Plasma-enhanced atomic layer deposition of transition metal phosphates
- Author
- Thomas Dobbelaere
- Year
- 2017
- Abstract & Cover
- Source of Information
- https://biblio.ugent.be/publication/8534242
- University
- Ghent University, Department of Solid State Sciences, CoCooN research group (Ghent, Belgium)
The work in this PhD thesis describes a new class of plasma-enhanced atomic layer deposition processes for the deposition of transition metal phosphates. They are all three-step pro cesses following the same basic structure, where a single cycle consists of the following three steps: 1. TMP plasma exposure 2. O2 plasma exposure 3. Metalorganic precursor vapor exposure The following reaction mechanism is proposed: 1. Creation of phosphate ester surface species through a plasma poly merization-like process. The created species are believed to be chem ically similar to TMP, but with some −O−CH3 groups replaced by −OH: PO(OCH3 )3 plasma −−−→ PO(OCH3 )2 (OH) (s) + 1 2 C2H4 (g) 2. Combustion of the remaining methyl ligands by the O2 plasma, con verting the species to (poly)phosphoric acids: PO(OCH3 )2 (OH) + 6 O • −−→ PO(OH)3 (s) + 2 CO2 (g) + 2 H2O (g) During steps 1 or 2, the species may also dehydrate, thereby cross linking phosphate groups by forming P−O−P bonds: PO(OH)3 + PO(OH)3 −−→ PO(OH)2−O−PO(OH)2 (s) + H2O (g) 3. The −OH sites on the (poly)phosphoric acid surface species react with the metalorganic precursor, yielding a metal (poly)phosphate by a proton exchange reaction: R1−(OH)x (s) + M−R2 (g) −−→ R1−OxM (s) + Hx−R2 (g) Here, M represents the metal with valency x, R1 represents the (poly)phosphate, and R2 represents the ligands of the organometal lic precursor. It is believed that this mechanism is generic, i.e. any metal phosphate can be deposited by choosing a suitable metal precursor. This was explic itly demonstrated for aluminum, zinc, iron, titanium, vanadium, cobalt, and nickel phosphate. Although the reaction proposed in step 3 implies that this metal precursor needs to be reactive towards −OH groups (i.e. it needs to react with water), experiments revealed that this was not neces sarily true: while the aluminium, zinc, titanium, and vanadium precur sors are indeed hydrolysed to their respective oxides when combined in an ALD process with water,[83, 122, 133, 156] this is generally not the case for the metallocene-type iron, cobalt, and nickel precursors.[157] It is therefore likely that the reaction in step 3 is in reality more complex and may be facilitated by extra reaction pathways which are enabled by the preceding and subsequent plasma exposures, i.e. by the presence of reactive radicals or by the plasma-assisted removal of the metal precursor ligands. In each case, it was found that the process also worked without the O2 plasma step, but this resulted in a lower deposition rate and more car bon impurities in the deposited material. The reason for this is easily understood from the proposed reaction mechanism: the presence of (un reactive) methyl groups hinders the attachment of the metal precursor and contributes to the carbon content.
