Negative corona discharges : a fundamental study
Citation for published version (APA):Gravendeel, B. (1987). Negative corona discharges : a fundamental study. Technische Universiteit Eindhoven. https://doi.org/10.6100/IR256688
DOI:
10.6100/IR256688
Document status and date: Published: 01/01/1987
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne
Take down policy
If you believe that this document breaches copyright please contact us at: openaccess@tue.nl
providing details and we will investigate your claim.
NEGAT+VE
CORONA DISCHARGES
NEGATIVE CORONA DISCHARGES
A FUNDAMENTAL STUDY
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op gezag van de rector magnificus, prof. dr. F.N. Hooge, voor een commissie aangewezen door het college van dekanen in het openbaar te verdedigen op
dinsdag 10 februari 1987 te 16.00 uur door
BASTlAAN GRAVENDEEL
geboren te Puttershoek
Dit proefschrift is goedgekeurd door de promotoren: Prof. dr. F.J. de Hoog
en
Prof. dr. ir. P.C.T. van der Laan.
These investigations in the program of the Foundation for Fundamental Research on matter (FOM) have been supported (in part) by the Nether-lands Technology Foundation (STW).
CONTEHTS SUMMARY SAMENVATriNG INTRODUeTION 1.1. 1.2.1. 1.2.2. General
The negative corona discharge The Trichel mechanism
1.3. Macroscopie aspects of negative corona
discharges in air
1. 4. Aim of the present work
References FAST CURRENT MEASUREMENTS 2.1. 2.2. 2. 3. 2.3.1. 2.3.2. 2.3.3. 2.3.4. 2.3.5. 2.4. Introduetion
Fundamental aspects of current measurements Measurements
The measuring setup
Electric field in a corona gap Laser triggered experiments Trichel pulses
Secondary mechanisms Conclusions
References
LIGHT MEASUREMENTS IN NEGATIVE CORONAS IN AIR 3 .1.
3.1.1. 3.1.2.
Introduetion General
Light intensity measurements in negative corona discharges
3.1.2.1. General
3.1.2.2. Determination of the electron density from
continuum measurements
3.1.2.3. A model descrihing the light pulse from a
corona discharge in oxygen in the Trichel regime iii 1 5 9 11 12 14 16 17 21 24 34 34 40 41 46 50 52 53 55 55 56 56 57 59
iv negattve corona discharges
3.1.2.4. Sputtering of needle material 3.2. Experimental setup and metbod 3.2.1. The discharge chamber
3.2.2. Optical system 3.2.3. 3. 3. 3.3.1. 3.3.2. 3.3.3. 3.3.4. 3.3.5. 3.3.6. 3.4. Measuring metbod
Measuring results in air
Identification of molecular and ionic spectra Development in time and position of the
337 nm emission for various pressures Continuurn measurements
Intensity decay of the 337 nm line for various pressures
Sputtering in a negative corona discharge in air
Verification of a negative corona discharge model Conclusions References 60 60 60 61 62 63 63 63 66 70 76 79 82 84
NEGATIVE IONS IN NEGATIVE CORONA DISCHARGES IN THE TRICHEL REGIME 4 .1. 4.1.1. 4.1.2. 4 • .2. 4.2.1. 4.2.2. 4.2.3. 4.2.4. 4.2.5. 4.2.6. 4.2.7. 4.2.8. 4.2.9. 4.2.10. . 4. 3. Abstract Introduetion Mass speetrometry
Negative ion mass identification in negative corona discharges in air
Experimental setup General
The discharge chamber The extraction hole The skiromer
The quadrupole mass filter The pboton trap
The channeltron Data handling
Gas handling system
Alignment of the ion trajectory Surface analysis of extraction foils
87 87 87 88 90 90 92 93 94 95 97 101 103 104 105 106
4.4. 4.4.1. 4.4.2. 4.4.3. 4.4.4. 4.4.5. 4.5. 4.5.1. 4.5.2. 4.5.3. 4.6. contents
Extracting near thermal negative ions from atmospheric discharges in supersonic flows General
The campargue expansion Intensity measurements
Clustering of negative ions with water molecules
The overall efficiency Results
General
Identification of the mass peaks
Reaction mechanisms in a negative corona discharge in air Conclusions Tables References CONCLUSIONS DANKWOORD CURRICULUM VITAE V 109 109 110 112 114 117 120 120 121 124 134 135 140 143 147 149
1
SUMMARY
In this thesis fundamental aspects of the negative corona discharge in air in the Trichel regime are studied using three different observation techniques: fast current measurements (resolution 0. 7 ns), mass speetrometry and fast light intensity measurements resolved in time (resolution 0.7 ns), inspace (resolution 0.05 mm) and in wavelength (resolution 0.16 nm).
An introduetion to the negative corona discharge is given in chapter I. In chapters II, III and IV the three different measuring techniques are described in detail. At the end of each chapter conclusions have been given. In a final chapter, chapter V, the conclusions are summarized.
The main problem to be solved in current measuring techniques used in inhamogeneaus electric field configurations is the interpretation of the externally measured current pulse. In chapter II a theoretica! expression is derived which relates the externally measured current with the movement of charges inside an inhamogeneaus field gap in the vicinity of space charge. The validity of the derived expression has been checked experimentally using laser triggered experiments in inhamogeneaus field gaps and found to be correct.
The rise time of Trichel pulses in atmospheric air is measured to be 1. 15 ns. Th is rise time is determined mainly by gas discharge processes. The current pulse measured is induced mainly by electrans rnaving in the high electric field gradient in front of the needle. Positive ions contribute only to the tail of the measured current pulse. In the setup used, negative ions do not measurably contribute to the externally measured current when the corona is just above onset potential.
The fast spectroscopie measurements are carried out using relatively simple tools: a time to pulse height converter, a spectrometer and a photomul tiplier. In a negative corona discharge in atmospheric air the positive ions N~ and 0~ were detected. Furthermore, the N2, 02,
2 negattve corona dtscharges
COz, HzO, Oa, and CO molecules were found. In an
atmospheric discharge, the active region in front of a needle with a radius of curvature of 0.03 mm along the axis of symmetry in the direction of the plane extends from 0 mm (needle tip) to 0.25 mm. The dimensions of the active region decrease with pressure.
By measuring photons in a very small time window, the dead count rate of the photomultiplier was suppressed sufficiently to cbserve phenomena simultaneously resolved in time, wavelength and position. The high time resolution of the system tagether with the low dead count rate made i t possible to measure continuum radiation from an atmospheric discharge. By calibrating the setup for absolute radiation intensities an estimate could be made of the electron density of the plasma formed. The electron density was found to be 1018
m-3
at maximum. The electron temperature in the plasma at the peak of the Trichel pulse was estimated to be 1 kK. The continuum radiation measurements are believed to be the first experimental proef of the streng attachment occurring in the peak of the Trichel pulse.
In a negative corona discharge in atmospheric air positive ions impinging on the needle surface cause sputtering of needle material. With the system used i t is possible to determine the density of sputtered needle atoms. Experimental values of these densities can only be explained when these sputtered atoms are assumed to be transported to the plane by the electric wind. Measuring the time averaged needle current and the discharge time an estimate of the sputtering coefficient can be made. For a capper needle the sputtering rate is estimated to be 10-4 capper atoms per impinging positive ion. This value is in agreement with a sputtering rate given in literature in a corresponding setup and using corresponding discharge conditions.
From the light intensity measurements at low pressure i t is found that the lumineus phenomena in front of a blunt needle correspond with the lumineus phenomema in a
summa.ry 3
glow discharge. The neg at i ve glow, Faraday' s dark space and the positive column can also be observed. In a corona discharge in air, however, these phenomena are found to be intermittent.
The model of a Trichel pulse in a low pressure corona discharge recently put forward by Morrow, descrihing the integral light intensity resolved in time and space, has been checked. The predicted · contraction of a lumineus volume in front of the needle has been confirmed experimentally. A dark space and a second contracting lumineus volume, the positive column were also observed, but are nat described by the model.
Finally, we want to emphasize that the setup used for the spectroscopie experiments is a strong and relatively simple tool to study gas discharge phenomena e.g., breakdown, at nanosecond time se ale. However, to avoid long measuring times, the event under study has to have a relatively high repetition rate.
To study the corrosive properties of negative coronas the negative ions originating from a negative corona discharge are determined. In doing sa oxidation and charging of the surface around the hole through which the ions are extracted must be avoided to minimize sampling errors. In literature this problem is seldom mentioned. In this thesis this problem is dealt with adequately. Furthermore, i t is necessary to properly extract the negative ions from the shock wave that almast unavoidably occurs in a vacuum vessel downstream from an extraction hole of an atmosphetric discharge. To fulfil this requirement a special shaped skimmer has been used. The basic ions in a negative corona discharge in clean and dry N2/02 mixtures (less than 5 vpm H20) in a ratio of 5 to 1 at atmospheric pressure are 03, OH-, N03 and C03. These ions are found to be clustered with three water molecules at maximum.
5
SAMENVAT'I'ING
In dit proefschrift worden fundamentele aspekten be-studeerd van de negatieve corona-ontlading in lucht in het Triohel regime. Bij de bestudering worden drie
verschil-lende observatietechnieken gebruikt: snelle stroommetingen (resolutie 0,7 ns), snelle lichtintensiteitsmetingen op-gelost in tijd (0,7 ns), plaats (0,05 mm) en golflengte
(0,16 nm) en massaspectrometrie.
In hoofdstuk I wordt een introductie tot de negatieve corona ontlading gegeven. In de hoofdstukken II, III en IV worden de drie bovengenoemde meettechnieken gedetailleerd beschreven. Ieder hoofdstuk wordt afgesloten met conclu-sies. In hoofdstuk V, een afsluitend hoofdstuk, wordt een samenvatting gegeven van de conclusies.
Het hoofdprobleem dat opgelost moet worden bij stroom-meettechnieken, als deze gebruikt worden in inhomogene electrische veldconfiguraties, is de interpretatie van de extern gemeten stroom. In hoofdstuk II wordt een theore-tische uitdrukking afgeleid, die een relatie geeft tussen de extern gemeten stroom en de beweging van ladingsdragers binnenin een volume waar een inhomogeen veld heerst en waar ruimtelading aanwezig is. De geldigheid van de afge-leide uitdrukking is gecontroleerd met behulp van experi-menten, uitgevoerd in inhomogene veldconfiguraties waarbij de bewegende ladingsdragers uit het elektrodeoppervlak vrijgemaakt worden met behulp van een korte-puls-laser. Er is een duidelijke overeenstemming gevonden tussen de theorie en het experiment.
Uit metingen is afgeleid dat de stijgtijd van Triohel pulsen in atmosferische lucht 1,15 ns bedraagt. Deze stijgtijd wordt in hoofdzaak bepaald door processen in het gas. De gemeten stroompuls wordt voornamelijk door elek-tronen geïnduceerd, die zich bewegen in een elektrisch veld met een sterke gradiënt. Positieve ionen geven slechts in de achterflank van de puls een bijdrage tot de gemeten stroompuls. Voor spitsspanningen juist boven de startspanning geven de negatieve ionen in de door ons
6 negattve corona discharges
gebruikte opstelling geen bijdrage tot de externe stroom. De snelle spectroscopische metingen
met relatief eenvoudige middelen: een
zijn uitgevoerd tijd-pulshoogte omvormer, een monochromator en een photomultiplier. In een negatieve corona-ontlading in atmosferische lucht zijn de positieve ionen N~ en o~ waargenomen. Ook zijn de molecu-len N2, 02, C02, H20, 0:~ en CO geïdentificeerd. In een atmosferische ontlading rondom een spits met een kromte-straal van 0,03 mm strekt het aktieve gebied van de ontla-ding zich uit vanaf 0 mm (de punt van de naald) tot o, 25 mm in de richting van de plaat en gemeten langs de symmetrie-as van de opstelling. De afmetingen van het aktieve gebied nemen af met de druk.
Door de fotonen in een smal tijdvenster te tellen, wordt de donkerstroomtelling zover onderdrukt, dat er ver-schijnselen waargenomen kunnen worden, die gelijktijdig in tijd, plaats en golflengte opgelost zijn. De hoge tijdre-solutie in combinatie met de lage donkerstroomtelling maakt het mogelijk om continuümstraling afkomstig van een atmosferische corona-ontlading te meten. Wanneer de op-stelling absoluut geijkt wordt, kan een schatting gemaakt worden van de elektronendichtheid in het gevormde plasma. De gevonden dichtheid blijkt maximaal 101 8
m-3
te bedra-gen. De temperatuur van de elektronen in het plasma tij-dens het maximum van de Triohel puls kan geschat worden op
1 kK. De continuüm metingen vormen waarschijnlijk het eerste experimentele bewijs voor de sterke attachment, die bij de piek van de Triohel puls optreedt.
In een atmosferische negatieve corona-ontlading ver-oorzaken de op de spits vallende positieve ionen sputteren van spits materiaal. Met de gebruikte opstelling is het mogelijk om de dichtheid van gesputtarde atomen te bepa-len. De waarden voor de gemeten dichtheid kunnen alleen verklaard worden, als aangenomen wordt, dat de elektrische wind voor het transport van deze atomen naar de plaat zorg draagt. Door de gemiddelde ontladingsstroom en de ontla-dingsduur te meten kan dan een schatting gemaakt worden ·van de sputtercoëfficiënt. Voor een koperen spits is de
samenvatting 7
sputtercoëfficiënt geschat op 10-" atomen per üpvallend positief ion. Deze waarde is in overeenstemming met lite-ratuurwaarden voor een overeenkomstige opstelling en onder overeenkomstige ontladingscondities.
Uit lichtintensiteitsmetingen bij lage druk blijkt, dat de lichtverschijnselen, die rondom een stompe punt optreden, corresponderen met de lichtverschijnselen in een glimontlading. In een corona-ontlading blijken deze ver-schijnselen intermitterend te zijn.
Het recentelijk in de literatuur verschenen model van de Trichel puls in een lagedruk ontlading is op zijn merites getoetst. Het model beschrijft de integrale licht-opbrengst als funktie van plaats en tijd. Het in het model voorspelde, zich samentrekkende, optisch aktieve gebied juist voor de spits treedt in werkelijkheid ook op. Uit het experiment blijkt, dat er een tweede, zich samentrek-kend, optisch aktief gebied optreedt, gescheiden van het eerste door een donkere ruimte. Dit tweede optisch aktieve gebied, de positieve zuil, wordt niet beschreven door het model.
Tenslotte willen we benadrukken dat met de opstelling, die gebruikt is bij de snelle spectroscopische experimen-ten, een krachtige en relatief eenvoudige diagnostiek is verkregen om op nanoseconde tijdschaal gasontladingen, zoals doorslag, te bestuderen.
Om de corrosieve eigenschappen van negatieve corona-ontladingen te bestuderen, worden de negatieve ionen uit een atmosferische corona gemeten. Hierbij moet oxydatie en oplading van het oppervlak rondom het gat, waardoor de ionen geëxtraheerd worden, voorkomen worden, teneinde een juiste bemonstering te realiseren. In de literatuur wordt van deze problematiek nauwelijks gewag gemaakt. In dit proefschrift wordt dit probleem op een adequate wijze aangepakt en opgelost. Tijdens. het bemonsteren is het bovendien nodig om de negatieve ionen ongeschonden door de schokgolf heen te loodsen. Het optreden van een dergelijke schokgolf in een vacuUmkamer achter een bemonsteringsgat van een hogedruk ontlading is vrijwel onvermijdelijk. Om
8 negative corona discharges
aan deze eis tegemoet te komen wordt een speciaal gevormd tweede bemonsteringsgat gebruikt. In een negatieve corona-ontlading in een droog N2/02 mengsel (minder dan 5 vpm H20) in een verhouding van 5 op 1 zijn bij atmosferische druk de velgende negatieve ionen gevonden:
o;,
OH-, N03 enco;.
Deze ionen blijken omringd te zijn door maximaal drie watermoleculen.9
CHAPTER I
INTRODUeTION
1.1. General
A corona discharge is a self-sustained electric gas discharge where the electric field confines the primary ionization processes to regions close to a high-field electrode [SIG83]. A widely studied form of corona discharge has a needle-to-plane configuration. There exist pure a.c. coronas (SAT84a and b] and a.c. coronas with superimposed voltage pulses [SAT83, HAY83]. A d.c. corona discharge is called positive or negative depending on the potential of the high field electrode with respect to the ether electrode. Positive or negative d.c. coronas can also be driven with superimposed voltage pulses (KORSO, GOL78]. The negative corona only occurs in electronegative gases [SIG78]. Corona discharges are driven at pressures from the J.tbar-range up to several bar. Electrode configurations used in corona discharges are e.g. wire-to-plane and point-to-plane configurations. We will limit ourselves to negative d.c. coronas in a point-to-plane contiguration. Under some condi t i ons d. c. corona discharges occur as pulsating discharges generating time dependent currents in the electrode leads. In case of a negative corona these current pulses are called Trichel pulses named after one of the pioneers in the field of corona discharges [TRI38
J.
We will come back to these Trichel pulses in sectien 1.2.2.Because i t is possible to run negative coronas in air, these discharges are widely applied. Applications are e.g. in radiation meters for weak ~-rays (AOY82], in hygrometers [BERSO] and in ozonizers [KOG83]. Furthermore, corona discharges are used in precipitators [MCDSO]. They are used to increase the heat transfer to certain objects from an ambient gas [KAD87]. The charging process in some dry photocopier systems also is based on the use of a negative corona discharge [LIC81]. The corona discharge is
10 negative corona discharges
used in plasma chemistry (surface treatment) [SIG83]. Corona discharges a lso
transmission lines causing disturbances [MAR82].
occur around high voltage transport losses and radio
Another application field is that of sanitation for men and animals (SUL74, TOM81, VIS86). Investigation pointed out that absence of negative ions in inhaled air may lead to depression and increases in suicide and crime [SUL74]. It is however unknown how the negative ions involved affect the human body and mood. one thing appears clearly: negative ions have a positive influence on the state of health of living beings.
Much work has already been done on the negative corona discharge in the Trichel regime. Up t i l l now various extensive reviews have been published. The first was written by Loeb in 1965 [LOE65]. In this contribution the state of the art in corona research up till 1965 is given. When reading this work one realizes that experimental techniques have been vastly improved ever since. The bandwidth of electrooie equipment, e.g. oscilloscopes, has increased with two orders of magnitude. The availability of fast data handling systems nowadays allows the recording of fast real time events. These improved techniques led to an extension of the knowledge and rnadelling of co~ona discharges. In 1978 M. and A. Goldman published a detailed study on coronas [GOL78]. A few years later R.S. Sigmond and M. Goldman gave a survey of corona work up to 1983 (SIG83].
A general survey of phenomena observed in neg a t i ve corona discharges is given in section 1.2.1. The physical aspects of the specific Trichel mode will be introduced in section 1. 2. 2. Macroscopie effects, e.g. electrode corrosion, are treated in section 1.3, whereas in 1.4 the scope of the experimental work reported on in this thesis will be outlined.
introduetion 11
1.2.1. The negative corona discharge
In a needle-to-plane system with the needle at a negative potential relative to the plane a choice of phenomena occur when, following the I-V characteristic, the voltage between the needle and the plane is increased (Fig. 1.2.1) [GAL77]. We will discuss the phenomena fora tip radius of 0.05 w~ and a gap width of 6 mm in atmospheric air.
At relatively low voltages spontaneous externally measurable current puls es occur due to e.g. cosmie rays creating electron-ion pairs. The time-averaged current is of the order of 50 fA (region A in Fig. 1.2.1). By increasing the voltage the discharge sets into the Geiger regime. At this voltage all charge carriers formed will be removed from the gap immediately. Here the externally measured current is proportional to the number of charges generated by external causes (region B in Fig. 1.2.1). Increasing the voltage between needle and plane still further (up to 2 kV) will lead to an electric field high enough for electrons to cause avalanches. In this region of the time-averaged current versus voltage plot irregular externally sustained current pulses, Trichel pulses occur
\
\
,_
- --' , G. unstable transition region ... I I I
---
---
____
-
... :;.____
f. breakdown feathers ; ; ; . . r -~le_?? __g_lo~ _ g u lar T richel pul ses !lf- sust ained J
-C. irregular trichel pulses (externally sustainedl
10 12
Fig. 1.2.1. Schematic
current-voitage reiatton
for air around
atmos-pheric pressure in a
needie·to-piane geometry
wtth negative de voLtage
12 negattve corona discharges
(section
c
in Fig. 1. 2 .1). The current pulses are again believed to be started by external sources. Increasing the voltage to several kV will lead to periodic pulses. External sourees are no longer necessary to create new Trichel pulses ( region D in Fig. 1. 2 .1) • Due to strong local electric field gradients and fast successive processes the description of the discharge becomes difficult. Provided the gap is sufficiently long the discharge comes into the pulseless glow mode if the voltage between needle and plane is increased up to 10 kV (region E in Fig. 1.2.1). Now the measured discharge current is not pulsed any longer. If the discharge gap is wide enough a further increase of the gap voltage may cause so-called 'negative feathers', a lumineus phenomenon superimposed on the glow present around the needle. These 'feathers 1 may become so long that they reach the anodeand cause breakdown or arcing in the gap dependent on the external circuit (regions G and J i n Fig. 1.2.1). In this thesis we have concentrated on studies of certain aspects of the self-sustained negative current pulses, the Trichel pulses.
1.2.2. The Trichel mechanism
Here we will give a qualitative description of the discharge mechanism leading to a Trichel pulse. Trichel pulses are characterized by fast rise times, believed to be smaller than a few nanoseconds at atmospheric pressures usually followed by two exponential falls with different time constants. Generally spoken a Trichel pulse starts with one or more electrens generated in the high field region. These electrens move rapidly away from the needle through this region and ionize the gas. The positive ions formed stay behind in the wake of the avalanche as in the start of an ordinary Townsend discharge. The number of positive ions increases rapidly with time and distance. In the case of atmospheric air and using a needle with a radius of curvature of 0 •. 05 mm at a gap width of 5 mm and a potential difference of 5 kV, the reduced electric field
introduetion 13
E/n, with n the gas density, can be as high as 1200 Td resulting in a primary ionization coefficient of 0.5 J..tm-1
[BAD72]. The electric field created between the electron and ion swarm is opposite to the original field in the gap, so after a number of ionizations this electric field reaches a value for which the sum of this field and the Laplacian needle-to-plane field is too low to generate ionization in the gap. A few nanoseconds after the start of the avalanche and at a position in the order of 30 Jlnl
from the start pos i ti on the avalanche is cut off. The electrens now have too little energy to create new ion-electron pairs. In the wake of this avalanche other electrens following close to the first generation of electrens will also contribute to the measured current. Due to the formed ion cloud the cut-off of the following avalanches will be subsequently closer to the needle leading the discharge to develop in the direction of the needle ( see also section 3. 3. 6) • At this point of the development of the Trichel pulse electrens will mainly attach to gas molecules and atoms and form a negative ion cloud. This ion cloud will move only slowly towards the plane. In the mean time the positive ion cloud moves towards the needle in an electric field that increases when the cloud approaches the needle. These ions impinge on the cathode surface and will release electrens from the needle. However the multiplication is now too low for these electrens to contribute to a new avalanche. As soon as the negative space charge cloud has drifted away far enough to restore the initial field conditions a new avalanche will be started. The initiating electrens may originate from secondary emission caused by positive ions or photons impinging on the needle tip, and are continuously being formed under the non-avalanche conditions between two Trichel pulses. This is illustrated by light measurements (see chapter 3) from which i t appears that the discharge is still active between two Trichel pulses. It is not necessary for the negative ion cl oud to re ach the anode before a new avalanche starts.
14 negative corona dtscharges
Taking into account the mobility of the negative ions it is very likely for the gap to be filled with a number of negative ion clouds as described by Gallo [GAL77].
Gallo assumes that the electrens starting the new avalanche are generated by field emission. This is unlikely because Trichel pulses also occur at needle tip field strengths lower than necessary for field emission
[CR086].
As stated above we assume that the cut-off of the Trichel pulse is caused by the sharp decrease of the resulting electric field in which the electrens move. The cut-off of the discharge is caused by a negative space charge cloud of electrens because negative ions have not yet been formed. This is contrary to Gallo [GAL77] and Goldman [GOL78] in whose opinions the cut-off is caused by the negative ion space charge cloud. This can be illustrated by the fact that the attachment time for electrens to oxygen molecules in the case of atmospheric air takes a time in the order of 10 ns when the electron energy is between 1 ev and 3 ev [BAD72].
1.3. Macroscopie aspects of negative corona discharges in air
Apart from the specific behaviour of the negative corona in the Trichel regime, coronas are known to be responsible for quite a number of other phenomena.
In 1962 Buchet et al. reported on the transport of needle material from gold and platinum needles to the plane during . a corona discharge in air at atmospheric pressure [BUC62]. Three years later A. Goldman published an artiele on anode effects due to a negative corona discharge in air at 346 mbar. This paper described the formation of crystals on a metal plane anode at corona currents of 250 ~A [GOL65].
In 1970 the sputtering rate of cathode material was related to the various stages of corona discharges at atmospheric pressures in air and mixtures of oxygen, nitrogen and carbon dioxide [BUC70]. It was demonstrated
i.ntroduction 15
that the crystals growing on the anode are built up from the anode material itself.
In 1979 A. Goldman and R.S. Sigmond came up with a paper dealing with the influence of negative corona discharges in humid air on thin anode foils [GOL79]. A thin aluminum foil of 0.016 mm thickness was found to be piereed under a corona discharge with cylindrical holes of approximately 0. 01 mm diameter. This hole formation appeared to be related with the relative humidity of the gas and increased with humidity. It was also found that, behind the foil, light from the second positive system of molecular nitrogen could be detected, probably caused by breakdown of the insulating and charged oxide layer. The hole formation disappeared when a grid was placed in front of the anode foil removing the negative ions from the electric wind. Later [SIG80] they concluded that the corrosion of the thin aluminum anode foil had something to do with the structure of the aluminum tri-oxyde (Al203 ) .
In 1982 A. Goldman and Krolikowski found that in corona discharges in air at atmospheric pressures not only cathode material precipitated on the plane anode but also anode metal itself, the latter being transported from one part of the anode to the other [KR082].
Le Ny reported in 1983 that pit corrosion of an aluminum anode, caused by
and atmospheric pressure Goldman and Sigmond, is
[LEN83].
a corona discharge in room air and earlier described by A. mainly caused by the N03 ion
In 1984 Le Ny, spectroscopy of
Fiaud and Nguyen reported on Raman corrosion products formed on the electrodes by negative coronas in atmospheric room air [LEN84]. Here hydroxinitrates appeared to be the dominant corrosive agents.
In a corona discharge in ambient air and at atmospheric pressure the temperature of the gas increases approximately 6 K for a discharge current of 150 ~A
[DOU82]. Fora current of 1 mA a temperature increase of 100 Kis found [DUB80].
16 negntlve corona dlscharges
The electric wind has been known for a long time but up till now it has notbeen really understood [VERSO]. One should realize that the electric wind dissipates some 50% of the corona input energy [SIGS3]. Typical values of the electric wind velocity for sharp needles are found to be 10 mjs [VERSO].
1.4. Ajm of the present work
From the above it is clear that much work has already been done on the subject of the negative corona discharge in the Trichel regime. However most research has been phenomenological and quite a number of problems are still unsolved. The aim of the present work is to contribute to the insight in the physical discharge processas of a negative corona discharge in the Trichel regime. We try to do so using three different experimental techniques.
1. Fast current measuring techniques. By measuring the fast Trichel pulses it is possible to investigate the fast build up of these pulses. We made use of a general expression that relates the externally measured current with the movements of charge carriers in an inhomogeneous electric field in the presence of space charge.
2. Fast spectroscopie techniques. We measured the light emission during a Trichel pulse resolved in time with a resolution of 0.7 ns, in space with a resolution of 0.05 mm and in wavelength with a resolution of 0.16 nm. Fr.om these measurements we drew conclusions on various unresolved aspects of the Trichel mechanism. It was also possible to deduce the electron density inside the discharge during a Trichel pulse as well as the geometrical dimensions of the active zone around the needle.
3. Mass spectrometry. Because of the interest in the corrosive properties of negative coronas we measured the negative ions originating from a negative corona discharge at atmospheric pressures. To do so we took care to extract the near thermal negative ions properly through the high pressure gradient.
introduetion 17
To generate a corona discharge one may use a wire-to-plane or a needle-to-plane geometry. We used a needle-to-plane geometry because this has several advantages. The discharge is sharply localized. On a wire several discharge points occur randomly distributed over the wire. Using a needie to plane configuration the current density at the anode is known from the Warburg Law (see sectien 4.4.5). The electric wind in a wire-to-plane configuration will be directed differently than in a needle-to-plane configuration. This will be of influence on the mixing of the formed ions and reactant products in the gas. So the results of mass spectrometric studies of a wire-to-plane set up are not directly comparable with the
results of a needle-to-plane set up.
In the following chapters we will subsequently discuss the three applied techniques and evaluate the resulting experimental data. In a final chapter we will summarize the conclusions.
References
AOY82 T. Ayoama and T. Watanabe, Nucl. Instrum. Methods, 197, 357 (1982).
BAD72
s.
Badaloni and I. Gallimberti, Basic Data of Air Discharges, Univ. di Padova, rep. nr. UPee-72/05(1972).
BER80 M.A. Berliner, in Feuchtemessung, hrsg. von G. Scholtz (VEB Verlag Technik, Berlin, 1980), p. 178. BUC62 G. Buchet, M. Goldman and mme. A. Fakinis-Zeitoun,
c.
R. Acad. Sci., 255, 79 (1962).BUC70 G. Buchet and A. Goldman, Proc. 1 st Int. Conf. Gas Disch., 459 (1970).
CR086 J.A. Cross, R. Morrow and G.N. Haddad, J. Phys. D, 19, 1007 (1986).
DOU82 N.G. Douglas, l.S. Faleener and J.J. Lowke, J. Phys. D, 15, 665 (1982).
18 negative corona discharges
GAL77 C.F. Gallo, IEEE Trans., IA-13(6), 550 (1977). GOL65 A. Goldman et al., J. Phys. (Paris), 26, 486
(1965).
GOL78 M. and A. Goldman, in Gaseous Electronics, ed. by M.N. Hirsch and H.J. Oskam (Academie Press), vol. 1, chap. 4.
GOL79 A. Goldman and R.S. Sigmond, J. Phys. (Paris), 40, C7-443 (1979).
HAY83
s.c.
Haydon and N. Sato, Proc. 16 th Int. Conf. Phen. Ion. Gas., 180 (1983).KAD87 H. Kadete, Ph. D. Thesis, Univ. Dar es Salaam, Enhancement of Heat Transfer by Corona Wind,
(1987).
KOG83
u.
Kogelschatz, Proc. 16 th Int. Conf. Phen. Ion. Gas., Inv. Lect., 1 (1983).KORSO Yu. D. Korolev, V.A. Kuz'min and G.A. Mesyats, Sov. Phys. Tech. Phys., 25(4), 418 (1980).
KR082
c.
Krolikowski and A. Goldman, Proc. 7 th Int. Conf. Gas Disch., 166 (1982).LENS 3 R. Le Ny, Proc. 6 th Conf. Electrost. Phen. , 17 3 (1983).
LEN84 R. Le Ny, C. Fiaud and A.T. Nguyen, J. Phys. (Paris), 45, C2-661 (1984).
LIC81 M. Lichtensteiger and
c.
Webb, Appl. Phys. Lett., 38(5), 323 (1981).LOE65 L.B. Loeb, Electrical Coronas (University of California Press, 1965).
MAR82 P.S. Maruvada, IEEE Trans. El. Insul., EI-17(2), 125 (1982).
MCDSO J.R. McDonald, M.H. Andersen and R.B. Mosley, J. Appl. Phys., 51(7), 3632 (1980).
SAT83 N. Sato and
s.c.
Haydon, Proc. 16 · th Int. Conf. Phen. Ion. Gas., 170 (1983).SAT84a N. Sato and
s.c.
Haydon, J. Phys. D, 17, 2009 (1984).SAT84b N. Sato and
s.c.
Haydon, J. Phys. D, 17, 2û23 (1984).introduetion 19
SIG78 R.S. Sigmond, in Elecrical Breakdown of Gases, eà. by J.M. Meek and J.D. Craggs (John Wiley and Sens, 1978) chap. 4.
SIGSO R.S. Sigmond, A. Goldman and D. Brenna, Proc. 6 th Int. Conf. Gas Disch., 82 (1980).
SIG83 R.S. Sigmond and M. Goldman, in Electrical Breakdown and Discharges in Gases, part B, ed. by E.E. Kunhardt and L.H. Luessen, NATO ASI Series, series B: Physics, vol. 89b (Plenum Press, 1983), chap. 1.
SUL74 F.G. Sulman et al., Int. J. Biometeor., 18(4), 313 (1974).
TOM81 G. Tom et al., Human Factors, 23(5), 633 (1981). TRI38
VERSO
G.W. Trichel, Phys. Rev., 54, 1078 (1938).
I.P. Vereshchagin and V.A. Zhukov, Power Eng., 18(2), 87 (1980).
VIS86 A.E.P. Visée, Internal Report, univ. of Utrecht, Faculty of Veterinary Science, 1986.
21
CHAPTER II
FAST CURRENT HEASUREMENTS
2.1. Introduetion
One of the methods to investigate an electrical discharge between electrades is described by Raether (RAE64], who calls i t the "electrical method". The basis of this methad is the observation of the time dependent current in the electrode leads caused by electron and ion movements between the electrodes. With this methad i t is possible to study the temporal development of e.g. a discharge inside the electrode volume. If this methad is used in a homogeneaus gap configuration (see Fig. 2.1.1) the interpretation is rather easy. Let us assume one electron crossing this gap with constant velocity. The externally measured current flows during the crossing of the gap by the electron and is constant in time. Fig. 2.1.2 gives this current as a function of time. At t t r the electron has crossed the entire gap and reaches the anode. At this time the current becomes zero. The area under the current pulse equals the total amount of charge that has
I
t
e
Gap
Cp
FIG. 2.1.1. SchematicaL drawing of an expertmentaL setup used with the eLectricaL method.
Rm Rd - damping resistor;
Rm - measuring resistor;
CP - parasttic capacitance of the gap and the cabLe.
FIG. 2.1.2. Current induced by a singLe
electron crossing a homogeneaus field ftr - t gap with constant veLocity.
22 negative corona discharges
crossed the gap. Since the electric field has a constant value space charge effects are neglected the time axis can be transformed to a position-in-the-gap axis: the width of the current versus time pulse then represents the electrode distance. So in a homogeneous gap configuration this method gives not only information of the time behaviour of a discharge, but also about the position at which phenomena e.g. ionization take place.
In an asymmetrical electrode configuration e.g. a needle to plane configuration, this transformation of the time axis to a position in the gap axis is more complicated and can be done analytically only in special cases.
An aspect not mentioned yet is the time constant of the measuring circuit. This must be small enough to follow the fast changes in the induced current. With the increased bandwidth of modern electronic measuring devices attention must be paid to the physical shape of the measuring electrode to keep parasitic capacitances as low as possible. In addition, the entire discharge circuit must be. constructed properly to ensure that the current can follow the physical processes without the delay caused by circuit response.
In the following, some hlstorical remarks on current measurements in a negative corona discharge will be made after a short characterization of the negative corona discharge itself. A negative corona discharge can have a pulsed character dependent on voltage and pressure wi th the current flowing in regularly spaeed 'Trichel' pulses [TRI38]. These pulses are characterized by a fast rise usually followed by two subsequent exponentlal decays with different time constants. In atmospheric air the rise time of such a Trichel pulse is approximately 1 ns. The charge per current pulse is approximately 50 pC. The peak value of the current pulse is in the order of 1 mA. The used gap voltages are in the order of a few kV. One of the first experimenters who measured currents in a negative corona discharges was Trichel, after whom the pulses were named
fast current measurements 23
[TRIJS]. He used a measuring resistor of 1 k!l in one electrode lead. The value of the parasitic capacitance was not mentioned. The highest registered frequency was 0. 2 MHz. Nothing is said about the rise time of the current pulses. Miller and Loeb did not mention their fastest possible rise time either [MIL51]. Somewhat earlier, English [ENG48] used a 2 MHz Dumont oscilloscope for his rise time measurements, with a fastest measurable rise time of approximately 200 ns. This is far to slow to measure the 1 ns rise of a Trichel pulse. In 1952 Loeb [LOE52] estimates a rise time of 10 ns at maximum on the ground of light measurements carried out by English [ENG50]. Two years laterAmin measured a current rise time of 10 ns [AMI54]. In 1969 Bugge and Sigmond used a sampling oscilloscope with a rise time of 0.3 ns and from low pressure experiments a rise time of
o.
75 ns for an atmospheric current pulse was estimated [BUG69]. In 1969 Zentner publisbed a thesis on the negative corona discharge [ZEN69]. He used a sampling oscilloscope with 0. 3 5 ns rise time and found a rise time of 1. 6 ns at 1000 mbar. This value was corrected for system influence. In 1985 we publisbed a measured rise time of 1. 3 ns at 1000 mbar [GRA85]. Corrected for the system influence a physical current rise time of 1.15 ns is found.In the following sectien we will discuss the physical interpretation of the externally measured current in detail. We will extend an equation for the electrode current derived by sato for a two electrode system to an equation valid fora multi-electrode system [SAT80]. Then we will check this equation using a laser triggered experiment. With this equation we will try to give an analysis of the shape of the current pulse. Further we will describe an experiment concerning the pboton coupling between two parallel corona needles. we close this chapter with a conclusion on the corona current measurements and their interpretation.
24 negative corona discharges
2.2. Fundamental aspects of current measurements
When charges are moving inside a gap between two or more electrades, a current is induced in each of the electrode leads. Ramo [RAM39] and Shockley (SH038] derived an expression for the current Ii in one of the electrode leads induced by a charge moving in a multi-electrode system. For their calculation method they made use of a hypothetical situation in which all electrades are grounded except the measuring electrode which is set at unit potential. This leads to an electrio field hypothetical but useful for the calculation. The
... velocity of the charge is v
0• The induced and externally
measured current Ii is shown to be:
(2.2.1)
More recently for a two-electrode system sato (SATSOJ pointed out some discrepancies in expressions used by various authors for electrode lead currents when a plasma or more generally moving space charges are present in between the electrodes. He considered a two-electrode system (Fig. 2.2.1) with a constant potential difference va, where E
0 is the Laplacian electrio field and ne, np, ...
we and wp are densities, respectively drift veloeities of positive and negative species. He derived from an energy balance concept in the inter-electrode volume V the following expression for a two-electrode configuration with pla~ma or space charge inside the inter-electrode
FIG. 2.2.1. A two-electrode system wtth moving space charge and externally measured current I.
fast current measurements 25
volume:
I
JIJ
(2.2.2)V
Assuming now N electredes with time independent potentials Vi, we will develop an expression for the current I, flowing into the discharge system through the
.1.
i -th ·lead. Th is measured current I i is brought about by charge carriers actually crossing the i-th electrode surface, but even more by the change in surface distributions on that electrode surface induced by the movement of the charge carriers within the discharge volume. Here we assume that retardation effects can be neglected. The amount qi of the dieleetrio flux leaving a unit charge at a position r within the inter-electrode volume and arriving at the i-th electrode is only dependent on the position of that unit charge. Let j(r) be the current density at position r, p(r) the charge density
...
and let the normal to the surface element, ds1 point from the inter-electrode volume outwards. The current into the discharge volume I1 from electrode i can be written as:
-I. l.
JIJ
V~
•q 1• (r)dV. öt (2.2.3)The boundary of the volume V is partly ohosen at the surface of the electredes
sufficiently far from the contribution to Ii (Fig. equation i t follows that:
and for the rest is situated charges to give a negligible 2.2.2). From the continuity
JJJaà~;>.qi<;>dv
-JIJ
(v•j(r))qi(r)dV.... .... ....
....
(2.2.4)V V
26 negattve corona discharges
FIG. 2.2.2. A N-electrode
system with space charge Qsc inside the inter-electrode
volume V. The boundary of
the integration volume has been dotted.
production or loss of e.g., a positive ion, is always accompanied by the production or loss of an electron. Combination of equation 2.2.3 and equation 2.2.4 gives:
...
v•(j(r)qi(r))dV +
(2.2.5)
The function qi is zero at all electrode surfaces except at the surface of the i-th electrode:
....
(i,j 1, . . . n), (2.2.6)
where rj are the coordinates at the surface of the j-th electrode. With Gauss' theorem applied to the second term on the right hand side of equation 2. 2. 5 the expression for the current 11 changes into
-Ii =
JIJ
....
j (r) •vq1 (r)dV (2.2.7)
V
In this expression the actual electric fields in the inter-electrode volume are implicitly contained in the current density. The function qi is purely geometrical and
fast current measurements 27
bears a streng relation to the concept of capacity in a multi-conductor situation.
In addi t i on to the N conducting electredes we can
assume that an additional, infinitesimally small spherical
....
conductor labelled conductor 0 is placed at a position r
in the electrode space. In a static configuration the
relation between the charges Qi and potentials Vj on these N+1 conductors can be written as:
.Q=~Y (2.2.8)
The capacity coefficient cij can be determined by a
measurement of the charge Qi on conductor i in case the
j-th electrode has unit potential and all other electredes including the i-th are at zero potential. Since the matrix ~ is symmetrie, any two static situations .Q', Y' and .Q" , Y" satisfy
(2.2.9)
In electrastatics this result is known as Gauss' identity [DUR66]. For the first case we choose:
.Q'
Y'
(1 ,Q{,Q~, .•• ,Ql,····Q~) (VQ
1 0 1 0 1 • • • 1 0 1 • • • 1 0 ) (2.2.10a) (2.2.10b)Clearly Ql is the charge induced on electrode i by the
....
unit charge at position r:
For the secend case we choose:
.Q"
Y"
(0 Q" Q" Q" Q")
I 1' 2 ' " ' ' ' i'""'' N
(Vö 1 0 1 Q 1 • • • 1 1 1 • • • 1 0 ) 1
from which fellows from Gauss' identity:
(2.2.11)
(2.2.12a) (2.2.12b)
28 negative corona discharges
Now it fellows that
q. = V" .
l. 0
(2.2.13)
(2.2.14)
Note that the apparent inconsistency in the dimensions of the above equations is caused by choosing the potential V~ l.
and the charge
QÓ
equal to unity.When we move the small conductor through the electrode volume and repeat our reasoning it fellows that -grad(qi)
....
equals the electric field E0i in the inter electrode space in the case that the i-th electrode has unit potential whereas all ether electredes are at zero potential (note that this field is the hypothetical field which is also used by Ramo and Shockley):
... ...
(2.2.15)
Using this expression in combination with equation 2. 2. 7 we write the externally measured current I
1 as:
(2.2.16)
In the case the i-th electrode is at potential V i, the current I. can be written as:
.1.
1
JIJ... .... ....
....
~ j(r)•E01 (r) dV •
.1. V
Wh en we mul tiply bath si des of equation ( 2. 2. 17) by the actual potential
v
1 of the i-th electrode, an expression for the power Pi into the system through the i-th lead shows up:fast current measurements
29
(2.2.18)
The total power input will be:
p
(2.2.19)
....
where E0 (r) is the souree-free electric field caused by
the superposition of all Eoi fields, when all electrades
are at their correct potential vi. The extension of Sato's
work from a two-electrode system to a multi-electrode
system is now complete. When the measured current I of a
two electrode system is written down according to equation
2. 2. 17 then Sato' s expression ( equation 2. 2 • 2) is found .
.... ....
The current density j(r) is taken as:
...
j (r)
(2.2.20)
We will now use a different approach to calculate the
power input in a system. For simplicity we will consider a
two-electrode system. The power input in this system is
given by the Poynting vector:
(2.2.21)
With Gauss' theorem we can write this as:
(2.2.22)
Using vector analysis we can write this as:
30 negattve corona dtscharges
If we assume the electric field to be curl free, then only the second term at the right hand side remains. With Maxwell's equation this can be changed into:
and this equals:
....
f.o öE)dV
öt (2.2.24)
(2.2.25)
The electric field E is assumed to be brought about by
....
superposition of two other fields: E0 and Esc' the space charge free field and the electric field due to the space charge respectively. We assume the space charge free field constant in time. The power input can now be written as:
....
Pin JJJÊ0 ·Îdv + JJJisc·Îdv +
~oJJJi·
8
=~cdV
• (2.2.26)V V V
....
The space charge field Esc is assumed to be curl free, so a potentlal •se can be defined as follows:
(2.2.27)
The potential •se is chosen to be zero at the boundary of the integration volume. The third term at the r.h.s. of equation 2.2.26 can now be written as:
-~ JJJ"""E·~
~0 at """v4 TSC dV 0 (2 2 28) 0 0V This can be written as:
fast current measurements 31
.(2.2.29)
Since ~se is chosen zero at the boundary of volume V, the first term of equation 2.2.29 becomes zero. The power input can now be written as:
The third r.h.s. term can be written as:
Using together with
...
v.j=
21?..
öt ... ... ... ...a. v•(j ~se> - j·v~sc we find for the input power Pin:(2.2.30)
(2.2.31)
(2.2.32)
(2.2.33)
Since ~se is zero at the boundaries of the volume V the fourth r.h.s. term equals zero. Using equation 2.2.27 again we can finally write for the input power in the electrode system:
32 negative corona discharges
The first r.h.s. term gives the power input needed for the
....
current density j to flow in the space charge free field E
0• The second and third r.h.s. terms cancel since the change in electric field energy, the third term [CHI71], causes a current density to flow in the space charge field and gain energy, the second term. In other words: the power needed for transport of particles through the electric field due to space charge is delivered by the change in energy of the space charge field. So the power that must be delivered by the souree of the electrode system is found to be:
=JIJ
(2.2.36)V
which is equal to equation 2. 2. 19, which was der i ved in another way. To derive this equation the following assumptions were made:
I. the total electric field in the gap is curl free. This is the quasi static approximation for electromagnetic fields in an electrode system. Later in this chapter we will discuss the validi ty of this approximation in our experimental situation;
II. the space charge free field is constant in time. This implies that the voltages across the gap are constant during the discharge;
III. the space charge electric field is curl free, which is correct since this field is caused by charges.
To understand better how the power input is related to the charge displacements in the gap, we consider a
a
..---1~V
fast current measurements 33
+ FIG. 2.2.3. Homogeneaus fieLd gap
wt th a charge Q movtng inside the
gap wtth constant veLocity v.
homogeneaus two-electrode system. A schematic view is given in Fig. 2.2.3. The gap voltage is kept constant and space charge effects are neglected. A charge Q is crossing
....
the gap. It takes a time dt to move from pos i t i on r to
....
__...r+dr. An energy balance learns that energy delivered by the power souree equals the work done by the electrio field E
0:
(2.2.37) __... If the electrio field inside the gap is curl free, E
0.dr
has the meaning of a potentlal difference:
E
0.dr = V(r) - V(r+dr) = dV. (2.2.38) The charge flowing from the souree in time interval dt as a resul t of a constant electrode potent i al V can now be written as:
dV
Idt
=
v-·Q .
(2.2.39)So the contribution of the charge Q to the externally measured current-time product Idt is only a fraction of the moving charge Q. Only when Q has crossed the full gap does Idt equal the charge Q.
34 negatiue corona discharges
known to get detailed information on the movements of the changes inside the gap. Furthermore we must realize that
...
EO i in equation 2. 2.16 is the electrio field in the gap when the measuring electrode is set at unit potential and all other electrades are grounded. In a measuring setup, care must be taken that the space charge free electrio field E
0 in the discharge gap is equal to the above mentioned EOi field in the region of interest where the charge carriers contribute measurably to the external current.
2.3. Measurements
2.3.1. The measuring setup
To measure the Trichel pulses a measuring setup has been built to study the fast phenomena involved. We used a set up as drawn schematically in Fig. 2. 3 .1. Opposi te a needle tip a plane is placed which is divided in two: an outer ring electrode and
electrode. The equipotential are plotted with the help of
a small inner measuring l in es drawn in Fig. 2 • 3 . 2 , resistance paper. The solid lines repreeent the equipotential lines when the needle is set at unit potential and both the outer ring and the measuring electrode are grounded. The dotted lines repreeent the equipotential lines when both the needle and the outer ring are set at unit potential and the measuring electrode is grounded. The two field line patterne
m
FIG. 2.3.1. Schematicat drawing of a measurtng setup to study fast
Trichet putses. I - measurtng
eLectrode; II -outer grounded ring; I I I - corona ttp.
fast current measurements 35
- -
----:::::"
/ ' Q J . .-
"1(-n
r
n
FIG. 2.3.2. Measured potenttal distributton inside a
needle-to-plane gap. I - measuring electrode; II -outer ring electrode; III - needle. Solid lines: needle at unit potenttal and bath the measuring electrode and the outer ring electrode grounded. Dotted lines: measuring electrode grouded and bath the outer ring electrode and the needle
at unit potential.
coincide very well in the region close to the needle, so we satisfy the requirement stated at the end of section 2. 2. To speed up the measuring system a cylinder was placed upon the outer ring. This allows the fast current to flow in a compact circuit. The back side of the ring and measuring electrode were bevelled. In Fig. 2. 3. 3 a detailed drawing of the setup is given. This setup is placed inside a stainless steel vacuum vessel which can be evacuated to 10-2
mbar. The setup has several parasitic capacitances as shown in Fig. 2.3.4. The capacitor c
2 is essential in the system since it allows the fast current pulse to flow in a very compact circuit. In Fig. 2.3.5 an overview of the complete setup is given. The resistor Rd decouples the measuring circuit from the H.V. supply lead for high frequencies. The measuring resistor Rm should be
36 negative corona discharges
I
I I
~
i
ri---='aluminumFIG. 2.3.3. Detaited
drawing of a compact and
fast measuring setup.
kept small and ought te be placed close te the gap. Tagether with the parasitic capacitance
c
1 the resistor Rm gives a time constant which gives the fastest measurable rise time.
The signal is taken from Rm and is fed te a Tektronix 7844 oscilloscope with 450 MHz bandwidth. If necessary a 500 MHz preamplifier can be used. The current pulses are stored using an oscilloscope camera and fast polaroid film of 20,000 ASA. The charge per pulse was measured .from the photographs by means of a planimeter.
To investigate the tip influence on the discharge
FIG. 2.3.4. Schematicat drawing
C2 of a fast measurtng setup with
its parasttic capacitances cl,
fast current measurements 37
D
~
N2 -laser I~-[a
Rm Rm 7844FIG. 2.3.5. Compact measuring setup for fast current measurements. I - measuring electrode; I I - earthed ring and screen;
c
1 :l:! 10pF,c
2 :l:! 10pF, Cg :l:! 0.6pF, Rm = 50 11,
Rd = 3 MO. N2 laser: 200 kW, 0.6 ns, À= 337.1 nm.
various needles from different materials with different tip radii were used. We did experiments in several gases. The needle to plane distance was kept constant at 5 mm. The pressure was varied over a wide range. The rise time of a current pulse is defined as the time interval in which the pulse rises from 10% to 90% of i ts amplitude. The parasi tic capaci tance
c
1 was 10 pF, which together with the 50 0 measuring resistor on both sides of the cable leads to a time constant ofo.
25 ns. With the measured rise time of the oscilloscope of 0.6 ns a system rise time of 0.7 ns was achieved.In order to check the theory of the current measurements we carried out laser triggered experiments with a short pulse laser. The laser used is a TEA pulsed N2 laser with a pulse width of 0.6 ns (FWHM) at a wavelengthof 337.1 nm delivering a power of approximately 200 kW [VER82]. With this short laser pulse electrens are liberated from the tip practically instantaneously. These electrens induce a current pulse in the measuring electrode according to equation 2. 2. 17. If the electrio field inside the gap and the electron drift velocity as function of the reduced electrio field are known, the external current can be calculated and compared with the
38 ned~tlue corona discharges
FIG. 2.3.6. Network
representation of a fast
measuring setup. The induced current is
represented by current
souree i •
g
expertntents. 'fhe tip voltage must be chosen low enough to
avoid ionization.
A network representation of the set up is given in Fig. ? .3.6. The gap current is represented by a current souree Under the conditions:
wC2Rd >> 1, Rd » Rn/2,
c
2 >>cg,
the output voltage v0 is found to be: ~/2 •i (2.3.1) (2.3.2) (2.3.3) (2.3.4)
The time consta~t of the measuring circuit is low enough to give only a small phase shift in the measuring signal. So the measured 0utput voltage is directly proportional to the current induced in the lead of the measuring electrode.
We will now check the assumptions made in the derivation of equation 2.2.36. An estimate of the inductance of the compact circuit of Fig. 2.3.3 is found from:
L (2.3.5)
Wi~h a needie having a length h of 5 mm, a diameter dn of
