Is CO a lewis acid?

Carbon monoxide is an excellent ligand towards low valent transition metals. Why?

The usual Lewis formulation is, #:C-=O:#. Given the triple bond, there is a formal negative charge on the carbon, and a formal positive charge on the oxygen. When carbon monoxide binds to a transition metal centre, usually the #C# is the donor atom. Now if the metal is already in a low oxidation state, further coordination to other donors is unlikely, unless the metal can redistribute (i.e. back donate) electron density to the ligand. The vacant anti-bonding orbitals on carbon monoxide are the right symmetry to ACCEPT electron density from filled metal d -orbitals.

For these reasons, carbon monoxide (and isoelectronic cyanide ion) are known as #pi#-acids.

The HOMO of carbon monoxide (above) is the donor orbital; it is conceived to be occupied, and shunts electron density towards the metal centre. Further ligation of carbon monoxide is unlikely, UNLESS the metal can back donate some electron density; and a mechanism for this process exists inasmuch as the anti-bonding orbitals on carbon monoxide (below) are UNOCCUPIED, and are of appropriate energy and symmetry to receive electron density from the filled non-axial #d# orbitals (#d_(xy)#, #d_(xz)#, #d_(yz)#) of the metal centre:

Both images were from this site.

This back-donation is ANTIBONDING with respect to the carbon monoxide ligand, hence its IR stretching frequency drops from the #2100# #cm^-1# of the free gas to below #1900# #cm^-1# or lower. Now while the interaction is antibonding with respect to the carbon monoxide ligand, the interaction is bonding with respect to the #M-CO# bond inasmuch as electron density has been shared between metal and ligand.

Because bound carbon monoxide has demonstrably accepted electron density, it is commonly known as a #pi#-acid ligand. There are many zerovalent transition metal complexes with the formula #M(CO)_6# or #M(CO)_4#.

anor has a good answer, but I wanted to provide a visual approach. Also, #"CO"# can be BOTH a Lewis acid and base, and is typically both in Ligand Field Theory. I go more into that here.

Although it may seem counterintuitive at first (seeing how there are lone pairs of electrons on both carbon and oxygen), carbon monoxide can accept electrons in its antibonding, #pi^"*"# orbitals, in addition to behaving like a Lewis base.

The MO diagram for CO is:

The high-lying pair of MOs (by the carbon #2p# AOs) that are close in energy are the #pi_x^"*"# and #pi_y^"*"# (if we take the horizontal coordinate axis to be the #z# axis). Those LUMOs would accept the electrons.

We can see, as an example, in the reaction to form carbonic acid, that #"CO"_2# accepts electrons (also in an antibonding, #pi^"*"# orbital on the carbon):

Similarly, carbon monoxide can do that too on the carbon (just focus on the first step).

Since carbon monoxide has a method for accepting electrons, it by definition can be a Lewis acid. But as implied earlier, it can also be a Lewis base.