Fermi–Walker Derivative

• Fermi–Walker derivative is a transport operator that preserves orthonormality along a timelike worldline while eliminating intrinsic spatial rotation.
• It underlies the construction of nonrotating frames for accelerated observers, connecting with Fermi coordinates, Thomas precession, and geodesic specializations.
• Generalized forms incorporate arbitrary proper rotation, extending applications to spinor transport, rotating frames, and infinite-dimensional symmetry algebras.

The Fermi–Walker derivative is the transport law along a timelike worldline that preserves orthonormality while removing intrinsic spatial rotation from an accelerated observer’s carried frame. In Minkowski spacetime, for a worldline $z^\mu(\tau)$ with proper time $\tau$, 4-velocity $u^\mu=\dot z^\mu(\tau)$, and 4-acceleration $a^\mu=\ddot z^\mu(\tau)$, a vector $w^\mu(\tau)$ is Fermi–Walker transported when

$$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$

Ordinary Fermi–Walker transport is the nonrotating baseline for accelerated motion; generalized versions include arbitrary proper rotation, while geodesic specializations reduce to parallel transport and underlie standard Fermi coordinate constructions (Llosa, 2017, Llosa, 2017, Klein et al., 2010).

1. Relativistic definition and geometric content

In the relativistic formulation, the Fermi–Walker derivative is built from the antisymmetric tensor

$\Omega^a{}_{b}=u^a a_b-u_b a^a,$

so that the transport law may be written

$\frac{dw^a}{d\tau}=\Omega^a{}_{b}\,w^b .$

Equivalently,

$$\left(\frac{D_{\rm FW}V}{d\tau}\right)^a = \frac{dV^a}{d\tau} - (u^a a_b-u_b a^a)V^b,$$

and Fermi–Walker transport is the condition $\frac{D_{\rm FW}V^a}{d\tau}=0$ (Llosa, 2017).

Its geometric role is sharply defined. The transport preserves orthogonality to the observer’s 4-velocity and excludes any intrinsic rotation of the spatial triad. In that sense it is the transport law appropriate to an accelerated observer carrying a nonrotating frame. If the worldline is inertial, $\tau$0, then $\tau$1, so Fermi–Walker transport reduces to parallel transport. The same source explicitly notes that $\tau$2 and $\tau$3 are Fermi–Walker transported along $\tau$4, but generally are not parallel transported (Llosa, 2017).

This nonrotation statement becomes especially transparent in tetrad form. For an orthonormal tetrad $\tau$5 with

$\tau$6

Fermi–Walker transport gives

$\tau$7

In comoving components,

$\tau$8

The vanishing of $\tau$9 is the precise algebraic expression of the absence of spatial rotation with respect to a gyroscopically nonrotating frame (Llosa, 2017).

2. Tetrads, coordinates, and accelerated reference systems

Fermi–Walker transport is not only a derivative operator; it is the structural input for coordinates adapted to accelerated observers. Given a timelike worldline $u^\mu=\dot z^\mu(\tau)$0 and a Fermi–Walker transported orthonormal tetrad $u^\mu=\dot z^\mu(\tau)$1, the time coordinate is defined implicitly by orthogonality to the 4-velocity,

$u^\mu=\dot z^\mu(\tau)$2

and the spatial coordinates by projection on the spatial triad,

$u^\mu=\dot z^\mu(\tau)$3

The inverse map is

$u^\mu=\dot z^\mu(\tau)$4

These are Fermi–Walker coordinates for an accelerated, nonrotating observer (Llosa, 2017).

The induced metric takes the form

$u^\mu=\dot z^\mu(\tau)$5

Several features follow directly. The spatial metric is Euclidean, there are no $u^\mu=\dot z^\mu(\tau)$6 cross terms, the frame is locally synchronous, and the factor $u^\mu=\dot z^\mu(\tau)$7 controls the position-dependent proper-time rate,

$u^\mu=\dot z^\mu(\tau)$8

The allowed coordinate domain is restricted by

$u^\mu=\dot z^\mu(\tau)$9

with equality defining the horizon of the Fermi–Walker coordinate system. Different stationary points $a^\mu=\ddot z^\mu(\tau)$0 have different proper accelerations,

$a^\mu=\ddot z^\mu(\tau)$1

Accordingly, a single accelerated Fermi–Walker frame does not assign one universal proper acceleration to all of its stationary observers (Llosa, 2017).

Within this framework, Einstein’s uniformly accelerated systems acquire a precise modern interpretation. They are Fermi–Walker coordinate systems, specifically the special case of rectilinear motion with constant proper acceleration. More strongly, if an accelerated reference system is required to admit an instantaneously comoving inertial system at every instant, then the transport law is forced to be Fermi–Walker: the only such accelerated systems belong to the Fermi–Walker class (Llosa, 2017).

The same analysis isolates Thomas precession. If one insists on axes remaining perpetually parallel to those of an initial inertial frame while the velocity changes direction, the result generally differs from Fermi–Walker transport. The relative rotation is Thomas precession, with angular velocity

$a^\mu=\ddot z^\mu(\tau)$2

Thus Fermi–Walker transport defines the nonrotating local frame, while successive non-collinear boosts relative to a fixed inertial standard produce precession relative to that frame (Llosa, 2017).

3. Generalized Fermi–Walker transport and rotating frames

A central extension replaces ordinary Fermi–Walker transport by a transport law that includes arbitrary proper rotation of the observer’s spatial axes. In this generalized form,

$a^\mu=\ddot z^\mu(\tau)$3

with

$a^\mu=\ddot z^\mu(\tau)$4

where $a^\mu=\ddot z^\mu(\tau)$5 is an arbitrary vector orthogonal to $a^\mu=\ddot z^\mu(\tau)$6, interpreted as the proper angular velocity 4-vector. The first term is the ordinary Fermi–Walker part determined by acceleration; the second adds arbitrary rotation of the spatial frame. Ordinary Fermi–Walker transport therefore describes an accelerated but nonrotating observer, whereas generalized Fermi–Walker transport describes an accelerated and arbitrarily rotating one (Llosa, 2017).

The associated generalized Fermi–Walker coordinates use the same simultaneity condition as the ordinary case,

$a^\mu=\ddot z^\mu(\tau)$7

but evolve the tetrad according to generalized transport. Their inverse map is

$a^\mu=\ddot z^\mu(\tau)$8

and comparison with ordinary Fermi–Walker coordinates built on the same worldline gives

$a^\mu=\ddot z^\mu(\tau)$9

The generalized system differs from the nonrotating one by a time-dependent rotation determined by the proper angular velocity (Llosa, 2017).

The metric in generalized Fermi–Walker coordinates is

$w^\mu(\tau)$0

Here $w^\mu(\tau)$1 and $w^\mu(\tau)$2 are the intrinsic proper acceleration and proper angular velocity in the comoving tetrad. The term $w^\mu(\tau)$3 is the acceleration or redshift factor; $w^\mu(\tau)$4 encodes rotational effects; and the cross term

$w^\mu(\tau)$5

is characteristic of rotating coordinates. The coordinate domain is restricted by

$w^\mu(\tau)$6

with equality defining the horizon of the generalized system (Llosa, 2017).

This generalization also has an algebraic consequence. Two generalized Fermi–Walker coordinate systems are related by generalized isometries rather than ordinary isometries, because the metric keeps the same functional form while the six functions $w^\mu(\tau)$7 and $w^\mu(\tau)$8 may change. The resulting infinitesimal symmetry algebra is an infinite-dimensional extension of the Poincaré algebra, but it is Abelian: $w^\mu(\tau)$9

$$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$0

The noteworthy point is that, unlike Lorentz boosts, acceleration and rotational boost generators commute with each other and with the Poincaré generators as well (Llosa, 2017).

4. Geodesic specializations, Galilean analogues, and moving frames

Not all uses of the adjective “Fermi” involve an explicit Fermi–Walker derivative formula. In Robertson–Walker cosmology, explicit Fermi coordinates have been constructed for comoving observers with timelike geodesic paths. In that setting the observer’s 4-acceleration vanishes, so Fermi–Walker transport reduces to parallel transport, and the tetrad used in the construction is taken to be parallel along the worldline: $$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$1 The coordinate system is then built by the exponential map from spacelike geodesics orthogonal to the observer’s worldline. This is the geodesic specialization of the usual nonrotating Fermi framework (Klein et al., 2010, Bolós et al., 2011).

This distinction matters conceptually. For geodesic observers, there is no need to separate Fermi–Walker transport from parallel transport. For accelerated observers, however, that distinction becomes essential: the nonrotating observer-adapted frame is Fermi–Walker transported, not merely parallel transported. The cosmological papers therefore illustrate the special case in which the Fermi–Walker machinery simplifies completely rather than a distinct alternative to it (Klein et al., 2010, Bolós et al., 2011).

A non-Lorentzian analogue appears in Galilean space $$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$2, where the metric is degenerate and vectors split into isotropic and non-isotropic classes. For a unit-speed curve with tangent $$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$3 and acceleration $$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$4, the Fermi–Walker derivative of a vector field $$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$5 is defined by

$$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$6

Because the scalar product depends on vector type, the explicit formulas split into cases: $$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$7

$$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$8

A vector field is Fermi–Walker transported when $$\frac{dw^\mu}{d\tau}=(u^\mu a_\nu-u_\nu a^\mu)\,w^\nu .$$9 (Şahin et al., 2018).

The Galilean treatment preserves the standard interpretation of nonrotation, but its geometry changes the transport equations substantially. For isotropic vector fields, the paper states that the Fermi–Walker derivative coincides with the Fermi derivative. For the Frenet and Darboux moving frames, the nonrotation criterion is very restrictive: the Frenet frame is nonrotating if and only if the curve is a line, and the Darboux frame is nonrotating if and only if the curve is a line. These results parallel the relativistic statement that commonly used moving frames generally rotate too much to qualify as Fermi–Walker frames (Şahin et al., 2018).

5. Spinors, curved spacetime, and quantum transport

Fermi–Walker transport also appears in spinorial form. In Schwarzschild spacetime, it has been used to describe the transport of spinors along accelerated worldlines under external force but no torque. The tensor transport law for tetrads is

$\Omega^a{}_{b}=u^a a_b-u_b a^a,$0

and the spinor transport operator is written as

$\Omega^a{}_{b}=u^a a_b-u_b a^a,$1

with

$\Omega^a{}_{b}=u^a a_b-u_b a^a,$2

Here $\Omega^a{}_{b}=u^a a_b-u_b a^a,$3 is the usual spin connection and $\Omega^a{}_{b}=u^a a_b-u_b a^a,$4 is the acceleration-dependent Fermi–Walker correction (Bakke et al., 2015).

For circular motion in Schwarzschild spacetime, this formalism yields a closed-form Wigner rotation angle,

$\Omega^a{}_{b}=u^a a_b-u_b a^a,$5

and the transported EPR state becomes

$\Omega^a{}_{b}=u^a a_b-u_b a^a,$6

The Bell-inequality violation then depends explicitly on $\Omega^a{}_{b}=u^a a_b-u_b a^a,$7. In this setting, Fermi–Walker transport provides a nonrotating local frame in which spin precession reflects acceleration and curvature rather than arbitrary axis rotation (Bakke et al., 2015).

A related but distinct line of work concerns Lie derivatives of spinors. "Spinor Lie derivatives and Fermion stress-energies" does not explicitly discuss the Fermi–Walker derivative by name, but it decomposes $\Omega^a{}_{b}=u^a a_b-u_b a^a,$8 into its antisymmetric Lorentz-rotational part, divergence, and trace-free symmetric shear. The spinor Lie derivative depends on

$\Omega^a{}_{b}=u^a a_b-u_b a^a,$9

which the paper identifies as essentially the anti-self-dual part of $\frac{dw^a}{d\tau}=\Omega^a{}_{b}\,w^b .$0, while the shear and divergence information is carried by

$\frac{dw^a}{d\tau}=\Omega^a{}_{b}\,w^b .$1

This suggests a natural formal language for spinorial formulations of Fermi–Walker-like transport, because the antisymmetric Lorentz generator is precisely the structure that ordinary Fermi–Walker transport isolates along a timelike curve (Helfer, 2016).

6. Characterizations, extensions, and recurring misconceptions

Several recurring identifications in the literature can be stated precisely. First, Fermi–Walker transport is not, in general, the same as parallel transport. The equivalence holds only in the inertial or geodesic case, when $\frac{dw^a}{d\tau}=\Omega^a{}_{b}\,w^b .$2. For accelerated motion, Fermi–Walker transport is the natural replacement for parallel transport because it preserves a nonrotating local frame (Llosa, 2017).

Second, ordinary Fermi–Walker transport does not include arbitrary rotation. Its defining property is exactly the exclusion of intrinsic spatial rotation. Arbitrary rotation enters only after the transport law is generalized by the addition of the proper angular velocity term $\frac{dw^a}{d\tau}=\Omega^a{}_{b}\,w^b .$3. Confusing ordinary Fermi–Walker transport with generalized Fermi–Walker transport obscures the distinction between accelerated nonrotating frames and accelerated rotating frames (Llosa, 2017).

Third, moving frames that are geometrically convenient are not automatically Fermi–Walker or nonrotating. In the Galilean treatment, the Frenet frame and the Darboux frame are nonrotating only in the trivial zero-curvature case of a line. That conclusion is the analogue of the relativistic fact that standard moving frames usually incorporate frame rotation not eliminated by the Fermi–Walker prescription (Şahin et al., 2018).

Fourth, the existence of Fermi coordinates does not always imply that a paper is explicitly studying the Fermi–Walker derivative as an operator. In Robertson–Walker cosmology, the central observers considered are geodesic, and the construction uses a parallel tetrad. Those results belong to the geodesic specialization of the broader Fermi–Walker framework rather than to a separate transport theory (Bolós et al., 2011).

Finally, generalized Fermi–Walker coordinates alter the usual symmetry picture. Their transformations are generalized isometries, and the associated infinite-dimensional extension of the Poincaré algebra is Abelian. The commuting character of the new acceleration and rotational generators is a distinctive result: unlike Lorentz boosts, they commute with one another and with the Poincaré generators (Llosa, 2017).

Taken together, these developments fix the Fermi–Walker derivative as the canonical transport law for nonrotating frames along timelike motion, delimit its geodesic reduction to parallel transport, distinguish it from rotating generalizations, and show how it extends into Galilean geometry, spinor transport, and the symmetry theory of accelerated reference systems.