Objective　Phase　change　materials　(PCMs)　have　demonstrated　significant　potential　in　the　application　of　non-volatile　integrated　photonic　devices,　garnering　considerable　attention　in　photonic　integrated　circuits.　Ge2Sb2Te5　(GST),　as　a　classic　phase-change　material,　exhibits　rapid　and　reversible　transitions　between　amorphous　and　crystalline　states　under　optical　or　electrical　pulses,　accompanied　by　a　substantial　difference　in　the　complex　refractive　index　between　the　two　states.　In　conventional　photonic　devices,　PCMs　are　typically　coated　directly　on　the　waveguide,　covering　the　top　and　both　sidewalls.　This　approach　offers　excellent　modulation　performance,　but　it　also　increases　the　transmission　loss　in　the　amorphous　state.　To　address　this　issue,　various　optimized　structures　have　been　proposed　by　researchers.　However,　the　influence　of　different　attachment　methods　of　PCMs　on　the　optical　transmission　loss　and　effective　refractive　index　of　waveguides　has　not　been　fully　investigated.　In　this　study,　simulations　and　analyses　are　conducted　on　the　light　transmittance　of　waveguides　coated　with　GST　in　different　configurations,　as　well　as　on　the　modulation　effects　of　phase　transitions　on　the　waveguide’s　transmission　loss　and　effective　refractive　index.　An　optimized　method　for　attaching　GST　to　the　waveguides　is　proposed　and　experimentally　verified.　Methods　A　ridge　waveguide　is　used　to　conduct　the　simulation,　which　has　a　width　of　450　nm,　a　thickness　of　220　nm,　and　a　ridge　height　of　90　nm.　The　waveguide　cross-section　and　optical　field　distribution　at　1550　nm　are　shown　in　Fig.　1(a).　When　the　upper　cladding　material　is　air,　the　optical　field　in　the　waveguide　is　well　confined　to　the　center,　exhibiting　low　optical　transmission　loss.　Four　different　coverage　configurations　of　GST　on　the　waveguide　are　compared:　full　coverage　(all),　top　coverage　only　(top),　side　coverage　only　(side),　and　trench　filling　(trench),　with　their　waveguide　cross-sections　illustrated　in　Fig.　1(b).　The　optical　field　transmission　in　the　waveguide　is　simulated　using　the　finite-difference　time-domain　(FDTD)　method,　with　the　test　structure　shown　in　Fig.　1(c).　This　simulation　yields　the　transmission　rate　of　the　waveguide　in　both　the　crystalline　and　amorphous　states　of　GST.　Additionally,　the　spectral　response　of　a　racetrack　micro-ring　resonator　is　used　to　analyze　the　modulation　differences　in　the　effective　refractive　index　and　transmission　loss　when　GST　is　attached　to　a　straight　waveguide　and　a　bent　waveguide.　Finally,　to　reduce　the　fabrication　process　requirements,　GST　is　applied　only　to　the　top　surface　of　the　waveguide.　An　optimized　method　is　proposed　and　experimentally　verified.　Specifically,　the　etching　groove　of　the　waveguide　is　filled　with　silicon　dioxide,　and　the　phase　change　material　is　deposited　on　the　plane　formed　by　the　silicon　dioxide　and　the　top　surface　of　the　waveguide.　Results　and　Discussions　The　transmission　rates　of　waveguides　under　different　GST　coverage　configurations　in　both　crystalline　and　amorphous　states　are　shown　in　Fig.　3.　The　simulation　results　demonstrate　that　covering　only　the　top　surface　of　the　waveguide　with　GST　yields　lower　transmission　loss　than　the　conventional　full-coverage　approach　while　retaining　strong　modulation　effects　during　phase　transitions　between　the　amorphous　and　crystalline　states.　GST　full　coverage　on　a　bent　waveguide　introduces　higher　transmission　loss　compared　to　a　straight　waveguide,　yet　fails　to　improve　the　effective　refractive　index　modulation　during　phase　transitions.　The　optimized　method　is　then　proposed:　filling　the　waveguide　trench　with　SiO2　and　depositing　GST　on　the　plane　formed　by　SiO2　and　the　top　of　the　waveguide.　As　illustrated　in　Fig.　7,　this　method　reduces　the　difference　in　the　effect　of　covering　the　phase-change　material　on　straight　and　bent　waveguides,　while　also　decreasing　the　waveguide　transmission　loss　in　the　amorphous　state　compared　to　the　full-coverage　approach.　The　optimized　method　also　features　lower　fabrication　process　requirements.　Experimental　spectral　tests　confirm　the　efficacy　of　this　method　in　minimizing　transmission　loss　while　achieving　over-30-dB　intensity　modulation　during　GST　transitions　from　amorphous　to　crystalline　states.　Conclusions　We　investigate　the　effect　of　different　coverage　configurations　of　GST　on　the　optical　transmission　loss　in　silicon　ridge　waveguides.　The　study　reveals　that　covering　only　the　top　surface　of　the　waveguide　outperforms　the　traditional　full-coverage　approach　in　reducing　transmission　loss　while　maintaining　excellent　modulation　performance　during　phase　transitions.　By　analyzing　the　resonant　spectral　shift　and　peak　intensity　changes　in　a　racetrack　micro-ring　resonator,　we　compare　the　modulation　effects　of　the　GST　material　on　the　effective　refractive　index　and　transmission　loss　between　straight　and　bent　waveguides.　Based　on　simulation　results,　we　propose　and　experimentally　validate　an　optimized　method　for　covering　phase-change　materials　on　waveguides.　This　method　involves　filling　the　waveguide　trench　with　SiO2　and　then　depositing　the　phase-change　material　on　the　plane　formed　by　the　SiO2　and　the　top　surface　of　the　waveguide,　which　reduces　the　complexity　of　device　fabrication　and　improves　the　uniformity　of　phase-change　material　coverage.　Experimental　results　demonstrate　that,　compared　to　the　full-coverage　approach,　this　method　effectively　reduces　transmission　loss　after　GST　material　attachment.　Additionally,　the　transmission　spectra　exhibit　strong　intensity　modulation　when　the　GST　material　transitions　from　the　amorphous　to　the　crystalline　state.　This　method　provides　a　new　approach　for　studying　low-loss　modulation　devices　based　on　phase-change　materials.　©　2025　Chinese　Optical　Society.　All　rights　reserved.