Why does a rolling sphere stop? I've just started studying Rotational Mechanics and one thing I'm confused about is that if friction is equal to zero when a body is rolling purely.. then why does a rolling body e.g. sphere.. stop rolling in reality
 A: This is a great question that, I think, naturally follows any explanation of rolling objects. Hopefully this webpage will help. 
http://www.lhup.edu/~dsimanek/scenario/rolling.htm
I will explain below, beginning with a brief summary of friction. 
Friction
Friction is defined as a force that opposes slipping or sliding between two surfaces. If you are dragging a bag of money:

You will notice it is hard. Why? There's just paper there, and you're pulling perpendicular to gravity so there shouldn't be any force you are fighting. Well, we know from physical experience that there is a force called friction:

On Earth's surface, gravity does its best to push objects downwards. And when objects that aren't perfectly smooth get pushed together all of their little bumps and holes will get pushed together as well and you will have something like the picture above. To move the green object you will have to go around all those contact points. How hard this will be depends on the roughness of the object (the coefficient of friction, µ) and how much the ayssmetrical surfaces are pressed together, $mg$.
The reality is that there are no perfectly smooth surfaces, and even if there were you would still experience some friction from intermolecular forces because the objects in contact are both made of atoms. 
The key thing to keep in mind is that friction opposes sliding between two surfaces, and that there is no realistic contact where this isn't present.
Perfect Rolling Objects
Now, lets apply this definition to a rolling object. A perfect sphere will only make contact at one point. And we have said there is always friction between two contact surfaces. So to move with friction we make the natural assumption that a sphere must be slipping or sliding. 

However, if we take a closer look at that single contact point we see immediately that this is not the case. For when the sphere moves forward at all, even just a little bit, that contact point rotates off the ground and there is a new contact point. A rolling object never has a surface in contact with the ground for a more than a split second. So, if there is not sliding (no two surfaces are continually fighting each other) then we conclude that there must not be friction.  
There seems to be some contradiction here. Two surfaces are moving relative to each other but are not experiencing the traditional definiton of friction.
To explain this we will take one more hypothetical before getting to the reality of this situation.
Imagine a pencil balancing on its tip. If you push at the center of mass, about halway up the pencil, the pencil will fall forward, essientially rotating about its tip. 
Now imagine this were repeated in space. Not even deep space, but the upper atmosphere of planet earth. Gravity is still present and pulling the same way on the pencil. The only difference is that there is no surface for the pencil to balance on. If you repeated the experiment and pushed on the center of mass the pencil would just move forward. It would not rotate. Things do not rotate unless there is a torque.
So where is the torque that causes the pencil to rotate forward on earth? It comes from the contact point with the surface, from friction. When you push on the center of mass there is an equal resistance force at the tip from the contact. There is now a torque and the pencil rotates around the fixed point, the tip. Once it begins to rotate gravity then helps with additional torque downwards at the center of mass.
So we now have an example when friction causes an object to move without sliding taking place between the surfaces. 
The reality of the perfect sphere is that it is made up of an infinite amount of pencil tips. When force is applied to the sphere the single contact point resists motion and the rest of the sphere begins to move around that point. But as soon as motion begins there is a new contact point, and torque from gravity and forward motion are causing the sphere to rotate about that new point. The sphere is continual falling over itself. Hopefully this picture helps to illustrate:

The reality is that rolling is caused by friction, though in a unique way. Rolling would not happen without those instantaneous contact points resisting motion. 
If a rolling object had zero friction it would not roll at all. It would just translate. But this is never the case because there is always friction. In the case of a rolling object, friction causes the object to be continually tipping over.
So we have identified a common misconception about the nature of rolling, but your question still remains, could an object roll forever.
Perpetual Motion
We have been looking at a perfect sphere with one contact point. There isn't any obvious reason why a rolling body should every stop in this case. But, as we have discussed, smooth rigid objects are an idealization. In reality both the body and the surface are slightly deformed by one another's presence.

Whether the object deforms or the surface does, there is now more than one contact point.  The normal force of these contact points are not all aligned with the center of the sphere, as was the case with a single contact point. If an object is rolling forward it is pushing against these new contact points. Since these reaction forces are off center there is now a torque opposite to the sphere's direction of motion. Over time this will bring the sphere to a stop.
Because all objects are made of atoms, there is not a case where this deformation  doesn't occur. No object in this scenario will roll forever. Even without deformation there are quantum mechanical reasons why the objects would slow, as Simanek explains in his article.  
