Just one more graph, as an addition: A 100nF 100V, 1206-size SMD multilayer ceramic cap. Above 20MHz the performance of the 100nF ceramic and the 15?F tantalum are very much the same, while below 20MHz the ceramic one gets progressively bad, while the tantalum one stays pretty good the low end of the frequency range measured.
Ergo: A tantalum electrolytic chip cap is a better, wider bandwidth bypass than a ceramic chip cap. From 20 to 300MHz there is no significant difference between them in bypassing performance, although I would expect the ceramic cap to handle more current over a longer time.
