153 Pages


Course Number: ETD 02262009, Fall 2009

College/University: Texas Tech

Word Count: 23260


Document Preview

OPTICALLY CONTROLLED DIFFUSE DISCHARGES FOR SWITCHING APPLICATIONS by GEORGE ZOHN HUTCHESON. B.S. in E.E.. M.S. in E.E, A DISSERTATION IN ELECTRICAL ENGINEERING Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY Approved August, 1986 ^ Mo , 5' ACKNOWLEDGEMENTS This work could not have been completed without the...

Unformatted Document Excerpt
Coursehero >> Texas >> Texas Tech >> ETD 02262009

Course Hero has millions of student submitted documents similar to the one
below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

Course Hero has millions of student submitted documents similar to the one below including study guides, practice problems, reference materials, practice exams, textbook help and tutor support.

CONTROLLED OPTICALLY DIFFUSE DISCHARGES FOR SWITCHING APPLICATIONS by GEORGE ZOHN HUTCHESON. B.S. in E.E.. M.S. in E.E, A DISSERTATION IN ELECTRICAL ENGINEERING Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment of the Requirements for the Degree of DOCTOR OF PHILOSOPHY Approved August, 1986 ^ Mo , 5' ACKNOWLEDGEMENTS This work could not have been completed without the assistance of many people. I would like to express my appreciation to Prof. Gerhard Schaefer for his guidance, support, and patience. I would also like to thank Prof Karl H. Schoenbach. Horn Prof. Magne Kristiansen. and Prof. Lynn Hatfield for their encouragement and technical advice. I must also thank Horn Prof. John F. Walkup for serving on my committee and for his encouragement. I would like to thank several of my fellow students who provided assistance for different tasks described in this dissertation and to whom I will always be indebted: Randy Cooper for his work on the ultraviolet ionization spark array and the high pressure discharge cell described in Chapter III; George X. Ferguson for his work on the high pressure discharge cell: Leo E. Thurmond for his work on photodetachment in the wall stabilized discharge device; Ed Strickland for setting up the fast photography equipment used to study the UV spark array; Rick Korzekwa for making the initial breakthrough on the optically enhanced attachment experiments: and, Courtney Holmberg for performing most of the optically enhanced attachment experiments. I must also express my appreciation to all of the technicians (especially Kim Zinsmeyer). undergraduate ii students, and graduate students of the Pulsed Power and Plasma Laboratory. Finally, I must thank my wife. Susan, for allowing me to pursue this effort. Only through her sacrifices could I have completed this project. Ill T A B L E OF C O N T E N T S ACKNOWLEDGEMENTS ABSTRACT LIST OF TABLES LIST OF FIGURES I. II. INTRODUCTION D I F F U S E D I S C H A R G E SWITCHES Inductive Energy Storage D i f f u s e Discharges Optical Control Considerations 18 21 22 25 30 ii vi vi i 1 ix 1 8 8 10 for Diffuse Discharges III. ULTRAVIOLET IONIZATION Theoretical Considerations D e s i g n C o n s i d e r a t i o n s for the UV Controlled Discharge Experiments Optical Investigations Discharge Investigations Using the UV Source IV. PHOTODETACHMENT Circuit Considerations for Opening Swi tches Theoretical Considerations Attachment and Photodetachment in Oxygen and Oxygen Bearing Molecules 39 47 48 51 60 IV Experiments Utilizing a Flowing Afterglow Apparatus Experiments Utilizing a Wall Stabilized Glow Discharge Apparatus Experiments Utilizing the UV Sustained/Initiated Discharge Apparatus V. OPTICALLY ENHANCED ATTACHMENT Dependence of the Attachment Cross-Section on Vibrational Excitation Methods of Vibrational Excitation Experiments Utilizing the Wall Stabilized Glow Discharge Apparatus IR Pumping in Discharges Sustained by Plasma Injection Photodissociation in Discharges Sustained by a Hollow Cathode Device VI. CONCLUSION 63 68 70 85 86 91 108 110 118 123 130 LIST OF REFERENCES ABSTRACT Diffuse discharges containing electronegative gases, at present, are the only means capable of fast, repetitive, long-life operation as opening switches. Optical control of diffuse discharge switches has been proposed as both a means of sustaining and of enhancing the performance of such switches. Processes considered in this dissertation are photoionization. photodetachment. and several approaches to optically enhanced attachment. Ultraviolet (UV) ionization has been used for several years now as a means of preionizing self-sustained diffuse discharges and, recently, has been used as a sustainment mechanism for diffuse discharge opening switches. Time-resolved measurements of the optical emission of a spark array, similar to those used for preionization or sustainment of diffuse discharges, are presented. Results of experiments in atmospheric pressure diffuse discharges containing admixtures of attachers, sustained and initiated by such a spark array, are also presented. Photodetachment is considered as a mechanism which could decrease switch losses and decrease switch closure time by counteracting dissociative attachment. Experimental results are presented demonstrating optically increased current densities, of as much as 900%, in VI externally sustained and externally initiated, atmospheric pressure, diffuse discharges containing 0^. proceeds through photodetachment of 0 This process by 590 nm light generated by a dye laser (0 +hu - 0+e). Optically enhanced attachment through the generation of vibrationally excited attachers appears particularly attractive as a means to decrease turn-off or opening times for diffuse discharge switches. For some molecules attachment cross-sections can be increased by orders of magnitude by vibrational excitation. The influence of this effect on the resistivity of a discharge through IR pumping of N . ^ and CpH^aF by a low power CO^ laser in continuous low H; pressure discharges is presented. UV enhanced attachment is also shown in gases containing molecules such as C^HpF^. CpHpClp, and CpH-Cl where UV photodissociation produced vibrationally excited, strongly attaching, molecules (e.g.. vii LIST OF TABLES 1. Decay Time Ratios for the Near UV and Visible Emissions from the UV Source Energy Density Loss Per Switch Cycle for the Curves Shown in Fig. 12 and 13 [21] Important Processes in 0 Discharges Photoenhanced Electron Attachment Summary of Resul ts 40 2. 56 64 93 124 3. 4. 5. viii LIST OF FIGURES 1. 2. Inductive energy discharge circuit Schematic representation of the optically controlled diffuse discharge switch experiment Streak photography of the ultraviolet source with a charging voltage of 33 kV (in air) Turn-on time (i.e. propagation of are development) of the ultraviolet source as a function of charging voltage Time resolved photograghs of the ultraviolet source (a) prior to complete turn-on and (b) after turn-on Relative intensity versus time relationship for the first spark gap of the ultraviolet source with a charging voltage of 33 kV Relative intensity versus time relationship for the whole ultraviolet source (20 gaps) with a charging voltage of 33 kV (visible spectral range) Current density, J. versus reduced field strength. E/N, for an E-beam sustained discharge and for a UV sustained discharge, both in argon with an admixture of 10% C2Fg Time dependence of current and voltage for a UV initiated discharge in a CO2 laser mixture(C02: N2: He=l: 16 : 23) Current density, J, versus reduced field strength, E/N, for an E-beam sustained discharge in argon with an admixture of 2% C2Fg. obtained 9 26 3. 32 4. 33 5. 35 6. 36 7. 38 8. 42 9. 45 10. ix at different times after E-beam initiation and with different load lines [28] 11. Time dependence of E/N (top), current density (middle), and power density loss (bottom) of an E-beam sustained, laser photodetachment assisted discharge in 1 atm N^ with an N^O fraction of 1%. The system impedance is Z=20 n. The E-beam and laser are on for 0<t^lOO ns. The variable parameter is the laser power density [21] 12. Time dependence of E/N for an E-beam sustained, photodetachment assisted discharge in 1 atm N^ with an N^O fraction of 1%. The system impedance is Z=20 n. The laser power density 7 2 is 10 W/cm. The E-beam is on for O^t^lOO ns. The variable parameter is the laser pulse length (see Table 2) [21] Time dependence of E/N for an E-beam sustained, photodetachment assisted discharge in 1 atm N^ with an NgO fraction of 1%. The system impedance is Z=20 n. The laser power density 7 2 is 10 W/cmT The E-beam is on for O^t^lOO ns. The variable parameter is the delay of the laser pulse with respect to the E-beam (see Table 2) [21] 14. Photodetachment cross-section for 0 and O" ( Smith, et al. [54]; o Branscomb. et al. [55]; Burch, et al. [56]) 15. Diagram of apparatus for photodetachment experiments in the flowing afterglow [20] 49 55 57 13. 58 61 65 16. Relative magnitude of optogalvanic effect as a function of laser pulse energy in the flowing afterglow [20] Diagram of apparatus used for photodetachment experiments in the wall stabilized glow discharge Optogalvanic effect (as ratio of peak optogalvanic voltage change to DC voltage) as a function of discharge current in the entire discharge Optogalvanic effect (as ratio of peak optogalvanic voltage change to DC voltage) as a function of discharge current in the positive column VI - characteristics of UV sustained and UV initiated discharges for several concentrations of 0^ with 2.6% N^ and balance of Ar to make 1 atm 67 17. 69 18. 71 19. 72 20. 75 21. VI - characteristics of UV sustained and UV initiated discharges in 7.9% O2. 2.6% N2. and 89.5% Ar at 1 atm without and with laser photodetachment 76 22. The influence of the laser on the current density (AJ/J , where o AJ = J,(with laser) - J (without La w laser)) versus charging voltage of the PFL for various concentrations of O2 with 2.6% N2 and balance Ar to make 1 atm 23. VI - characteristics of UV sustained and UV initiated discharges for several concentrations of N2 with 5.3% 0^ and balance of Ar to make 1 atm 78 79 XI 24. The influence of the laser on the current density (AJ/J , where AJ = J^(with laser) - J (without laser)) versus charging voltage of the PFL for various concentrations of N2 with 5.3% 2 and balance of Ar to make 1 atm 80 25. Current density change (AJ = J. (with laser) - J^(without laser)) for varying laser intensities in 13.2% O2. 2.6% N2. and 84.2% Ar at 1 atm and an E/N = 53 Td 82 = 26. Change in discharge resistances (AR/R o R^(without laser) - R.(with laser)) versus PFL charging voltage for 13 kn system impedance. Gas mixture contained 2.6% Np, varying concentrations of 0^> and a balance of Ar to make 1 atm 27. Potential energy versus internuclear distance curves illustrating resonant dissociative attachment Dissociative attachment cross-section for HCl (e+HCl-*H+Cl") for different vibrational quantum numbers, u, [84] and [93] 29. Potential energy versus internuclear distance curves for a single bond of a triatomic molecule. Subscripts s and b refer to straight and bent species, respectively Potential energy versus internuclear distance curves for I2 and I2. illustrating the method of electronically enhanced attachment for Ig [93] 92 84 87 28. 101 30. 107 xii 31. Influence of CO^ laser on the discharge voltages (AV/V , where AV = Vj^(with laser) - V (without laser)) for various pressures of 57% NH3 and 43% Ar 109 32. Illustration of apparatus used to study IR enhanced attachment in externally sustained DC discharges Current versus E/N for several of the gas mixtures investigated in the apparatus of Fig. 32 Influence of CO^ laser on the discharge voltages (AV/V . where AV = V. (with O lirf Ill 33. 113 34. laser) - V (without laser)) for various concentrations of NH., with 6% He and balance of Ar to make 32 torr Influence of CO^ laser on the discharge voltages (AV/V , where AV = V. (with O I' 115 35. laser) - V (without laser)) for 7.6% C2H3F. 87.4% Ar. and 5% He at 36. various pressures Influence of CO2 laser on the discharge voltages (AV/V , where AV = Vj^(with laser) - V (without laser)) for ' o 15% C2H3F. 80% Ar, and 5% He at various pressures 37. 38. Apparatus for photodissociation enhanced attachment experiments Change in relative discharge resistances versus E/N containing CH.^C1 caused by enhanced attachment through UV photodissociation 117 119 116 121 xiii CHAPTER I INTRODUCTION Optical control of a gas discharge is the ability to change the conductivity of a gas discharge by optical means. Space dependent illumination of a discharge also It has been known provides spatial control of a discharge. since the beginning of the century that light emitted from a spark will cause a gas to become electrically conductive [l]-[3]. However, the need for optical control was not A necessary realized until the invention of the gas laser. condition for the proper operation of high power, pulsed, gas lasers is the creation of a stable volume discharge, or "diffuse discharge." as a lasing medium. One requirement for creating such discharges is to preionize the laser gas before the voltage pulse is applied to the gas. Ultraviolet (UV) spark sources have successfully been used for this purpose [4]. Such preionization allows the initiation of homogeneous volume discharges and avoids the formation of arcs. Denes and Kline [5], for example, demonstrated that a volume discharge suitable for laser excitation would only form where the gas had been illuminated by a UV preionizer. With the advent of the laser it became possible to use the laser as a light source to influence the properties of a gas discharge. For example, laser triggered spark gaps. may be considered a form of optical controlled conductivity. Here a spark gap is illuminated by an intense laser pulse focused somewhere in the middle of the gap (depending on the wavelength of the laser used). Through multi-step ionization and inverse bremsstrahlung the light forms a plasma which quickly expands, due to the electric field of the gap. and closes the gap. Laser triggering initiates conduction through the gas by converting an insulting gas into a conducting plasma. initiated, the plasma is only weakly influenced by the laser. Therefore, this treatise will not cover this Once subject; the interested reader is referred to references [6]-[8]. In the late 1960's and more vigorously in the mid 1970*s, again after the invention of the laser, wavelength dependent changes of the conductivity of discharges was demonstrated. Many investigators studied how the conductivity of various gas discharges would change when illuminated by light having a wavelength corresponding to a specific optical transition in the gas [9]-[17]. This process, most commonly referred to as the optogalvanic effect , was of interest as a spectroscopic tool and Throughout this paper, however, "optogalvanic effect" will be used for any mechanism in which conductivity changes can be induced by optical illumination. recently has been investigated by Lawler [16], [17] as a switching mechanism. Very recently, Guenther [18] and Schoenbach. Schaefer. Kristiansen, Hatfield and Guenther [19] proposed the use of optical control techniques for switching purposes. Specifically, they were interested in using such techniques for high power, fast, repetitive opening switches for pulsed power applications. In addition to using transition induced optogalvanic effects, described above, and multi-step ionization proposed by these authors. Schaefer and Schoenbach et al. have also proposed the use of photodetachment [20], [21] and optically enhanced attachment [19] as control mechanisms. These authors argue that, although a diffuse discharge switch could be built using optical control mechanisms alone, optical control will prove more efficient as additional control mechanisms for other types of diffuse discharge switches. The main reason for this approach is that laser photons, especially high energy photons necessary for ionization, are too "expensive." Optical control techniques may prove to enhance the performance of other types of diffuse discharge switches in at least four ways. Firstly, the conductivity of diffuse discharges may be increased (say through photodetachment) during the steady state or conducting phase of the discharge in order to lower the power losses during switch conduction. Secondly, the switch turn-on or risetime may be shortened by using mechanisms which, again, increase the conductivity of the discharge (e.g., photodetachment). this case, however, the control mechanism would only be utilized during the formation or turn-on time of the discharge. A third possible control technique would In involve decreasing the discharge conductivity during the termination or turn-off phase of the discharge in order to decrease the opening time of the switch. Optically enhanced attachment or optical quenching of important metastable states could be used to decrease discharge conductivities. Lastly, optical control may be able to lower the probability of diffuse discharges transforming into arcs during conduction. The ability to limit this so called glow-to-arc transition may allow longer switch conduction times and possibly higher repetition rates. It is the purpose of this dissertation to analyze the influence of these processes on a discharge and to describe proof-of-principle experiments which demonstrated that the mechanisms for these optical control techniques proposed for switching applications do indeed work. In order to fully understand how these control techniques can be used, the following chapter presents a discussion of the present work in diffuse discharge opening switches and the properties of the gases used for such switches. The chapters which follow present theoretical discussions and experimental results of three optical control processes: ultraviolet sustainment and initiation of diffuse discharges: photodetachment of oxygen bearing discharges; and optically enhanced attachment in low pressure discharges. Specifically, the research conducted for this dissertation found the following: 1. Spark arrays can be used as ultraviolet sustainment sources for diffuse discharges containing attachers. UV sustained discharges containing argon and C^F^ and containing argon, nitrogen, and oxygen were investigated. These gas mixtures exhibited current density versus reduced field strength (E/N) characteristics having regions of negative differential conductivity as a result of increasing attachment rates with increasing E/N at intermediate E/N ranges. 2. However, it will be shown that the turn-on times of such UV sustained diffuse discharges are limited by the serial breakdown time of the spark array. Also, the decay time of the UV emission of the spark array (which is determined by the recombination rate of free electrons in the gas) limits the turn-off times of the diffuse discharges it sustains. A simpler circuit for initiating selfsustained laser discharges can be utilized by choosing a laser gas mixture, an operating pressure, and a gap spacing such that the breakdown voltage of the laser gap is twice the glow voltage of the gap. In this way, a transmission line may directly drive the laser without any intermediate swi tches. Increased current densities can be induced through photodetachment of negative ions by illuminating atmospheric pressure, diffuse discharges containing oxygen with moderate 2 power, pulsed, visible laser (800 kW/cm , 590 n m ) . Current density changes of as much as a factor of 9 were observed. 5. Photodetachment can enhance the stability of atmospheric diffuse discharges containing oxygen. This was demonstrated by illuminating such discharges with a pulsed visible laser. 2 6. Low power laser illumination (< 1 W/cm ) of low pressure, continuous discharges sustained by a plasma injection device can cause enhanced dissociative attachment via vibrationally excited attachers in such discharges. Discharge voltage increases of nearly 1% were observed in argon with admixtures of either NH.^ or CpH~F. 7. Similarly, low pressure, continuous discharges, also sustained by a plasma injection device, showed increased voltages when illuminated by UV light from a spark array. Changes of as much as 30% in discharge resistances were observed in gases of helium with either C2H2F2. C2H2CI2. or C2H3CI. CHAPTER II DIFFUSE DISCHARGE SWITCHES Before examining optical control mechanisms for diffuse discharges, a discussion of the motivation for using such discharges as switches must be presented and some information on the properties of diffuse discharges is necessary. Inductive Enerev Storage In recent years there has been considerable interest in using inductive energy storage elements for pulsed power applications. At present inductive energy stores have significantly higher energy storage densities than capacitive stores thus allowing more compact high energy pulsers [22]. in Figure 1. An inductive energy storage circuit is shown Initially, the opening switch, S , is closed Energy is slowly At an and the closing switch. S . is open. transferred from the source to the inductor. appropriate time (generally the time at which the current in the inductor has reached its maximum value), S opened and simultaneously S is is closed, and the energy stored in the inductor is rapidly transferred to the load. For a real system, the opening switch is not ideal and opens with some finite time. As the switch opens, current 8 Inductor CTJO Source 6 A^5 Fig. 1. Inductive energy discharge circuit 10 through the inductor is forced to change. This changing current induces an increased voltage pulse across the inductor, due to L(di/dt). which in turn impresses an increased voltage pulse across the load. If the impedance of the load is smaller than the impedance of the source, some of the energy stored in the inductor will be delivered to the load in a shorter time than the time for charging the inductor. Thus, the circuit will have delivered a Of course, voltage and power amplified pulse to the load. for high voltage multiplication, the opening time of the switch should be as small as possible. Unfortunately, opening switch technology is. as yet. not a mature technology. Opening times on the order of nanoseconds have been achieved using single shot devices [22]. [24] and repetitive circuits have been built with opening times in the milliseconds. At this time, only one devicethe diffuse discharge switchappears capable of achieving both fast opening times and high repetition rates. Diffuse Discharges A diffuse discharge is characterized as a cold, weakly 14 -3 ionized (electron densities ^ 1 0 cm ) gas discharge 3 which is bounded by a large volume (10 cm 3 to 1000 cm ) . These discharges do not exhibit Local Thermodynamic Equilibrium (LTE); that is the electron temperatures 11 (typically a few eV) are much higher than the molecular and ion temperatures (typically room temperature). As a result, collisions of electrons with neutrals play a major roll in the behavior of diffuse discharges. More importantly, gas additives can be used in these discharges in such ways that the additives may enhance or diminish the conductivity of the discharges in desirable ways. Typically, the electron velocity distributions of diffuse discharges are not Maxwellian since the discharge electrons are strongly influenced by elastic collisions. Since diffuse discharges are bounded by large volumes and, therefore, have large cross-sections perpendicular to the E-fields, diffuse discharges have lower inductances and lower electrode erosion rates than other forms of discharge swi tches. As a result of these characteristics, the conductivity of diffuse discharges can be influenced or controlled. The electron collision processes in these discharges allow the molecules in the gas to influence the conductivity. Control can take the form of an external device which modifies a collisional process in the gas. or the control can take the form of using a gas mixture which gives the discharge certain desired electrical characteristics. present, the most common control mechanisms for diffuse At 12 discharge switches uses externally sustained discharges in appropriately designed gases. Diffuse discharges may be operated in one of two modes: self-sustained, or externally sustained. A self-sustained discharge is operated at a reduced field strength (the electric field strength divided by the number density of particles), E/N, above the threshold for ionization. In this mode, internal ionization forces the discharge to remain conducting until the electric field applied to it terminates. In an externally sustained discharge. E/N is below the threshold for internal ionization. Conducting electrons are produced by some device, external to the discharge, which ionizes the gas. By controlling the ionization rate externally, the conductivity of the discharge can be controlled. For opening switches, the external ionization source would be on during the conduction phase, and, at the moment when the switch is to open, the ionization source would be turned-off. Demonstrated external ionization sources for externally sustained diffuse discharges include electron beams (E-beams) [25]-[28] and ultraviolet (UV) light sources [29]-[37]. Unfortunately, for opening switches, recombination alone will not remove the charge carriers fast enough at the time that the ionization source is turned-off and the 13 switch is to open. In order to increase the rate of charge carrier loss at switch opening, gases which have strong resonant dissociative attachment rates are added to the gas mixture. In resonant dissociative attachment an electron collides with either a diatomic or polyatomic molecule. The electron is absorbed by the molecule forming a temporary, excited, negatively charged molecule. The molecule quickly dissociates into a negative ion and a neutral fragment: e + AB - AB~ * - A + B"". > * > Of course, the ions, both positve and negative, have much lower mobilities than the electrons. Thus, the electrons previously associated with conduction have been absorbed into negative ions with lower mobilities: conduction is greatly reduced and the discharge, as a switch, opens. The discharge kinetics can be modeled in a simplified way by considering the discharge as spatially homogeneous and only considering the charged particles [38]. All ion species have such low drift velocities that all of the current is carried by the conduction electrons. current density. J. is then given by: J = en^Vp(E/N), (1) The 14 where e is the charge of an electron, n e electrons, and Vp(E/N) is the drift velocity of electrons which is dependent upon the reduced field strength. E/N. Again, the variable in Eq. (1) that allows control of the conductivity, or. alternatively, the current density, is n^. The rate equations which determine the densities of is the density of the charged particles are given as: dn dF^ = S " ^ion'^ % - ^ e c ' * . ^ - ^att'^att-e' ^2) (3) ^ ' (4) * ' dN^ ^- = S + k . Nn -k N^n - k. N^N . dt ion e rec + e lonrec + dN X T - = k ,^N ^^n - k, N^N . dt att att e ionrec + where n , N , and N_ are the electron, positive ion. and negative ion densities, respectively: k. . k . k . and k. are the ionization, electron-ion recombination, ionrec attachment, and ion-ion recombination rate coefficients, respectively; N is the density of the gas: N is the density of the attacher; and S is the source function (rate of electron-ion pair production per volume by the external ionization source). All of the rate coefficients are dependent on E/N except ^^onrec' For an externally sustained discharge, internal ionization, k. N n . is neglected: while for a ion e self-sustained discharge, the source function, S. does not exist. Similarly, some assumptions about the loss terms 15 can be made. If no attaching gas is present the attachment term. ^^^^^3^^. ^^^ ay be removed from the above equations. However, if there is a sufficient amount of attaching gas in the discharge, the attachment term may dominate the electron-ion recombination term, k >> k. N n . att att e ion + e N n Thus, for an externally sustained, attachment dominated, discharge (an opening switch discharge) the rate equations are given as: dn J T = S - ^att^tt'^e' dN^ dt dN TTdt (5) (6) (7) ^ ^ = S - k ionrec N+N_. ,__ = k , N ^^n - k, N^N . att att e ionrec + - It is interesting to compare the time dependent electron losses between electron-ion recombination and attachment during the time immediately after the ionization source has terminated. For a recombination dominated n = N , and discharge, charge neutrality requires that Eq. (2) can be rewritten as: dn /dt = - k n e n rec e Since the relative rate of change of the electron density, (dn /dt)/n , is dependent on the electron density, the rate of electron loss decreases with lower electron densities; 16 that is, for n < 10 14 cm 3 the loss is relatively slow. I^ ^rec ^^ assumed constant, then the solution to the above equation is given by: TT = ;r + k t. n n rec e eo (8) ^^ where n^^ is the electron density at the time immediately before the source was terminated. Now. for an attachment dominated discharge (k N >> k N ^ ) , charge neutrality att act rec^ requires N = n + N_, and Eq. (2) is changed to read: dn /dt n = -k att att ,,N ,,. e Here the relative rate of decay is not dependent on the electron density. The loss of electrons can be controlled entirely by choosing an attacher with an appropriate k att and by selecting the correct concentration of the attacher. N att . If k ^ is assumed constant, then the above equation att ^ has the solution [39]: From the above analysis, for the case of constant coefficients, it can be seen that attachment can be made a faster electron decay process than recombination. Although fast electron decay through dissociative attachment is an important consideration, there are 17 additional gas properties to consider. Basically, the opening switch gas mixture should meet three conditions: 1. During the conduction phase, the switch voltage and E/N will be low. At this low E/N, J should be high in order to minimize power loss in the discharge. For high J at low E/N, the attachment rate should be low and the electron drift velocity should be high. 2. During the opening phase, the switch voltage and E/N will be increasing with time. current density should decrease with increasing E/N in order to insure fast opening. In order to have decreasing J with The increasing E/N, the attachment rate should increase and the electron drift velocity should decrease with increasing E/N. 3. Finally, the gas should have a high hold-off voltage for the time that the switch is open. 18 These gas conditions can be met by selecting an attaching additive that has the k ^^ versus E/N characteristics att described above. The drift velocity requirements can be achieved by selecting a buffer gas for the switch mixture that has a decreasing v_. with increasing E/N. To meet these requirements. Christophorou et al. [40], [41] have proposed using perfluoro alkanes (e.g.. CF., CgFg. and C^Fg) as attachers and argon and methane as buffer gases. Optical Control Considerations for Diffuse Discharges The above equation for the conductivity and the rate equations. Eqs. (l)-(4), can be modified in a variety of ways by considering optical control mechanisms. Although E-beam ionization sources are the most common form of external sustainment devices, there are advantages in using ultraviolet sustainment sources. Electrons with sufficient energies for ionization of a gas can be produced relatively efficiently by electron beam guns. Unfortunately, photons with energies necessary for ionization of a gas (wavelengths in the UV) cannot be produced with the same degree of efficiency as ionizing electrons. However. UV sources may be more reliable and simple to operate than E-beam sources. For UV sustained discharges, the controlled parameter in Eqs. (l)-(4) is the source function. S. Additionally, the carrier balance may be 19 considered to be modified by additional external control mechanisms such as photodetachment of negative ions by laser radiation. The generation rate of electrons can also be optically modified through the ionization coefficient, k. ion The ionization coefficient actually represents the sum of many collisional excitation coefficients. Most molecules in a self-sustained discharge are not ionized by a single electron collision. Ionization occurs through multiple collisions in which the molecule is excited to an energy level closer to ionization than the previous level. These excitation coefficients in k. are coefficients for the ion intermediate excitations necessary to ionize the gas. If one of the energy levels that is an intermediate step to ionization is populated by optical pumping (i.e.. using laser light) from a lower level, the ionization rate will appear to have been increased [6]-[14]. The electron depletion rates can also be influenced by optical means. The electron-recombination [42] and This attachment [43] rates are temperature dependent. temperature dependence is due to vibrational excitation dependence of the recombination and attachment collision cross-sections. By vibrationally exciting molecules responsible for recombination or attachment, by optical means, these electron decay rates can be increased. Since 20 recombination is slower than attachment, there is generally little advantage in enhancing the electron-recombination rate. The following chapters describe in more detail how some of these optical control techniques can be used in discharges similar to opening switch discharges (i.e.. externally sustained discharges in attaching gas mixtures). Specifically. UV ionization as an external sustainment source and initiator of self-sustained discharges: photodetachment as a conductivity enhancement mechanism; and optically enhanced attachment as a conductivity diminishing process are described. CHAPTER III ULTRAVIOLET IONIZATION The apparatus described in this chapter was designed to investigate optical control mechanisms using lasers (i.e., photodetachment and optically enhanced attachment). These processes have to be investigated over a wide E/N range from below 1 Td to up to the self-breakdown voltage of the discharge. These variations of E/N require the possibility to operate the discharge in the self-sustained and in the externally sustained mode. In the externally sustained discharge mode, ionization was provided by a UV source. This chapter describes the properties of the UV ionization source used in this apparatus and describes self-sustained and externally sustained discharges generated by the UV source. Although the device was built to study other optical control techniques, the design features and the results of the experiments described in this chapter are applicable to self-sustained laser discharges and externally. UV sustained laser and switch discharges. There is \ considerable evidence that UV sustained discharges for lasers provide discharge energy loadings that are comparable to E-beam sustained discharges [29]. [30]. Ultraviolet sustained discharges for opening switches have also achieved current densities approaching the values 21 22 achieved with E-beams [35]. Combined with the simplicity and reliablity of UV sources, these characteristics make UV ionization sources competitive with E-beam sources. The research described in this chapter was conducted with the assistance of several people. James R. Cooper designed and built the pulse forming system for the ultraviolet source. Bryan E. Strickland set up and helped perform the fast photography experiments described below. The construction and operation of the remaining apparatus described in this chapter was assisted by George X. Ferguson, James R. Cooper, and Leo E. Thurmond. Theoretical Considerations Photoionization is a mechanism that may proceed through either single step or multi-step processes. However, for preionization or sustainment of diffuse discharges, single step ionization (hu+A - A +e) is the major mechanism [4]. Unfortunately, most molecules and atoms have relatively high ionization energies that are not easily accessible by lasers. For example, cesium has the lowest ionization energy of any atom at 3.87 eV [44] which corresponds to 319.8 nm - well into the ultraviolet. alkali [45], [46] and noble atoms [46]-[49] exhibit a strong maximum in the photoionization cross-section just above the threshold energy. The cross-sections decrease Vibrational and Most for photon energies above the threshold. 23 rotational states for polyatomic molecules cause the photoionization cross-sections to be smeared out over wide absorption bands [50], [46]. [47], [51], [52]. Extensive studies were conducted in the 1970's of low ionization potential additives for C0 laser preionization [4]. It was found that most of the ionizing radiation generated by spark sources was absorbed very close to the sources and that little of the radiation entered the active volume needing ionization. However, a narrow window around Organic additives were 120 nm allowed UV penetration. found with ionization energies corresponding to this narrow transmission window. Most of these additives are complex hydrocarbons with amines and anilines being the most common forms used for diffuse discharges. These compounds exhibit wide photoionization cross-sections and have threshold energies between 7 and 9.5 eV. Although diffuse discharges have gained recent interest as switches, they have been employed as lasing media since the early 1970*s [4]. A preionization source is usually required to generate stable, arc-free, self-sustained discharges. Preionization generates a homogeneous distribution of electrons at low densities. These electrons, under the force of an electric field, expand into avalanches. Arcs are prevented from forming if the avalanche heads overlap before the time that the 24 electron density of the avalanche heads have reached the critical value for streamer formation [53]. [54]. electron balance equations. Eqs. (2)-(4). for a self-sustained attachment dominated discharge, are written as : file = k, ^Nn - k ^ N ^ n J ion e att att e, + j^ j^ ut = k, Nn - k. N^N , ion e ionrec + = katt att n e - k. ^^N ^ ionrec N^N - . + (10) ^ ^ (11) ^ ' (12) ^ ' The For steady-state conditions, a self-sustained discharge will only operate at a particular value of E/N which is determined from Eq. (10) and given by: k. N = k ^ N ^^ ion at t at t. (13) ^ ' While self-sustained discharges require preionization, in an externally sustained discharge ionization totally depends on the external source. Such an externally sustained discharge can be used as an opening and/or closing switch. In an externally sustained diffuse discharge opening switch the controlling mechanism ionizes the gas in the discharge, forming a conducting medium which continues to conduct as long as charged particles are present. Once the ionization source is removed, electron attachment and recombination processes in the gas cause the 25 electron density to decrease until the discharge ceases. thus forming an insulating medium. The electron balance equations for an externally sustained attachment dominated discharge are given by Eqs. (5)-(7). The requirements for the operation of the external ionization source strongly depend on the specific application. In the case of an externally initiated discharge it is mainly the turn-on characteristics of the ionization source. S, which determine the performance of the discharge, while in the case of an opening switch the turn-off characteristics and the charge carrier loss mechanisms, k N n , determine the opening time. a c L act e Design Considerations for the UV Controlled Discharge Experiments The optically controlled switch experiment used to study optogalvanic effects is shown in Fig. 2. The major component is the discharge chamber with a TEA laser electrode configuration (variable gap distance, d = 3.510 mm, active electrode width, w = 20 mm, electrode length. 1 = 200 mm). The electrodes of this chamber are connected A resistor in series to a 125 Q coaxial transmission line. with the line can be installed to vary the system impedance. Additionally, a laser triggered spark gap may be placed between the line and the discharge chamber in order to synchronize precisely the application of a 26 LASERJIBEAM ^^NWZZ oJ\/sj< LASER TRIGGERED SR\RK GAP Fig. 2. Schematic representation of the optically controlled diffuse discharge switch experiment. 27 rectangular voltage pulse to the discharge gap. The time dependence of current, voltage, and impedance of the generated discharges are measured with current and voltage probes in the main line. The discharge volume can be illuminated with a laser through side-on windows in the discharge chamber. This experiment utilizes an ultraviolet radiation source which can be used as a preionizer for self-sustained discharges or as the external ionization source for an externally sustained discharges. To produce a homogeneous self-sustained discharge at high pressure, such as in TEA lasers, it has been proven that the transverse homogeneity of the ionization source is one of the key requirements for arc free operation [4]. The best way to incorporate an ultraviolet source into a discharge device is. therefore, to illuminate the discharge volume from behind one electrode. This requires that the electrode be optically Placing the ultraviolet source transparent or a mesh. behind the electrode also assures that the maximum photoionization effect occurs directly between the parallel surfaces of the switch electrodes. When operating discharges for lasers or switches it is desirable to maintain a constant gas environment in the discharge chamber. Arcs are good sources of ultraviolet radiation and can be created by surface discharges or bare 28 sparks [4]. Surface discharges are actually better sources of ultraviolet light than bare sparks [4]: however, surface discharges erode some part of the insulator surface with each discharge. The by-products of surface discharges can Therefore, a bare contaminate the discharge environment. spark source should be used whenever it is important to maintain the integrity of the gas in the discharge chamber over a large number of shots. In order to optimize the turn-on and turn-off characteristics and the efficiency of the source, a pulse forming network providing a rectangular current pulse should be used and the impedance of the ultraviolet source after breakdown should be matched to the impedance of the pulse forming network which excites the source. To achieve fast breakdown of the spark gaps of the UV source, high applied voltages are also required. A stripline design Since the allows optimization of these conditions. impedance of an individual spark is on the order of 25 milliohms [35], many spark sources in series have to be incorporated into the transmission line. Nearly the full charging voltage will then be applied to each individual spark source before breakdown and the impedance of the line can be matched to all spark sources in series after breakdown. 29 The impedance Z o available for this experiment was 9.4 Q [55], [56]. Therefore, approximately 375 individual spark sources would be needed to match the impedance using only the resistance of the sparks. Given the physical size limitations of the of the pulse forming network discharge chamber, only 20 spark sources in a linear array were used. Therefore, an additional resistance was incorporated at the end of the array to match the impedance of the ultraviolet source to the pulse forming network and to avoid wave reflections. A practical and efficient system can operate with a 2.5 0 pulse forming network and approximately 100 sparks in series. Utilizing the above considerations, an ultraviolet source was designed and constructed which consists of 20 individual spark gaps with approximately 1 mm gap spacing in a single linear array and terminated with an additional matching resistance. When the switch spark gap shown in Fig. 2 was triggered by a ruby laser, a pulse with a voltage of V /2, a current pulse of I^=V^/2Z . and a pulse length of approximately 10 ns was delivered to the transmission line. The magnitude of the charging voltage was usually kept below the DC breakdown voltage of the source as a whole (the sum of the breakdown voltages of the spark gaps) but well above the DC breakdown voltage of a single spark gap. Since the individual spark gaps were 30 incorporated into the transmission line, the array would break down in a serial manner. The turn-on time of the entire ultraviolet source was, therefore, determined by the delay between the light emission from the first and the last spark gap. This delay time included the transit time of the pulse along the transmission line from the first to the last spark gap and the sum of the breakdown times of all spark gaps. Optical Investigations In order to understand the turn-on and turn-off characteristics of the ultraviolet source as a function of the charging voltage V and to determine the time dependent, relative intensity of the emitted light, three sets of experiments were performed using optical diagnostics. Firstly, streak photography was used to obtain information about the turn-on time of the source as a function of charging voltage, V . Secondly, high-speed, high-resolution time resolved photography was employed to determine the time dependent relative intensity characteristics of the source. Lastly, the optical decay times of the spark gaps were measured in two spectral ranges. For each of the optical investigations performed, the experimental set-up was the same with the exception of the camera. The ultraviolet source was positioned so that it 31 directly illuminated the lens of the camera. Delay generators were incorporated into the trigger circuitry of the experiment to allow for variation in the timing of the event with respect to the camera shutter. The first two sets of experiments were performed in open air. A TRW streak camera was used to investigate the turn-on characteristics of the source. A streak photograph of the ultraviolet source with a charging voltage of 33 kV is shown in Fig. 3. The streak photograph illustrates that The the linear array of spark gaps break down serially. dim area in the center region of the streak photograph is due to an imperfection in the photocathode of the streak camera and not due to the source itself. The turn-on time of the source for a given charging voltage can be determined by considering the writing rate or streak rate of the camera and the distance between the beginning of the streaks made by the first and the last spark gaps of the source. The picture shown in Fig. 3. for example, shows that the turn-on time of the source with a charging voltage of 33 kV was 15.5 ns. Similar streak photographs were taken of the source with charging voltages varying from 23-35 kV. Figure 4 shows the dependence of the turn-on time of the ultraviolet source on the charging voltage of the pulse forming network. 32 n20 CE LU CO - 10 DC < ^1 i i CL CO 0 10 20 30 TIME / (ns) Fig. 3. Streak photography of the ultraviolet source with a charging voltage of 33 kV (in air) 33 ? 20 LU o Q < UJ 15 >< < J L 25 30 CHARGING VOLTAGE / (kV) 35 Fig. 4. Turn-on time (i.e. propagation of arc development) of the ultraviolet source as a function of charging voltage. 35 c o 1 3 A c u 4^ o 0) 4-> 4-) 0) -^ o & E C O O Xi u bO N (0 4>> o u hO O 4^ u O ^ A U (0 o a a -o ^ ^ 0 > ^ 4 O 1 d C 0) VMM' w u u 0) u 3 u 3 w a; u E F^ o c o U hJ t- 0) w 0) u ^H \n iH f^ o c d ^> > b O e ^.^ d u tm> P ^ ua U. 3 a c d 34 The time resolved photographs of the ultraviolet source were made using a high-speed, high-resolution image converter camera. The camera consists of a proximity focusing diode manufactured by ITT with a Tektronix roll film back mounted on the rear of the diode . A krytron switched transmission line pulser was used to deliver a 10 kV trapezoidal pulse with a rise and fall times of approximately 5 ns and a half width of approximately 10 ns to the diode. of 10 ns. Consequently, the camera had a shutter speed The diode had a high resolution of 45 Ip/mm. Figure 5 contains two photographs which were taken at different times. Figure 5a was taken during the turn-on phase and illustrates again that the spark gaps would break down in succession. Figure 5b. taken with reduced sensitivity, shows the ultraviolet source with all 20 spark gaps emitting light after breakdown. Figure 5a and 5b were Figure 6 depicts taken at 10 ns and 290 ns, respectively. the relative intensity versus time for the first spark gap in the ultraviolet source. The data for this figure were obtained by making photodensitometer plots and correcting them for the different aperture settings and film speeds. The total light emitted from the UV source was the superposition of the light emitted from the 20 individual 2 The image converter camera was designed and built by M. Michel. Technische Hochschule Darmstadt. FRG. 36 CO z z > LU ILLJ LU IT 10 100 1000 TIME / (ns) Fig. 6. Relative intensity versus time relationship for the first spark gap of the ultraviolet source with a charging voltage of 33 kV. 37 sparks, as shown in Figure 7. The linear section of the turn-on characteristic corresponds to the serial breakdown of the individual spark gaps and makes up for the most important part of the risetime (compare Fig. 3 ) . The minimum risetime achieved was 14 ns for the array of 20 sparks with a charging voltage of 35 kV. Considering that approximately 100 sparks in series are needed to match a 2.5 n line, one should observe a risetime of approximately 70 ns. It should be mentioned that this is the risetime of An the source function, S. for the discharge as a whole. individual volume element will experience a source function with a significantly shorter risetime since not all sparks contribute to the ionization in a given volume element. It is also important to note that the ultraviolet source continued to emit light for some time after the relatively short current pulse, which excited the source, had terminated. The experimental results shown in Figures 6 and 7 were taken from the spectral range of 225 to 665 nm (FWHM) corresponding to the sensitivity range of the image converter camera. The decay time (90% to 10%) was The breakdown of a spark gap approximately 600 ns. produces a hot plasma, with an increased free electron density which remains in the arc region until electrons are removed by recombination processes. Recombination dominates over attachment in arcs due to the high electron 38 0 10 20 30 40 50 50 250 500 750 1000 1250 1500 TIME / (ns) Fig. 7. Relative intensity versus time relationship for the whole ultraviolet source (20 gaps) with a charging voltage of 33 kV (visible spectral range). 39 density. Through recombination, highly excited atoms or Thus, until molecules are produced which decay optically. the free electrons are removed by recombination, light will be emitted from the plasma remaining in the arc region. Since the ionization only occurs with vacuum UV light, wavelength dependent measurements were also performed using a fast vacuum photodiode with different spectral filters: a near UV bandpass filter with 285 nm < X < 370 nm; and a visible bandpass filter with 400 nm < X < 650 nm. The decay times for the light emitted by the source in the two spectral ranges were measured in different gases. The results shown in Table 1 demonstrate that the decay time in the UV spectral range was shorter than in the visible range. The optical decay time for the ionizing UV radiation below 180 nm [50] is probably even shorter. Discharge Investigations Using the UV Source The ultraviolet source, installed in the discharge chamber, was used to investigate three types of diffuse discharges. One set of experiments demonstrated the ultraviolet source as a preionizer. allowing the system to operate in a self-sustained discharge mode [56]. Another set of experiments illustrated that the ultraviolet source could be used as a sustainment device for diffuse discharges containing attachers. This enabled the system 40 Table 1 Decay Time Ratios for the Near UV and Visible Emissions from the UV Source Lab Air He Ar T UV /T . VIS 0.86 0.92 0.88 41 to be operated as a ultraviolet sustained, diffuse discharge switch. It should be noted here again that a switch which utilizes a spark gap as a sustainment mechanism may have limitations with respect to short turn-off times since the ultraviolet source continues to emit ultraviolet radiation after the excitation current has been removed [35]. The third set of experiments demonstrated the generation of UV initiated discharges. Externally sustained discharges were used to investigate the influence of specific attachers on the discharge characteristics. Figure 8 shows the steady state J versus E/N characteristic for UV sustained discharges in 10% C^F^, 90% Ar, and approximately 350 ppm N.N 2. b Dimethylaniline, CgHgN-(CH3)2. The UV sustained results are shown with results of E-beam sustained discharges also in 10% C^F^ and 90% Ar. 2 o The E-beam results were generated by the E-beam Sustained Diffuse Discharge Opening Switch Experiment at Texas Tech University [27]. [28]. As pointed-out above, gas mixtures composed of argon with a small percentage (0.5% to 20%) of flourinated alkanes (such as CF.. C^Fg, and C^Fg) have been proposed by Christophorou et al. [40], [41] to exhibit the attachment and mobility dependences with.respect to E/N for opening switches outlined previously. Dimethylaniline has a low ionization potential and is commonly used in UV sustained and UV 42 10 I E O a E-BEAM SUSTAINED S 1.3*1020 cm'^ "* z a - UJ . 1 o z tu 3 U cr a: 01 UV SUSTAINED S - 1. O'lO^" c * " ' m'^ 10 15 20 25 30 35 40 REDUCED FIELD STRENGTH C E/N )/Td Fig. 8. Current density, J, versus reduced field strength. E/N, for an E-beam sustained discharge and for a UV sustained discharge, both in argon with an admixture of 10% O^Fg. 43 preionized discharges to increase the electron yield from the UV photons [50], [31]. The discharge characteristics show a region of negative differential conductivity (NDC). At low values of E/N (E/N < 2 Td) the discharges were recombination dominated and the current densities increased linearly with E/N. At about 3 Td the discharges became attachment dominated and J was forced to decrease as E/N increased. This variation of J with E/N matches the requirements of low attachment and high mobility at low E/N (high J) and high attachment and low mobility at high E/N (low J). One would expect the onset of the negative differential conductivity to be more pronounced if the source function decreases since the influence of recombination is reduced [21]. opposite is found. According to Fig. 8 the A significant difference between the two ionization sources is that the average energy of the electrons produced by the UV source is very low (<<1 eV) while the electrons produced by the E-beam have an average energy in the order of several eV. This effect in combination with attaching gases may have a significant effect on the electron energy distribution function and consequently on the onset of attachment with increasing E/N. Also the addition of dimethylaniline may contribute to a change of the electron energy distribution. 44 The third set of discharge experiments investigated UV initiated self-sustained discharges for laser applications. In this UV intiation mode, a DC voltage below self-breakdown was applied to the discharge gap. The ultraviolet source was then triggered and a diffuse discharge formed in the gap. In this mode, the externally initiated discharges acted as both the switch for the discharges' pulse forming circuit and as the active media of the laser. Several methods have already been used to initiate discharges, such as overvolting with a high voltage trigger pulse [57]. [58] or UV or X-ray photoionization [59], [60]. Using this UV initiating technique, discharges containing C0 were generated that were matched to the impedance of the pulse forming line which drove the discharges. As shown in Fig. 2, the discharge gap was This line was directly connected to the transmission line. DC charged to a value V which was below the self-breakdown ^ o voltage and above the operating voltage or "glow voltage," VJ, of the self-sustained discharges. d matched operation, the condition V fulfilled. Figure 9 shows the typical time dependence of current and voltage for matched operation. was approximately 20 ns. The current risetime For impedance = 2V, must be With preionized discharges. 45 LU DC CC O O < LU TIME(100ns/div) Fig. 9. Time dependence of current and voltage for a UV initiated discharge in a C0 laser mixture ( C 0 2 : N 2 rC0:N:He=l:16:23 at 300 torr) 46 triggered by a laser triggered spark gap, and utilizing the same experimental device, risetimes down to approximately 10 ns were achieved. It can be concluded, therefore, that the risetime of the ultraviolet source, caused by the sequential breakdown of the spark array, imposed a limitation for the lowest possible risetime of UV initiated discharges using spark arrays. CHAPTER IV PHOTODETACHMENT The presence of attachers in opening switch discharges generally increases the power losses in the switches. Additionally, the closure phase, where the gases change from non-conducting to conducting, can be inhibited by the presence of attachers. One method of overcoming these effects of attachment during specific switch periods is to use photodetachment: hi) + B " - B + e. > Photodetachment is a non-resonant process similar to photoionization but typically requires much lower photon energies than photoionization. In this chapter, theoretical considerations for improved switching by diffuse discharges utilizing photodetachment as an additional control mechanism are presented. Also presented are the results from three separate experimental devices, built and operated to demonstrate the feasibility of photodetachment as a control mechanism. The experiments described in this chapter were also performed with the assistance of several people. The experiments in the flowing afterglow apparatus were performed by Professors P. Frazier Williams and Gerhard 47 48 Schaefer. Leo E. Thurmond conducted the investigations using the wall stabilized discharge apparatus. Circuit Considerations for Opening Swi tches Attachers with high attachment rates at high values of E/N and low attachment rates at low values of E/N will allow fast opening when the external ionization source is turned off and will have low losses in the conduction phase. Such attachers, however, increase the closing time and increase the switch losses during closure if switch closure is performed in a high impedance system. This effect has been predicted in calculations on discharges in N t containing N0 [21] and demonstrated in experiments on r discharges in argon containing C^Fg [28]. The effect of a high impedance circuit on a diffuse discharge can be understood by examining the steady state J versus E/N characteristics for an electron beam sustained discharge, as shown in Fig. 10 [28]. These characteristics were generated in Ar with 2% C2Fg using two different circuit impedances. The E-beam had a risetime of approximately 10 ns and a nearly flat maximum over a pulse length of 400 ns. The curve for the low impedance system For the low is the true discharge J-E/N characteristics. impedance case the characteristics exhibit a strong negative differential conductivity (NDC). Shown in Fig. 10 49 3.0 (M I E U I T 1 1 I I I I 2 OHH LOAD 23 aH LOAD i a z UJ Q Z UJ a: a: U 0.5 % o A g ^ 0.0-t REDUCED FIELD STRENGTH ' 10 15 C E/N )/Td Fig. 10. Current density, J, versus reduced field strength. E/N, for an E-beam sustained discharge in argon with an admixture of 2% C2Fg. obtained at different times after E-beam initiation and with different load lines [28]. 50 are the load lines for the low (2 Q) and high (100 Q) impedance circuits. Of course, the discharge J-E/N must Initially, when the move along the appropriate load line. discharge is not conducting and the switch is open, the discharge will be at the point where the load line crosses the E/N axis. As the E-beam is turned-on. the discharge will move up the line until it reaches the first point where the load line intersects the J-E/N characteristics. For the low impedance case, the discharge will operate at the peak current density shown in Fig. 10. However, for the 100 n load line plotted in Fig. 10. the discharge will operate at a relatively high E/N and a relatively low J; the discharge never reaches the region of maximum current densi ty. Figure 10 also illustrates another feature of the high impedance circuit interaction with the discharges. Assume that the initial E/N (i.e.. before the E-beam is turned on) is in a region where the discharge can reach the characteristics of high current density: say an initial E/N < 7 Td. At the time the E-beam is initiated, the discharge is already in a region of high attachment and low mobility. The result is that the closing process of the discharge is obstructed and the time necessary to reach the characteristics is dramatically increased (i.e.. the current risetime is increased). For the 2 Q system the 51 steady state values were obtained in less than 100 ns. For the 100 n system the J-E/N values are plotted for 75 ns and for 350 ns after E-beam initiation. For the 100 Q case, the discharges never reached the steady state characteristics. It has been proposed that photodetachment can be used to overcome these losses during closure [20], [21]. influence of photodetachment will mainly affect the discharge characteristics at an intermediate E/N. At low The values of E/N attachment will not dominate the discharge if the attacher used has the properties mentioned above. high values of E/Nabove self-breakdownionization At through discharge electrons will dominate [61]. Theoretical Considerations Before examining photodetachment as a means of overcoming circuit interactions, an analysis of the species' balance equations is in order. A term must be placed within Eqs. (l)-(4) to account for increased electron production through photodetachment. given by: R = CT j^cN_q. where R is the rate of electron production through photodetachment. a , is the photodetachment cross-section (14) This rate is for the negative ion present which is a function of photon 52 energy, c is the speed of light. N_ is the negative ion density, and q is the density of photons. With: q = hue' Eq. (14) can be rearranged to read IN CT , R = : p^ (15) hi) where I is the light source (i.e., laser) intensity, and hu is the photon energy. dn ^ dt Now Eqs. (2)-(4) may be changed to: = S + k, Nn ion e IN CT , k rec N^ne + att att e (16) + dN hi) - Ph dT = S + k . ion Nn e - k N.n rec + e - k ionrec N^N - . + IN CT , (17) ^ ' (18) dN_ TT" = k ^^N ^^n - k. N N dt att att e ionrec + - J-^ hu It is important to note that the negative ion density will vary with E/N since the attachment coefficient varies with E/N. Therefore, the photodetachment term in the above Now for an equations will vary, indirectly, with E/N. attachment dominated, externally sustained discharge Eqs. (16)-(18) are given as: dn IN_CT , T T = S - ^att\tt-e ^ -^n^ (19) 53 dN^ ;r-^ = S - k, N N . dt ionrec + dN_ dT- = ^^tt^^tt^^ - k, N N at att att e ionrec + IN_CT , r-^. hu (20) ^ ^ (21) ^ Photodetachment can be used to overcome the high impedance circuit problems by forcing an increased generation of electrons, over that amount normally produced by the external source. If in Eq. (19) the photodetachment term is made large compared to the attachment term (IN_CTpn ,/hu > k^..N^.^n e ) then the NDC observed in the J-E/N attauL characteristics will be reduced or eliminated. Furthermore, since the generation rate of electrons is increased, the closure time for the switch is reduced. The light source would be active only during the closure phase of the discharge. Due to the small values of the photodetachment cross-section for most ions, intense light sources (i.e.. lasers) would be required. Once the discharge has reached the region of the J-E/N characteristics of peak current, the laser would not be needed. In this way. only a short pulse of intense light but of small energy would be necessary Schaefer et al. [21] did a series of calculations that modeled E-beam sustained discharges in N2 with small percentages of N^O. The discharges were modeled in an As an additional control inductive energy storage circuit. 54 mechanism, the authors used photodetachment of 0~. negative oxygen ion. 0~, is the major byproduct of dissociative attachment in N2O. Table 2 Figures 11 through 13 and The show the results of these calculations. In Fig. 11 the theoretical discharge is sustained by an electron beam and illuminated by a laser for 100 ns. The discharge operates at a high E/N and low J for the no laser case. Additionally, the power losses in the When the discharge is illuminated discharge are very high. by the laser, E/N decreases, J increases and the power losses drop by nearly an order of magnitude. Additional calculations were performed for short laser pulses occurring around the first 20 ns after E-beam initiation. Figure 12 and Table 2 show the time dependence of E/N and the energy density losses when the discharge is illuminated by laser pulses of varying widths but always beginning at the same time as E-beam initiation. By illuminating the discharge with a short {Z 20 ns) intense (107 W/cm 2 ) laser pulse, the discharge, at closure, moves through the high E/N range and settles to a steady state operating point where E/N is low. The short laser pulse achieved the same results as in the case of the long laser pulse; however, the total laser energy is reduced. Of course, during the first few nanoseconds after the E-beam is initiated, the density of negative ions is very 55 200 150 l?ioo|. "I NO LASER so 100 150 2 250 3O0 100 400 30010 ixr Vi (/) 200 250 300 O ^ 200 IP ' \ \10*W/cm||{r. N 100^ .;/,, \ J:;.. 50 100 150 200 TIME/ns 250 300 Fig. 11. Time dependence of E/N (top), current density (middle), and power density loss (bottom) of an E-beam sustained, laser photodetachment assisted discharge in 1 atm N^ with an N^O fraction of 1%. The system impedance is Z=20 Q. The E-beam and laser are on for 0<t<100 ns. The variable parameter is the laser power density [21]. 56 Table 2 Energy Density Loss Per Switch Cycle for the Curves Shown in Fig. 12 and 13 [21] # in Fig. 12 Laser On From-To (ns) 1 2 3 4 5 6 7 8 # in Fig. 13 no laser 0-5 0-7.5 0-10 0-12.5 0-15 0-20 0-100 Laser On From-To (ns) 1 2 3 4 5 6 7 no laser 0-10 2.5-12.5 5-15 7.5-17.5 10-20 0-100 Energy Density Loss In mj/cm 22.0 28.9 21.65 20.7 13.2 10.5 9.4 9.0 Energy Density Loss 1/ In mJ/cm 3 22.0 20 7 13 2 10.8 10.0 10 1 9 0 57 Td 50 TiME/ns 75 Fig. 12. Time dependence of E/N for an E-beam sustained, laser photodetachment assisted discharge in 1 atm N^ with an NjO fraction of 1%. The system impedance is Z=20 fl. The laser power 7 2 density is 10 W/cm . The E-beam is on for O^t^lOO ns. The variable parameter is the laser pulse length (see Table 2) [21]. 58 EIN Td Fig. 13. Time dependence of E/N for an E-beam sustained, laser photodetachment assisted discharge in 1 atm N^ with an 2 "2 fraction of 1%. The system The laser power impedance is Z=:20 Q. and the laser density is 10^ W/cm ns. The E-beam is pulse length is 10 on for 0^t<100 ns. The variable parameter is the delay of the laser pulse with respect to the E-beam (see Table 2) [21]. 59 low and the effect of the laser is correspondingly small. Figure 13 and Table 2 show the results where the laser pulse length is kept constant (10 ns) but the time of laser initiation is varied with respect to time of E-beam initiation. Again, by illuminating the discharge during closure with a short intense laser pulse but, in this case, appropriately timed, the discharge will reach low E/N and high J. Photodetachment can also be used as a spatial control mechanism. As with UV ionization, electron generation can be increased in regions illuminated by a light source. Such control appears particularly attractive for maintaining the stability of diffuse discharges. All diffuse discharges will, after some time, undergo a glow-to-arc transition [62]. [63]. Small spatial perturbations in the electron density due to thermal perturbations or electrode protrusions will cause non-isotropic currents in a discharge. Higher current densities in a volume element of the discharge due to electron density perturbation will heat that volume element. Increased heating leads to increased ionization, and a runaway effect develops for the current in the volume element. Eventually, all of the current of the discharge is carried through the volume element by a constricted or 60 filamentary discharge; the volume discharge is replaced by an arc. Photodetachment can increase the overall electron density in the discharge if the entire discharge region is illuminated. The ratio of the electron density in the perturbed region to the background density is thus lower than the ratio for a discharge where photodetachment does not occur. Consequently, the ratio of the current through the perturbed element to the overall discharge is lower for the illuminated discharge case than for the not illuminated discharge case. The result appears as a longer time for the glow-to arc transition (i.e., longer conduction times). As an additional bonus for this scheme, negative ions can act as a reservoir of electrons. Cooper et al. [56] showed, through very general calculations, that the negative ion density in a UV sustained discharge could increase by as much as a factor of 10 over the electron density. This reservoir could then be utilized for enhancing the stability of diffuse discharges. Attachment and Photodetachment In Oxvgen and Oxvgen Bearing Molecules Possibly, the oxygen anion, 0, is the ion for which photodetachment has been most studied [64]-[68]. photodetachment cross-sections for 0 Fig. 14 [64]-[66]. The and O2 are shown in has a The cross-section for 0 61 i 10 z o p u Ul M (A A O s Z w 1 Z u u < o o o z U 2.0 2.5 3.0 PHOTON ENERGY/*V Fig. 14. Photodetachment and 0^ ( cross-section for 0 Smith, et al. [54]; o Branscomb. et al. [55]: Burch. et al. [56]). 62 threshold of about 1.47 eV photon energy and reaches a plateau of 6.3X10 18 cm 2 at around 2 eV. As with most photodetachment cross-sections, the plateau is fairly broad and. unlike transition-induced optogalvanic effects, does not require a light source tuned to a specific and narrow wavelength in order to cause the effect. The negative ion of atomic oxygen can be formed from dissociative attachment collisions with many different oxygen bearing molecules. The formation of 0 reaction e + 0^ - 0 > from molecular oxygen. 0^, by the + 0 has been studied by several Other molecules that can be used investigators [69]-[72]. to form 0~ are N2O, SO2. CO2. NO, CO, and H2O [43]. Because of the fairly readily available sources of information on the creation and photodetachment of 0 , it was selected as the negative ion for studying photodetachment. Additionally, the photodetachment of 0 occurs at spectral wavelengths that are easily accessible with visible untuned lasers. Maximum power for dye lasers occurs at 590 nm which corresponds to 2.1 eV photons or well within the peak photodetachment cross-section of 0 . Oxygen was choosen as the attacher for these experiments, since the conditions for optimum 0 available in the literature. It should be noted, however, that 0^ is not an optimal attacher for opening switches. In the gas mixtures production were 63 containing CgFg the major negative ion is F~ [41]. Unfortunately, the photodetachment cross-section for F a high threshold (X 4.3 eV), requiring the use of light sources in the ultraviolet [73]-[75]. for the F for 0~. Table 3 lists the important reactions that can occur in oxygen discharges [20]. Not only is it possible for 0~ At low E/N The plateau values has cross-section are only slightly lower than those to form but also O^. 0^, and in some cases 0.. or low E/p (E/p i 2 Vcm torr ), and especially for higher pressures [69], the dominant attachment reaction is the three body attachment (reaction 2 ) . of E/p (E/p > 2 Vcm torr For the higher values ), reaction 1, dissociative At around E/p = 10 Vcm torr attachment, dominates [70]. the production of 0 E/p (E/p > 20 Vcm peaks [70]. torr At still higher values of ) ion molecule reaction 3 forces favored production of O2 [70]. Experiments Utilizing a Flowing Afterglow Apparatus The first experiments to investigate the use of photodetachment as a control mechanism were performed in the flowing afterglow of DC discharges [20]. is illustrated in Fig. 15. The apparatus This device allowed the measurement of the photodetachment effect in a region that was separate from the discharge that created the negative 64 Table 3 Important Processes in 0^ Discharges Reaction Attachment (1) (2) e + 2 - O" + 0 > e + 2O2 - 2 + 0 Ion-Molecule Reactions (3) O" + O2 - O2 + 0 > Charged Product 0~ O2 O2 (4) 0" + 2O2 -^03 + 02 Photodetachment ^3 (5) (6) (7) 0" + hu - 0 + e > O2 + hu - O2 + e > O3 + hu - O3 + e > (hu > 1.47 eV) (hu > 0.434 eV) (hu > 2.103 eV) Photodissociation e e e (8) (9) O3 + hu - 0" + O2 > O3 + hu - O2 + 0 > (hu > 1.7) (hu > 2.7) 0" Og 65 LASER f I . . . .1 . ' ^ . .. . . . , . . FLOW I '. .. I . .* 'J ..! . I I / . . . . . . . .. ' . I.' '....: .-v.: i Fig. 15. Diagram of apparatus for photodetachment experiments in the flowing afterglow [20]. 66 ions. For most of the experiments, a self-sustained, low pressure, DC glow discharge was established between grids 1 and 2. The region between grids 2 and 3 was either field free or a small bias voltage was applied to prevent free electrons from entering the region of photodetachment. Negative ions produced either in the discharge or in the drift region between grids 2 and 3 were carried by the gas flow into the region between 3 and 4. The gas was illuminated by a dye laser between grids 3 and 4. Electrons given off by photodetachment were collected by a small bias voltage (5-100 V) between 3 and 4. The photodetachment was measured by monitoring the voltage across a current viewing resistor, Rp. The dye laser was pumped by a frequency-doubled Nd:YAG laser and could produce 560 nm, 10 mJ, 7 ns pulses. The results of the experiments performed with this device Indicated that the various negative ions of oxygen (0~, 0~ and 0~) were produced in the ranges of E/p Experiments were also performed to described above. determine the laser energy flux necessary to photodetach a fair percentage of the negative ions. experiments are shown in Fig. 16. The results of these The solid line shows the calculated results, assuming a photodetachment cross-section of 6.5X10 -18 , and the circles represent the 2 measured results. At an energy flux of ~ 150 mJ/cm all of 67 200 LASZR PULSE ENERGY DENSITY i.2 nJ / cxD Fig. 16. Relative magnitude of optogalvanic effect as a function of laser pulse energy in the flowing afterglow [20]. 68 the negative ions were photodetached or the effect 2 saturated. A light energy flux of only 35 mJ/cm was required to photodetach half of the ions. negative The results from this set of data encouraged further work. Experiments Utilizing a Wall Stabilized Glow Discharge Apparatus A second set of experiments were performed in self-sustained wall stabilized glow discharges in oxygen [76]. The apparatus for these experiments, shown in Fig. 17, consisted of a quartz tube terminated at the ends by two sets of hollow electrodes. The outer two electrodes maintained the discharges, and the inner two electrodes allowed the voltage measurements of the positive column region of the discharges. A laser beam was directed along the length of the quartz tube and through the hollow regions of the electrodes. A flashlamp pumped dye laser was used with the beam focused to produce a peak intensity o of 6 MW/cm in the discharge tube. The discharge tube was cooled by an outer jacket of flowing transformer oil. Experiments were performed in which the photodetachment effect was measured for the whole discharge region and for the positive column alone. The experiments were conducted with both flowing and non-flowing gases. Since the flowing gas removed 0 ions before the photodetachment effect could be measured, the 69 PRESSURE METER COOLANT GAS AS FLOW LOW VACUUM FLOW COOLANT / i -F= t AUXIUARY ELECTRODE ( GAS FLOW LASER -Q rr-O- fi ==cnf^ MAIN ELECTRODE COOUNG JACKET DISCHARGE I MAIN ELECTRODE F i g . 17. Diagram of a p p a r a t u s used for photodetachment experiments in the wall s t a b i l i z e d glow d i s c h a r g e . 70 flowing gas results showed a smaller effect than the non-flowing gas results. Figure 18 shows the change of the total discharge voltages versus discharge DC currents for different gas pressures. The largest effect occurred at This 12 torr pressure and 2 mA discharge current. discharge condition corresponds to a bistable region in the discharge voltage and current (VI) characteristics. bistable region is caused by multistep ionization proceeding through long lived metastables in 0^ [77]. Photodetachment allowed the discharges to make the transition from the bistable region of high voltage to the region of low voltage. Other conditions shown in Fig. 18 The exhibit a smaller effects. Figure 19 shows photodetachment results for the positive column only. The strongest effect ocurred for 18 torr and 2.4 mA, and may again be related to the bistable properties of oxygen discharges. The change of the discharge voltage under these conditions was 37% of the DC voltage which corresponds to a factor of 1.6 decrease in the discharge resistance. Experiments Utilizing the UV Sustained/Initiated Discharge Apparatus The last set of experiments investigating photodetachment were performed in the apparatus described in Chapter III for producing UV sustained or UV initiated 71 10 9 > T T T 1 r T T e 7 6 5 4 3 2 1 0 > 4J O 9 (4<4- liJ O c o > 4 torr O D) O 4^ Q. o J Discharge Currant I/CmA) Fig. 18. Optogalvanic effect (as ratio of peak optogalvanic voltage change to DC voltage) as a function of discharge current in the entire ' discharge. 72 40 T I 1 1 1 1 1 T ^^ >^ > <l > U 01 <4. 30 - QUJ O 20 c o > *^ Q. O o o 10 Discharge Current I/CmA) Fig 19. Optogalvanic effect (as ratio of peak optogalvanic voltage change to DC voltage) as a function of discharge current in the positive column. 73 diffuse discharges. (See Fig. 2.) With this device, photodetachment could be investigated in discharges similar to those in diffuse discharge opening switches. spacing was adjusted to 3.5 mm. The gap The same laser used for the wall stabilized glow discharge experiments was also used in these experiments. The laser produced pulses of approximately 1000 ns length and nearly constant power of 1 MW over a central plateau of 400 ns and at about 590 nm. The laser beam was focused so that the maximum discharge volume with the maximum laser power density was illuminated. The laser Intensity in the chamber was 5 approximately 8X10 W/cm 2 for most of the results reported. The influence of photodetachment was recorded by the voltage and current probes shown in Fig. 2. The measurements reported here were performed in mixtures of argon and nitrogen with admixtures of oxygen as the attacher. Argon, as discussed above, has been proposed as a buffer gas for diffuse discharge opening switches and was used as a buffer in these experiments to more closely simulate opening switch discharges. The mixture of argon and nitrogen optimized the ionization performance of the UV source. Nitrogen is known to increase the UV yield of spark sources [78] while argon allows increased penetration depth of the ionizing radiation. The ionization efficiency 74 was further enhanced by using N.N dimethylaniline as a low ionization-potential additive [50]. The first experiments were conducted to evaluate the influence of attachment on the steady state J-E/N characteristics of the oxygen bearing discharges. For these experiments the impedance of the pulse forming line that drove the discharges was kept small compared to the Impedances of the discharges. the laser was not used. For this set of experiments Figure 20 shows the J-E/N characteristics for gases with varying concentrations of Op. The mixtures used with higher 0^ concentrations generated regions of negative differential conductivity in intermediate E/N ranges of the characteristics. As discussed above, this effect is the consequence of an attachment coefficient that strongly increases with E/N. At high values of E/N the currents increased drastically. In this E/N range internal ionization through discharge electrons became significant and, hence, the discharges became self-sustained. Figure 21 shows typical J-E/N characteristics for discharges exposed to the the laser and not exposed to the laser. No influence of the laser on the discharge characteristics was observed in the low E/N range where no dissociative attachment occurs. The influence of the laser starts in the range where the current density is at a 75 1.000 (NJ I E U / OXi e 2.6X1 a 7. gxi 13.2Xi A 0. 100 0) z LU Q 0.010 UJ Q: u 0 . QOl I 1 1 1 I I I 1 I I I I 0 10 20 30 40 50 60 REDUCED FIELD STRENGTH (E/N)/CTd3 Fig. 20. VI - characteristics of UV sustained and UV initiated discharges for several concentrations of O2 with 2.6% No and balance of Ar to make 1 atm 76 1.000 1 1 i 1 1 1 1 1 T I E 0 WITHOUT LASERi e WITH LASERi a P 4" 0. 100 - 1> / 3 LU Q 0.010 LXJ - a: on U 0.001 ' ' ' ' ' ' 10 20 30 40 50 REDUCED FIELD STRENGTH (E/N)/CTdl Fig. 21. VI - characteristics of UV sustained and UV initiated discharges in 7.9% 0. 2.6% N. and 89.5% Ar at 1 2* 2' atm without and with laser photodetachment. 77 minimum. In this E/N range attachment is strong and the The strongest changes of density of negative ions is high. the current density at constant values of E/N were observed in the transition regime to the self-sustained discharges. As internal ionization becomes strong, the generation of electrons increases. This increased electron density along with the high E/N causes a high density of negative ions which, in turn, gives a stronger photodetachment effect. The influence of photodetachment for varying concentrations of 0^ is shown in Fig. 22. No influence of the laser was observed for discharges that did not contain oxygen. The higher oxygen concentrations produced a more pronounced change in current density, and. again, the strongest influence was always at the transition from externally sustained to self-sustained. Figures 23 and 24 show the experimental results for varying amounts of nitrogen. The current densities of the externally sustained discharges increased significantly with increased N^ concentration. This effect is a consequence of the higher UV yield from the spark sources. The strongest NDC was found for an N2 concentration of 2.6%. This mixture also showed the strongest influence of This may be due to the fact that the attachment the laser. cross-section for dissociative attachment of 0^ has its maximum above the N^ cross-section for vibrational 78 10.00 r f 1 1 r t 1 V r f r - T I o -) < UJ 2.8Xi e 5.3X1 e 7. O X I W laSXi 13.2Xi A o < I . 00 - u > I(/) 2 UJ I a UJ Q: 0.10 - 0.01 1 ' ' 1 1000 2000 3000 4000 5000 6000 7000 CHARGING VOLTAGE V^/CV] Fig. 22. The influence of the laser on the current density (AJ/J^. where AJ = Jj^(with laser) - J (without laser)) versus charging voltage of o the PFL for various concentrations of O2 with 2.6% N2 and balance Ar to make 1 atm. 79 1.000 1 1 1 r- 1 T r 1 ' (\J I E U OXI O 2.6Xi a 5.3Xi 10.5Xi A 0. 100 --1 - z UJ a 0.010 UJ A ^ rn/o a: U _U_^.^ 0.001 ~ i 1 1 i_ l _ 1 1 10 20 30 40 50 REDUCED FIELD STRENGTH CE/N)/CTd3 Fig 23 VI - characteristics of UV sustained and UV initiated discharges for several concentrations of N2 with 5.3% Oo and balance of Ar to make 1 atm. 80 10.00 \ I 1 I I I r r I 1 ? r OXI O UJ u z < I 2.6Xi o 5.3Xi # 10.5X1 A 1.00 u z a UJ 0. 10 UJ (T u 0.01 1000 2000 3000 4000 5000 6000 7000 CHARGING VOLTAGE V^/CV] Fig. 24. The influence of the laser on the current density (AJ/J^, where AJ = J^(with laser) - J (without laser)) versus charging voltage of the PFL for various concentrations of N2 with 5.3% O2 and balance of Ar to make 1 atm. 81 excitation of N^According to the two electron group model developed by L. Vriens [79], the threshold of the lowest excitation cross-section within a gas mixture will rapidly deplete the tail of the electron energy distribution. In this case, the lowest excitation As the N^ concentration is cross-section is that for Np. increased, the high energy tail of the electron energy distribution represents increasingly smaller proportions of the total number of electrons. As a result, the number of attaching collisions decreases with increasing N^ concentrations. The influence of the laser power on the current density was also measured. Figure 25 shows the results of these measurements for the discharge conditions where the laser had its strongest effect (E/N = 53 Td in 13.2% O2. 2.6% N2. and 84.2% Ar at 1 Atm). The change in current density was approximately linear up to the maximum laser 5 2 intensity of 8X10 W/cm indicating that the For increased photodetachment effect was not saturating. laser powers, even stronger current changes could have been measured. Further experiments involved determining the effect of photodetachment on a high impedance circuit. A 13 kfl resistor was placed in series with the transmission line of the diffuse discharge circuit (as described in Chapter 82 J r- 1 1 T 1 1 1 I O -) \ -> o 4 o <J UJ (J) z < o o o 3 u 0) I o o z UJ Q UJ 2 o o 1 o - o O o 400 600 800 o nHI L. 1 200 LASER INTENSITY I/CkW cm~2] Fig. 25. Current density change (AJ = J (with laser) - J^(without laser)) for varying laser intensities in 13.2% O2. 2.6% No. and 84.2% Ar at 1 atm and an E/N = 53 Td. 83 III). The resistor lowered the resolution of the current measurements but improved the voltage probe resolution. The results, plotted as changes in discharge resistance, are shown in Fig. 26 for 1.3% and 2.6% 0^ concentrations, respectively. For low values of E/N, laser illumination of the discharges lowered the resistances of the discharges. These last results indicate that photodetachment may be able to enhance the closure of diffuse discharges in attachers having high attachment rates at high E/N and low attachment at low E/N. As a final demonstration of the utility of photodetachment, laser enhanced stability experiments were performed. Sudden voltage collapse and current changes in These the discharges indicated formation of arcs. experiments were conducted in a mixture of Ar with 5.3% 0^ and 340 ppm dimethylaniline at 1 atm. As in the previous experiments, the flashlamp pumped dye laser was used with a 5 2 peak intensity of approximately 8X10 W/cm . In five out of five cases where the discharges were not illuminated by the dye laser, arcs development occurred anywhere from 350 ns to 1 j s after discharge initiation. i However, for five out of five cases where the discharges were illuminated, no arcs were observed, and the discharge stability was enhanced. 84 o a: o: < Ul I- u z < z u lit u z < w V) UJ ct UJ C3 QC < z u (/) J ' I L_ 2000 3000 4000 5000 CHARGING VOLTAGE V - / C k V ] F i g . 26. Change in d i s c h a r g e r e s i s t a n c e s (AR/R = R (without laser) * > o o*^ ^ - R. (with laser)) versus PFL charging voltage for 13 kfl system impedance. Gas mixture contained 2.6% N^. varying concentrations of Op. and a balance of Ar to make 1 atm. CHAPTER V OPTICALLY ENHANCED ATTACHMENT Optically enhanced attachment provides a means of externally increasing the depletion rate of electrons in a diffuse discharge. The technique Involves increasing the attachment cross-section by vibrationally exciting attaching molecules. If this technique is employed, for an opening switch, at the moment of switch opening, the turn-off time could be significantly reduced while avoiding high attachment rates and power losses during the switch conduction phase. After a presentation of the physics of optically enhanced attachment, various methods of vibrationally exciting molecules will be described. Experiments utilizing two of these vibrational excitation methods in three separate experimental devices will be di scussed. The optically enhanced attachment results using the wall stabilized discharge device, described below, were obtained in collaboration with Courtney Holmberg and Richard A. Korzekwa. Courtney Holmberg also performed the experiments using IR excitation in the plasma injection sustained discharge apparatus. Professor Gerhard Schaefer conducted the photodissociation experiments in the device using discharges sustained by a hollow cathode device. 85 86 Dependence of the Attachment Cross-Section on Vibrational Excltation Dissociative attachment cross-sections are known to increase with increasing gas temperatures [43]. This effect is attributed to vibrational and rotational excitation of the attaching molecules. The attachment cross-section in turn increases with Increasing excitation. A qualitative understanding of this process can be made by examining the potential energy curves, as shown in Fig. 27. As described Chapter II, dissociative attachment actually proceeds in two steps. The first step Involves the electron capture and the formation of a temporary negative molecular ion (e + AB - AB > ). This process can be represented by an electron capture cross-section, a cap . and * ^ ^ * ^ is shown in Fig. 27 as the vertical transitions from the lower AB curve to the upper AB curve. Molecular dissociation and the formation of a stable negative ion follows in the second step (AB represented by the AB - A + B ) and is > However, there curve in Fig. 27. is a certain probability that the intermediate negative molecule, AB , will return to the AB state (autodetach) - AB > + e). In and simply form an excited molecule (AB order for dissociation and the creation of the negative ion fragment to occur, the intermediate negative molecular ion must survive until it reaches the critical distance R where the AB and AB potential energy curves cross. The 87 AB* o > CC LLl Z UJ UJ O Q. 0 c INTERNUCLEAR DISTANCE. R Fig. 27. Potential energy versus internuclear distance curves illustrating resonant dissociative at tachment. 88 attachment cross-section, a, , can now be divided into two aa parts given by [80]-[83]: ""d^i^^^ where: T [6] 7^ . a = ^cap^P(-PC^^)' (22) P[e] = (23) The term exp(-p[fc]) is the survival probability of the _ ^ AB molecule surviving to R : T is the autodetachment c a _ ^ lifetime of the electron from AB and is treated as a constant: and T [fe] is the electron energy dependent stabilization time necessary for the nuclei to reach R . c _ ^ the crossing point of the AB and AB potential curves. Of course, this process is only possible for R < R . Most of the energy dependence in a, is associated with the energy dependence of a . which is determined ^^ " ^ cap from the probability distribution of the internuclear distance of the nuclei of AB. The energy for the peak in the capture cross-section is the energy for the transition from the AB to the AB the highest probability. curves at the value of R having For the ground vibrational state. the most probable R is at the lowest point on the AB curve. R . Since the probability of higher or lower values of R o decreases on either side of R , the capture cross-section decreases on either side of the peak energy for capture. 89 Therefore, attachment cross-sections exhibit increasing cross-sections up to a peak energy and then decreasing cross-sections with increasing energies. Vibrationally excited molecules show similar a cap energy dependence but the survival probability is strongly enhanced. Vibrationally excited molecules will have higher probabilities of existing at values of R greater than R ; o that is, a vibrationally excited molecule will have an extended nuclear wave function. If electron capture occurs _ ^ at an R > R , the time spent for AB to reach the o critical distance. R . (i.e.. T [e]) is reduced, and the o s survival probability is increased. For higher vibrational states, values of R of highest probabilities will be even closer to R . and the survival probability will be higher still. From Eq. (23). the survival probability is exponentially dependent on the time necessary for the AB molecule to reach R . c Therefore, the attachment cross-section is very sensitive to the vibrational quantum number of the attacher. Rotational excitation will also increase the attachment cross-section through centrifugal stretching of the attacher: however, attachment enhancement is much weaker for rotational excitation than for vibrational excitation [84], [85]. From Fig. 27 it is also obvious that as the attacher is vibrationally excited, the vertical distance from the AB 90 ^ curve to the AB curve is smaller. The energies for electron attachment are correspondingly smaller, and the threshold energy for the dissociative attachment cross-section occurs at a lower energy. Although the arguments presented above have been given for a diatomic molecule, the arguments hold true for polyatomic molecules as well. For polyatomic molecules, one need only consider the bond between the portions of the molecule that will later form the dissociated fragments. Under equilibrium, the vibrational state of a molecule is coupled to the temperature of the gas that the molecule is in. Thus, the attachment cross-section is correspondingly coupled to the gas temperature. The temperature dependence of attachment cross-sections have been measured for a variety of compounds: Op [71]; COp [71], [86]: NgO [87]; SFg, SF^ [88]. [89]; H2. HCl. HF [84], [85]. Fite and Brackmann also measured this effect in Op [90], and, later. O'Malley [91] was able to theoretically calculate their results with a fair degree of accuracy. Additionally, O'Malley showed, theoretically, that the attachment cross-section for Op would increase exponentially with increasing vibrational quantum number, u, and that the threshold energy for attachment would decrease with increasing u. 91 Measurement of attachment for specific vibrational states has not been done. However. Srivastava and Orient [92] have measured negative ion yields from COp after excitation of specific vibrational modes of these molecules by a monoenergetic electron beam. Enhanced negative ion production was found for E-beam energies equal to the energies corresponding to the peaks of the COp vibrational excitation cross-sections. Bardsley and Wadehra [93] were able to calculate the attachment cross-sections of HCl as a function of u from temperature dependent cross-sections measured by Allan and Wong [84]. (See Fig. 28.) Again these data show that, for the attachment, the magnitude of the cross-section increases and the threshold energy for attachment decreases for increasing vibrational excitation. Methods of Vibrational Excitation A wide variety of techniques have been employed to vibrationally excite molecules for laser applications. Several of these techniques can be used for enhanced attachment. Table 4 lists and catagorizes the various However, a technique vibrational excitation techniques. used for switching purposes must be fast enough to achieve current depletion on a time scale suitable for switch opening. Additionally, synchronization with switch opening is important. 92 0.5 1.0 ELECTRON ENERGY/eV Fig. 28. Dissociative attachment cross-section for HCl (e+HCl-H+Cl~) for different vibrational quantum numbers. I). [84] and [94]. 93 Table 4 Photoenhanced Electron Attachment (1) Vibrational Excitation (a) Single Photon Excitation plus Energy Transfer (b) Overtone and Combination Band Excitation (c) (2) (a) (b) (3) (a) (b) Multi-step Excitation Photodissociation Single Photon Multi-Photon Electronic Excitation Collisional Transfer of Electronic to Vibrational Energy Transitional or Radiative Transfer of Electronic to Vibrational Energy 94 Electron collisions in cold discharges, like diffuse discharges, cause the lowest vibrational state to be the only excited state of considerable density [93]. Therefore, for controlled attachment, the technique used for vibrational excitation must cause enhanced attachment for excitation states above the first excited state (u > 1). Perhaps optical pumping is the most direct approach to vibrationally exciting a molecule*. AB(i)=:0) + hu - AB(u>l) e + AB(i)>l) - A + B T This technique has been used by many investigators to pump infrared (IR) lasers. Tlee and Wittig [94], for example, have demonstrated laser action in CF., NOCl, CF.^1, and NH.^ pumped by a COp laser. Two of these gases. CF. and NH.^. are known attachers, and CF. has been proposed as an attacher for opening switches [40], [41]. In 1979, Chen and Chantry [95] reported the first experiments in which the attachment rate was increased by laser excitation. Their experiments showed a significant Increase in the dissociative attachment branch of SFg: e + SFg - SF~ + F. > 95 Since their results were dependent on the S isotope used, the effect was shown not to be due to general gas heating by the laser. Sulfur hexafluoride has a high attachment rate at low values of E/N and is, therefore, not a good candidate for an opening switch attacher. To analyze the effect of optical pumping of the attacher on the kinetics of a diffuse discharge, the species balance. Eqs. (2)-(4). must be modified. The enhanced attachment rate must be accounted for. and the rate of attacher excitation must be Included. As before. the dissociative attachment coefficient is E/N dependent. but now ka t t is also v dependent. wr1tten: dn dt = S + k , Nn - k N.n ion e rec + e Equations (2)-(4) are - I KtttE/"-"]"attf">e)dN T-^ = S + k. Nn e - k rec N^n e - k, dt ion + ionrec N^N - . + dN NV r 1 = > k [E/N,ij]N ^^[u]n - k. N^N , dt Z I att"- att*- - ej " ' ionrec + I) ^ (24) (25) ^ ' (26) ^ ' where the summations, 2 , are over all the vibrational quantum numbers, u. for the attacher, and ^ [u] is the The density of the attaching molecules in the v state. excitation rate for a specific state of the attacher can be written as: 96 dI = Z^,KibtE/N.i>-.]N^^^[u-]n INatt*- - vlbex*^Av'la * ,, [i>*-u] * h(i)-D*) ) - katt*^.[E/N,u]N^^ J u ]" e - '^att'^"^ . ^^^ . - att*- -n " TqL J, \V\ ^(27) where u'and v are the initial and final vibrational states of the attacher. respectively. 2 . is the summation over all of the initial states, k .,[E/N,u'-u] is the vib* coefficient for vibrational excitation from u' to v by electron collision, a . , [u*-i>] is the vibrational vlbex " excitation cross-section from v' to v by photon interaction with N att ^^[u*], and Tq [ U ] is the mean time for relaxation out of the V state. The first term of Eq. (27) accounts for electron collisional excitation of the attachers from state I)* to V. Notice that the collisional vibrational ,[E/N,u*-u]. is E/N dependent. excitation coefficient, k The second term is the rate of optical excitation from v' to V via a radiation field of Intensity I. The optical excitation term was derived in the same way as the photodetachment term in Chapter IV. Eq. (15). Since the attachers may already be in various vibrationally excited states, the first two terms must be summed over all possible initial states. Loss of attachers in the uth The state due to attachment is given by the fourth term. 97 last term represents the relaxation out of the vibrational state, u, due to collisional quenching, vibrationalvibrational energy exchanges, vibrational-translational exchanges, and possibly other mechanisms. Of course, Eqs. (24)-(27) above, are strongly coupled and have non-constant coefficients which makes them very difficult to solve. However, these equations can be simplified by applying the appropriate assumptions, described in Chapter II, for external or self-sustained and recombination or attachment dominated discharges. Additionally, as pointed-out before, only the first vibrational state is excited by electron collisions to any appreciable density. Therefore, for the electron collisional vibrational generation term in Eq. (27). one need only consider the transition from u*=0 to u=l. Also, the optical pumping term in Eq. (27) need only apply to the initial states u*=0.1. but, since only one optical transition from the ground state to the highest possible vibrational state would be used for enhanced attachment, only the u'=0 to one particular higher v need be considered. Equation (27) can now be re-written as: TT dt dNa t,t ,*[- u ]-* = INa t,t ,*[- 0 ]-* Vlbex*-[ 0 - u -* a ., ] rhu , n?/M TM r n " * ^ [E/N,u]N ..^[u]n att*-" a t t - - < "atti:"J (28) 98 for all u>l and for i = l the term: ) k^., [E/N,0-l]N^^JO]n e viD^ - att " * must be added to Eq. (28). Although it may be difficult to calculate the exact expression for the attachment coefficient k ^^[E/N,u] as a att"^ function of v, some generalizations may be made. Since the stabilization time for the attachment cross-section, given in Eq. (23), decreases with increasing vibrational quantum numbers, the attachment cross-section exponentially with v. Increases coefficient, Thus, the attachment given by <a, [v ,i)]v >, where v is the electron speed, da e e e will also increase exponetially. attachment Optically enhanced through vibrational excitation should have an extremely strong influence on the conductivity of diffuse discharges. Most compounds of Interest for optically enhanced attachment exhibit strong IR absorption peaks at bands easily accessible by common IR lasers (e.g., COp lasers). These absorption peaks are associated with excitation of vibrational modes of specific bonds and may be identified by the type bond and the functional groups connected by the bond. For example, the vinyl group: H>C=C<" 99 absorbs strongly around 10.6 yim. Fortunately, vibrational excitation of one bond of a molecule quickly equilibrates to all other bonds of the same molecule. This allows excitation of the bond bearing the electronegative component, which forms the negative fragment, by exciting the molecule through any convenient absorption peak. Additionally, by selecting the proper concentration of the attacher, nearly all of the optical power for enhanced attachment can be delivered to the attacher. Eisele recently attempted to utilize optical pumping of several attachers to measure this effect in E-beam sustained diffuse discharges [96]. A COp laser tuned to the lines for optimum absorption in NF.^, C^FQ , or COp was used. Additionally, HCl and HCN were pumped using light generated by optically mixing 1.05 ]im radiation from a Nd: YAG laser with 0.66 pun light from a dye laser. No enhanced attachment was observed in these experiments. There are several possible reasons for Eisele's results. Since Np was used as a buffer gas. Eisele suggested that collisional quenching of the excited states of the attacher by Np may have counteracted the laser induced excitation. Additionally, the discharge region illuminated in these experiments was only a small volume near the anode and accounted for "probably" [88] an area of about half of the total discharge current. The discharges 100 quite possibly increased the conductivity in the region not illuminated in order to compensate for the lower conductivity in the illuminated region. Eisele conducted his experiments at very low current densities and very low attacher concentrations. The data reported later in this chapter indicate that enhanced attachment, using optical pumping, performed best at high current densities and with attacher concentrations of 7.5-45%. Finally, the optical pumping was attempted in all but a few cases at steady state currents and several microseconds after discharge initiation, which resulted in equilibrium conditions in the discharges. As a result, optical pumping was forced to compete with collisional excitation of the attachers. Chantry [87] has pointed-out that enhanced attachment occurs not only for excited stretching modes but also for bending modes. Figure 29 Illustrates this mechanism for a The states X s hypothetical triatomic molecule. and A s are the ground states of the neutral and negative ion of the straight molecule, respectively: whereas, X, and A, are. correspondingly, the ground states of the neutral and negative ion of the bent molecule. The X and A curves are closer together for the bent molecule than for the straight molecule. Consequently, the threshold energy for attachment decreases and the survival probability 101 01 > O OC LU z < LU z UJ H o QL INTERNUCLEAR DISTANCE, R Fig. 29. Potential energy versus internuclear distance curves for a single bond of a triatomic molecule. Subscripts s and b refer to straight and bent species, respectively. 102 increases. This effect may account for the very low energies for attachment by NpO. Direct vibrational pumping of attachers can be done not only by single step excitation but also by multi-step processes [93]. Additionally, enhanced attachment excitation can be accomplished through overtone excitation or excitation of combination states [93]. Another vibrational excitation method Involves photodissociation or photoellmination of polyatomic molecules [97]-[100]. In this case, a vibrationally excited attaching fragment is formed from the dissociation of a weakly attaching parent molecule: ABC + hu - AB(u>l) + C e + AB(u>l) - A + B 7 Halogenated ethenes are the most promising candidates for the parent molecules since the attaching fragments formed after photodissociation. such as HF and HCl. are already strong attachers in their ground vibrational states. The dissociation reactions of fluorinated and chlorinated ethanes and ethenes to form hydrogen halides are only slightly endothermic. but the activation energies necessary for dissociation are relatively large [99]. [101]. The difference in the activation energies and the reaction energies must be absorbed by the molecular system. 103 Thus, large amounts of energy are made available for translational. rotational, and vibrational excitation of the attaching fragment [99]. Fortunately, the high activation energies of these reactions will prevent dissociation of the parent molecules by discharge electrons. Photodissociation may proceed either through Since the values for single step or multi-step processes. the activation energies for dissociation are large, the photon energies for dissociation are large. Thus, single step photodissociation requires light sources with wavelengths in the ultraviolet. Most organic compounds absorb UV in the 200 nm range by promoting a weakly held electron of a i bond to an r antlbonding orbital, ir -* ir [101]. For halogenated ethenes, the i orbital would be one of the two bonds r between the carbon atoms. molecule becomes unstable. In the excited state, the As a result, a hydrogen atom and a halogen atom on the same side of the molecule but bonded to opposite carbon atoms are brought together. The hydrogen and halogen atoms break from the molecule, forming the attacher, and the parent molecule transforms into an alkyne. The species rate equations for the production of vibrationally excited attachers through photodissociation are the same as Eqs. (24)-(26). Of course, the generation 104 of the attachers to each vibrational state is different than before. This generation may be represented by rate equations that describe the reactions that produce the attachers in the specific excited states. are given by: These equations dr IJShdt"]%ar]in7^ k,[E/N.]N^,,[]n^- -|ifj3-. (29) wh ere, as before, N ^^[u] is the density of attachers in a tL state u, N is the density of the parent molecules, par " ^ ' ^ k V,JC^] is the rate coefficient for the production of attachers in state u from the parent molecule, and hu*' is the energy of the photons causing dissociation. Unfortunately, the rate coefficients, k I.JCI']. have not been measured for most molecules. However, Sirkin and Pimentel [98] have found that for the parent molecules of Interest for optically enhanced attachment (halogenated ethenes), the rate coefficients decrease monotonically with increasing vibrational quanta. As with IR pumping of molecules, photodissociation has been used to generate rotationally and vibrationally excited compounds for lasing media. Sirkin and Pimentel [98] used photolysis of CJH^F and C2H2F2 by a UV flashlamp to generate a significant fraction of HF molecules in 105 states u > 1. Quick and Wittig [99] photodissociated various fluorinated ethanes and ethenes by multi-step processes using a COp laser to obtain laser action from excited HF molecules. Very recently, Rossi, Helms, and Lorents [100] have demonstrated enhanced attachment through photodissociation of CgH^Cl and CgHF^. using an excimer laser. These authors estimated vibrational excitation of HCl from CpH^Cl to be u = 6 + 1 and vibrational excitation of HF corresponding to the thermal activation energy for the dissociation of CpHF.^. Their results showed a dramatic increase in the 4 attachment rate by a factor of > 2X10 at 0.1 Td in 4 mtorr 2 of CpH^Cl with 500 torr of He and a factor of 5X10 at 2.2 Td in the same gas. They also reported attachment rate changes of as much as three orders of magnitude in 100 mtorr of CpHF^ with 100 torr of He at low E/N and a factor of 40 at 9 Td. Diatomic molecules composed of the same nuclei (i.e.. Of,, Ff,. Clrt. Ip, etc.) have no dipole moment and, as a result, cannot be directly vibrationally excited by optical means. However, electronic excitation followed by radiative decay to highly excited vibrational states is one method of exciting such homonuclear diatomic molecules. This electronic excitation method is also potentially useful for excitation of all forms of attachers. Beterov 106 and Fateyev [102] demonstrated enhanced attachment of Ip through electronic excitation of that molecule. Radiation from a frequency doubled Nd:YAG laser (532 nm) induced electronic excitation from the Ip ground state to the B state, as shown in Fig. 30. Radiative transition from the B state to vibrationally excited states of the electronic ground state followed the laser induced excitation. Stokes flouresence confirmed the radiative decay of the upper electronic state. Their results indicated an increase in 3 the attachment cross-section of Ip by a factor of 10 to Electronic excitation can also be transformed into vibrational energy by collisions of electronically excited molecules with unexcited molecules. This technique has been used as a pumping mechanism for molecular lasers. Peterson, Wittig, and Leone [103] demonstrated laser action from COp pumped by excited bromine atoms. pumping involved a two step process: Brp + hu ^ Br + Br Br^ + CO2 - Br + C02(u>0) + AE. < - 1 The reaction for The energy of the Br atoms was 3685 cm and the vibrational levels of COp of comparable energy, the (101) and (021) states, were excited. Although this technique 107 STIMULATED RAMAN PROCESS > B*7r(co ^ flC 111 TENTIAL E z T TrpLuof^EscENCE DISSOCIATIVE ATTACHMENT LASER S320A \ O INTERNUCLEAR DISTANCE Fig. 30. Potential energy versus internuclear distance curves for I2 and I2 illustrating the method of electronically enhanced attachment for I2 [93]. 108 may work well for lasers, the molecule-molecule collision rate may be too slow for fast switching action. In order to determine the feasibility of optically enhanced attachment as a switch control mechanism, a series of small, low pressure experiments were conducted. Two sets of these experiments investigated enhanced attachment through direct pumping of attachers by a low power CW COp laser. The third set examined photolysis of halogenated ethenes by UV radiation generated from spark sources. Experiments Utilizing the Wall Stabilized Glow Discharge Apparatus The first set of experiments were conducted in the wall stabilized discharge aparatus used to investigate photodetachment. (See Fig. 17.) For these experiments, a low power (< 1 W) CW COp laser was used to pump the attacher, NH.^, directly at a wavelength of 10.6 iim. chopper wheel was used with the laser to study the transient effect of the optical process. Figure 31 shows the results of these experiments. peak increase in the discharge voltages as a function of the discharge DC currents are shown for different pressures. The measurements were performed with the The A current as the variable parameter since the discharge voltage was nearly constant. The value of E/N at the stability limit with low currents was approximately 50 Td 109 0.030 i-j -1 r 1 1 1 1 T - ;^ 0 > > 0.025 ~ \ \ \ ei 32. 2 t o r r i 29. 0 t o r r Al 26. 0 t o r r - S UJ 13 Z < 0.020 I z u UJ 13 < 0.015 - 1 1 a > UJ 13 0.010 - z u 1 1 0.005 - 1 < 1 1 0.000 150 1 1 1 1 200 250 300 350 0ISCHARGE CURRENT/C10"^A3 I n f l u e n c e of COp l a s e r on t h e Fig. 31 discharge voltages (AV/V , where AV = dischar V.(with laser) - V (without laser)) for va various pressures of 57% NH^ and 43% Ar 1 i 1 1 110 and decreased with increasing currents by less than 10% in the current range investigated. The change in the voltage increased with increasing pressures due to the increased partial pressures of the attacher. As the DC currents were varied, attachment had to compete with collisional vibrational excitation and detachment. At low currents, vibrational excitation and detachment were weak due to the low density of electrons. At higher currents, the higher electron density increased the collisional excitation and detachment rate which appeared as a smaller voltage change. 2 The low laser powers (< 1 W/cm ) insured that the optogalvanic effect was due to the vibrational pumping of the attachers and not due to general gas heating followed by Increased attachment rates. IR Pumping in Discharges Sustained by Plasma Injection A second set of experiments involving direct optical pumping of attachers by the COp laser described above utilized a plasma mixing device to sustain low current continuous discharges over a wide E/N range. An The Illustration of the apparatus is shown in Fig. 32. main discharge region where the optogalvanic effect was measured was composed of two parallel electrodes. 11.3 cm long. Brewster windows at either end of the device allowed The illumination of the discharge region by the COp laser. Ill GAS INLET FLOWING HOLLOW CATHODE t PUMP LASER BEAM Fig. 32. Illustration of apparatus used to study IR enhanced attachment in externally sustained DC discharges. 112 plasma mixing or hollow cathode device was embedded within the anode, behind a steel mesh. A pin anode surrounded by The end of a steel tube made-up the plasma mixing device. the tube was tapered to form the hollow cathode. Typically, helium was blown through the hollow cathode while argon flowed around the outside of the tube. A mixture of Ar and the attacher flowed through the main discharge region. To operate the apparatus, continuous discharges were established between the pin anode and the hollow cathode of the plasma mixing device at relatively high currents (0.5-2.5 mA). Gas flow would force the plasma generated by The these discharges into the main discharge region. helium plasma would mix with the gases in the main discharge region. Owing to its high density of metastable atoms, the plasma would further ionize the gases in the main discharge region. This arrangement allowed the discharges in the main region to be operated as externally sustained, continuous discharges over a wide E/N range (0-100 Td). Currents in these externally sustained discharges were typically on the order of microamperes. Figure 33 shows the current versus E/N for several gas mixtures investigated in these experiments. As in the experiments with the wall stabilized discharge apparatus, the laser was optically chopped and 113 16 14 < T 1 1 r T r a. :: u a 12 10 - z UJ tr 3 U 8 - u a 6 - UJ (3 < z u 4 2 - 100 REDUCED FIELD STRENGTH (E/N)/CTd3 F i g . 3 3 . Current v e r s u s E/N for s e v e r a l of the gas m i x t u r e s i n v e s t i g a t e d in the a p p a r a t u s of F i g . 32 (0 15% C2H3F. 5% He, 80% Ar. a t 21.8 t o r r ; ^ : 7.6% C2H3F. 5% He. 79.4% Ar. at 21.8 torr; A: 33% NH^, 6% He. 61% Ar. at 32 torr). 114 the transient response of the discharge voltages were measured as functions of the DC values of E/N. For comparison purposes, ammonia was the first attacher investigated. 34. The experimental results are shown in Fig. These curves indicate a weak response over a very wide Above 50 Td sharp increases E/N region, Z 15-50 Td. occurred at reduced field strength values which depend on the NH3 concentration. The threshold for internal ionization or self-sustainment corresponds to these values of E/N for strong response. The increased electron densities increased the total attachment rate, k ^ N ^^n . att att e Consequently, the ratio of the change in the discharge voltage (AV) to the discharge DC voltage (V ) increased. Further experiments involving direct pumping of attachers were performed in gas mixtures containing vinyl flouride, CpH^^F. This compound was choosen for its very Although its attachment strong absorption at 10.6 jim. properties are not known, the similar structure of CpH.^F to other attaching compounds and the presence of fluorine in the compound led to the belief that the molecule would be an attacher. Figures 35 and 36 show the results of the As with NH^. the response experiments with this compound. of CpHoF was present but weak at intermediate values of E/N z, 15-60 Td. At higher E/N the effect became stronger for Additionally, higher the same reasons given for NH.^. 115 15 0 1 ^ I" I I I I I I I I 1T 1 1I 1 I I 1r M N < % O > A <^ 0. 12 oi 45X NH3 33X NH3 I > 0. 10 u (3 Z < Al 24X NH3 z u 0.08 r UJ < -I a > UJ 0.06 (3 < Z 0.04 u S 0. 02 h 0.00 J I I I I I I 1 I Ll I I I 10 20 30 40 50 60 70 80 90 100 REDUCED FIELD STRENGTH (E/N)/CTdl I n f l u e n c e of COp l a s e r on t h e Fig. 34 discharge voltages (AV/V , where AV = dischar V.(with laser) - V (without laser)) for various concentrations of NH^ with 6% He and balance of Ar to make 32 torr. 116 0.7 M I I r T 1 I I I r 0. 6 0 > OI 25. 8 t o r r 0.5 I 21. 8 torr Al 18. 1 t o r r 0.4 UJ U z < z u UJ (3 < I_J 0.3 a > UJ (3 < < Z 0.2 u ^ 0. 1 0.0 I I I 0 10 20 30 40 50 .60 70 80 REDUCED FIELD STRENGTH (E/N>/CTd3 Fig. 35. I n f l u e n c e of CO2 l a s e r on t h e d i s c h a r g e v o l t a g e s (AV/V^. where AV = V, ( w i t h l a s e r ) - V ^ ( w i t h o u t l a s e r ) ) f o r 7.6% C2H3F, 87.4% Ar. and 5% He a t various pressures. 117 1.2 1rIrIIIIjr TIr-ii-TI I I I I o > OI 25. 8 t o r r I 21. 8 torr UJ 0. 8 - Al 18. 1 t o r r (3 z < z u u < UJ 0.6 a > 0. 4 - UJ 13 K < Z u (A 0. 2 - 0.0 ^^ ^ -G e I i I I I I 20 40 60 80 100 120 REDUCED FIELD STRENGTH (E/N)/CTdD Fig. 36.1 Influence of CO2 laser on the discharge voltages (AV/V^, where AV = V, (with laser) - V^(without laser)) for 15% C^n^T. 80% Ar, and 5% He at various pressures. 118 concentrations of the attacher and higher operating pressures gave stronger responses due to the higher partial pressures of the attacher. Vinyl fluoride responded much stronger to IR pumping than ammonia. This result is not surprising since CpH.5F has a stronger absorption at 10.6 ^ m than NH.^ and since it i may be estimated that CpH.^F is a stronger attacher than NH3. Additional experiments using this apparatus are planned for future investigations. Other gases similar to Ce^U^Y, such as CpH^^Cl and CpHF.^, will be studied utilizing a higher powered tunable COp laser. Of particular Interest is the effect of IR pumping of C.^Fj. as proposed by Eisele [96]. This compound has been proposed as an attacher for opening switches and, therefore, would be well suited for optically enhanced attachment in working opening switches. Photodissociation in Discharges Sustained bv a Hollow Cathode Device The final set of experiments investigated photodissociation as a control mechanism. These experiments were performed in an apparatus similar to that used for the IR pumping experiments. The apparatus for the The photodissociation experiments is shown in Fig. 37. region of optical interaction with the discharge was between two wire meshes. Behind one mesh was placed a tube 119 GRID-ANODE HOLLOW CATHODE \ GRID" SECONDARY ANODE H. PUMP ATTACHER QUARTZ TUBE QUARTZ WINDOW Fig. 37. Apparatus for photodissociation enhanced attachment experiments. 120 which acted as a hollow cathode. A DC voltage below the As breakdown voltage of the gas was applied to the meshes. in the IR pumping experiments, a discharge was established between the hollow cathode and the closer mesh. A gas flow of He through the hollow cathode forced the plasma of the discharge into the region between the meshes. The injected plasma ionized the surrounding gas and acted as a sustainment source for the inter-mesh region. Light from a spark source illuminated the externally sustained discharges through a side window made of quartz. voltages were measured between the meshes. Measurements were conducted in He with admixtures of HpC=CFp. HpC=CClp. and H2C=CHC1. Typical transient voltage Transient signals showed an initial decrease in the discharge voltage, lasting approximately 20 jis. caused by photoionization followed by an Increase in discharge voltage. This increased voltage would rise to peak values much larger than the voltage decrease pulse within several tens of microseconds followed by an exponential decay of approximately 500 ^ts. The increased voltage was caused by optically enhanced attachment through photodissociation. as described above. Increases in the discharge resistances of Figure 38 presents the data The enhanced as much as 30% were observed. from photodissociation experiments in CpH^Cl. attachment appears to reach a point of saturation as the 121 10 Vo C2H3CI 20 Fig. 38. Change in relative discharge resistances versus E/N for an externally sustained discharge containing C2H3CI caused by enhanced attachment through UV photodissociation. 122 concentration of CpH.^Cl was increased, perhaps due to the decreased penetration depth of the UV photons with increased CpH.^Cl partial pressures. effect is strongest at low E/N. Additionally, the This result is not surprising since most attachers containing CI have their highest attachment rates at very low values of E/N [43]. CHAPTER VI CONCLUSION The mechanisms described in this dissertation demonstrate the feasibility of optical control of diffuse discharges. (See Table 5.) These control methods, which are not all of the possible control methods [93]. are capable of large conductivity changes under moderate laser powers and may cause increased or decreased conductivity. UV ionization has been known for sometime to be an effective means of controlling the spatial properties of self-sustained diffuse discharges and has recently been demonstrated as a sustainment source for opening switches. The results of the work presented here show that spark arrays can be used as efficient ionization sources for externally initiated and externally sustained discharges. Ultraviolet sustained discharges in Ar with admixtures of either C^F^ or 0^ as attachers were investigated. At intermediate ranges in E/N these discharges were found to exhibit negative differential conductivity. These ranges began at about 3 Td for CpFg bearing discharges and at about 5 Td for Op bearing discharges. This effect is believed to be the consequence of an attachment coefficient for the dissociative process which increases strongly with E/N in this range. Ultraviolet initiated discharges were investigated as a means to simplify the circuits which 123 124 w < JN 8 M lb It X M O n X X u M n X U. U M X M u *0 e M e 1. ** a 3 JS M >t V X B b w O b >, B -9 9JS B o *M o (0 ^ M u w u B o > ** 9* M M M 0M m >M t-ja m tt tt IM e > > t * > t "O * o *< B < * N - Bw a ^ V m ** B a u 3 m a w e e in o >> H 9 111 M a o l w B I . ^ > m o tt Ji B N b O > M e a em b e e M >t u B a B a a B a a M M o a b . WB M M en 3 a E e M la 1. M lA e < "9 ^ 9 k> e tt a >t u a a w b B b M M b b a - e >> a e b b 9 IM W jijz o ^ (I a u u b B U tt ja tt ifi ** & J3 b O x a a a ^ u o ^ X( a ^ a . M a O b M a B b w w B a .B u a a 3 B a u u B a S U M ^ M V V > tm -M . a ^ - >. e M . M e s ** ^ ^. > ^ ^ * o a a "^ a * o u 3 V > ' 9 e Jt e - > > - . 3 -M W W M IM W O >t O O'* 3 V m B < ^ a w e ** M 3 "O o M e b 3 M a b . B M ja b B b a 6 ^ O 3 O U B U " o o a g 0 >> 6 a u .e "OO BM ao O >t o o a > a a M a u a k B a o a M a > b a b 1 a ^ * a b u > b O a b b 3 U e a > a M b * .a a w (A a a Ul Ul H Z W M X U H H 111 N H < a Ul o o O X Ul a Q Ul o w o u > 125 drive laser discharges. It was found that by adjusting the discharge pressure (300 torr), gap spacing (0.35 cm), and the ratio of the gas components (COp:Np:He=l:16:23) the glow voltage of the discharges was less than half the breakdown voltage. With this arrangement, a laser could be driven directly by a transmission line without any intermediate switches. The turn-on time of discharges sustained or initiated by spark arrays will be limited by the sequential breakdown of the entire array. Measurements of the 20 gap array built for the research described here, indicated a decreasing turn-on time with increasing voltage of the pulse forming line used to drive the array. A minimum turn-on time of about 13.5 ns was observed for a voltage of 35 kV and a maximum turn-on time of 18.3 ns was observed for a voltage of 22 kV. Efficient (impedance matched) spark arrays require the operation of a large number of spark gaps in series ( 100). The turn-on time of such a ^ device could be as much as 70 ns. The turn-off time is limited by the decay of the plasma produced by the sparks. Photodetachment provides a means of overcoming the attachment process and allows decreased turn-on time and lower switch losses. The experiments presented demonstrated that externally sustained discharges in mixtures of argon and nitrogen with admixtures of oxygen exhibit a negative differential conductivity in an 127 The current density changes were found to increase linearly with laser intensity. This indicated that only a fraction of the negative ions were being photodetached and 5 2 that with higher laser power densities (>8X10 higher current increases could be obtained. W/cm ) even In similar discharges, but using a 13 kfl pulse forming line to drive the discharges, discharge resistivity changes were measured for initial E/N values as low as 16 Td. Glow-to-arc transitions were eliminated in atmospheric pressure, self-sustained discharges by using photodetachment. Conditions for these experiments included a gas mixture of 5.3% Op and 94.7% Ar and a laser power density of 800 kW/cm^ at 590 nm. Optically enhanced attachment, the least understood of the processes, could provide increased electron depletion and decreased opening times from gas additives that are normally weak attachers. Infrared pumping of attachers. such as SF-, have shown increased attachment rates with increased vibrational excitation [95]. Experiments in continuous, low pressure discharges in NH^ and CpH.^F have shown increased transient discharge voltages of as much as 1% when illuminated by very low COp laser power densities 2 (less than 1 W/cm ) . Such favorable response from a weak control source makes optically enhanced attachment appear very attractive. Perhaps even more attractive is the production of highly vibrationally excited, strongly 126 intermediate E/N range. Significant changes in the discharge characteristics caused by illumination of the discharges by a visible laser were predicted from calculations of discharges containing NpO [21] and were demonstrated for Op bearing discharges presented here. Current density increases of as much as 900% were observed in atmospheric pressure, diffuse discharges containing 13.2% Og. 2.6% N2. and 79.3% Ar. The laser power density 2 for this degree of change was 800 kW/cm . It was found that the current density change increased with increased Op concentration as a result of the Increased negative ion production with increasing Op concentration. The value of E/N for peak current density change also increased with Increased Op concentration. This is a consequence of the optimum ratio of electron density to negative ion density for peak current changes occurring at the value of E/N where the discharges changed from externally sustained to self-sustained. Below this value of E/N. too few electrons are present to build-up a substantial negative ion density: above this value of E/N the electron production from photodetachment must compete with internal ionization. Since ionization must balance attachment (k. N=k ^^N ^^) ion att att for a self-sustained discharge, the value of E/N where the discharges changed from externally to self-sustained increased with O2 concentration. 128 attaching compounds through photodissociation of weakly attaching molecules using UV lasers. Results from experiments on CgH^Cl and CpHF^ reported by Rossi. Helms, and Lorents [100] showed attachment rate changes of several orders of magnitude. The results reported here on CpHpFg. ^2^2^^2* ^^^ ^2^3^^ ^^ ^ ^ pressure, externally sustained, continuous discharges indicate changes in the discharge resistances of as much as 30% when illuminated by a weak UV spark source. The large variety of processes allows one to choose a mechanism proper for the particular application needed. For many cases, control techniques other than optical techniques may be more favorable either due to higher efficiencies or lower cost or due to unsuitable light sources in the necessary spectral range for the process to be controlled. As an example of this, E-beams are at this time more efficient sustainment sources than lasers. However, optical control techniques provide means of influencing discharges that cannot be achieved by any other means (such as optically enhanced attachment), and. in this way, switches with controllable aspects, not realizable by any other switching technique, can be constructed. In some cases, such as gas laser discharges sustained by UV spark sources rather than E-beams, optical control provides less expensive and more reliable operation than other methods. The simple controllability of both spatial and temporal 129 characteristics of light are two of the major advantages of optical sources over all other discharge control devices. However, much further work is necessary before many of these optical control methods can be utilized in practical switches. The results presented in this treatise are only proof-of-principles and are still quite far from realistic switch conditions. Questions of scaling and timing have not been fully addressed for many of these processes, and the results of Eisele's experiments [96], when compared to the results presented in this dissertation, are still unexplained. Additionally, although the basic optical processes are understood, the interaction of these optical processes with other discharge processes and the effects of the circuit are not fully understood. The promise of optically controlled diffuse discharges along with the many, as yet, unresolved questions could provide much further fruitful research in both basic science and engineering. LIST OF REFERENCES 1. P. Lenard, "Ueber die Elektricitaetszerstreuung in ultraviolett durchstrahlter Luft," Ann, de Phvs.. vol. 308, pp. 298-319, 1900. P. Lenard. "Ueber Wirkungen des ultravioletten Lichtes auf gasfoermige Koerper," Ann, de Phvs.. vol. 306, pp. 486-507. 1900. J. S. Townsend, "The Conductivity Produced in Gases by the Aid of Ultra-Violet Light." Phil. Mag.. vol. 3, pp. 557-576. June 1902. L. E. Kline and L. J. Denes, "Preionized SelfSustained Laser Discharges," Applied Atomic Collision Phvsics. Volume 3: Gas Lasers, eds. E. W. McDaniel and W. L. Nighan, New York: Academic Press, pp. 387-422. 1982. L. J. Denes and L E. Kline. "Electrode and Preionizer - Geometry Effects on TEA Laser Discharge Formation." Appl. Phvs. Lett.. vol. 30, pp. 197-199, Feb. 1977. A. H. Guenther and J. R. Bettis. "The Laser Triggering of High Voltage Switches." .1. Appl. Phvs. D: A P P I . Phvs.. vol. 2. pp. 1577-1613. 1978. R. A Dougal. P. F. Williams, and A. H. Guenther. "Breakdown Processes in Laser-Triggered Switching." Proc. 4th IEEE Pulsed Power Conf.. Alb.. N.M.. pp. 443-446. 1983. H. C. Harjes. K. H. Schoenbach. M. Kristiansen. A. H. Guenther. and L. L. Hatfield. "Laser Triggering Through Fiber Optics of a Low Jitter Spark Gap." IEEE Trans, on Plasma Sci.. vol PS-8. pp. 170-176. Sept. 1980. J. E. M. Goldsmith and J. E. Lawler. "Optogalvanic Spectroscopy," Contemn. Phvs.. vol. 22, pp. 235-248, 1981. R. B. Green, R. A. Keller. G. G. Luther. P. K. Schenck. and J. C. Travis, "Galvanic Detection of Optical Absorption in a Gas Discharge," Appl. Phvs. Lett. vol. 29. pp. 727-729. Dec. 1976. 2. 3. 4. 5. 6. 7. 8. 9. 10. 130 131 11. W. B. Bridges. "Characteristics of an Opto-Galvanic Effect in Cesium and Other Gas Discharge Plasmas." J. Opt. Soc. Am.. vol. 68, pp. 352-360. March 1978. D. M. Pepper, "Analysis of th Optogalvanic Effect in Discharge Plasmas Using Rate Equations in a Modified Schottky Formalism." IEEE J. Quantum Electron.. vol. QE-14. pp. 971-977, Dec. 1978. G. Erez, S. Lavi, and E. Miron, "A Simplified Theory of the Optogalvanic Effect," IEEE J. Quantum Electron.. vol. QE-15, pp. 1328-1332, Dec. 1979. A. Ben-Amar, G. Erez, and R. Shuker, "Pulsed Resonant Optogalvanic Effect in Neon Discharges." . . Appl. T Phvs.. vol. 54. pp. 3688-3698. July 1983. D. K. Doughty and J. E. Lawler. "Spatially Resolved Electric Field Measurements in the Cathode Fall Using Optogalvanic Detection of Rydberg Atoms," Appl. Phvs. Lett.. vol. 45, pp. 611-613, Sept. 1984. J. . Lawler and A. H. Guenther, "Applications of Optogalvanic Effects in Opening Switches," Proc. 3rd IEEE Pulsed Power Conf.. Alb., N.M.. pp. 147-150. 1981. D. K. Doughty and J. E. Lawler, "Optogalvanic Effects in Obstructed Glow Discharge." Appl. Phvs. Lett.. vol. 42, pp. 234-236. Feb. 1983. A. H. Guenther. "Optically Controlled Discharges," Proc. ARO Workshop on Repetitive Opening Switches. Tamarron. CO. pp. 47-64. 1981. (DTIC No. AD-Al10770) K. H. Schoenbach. G. Schaefer. M. Kristiansen. L. L. Hatfield, and A. H. Guenther, "Concepts for Optical Control of Diffuse Discharge Opening Switches," IEEE Trans, on Plasma Sci.. vol PS-10. pp. 246-251, 1982. G. Schaefer, P. F. Williams. K. H. Schoenbach. and J. T. Mosely. "Photodetachment as a Control Mechanism for Diffuse Discharge Switches." IEEE Trans, on Plasma Sci.. vol PS-11. pp. 263-265. Dec. 1983. 12. 13. 14. 15. 16. 17. 18. 19. 20. 132 21. G. Schaefer, K. H. Schoenbach. H. Krompholz. M. Kristiansen. and A. H. Guenther. "The Use of Attachers in Electron Beam Sustained Discharge Switches - Theoretical Considerations," Lasers and Particle Beams, vol. 2. pp. 273-291, 1984. K. H. Schoenbach, M. Kristiansen. and G. Schaefer. "A Review of Opening Switch Technology for Inductive Energy Storage." Proc. IEEE, vol. 72. pp. 1019-1040. Aug. 1984. C. W. Mendel and S. A. Goldstein. "A Fast-Opening Switch for Use in REB Diode Experiments," J. Appl. Phvs.. vol. 48. pp. 1004-1006. 1977. E. M. Honig, "Generation of a 75-MW 5-kHz Pulse Train From an Inductive Energy Store." IEEE Trans. Plasma Sci.. vol PS-12. pp. 24-28. March 1984. R. J. Commisso. R. F. Fernsler, V. E. Scherrer, and I. M. Vitkovitsky, "Inductively Generated High Voltage Pulse Using an Electron Beam Controlled Opening Switch," Appl. Phvs. Lett., vol. 47, pp 1056-1058. 1985. J. F. Lowry. L. E. Kline, and J. V. R. Heberlein. "Electron-Beam Controlled On/Off Discharge Characterization Experiments." Proc. 4th IEEE Pulsed Power Conf.. Alb.. N.M.. pp. 94-97. 1983. K. H. Schoenbach, G. Schaefer. M. Kristiansen, H. Krompholz, H. Harjes, and D. Skaggs, "An Electron-Beam Controlled Diffuse Discharge Switch." J. Appl. Phvs.. vol. 57. p. 1618. 1985. G. Schaefer. K. H. Schoenbach. M. Kristiansen, B. E. Strickland. R. A. Korzekwa. and G. Z. Hutcheson. "The Influence of the Circuit Impedance on an Electron-Beam Controlled Diffuse Discharge with a Negative Differential Conductivity." Appl. Phvs. Lett.. vol. 48. pp. 1776-1778. June 1986. H. Seguin and J. Tulip. "Photoinitiated and Photosustained Laser." Appl. Phvs. Lett.. vol. 21. pp. 414-415. Nov. 1972. R. C. Lind. J. Y. Wada. G. J. Dunning, and W. M. Clark. "A Long-Pulse High-Energy COp Laser Pumped by an Ultraviolet-Sustained Electric Discharge." IEEE . . Quantum Electron. . vol. QE-10. pp. T 818-821. Oct. 1974. 22. 23. 24. 25. 26. 27. 28. 29. 30. 133 31. D. B. Gurevich. M. A. Kanatenko, and I. V. Podmoshenskii,"Breakdown in an Externally Sustained Volume Discharge with External Photoionization," Sov. J. Plasma Phvs.. vol. 5. pp. 760-764. Nov.-Dec. 1979. D. R. Suhre. "UV Sustained CO Laser Discharge I. Photoionization Studies." J. Appl. Phvs.. vol. 52. pp. 3855-3857. June 1981. D. R. Suhre and S. A. Wutzke. "UV Sustained CO Laser Discharge II. Discharge Studies." J. Appl. Phvs.. vol. 52. pp. 3858-3860. June 1981. Yu. I. Bychkov. D. Yu. Zaroslov. N. V. Karlov, G. P. Kuz'mln. V. V. Osipov, A. M. Prokhorov. and V. A. Tel'nov. "Initiation of a High-power Nonself-Sustained Volume Discharge in Molecular Gases by Ultraviolet Radiation from a Plasma Cathode." Sov. J. Quantum Electron.. vol. 12, pp. 1105-1107. Aug. 1982. W. M. Moeny and M. von Dadelszen. "UV-Sustained Glow Discharge Opening Switch Experiments." Proc. 5th IEEE Pulsed Power Conf.. Alexandria. VA, pp. 630-632, June 1985. C. M. Young, M. von Dadelszen, J. M. Elizondo. J. W. Benze, W. M. Moeny, and J. G. Small, "Investigation of a UV-Sustained Radial Glow Discharge Opening Switch," Proc. 5th IEEE Pulsed Power Conf.. Alexandria. VA, pp. 633-635, June 1985. G. Z. Hutcheson, J. R. Cooper, B. E. Strickland, G. Schaefer. and K. H. Schoenbach. "Time Dependent Emission from a Spark Array Used as an Ionization Source for Diffuse Discharges," to be submitted for publication in 1986. G. Schaefer and K. H. Schoenbach, "A Review of Diffuse Discharge Opening Switches," to be published in IEEE Trans. Plasma Sci.. Oct. 1986. B. E. Cherrington, Gaseous Electronics and Gas Lasers. New York: Pergamon Press, 1979, ch. 7. L. G. Christophorou, S. R. Hunter, J. G. Carter, and R. A. Mathis. "Gases for Possible Use in Diffuse-Discharge Switches," Appl. Phvs. Lett.. vol. 41, pp. 147-149, July 1982. 32. 33. 34. 35. 36. 37. 38. 39. 40. 134 41. L. G. Christophorou. S. R. Hunter. J. G. Carter, S. M. Spyrou, and V. K. Lahdawala, "Basic Studies of Gases for Diffuse-Discharge Switching Applications," Proc. 4th IEEE Pulsed Power Conf.. Alb., NM, pp. 702-708, 1983. M. A. Biondi, "Electron-Ion Processes in Externally Controlled Diffuse Discharges," Report on US-FRG Joint Seminar on Externallv Controlled Diffuse Discharges. Bad Honnef. Ger., pp. 134-140. Aug. 1983. L. G. Christophorou, Atomic and Molecular Radiation Phvsics. New York: Wiley-Interscience, 1971. ch. 6 and 7. E. Nasser. Fundamentals of Gaseous Ionization and Plasma Electronics. New York: Wiley-Interscience. p. 82. 1971. F. L. Mohler and C. Boeckner. "Photoionization of Some Alkali Vapors." J. Res. Nat. Bur. Standards, vol. 3, pp. 303-315. 1929. G. L. Weissler. "Photoionization in Gases and Photoelectric Emission from Solids." Encvclopedia of Phvsics. Volume XXI. Berlin: Springer, pp. 304-341, 1956. G. L. Weissler, W. C. Walker, and N. Weinfan, "Photoionization Efficiencies and Cross Sections in Op. N2. COg. A (sic). H2O. Hg. and CH^," Phvs. Rev.. vol. 99, p. 542, 1955. G. L. Weissler and P. Lee, "The Photo-Ionizatlon Cross-Section of Neon," Proc. Rov. Soc. A. vol. 220, p. 71, 1953. D. J. Baker, D. E. Bedo, D. H. Tambaulian. "Continuous Photoelectric Absorption Cross Section of Helium," Phvs. Rev., vol.124, p. 1471, 1961. D. F. Grosjean and P. Bletzinger, "Photoionization and Photoabsorption Characteristics of Laser Seed Compunds." IEEE J. Quantum Electron., vol. QE-13. pp. 898-904. Nov. 1977. 42. 43. 44. 45. 46. 47. 48. 49. 50. 135 51. J. Samson and R. B. C a l m s . "Total Absorption Cross-Section of CO and COp in the Region 550-200 A and Ionization Potential of Oxygen," J. Opt. Soc. Am.. vol. 56, p.527. 1966. K. Watanabe. "Ionization Potential of Some Molecules," J. Chem. Phvs.. vol. 26. pp. 542-547, March 1957. J. I. Levatter and S. C. Lin, "Necessary Conditions for the Homogeneous Formation of Pulsed Avalanche Discharges at High Gas Pressusres," J. Appl . Phvs.. vol. 51, pp. 210-222, Jan. 1980. A. J. Palmer. "A Physical Model of Initiation of Atmospheric Glow Discharges," Appl. Phvs. Lett.. vol. 25, pp. 138-140. 1974. J. R. Cooper, A Fast Preionization Source for Diffuse Discharges Containing Attachers. M. S. Thesis, Texas Tech University, 1983. J. R. Cooper, G. Z. Hutcheson, M. Kristiansen, G. Schaefer, K. H. Schoenbach. and A. H. Guenther, "A Multispark Preionization Source for Diffuse Discharges Containing Attachers, " Proc. 4th IEEE Pulsed Power Conf.. Alb., NM, pp. 726-728, 1983. R. S. Taylor and K. E. Leopold, "Magnetically Induced Pulser Laser Excitation," Appl. Phvs. Lett.. vol. 41, pp. 335-337. Sept. 1985. R. S. Taylor and K. E. Leopold. "Microsecond Duration Optical Pulses From a UV-Preionized XeCl Laser," A P P I . Phvs. Lett., vol. 47, pp. 81-83, July 1985. 0. de Witte, B. Lacour, and C. Vannier, "Switchless TE Gas Lasers," Conf. Lasers and Electro-Optics. Digest of Technical Papers. Washington.DC, April 1982. B. Lacour and C. Vannier, "X-ray Photoswltching of a 1-J Excimer Laser," Conf. Lasers and Electro-Optics. Digest of Technical Papers. Baltimore. MD, May 1985. R. T. Van Brunt and M. Misakian, "Role of Photodetachment in Initiation of Electric Discharges in SFg and 02'" - Appl . Phvs. . vol. T 54, pp. 3074-3079. June 1983. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 136 62. D. H. Douglas-Hamilton and S. A. Mani. "Attachment Instability in Externally Ionized Discharge." J : _ A P P I . Phvs.. vol. 45. p. 4005, Oct. 1974. R. A. Haas. "Stability of Excimer Laser Discharges," Applied Atomic Collision Phvsics. Volume 3: Gas Lasers. eds. E. W. McDaniel and W. L. Nighan, New York: Academic Press, PP. 423-452. 1982. S. J. Smith, Proc. Int. Conf. on Ionization Phenomena in Gases, p. IC 219, 1960. L. M. Branscomb, S. J. Smith, and G. Tiscone, "Oxygen Metastable Atom Production Through Photodetachment," J. Chem. Phvs.. vol. 43, pp. 2906-2907, 1965. D. S. Burch, S. J. Smith, and L. M. Branscomb. "Photodetachment of O9." Phvs. Rev.. vol. 112. p. 171, 1958. 67. P. C. Cosby. R. A. Bennett, J. R. Peterson, and J. T. Mosely, "Photodissociation and Photodetachment of Molecular Negative Ions. II. Ions Formed in Oxygen." J. Chem. Phvs.. vol. 63. pp. 1612-1620. Aug. 1975. L. C. Lee and G. P. Smith, "Photodissociation and Photodetachment of Molecular Negative Ions. VI. Ions in O2/CHVH2O Mixtures From 3500 to 8600 A," J. Chem. Phvs.. vol. 70, pp. 1727-1735, Feb. 1979. L. M. Chanin, A. V. Phelps, and M. A. Biondi, "Measurements of the Attachment of Low-Energy Electrons to Oxygen Molecules," Phvs. Rev.. vol. 128, pp. 219-230. Oct. 1962. J. L. Moruzzi and A. V. Phelps, "Survey of Negative-Ion-Molecule Reactions in 0^* ^^o' ^2^' CO, and Mixtures of These Gases at High Pressures," J. Chem. Phvs.. vol. 45, pp. 4617-4627, Dec. 1966. D. Spence and G. J. Schulz, "Temperature Dependence of Dissociative Attachment in O2 and CO2." Phvs. Rev.. vol. 188. pp. 280-287. Dec. 1969. 63. 64. 65. 66. 68. 69. 70. 71. 137 72. W. R. Henderson, W. L. Fite. and R. T. Brackman, "Dissociative Attachment of Electrons to Hot Oxygen," Phvs. Rev.. vol. 183. pp. 157-166, July 1969. R. S. Berry and C. W. Reimann. "Absorption Spectrum of Gaseous F and Electron Affinities of the Halogen Atoms," J. Chem. Phvs.. vol. 38, pp. 1540-1543, Apr. 1963. 74. E. J. Robinson and S. Geltman, "Single- and Double-Quantum Photodetachment of Negative Ions." Phvs. Rev.. vol. 153, pp. 4-8, Jan. 1967. A. Mandl. "Electron Photodetachment Cross-Section of the Negative Ion of Fluorine," Phvs. Rev. A. vol. 3, pp. 251-255. Jan. 1971. L. E. Thurmond III, Photodetachment as a Discharge Control Mechanism in Gases Containing Oxygen. M. S. Thesis, Texas Tech University, 1985. J. W. Dettmer and A Garscadden. "Role of 02(a*A) in the Oxygen Plasma," .T. de Phvs. . vol. 40, pp. C7-135 - C7-136, 1979. H. J. J. Seguin, D. McKen, and J. Tulip, "Photon Emission and Photoionization Measurements in the COp Laser Environment," .T. Appl . Phvs. . vol. 28, pp. 487-489, May 1976. 79. 80. B. E. Cherrington, Gaseous Electrolncs and Gas Lasers. New York: Pergamon Press, 1979, ch. 4. K. J. Nygaard, H. L. Brooks, and S. R. Hunter. "Negative Ion Production Rates in Rare Gas-Halide Lasers," IEEE J. Quantum Electron., vol. QE-15, pp. 1216-1223, Nov. 1979. T. Holstein, "Dissociative Attachment of Electrons to Diatomic Molecules," Phvs. Rev.. vol. 84, p. 1073, Dec. 1951. T. F. O'Malley, "Theory of Dissociative Attachment," Phvs. Rev.. vol. 150, pp. 14-29, Oct. 1966. F. Fiquet-Fayard. "Theoretical Problems in the Interpretation of Dissociative Attachment Experiments." Vacuum, vol. 24, pp. 533-547, 1974. 73. 75. 76. 77. 78. 81. 82. 83. 138 84. M. Allan and S. F. Wong. "Dissociative Attachment from Vibrationally and Rotationally Excited HCl and HF." .T. Chem. Phvs. . vol. 74, pp. 1687-1681, Feb. 1981. M. Allan and S. F. Wong. "Effect of Vibrational and Rotational Excitation on Dissociative Attachment in Hydrogen," Phvs. Rev. Lett.. vol. 42, pp. 1791-1794, Dec. 1978. P. J. Chantry, "Dissociative Attachment in COp." J Chem. Phvs.. vol. 57, pp. 3180-3186, Oct. 1972. P. J. Chantry, "Temperature Dependence of Dissociative Attachment in NpO." .T. Chem. Phvs. . vol. 51, pp. 3369-3379, Oct. 1969. C. L. Chen and P. J. Chantry, "Temperature Dependence of SFg, SF~ and F~ Production from SFg," Bull. Am. Phvs. Soc.. vol. 15, pp. 418-419, 1970. 89. C. L. Chen and P. J. Chantry. "Electron Attachment in SF.." Bull. Am. Phvs. Soc.. vol. 17, pp. 406-407. 1972. W. L. Fite and R. T. Brackman, Proc. VI Int. Conf. lonz. Phenom. in Gases. Vol. I. Paris, pp. 21-26. 1963. T. F. O'Malley, "Calculation of Dissociative Attachment in Hot 02" Phvs. Rev.. vol. 155, pp. 59-63, March 1967. S. K. Srivastava and 0. J. Orient. "Double E-Beam Technique for Collision Studies from Excited States: Applications to Vibrationally Excited COp," Phvs. Rev. A. vol. 27. pp. 1209-1212, Feb. 1983. 93. G. Schaefer, "Optically Controlled Diffuse Discharges," USAF Pulsed Power Lecture Series No. 49. Texas Tech University. J. J. Tiee and C. Wittig, "Optically Pumped Molecular Lasers in the 11-17 ^m Region," J. Appl. Phvs.. vol. 49. pp. 61-64. Jan. 1978. 85. 86. 87. 88. 90. 91. 92. 94. 139 95. C. L. Chen and P. J. Chantry, "Photon-Enhanced Dissociative Electron Attachment in SF^ and Its D 96. Isotope Selectivity." J. Chem. Phvs.. vol. 71. pp. 3897-3907. Nov. 1979. F. L. Eisele. Photon Induced Electron Attachment. Final Report to Air Force Wright Aeronautical Laboratory on Contract No. F33615-82-C-2202, Georgia Institute of Technology, Atlanta, GA, Dec. 1984. M. J. Berry and G. C. Pimental, "Hydrogen Halide Photoellmination Chemical Lasers," J. Chem. Phvs.. vol. 51, pp. 2274-2275, Sept. 1969. E. R. Sirkin and G. C. Pimental, "HF Rotational Laser Emission Through Photoellmination from Vinyl Fluoride and 1,1-Dlfluoroethene," J. Chem. Phvs.. vol. 75. pp. 604-612, July 1981. C. R. Quick and C. Wittig, "IR Multiple Photon Dissociation of Fluorinated Ethanes and Ethylenes: HF Vibrational Energy Distributions," J. Chem. Phvs.. vol. 72, pp. 1694-1700. Feb. 1980. 97. 98. 99. 100. M. J. Rossi. H. Helm, and D. C. Lorents, "Photoenhanced Electron Attachment of VinyIchloride and Trifluoroethylene at 193 nm." A P P I . Phvs. Lett.. vol. 47. pp. 576-578, Sept. 1985. 101. T. W. G. Solomons, Organic Chemlstrv. New York: John Wiley & Sons, 1978. 102. I. M. Beterov and N. V. Fateyev. "Optogalvanic Demonstration of State-To-State Dissociative Electron Capture Rate in I Q * " Optics Comm.. vol 40, pp. 425-429, Feb. 1982. 103. A. B. Peterson, C. Wittig, and S. R. Leone. "Electronic-To-Vibrational Pumped CO2 Laser Operating at 4.3, 10.6, and 14.1 jim," J. A P P I . Phvs.. vol. 47, pp. 1051-1054, March 1979.