Is there a simple way to convert each cluster detected by an k-means to an adjacency matrix in R?
Sure, you could bind the clusters to the data.
res <- kmeans(as.matrix(iris[1:2]), 2)
cbind(res$cluster, iris[1:2])
To split the matrix into separate clusters
split(cbind(res$cluster, iris[1:2]), res$cluster)