This is very interesting. I'm kind of of the opinion- although I have nothing to base this on- that emotions evolved alongside various language trees (because language isn't universal- I'll cite the source if anybody's interested) as a very useful social tool to allow better cohesion of groups.
Since this in many ways mirrors a similar theory I have about religion being a special case of the unlikely phenomena of group selection (the basic premise of my argument in this is that religion allows a group to grow and propagate it's belief without the need for biological reproduction- so the 'god of battles' case where each successive generation has people less and less willing to die for the idea fails to happen, because it is not truly natural selection that is acting upon them- it's behavior and group selection) it would fit that religious groups that propagate widely would also propagate similar emotional patterns- which would, among other things, be one possible cause for the disparity of facial expressions between the far east and the western world (as only the most visible example of this).
The argument that one or both of these could be useful for some kind of an artificial intelligence is one that has also occurred to me, but only if that artificial intelligence works cooperatively in groups; even then, if I were to program this, I would make considerable changes as to how the emotion is experienced, dealt with, and what it can communicate- but this presupposes a lot of thought on things such as an AI's sense of it's own social image as well as the social images of everything else (in the context of social images of certain objects, actions, states or events, or by association with other entities of different social states, etc).
First I would caution against referring to a "science of emotions", certainly there has been an increase in the scientific investigation of emotions; however, there are huge gaps in our understanding that would preclude being able to reference a SoE. Rest assured however that people have been very interested in understanding how emotions work for some time.
Aspects of emotions that are, to some degree scientifically understood;
So you see there actually has been a good deal of concerted effort to understand emotions in humans.
Emotions and spirituality are definitely not the same thing, after all spirituality is an invention of homo-sapiens, and the emotional centers of the brain are vastly older than our species. Spirituality or religiosity, begins as a function of cognition, with immediate and sometimes huge emotional stimulation potentials within the individual. Religion taps into this feedback loop and exploits it better than any other thing I can think of offhand. Emotions are primal and extremely powerful, the have the ability to easily override higher brain functions. How is it people continue to believe the world is 6000 thousand years old, or that eternal life is a good idea?
So in my opinion they are not the same, but are intimately related, so much so that emotional content, may be a necessary condition for spirituality.
As far as emotions being useful to robotics, I would say no, not at this time, nor until there are some major breakthroughs in AI, and just to be clear, there are many orders of magnitude difference between AI and machine consciousness.