Register to reply 
What is the angularmomentum 4vector? 
Share this thread: 
#1
May1011, 02:51 PM

P: 23

Uh, the title pretty much says it: I'm wondering what the 4vector analog to the classical 3angular momentum is. Also, is the definition
L = r [tex]\times[/tex] p still valid for the 3angular momentum in special relativity? 


#2
May1011, 03:38 PM

Sci Advisor
Thanks
P: 4,160

In four dimensions, angular momentum is no longer represented by a vector. The position is a vector, x_{μ}. The momentum is a vector, p_{ν}. But when you take the cross product you get a rank two tensor, J_{μν} = x_{μ} p_{ν}  p_{μ} x_{ν}.
Why does this happen in four dimensions? Well actually the cross product is supposed to do that. It's only in three dimensions that we can replace it with a vector. J_{μν} has six components. If you look at three of them in which μ and ν are both space indices, J_{ij} = x_{i} p_{j}  p_{i} x_{j}, then these values correspond to the three components of what we call the angular momentum vector: L_{x} = J_{yz}, L_{y} = J_{zx}, and L_{z} = J_{xy}. 


#3
May1011, 04:37 PM

P: 23

Thanks for the helpI assume then that this tensor transforms by applying the Lorentztransformation matrix twice?



#4
May1011, 04:55 PM

Sci Advisor
HW Helper
Thanks
P: 26,148

What is the angularmomentum 4vector?
Hi Zarathustra0! Welcome to PF!
to transform it, it's easiest to write it as a 2form rΛp, with the six (antisymmetric) basis elements tΛx tΛy tΛz yΛz zΛx and xΛy so yes you apply the Lorentz transformation matrix twice to each basis element (of course, it only makes a difference to t and x) 


#5
May1011, 10:52 PM

P: 23

To make sure I understand (and I may not because the notions of forms and wedge products are still a bit new to me), does this mean I should express the tensor as a linear combination of the tensors
[0 1 0 0] [1 0 0 0] [0 0 0 0] [0 0 0 0], [0 0 1 0] [0 0 0 0] [1 0 0 0] [0 0 0 0], [0 0 0 1] [0 0 0 0] [0 0 0 0] [1 0 0 0], [0 0 0 0] [0 0 1 0] [0 1 0 0] [0 0 0 0], [0 0 0 0] [0 0 0 1] [0 0 0 0] [0 1 0 0], and [0 0 0 0] [0 0 0 0] [0 0 0 1] [0 0 1 0], perform the Lorentz transformation on each by two applications of the transformation matrix, and then add up the resulting six matrices? In other words, do the above tensors form the basis in question? Earlier I naively tried to just apply the transformation matrix twice to the Jtensor itself, but the result wasn't antisymmetric, so I don't think it worked. Then again, given the nature of matrix multiplication, it seems like writing the matrix as a linear combination of other matrices and multiplying each of them by the square of the transformation matrix would have the same effect as just multiplying the square of the transformation matrix directly with the original angularmomentum tensor itself, and I already tried that, so maybe I'm still not understanding. 


#6
May1111, 03:32 AM

Sci Advisor
HW Helper
Thanks
P: 26,148

Hi Zarathustra0!
Use one or the other. 2forms, in my opinion, are easier. There isn't much call for writing out the Minkowski angular momentum but the similar electromagnetic field is regularly used … we can write it either as the antisymmetric tensor (matrix) 0 B_{y} B_{z} E_{x} etc (which takes up four lines, and doesn't say anything intuitive), or as a neat 2form … E_{x}tΛx + E_{y}tΛy + E_{z}tΛz + B_{x}yΛz + B_{y}zΛx + B_{z}xΛy but once you get used to the idea, you can just write that 2form as (E_{x},E_{y},E_{z},B_{x},B_{y},B_{z}), and you'll know what that means just as you know what (E,p_{x},p_{y},p_{z}) means without having to write the basis vectors! (try writing out the Lorentz force expression F* = q(t,x,y,z)Λ(E_{x},E_{y},E_{z},B_{x},B_{y},B_{z}), to get a feel for how it all works ) 


#7
May1111, 07:50 AM

P: 23

OK, in that case to apply the transformation twice to each basis element, should I simply make the substitutions
ct' > [tex]\gamma[/tex]ct  [tex]\beta\gamma[/tex]x, x' > [tex]\beta\gamma[/tex]ct + [tex]\gamma[/tex]x, y' > y, and z' > z twice to the six basis elements? 


#8
May1111, 08:00 AM

Sci Advisor
HW Helper
Thanks
P: 26,148

yes!
so for example x'Λt' = … ? (btw, you may find it easier to use the rapidity α, where tanhα = v = β, and coshα = γ) 


#9
May1111, 08:22 AM

P: 23

OK, coolthanks for all the help!
Actually I think I have one last (perhaps trivial) question. In the position 4vector (in r [tex]\wedge[/tex] p), should the temporal component ct be 0? This seems intuitive since we're just talking about two points in space (the position of the orbiting particle and the center of rotation) and not events at different times. 


#10
May1111, 10:40 AM

Sci Advisor
HW Helper
Thanks
P: 26,148

I can't remember.
I can't remember ever seeing a useful application of rΛp ! 


#11
May1111, 06:32 PM

Sci Advisor
PF Gold
P: 1,847




#12
May1211, 10:35 AM

P: 95

The timespace components of the angular momentum tensor give the center of mass of the system (minus t times the momentum, if you are in a frame with nonzero momentum). You can think of the angular momentum tensor as a way of combining the ordinary angular momentum and the center of mass into a single 4D object, much as the relativistic energymomentum vector combines energy and momentum into a single 4D object.



Register to reply 
Related Discussions  
Heisenberg Uncertainty Relations  angular momentum and angular displacement  Advanced Physics Homework  1  
Vector products Finding the direction of angular momentum  Advanced Physics Homework  7  
Physics Rotating Disk, Moment of Inertia, Angular Momentum, Angular Velocity  Introductory Physics Homework  12  
Angles made by angular momentum vector with magnetic field  Introductory Physics Homework  2  
Angular momentum and orbital angular momentum problems  Introductory Physics Homework  3 