COMMUNICATIONand as self-organized nanopatterns induced on a surface, [ 30 ] are among other polarization-related phenomena demonstrated recently.The infl uence of beam polarization orientation on laser processing has been thoroughly studied in the cases of conductive and dielectric solid targets. [ 31,32 ] There are known effects of polarization on the scalar parameters of laser-matter interaction, such as absorption coeffi cient and ionization rate. [33][34][35][36] It is also known that heat conduction fl ux (vector) in plasma might be dependent on the direction of the imposed fi eld. [ 37 ] In what follows, these effects are considered in succession: (i) accumulation from multiple pulses, (ii) effects of polarization under high-numerical aperture (NA) focusing, and (iii) the infl uence of the external high-frequency electric fi eld on electronic heat conduction. The latter contribution has not yet been considered in laser fabrication under tight focusing. All these photomodifi cation mechanisms occur simultaneously and affect polymerization, which takes approximately a millisecond for common photoresists [ 38 ] at a >90% voxel overlap at typical writing velocity of 100 µm s -1 for widespread laser 3D nanolithography. [ 39 ] In this paper, a systematic analysis via modeling and experiments is presented in order to reveal polarization effects, their infl uence on the feature size (resolution), and the coupling between thermal gradient and polarization in DLW.To study and demonstrate the polarization effects, 3D suspended resolution bridges at various angles, α , between the linear polarization and scanning direction were fabricated on a glass substrate ( Figure 1 inset in (a); see details in the Experimental Section). The line-width difference was 10%-20% (varying exposure) under typical polymerization conditions for the linearly polarized pulses. The largest width of a 3D bridge was observed when the scan direction was perpendicular to the orientation of linear polarization, α = π / 2 ( E ⊥ v s ). The heightto-width ratio of the suspended lines was dependent on the orientation of linear polarization and changed between 3.07 and 3.44; self-focusing was not present under our experimental conditions. [ 40 ] Detailed analysis of polarization, threshold, and heat accumulation effects, which are all important, are discussed next.For the experimental conditions used, the focal spot diameter (at 1/ e 2 ) can be calculated as d f = 1.22 λ /NA = 898 nm assuming, for simplicity, a Gaussian intensity profi le. However, at such tight focusing it is necessary to use the vectorial Debye theory (specifi cs [ 41 ] can be found in Section A, Supporting Information), which predicts an ellipsoidal focal spot with two lateral cross sections: W l = 790 nm and W s = 572 nm for long and short cross sections, respectively (or 500 and 360 nm at full width at half maximum (FWHM)) for the actual experimental