Different R-Square using LOGEST and graph

Hi,

I have calculated the R-Squared value of an exponential regression using LOGEST and using a graph (by adding the exponential trendline and its RSquared value). 

Despite Xs and Ys values being the same in both calculations, I get different R-Squared values.

I have tried using a linea regression and RSQ formula and the result perfectly matches with the result in the graph, but the values follow an exponential trend, so I need to use LOGEST.

Could anyone help me with this? why my R-Squared value using LOGEST does not match with the result in the graph using an exponential trendline?

I have attached the values in Excel file.