The H2O layer thickness on flat hydroxylated SiO2 surfaces was measured at various H2O pressures and substrate temperatures using in situ real-time spectroscopic ellipsometry (SE). The in situ SE measurements were conducted at 18.1, 27.2, and 30.4 degrees C (291.25, 300.35, and 303.55 K) in a warm-wall vacuum chamber designed with a cooled sample stage. The H2O pressures were varied up to the saturation H2O vapor pressures of 15.6, 27.0, and 32.5 Torr at 18.1, 27.2, and 30.4 degrees C, respectively. The SE measurements showed that there were two distinct types of H2O layers on the hydroxylated SiO2 surface: a thin strongly adsorbed layer and a weakly adsorbed layer that was much thicker at high H2O pressures. The strongly adsorbed layer had thicknesses <= 1.2 & Aring; and was not lost by removing the H2O pressure. The strongly adsorbed layer could be desorbed by heating the sample stage to 124 degrees C (397.15 K). The stability of the strongly adsorbed layer was consistent with an adsorption energy of >20 kcal/mol. In contrast, the weakly adsorbed layer could be added or removed by increasing or decreasing the H2O pressure. The weakly adsorbed layer obtained much higher multilayer H2O thicknesses at larger H2O pressures. For example, the weakly adsorbed multilayer thickness was 7.5 & Aring; at 92% relative humidity at 30.4 degrees C (32.5 Torr). Complementary in situ Fourier transform infrared (FTIR) studies were also performed on hydroxylated SiO2 nanoparticles that were in qualitative agreement with the SE results.

Selectivity was examined between SiO2 and SiNx during thermal atomic layer etching (ALE) and spontaneous etching. Thermal ALE of SiO2 and SiNx was explored using sequential trimethylaluminum (TMA) and hydrogen fluoride (HF) with reactant exposures of 3 Torr for 45 s at 275 degrees C. SiO2 thermal ALE achieved an etch per cycle (EPC) of 0.20 & Aring;/cycle and near-ideal synergy up to 95%. SiNx thermal ALE exhibited a higher EPC of 1.06 & Aring;/cycle. The selectivity factor was similar to 5:1 for SiNx etching compared to SiO2 etching (preferential SiNx removal) during thermal ALE using TMA and HF. Spontaneous etching was then quantified using repeated exposures of HF vapor alone at 3 Torr and 275 degrees C. SiO2 spontaneous etching was minor at an etch rate of 0.03 & Aring;/min, enabling near-ideal synergy for SiO2 thermal ALE. In contrast, major SiNx spontaneous etching displayed an etch rate of 1.72 & Aring;/min and predominated over SiNx thermal ALE. The selectivity factor was similar to 50:1 for SiNx spontaneous etching compared to SiO2 spontaneous etching using an HF pressure of 3 Torr. This selective SiNx spontaneous etching was attributed to F- surface species during HF exposures. NH3 codosing with HF was then examined during thermal ALE and spontaneous etching. Thermal ALE of SiO2 and SiNx was examined using sequential TMA and HF + NH3 codosing with reactant exposures of 3 Torr for 45 s at 275 degrees C. SiO2 thermal ALE with HF + NH3 codosing had a high EPC of 8.83 & Aring;/cycle. In contrast, SiNx thermal ALE with HF + NH3 codosing was negligible. The selectivity factor was reversed and much higher at >1000:1 for SiO2 etching compared to SiNx etching (preferential SiO2 removal) during thermal ALE with HF + NH3 codosing. Rapid SiO2 spontaneous etching with HF + NH3 codosing at 3 Torr had an etch rate of 27.50 & Aring;/min. In contrast, SiNx spontaneous etching with HF + NH3 codosing produced a very low etch rate of 0.02 & Aring;/min. The selectivity factor was >1000:1 for SiO2 spontaneous etching compared to SiNx spontaneous etching with HF + NH3 codosing. This selective SiO2 spontaneous etching was attributed to HF2- surface species during HF + NH3 exposures. These studies revealed that the NH3 coadsorbate during HF exposures modified the active etch species and dramatically influenced the etch selectivity between SiO2 and SiNx. Reciprocal etch selectivity should be important for the selective removal of SiO2 or SiNx in composite structures.

Ruthenium (Ru) thin films were deposited utilizing electron-enhanced atomic layer deposition (EE-ALD). Sequential exposures of Ru(DMBD)(CO)3 (DMBD = 2,3-dimethylbutadiene) and low-energy electrons at ∼125 eV were used to grow the Ru films at temperatures ≤160 °C. The electrons were obtained from a hollow cathode plasma electron source that provided an electron current of ∼200 mA over a surface area of ∼4 cm2. Low-energy electrons can desorb surface ligands derived from Ru(DMBD)(CO)3, such as CO, through electron-stimulated desorption. The desorbed surface ligands leave chemically reactive sites for subsequent Ru(DMBD)(CO)3 precursor absorption. Ru EE-ALD film growth was monitored utilizing in situ spectroscopic ellipsometry (SE). The electron exposures resulted in rapid Ru film nucleation and growth. Under saturation conditions at 160 °C, the growth rate for Ru EE-ALD was 0.2 Å/cycle. The electron efficiency factor for Ru EE-ALD was ∼21 500 electrons/deposited Ru atom. There was no film growth without electron exposures. Ru growth was observed on various substrates including silicon with native oxide and titanium. Ru growth was also obtained on insulating substrates such as 400 nm thick thermal SiO2 substrates. XPS analysis measured <1 at. % oxygen in the deposited Ru films. XRD, x-ray reflectivity, and SE were used to characterize the Ru films before and after forming gas anneal (FGA). FGA successfully removed carbon impurities from the as-deposited Ru films. The resistivity of the Ru EE-ALD films after FGA was determined to be as low as 17 μΩ cm for a film thickness of 6.7 nm. SE measurements of the imaginary part of the pseudodielectric function, 〈ɛ2〉, were utilized to characterize the as-deposited Ru films and the high purity Ru films after FGA. The low resistivity of the Ru films after FGA was consistent with a prominent Drude absorption in the ⟨ε2⟩ spectrum at ≤1 eV. Various reactive background gases such as H2, NH3, and H2O were utilized during EE-ALD to attempt to remove the carbon from the as-deposited Ru EE-ALD films.

A survey of known gold-containing chemical vapour deposition (CVD) and atomic layer deposition (ALD) precursors, with a focus on collecting their volatilization and decomposition data. These data were applied to a figure of merit (σ) developed to easily assess the thermal characteristics.

In microelectronic or nanoelectronic manufacturing, pattern transfer by directional reactive ion etching (RIE) progressively erodes amorphous carbon (aC) hard masks. To maintain critical dimensions and tolerances of high-aspect-ratio device structures, new carbonaceous materials may be added repeatedly to replace the eroded aC hard mask. Such a mask repairing step during RIE needs self-aligning growth of organic materials. Area selectivity is required to deposit the organic material on the aC hard mask exclusively. Deposition on the dielectric or semiconductor device structures underlying the mask would complicate their precise etching or later cleaning. When ashing the aC hard mask, all-organic materials are preferable to organic-inorganic hybrid materials because they leave no residue. In this work, area-selective molecular layer deposition (MLD) was developed for the all-organic polyamide nylon 6,2. The monomer reactants for nylon 6,2 MLD were ethylene diamine and adipoyl chloride. Nylon 6,2 MLD was studied in the homogeneous, steady-state growth regime and during nucleation on various starting surfaces utilizing in situ spectroscopic ellipsometry. Area-selective MLD of nylon 6,2 was achieved on the “growth” carbon surface in the presence of silica by functionalizing aC via mild oxidation. In addition, a surface passivant was selectively attached to silica by using an amine-catalyzed coupling chemistry. The passivant inhibited the nylon 6,2 MLD on the “nongrowth” silica surface. A single passivation pretreatment was sufficient to restrict the MLD on the silica surface. The passivant, however, did not substantially impact the MLD nucleation and growth on the aC surface. This strategy yielded area selectivity with exceptionally high quality and over a wide range of MLD cycles. The area-selective MLD of nylon 6,2 was further applied on industrial test features with aC patterns masking trenches in silica. This demonstration illustrated the capability of area-selective MLD to repair RIE-eroded aC hard masks and to maintain the critical dimension.

The copper electrochemical deposition (Cu-ECD) filling capability of high aspect ratio through silicon vias (HAR-TSVs) and homogeneity over 300 mm wafers were investigated on a film stack of thermal ALD (thALD) TaxNy barrier with thermal ALD Ru seed in comparison to TixNy barrier with a standard Cu i-PVD seed layer using a commercial 300 mm plating tool. As a first step, Cu-ECD was conducted on wafers with TSV blind holes with aspect ratios (AR) of 10 to 12. To achieve this, a thermal ALD film stack of approximately 6 nm TaxNy and 9 nm Ru (with a sheet resistance of [25.6 +/- 1.4] Omega/Upsilon) were deposited at 250 degrees C. The reactants for the barrier layer were (tert-butylimido)tris(diethylamino)tantalum(V) (TBTDET) and ammonia (NH3) as co-reactant. For the Ru seed layer deposition (ethylcyclopentadienyl)(pyrrolyl)ruthenium(II) (ECPR) and molecular oxygen as co-reactant were used supplemented by a hydrogen purge step after every third ALD cycle. The corresponding ALD growth was observed during the entire process by in-situ real-time spectroscopic ellipsometry (irisE). Blister-free deposition and satisfactory film stack adhesion with no delamination was verified ex situ by scanning electron microscopy (SEM). The deposited copper inside the TSVs was analyzed by focused ion beam (FIB) imaging and X-ray tomography. The Cu ECD filling capability in HAR-TSVs was shown on a film stack of thALD TaxNy thALD Ru seed using a commercial industry standard 300 mm plating tool. A novel blister-free ultra-thin Ru ALD film having good adhesion properties and unique advantages, e. g. high conformity in high-aspect-ratio through-silicon vias large-scale film uniformity over 300 mm wafers, as well as good reproducibility was developed.

Nanoimprint lithography (NIL) has the potential for low-cost and high-throughput nanoscale fabrication. However, the NIL quality and resolution are usually limited by the shape and size of the nanoimprint stamp features. Atomic layer etching (ALE) can provide a damage-free pattern transfer with ultimate etch control for features of all length scales, down to the atomic scale, and for all feature geometries, which is required for good quality and high-resolution nanoimprint stamp fabrication. Here, we present an ALE process for nanoscale pattern transfer and high-resolution nanoimprint stamp preparation. This ALE process is based on chemical adsorption of a monoatomic layer of dichloride (Cl2) on the silicon surface, followed by the removal of a monolayer of Cl2-modified silicon by argon bombardment. The nanopatterns of different geometries, loadings, and pitches were fabricated by electron beam lithography on a silicon wafer, and ALE was subsequently performed for pattern transfer using a resist as an etch mask. The post-ALE patterns allowed us to study the different effects and limitations of the process, such as trenching and sidewall tapering. The ALE-processed silicon wafers were used as hard nanoimprint stamps in a thermal nanoimprint process. Features as small as 30 nm were successfully transferred into a poly(methyl methacrylate) layer, which demonstrated the great potential of ALE in fabricating nanoimprint stamps with ultrahigh resolution.

Area selectivity is an emerging sub-topic in the field of atomic layer deposition (ALD), which employs opposite nucleation phenomena to distinct heterogeneous starting materials on a surface. In this paper, we intend to grow Ru exclusively on locally pre-defined Pt patterns, while keeping a SiO2 substratum free from any deposition. In a first step, we study in detail the Ru ALD nucleation on SiO2 and clarify the impact of the set-point temperature. An initial incubation period with actually no growth was revealed before a formation of minor, isolated RuOx islands; clearly no continuous Ru layer formed on SiO2. A lower temperature was beneficial in facilitating a longer incubation and consequently a wider window for (inherent) selectivity. In a second step, we write C-rich Pt micro-patterns on SiO2 by focused electron-beam-induced deposition (FEBID), varying the number of FEBID scans at two electron beam acceleration voltages. Subsequently, the localized Pt(C) deposits are pre-cleaned in O-2 and overgrown by Ru ALD. Already sub-nanometer-thin Pt(C) patterns, which were supposedly purified into some form of Pt(Ox), acted as very effective activation for the locally restricted, thus area-selective ALD growth of a pure, continuous Ru covering, whereas the SiO2 substratum sufficiently inhibited towards no growth. FEBID at lower electron energy reduced unwanted stray deposition and achieved well-resolved pattern features. We access the nucleation phenomena by utilizing a hybrid metrology approach, which uniquely combines in-situ real-time spectroscopic ellipsometry, in-vacuo x-ray photoelectron spectroscopy, ex-situ high-resolution scanning electron microscopy, and mapping energy-dispersive x-ray spectroscopy.

Graphene based electron field emitter arrays consisting of cone-shaped silicon tips, a thin Al2O3 tunnel barrier, and graphene top electrode are fabricated. Due to the monolayered graphene top electrode, the electrons are able to tunnel through the Al2O3 layer and emit into the vacuum. The temperature behavior of the tunnel leakage current as well as the emission current is characterized.

Graphene has been considered for a variety of applications including novel nanoelectronic device concepts. However, the deposition of ultra-thin high-k dielectrics on top of graphene has still been challenging due to graphene's lack of dangling bonds. The formation of large islands and leaky films has been observed resulting from a much delayed growth initiation. In order to address this issue, we tested a pre-treatment with NF3 instead of XeF2 on CVD graphene as well as epitaxial graphene monolayers prior to the Atomic Layer Deposition (ALD) of Al2O3. All experiments were conducted in vacuo; i. e. the pristine graphene samples were exposed to NF3 in the same reactor immediately before applying 30 (TMA-H2O) ALD cycles and the samples were transferred between the ALD reactor and a surface analysis unit under high vacuum conditions. The ALD growth initiation was observed by in-situ real-time Spectroscopic Ellipsometry (irtSE) with a sampling rate above 1 Hz. The total amount of Al2O3 material deposited by the applied 30 ALD cycles was cross-checked by in-vacuo X-ray Photoelectron Spectroscopy (XPS). The Al2O3 morphology was determined by Atomic Force Microscopy (AFM). The presence of graphene and its defect status was examined by in-vacuo XPS and Raman Spectroscopy before and after the coating procedure, respectively.

The paper presents atomic layer deposition (ALD) processes and process integration for the deposition of insulating liners, copper diffusion barriers, and seed layers for direct copper plating in high aspect ratio (>20:1) through silicon vias. A TaN-based copper diffusion barrier was deployed on an aluminum oxide insulating liner. The latter has the potential to act also as a dielectric barrier against copper diffusion according to BTS and TVS measurements. Furthermore, ruthenium ALD films applied as seed layers for direct copper plating were deposited with an intermediate annealing step to improve film adhesion and ensure the deposition of thick films without any delamination. The step coverage of the presented ALD processes was confirmed by SEM measurements on cross-sections of coated TSV structures. Finally a subsequent electrochemical deposition (ECD) of copper was conducted revealing the satisfying functionality of the Ru seed layer.

The initial growth behavior of ruthenium during the thermal-activated atomic layer deposition (ALD) using [(ethylcyclopentadienyl)(pyrrolyl)ruthenium(II)] (ECPR) and molecular oxygen was investigated in a cluster tool combining an ALD reactor with a surface analysis unit under high vacuum conditions. A direct qualification and quantification of the chemical surface composition by X-ray photoelectron spectroscopy (XPS) and a determination of the surface topography by atomic force microscopy (AFM) were conducted in the course of the ALD cycles without vacuum break. XPS revealed a substrate-inhibited Ru growth on a hydrogen-terminated silicon surface, which was preceded by an incubation period of 20ALD cycles. The Si surface oxidized during the first 50cycles. AFM measurements showed a roughness maximum around the 40th ALD cycle, which suggested an island growth mode and thus corresponded with the substrate inhibition. As verified from the AFM data with an analytical model by Nilsen et al. the Ru islands coalesced between the 40th and 50th ALD cycle. The ALD growth initiation of Ru was also investigated on aluminum oxide and tantalum nitride. XPS revealed a similarly inhibited growth behavior on all the investigated substrates. However, the ECPR adsorption during the very first Ru precursor pulse differed as the amount of chemisorbed Ru on the NH"y-terminated TaN"x(O, C) surface was much higher compared to the OH-terminated Al"2O"3 and the H-terminated Si surface. Summarizing, we demonstrated that in-vacuo XPS and AFM as well as the combination of both are ideally suited for studying the ALD growth initiation of Ru. Furthermore, we provided important chemical information about the initial Ru precursor adsorption on several foreign substrate materials, which will direct further investigations towards a non-inhibited Ru growth.

In this work, a process for the thermal activated atomic layer deposition (ALD) of ruthenium from the organometallic heteroleptic precursor [(ethylcyclopentadienyl)(pyrrolyl)ruthenium] with molecular oxygen was developed and characterized. Silicon substrates were precleaned in hydrofluoric acid and preheated to a specific temperature before coating with ruthenium. The corresponding cycle-by-cycle growth was monitored throughout the entire ALD process time, utilizing an in-situ real-time spectroscopic ellipsometer. Transmission electron microscopy and atomic force microscopy were applied at a reference sample to generate an appropriate optical model for the translation of the ellipsometric spectra into Ru film thicknesses. Given a representative set of process parameters the cycle-by-cycle growth was studied in detail, obtaining information about incubation, nucleation, linear growth and delamination. In order to determine the ALD characteristic dependencies, the following process parameters were varied while applying ellipsometry during the linear film growth regime on as-deposited ruthenium film surfaces; thus excluding effects from the initial foreign substrate material: both reactant doses and purging times, the substrate temperature and the total pressure. During the respective film growth experiments, one process parameter-setting was changed each 15 ALD cycles, which enabled a fast and extensive process development.

The diffusion barrier properties of PVD Ru and PECVD / PEALD Ru-C films, deposited by RuEtcp2 precursor and N2/H2 plasma, were compared on the basis of bias temperature stress measurements. An MIS test structure was used to distinguish between thermal diffusion induced by annealing and a Cu field drift due to applied electric fields. BTS-CV, TZDB and TDDB measurements revealed that the barrier performance is significantly better for PEALD and PECVD Ru-C films. This improvement is associated with carbon impurities in the Ru films with a concentration in the order of several percent according to ToF-SIMS and ERDA. The TDDB mean time to failure at 250°C, +5MV/cm was 7s for PVD Ru samples, ≈500s for PECVD Ru-C, ≈800s for PEALD Ru-C and>3600s for PVD TaN. Triangular voltage sweep measurements at 300°C, 0.1V/s confirmed the presence of Cu ions inside the SiO2 for degraded dots, in contrast to the Al reference sample and to PVD TaN, which performed best among all the Cu barriers under test. XRD data suggests that PEALD and PECVD Ru-C films are only weakly crystalline.

PECVD and PEALD of ruthenium films using RuEtcp(2) as a precursor and N-2/H-2/Ar plasma as a reducing agent were characterized. A self-adjusting process to overcome the previously reported inhibition of Ru PEALD on TaN substrates was investigated. Ellipsometric modelling of Ru films was demonstrated providing information on both film thickness and estimated Ru content. The physical properties of PECVD/PEALD Ru films were compared to characteristics of sputtered Ru films within the categories resistivity, impurites, crystal structure, conformity and Cu plating. As a result, ToFSIMS, ERDA and 3D atomprobe revealed the presence of carbon impurities in PECVD and PEALD Ru films, dependent on deposition temperature and plasma power. Nevertheless, highly conductive Ru-C films were produced via PECVD and PEALD achieving resistivities equal to PVD Ru. For all types of Ru films, the size effect played a significant role at thicknesses below 10 nm; Cu plating and crystallization behaviour appeared similar. Direct Cu fill potential of different Ru films was discussed for damascene structures and through silicon vias. (C) 2011 The Electrochemical Society. [DOI: 10.1149/2.066202jes] All rights reserved.

Depositing ultra-thin metallic films with an accurate control of the film properties (like film thickness, surface roughness and electrical properties) both in the initial and in the progressed film growth regime is a critical challenge. Monitoring atomic layer deposition processes by spectroscopic ellipsometry allows film thickness control in the sub-nanometer range. In addition, ellipsometry serves as a powerful technique for process analysis as it covers many relevant issues, like the evaluation of substrate temperatures as well as the quantification of film properties during the entire ALD process (i. e. in all relevant growth regimes).