The rate constants for a bimolecular gas phase reaction have been measured at several temperatures.

kr $/$M$-1$s$-1$

$0\mathrm{.}138$x$103$

$1\mathrm{.}63$x$103$

$7\mathrm{.}2$x$103$

$36\mathrm{.}8$x$103$

$81$x$103$

T$/$C

$155\mathrm{.}3$

$208\mathrm{.}3$

$246\mathrm{.}5$

$295\mathrm{.}8$

$330\mathrm{.}8$

a$)$ Use your favorite data analysis software $[$NOT Excel$]$ to plot kr vs T $($in Kelvin$)$

Without linearizing the plot $($i$.$e$.$ don't change the above plot$)$

b$)$ Fit data to the Arrhenius equation $($equation $6$D$\mathrm{.}1$c$)$ and extract values for the collision frequency, A$,$ and activations energy, Ea$,($be careful of units$)$

Now linearize the plot by plotting ln$($kr$)$ vs $1/$T $($in $1/$Kelvin$)$

c$)$ Do a linear regression fit and then algebra to extract the collision frequency, A$,$ and activations energy, Ea$.$

d$)$ Compare the values you got from the two different fits you did $-$ are they functionally the same?

e$)$ With d$)$ in mind $-$ historically we linearized plots so that fits were always linear regression which computationally was easier to be done by hand, then later by the early computers. Statistically this unintentionally weights data for the fit. Nowadays computers can easily handle non$-$linear fits $-$ do you think the practice of linearization is appropriate now that we have the computing power? Justify your answer.