Zirconium phosphinimide complexes of the form CpZr(NP-t-Bu 3 )Cl 2 (1) and Cp*Zr(NPR 3 )-Cl 2 (R ) i-Pr (2), t-Bu (3)) were readily prepared under ambient conditions via the reaction of [CpZrCl 3 ] n or Cp*ZrCl 3 with the appropriate trialkylphosphinimide lithium salt (R 3 PNLi). A series of derivatives were readily obtained via alkylation or arylation of the above dihalide precursors. These included CpZr(NP-t-Bu 3 )Me 2 (4), Cp*Zr(NPR 3 )Me 2 (R ) i-Pr (5), t-Bu (6)),. Reaction of 17 with the borane B(C 6 F 5 ) 3 yielded the zwitterionic and cationic complexes Cp*Zr(NP-t-Bu 3 )(CH 2 C(CH 3 )C(CH 3 )CH 2 B(C 6 F 5 ) 3 ) (18) and Cp*Zr(NP-t-Bu 3 )(THF)-(CH 2 C(CH 3 )C(CH 3 )CH 2 B(C 6 F 5 ) 3 ) (19). A number of the above compounds were screened for their potential as catalyst precursors in ethylene polymerization. In general, upon activation with methylaluminoxane, the resulting catalysts exhibit low activity. Efforts to understand the deactivation pathway for these zirconium catalysts involved investigating the interactions of catalyst precursors with activators. For example, reaction of 4 with the borane B(C 6 F 5 ) 3 leads to aryl group transfer and formation of catalytically inactive CpZr(NP-t-Bu 3 )(C 6 F 5 ) 2 (20). Interactions with MAO were modeled via reaction with AlMe 3 . The Zr clusters (Cp*Zr) 4 -(µ-Cl) 5 (Cl)(µ-CH) 2 ( 21) and (Cp*Zr) 5 (µ-Cl) 6 (µ-CH) 3 ( 22) were two of the products that were characterized from these reactions. The isolation of 21 and 22 infers that aryl for methyl exchange, ligand abstraction, and C-H bond activation may be catalyst deactivation pathways.