Is the difference between the acyclic chromatic number and the star chromatic number unbounded?

Question

Is $\chi_s(G)-\chi_a(G)$ unbounded in general graphs?

I thought $\chi_s(G)-\chi_a(G)$, the difference between the star chromatic number and the acyclic chromatic number, is unbounded for general graphs. But, I am unable to show this. I think I have seen a paper on star coloring which give $\chi_a$-binding functions for the star chromatic number and claim that these functions are best possible (but I don't recall the name of the paper; I think the paper dealt with a number of colouring variants including acyclic coloring and star coloring). If this is true, then $\chi_s(G)-\chi_a(G)$ must be unbounded.

Definitions
A coloring of a graph $G$ is a function $f$ from the vertex set of $G$ to a set of colors such that $f(u)\neq f(v)$ for every edge $uv$ of $G$. A coloring $f$ of $G$ is an acyclic coloring if there is no cycle in $G$ 2-colored by $f$. A coloring $f$ of $G$ is a star coloring if there is no 4-vertex path in $G$ 2-colored by $f$. The acyclic chromatic number $\chi_a(G)$ (resp. the star chromatic number $\chi_s(G)$) is the minimum number of colors required to acyclic color (resp. star color) the graph $G$.

Answer 1

While I know nothing about these measures, quick googling led me to the paper

Michael O. Albertson, Glenn G. Chappell, H. A. Kierstead, André Kündgen, Radhika Ramamurthi: Coloring with no $2$-colored $P_4$’s, Electronic Journal of Combinatorics 11 (2004), no. 1, art. no. R26, doi: 10.37236/1779.

They prove that $\chi_s(G)\le\chi_a(G)\bigl(2\chi_a(G)-1\bigr)$ (Corollary 4.6), and exhibit a family of graphs $G_t$ such that $\chi_a(G_t)=t$ and $\chi_s(G_t)=\binom{t+1}2$ (Theorem 6.3 and the comment below).