Sono-Electroplating of Bismuth Films and the Low-Field Hall-Effect Dispute

This dossier page gathers several unrelated items that share one material: bismuth, and one measurement: the Hall effect. Its most substantial content is a published disagreement from the 1920s about what happens to the Hall coefficient of bismuth when the applied magnetic field becomes very weak. Around that dispute sit two modern research abstracts — one on ultrasound-assisted electroplating of bismuth films, one on the magnetotransport properties of electrodeposited bismuth — plus a forum thread on measuring film thickness. The items are adjacent by subject, not by argument: the modern papers do not test the 1920s claim, and nothing here shows that the older author's proposed device was ever built.

The Hall effect and the Hall coefficient

When a current-carrying conductor sits in a magnetic field perpendicular to the current, charge carriers are deflected sideways and a transverse voltage appears across the conductor. This is the Hall effect, discovered in 1875. The Hall coefficient R quantifies its size for a given material and geometry: it is the transverse field divided by the product of current density and magnetic field, and its sign indicates whether the dominant carriers behave as electrons or as holes.

Bismuth is a semimetal with an unusually large Hall coefficient among ordinary metals, which is why it became a standard test material. The dossier's central question is how R behaves as the applied field is reduced toward zero — a regime that is experimentally awkward because the voltages to be measured become very small.

Craig's 1926 claim: an abnormally large low-field coefficient

Palmer H. Craig, working from a 1926 University of Cincinnati doctorate, published "The Hall Effect in Bismuth with Low Magnetic Fields" in Physical Review 27: 772–778 (June 1926). He set out to measure the effect between 0.07 and 1.00 gauss, a range far below what most earlier investigators had used; he notes Righi as a rare exception who worked at fields comparable to the Earth's.

Two methodological problems dominate the paper. First, the films must be extremely thin, since the Hall effect grows as the strip gets thinner, yet they must remain electrically continuous. Craig compared casting, dipping, spraying, electroplating, evaporating and sputtering, and reported that evaporating, sputtering and metallic spraying were the most successful. Casting required pressure on the cooling metal and allowance for lateral expansion; dipping mica sheets into molten bismuth reportedly gave surprisingly thin, uniform films. Second, the voltages are tiny, so he refined his potentiometer and measuring circuits to read reliably to one-tenth of a microvolt.

His headline result is an anomaly. For one film, R was reported as −171 at 0.07 gauss, against −11 at 15 gauss and −29 at 4220 gauss, with a plotted curve showing −R falling rapidly between 0.07 and 0.30 gauss. Craig treats this as a genuine low-field phenomenon that theory would have to accommodate.

He also proposed an application. Stacking the Hall potentials of several films in series, he argued, yields a comparatively high Hall emf that could serve as an alternating-current rectifier in radio and similar uses. He describes using this "additive principle" for Hall-effect rectification in a manner similar to a method attributed to Descoudres, and claims a Hall potential of several volts in low fields with thin bismuth films. The dossier contains no record that such a rectifier was built or independently tested; the proposal is an assertion about what the effect might be used for, not a demonstrated performance.

Heaps's 1927 rebuttal: a null result and an error mechanism

C. W. Heaps replied in Physical Review 29(2): 332–336 (1927), "The Hall Effect in Bismuth with Small Magnetic Fields." He measured a bismuth plate of dimensions 0.011 × 0.9 × 2.0 cm over 0.07 to 2.40 gauss, with a primary current of 1.3 A, each recorded R being the average of at least five values. The average R in that range was 11.5, and the variation of R with field within the range was smaller than the experimental error. At higher fields the coefficient of the same specimen fell from 13.5 at 650 gauss to 5.9 at 8600 gauss.

Heaps concludes that Craig's data for comparable field ranges are erroneous, and offers two candidate causes: insulation leakage, and uncompensated thermomagnetic effects. He adds a procedural criticism — Craig's paper does not show that spurious temperature effects were averaged out by reversing the magnetic field, and does not test the effect of the field on the relevant potentials. Heaps also describes a simple method of making very thin bismuth plates: a slightly drawn-down glass tube is clamped vertically, a bismuth rod inserted and held by the constriction, and a clean surface placed about 10 cm below the tube end. The plate he used was supplied by Eimer and Amend and listed as c.p. (chemically pure).

The two papers therefore stand in direct contradiction on the same quantity in the same field range. Heaps's result is an independent measurement with a stated error budget and a plausible physical mechanism for the discrepancy; Craig's is the original anomaly. The dossier presents both without resolving them, and that is how the disagreement should be read.

Sono-electroplating of bismuth from a Bi(III)-EDTA bath

A separate item is A. Chiba and T. Kojima, "Sono-electroplating of Bismuth Film From Bi(III)-EDTA Bath," e-Journal of Surface Science and Nanotechnology 7 (2009), 688–692. Sono-electroplating means electroplating with ultrasonic agitation applied to the bath. The reported rationale is that ultrasound generates micro-jets and shock-wave pressure at the electrode surface, accelerating mass transfer so that bismuth ions are supplied to the surface more effectively, and modifying the crystallization process.

The electrolyte and operating conditions are reported as follows:

The authors report that sono-electroplating proceeded smoothly because mass transfer was accelerated by ultrasonic agitation, that the exchange current density and reaction rate constant were higher under sonication than in the stationary state, and that the deposited film was smoother and denser than under stationary conditions. They attribute the smoothing mainly to shock-wave pressure, with the accelerated electron reaction attributed to micro-jets or shock-wave pressure.

These are process and morphology results. The paper does not report Hall measurements on the films, so it neither supports nor contradicts Craig's low-field anomaly.

Magnetotransport in electrodeposited bismuth films

A third item is B. O'Brien, M. Plaza, L. Y. Zhu, L. Perez, C. L. Chien and P. C. Searson, "Magnetotransport Properties of Electrodeposited Bismuth Films," J. Phys. Chem. C 2008, 112(31), 12018–12023. Polycrystalline bismuth films were deposited on gold from bismuth nitrate solution. The reported findings: film texture depends strongly on deposition potential and Bi(III) concentration but only weakly on thickness; film morphology depends strongly on deposition potential and on thickness; and the magnetoresistance of the as-deposited films depends strongly on morphology. The abstract also refers to a quantum size effect (QSE) and to thickness dependence of the Hall coefficient that disagrees with the predictions of an infinite-potential-well model, with better agreement under a less rigid boundary condition; temperature dependence of the transport coefficients was found to differ from bulk bismuth.

This is a modern, peer-reviewed study of how deposition conditions shape bismuth film structure and transport. It is not a replication of Craig's experiment and does not address the 1926 anomaly.

A forum aside on thickness measurement

The dossier also includes a Physics Forums thread on measuring the thickness of a thin bismuth film for a Hall-effect experiment. One participant suggests that placing a coil next to a thin conducting film and applying AC will make the coil's inductance depend on film thickness, so long as the film is thinner than the skin depth at the driving frequency, potentially giving about 1% precision with an oscilloscope and a function generator. This is informal commentary rather than evidence, and it is included here only to note that film thickness — the quantity Craig and Heaps both had to control — remains a practical measurement problem.

What the evidence does and does not establish

The Craig–Heaps exchange is a genuine published dispute in a mainstream physics journal, with an independent measurement and a named error mechanism set against an anomalous result. That much is documented. What is not documented here is any resolution: no later replication of Craig's low-field values appears in this material, and the absence of such data in a partial excerpt cannot be read as proof that none exists.

The rectifier proposal is likewise unverified. Craig asserts that series-stacked films give a usable Hall emf for AC rectification; the dossier contains no built device, no efficiency figure and no independent test. The sono-electroplating and magnetotransport papers are sound work on bismuth film preparation and properties, but they are separate studies with separate aims, and neither was designed to test the 1920s claim. Reading them as confirmation of Craig would be a category error.

Related work and context

The dossier is hosted as an aggregated page on the source archive, the inventor-and-article archive described in sources/main-archive and sources/library-annex. Its structure — an original claim, a rebuttal, and later adjacent research — resembles other claim-versus-evidence dossiers in this collection, though the subject here is a conventional measurement dispute rather than a free-energy or perpetual-motion assertion. The distinction matters: Craig's anomaly, if real, would be a refinement of accepted transport theory, not a violation of thermodynamics.