John R. Abelson

Although it has long been known that metal-containing compounds can serve as catalysts for chemical vapor deposition (CVD) of films from other precursors, we show that metal-containing compounds can also inhibit CVD nucleation or growth. For two precursors A and B with growth onset temperatures TgA < TgB when used independently, it is possible that B can inhibit growth from A when the two precursors are coflowed onto a substrate at a temperature (T) where TgA < T < TgB. Here, we consider three precursors: AlH3⋅NMe3 (Tg = 130 °C, Me = CH3), Hf(BH4)4 (Tg = 170 °C), and AlMe3 (Tg = 300 °C). We find that (i) nucleation of Al from AlH3⋅NMe3 is inhibited by Hf(BH4)4 at 150 °C on two oxide surfaces (Si with native oxide and borosilicate glass), (ii) nucleation and growth of HfB2 is inhibited by AlMe3 at 250 °C on native oxide substrates and on HfB2 nuclei, and (iii) nucleation of Al from AlH3⋅NMe3 is inhibited by AlMe3 at 200 °C on native oxide substrates. Inhibition by Hf(BH4)4 is transient and persists only as long as its coflow is maintained; in contrast, AlMe3 inhibition of HfB2 growth is more permanent and continues after coflow is halted. As a result of nucleation inhibition, AlMe3 coflow enhances selectivity for HfB2 deposition on Au (growth) over Al2O3 (nongrowth) surfaces, and Hf(BH4)4 coflow makes it possible to deposit Al on Al nuclei and not on the surrounding oxide substrate. We propose the following criteria to identify candidate molecules for other precursor–inhibitor combinations: (i) the potential inhibitor should have a higher Tg than the desired film precursor, (ii) the potential inhibitor should be unreactive toward the desired film precursor, and (iii) at the desired growth temperature, the potential inhibitor should adsorb strongly enough to form a saturated monolayer on the intended nongrowth surface at accessible inhibitor pressures.

In the following review, the authors describe how the kinetics of film growth can be controlled in innovative ways to achieve the deposition of conformal, superconformal, area-selective, and ultrasmooth films by low-temperature CVD. Illustrative experimental data and kinetic models are given for the growth of a variety of materials, including transition metals, metal oxides, and metal diborides. Key results are as follows: (i) For precursors that exhibit site blocking on the growth surface, CVD can afford a high degree of conformal coverage on very high aspect ratio features. (ii) The problem of conformal coverage has been solved analytically and presented as a conformal zone diagram. (iii) “Nonconformal” precursor molecules can be made to afford highly conformal films by adding a suitable neutral molecule to enhance surface site blocking. (iv) An inhibitor that adsorbs strongly to the growth material, but not to the substrate surface, can be used to moderate the size distribution of nuclei such that the film is ultrasmooth at coalescence. (v) An inhibitor that binds preferentially to hydroxyl sites on an oxide surface can be used to completely suppress film nucleation and afford area-selective growth. (vi) Superconformal growth, which affords a V-shaped coating and complete fill of a deep trench, can be achieved using one of three approaches, depending on the precursor chemistry. (vii) The science and technology of CVD can further be expanded by the chemical design of new precursors that have a favorable combination of high partial pressure, suitable reactivity at low substrate temperature, and ligand groups that desorb cleanly.

The authors describe the enhancement of the area-selective chemical vapor deposition of cobalt films on one oxide surface over another from the precursor Co2(CO)8 by addition of the nucleation inhibitor ammonia (NH3). In the absence of an NH3 coflow, the Co2(CO)8 precursor exhibits a weak intrinsic selectivity: at 70 °C, Co nucleates quickly on Al2O3 but more slowly on SiO2. The addition of an NH3 coflow, however, greatly amplifies the selectivity between different oxide surfaces. Thus, NH3 significantly inhibits nucleation on acidic oxides such as SiO2 and WO3 but has little effect on more basic oxides such as Al2O3, HfO2, and MgO. Comparison of growth on fully hydroxylated and dehydroxylated SiO2 suggests that hydroxyl groups are the nucleation sites that are affected by the addition of NH3. The mechanism of nucleation appears to be disproportionation of Co2(CO)8 to Co2+ (the intermediate that leads to nucleation) and Co(CO)4−: this disproportionation occurs readily on basic oxides but not on acidic oxides. The addition of NH3 has little effect on Co nucleation on basic oxides, probably because ammonia binds poorly to such surfaces, but NH3 greatly retards nucleation on acidic oxides such as SiO2; the authors propose that the latter result is either a site blocking effect or the result of conversion of Co2+ to inactive Co(NH3)x2+ species. Nucleation of cobalt is facile on gold (a very unreactive metal) even in the presence of NH3. The authors have found, however, that the deposition of Co on tungsten can be inhibited by exposing the surface briefly to ozone; no deposition occurs on the resulting thin tungsten oxide overlayer from Co2(CO)8 in the presence of NH3. In other words, this thin oxide overlayer affords the same selective inhibition behavior as seen on bulk WO3. In this way, both metal-on-metal and metal-on-oxide selectivity can be achieved. Cobalt films grown in the absence and presence of ammonia have resistivities of 11–20 and 15–25 μΩ cm, respectively.

Superconducting thin films of vanadium nitride have been grown by low temperature (250–300 °C) chemical vapor deposition from tetrakis(dimethylamido)vanadium (TDMAV) and ammonia. For example, films grown from TDMAV (1 sccm Ar as carrier gas) and 7 mTorr ammonia at 300 °C are nanocrystalline (cubic δ-phase) with an average crystal size of 20 nm, have relatively low room temperature resistivities of 250 μΩ cm, and are superconducting with critical temperatures as high as 7.6 K (versus a bulk value of 9 K). The films have a V:N ratio of 1:1, with a carbon content of <5 at. % and an oxygen content of <3 at. % (as determined by high resolution XPS). The V 2p3/2 and N 1 s XPS binding energies of 513.5 and 397.3 eV, respectively, are consistent with the presence of a nitride phase. In contrast, films grown at lower temperatures 95% in trenches of an aspect ratio of 4:1; for superconducting films grown at 250 °C, the step coverage is 65% for an aspect ratio of 3:1. At 150 °C, near-stoichiometric films with <2 at. % carbon and <3 at. % oxygen can be deposited if the gaseous ammonia is precracked by a remote plasma source; the resulting films have low resistivities of 320 μΩ cm but are not superconducting down to 4 K.

Attempts to fill deep trenches by chemical vapor deposition often result in a “bread-loaf” profile, an overhang near the trench opening that arises whenever the growth rate is slightly higher near the opening than deeper in the feature. Continued growth leads to premature pinch-off at the opening, which leaves an undesirable void or seam along the centerline. Bread-loaf profiles can form even under superconformal growth conditions, as the authors recently found for the growth of HfO2 from the precursor tetrakis(dimethylamino)hafnium and a forward-directed flux of H2O coreactant. The current paper describes a method that can reduce or eliminate the bread-loaf problem: addition of an isotropic flow of a reactant that inhibits growth near the trench opening but leaves the growth rate unchanged deeper in the trench. A Markov chain model for ballistic transport of the inhibitor inside trenches is developed to account for this behavior: the model reveals that suppression of a bread-loaf profile is best accomplished with growth inhibitors that have a high sticking probability (>0.1 per wall collision) and that are consumed during growth. Four molecules are investigated as potential inhibitors during HfO2 growth: tris(dimethylamino)silane, 3DMAS; methoxytrimethylsilane, MOTMS; hexafluoroacetylacetone, H(hfac); and acetylacetone, H(acac). The molecules 3DMAS and MOTMS inhibit growth but do so everywhere. As a result, they improve conformality, but are unable to eliminate the bread-loaf profile. In contrast, relatively small partial pressures (fluxes) of H(hfac) or H(acac) strongly inhibit HfO2 growth and do so selectively on the upper substrate surface and near trench openings. In conjunction with the use of a forward-directed water flux that affords superconformal growth, the use of H(hfac) or H(acac) enables seamless fill of HfO2 in trenches with aspect ratios as large as 10.

Thin films of molybdenum carbonitride, MoCxNy, are deposited by low temperature chemical vapor deposition from Mo(CO)6 and NH3 in the temperature range 150–300 °C. At a substrate temperature of 200 °C and Mo(CO)6 pressure of 0.01 mTorr, the composition varies from MoC0.48N0.20 to MoC0.36N0.33 (i.e., greater nitrogen and less carbon content) upon increasing the ammonia pressure from 0.3 to 3.3 mTorr. At a constant Mo(CO)6 pressure of 0.01 mTorr and an NH3 pressure of 2 mTorr, the composition varies from MoC0.50N0.30 to MoC0.12N0.40 with increasing substrate temperature from 150 to 300 °C. Selected films grown at substrate temperatures of 150, 200, and 250 °C are superconducting with critical temperatures of 4.7, 4.5, and 5.2 K, respectively. Grazing incidence x-ray diffraction data indicate that the films are crystalline and isomorphous with the cubic phases of Mo2N and Mo2C. With a forward-directed flux of precursors toward the surface, film growth is highly conformal in microtrenches of aspect ratio 6, with step coverages of ∼0.85 and 0.80 at growth temperatures of 150 and 200 °C, respectively.

We use aligned carbon nanotube (CNT) forests as scaffolds to deposit hafnium diboride (HfB2) and fabricate millimeter-thick ultra-high temperature composite coating. HfB2 has a melting temperature of 3250°C, which makes it an attractive candidate for applications requiring operation in extreme environments. Compared to typical refractory HfB2 processing, which requires temperatures exceeding 1500°C, we use conformal HfB2 chemical vapor deposition (CVD) to coat CNT forests at a low temperature of 200°C. During this process, nanometer thin HfB2 films grow on the CNT surface and uniformly fill tall CNT forests, thus transforming nanometer film deposition to a scalable HfB2 coating technology. The conformal HfB2 coating process uses static (S-) CVD, where the precursor is fed into a closed-system, enabling highly conformal coating and economically efficient utilization of the HfB2 precursor reaching 85 %. The modulus and compressive strength of the composites are measured using flat-punch indentation of micropillars having various coating thickness. Filling the CNTs with HfB2 strengthens their node morphology, and effectively enhances the mechanical properties. We study the nonlinear behavior of the material to extract unique modulus value that describes the stress-strain response at any applied compression. At the highest HfB2 coating thickness of 45 nm, the solid fraction is increased from 2% for the bare CNTs to 36% for the composite; the modulus and strength reach 107 and 1.5 GPa, respectively. An analytical model is used to explain the mechanism of the measured structure-mechanical property scaling. Finally, the process is used to fabricate CNT-HfB2 films having 1.7 mm height, a centimeter square area and only 5.8 x 10-6 nm/nm thickness gradient to demonstrate the potential for scalability.

In previous work, it was demonstrated that FexCo(1−x) alloy thin films with near ideal magnetic properties can be grown by chemical vapor deposition (CVD) from the precursors Fe(CO)5 and Co2(CO)8; previous attempts to grow such films by CVD, using these or other precursors, had not been able to afford high saturation magnetization. However, it was found that the morphology and composition were extremely sensitive to small variations in the deposition temperature and the precursor partial pressures. In a second work, it was showed that the CVD of pure iron films from Fe(CO)5 is subject to a self-poisoning effect in which the growth surface accumulates carbon, which causes the growth rate to decline progressively to zero. Then it was shown that the poisoning effect can be eliminated by adding a coflow of NH3 during CVD, which does not introduce measurable quantities of nitrogen into the film. In the current work, the authors return to the compositional instabilities in FexCo(1−x) alloy growth and show that, as seen for pure Fe growth, these instabilities can be as attributed to a surface poisoning effect involving dissociative chemisorption of carbon monoxide. It was found that a coflow of ammonia, which inhibits CO adsorption, enables the growth of FexCo(1−x) films over a wide temperature window with highly reproducible morphology and stoichiometry. Alloys that were grown under the NH3 coflow with suitable compositions (x ∼ 0.6) achieve near ideal values of the saturation magnetization.

We describe an atomic layer etching (ALE) method for copper that involves cyclic exposure to an oxidant and hexafluoroacetylacetone (Hhfac) at 275°C. The process does not attack dielectrics such as SiO₂ or SiN_x, and the surface reactions are kinetically self-limiting to afford a precise etch depth that is spatially uniform. Exposure of a copper surface to molecular oxygen, O₂, a weak oxidant, forms a ∼0.3 nm thick layer of Cu₂O, which is removed in a subsequent step by exposure to Hhfac. The etch reaction involves disproportionation of Cu(hfac) intermediates, such that ∼0.09 nm copper is removed per cycle. Exposure of copper to ozone, a stronger oxidant, affords ∼15 nm of CuO; when this oxidized surface is exposed to Hhfac, 8.4 nm of copper is removed per cycle. The etch products, Cu(hfac)₂ and H₂O, are efficiently pumped away; H₂O, a poor oxidant, does not attack the bare Cu surface. The roughness of the copper surface increases slowly over successive etch cycles. Thermochemical and bulk etching data indicate that this approach should work for a variety of other metals.