We show that the Fe xvii I ( 17.10 ~ { } { \AA } ) / I ( 17.05 ~ { } { \AA } ) line ratio observed in the Chandra HETG spectrum of the intermediate polar EX Hydrae is significantly smaller than that observed in the Sun or other late-type stars .
Using the Livermore X-ray Spectral Synthesizer , which calculates spectral models of highly charged ions based on HULLAC atomic data , we find that the observed I ( 17.10 ~ { } { \AA } ) / I ( 17.05 ~ { } { \AA } ) line ratio can be explained if the plasma density n _ { e } { \lower 3.225 pt \hbox { $ > $ } \atop \raise 2.15 pt \hbox { $ \sim$ } } 3 \times 1 % 0 ^ { 14 } ~ { } cm ^ { -3 } .
However , if photoexcitation is included in the level population kinetics , the line ratio can be explained for any density if the photoexcitation temperature T _ { bb } { \lower 3.225 pt \hbox { $ > $ } \atop \raise 2.15 pt \hbox { $ \sim$ } } 55 kK .
For photoexcitation to dominate the Fe xvii level population kinetics , the relative size of the hotspot on the white dwarf surface must be f { \lower 3.225 pt \hbox { $ < $ } \atop \raise 2.15 pt \hbox { $ \sim$ } } 2 % .
This constraint and the observed X-ray flux requires a density n { \lower 3.225 pt \hbox { $ > $ } \atop \raise 2.15 pt \hbox { $ \sim$ } } 2 \times 10 ^ { 14 } ~% { } cm ^ { -3 } for the post-shock flow .
Either way , then , the Chandra HETG spectrum of EX Hya requires a plasma density which is orders of magnitude greater than that observed in the Sun or other late-type stars .