This error does not have much to do with igraph. After get.adjacency
you have a sparse graph and
adj <- adj + 1
essentially converts it to a dense matrix. To have a dense matrix with 80,000 rows/columns you need 80,000 * 80,000 * 8 bytes of memory, which is about 48GB. Actually, even if you have that much memory, most probably it will not work, because R will want to copy the matrix at least once, so you need twice of this.
And then if you want to create an igraph graph from it, you'll need much more memory than this ~100GB, because igraph was designed for sparse graphs, and it is not very efficient with memory if the graph is dense. It needs (2*n+4*m)*8 bytes of memory, where n is the number of vertices and m is the number of edges. You will need another m*8 bytes of memory for the weights, and n*4 bytes for the other (integer) attributes, per attribute.
So I would suggest to use some graph analysis software that stores the data on the disk, or consider another data representation, e.g. not storing the edges that have zero weight.