1: \begin{proof}

2: 	The first part has been proved in \cite{WuTang2015},

3: %	the above  {\em Lax-Friedrichs splitting property} holds, see Lemma 2.3(iii) in

4: %	and has played an important role in analyzing the PCP property of the numerical schemes for the RHD equations. 	

5: 	while the second part  is proved by contradiction as follows.

6: 	

7: 	Assume that the LxF splitting property holds for $\vec U\in \mathcal G$ and $\Gamma \in(1,2]$.

8: 	For any $\vec V=(\rho,\vec v,\vec B,p)^{\top} $ satisfying $\rho>0,~p>0$, and $v<1$,  one has

9: 	$$\vec U(\vec V) \pm \alpha^{-1} \vec F_i (\vec U(\vec V)) \in {\mathcal G},\quad \forall \alpha  \ge \varrho_i,~i=1,2,3.$$

10: Because	the speed of light $c =1$ is a rigorous bound of the spectral radius of $\partial \vec F_i/\partial \vec U$,

11: one can specially take $\rho=p=\epsilon>0,\vec v=(0.5,0,0), \vec B = (1,0,0)$, $\alpha = 1/\theta$ for $\theta \in (0, 1]$, and $\Gamma=5/3$, such that

12: 	\begin{equation*}

13:     \begin{split}

14: 	\vec U^\pm(\epsilon,\theta) := &\vec U \pm \alpha^{-1} \vec F_1 (\vec U)

15: 	\\

16: 	=& \left( \frac{\sqrt{3}}{3} (2+\theta) \epsilon,

17: 	\frac{14 \pm 13 \theta}{6} \epsilon \mp \frac{ \theta}{2}

18: 	,0,0,1,0,0, \frac{ 11 \pm 7 \theta }{3} \epsilon + \frac{1}{2}  \right)^{\top} \in {\mathcal G}={\mathcal G}_0.

19:     \end{split}

20: 	\end{equation*}

21: 	According to Remark \ref{rem:rmhd:equal}, one has $\tilde q(\vec U^\pm(\epsilon,\theta) )>0$, for all $\epsilon>0$

22: 	 and $\theta \in (0, 1]$.

23: 	The continuity of $\tilde q(\vec U)$ with respect to $\vec U$ further implies that for any fixed $\theta$,

24: 	$\vec U^\pm(\epsilon,\theta)$ is also continuous with respect to $\epsilon$. Therefore

25: 	$$

26: 	0 \le \mathop {\lim }\limits_{\epsilon \to 0^+ } \tilde q(\vec U^\pm(\epsilon,\theta) ) = \tilde q(\vec U^\pm(0,\theta) ) = - \frac{27}{64} \theta^2 (\theta^2+4)^2 <0,

27: 	$$

28: 	which leads to the contradiction. Hence the LxF splitting property does not hold for the admissible state set  $\mathcal G$  for the RMHD equations  \eqref{eq:RMHD1D} in general.

29: \end{proof}

30: