Abstract　 　　  

    Until about a decade ago, star clusters were believed to resemble"simple" stellar populations: all of their member stars had presumablyoriginated from a common progenitor molecular cloud that collapsedwithin a very short time period. This implies that they should sharethe same metal abundance and be characterized by a very narrow agerange. However, this simple picture no longer holds, in particular notfor most old globular clusters, which often show clear evidence ofhosting multiple stellar populations. I will introduce my PhD work,focusing on the stellar population properties of seven star clusterswith ages spanning from extremely young (∼ 25 Myr) via intermediateages (∼ 1-2 Gyr) to the cosmic age (≥ 10 Gyr). We found for theextremely young star clusters that even though they can still be verywell described by the single stellar population model, evidence offrequent interactions involving binary stars has already becomeapparent. In the color-magnitude diagrams of our intermediate-agesample clusters, complex extended main-sequence turn-off regionsappear, indicating that their initial star-forming processes may havelasted in excess of 300 Myr. This would dramatically change ourtraditional understanding of star cluster formation scenarios. However, upon careful analysis of the subgiant-branch morphology of the 2 Gyr-old cluster NGC 1651, to our surprise we foundthat the single-stellar-population origin might still apply. Weinterpret the observed extended main-sequence turn-off as evidence ofthe presence of a population of rapidly rotating stars in thecluster. In extremely old globular clusters we confirm the existenceof rejuvenated and secondary stellar populations. Our comprehensiveanalysis has clearly revealed the stellar population distribution instar clusters of different ages, as well as their formation processes.