Atomic Layer Deposition of Late First-Row Transition Metals: Precursors and Processes

Author
Katja Väyrynen
Year
2019
Abstract & Cover

Late first-row transition metals, namely copper, nickel, and cobalt, are pivotal materials in many modern and future applications. Because of its low resistivity, Cu has for long been the metal of choice for interconnects in microelectronic devices. Co is needed in the smallest features of the 10 nm technology node interconnects, as it is more robust than Cu toward electromigration, a phenomenon causing damage to the interconnects. Being ferromagnetic, Co and Ni are in the focal point of developing faster and more durable magnetic memories capable of handling the exponentially increasing amounts of data being generated annually. The development of faster yet smaller electronic devices requires a constant increase in computational power. To improve the performance without increasing device size, the components on integrated circuits should be shrunk and packed more closely. The shrinking is achieved by using thin films with nanoscale thicknesses preferably arranged in threedimensional forms. For downscaling to continue, accurate thin film deposition methods are needed. Atomic layer deposition (ALD) provides atomic level accuracy and is thus the number one thin film deposition technique for modern and future devices. ALD is based on a cyclically repeated alternate supply of gaseous precursors that react on a substrate and form a uniform layer of material, atom by atom, even on complex three-dimensional structures. ALD is based solely on chemistry; to benefit from the many advantages the method has to offer, suitable precursors must first be found for each of the desired materials. ALD has been employed to deposit a myriad of materials ranging from pure elements to, for example, oxides, nitrides, and chalcogenides, but the deposition of metals has been hindered by a lack of reactive precursors and reducing agents. Thermal ALD processes exist mostly for noble metals, but mere thermal activation has often proven insufficient for the reduction of the late first-row transition metals. The aim of this thesis was to find and develop new precursors and processes for the ALD of high-quality Cu, Ni, and Co thin films, thus promoting the development of better microelectronics. Within the scope of this thesis, several new metal precursors for the ALD of the late firstrow transition metals were developed and tested. Out of all of them, the diamine adducts of Co(II) and Ni(II) chlorides showed the best performance in the ALD experiments. In addition to the new metal precursors, the focus of this thesis was also on finding more efficient alternatives for the conventional reducing agents, H2 and NH3. Tert-butylhydrazine showed high reactivity to produce Cu and Ni3N by ALD, providing significant improvement on film purity and resistivity over the existing processes. Tributyltin hydride, another powerful reducing agent, was studied for the ALD of Co and Ni. Instead of producing metallic Co or Ni, intermetallic Co3Sn2 and Ni3Sn2 were deposited unveiling a new field of ALD: the ALD of intermetallics. The same approach was also applied to the ALD of Ni2Ge thin films. Postdeposition reduction of the corresponding metal oxides and nitrides was also explored as an alternative route for the preparation of metal thin films. 

Source of Information
website (via riikka)
University
University of Helsinki
(Helsinki, Finland)
External Link
Read Thesis
LinkedIn
linkedin invite