Why is the density plot using noob normalization not smoother than Quantile normalization? Does it indicate something regarding the data? `

It's not clear what you mean by noob normalization, since that's not actually a thing. NOOB stands for normal-exponential out-of-band background correction, and is simply a way to estimate and correct for background binding. In other words, preprocessQuantile normalizes without background correcting first, and preprocessNoob background corrects only. The normal workflow that uses preprocessNoob is preprocessFunnorm, which first background corrects, and then uses the first two principal components of the control probes to normalize the type I and II probes separately. So if you wanted to compare things, I would imagine it should be preprocessQuantile versus preprocessFunnorm.