Ultra Low-k Research Articles

Effects of Cu surface roughness on TDDB for direct polishing ultra-low k dielectric Cu interconnects at 40nm technology node and beyond

The time-dependent dielectric breakdown (TDDB) reliability performance of back-end-of-line (BEOL) structures has been demonstrated to obviously correlate with the surface roughness of Cu metal line. Larger Cu surface roughness could induce the cracking of metal capping layer and the degradation of TDDB. More ultra-low k (ULK) polishing amount and higher de-ionized water (DIW) dilute ratio post-Cu chemical mechanical polishing (Cu-CMP) chemical cleaning process have been developed to reduce Cu surface roughness for better TDDB performance read-out. An optimized post-cleaning process with adjusted brush gap condition has been demonstrated to eliminate microscratch level with the improvements of the yield of Cu wide line serpent open test structures and TDDB performance at 28nm technology node.

Synergistic combinations of dielectrics and metallization process technology to achieve 22nm interconnect performance targets

An improvement in interconnect performance implies a reduction of the resistance-capacitance (RC) time constant. Instead of scaling the capacitance at each technology node through a reduction in the dielectric constant of the interlayer dielectric (ILD), the interconnect aspect ratios could be scaled holding the ILD fixed. In this case, the material properties of the ILD must be robust to process-induced damage and amenable to the creation of high aspect ratio features. In addition, a metallization scheme that can provide void free Cu fill in high aspect ratio features is required. Characterization, patterning, and integration results collected on such an ultra-low-k (ULK) ILD material and void free metallization is presented. A measured reduction in the resistance of a 22nm node interconnect in this ILD was observed as a function of increasing aspect ratio. The copper seed deposition process, capable of enabling the fill of even higher aspect ratio features, is also discussed.

Advanced Direct-Polish Process on Organic Non-Porous Ultra Low-k Fluorocarbon Dielectric on Cu Interconnects

A direct-polish process has applied to the ultra low-k dielectric without the etch stop layer to reduce effective dielectric constant in damascene interconnects. In this study, the direct-polish process on organic non-porous dielectric, fluorocarbon (k=2.2), was demonstrated and an optimum direct-polish process condition was investigated. Mechanical effect, not chemical effect, on fluorocarbon degraded electrical properties by changing of chemical structure of fluorocarbon. A surface plasma treatment of fluorocarbon before polishing was applied to avoid the degradation of electrical characteristics during direct-polish and this result revealed that the surface plasma treatment of fluorocarbon is a practical technique in advanced Cu interconnects.

Development of Porosimetry Techniques for the Characterization of Plasma-Treated Porous Ultra Low-K Materials

not Available.

Profiling of the Mechanical Properties of Ultralow-k Films Using Nanoindentation Techniques

ABSTRACTSeveral nanoindentation techniques were applied to the surface, the reverse side and cross-sections of PECVD ultralow-k (ULK) film stacks to characterize their elasto-plastic properties quantitatively. Results showed good agreement of the reduced modulus (Er) values measured from above and on cross-sections, respectively. Er decreased by 10-22% from the upper to the lower surface of the films. This gradient suggests that UV light absorption inside the film leads to slightly reduced curing at the rear side of the films compared to the surface of the ULK layers. Both quasi-static nanoindentation and dynamic mechanical mapping showed this trend. It is demonstrated that quantitative mechanical mapping can be performed with a lateral resolution ≤ 100nm. Slight local variations of Er were detected on ULK/SiCxNy films stacked on top of Cu-low-k interconnect structures.

Fine Pitch Wirebonds on Ultra Low-k Device

The mechanical integrity of wirebonds are sensitive to structures under the bond pads of ultra low-k dielectric devices. The authors studied the mechanical performance of wirebonds on 32-nm test chips with various layouts of lines and vias under the 35-μm-pitch bond pads, various stacks of dielectric layers, and a range of bonding process conditions. Poor mechanical integrity resulted in the pad tearout failure mode at wire pull testing. The thickness of the SiO2/FTEOS layer and the density of the vias in the ULK layer are the key factors for good wirebond integrity, manufacturability, and module-level reliability with ultrafine pitch wirebonds. The wirebond experimental results were analyzed by capillary indentation, excessive bond force parameter and finite element method (FEM). The authors found that only excessive bond force does not increase the rate of pad tearout and 2D FEM model can correlate the location of pad tearout in the ULK layers with calculated highest stress point which occurs by the ultrasonic vibration during wirebonding process.

Ultra Low-k Materials Based on Self-Assembled Organic Polymers

ABSTRACTThe material properties of two ultra low-k organic polymers are characterized for copper interconnect integration. The k-values are 2.2-2.3 for both. Compared to OSG materials of similar k-values, these polymers have lower porosity and smaller pore size, achieved using selfassembled chemistry. Both materials demonstrate excellent resistance to plasma damage: no water uptake was detected after exposure to selected etching plasmas. This characteristic, combined with the small pore size and low porosity, results in the successful integration of the organic low-ks in 80 nm spacing with no significant increase in the integrated k-values.It is found that higher open porosity in polymer A is accompanied by higher leakage current, which is not however linked to lower dielectric breakdown lifetimes.

A Less Damaging Patterning Regime for a Successful Integration of Ultra Low-k Materials in Modern Nanoelectronic Devices

ABSTRACTThis paper describes a three-step process regime for the integration of porous SiCOH based ultra low-k materials in existing copper damascene technologies. During the work with these complex and sensitive materials, it became more and more clear, that a successful patterning depends not only on the etch step but also on the adjustment between the etch and the following cleaning and k-restore processes. The presented process regime starts with a reactive ion etch process for trench patterning followed by a post etch clean to remove etch residues. Finally a k-restore process was performed to repair the damaged regions in the trench sidewalls. In this work it became clear, that the etch chemistry influences not only the results of the etch process ostensibly sidewall damage but also kind and effect of the post etch clean. Each plasma composition results in the necessity of a customized post etch cleaning solution. Finally a k-restore process using Hexamethyldisilazane (HMDS) as restore chemical was demonstrated successfully. Enhanced temperatures and an additional UV-treatment are possibilities to promote the restore effect.

Ultra low-k die crack study for lead free solder bump flip-chip packaging

Ultra low-K (ULK) dielectric has lower mechanical strength (E < 8 GPa), lower cohesive strength and lower adhesion (<5 J/m2) than low-K and SiO2 dielectric material. The packaging reliability test has shown that delamination between copper (Cu) and ULK is a major concern. In addition to the Cu and ULK delamination issue, the ULK die crack after temperature cycling test (TCT) showed die crack failure to be another issue. ULK die crack failure can be detected by C-mode scanning acoustic microscopy (CSAM). The CSAM image of the die crack mostly shows a crescent moon shape. The crack initiates at the upper edge of the underfill fillet penetrating the sidewall of the die, and then propagates toward the inside of the die. Finite element simulation indicates that the die crack failure starts at the backside edge of the die. The ULK die crack is caused by two mechanisms. First is the bending stress at the backside of the die, which is the result of the coefficient of thermal expansion (CTE) mismatch between the silicon die and the organic substrate. Second is the thermo-mechanical stress, which is the result of the local CTE mismatch between the silicon die and the underfill. The finite element simulation parametric study performed in this paper shows that a low underfill fillet height and a small fillet tip angle reduce the thermo-mechanical stress. In addition, attaching a heat sink to the surface of the flip-chip die increases packaging stiffness and resists the bending stress induced by the shrinkage of the substrate. Experimental results demonstrate that lowering the height of the underfill fillet, reducing the angle of the fillet tip and attaching a heat sink to the flip-chip die are effective ways to solve the ULK die crack issue.

Surface and near-surface modifications of ultralow dielectric constant materials exposed to plasmas under sidewall-like conditions

The authors describe the temporal evolution of the surface and near-surface regions of a porous SiCOH ultralow k (ULK) dielectric during exposure under sidewall-like exposure conditions to various plasma processing environments. The authors studied the exposure of the ULK material to Ar plasma, C4F8/Ar-based etching plasma, and O2 or CO2 ashing plasmas, as well as various sequences of these processes. Real-time monitoring of the ULK surfaces during plasma processing was performed by in situ ellipsometry employing a novel gap structure. Additionally, changes in ULK surface properties were characterized by x-ray photoelectron spectroscopy and selective dilute hydrofluoric acid wet etching in combination with ex situ ellipsometry measurements. Pristine ULK material exposed to O2 plasma without ion bombardment shows the formation of a near-surface porous layer. For exposure of the ULK to CO2 plasma operated at comparable plasma operation conditions, the modification depth for a given exposure time is reduced relative to O2, but otherwise an identical ellipsometric trajectory is followed. This is indicative of a similar ULK damage mechanism for the two discharges, although at different rates. Energetic (∼400 eV) ion bombardment on the surface of ULK with line-of-sight Ar plasma exposure introduced a ∼12 nm thick SiO2-like densified layer on the ULK surface meanwhile sputtering off the ULK material. The sidewall-like modifications of ULK due to metastable Ar, if present, were too subtle to be measurable in this article. For ULK exposed under sidewall-like geometry to C4F8/Ar-based etching plasma, fluorocarbon quickly permeated into the subsurface region and showed saturation at a mixed layer thickness of about 14 nm. For additional exposure to O2 or CO2 discharges, a strong decrease of the CO2 plasma induced ULK surface modifications with increasing fluorocarbon (FC) film thickness was found, indicative of surface protection by FC surface deposition along with pore-sealing by the FC material. Attempts to increase the protective nature of the FC film by additional plasma processing, e.g., by exposure to Ar or He plasma after FC plasma etching, did not reduce CO2 plasma induced ULK surface modifications further.

Ultra Low-k Materials: Challenges of Scaling

High porosity and carbon doing are required to achieve low dielectric constant, k values for ultra low-k (ULK) materials based on silicon dioxide. Those characteristics create new challenges in the integration of ULK materials that are being developed for future interconnect scaling. Four of those challenges based on electrical characterizations of ULK films will be presented in this paper. First, a high concentration of porogen residue is typically present in ULK materials fabricated by Plasma Enhanced Chemical Vapor Deposition (PECVD). The high porogen residue concentration leads to a high leakage current density. Second, ULK materials based on Spin-on Glass (SOG) process demonstrate a low leakage current density. But early dielectric breakdown has been observed in SOG k=2.2 material. Third, PVD TaNTa barrier process can introduce significant damage to ULK materials. The damage level increases as the k value decreases or porosity increases. Four, direct deposition of ALD Ru barrier on ULK films results in a high leakage current density.

Interactions of Metal-Organic PEALD TaN with Ultra-Low k Dielectric Materials

Both dense and porous patterned ultra-low k (ULK) substrates have been characterized in terms of their interaction with TaN grown by plasma enhanced atomic layer deposition (PEALD). The conformal nature of the deposited TaN films by PEALD was demonstrated on a patterned dense oxide substrate. For the case of ULK dielectric materials with a dielectric constant of 2.55 (ULK2.55), penetration of TaN into the ULK matrix was observed for porous ULK2.55 while dense ULK2.55 did not exhibit any significant penetration of TaN. Both hydrogen plasma-based PEALD TaN and ammonia plasma-based PEALD TaN processes were shown to display a similar interaction with ULK2.2. A dielectric pore-sealing layer applied on patterned ULK2.2, prior to the deposition of TaN, enabled a significant reduction in PEALD TaN penetration as observed at ULK sidewalls.

The TDDB Study of Post-CMP Cleaning Effect for L40 Direct Polished Porous Low K Dielectrics Cu Interconnect

In order to realize the high performance of RC delay, direct polished porous type ultra low-K film (ULK) is integrated into Cu dual-damascene interconnects for 45 nm and beyond. In the past, metal line seems not an important role in the world of TDDB. However, because of the minimized feature size and the relative fragile ULK film properties, it was found that the Cu roughness of polished surface also plays an important role to effect on the reliability such as TDDB from this generation. Post-cleaning of Cu CMP is a major process to manage the Cu surface roughness. In this paper, the correlation for post-cleaning time to the TDDB was discussed. The mechanism with two models was built up to explain this behavior as well.

Mechanistic study of ultralow k-compatible carbon dioxide in situ photoresist ashing processes. II. Interaction with preceding fluorocarbon plasma ultralow k etching processes

Process interactions between fluorocarbon (FC) plasma ultralow k (ULK) dielectrics etching followed by carbon dioxide (CO2) in situ photoresist (PR) ashing on ULK damage have been studied in a dual frequency, capacitively coupled plasma reactor. Introduction of ULK trench/via sidewall damage was simulated employing blanket ULK films by exposing them in a non-line-of-sight fashion in a small gap structure to the plasma environment. ULK damage was quantified using the dilute hydrofluoric acid (0.5%/15 s) selective etching method. CO2 in situ ashing processes showed a chamber memory effect due to prior FC plasma etching, significantly increasing damage of pristine ULK films. For ULK plasma etching/PR ashing process sequences, ULK material surfaces were modified by FC plasma etching prior to the CO2 plasma exposure. X-ray photoelectron spectroscopy studies showed that the modifications consisted primarily of 1–2 nm FC coverage of the ULK. This FC deposit remained on the ULK surfaces during in situ CO2 processing and provided protection of the underlying ULK material. PR-patterned ULK structures were also processed employing the same processing conditions. The results obtained from the characterization of the resulting trench structures support the findings obtained with blanket films. CO2 in situ PR ashing processes performed at low pressure (10 mTorr) and enhanced by rf biasing provided a good combination of high PR stripping rate and low ULK damage introduction.

Mechanistic study of ultralow k-compatible carbon dioxide in situ photoresist ashing processes. I. Process performance and influence on ULK material modification

In situ photoresist (PR) ashing processes are attractive because of the ease of process integration with plasma etching processes. The authors have examined the performance of carbon dioxide (CO2) as a source gas for in situ PR ashing processes compatible with ultralow k (ULK) materials and compared it with the results obtained using O2. They performed measurements of 193 nm PR ashing rates in a dual frequency capacitively coupled plasma reactor. The damage to porous ULK feature sidewalls was simulated by exposing blanket ULK films in a non-line-of-sight fashion in a small gap structure to the plasma-generated reactants. The pressure for the in situ ashing processes was varied from 10 to 100 mTorr, and the self-bias voltages ranged from floating potential to ∼−400 V. To increase line-of-sight etching of PR by inert ion bombardment, Ar/CO2 mixtures with up to 75% Ar were investigated. The ULK material modifications were analyzed by x-ray photoelectron spectroscopy and time-of-flight secondary ion mass spectroscopy (ToF-SIMS). Plasma-damage of the ULK material primarily is detected as the removal of carbon from the SiCOH ULK films. To compare the performance of different ashing processes for PR stripping from ULK material, the authors introduced an ashing efficiency (AE) parameter which is defined as the thickness of PR removed over the thickness of ULK simultaneously damaged, and can be considered a process figure of merit. AE with CO2 was about three times greater than AE with O2 for the same process conditions. When a 75% Ar/CO2 gas mixture was used and a −100 V substrate bias was applied during PR ashing, a PR ashing rate of 200 nm/min could be achieved for a 10 mTorr Ar/CO2 plasma. For this process, the measured AE was 230, more than 10× greater than AE achieved with O2 discharges using the same conditions. The authors found that ULK damage showed a direct dependence on the atomic oxygen densities of both CO2 and O2 discharges which was characterized by optical emission of discharges. The question whether in-diffusion of carbon species from CO2 discharges into ULK material was significant was also examined. For this the authors substituted C13O2 for C12O2 and performed ToF-SIMS analysis of the exposed ULK films. No significant amount of C13 from C13O2 plasmas was detected either on the surface or in the bulk of the C13O2 plasma-exposed ULK.

Soft breakdown characteristics of ultralow-k time-dependent dielectric breakdown for advanced complementary metal-oxide semiconductor technologies

During technology development, the study of ultralow-k (ULK) time-dependent dielectric breakdown (TDDB) is important for assuring robust reliability. As the technology advances, the increase in ULK leakage current noise level and reversible current change induced by soft breakdown (SBD) during stress has been observed. In this paper, the physical origin of SBD and reversible breakdown, and its correlation to conventional hard breakdowns (HBDs) were extensively studied. Based on constant voltage stress (CVS) and constant current stress (CCS) results, it was concluded that SBD in ULK is an intrinsic characteristic for ULK material, and all first breakdown events most likely are soft instead of hard. Therefore, a unified understanding of SBD and HBD for low-k TDDB was established. Furthermore, the post-SBD and HBD breakdown conduction characteristics were explored and their impacts on circuit operation were discussed. Based on current limited constant voltage stress studies, it was found that the power dissipation, not the stored energy, determined the severity of ULK dielectric breakdown, and the postbreakdown conduction properties. A percolation-threshold controlled, variable-range-hopping (VRH) model was proposed to explain all postbreakdown aspects of SBD and HBD of ULK material.

Controllable Ultra Low-k by Via-Typed Air Gap with the Better Design Margin for Logic Devices below 45 nm Node

By changing the air gap to a via-typed air gap, the height of the air gap is reduced up to about 50% compared to the height of the trench-typed air gap. The controllable (1 ≤ k < 2.9) ultra low-k is also easily fabricated by adjusting the space of the via-typed air gap. The via-typed air gap makes the design margin better due to its lower height in the dense and narrow lines.

Adsorption mechanisms of fluorocarbon polymers at ultra low- k surfaces

We report ab initio investigations for the adsorption mechanisms of fluorocarbon polymers at ultra low- k (ULK) surfaces. As prototypical example we study a C 2F 4-molecule adsorbed at a passivated SiO 2:CH 3(001) surface. The calculated adsorption energies between 1.37 and 1.95 eV imply a chemisorption bonding mechanism. The detailed analysis of all investigated structures provides a simple rule to make a rough estimate of the stability of adsorbed fluorocarbon polymers. In addition, an adsorption pathway is proposed and used to suggest a possible adsorption mechanism, triggered by an electron transfer between the ULK surface and the C 2F 4 molecule.

Ultra Low-k Materials: Challenges of Scaling

not Available.

Multi-solvent ellipsometric porosimetry analysis of plasma-treated porous SiOCH films

Porous ultra low- k (ULK) dielectrics are used to reduce resistance–capacitance delay for advanced complementary metal oxide semiconductor interconnects. Since the porosity leads to an increased sensitivity of the material to plasma processes (etching and post-etching plasmas), dedicated characterization techniques are needed to assess the ULK properties during its integration. This study shows that ellipsometric porosimetry, employed with an appropriate multi-solvent protocol, can be effectively used to characterize different plasma-treated porous SiOCH (p-SiOCH) films in terms of porosity, pore sealing and hydrophobicity. It was found that after exposure to fluorocarbon, NH 3, H 2, CH 4 or O 2 based plasmas the p-SiOCH surface is modified leading to moisture uptake. According to the plasma, the solvent adsorption is also modified due to a densification of the surface. In that case solvent adsorption measurements were performed in a kinetic mode to quantify the plasma-sealing effect.