Trichel · Source catalogue

A 2D finite-element (FEM) plus flux-corrected-transport (FCT) simulation of Trichel pulses, the regular current pulses of a negative point-to-plane corona discharge in air. Geometry: hyperbolic needle 1 cm long, tip radius 100 um, 1 cm from a grounded plane, an external resistance Rext in series (value not given). Three charged species are tracked (electrons, positive ions, negative ions) with Poisson's equation and three continuity equations; constants (mobilities, ionization, attachment, recombination) are cited to a 1997 TRINITI contract report for ABB (ref. 16), secondary emission gamma=0.01 to Morrow 1985. Results for -7, -8, -9 and -11 kV: at -7 kV six pulses with intervals 11.5, 13.38, 10.67, 10.85, 10.4 us and peaks of about 25 mA (first) then about 1.8 mA; at -9 kV peaks about 37 mA then 1.7-2.0 mA, intervals 19.1 us then 3.3-3.5 us. Authors' mechanism: an avalanche near the tip makes electron and positive-ion clouds; positive ions are absorbed by the needle within nanoseconds, which switches the ionization off; slowly drifting negative ions (mobility 2e-4 m2/Vs) must clear away before the tip field recovers to threshold, setting the multi-microsecond period. A small post-pulse (secondary avalanche from returning positive ions) appears in the simulation. The first pulse is much larger because it forms in charge-free space. Table 1 compares simulated average periods with experimental ones from Lama and Gallo (1974): 11.33 vs 21.85 us (7 kV), 6.31 vs 13.32 (8 kV), 3.39 vs 9.09 (9 kV), 2.46 vs 5.07 (11 kV). The authors say the model follows the trend but underestimates periods by about 50%, cause 'currently being investigated'. Deep-read observations (reader's checks, not authors' claims): the paper reports no rise time or width for its own simulated pulses; the printed ionization law ki=10^(11.93-4.493/EN) gives no threshold (a misprint of 449.3 is a possible but unconfirmed reading); eqs (6)-(7) print source terms in Np and Nn instead of Ne; N=2.46e25 m^-3 is total air density but is called oxygen density; N as number versus charge density is ambiguous; ratios num/exp are 0.52, 0.47, 0.37, 0.49 (so 'approximately 50%' is loose at 9 kV); the experimental geometry (tip radius, gap, Rext) is not stated, so the comparison's equivalence is unverified; no mesh or time-step convergence study; delta1 is 0.01 in text and 0.05 in the flowchart. The paper cites 19 references, from Trichel 1938, Loeb 1941, Zentner 1970, Lama and Gallo 1974, Morrow 1985, to Napartovich 1997 and Soria 2007. The text notes G. W. Trichel's original explanation: space charge cycle with a negative cloud far from the point and a positive cloud near it. QE deep read of pp.1-3 (introduction and literature review): the paper states that Trichel pulses 'only occur in electronegative gases' and that in electropositive gases such as nitrogen 'the corona current in the external circuit would be continuous without pulses' [read: p.1]; it says 'Loeb et al [6] confirmed that these regular pulses exist only in electronegative gases' (ref. 6 = Loeb, Kip and Hudson, Phys. Rev. 60, 714-722, 1941; the paper gives no gases, purities or data) [read: p.2]. Goldman and Goldman [1] classify negative corona into (1) auto-stabilization, (2) regular pulses, (3) continuous current; pulses are very regular in oxygen and air but more irregular in SF6 (Van Brunt and Leep [3]) [read: p.1]. Trichel's own account as the authors report it gives the positive-ion cloud the central role: a negative cloud forms far from the point and a positive cloud near it; the positive ions raise the field near the point and lower it between the two clouds, so no ionization can occur beyond the positive cloud; the positive ions then move into the point, 'narrow the region of enhanced electric field and finally the ionization process stops'; the paper also credits Trichel with explaining 'the shielding effect produced by the positive ion cloud near the cathode' [read: p.1-2]. The negative-ion-clearing emphasis of the authors' mechanism is not attributed to Trichel in these passages (the paper's text does not say who first stressed it). Earlier models as the authors describe them [read: p.2]: Loeb et al. [6], successive avalanches (each giving rise to three successors), could not explain the 1.5 ns rise in air; Aleksandrov [8] (1963), parallel avalanches, can predict the faster rise and the slowly rising pre-pulse current; Kekez, Savic and Lougheed [9] (1982), equivalent circuit, describes the pulse succession but not the detailed mechanism; Ogasawara [10] (1966), analytic early-stage solution, gives only exponential growth, no Trichel pulses and no role for negative ions; Morrow [11,12] (1985), 1D finite-difference plus FCT, best-known, agreed with current-voltage and pulse-initiation data, modelled only the first pulse (current then decays), ignored ion secondary emission; Napartovich et al. [13] (1997), 1.5D, sequence of pulses in dry air for gaps under 1 cm; other simplified chain-of-pulses models [9],[14],[15] (ref. 14 = MacAlpine and Yim, 'Computer modeling of Trichel pulses in air'); Soria, Pontiga and Castellanos [5] (2007), particle-in-cell in pure oxygen, is cited only for pulse duration and period. Lama and Gallo [7] are described as 'a series of carefully designed experiments' on frequency, charge per pulse and mean current against voltage, tip radius and spacing, who also tried to explain the mechanism [read: p.2]. Small internal oddities: the abstract says that with increased voltage 'the pulse frequency decreases' although the results show the period falling (frequency rising); rise time is quoted as 1.3 ns [4] on p.1 and 1.5 ns on p.2; ref. 2 is printed with the title 'The mechanism of the positive point-to-plane corona in air' although the subject is negative corona. None of the literature claims in the introduction is accompanied by data in the paper.

Key claims

• Trichel pulses occur only in electronegative gases and only in a limited voltage range, with rise times as short as 1.3 ns, durations of tens of ns and inter-pulse periods of tens of us (literature values quoted in the introduction, not simulated results).
• Trichel pulse frequency varies linearly with corona current and inversely with needle-tip radius, independent of gap length (quoted from earlier literature).
• A 2D FEM-FCT model with electrons, positive ions and negative ions in a point-plane geometry (needle tip radius 100 um, gap 1 cm) reproduces a regular sequence of Trichel pulses.
• Pulse repetition frequency rises with applied negative voltage: simulated average period falls from 11.33 us at -7 kV to 2.46 us at -11 kV.
• The inter-pulse period is set by the time slowly drifting negative ions (mobility 2e-4 m2/Vs) take to clear the region near the needle so the tip field can recover to the ionization threshold.
• Each main pulse ends because positive ions are absorbed by the needle within nanoseconds, dropping the field in the ionization layer below threshold.
• A small post-pulse (secondary avalanche from positive ions returning to the needle) follows every main pulse in the simulation.
• The first pulse is 12-20 times larger (25 mA at -7 kV, 37 mA at -9 kV, versus about 2 mA steady) and is followed by a longer interval because it forms in charge-free space.
• The simulated periods underestimate Lama and Gallo's experimental periods by about 50%, with the same trend; the authors attribute this to unspecified numerical approximations.
• The authors claim this is the first successful application of FEM-FCT to 2D simulation of Trichel pulses at several voltages.
• Loeb, Kip and Hudson (1941) confirmed that regular Trichel pulses exist only in electronegative gases, and in electropositive gases such as nitrogen the external-circuit corona current is continuous without pulses (stated by the authors without data or gas details) [read: p.1-2].
• In Trichel's own account (as the authors report it), the positive-ion cloud near the point shields the field and, by drifting into the point, narrows the enhanced-field region until ionization stops; a negative cloud far from the point is the other half of the space-charge cycle [read: p.1-2].
• Earlier models failed in specific ways: Loeb's successive avalanches could not explain the 1.5 ns rise; Ogasawara gave no pulses; Morrow (1D, 1985) gave only the first pulse and ignored ion secondary emission; Napartovich (1.5D, 1997) already modelled pulse trains in dry air for gaps under 1 cm [read: p.2]. This is the context for the authors' priority claim.

Relevance: It is the only document in the Trichel folder and so is the whole evidential base for the project. It is a modern mainstream computational model, not an alternative-science work: it explains the self-pulsing of a negative corona through space-charge dynamics (Trichel's own 1938 idea). Its usefulness for the reader is as a worked physical mechanism (avalanche, attachment, ion drift, space-charge field quenching) and as a case study in model validation (a factor-of-two period discrepancy, incomplete parameter disclosure, printed-equation inconsistencies). Nothing by G. W. Trichel himself is in the folder.

Mainstream conflict: none (it is mainstream corona physics). Internal weaknesses noted in the deep read: printed equations/units inconsistencies, absent rise-time/width comparison, 50% discrepancy with experiment, undisclosed Rext and experimental geometry. Several constants rest on an unpublished contractor report (ref. 16, for ABB, an interested commercial party).