λ/30 inorganic features achieved by multi-photon 3D lithography

It’s critically important to construct arbitrary inorganic features with high resolution. As an inorganic photoresist, hydrogen silsesquioxane (HSQ) has been patterned by irradiation sources with short wavelength, such as EUV and electron beam. However, the fabrication of three- dimensional nanoscale HSQ features utilizing infrared light sources is still challenging. Here, we demonstrate femtosecond laser direct writing (FsLDW) of HSQ through multi-photon absorption process. 26 nm feature size is achieved by using 780 nm fs laser, indicating super-diffraction limit photolithography of λ/30 for HSQ. HSQ microstructures by FsLDW possess nanoscale resolution, smooth surface, and thermal stability up to 600 °C. Furthermore, we perform FsLDW of HSQ to construct structural colour and Fresnel lens with desirable optical properties, thermal and chemical resistance. This study demonstrates that inorganic features can be flexibly achieved by FsLDW of HSQ, which would be prospective for fabricating micro-nano devices requiring nanoscale resolution, thermal and chemical resistance.


Supplementary
by FsLDW with the scanning speed from 10 to 100 m/s. For shorter exposure (exposure time smaller than 270 ms, corresponding to a laser scanning speed higher than 50 m/s), the photocuring laser threshold power versus the scanning speed obey the rule 1-3 : Where Pth is the laser threshold power at a given scanning speed, C is a coefficient associated with the characteristic of the photoresist, v is the scanning speed, and N is the nonlinear absorption order in HSQ. For HSQ line array fabricated by a 780 nm femtosecond laser, N is determined to be 3.83 (Supplementary Figure 3k), indicating multi-photon absorption in HSQ by FsLDW. As a result, we depict that FsLDW of HSQ is attributed to multi-photon lithography.

Supplementary Note 2
(1) AFM images of HSQ nanowires by FsLDW using single-scanning method with different scanning speeds Morphology of the HSQ nanowires by FsLDW with different scanning speeds are investigated with atomic force microscope (AFM). Supplementary Figure 4 shows the morphology and sectional profile information of the HSQ nanowires by FsLDW with different scanning speeds.
According to the AFM results, the width of the HSQ nanowires gradually drops with the increased scanning speed, which agrees well with the SEM measurement ( Figure 2c). Nevertheless, the height of the HSQ nanowires keeps almost constant at ≈80 nm when the laser scanning speed is the HSQ oligomers appear at the fringe of the focus spot. The former will result in the formation of HSQ features, while the latter will not lead to any feature after development since it's not adequate to support the formation of microstructures. The interference between adjacent lines is neglegible since the spacing of adjacent lines is 2 m, which is much larger than the laser focus spot. In this case, 33 nm feature size can be achieved due to the threshold effect. Nevertheless, when we perform cross-scanning method with 0.5 m spacing between adjacent lines (Supplementary Figure 5b), the interference between adjacent lines is no longer negligible. For the laser scanning in the vertical direction, the crosslinking degree of HSQ in the laser exposed region will be enhanced by the existed HSQ oligomers caused by the fringe of the focus spot in the first laser scanning. HSQ features with adequate crosslinking degree can be constructed by reducing the laser intensity closer to the laser intensity threshold. As a result, much smaller region of HSQ can be photocured with laser intensty approaching closer to the threshold, allowing for the formation of HSQ feature with 26 nm feature size.

(3) SEM images of 2D HSQ features by FsLDW employing single-scanning method and cross-scanning method
To collaborate the formation mechanism of the HSQ features by FsLDW through different scanning methods, we studied the relationship between the feature size and the line spacing of 0.5,

(4) SEM images of 2D and 3D HSQ features by FsLDW employing cross-scanning method
We performed FsLDW to fabricate 2D and 3D HSQ grids by employing cross-scanning method.  drop of the laser intensity will not allow the formation of HSQ nanowire.
Supplementary Figure 8d illustrates the dependence of the linewidth of the HSQ nanowires on the laser intensity with different scanning speeds. The dependence of the linewidth on laser intensity can be fitted by equation (2), which has been employed to describe the resolution of organic photoresist without photoinitiator through multi-photon ionization (MPI) process 4 .
Where D is the linewidth of the cured HSQ nanowires, w0 is the focal radius of the focused fs laser beam, I is the laser intensity of the incident laser, and Ith is the threshold laser intensity for photocuring of HSQ. Supplementary Figure 8d shows good agreement of experimental data and fitting curves according to equation (2), indicating a nonlinear multi-photon absorption process of HSQ by FsLDW. As a result, the dependence of linewidth of HSQ on laser intensity can be theoretically predicted, which will facilitate the construction of microstructures with required nanoscale size.     (4) The refractive index of pristine HSQ, fs laser exposed HSQ, fs laser exposed HSQ after thermal treatment at 400 and 600 ℃

(3) Focusing property of the HSQ Fresnel lens after chemical treatment
We investigated the focusing property of the heated HSQ Fresnel lens after being exposed to typical chemical reagents for 1 hour (Supplementary Figure 16). The focus images of the HSQ Fresnel lens exposed to typical chemical reagents are exhibited in Supplementary Figure 16b2 The change of FWHM after being exposed to the chemicals is less than 3.8 %.
It's worth noting that the focusing property of the Fresnel lens did not change much for even being exposed to 98% H2SO4, indicating the good chemical resistance of the HSQ microstructure.  Optical microscopy images of the HSQ grating microstructures by FsLDW with the periodicity of a 1 m, d 2 m, and g 4 m, respectively. The insets are the AFM images of the HSQ gratings. SEM images of the HSQ gratings with the periodicity of b 1 m, e 2 m, and h 4 m, respectively. The insets are the magnified SEM images. The diffraction pattern, diffraction intensity and diffraction angel observed in transmission for the HSQ grating microstructures with the periodicity of c 1 m, f 2 m, and i 4 m, respectively. The number refers to the diffraction order of each HSQ grating, and the zero order of the diffraction pattern is covered for clarity.
Moreover, we fabricated diffractive optical device, i.e. gratings with the periodicity of 1, 2, and 4 m, respectively. The size of each diffractive grating structure was set to 100 m × 100 m.
Due to the good reproducibility and stability of the photocuring of HSQ by FsLDW, uniform diffractive grating microstructures were fabricated on glass substrates. The uniformity of the HSQ diffractive gratings is verified by the optical transmittance micrograph, AFM and SEM images, as  Table 2). As the periodicity of the HSQ grating grew to 2 m, we observed increased laser diffraction beam spots of the first, second, and third order at 15.3º, 31.1º, and 52.6º, respectively. As for the period width of 4 m, we recorded the diffracted laser beam spots up to 6th order with the diffraction angles from 7.5º to 52.5º, while the 7th order diffraction spot is too weak to be recognized by the mobile phone camera. The diffraction angles can be theoretically estimated by equation (3):

= sin 
In Equation 3, m refers to the diffraction order, λ is the wavelength of incident light, d is the period width of the grating, and  is the diffraction angle for different orders. It is obvious that when the periodicity of the diffractive gratings decreases, the number of supported orders also declines, and each supported order covers a larger diffraction angle. Moreover, the deviation of the diffractive angels between experimental results and theoretical calculation is not more than 1 º (Supplementary Table 2). The agreement of the diffraction order and the diffractive angle between the experimental results and theoretical calculation verifies high quality of the HSQ gratings fabricated by FsLDW.
(2) Experimental data and theoretical calculation of the diffraction angle of the HSQ gratings Supplementary  We further demonstrate the fabrication of HSQ macrostructure by fabricating millimeter-sized HSQ grating structure employing FsLDW. As shown in Supplementary Figure 18a, HSQ grating macrostructure with 1 mm × 0.2 mm has been successfully constructed. SEM image reveals no defect in the uniform HSQ grating structure, in which the periodicity is 4 m (Supplementary   Figure 18b). Supplementary Figure 18c shows the diffraction pattern observed in transmission for