The potential of solid-state materials comprising Co(salen)
units
for concentrating dioxygen from air was recognized over 80 years ago.
While the chemisorptive mechanism at the molecular level is largely
understood, the bulk crystalline phase plays important, yet unidentified
roles. We have reverse crystal-engineered these materials and can
for the first time describe the nanostructuring requisite for achieving
reversible O2 chemisorption by Co(3R-salen) R = H or F,
the simplest and most effective of the many known derivatives of Co(salen).
Of the six phases of Co(salen) identified, α-ζ: α
= ESACIO, β = VEXLIU, γ, δ, ε, and ζ
(this work), only γ, δ, ε, and ζ are capable
of reversible O2 binding. Class I materials (phases γ,
δ, and ε) are obtained by desorption (40–80 °C,
atmospheric pressure) of the co-crystallized solvent from Co(salen)·(solv),
solv = CHCl3, CH2Cl2, or 1.5 C6H6. The oxy forms comprise between 1:5 and 1:3
O2:[Co] stoichiometries. Class II materials achieve an
apparent maximum of 1:2 O2:Co(salen) stoichiometries. The
precursors for the Class II materials comprise [Co(3R-salen)(L)·(H2O)
x
], R = H, L = pyridine, and x = 0; R = F, L = H2O, and x = 0; R = F, L = pyridine, and x = 0; R = F, L =
piperidine, and x = 1. Activation of these depends
on the desorption of the apical ligand (L) that templates channels
through the crystalline compounds with the Co(3R-salen) molecules
interlocked in a Flemish bond brick pattern. The 3F-salen system produces
F-lined channels proposed to facilitate O2 transport through
the materials through repulsive interactions with the guest O2. We postulate that a moisture dependence of the activity
of the Co(3F-salen) series is due to a highly specific binding pocket
for locking in water via bifurcated hydrogen bonding to the two coordinated
phenolato O atoms and the two ortho F atoms.