May I extract from the function fviz_cluster() the PCs composition?

Dim1 and Dim2 corresponds to principal components 1 and 2, respectively.

No, k-means does not compute PCA. The K-means clustering identifies only the grouping of the samples.

Visualization is performed using the function fviz_cluster().

1/ If you have only two variables (x, y) in your dataset, then a standard scatter plot of x vs y is drawn by fviz_cluster() and points will be colored by groups.

2/ If you have more than two variables in the data set, then the function fviz_cluster() performs PCA to reduce the dimensionality of the data, and uses PC1 (Dim1) and PC2 (Dim2) to visualize the data.

The function fviz_cluster() [factoextra package] can be used to easily visualize k-means clusters. It takes k-means results and the original data as arguments. In the resulting plot, observations are represented by points, using principal components if the number of variables is greater than 2.