Its row in the atlas →

White et al. 2021: the worldline-numerics Casimir pattern

A DARPA-funded numerical study of Casimir energy around a micron pillar, whose plot resembled the Alcubierre toroid and was reported as the first real warp bubble.

Kind: physics · Loophole: L2 · Standing: K1 · Bill: B-neg · Last reviewed: 2026-09-12

The claim

Harold White and four co-authors were modelling custom Casimir cavities for a DARPA project on vacuum energy storage. The cavity of interest is a pair of parallel plates about 4 µm apart with 1 µm pillars along the mid-plane, designed to test White's "dynamic vacuum model", which predicts that the negative Casimir energy density in a cavity is not uniform but concentrated along the mid-plane, and that this structure comes with a small electrostatic field that a high-impedance oscilloscope could see as a transient voltage. The engineering question was whether a pillar would screen itself from that field. To answer it the team implemented worldline numerics, a Monte Carlo method for Casimir energies developed by Gies, Langfeld and Moyaerts, validated it on plate-plate and plate-sphere cases and ran it on the plate-pillar geometry.

The paper's claim is what they saw when they plotted the result. The negative energy density around the pillar, in a two-dimensional section, "was observed to be qualitatively quite similar to a two-dimensional representation of energy density requirements for the Alcubierre warp metric": a lenticular concentration of negative energy either side of a central body, which in the Alcubierre case is the toroidal ring of negative density around the flat interior. A toy model of a 1 µm sphere in a 4 µm cylinder gave a three-dimensional toroidal Casimir density that "correlates well with the Alcubierre warp metric requirements". On that basis the authors suggest a chip-scale experiment: route a pulse of electrons or photons through the sphere and compare its transit time with free space; a difference "would be an empirical confirmation of the generation of a real nano scale warp bubble on a chip", and they insist this would be "a genuine implementation of the idea in physical fact", if only "a real, albeit humble, warp bubble".

What the paper does not claim is as important. It does not solve the Einstein equations for the cavity, does not write down a metric sourced by the computed density, does not assign the notional bubble a velocity, does not compare the magnitude of the Casimir density with what any Alcubierre bubble of that size would need and reports no measurement. The claim is a qualitative resemblance between two plots and a proposal to look.

Origin and lineage

White, Vera, Han, Bruccoleri and MacArthur, Limitless Space Institute, Texas A&M and Izentis, Eur. Phys. J. C 81, 677 (2021), received 17 May, accepted 25 July, published 31 July 2021, open access under SCOAP3, funded by the DARPA Defense Science Office QUEST programme; no arXiv version exists [HIGH] S1. It is the continuation of White's earlier warp programme at NASA, the Eagleworks warp field interferometer, and of his "Warp Field Mechanics 101" and "102" reports, which it cites for the thick-wall energy reduction and for the "boost" reading of the Alcubierre 1994 metric. The Casimir idea as an exotic-matter source goes back to Morris and Thorne, Visser and Alcubierre himself, as the paper notes. It is priced by ENE-6, the register's Casimir floor, and by ENE-3, the Alcubierre bill. The press lineage runs through White's talk at the AIAA Propulsion and Energy Forum in August 2021 to the December 2021 headlines about the world's first warp bubble.

The mechanism

Worldline numerics evaluates the Casimir interaction energy of a scalar field with Dirichlet boundaries by averaging over an ensemble of closed random loops. The effective action for a field of mass $m$ in a background potential $V$ representing the geometry is

$$\Gamma[V] = -\frac{1}{2(4\pi)^2}\int_{1/2}^{\infty}\frac{dT}{T^3}\,e^{-m^2T}\int d^4x\,\big\langle W_V[y(t);x] - 1\big\rangle_y,$$

with the Wilson-loop factor $W_V = \exp\big(-T\int_0^1 dt\,V(x + \sqrt T\,y(t))\big)$ averaged over unit loops $y$ with Gaussian measure [HIGH] S1. In the Dirichlet limit for a massless field in 3+1 dimensions the interaction energy becomes

$$E_{\text{Casimir}} = -\frac{1}{2(4\pi)^2}\int_0^\infty\frac{dT}{T^3}\int d^3x_{\text{CM}}\,\big\langle\Theta[x(\tau)]\big\rangle_x,$$

where $\Theta = 0$ if the rescaled loop misses the geometry and $1 - n$ if it pierces $n \ge 1$ bodies, discretised over $n_L$ loops of $N$ points each generated by the "v-loop" algorithm [HIGH] S1. The quantity at each grid point is the loop-weighted interaction energy density, with single-body self-energies subtracted; the method "can currently only assess idealized behaviour for bounding geometry and cannot assess any frequency dependence of materials", and "does not account for the impacts of temperature" [HIGH] S1.

The runs: a plate-sphere validation on a $50^3$ grid with 2000 loops on 100 CPUs; the plate-pillar case, a 4 µm cavity with a 1 µm pillar, on a 35 × 35 grid over ±4 µm; and the sphere-in-cylinder toy model, 1 µm sphere in a 4 µm cylinder, on a 100 × 100 grid with 2000 loops on 660 CPUs [HIGH] S1. The pillar does not screen itself; it raises the negative density inside itself by a factor of 3 to 5 over the plates-only value [HIGH] S1. The dynamic vacuum model's predicted mid-plane potential for a 4 µm gap is about 0.7 mV [HIGH] S1.

The comparison is to the Alcubierre metric in its original form with the tanh shape function, its York time $\theta = v_s\,(x_s/r_s)\,df/dr_s$, and its Eulerian energy density

$$T^{00} = -\frac{1}{8\pi}\,\frac{v_s^2\,(y^2 + z^2)}{4 r_s^2}\Big(\frac{df}{dr_s}\Big)^2,$$

whose toroidal shape the paper's figure 9 sets beside the pillar plot [HIGH] S1. The paper notes that the Casimir plot is a linear extrusion (rod-like) while the Alcubierre density is a solid of revolution (toroidal), which is why the sphere-in-cylinder model was built [HIGH] S1. It also repeats White's earlier finding that thickening the Alcubierre wall lowers the peak density and total energy at the cost of interior volume [HIGH] S1.

What the resemblance is not. The Casimir density is a quantum expectation value for a scalar field between conductors, and the mean value in a 4 µm gap follows from the register's formula $-\pi^2\hbar c/720d^4$ as about $-1.7 \times 10^{-6}$ J m$^{-3}$, so the pillar's factor of 3 to 5 puts the peak near $-10^{-5}$ J m$^{-3}$ [HIGH] S1. An Alcubierre bubble the size of the pillar, radius half a micron and wall thickness of the same order, would need by the density formula above a peak $|T^{00}|$ of order $(c^4/8\pi G)\,(v_s/c)^2\,(1/4)\,(1/\Delta)^2 \approx 5 \times 10^{54}\,(v_s/c)^2$ J m$^{-3}$ for $\Delta = 0.5$ µm, an estimate made here from the Alcubierre expression: at $v_s = c$ that is about sixty orders of magnitude above the Casimir figure, and the gap closes only as $(v_s/c)^2$, so even a bubble crawling at a metre per second would need some forty orders more than the cavity holds [MED] S1. The paper never states a velocity for its notional bubble and never makes this comparison, so the "warp bubble" is a pattern of the right shape at a magnitude some sixty orders too small. The proposed transit-time test would be a measurement of light or electron propagation through a micron-scale conducting structure, which the register prices under LOR-3 and LOR-4: a nanostructure can shift a pulse's group delay without any front outrunning $c$, and light between Casimir plates has a phase velocity above $c$ by a part in $10^{32}$, unmeasurable and causally harmless.

How it was reported. In December 2021 The Debrief ran the story under the headline that DARPA-funded researchers had accidentally created the world's first warp bubble, quoting White that the analysis had identified a structure "predicted to generate a negative vacuum energy density such that it would manifest a real nanoscale warp bubble", and the story was picked up across the technology press; White added that this did not mean a working warp drive was near [LOW] S4. The paper's abstract says "qualitatively quite similar" and "would suggest that chip-scale experiments might be explored"; nothing was created and nothing was measured [HIGH] S1.

What it costs

B-neg, and nothing to pay it with. The proposal is the Alcubierre metric, which the paper cites for its negative-energy toroid and which needs the energy conditions to fail; the Casimir effect is the only negative energy density anyone has made, and the register's floor puts it tens of orders of magnitude below any warp bill. The paper's own numbers, once the comparison it omits is made, put the shortfall at roughly forty to sixty orders for a bubble of the pillar's size at any speed worth the name. No B-mass mark is given because no mass figure is claimed or derivable: the paper assigns its bubble no velocity. No B-new: the physics invoked, Casimir energy and the Alcubierre metric, is standard; the "dynamic vacuum model" that motivated the cavity is White's own and is not the mechanism of the warp claim. No B-caus because no speed is asserted.

Constraint scoring

Every entry in the register, one row each, verdict from the physics vocabulary in GRADING.md, and a note that says why.

ConstraintVerdictNote
CAU-1SILENTAn Alcubierre bubble on a chip is proposed with no velocity and no word on causality; the entry applies to any such bubble that is superluminal
CAU-2SILENTChronology protection is never mentioned
CAU-3SILENTTwo devices are never considered, though the proposed experiment is an array of many
CAU-4SILENTNo frame is discussed
CAU-5N/ANo entanglement signalling
ENE-1VIOLATESThe bubble sought is Alcubierre's, which needs negative energy density by the very formula the paper reproduces; the paper embraces that need and offers Casimir energy to meet it
ENE-2SATISFIESMakes no claim to beat the theorem: it accepts that a superluminal bubble needs negative energy and goes looking for some
ENE-3SILENTNever computes the energy or wall thickness an Alcubierre bubble of the pillar's size would need at any speed, so the Pfenning–Ford bounds are never confronted; the estimate above shows they would not be met
ENE-4N/ANo pocket geometry
ENE-5SATISFIESClaims no positive-energy drive; its bubble is Alcubierre's and violates the null condition as the entry says, which the paper does not dispute
ENE-6VIOLATESThe proposal is precisely to use Casimir energy as the bubble's source; the entry records the floor as tens of orders short of any bill, and the pillar's factor of 3 to 5 over $-1.7 \times 10^{-6}$ J m$^{-3}$ does not change that
ENE-7N/AA nanoscale warp bubble of trivial topology; no wormhole, so the theorem does not bear. The negative energy any superluminal bubble needs is ENE-2's verdict, and ENE-7 is that entry's wormhole form.
CON-1SILENTHorizons and control are never discussed
CON-2SILENTThe bubble is to be fabricated in place, which is the entry's point, but the paper never addresses how a bubble would be made to move or what lies ahead of it
CON-3N/ANot a tube
STA-1SILENTSemiclassical stability never discussed
STA-2N/ANo throat
STA-3N/ANo time machine
HAZ-1SILENTNever discussed
HAZ-2SILENTNever discussed
HAZ-3N/ANo throat
LOR-1SATISFIESNothing is pushed toward $c$; the mechanism invoked is geometric
LOR-2N/ANo tachyons
LOR-3SILENTThe proposed test is a transit-time comparison for a pulse through a micron structure, exactly the measurement the entry warns about: group delay in a structure is not a signal velocity, and the paper does not distinguish them
LOR-4SILENTLight between Casimir plates is the Scharnhorst case, a phase velocity above $c$ by a part in $10^{32}$, unmeasurable and benign; the paper proposes to time photons through a Casimir structure and never mentions it
WRP-1SATISFIESThe Alcubierre metric verbatim, with its shape function, York time and energy density
WRP-2SATISFIESAccepts that a warp bubble needs exotic matter and proposes Casimir energy as the exotic material
WRP-3N/ANo positive-energy claim is made
MAN-1N/ANo extra dimensions in this paper; White's earlier Chung–Freese argument is not invoked here

Status of the argument

Sources