2026-10-04 Western University Open Repository (uwo.scholaris.ca / Scholarship@Western) - Sattari 2011 PhD thesis

URL [web]: https://uwo.scholaris.ca/server/api/core/bitstreams/0c5eda5f-5b90-43e3-8010-ff0188ca2b1a/content (open access; citation https://ir.lib.uwo.ca/etd/215) Paria Sattari, "FEM-FCT Based Dynamic Simulation of Trichel Pulse Corona Discharge in Point-Plane Configuration", PhD, Electrical and Computer Engineering, The University of Western Ontario, 28 July 2011. Supervisor K. Adamiak, joint supervisor G.S.P. Castle. How obtained: Chrome's PDF viewer gives no text, so the fallback WebFetch was used. The PDF was cached by the tool and copied to research/files/Sattari-2011-PhD-thesis-UWO.pdf, then read with pdftotext. Page numbers below are the thesis's printed page numbers.

Findings (all [web] from this thesis)

1. Ionization law, misprint resolved: Ch. 6 (p. 91) prints ki = 10^(11.93 - 449.3/EN) (1/s), ka = 6e6 1/s, citing [99] (Akishev, Grushin, Deryugin et al., TRINITI report for ABB, 1997). The ESA 2010 paper prints 4.493. The deep read's suspected misprint (449.3) is confirmed by the authors' own later text.
2. Mobilities (p. 90-91): mu_p = 2.24e-4 and mu_n = 2.16e-4 m^2/Vs. The ESA paper used 2e-4 for both.
3. Gas density: N_O2 = 2.46e25 1/m^3 is again called "oxygen molecules" (same issue as in the ESA paper).
4. External resistance and circuit: Ch. 6 assumes R_ext = 100 kOhm and gap capacitance Cd = 0.055 pF, with a circuit ODE dVc/dt = (Vapp - Rext ic - Vc)/(Rext Cd) (eqs 6.12-6.13). The ESA paper gives no R_ext.
5. Onset: -4.7 kV is "evaluated from experimental Peek's formula" (p. 96).
6. Parametric study, Ch. 7 conclusions (p. 124): - lower R_ext gives a longer period; - higher ion mobility gives a shorter period; - the secondary emission coefficient has no significant effect. Contrast: Trichel 1938 reported the pulses as unaffected by the external circuit constants (APS abstract).
7. Experimental reference values differ between the two documents by the same authors. Thesis Table 7.1 (R_ext = 100 kOhm, r = 100 um, d = 1 cm):

Average currents: simulated 13.6-33.3 uA against Gallo's 17.1-40.0 uA. The ESA 2010 Table 1 gives experimental periods of 21.85 / 13.32 / 9.09 / 5.07 us at -7 / -8 / -9 / -11 kV. So the "experimental" column at -7 kV is 21.85 us in one document and 11.3 us in the other, and the direction of the model error flips: the ESA paper underestimates by about 50%, the thesis overestimates. 8. Ch. 8.2 contradicts itself: it says the numerical results "considerably overestimate the pulse period", then that they "underestimate the pulse periods, especially at low voltages". Ch. 8.7 says "overestimates the pulse periods at low voltages". 9. Lama and Gallo's empirical laws, quoted in Ch. 8.2: - f = 1/T ~ K1 V (V - V0)/(r d^2), with K1 ~ 27 kHz mm^3 kV^-2; - I ~ K2 V (V - V0)/d^2, with K2 ~ 52 uA mm^2 kV^-2; - V0 = -2.3 kV, "comparatively independent of tip radius which is unexpected and unexplained". 10. Previously unpublished observations by C. F. Gallo (Ch. 8.3, Fig. 8.3, Table 8.1): - Under a microscope the negative needle tip shows "an extremely narrow bright ionization channel" that abruptly expands into the usual corona cone. It is absent in positive corona. - Very sharp tips make the channel "wobble", giving irregular pulses. This is the source's explanation of the "mode shifts" Loeb and Trichel described. - Slightly blunted tips give stable pulses. - Estimated sizes: tip radii 3-50 um, gaps 3-16 mm, channel diameter 0.12-2 um, channel length 1.2-20 um. - The authors write that the channel "is not described by present theory or simulations and may possibly be confined magnetically by the high current density". They add that localized electron field emission may explain the low -2.3 kV threshold, against a calculated -4.7 kV. Source and stake: these are C. F. Gallo's own observations, reported second-hand by the thesis authors, with whom he collaborated. No independent publication was found in this run. 11. Revised comparison (Ch. 8.4): the model was re-run with a fitted "effective tip radius" of 10 um (the channel), d = 6 mm, -4 kV, V0 assumed -2.3 kV. Results: T = 2.3 us against Gallo's 2.0 us, and I = 12.1 uA against 9.8 uA (15% and 23% high). Note: the agreement depends on a chosen effective radius. 12. Pressure (Table 8.3, -9 kV, R_ext = 100 kOhm):

The simulated pulse duration is about 0.2-0.3 us, which bears on the "tens of ns" literature value. 13. Charge per pulse (Table 8.4): about 115 pC at r = 100 um and about 22 pC at r = 10 um. The ratio Qi/Q > 0.5 disagrees with Haidara, Denat and Atten's data (< 0.5, high-pressure air, no R_ext); the authors attribute this to the surge resistor. 14. Trichel citation: thesis ref. [16] is again the positive-corona paper (Phys. Rev. 55, 382).

Bearing: claims 3-5 directly. This is the inventor-side priority-1 source (the authors' own fuller work). It fixes the printed ki and supplies R_ext. It also changes the experimental comparison set, and the reviewers need to reconcile the two sets. Recheck: yes (the IEEE TIA 2011 journal version of the ESA paper, 47(4), was not read).