Research with older adults is no different from other populations in that large sample sizes are going to be most desirable. There are a few groups/settings that you can probably get away with smaller samples... senior housing settings or specialty healthcare setting (e.g., day center, hospice, etc.) you can probably get away with a smaller sample size because of the difficulty involved in collecting data in these settings. Same with oldest old samples (about 80+), you can get away with smaller samples if this is the focus of your study because this group is harder to collect data with. The exact number in your sample will depend on your research question/statistical analyses. Cutoff ages will also depend on your research question, but most people use 65 years old to determine "older adults".