This is a blog about Applied Vacuum.net-theory.
J.Konstapel,Leiden,6-10-2026.
© 2026 J. Konstapel, Constable Research, Leiden. All rights reserved. Quotation with attribution is
permitted; reproduction of the whole requires written permission
For a century, physics treated the vacuum as a floor. It was set to zero. Everything that happened was something lifted above it and allowed to fall back. Light, motion and heat came at the end of a chain that began with fire. Coal was burned. Water became steam. Uranium was split.
Three lines of work now describe a different situation. They started separately. They were not coordinated. They do not share a vocabulary, a method or a starting point. Yet they arrive at the same structural claim. The vacuum has a local state. That state is set locally by what sits in it. What persists is what fits.
The first line is a measurement. Garret Moddel, at the University of Colorado Boulder, has measured electrical current from a metal-insulator-metal structure built against an optical cavity. The result was published in a peer-reviewed journal. It carries eight artefact tests and a six-month ageing record.
The second line is a spectrum. Harold White and colleagues at Casimir, Inc. in Houston have derived the hydrogen spectrum from a dynamic vacuum with a constitutive profile. The paper appeared in Physical Review Research in March 2026. Quantization is not postulated. It emerges from symmetry, boundary conditions and causal response in a medium that varies in space and time.
The third line is a net. The Vacuum.Net theory, developed in Leiden by J. Konstapel, derives the same kind of result from five axioms about a single strand, a ternary register and closure. Its picture is a fishnet stretched in water. Strands run everywhere, under tension. Where strands wind around each other, knots appear: matter. Between strands lie meshes: openings with a shape and a size. A winding persists only if it returns in phase inside its mesh. What does not fit does not stay.
A fourth document completes the picture. The Dutch Vacuum Landscape maps the research groups in the Netherlands whose running work touches the same quantities: switching the force, tuning the cavity, moving a medium toward its critical point, and measuring membranes in continuous motion. It finds that three of the four needed quantities already have Dutch programmes behind them. The fourth is untouched.
This essay sets the material side by side. It states where the lines agree. It states where they do not. And it names the one number that stands between a theory of the vacuum and a device.
2. The measurement: a device that produced current
Garret Moddel is Professor Emeritus of Photonics and Quantum Engineering at the University of Colorado Boulder. His group worked for years on ultra-fast metal-insulator diodes for optical rectennas. Separately, it worked on the question of whether the quantum vacuum can be tapped.
His record is worth stating in full, because it is not the record of a believer. In 2009 he published an evaluation of vacuum energy extraction methods. He analysed the published patents and papers from first principles. He concluded that most of them contain fundamental errors and cannot work. A person who has publicly dismantled most of a field, and then publishes a measurement inside it, is not easy to dismiss.
His 2021 paper in Symmetry reports current from a structure of an aluminium mirror, a thin filling and a palladium electrode. The structure produced optical-cavity-induced current, with no applied voltage. The paper carries eight tests for measurement artefacts. It carries a six-month ageing record. His own later summary describes thousands of trials, dozens of variations and replications.
What the measurement establishes is small and exact. A cavity structure of this kind produces current. The number is 1.4 × 10⁻¹² W from an area of 0.02 μm². Scaled to an opening of one natural unit of space, that is 1.5 × 10⁻¹³ W. This is the only measured value of its kind in the literature. It functions as a zero: what one particular arrangement delivers, against which everything else is measured.
There is a second independent structure. Thibado and colleagues measured fluctuation-induced current from free-standing graphene, published in Physical Review E in 2020. A free-standing membrane moves continuously at room temperature. Its motion is not classical Brownian motion. It shows Lévy flights with curvature inversion, and it drives a current through a circuit of two opposed diodes. It is the nearest thing to a replication of the asymmetry principle in another laboratory, in a completely different material system.
3. The spectrum: quantization that was not postulated
Harold White led NASA’s Advanced Propulsion Physics Laboratory at Johnson Space Center. He then founded Casimir, Inc. in Houston. In March 2026, White, Vera, Sylvester and Dudzinski published “Emergent quantization from a dynamic vacuum” in Physical Review Research 8, 013264 — a peer-reviewed, open-access journal of the American Physical Society.
The paper does something structurally unusual. It models the vacuum as an acoustic medium with quadratic temporal dispersion and a radially varying profile imprinted by a proton. From that profile it writes an inverse sound speed of the form 1/c_s²(r) = A(ω) + C(ω)/r. This makes the time-harmonic operator Coulombic at each bound eigenfrequency. Separation of variables then yields the hydrogenic eigenfunctions exactly. Identifying the spatial scale with the frequency gives ω_n = 1/n², which is the Rydberg ladder. The angular labels follow from rotational symmetry on the sphere.
The conclusion matters here more than the formalism. Quantization is not put in as a postulate. It emerges from symmetry, boundary conditions and causal response in a vacuum that varies in space and time.
The paper does not prove that the vacuum is such a medium. It proves that if it is one, the discrete structure of matter follows without a separate quantum postulate. That is the first such statement to pass peer review at the American Physical Society.
The commercial context is part of the record. Casimir, Inc. raised a twelve million dollar seed round in May 2026. It has since received a SpaceWERX Phase I award from the Air Force Research Laboratory. Its stated target is a first chip in 2028, for ultra-low-power devices. The company is explicit that the development trajectory is early. The sceptical view — that net extraction runs into hard thermodynamic constraints, and that early prototypes delivered output in the picoampere range — is widely held and belongs in any honest account.
4. The net: a theory built from five axioms
The Vacuum.Net theory was developed in Leiden, along a line that goes back through context-sensitive physics to work on knots and strands. Its published foundation does not begin with the vacuum. It begins with distinction.
Five axioms carry it:
- N1, Strand. There is one strand. Every distinction is the strand touching itself.
- N2, Closure. A configuration persists only if it closes on itself.
- N3, Memory. The net keeps the orientation of what was first laid down.
- N4, Context. Constants are state functions of the local vacuum, with universality as the limiting case.
- N5, Scale. The same winding appears at every depth.
The register is ternary. Each strand state is +1, 0 or −1.
From these axioms the theory derives a constitutive law: g_eff = g(1 + γδ), where δ is the local density deviation. The sensitivity is γ. And γ equals the inverse of the distance of the net to its critical point, ε. Calibrated against the acceleration threshold measured in galaxy rotation curves, that distance comes out near 0.159 and the sensitivity near 2π. The theory records this as a numerical correspondence, not as a proof.
The theory keeps a five-way status ledger for every claim: assumed, conditionally proved, corresponded, measured, or open. That discipline is rare. It is the reason the comparison in this essay can be made at all.
The theory also fixes a natural unit of space. From the Rydberg frequency and the speed of light, the unit of time is 1.52 × 10⁻¹⁶ s and the unit of space is 45.6 nm. Inside the time region, divided by Larson’s inter-regional ratio of 156.444, the unit becomes 0.291 nm — which is the nearest-neighbour spacing of aluminium, silver and gold. The theory arrives at these figures without fitting them to anything.
5. The generator: a design taken from the axioms alone
Applied Vacuum Theory, Part Five turns the theory into a design. Its element is a mesh defined by the natural units of the theory, with deliberately different boundary conditions on its two sides.
A generator on this ground is not a machine that makes energy. Energy is a winding, a strand closed on itself. It is not made; it passes. A generator is a structure in which the boundary conditions on one side differ from those on the other. The difference between the two sides is the proposed output.
The system has four conceptual parts. An element, which is one mesh. A count, which is how many elements are switched in. A wall, which determines the boundary response. And a hold, which maintains the relevant configuration. The published design deliberately reduces the number of independent variables.
The closed wall is not treated simply as a mirror. It is a structured boundary whose response is periodic on the scale specified by the theory. The open wall is deliberately different and does not reproduce the same boundary condition. The asymmetry is the essential property of the element.
The important quantity is the fraction of the closed configuration that is returned in phase at the open boundary. This quantity is material- and structure-dependent. It is therefore the experimentally unknown parameter, rather than something that can legitimately be assumed from the theory.
Output is proposed to scale with the number of active meshes. This makes the architecture scalable in principle, but the scaling remains a prediction until independently demonstrated. The existence of a scaling relation must not be confused with a demonstrated power source.
The published design therefore establishes a testable architecture rather than a working product. The numerical performance estimates are conditional on the unknown coupling parameter and are not experimental results.
The build order follows the same logic: first establish a controlled single-element response; then vary the relevant geometrical and material parameters; then test scaling, combinations and long-duration stability. The essential requirement is that each stage be independently measurable and compared with an appropriate control.
The specific geometry, material sequence, layer thicknesses, dimensional tolerances, coupling arrangement and implementation parameters are intentionally not reproduced here. They belong to the engineering implementation rather than to the public statement of the physical hypothesis.
6. The Dutch landscape: who could build it
The Dutch Vacuum Landscape asks a practical question. Which existing research groups already work on the quantities the net generator needs? Four quantities matter: the size of the mesh, the asymmetry between its two walls, the distance of the medium to its critical point, and the coupling on the open side. Dutch research covers the first three. The fourth is untouched.
Groningen. Prof. George Palasantzas holds the chair of Surface Interactions and Nanostructures at the Zernike Institute. He chaired the European Casimir network of more than sixty groups. His programme switches the Casimir force in place, by reversible amorphous-to-crystalline phase transitions in phase-change materials. The reported contrast is 20 to 25 per cent between gold and AIST on crystallisation. In the net’s terms, this is a measured δ-setting. It is the closest thing in the Netherlands to setting the local state of the medium. Funding is current, through two NWO grants, and a 2025 publication extends the work to GST thin films.
Twente. Dr Vitaly Svetovoy is the theoretical partner on much of the Groningen work. His roughness correction to the Casimir force, written with Broer, Palasantzas and Knoester, addresses exactly the case where surface roughness becomes comparable with the separation — the regime of a mesh at one unit of space.
Eindhoven and AMOLF. Prof. Jaime Gómez Rivas pioneered metal nanoparticle arrays for strong light-matter coupling. His field states its own premise plainly: its founding paper is titled “Modifying Chemical Landscapes by Coupling to Vacuum Fields.” Reaction rates, ionic conductivity and molecular polarisability have been reported to change under coupling to a cavity field, with no illumination. Dr Said Rodriguez at AMOLF works with open-access tunable microcavities, where the mirror separation is a continuously adjustable parameter — in net terms, a mesh whose size is a control variable. His group’s stochastic thermodynamics work addresses the exact objection a reviewer raises against the generator.
Amsterdam. Prof. Peter Schall works with the critical Casimir force: the thermodynamic analogue of the quantum effect, arising when two surfaces sit in a binary liquid mixture near its critical point. The correlation length depends on temperature, so the force is tunable by temperature alone. This is the one Dutch system where the net’s ε is set by a dial. If the claim that sensitivity equals 1/ε is right, it should hold there, and it is measurable there.
Delft. Prof. Peter Steeneken measures suspended graphene and MoS₂ membranes: resonance, thermal motion, nonlinear dynamics. This is the Dutch group that holds the material Thibado uses. In net terms, a rippling membrane with two opposed diodes is a mesh with one open passage. The capability is there. The question is not asked.
The comparison yields a hard finding. Of the four quantities the net generator needs, three have Dutch programmes behind them. The fourth — how much of what the mesh closes leaves as current — has none. That is where the experimental uncertainty remains, and it is unworked in this country.
7. Where the three lines converge
All three give the vacuum a state. Moddel’s structure works because a cavity on one side and an open electrode on the other make the two sides different. White’s model works because a profile imprinted by a proton makes the medium vary with radius. The net’s χ is the same move stated as an axiom. None of the three treats the vacuum as uniform.
All three replace a postulate with a shape. White derives the energy levels of hydrogen from boundary conditions in a dispersive medium, instead of postulating quantization. The net derives which windings persist from what returns in phase in a mesh, instead of postulating a particle. These are the same manoeuvre on the same problem, reached from opposite directions — White from Madelung hydrodynamics, the net from strands, closure and a ternary register.
All three produce a constitutive relation rather than a mechanism. White’s 1/c_s²(r) = A(ω) + C(ω)/r relates the medium’s response to its profile at every point simultaneously. The net’s g_eff = g(1 + γδ) does the same. Neither is a causal chain. Both say what holds at once. That is a notable agreement, because it is the kind of statement a mechanism-based physics does not naturally produce.
All three locate the action below a few hundred nanometres. Moddel’s filling is 33 nm. The net’s unit of space is 45.6 nm, taken from the Rydberg frequency and c without fitting. The Casimir regime generally sits below a few hundred nanometres. Three independent routes agree on the size range.
And the convergence is strongest where the lines did not borrow. Moddel’s device is prior art to Casimir, Inc., and the two American lines share a research tradition. The Leiden line is independent of both. Its published foundation predates the March 2026 paper, and it reaches the constitutive form from axioms that have nothing to do with acoustics. Convergence between lines that did not borrow from each other is worth more than agreement between lines that did.
8. Where they do not converge
The three are not one theory. Saying so would weaken all of them.
Different objects. White’s medium is acoustic, with a sound speed and a dispersion relation. The net has one strand, a ternary register and closure. These are not translations of each other. No translation has been written. Whether one can be written is an open and interesting question.
Different status. Moddel has a measurement. White has an analytic result that passed peer review. The net has a derivation with a status ledger and one numerical correspondence. These are three different kinds of claim. They cannot be added together as if they were evidence of the same proposition.
Different accounts of where the energy is. Moddel’s framing is a borrowed fluctuation, captured before it is returned. White’s framing is resonance in a dispersive medium. The net’s framing is that nothing is made or taken at all: a winding is the strand, closed, and the only question is where and in what form a mesh lets tension close. These readings would make different predictions about what a device can do over time.
An asymmetry of framing. The Dutch lines treat the vacuum field as a boundary condition on something else: a force between bodies, a shift in a molecular level, an attraction between colloids. The net treats it as the one strand, and the bodies as windings in it. The measurements do not distinguish the two readings. The difference shows only in what one asks next. Within the first framing, an asymmetric mesh is not a natural question. Within the second, it is the only question.
9. The one open number
Every one of the three lines stops at the same place. A structure with two unlike sides closes something on one side and produces a measurable response on the other.
The closed side is settled. Lifshitz theory computes it from measured dielectric functions. Casimir’s original count computes it from the geometry. The net computes it from the ratio of space to time. All three agree to within what the measurements can resolve.
The open side is not settled. Of what a structure closes, how much leaves as current? Moddel’s measured structure gives the only number there is. It is small: a fraction of order 10⁻⁹. Nobody has shown whether that fraction is a ceiling, a property of the particular materials he used, or the first point on a curve nobody has mapped.
That single unknown is where the distance lies. Not in the closed wall. Not in the general size range, which three independent routes now agree on. In the coupling on the open side — where no group in Europe is working, and where the measured record consists of one device.
The exact implementation of that coupling is deliberately not specified here. It is the subject of experimental work, not a conclusion of this essay.
10. What would settle it
The question is narrow enough to be answered by an experiment that is small by the standards of a semiconductor laboratory. It needs a thin-film line, a structure with two unlike boundary conditions, a source meter, and discipline. The discipline is the part Moddel already brought to his own measurement: a blind assignment, a sham structure with no phase relation, a galvanically isolated reading, a full current-voltage curve rather than a single point, and a prediction fixed before the run.
The work in Leiden has reached the point where the lattice side can be settled computationally. The physical side cannot be settled from a desk. It needs someone with a line.
The practical routes run through the Dutch groups named above. Groningen can answer what a structured boundary returns compared with a conventional reflective boundary. AMOLF can answer what happens when the two sides of a tunable cavity are made unlike each other — for that group, asymmetry is a change of sample, not of apparatus. Amsterdam can answer whether sensitivity equals inverse distance to the critical point, in the one medium where that distance is a temperature setting. Delft can answer what a suspended membrane does when the circuit behind it is asymmetric.
The precise implementation proposed by Vacuum.Net is not required to establish the scientific question. The decisive result is a reproducible, independently controlled output whose dependence on the relevant structural variables can be predicted before measurement.
11. Conclusion
Three lines of work, started separately and for different reasons, now describe the vacuum the same way: as a medium with a local state, not as an empty background. One has a measurement. One has a peer-reviewed derivation. One has an axiomatic theory with an explicit status ledger. They agree on the structure of the claim, on the constitutive form it takes, and on the size range where it acts. They disagree on the objects, on the status of the claims, and on where the energy sits.
Between that description and a device stands one quantity: the coupling on the open side of an asymmetric structure. It is measurable. The experiment is small. The existing record consists of one device, built by one group, reported in one paper.
Nothing in this record is offered as a working product. What is stated is narrower and more useful. Three lines arrived independently at the same description of the vacuum. One quantity separates that description from a device. And the quantity is measurable — by any laboratory with a thin-film line and the discipline to fix its prediction before the run.
The engineering implementation remains undisclosed.
Annotated references
The three source documents
Konstapel, J. (2026). Three Roads to a Full Vacuum: A Measurement, a Spectrum and a Net. Constable Research, Leiden, 6 October 2026. Why read? The essay that first set the three lines side by side and named the open question they share. It is the backbone of this synthesis. Reading advice: Sections 5 and 6, on convergence and non-convergence, are the core; Section 8 states the experiment that would settle the open side.
Konstapel, J. (2026). The Dutch Vacuum Landscape. Constable Research, Leiden, 6 October 2026. Why read? The mapping of four Dutch research lines onto the quantities the net needs, with named groups, current funding and a contact list. Its central finding: three of four quantities have programmes behind them; the open side has none. Reading advice: Section 7, on what the comparison yields, and Section 9, which states the one question each group is positioned to answer.
The measurement
Moddel, G., Weerakkody, A., Doroski, D., and Bartusiak, D. (2021). “Optical-cavity-induced current.” Symmetry 13, 517. Why read? The one measured device of its kind: the stack, the 1.4 pW from 0.02 μm², the eight artefact tests and the six-month ageing record. It fixes the zero of the open-side coupling. Reading advice: Table 1 and Figure 4, then the artefact section, which is the part that makes the rest worth reading.
Moddel, G. (2009). “Something from nothing: an evaluation of vacuum energy extraction methods.” University of Colorado. Why read? The author takes apart most of the published proposals in his own field from first principles and concludes they cannot work. It establishes the standard he applies to his own later measurement.
Moddel, G. (2022). “Zero-point energy: capturing evanescence.” Journal of Scientific Exploration 36(3). Why read? His own account of the operating principle he proposes, and of where he thinks the thermodynamic objections do and do not apply. Reading advice: read the thermodynamics section against his 2009 paper.
Thibado, P. M., et al. (2020). “Fluctuation-induced current from freestanding graphene.” Physical Review E 102, 042101. Why read? A second independent structure with one open passage, in a completely different material system. The nearest thing to a replication of the asymmetry principle in another laboratory.
The spectrum
White, H., Vera, J., Sylvester, A., and Dudzinski, L. (2026). “Emergent quantization from a dynamic vacuum.” Physical Review Research 8, 013264, published 9 March 2026. Why read? The derivation of the hydrogenic spectrum from a dynamic vacuum with a constitutive profile, with no free parameters and no separate quantization postulate. This is the paper that makes the convergence something other than a coincidence of vocabulary. Reading advice: the constitutive profile and the mapping to the Coulomb operator.
Casimir, H. B. G. (1948). “On the attraction between two perfectly conducting plates.” Proceedings KNAW 51, 793. Why read? The older notation for a two-wall opening, and the count against which every later treatment is checked.
The net
Konstapel, J. (2026). The Vacuum.Net Theory: Foundational Paper, sixth edition. Constable Research, Leiden. Why read? The five axioms, the theorems proved from them, and the five-way status ledger that keeps assumed, proved, corresponded, measured and open apart. Reading advice: the ternary ground first.
Konstapel, J. (2026). Applied Vacuum Theory: Setting the State of the Net, fourth edition. Constable Research, Leiden, 25 September 2026. Why read? The control variables, the constitutive law, the identity γ = 1/ε, and the calculation showing that the fine-structure constant is immovable while the effective coupling is not. Reading advice: Sections 4 and 5.
Larson, D. B. (1959). The Structure of the Physical Universe. Portland: North Pacific Publishers. Reworked as Nothing But Motion (1979). Why read? The natural units of space, time, energy and power from the Rydberg frequency and c, and the inter-regional ratio 156.444. Reading advice: Chapter 13 only.
Williamson, J. G., and van der Mark, M. B. (1997). “Is the electron a photon with toroidal topology?” Annales de la Fondation Louis de Broglie 22, 133–160. Why read? The winding as a photon on a closed loop: the thing the element passes. Reading advice: the geometric first half.
The threshold used in the calibration
Milgrom, M. (1983). “A modification of the Newtonian dynamics as a possible alternative to the hidden mass hypothesis.” Astrophysical Journal 270, 365. Why read? The first statement of the acceleration threshold against which the net’s sensitivity is calibrated.
McGaugh, S. S., Lelli, F., and Schombert, J. M. (2016). “Radial acceleration relation in rotationally supported galaxies.” Physical Review Letters 117, 201101. Why read? The measured relation that fixes that threshold, and the reason the correspondence can be stated as a number rather than an impression.
The Dutch groups
Torricelli, G., et al. (2010). “Switching Casimir forces with phase-change materials.” Physical Review A 82, 010101(R). Why read? The measurement that a Casimir force can be switched in place by a phase transition. The nearest existing realisation of setting δ.
Broer, W., Palasantzas, G., Knoester, J., and Svetovoy, V. B. (2011). “Roughness correction to the Casimir force beyond perturbation theory.” Europhysics Letters 95, 30001. Why read? The case where roughness is comparable with the separation. The nearest treatment of what a non-flat wall returns.
Ebbesen, T. W. (2012). “Modifying chemical landscapes by coupling to vacuum fields.” Angewandte Chemie International Edition 51, 1592. Why read? The founding statement of the field that changes matter with a cavity field and no illumination. The closest existing premise to the net’s own.
Nguyen, V. D., Faber, S., Hu, Z., Wegdam, G. H., and Schall, P. (2013). “Controlling colloidal phase transitions with critical Casimir forces.” Nature Communications 4, 1584. Why read? A fluctuation force used as a dial, with the correlation length set by temperature. The experimental form of the lever 1/ε.
Davidovikj, D. (2018). Dynamics of Interacting Graphene Membranes. Doctoral thesis, Delft University of Technology. Why read? What a suspended membrane does, measured: thermal motion, nonlinear dynamics, squeeze-film effects. The material Thibado’s result is built on.