profiles along the dashed lines labeled Top, Gap, and Bottom. The electric field intensity along the dashed line Gap is significantly higher than those along Top and Bottom. The maximum intensity along Gap (peaking at 2.59 V/m) occurs at an X-coordinate of 0.048. In contrast, the intensities along the dashed lines Top and Bottom remain nearly constant, reaching their maximum values of only 0.05 V/m and 0.005 V/m, respectively, at an X-coordinate of 0.10.

Fig. 5.

Electric field strength at the three dashed lines Top, Gap, Bottom.

Due to the limited recrystallization volume in thicker nanowires, quantitative grain size analysis was primarily conducted on the AgNW-60 samples, which exhibit the most extensive microstructural evolution. To reveal the internal microstructural evolution, FIB-prepared cross-sectional samples of the representative AgNW-60 nanojoints were analyzed using high-resolution TEM (Fig. 6). Quantitative grain size analysis was performed using Nano Measurer 1.2 software on over 50 individual grains collected from the three distinct regions shown in Fig. 6(b–d).

Fig. 6.

TEM image of Ag nanowire material after fs laser irradiation. White squares indicate polycrystalline grains of the nanomaterials. (a) TEM image of the interconnections of Ag nanowire materials. (b) TEM image corresponding to the red circle in (a), with the corresponding SAED image in the upper right corner. (c) TEM image corresponding to the light cyan circle in (a), with the corresponding SAED image in the upper right corner. (d) TEM image corresponding to the purple circle in (a), with the corresponding SAED image in the upper right corner. (e) Statistical box plots of grain sizes measured from the regions corresponding to (b), (c), and (d). The numeric values and green squares indicate the mean grain size.

Figure 6(a) shows the cross-section view of intersect AgNWs with laser irradiation. Figures 6(b-d) present TEM images showing different grain orientations within the Ag nanocrystals. The region in Fig. 6(d) corresponds to the purple square in Fig. 6(a), located at the bonding interface with the strongest localized field enhancement. The material here exhibits a highly polycrystalline nature with ultrafine grains. The average grain size is measured to be 4.42 ± 2.48 nm (n = 41). The corresponding SAED pattern (upper right corner) confirms the random crystallographic orientations. The region in Fig. 6(b) marked by the red square, the grain size is significantly larger, averaging 15.69 ± 3.60 nm (n = 28). Lattice fringe measurements reveal spacings of 0.23 nm and 0.24 nm, corresponding to Ag(111) and Ag(200) planes, respectively. The region in Fig. 6(c) far from the junction center (marked by the light cyan square), the grains are the coarsest, with an average size of 26.27 ± 7.88 nm (n = 22). Figure S3 displays the TEM image and SAED pattern of pristine AgNWs without laser irradiation, confirming their single-crystalline character. This stands in stark contrast to the polycrystalline structures observed in the laser-irradiated regions (Fig. 6).

The statistical analysis (summarized in box plots in Fig. 6e) reveals a clear gradient: grain size is smallest in the Fig. 6(d) region and largest in the Fig. 6(c) region. This distribution strongly correlates with the non-uniform electric field intensity: the intense localized field at the bonding interface (Fig. 6d) drives rapid melting and deep supercooling, leading to high nucleation rates and grain refinement, whereas the lower field intensity at the surface (Fig. 6c) allows for slower cooling and grain growth.

The distinct melting and solidification behaviors are further governed by the ultrafast thermal dynamics under fs laser irradiation. Upon irradiation, the laser energy is first absorbed by the electron system, followed by energy transfer to the lattice through electron-phonon (e-ph) coupling within several picoseconds³⁵. In thinner nanowires (e.g., 60 nm), the high localized electric field leads to a higher energy density deposition per unit volume (ω). Due to their smaller thermal mass and limited heat sink capacity compared to 300 nm nanowires, the lattice temperature rises rapidly above the melting point throughout the cross-section. Subsequently, the rapid heat dissipation to the surroundings and the cold nanowire segments triggers an extreme cooling rate (exceeding 10¹² K/s), as predicted by the MD models. This ultra-fast quenching induces a state of deep supercooling, which, according to classical nucleation theory, exponentially increases the nucleation rate. This facilitates the formation of a high density of nuclei, resulting in the observed grain refinement (4.42 nm) at the bonding interface where the localized field is strongest.

Based on the observed diameter-dependent mechanisms, certain guidelines for practical parameter optimization can be inferred. For fine NW networks (e.g., AgNW-60), the laser fluence should be precisely controlled at a lower level (e.g., near 6 mJ/cm²) to prevent the complete volumetric melting and morphology distortion (spherical collapse) triggered by high LSPR enhancement. In contrast, thicker NWs (e.g., AgNW-300) exhibit a higher tolerance to laser energy due to lower field enhancement and require increased fluence (e.g., up to 10 mJ/cm²) to promote sufficient material diffusion across the larger contact area. These trends suggest that laser parameter selection must be tailored to the specific NW dimensions to balance joint strength and network structural integrity.

Overall, the intrinsic nanowire diameter serves as the governing factor in the nanojoining process by modulating the LSPR intensity. This, in turn, dictates the volumetric energy deposition density (ω), which triggers distinct melting modes—ranging from localized surface melting in 300 nm NWs to complete volumetric melting in 60 nm NWs. The subsequent ultrafast thermal quenching dynamics, characterized by extreme cooling rates and deep supercooling, further drive the evolution of the grain structure. These findings bridge the gap between microscopic plasmonic excitation and macroscopic joint properties, providing a mechanistic framework for precision nano-manufacturing using ultrashort pulse lasers.

Conclusion

In summary, AgNWs nanojoints were fabricated by fs laser irradiation. It was found that the electric field strength has a significant effect on the formation of nanojoints and the refinement of polycrystalline grains. AgNW-60, exhibiting the highest electric field strength at nanowire terminals, produced nanojoints with a projected area ratio significantly exceeding that of the thicker nanowires (AgNW-200 and AgNW-300), attributed to the volumetrically enhanced melting and spherical collapse mechanism. In addition, a distinct site-specific grain size distribution was revealed: the measured polycrystalline grain size at the bonding interface (where the electric field is maximized) is drastically refined compared to the grain sizes at the non-bonding surface regions. This study provides new physical insights into the relationship between electric field strength at nanomaterial surfaces and the resulting nanojoint morphology and polycrystalline grain structure after fs laser irradiation.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Q.Z. designed and performed the experiments, collected the data, and wrote the manuscript. M.C. and X.L. helped in experimental data analysis. Y. Z. and B. S. conducted the molecular dynamics simulations and analyzed the theoretical data. X. L. contributed to manuscript. All authors discussed the progress of the research and reviewed the manuscript.

Funding

This research was funded by Key Research Projects of Higher Education Institutions in Henan Province (Grant No. 24A460003), the General program of Henan Natural Science Foundation (Grant No. 252300420022 and No. 262300420067), Central Guided Local Science and Technology Development Fund Program of Henan Province (Grant No. Z20241471072), and Henan Institute of Technology High-Level Talent Research Start-up Fund (Grant No. KQ2307).

Data availability

The data that support the findings of this study have been deposited in the scientific data bank. The private access link is: https://www.scidb.cn/en/s/JFN3ey.

Declarations

Competing interests

The authors declare no competing interests.