# Building huge magnetic field gradient

Looking at how MRI works, I came across the fact that the spatial resolution depends on the magnetic field gradient, this gradient being created by "gradient coils".

I was not able to find what the shape of these coils is! More generally, my question is

How do you achieve a huge magnetic field gradient?

• Took me less than sixty seconds on Google using "gradient coil design" to find several papers. One that's in the open is mri-q.com/uploads/3/2/7/4/3274160/gradient_coil_design_2010.pdf – CuriousOne Dec 29 '14 at 16:08
• Very interesting, I saw the abstract of this article many times but was never able to have access to it – agemO Dec 30 '14 at 21:19
• It's hit and miss. May be a temporal IP violation, too! Download a copy before it disappears. Good reading! – CuriousOne Dec 30 '14 at 21:24
• @CuriousOne The presence of paywalls for those not in academic settings is something we must always keep in mind, unfortunately. – probably_someone Jan 30 '18 at 18:57

In MRI, we have large homogenous field $\vec B_0 =B_0 \hat z$ always present, $\hat z$ being a unit vector in $z$ direction. This field is huge compared to the gradient fields which is why we are only interested in the change of the $z$ component of the total field. The word gradient actually refers to the gradient of this particular field component, i.e., $\nabla B_z$.
To achieve a gradient of $B_z$ in $z$ direction, one can use a simply a Helmholtz coil pair or a Maxwell coil. For the trasverse gradient (in $x$ or $y$ direction), one has to choose another type of coil, called Golay coil. You can find the coil geometries by googling.
Because static magnetic fields are curl-free in free space, having gradient in of $z$ component, the field must also a gradient of another component. Since the $B_0$ field dominates the field direction and magnitude, we can always neglect these so called concomitant components.