ILD 6-1 Deposition - 40nm

ILD 6-1 deposition forms a dense amorphous silicon-based capping layer to seal exposed Cu and Ta-based liner and prevent diffusion and oxidation .

After the Cu CMP and post-CMP clean of the MET6 module, the top surface of the copper interconnects is exposed and highly susceptible to oxidation and

outward metal diffusion . ILD 6-1 functions primarily as a dielectric diffusion barrier (DB) and copper capping layer (CCL) to seal the exposed Cu and Ta-based liner . Because the subsequent steps involve depositing the bulk ILD 6-2 and forming the bond pad cavity, ILD 6-1 must provide a robust hermetic seal against moisture ingress while acting as a reliable etch stop during the eventual bond pad patterning (Engineering Practice). This distinct structural separation between a thin, dense capping layer (ILD 6-1) and a thicker bulk dielectric (ILD 6-2) is necessary to balance diffusion blocking with overall capacitance reduction, directly addressing the fundamental resistance-capacitance (RC) delay constraints of nanoscale interconnects . The deposition is typically performed using Plasma Enhanced Chemical Vapor Deposition (PECVD), which utilizes radiofrequency (RF) energy to dissociate precursor gases into highly reactive plasma species . This plasma-assisted dissociation significantly lowers the reaction activation energy, enabling the growth of dense, amorphous silicon-based thin films at relatively low temperatures to prevent thermal degradation of the underlying device structures . By tuning the RF power and reactant chemistry, a dense amorphous silicon carbon nitride (a-SiNC:H) or oxygen-doped silicon carbide (a-SiCO:H) network can be synthesized . The initial atomic layers must intimately bond with the post-CMP Cu surface without oxidizing it, often requiring a highly controlled ambient or a reductive pre-treatment (such as a reductive N2/H2 plasma) to remove native Cu oxides before the dielectric network polymerizes and crosslinks . The selection of carbon-doped silicon nitrides or carbides for ILD 6-1 is driven by a fundamental thermodynamic and electrical trade-off . Pure silicon nitride provides excellent diffusion blocking and high mechanical hardness, but possesses a prohibitively high dielectric constant that exacerbates parasitic capacitance . Introducing carbon and organic groups, such as via cycloalkyl-containing organosilicon precursors, reduces the material's polarizability and successfully lowers the effective dielectric constant . However, this incorporation of carbon increases film porosity and alters intrinsic stress, which can lead to mechanical instability or cracking during subsequent thermal cycles . Therefore, the PECVD process parameters—such as gas precursor ratio, chamber pressure, and deposition rate—must be directionally optimized to maintain a sufficiently dense Si-network that blocks Cu atoms while preserving acceptable low-k electrical characteristics . Compared to intermediate steps like ILD 2-1, which serves lower-level BEOL where line-to-line capacitance tightly dictates ultra-low-k requirements, ILD 6-1 is situated just below the terminal bond pad module (Engineering Practice). Consequently, mechanical robustness and hermeticity take much higher precedence over extreme k-value scaling at this upper metal layer . Furthermore, because the subsequent bond pad photo and etch processes subject the dielectric stack to significant physical and chemical stress, ILD 6-1 is engineered with higher intrinsic compressive stress via controlled ion bombardment during deposition, ensuring multilayer mechanical stability and preventing stress reversal phenomena .

[High] Copper outward diffusion and dielectric degradation: If the ILD 6-1 film density is insufficient or the carbon content is excessively high, the dielectric barrier fails to block Cu atoms from migrating into the overlying ILD 6-2 under electric field stress . This metal diffusion creates deep trap levels in the dielectric, generating leakage paths and ultimately causing interconnect dielectric breakdown .
[Medium] Dielectric constant (k-value) drift due to moisture ingress: Porosity introduced to lower the k-value can make the film susceptible to adsorbing ambient moisture if the Si-O-C-H network lacks sufficient crosslinking . Because water has a significantly higher dielectric constant, this moisture uptake severely increases polarizability and degrades the RC performance of the MET6 structures .
[Medium] Interfacial delamination and stress-induced cracking: The mismatch in the coefficient of thermal expansion between the thick MET6 Cu lines and the deposited capping layer generates significant interfacial mechanical stress . If the PECVD process fails to tune the intrinsic compressive stress of the a-SiNC:H or a-SiCO:H film adequately, the resulting tensile stress transition during subsequent thermal or UV processing can cause the film to crack or delaminate from the substrate .
[Low] Plasma-induced oxidation of the Cu surface: Excessive high-RF power or the presence of oxidizing species in the initial deposition phase can chemically react with the underlying Cu, forming non-conductive copper oxides . This unintended interfacial reaction increases via contact resistance and degrades the wetting and adhesion between the Cu interconnect and the capping layer, potentially leading to copper agglomeration .

ILD 6-1 deposition forms a dense amorphous silicon-based capping layer to seal exposed Cu and Ta-based liner and prevent diffusion and oxidation .

After the Cu CMP and post-CMP clean of the MET6 module, the top surface of the copper interconnects is exposed and highly susceptible to oxidation and

[High] Copper outward diffusion and dielectric degradation: If the ILD 6-1 film density is insufficient or the carbon content is excessively high, the dielectric barrier fails to block Cu atoms from migrating into the overlying ILD 6-2 under electric field stress . This metal diffusion creates deep trap levels in the dielectric, generating leakage paths and ultimately causing interconnect dielectric breakdown .
[Medium] Dielectric constant (k-value) drift due to moisture ingress: Porosity introduced to lower the k-value can make the film susceptible to adsorbing ambient moisture if the Si-O-C-H network lacks sufficient crosslinking . Because water has a significantly higher dielectric constant, this moisture uptake severely increases polarizability and degrades the RC performance of the MET6 structures .
[Medium] Interfacial delamination and stress-induced cracking: The mismatch in the coefficient of thermal expansion between the thick MET6 Cu lines and the deposited capping layer generates significant interfacial mechanical stress . If the PECVD process fails to tune the intrinsic compressive stress of the a-SiNC:H or a-SiCO:H film adequately, the resulting tensile stress transition during subsequent thermal or UV processing can cause the film to crack or delaminate from the substrate .
[Low] Plasma-induced oxidation of the Cu surface: Excessive high-RF power or the presence of oxidizing species in the initial deposition phase can chemically react with the underlying Cu, forming non-conductive copper oxides . This unintended interfacial reaction increases via contact resistance and degrades the wetting and adhesion between the Cu interconnect and the capping layer, potentially leading to copper agglomeration .

40nm BSI CMOS Image SensorPreview

Process Cross-Section

Step highlight

In depth

Risks & Challenges

40nm BSI CMOS Image SensorPreview

Process Cross-Section

Step highlight

In depth

Risks & Challenges