Blaze Labs EHD Thrusters Research© Engineer Xavier Borg - Blaze Labs Research
Home of the highest performance EHD thruster cells
- What is an EHD Thruster
- De Seversky video footage and article scan from 1964
- How do they work
- Hints for a good thruster
- Peek's Corona Inception Voltage (CIV)
- Corona Mechanism
- Corona Losses & Efficiency
- Calculating collector radius of curvature
- Streamer Breakdown in EHD thrusters
- Improving thrust by modifying ion drift velocity
- A word about health hazards
What is an EHD thruster?
EHD stands for Electro Hydro Dynamics which is the study of the flow of a fluid under the effect of an electric field. The principle of ionic air propulsion with corona generated charged particles has been known as from the earliest days of the discovery of electricity. One of the first reference to sensing moving air near a charged tube appeared in year 1709 in a book titled 'Physico-Mechanical Experiments on Various Subjects' by F.Hauksbee. Many other pioneers of electricity, including Newton, Faraday, and Maxwell, studied this phenomenon. Unfortunately, EHD is not a common topic in most high school syllabus, which is the main reason why most of the general public get confused when seeing such devices in action.
An EHD thruster is an electrohydrodynamic device which ionises air and moves the charged ion cloud in a way and direction to transfer momentum to neutral air molecules. By Newton's third law of motion, action is equal and opposite to reaction, and the EHD thruster will move in the opposite direction of the ion cloud. Ionocrafts was the name given to the first kind of vertical takeoff EHD thrusters designed during the early 60's, and form part of the EHD thrusters family. Recently, this effect has gained popularity under the less appropriately titled 'lifter' which due to the lack of knowledge of EHD by most people, has been related to some sort of antigravity effect. It is a well known fact that these devices produce thrust along their own axis, and not against the force of gravity as would be expected from an antigravity device. An EHD thruster in its simplest form is made up of two electrodes, one with a sharp edge, the ioniser and one with a smooth edge, the collector, which when powered by a high dc voltage (a few kV ) produces thrust against the surrounding medium, normally air. Shown above is a diagram of how it works, and the photo below shows a simple EHD thruster utilising this effect. This is the basic EHD flying machine.
An EHD thruster works without moving parts, flies silently, uses only electrical energy and when immersed in a fluid (air, oil, etc..) is able to lift its own weight together with additional payload. The basic design of the simple lifter has been fully described in the Townsend Brown US Patent N°2949550 filed on 3rd July 1957 and titled "Elektrokinetic Apparatus". Even though T.T.Brown was fully aware that the thrust from his devices was due to ion interactions, lack of EHD knowledge at that time, resulted in very low efficiency operation of these first EHD thrusters. Shown below is one of Brown & Bahnson's designs described in US Patent #3,223,038.
While extensive research was performed in the 1950's and 1960's on the use of electric propulsion for interplanetary spaceflight, many promising concepts had to be abandoned due to the technological limitations of the power conditioning systems in use at the time. It is also understood that the research & development of ionic thrusters by NASA at those days was aimed mainly for interplanetary space flights, and the fact that ionocrafts need a fluid medium to work has led these fantastic devices to be largely abandoned by the scientific community since the late 1960's. To date no consistent effort has been made to reevaluate these approaches in light of modern power processing technologies and develop flying machines to operate within the atmosphere. During the 1960's a lot of work on EHD technology had been accomplished by Major De Seversky (pictured below). De Seversky noticed an air flow developing between the two electrodes of an air ioniser commonly used to clean air. "To an old flyer like me," said the major, "anything that stirs up a wind is a flying machine. So I began to develop the idea." In fact a few years later, he patented The Ionocraft patent no:US3130945 in April 28, 1964. The major seemed concerned that the Ionocraft might be mistaken for a kind of space vehicle. "This is not a spacecraft," he explained emphatically to forestall any possible misunderstanding. "It's an airplane, designed to operate within the atmosphere. But it will be able to do things that no present type of aircraft can accomplish." Indeed, this misunderstanding still prevails in the present days, but our research clearly shows that EHD thrusters & lifters do not work in vacuum.
De Seversky video footage and article scan from 1964
How do they work
The top sharp electrode ionises the air. If the electrode is positive, free electrons in the vicinity will accelerate towards it, and strip off other electrons from the air molecules around the sharp wire. A cloud of heavy positive charges is thus formed, and the avalanche of electrons approaching the sharp electrode account for the corona & ionisation current. In their mad rush from the ion emitter to the smooth negative electrode, the positive ions bump into neutral air molecules-air particles without electric charge. The force exerted on them by the electric field is offset by the force of friction caused by collisions of the ions with the neutral air molecules. As a result, ions drift through the air gap with an approximately constant velocity Vd, that is proportional to the electric field given by Vd=kE, where the proportionality constant K is called the ion mobility, the highest the value the more mobile (faster) and the less friction is offered.
The terrific wallop in these collisions hurls a mass of neutral air downward along with the ions. The distance in cm travelled by an ionised air molecule until it hits a neutral air molecule is given by the mean free path and is equal to 5E-3/P, where P=760 Torr at sea level. The larger the air gap relative to the mean free path, which works out to be equal to 6.6E-6cm, the more probability there is of an ion repeatedly hitting neutral molecules, and therefore the more impacts and thus effective thrust we get. During these collisions, the ion charge is not transferred to the neutrals. When they reach the lower smooth electrode, the ions, still being positive, hit it and neutralise themselves. But the grid has no attraction for the neutral air particles that got bumped along. So the air flows right along the sides of the lower electrode, making a downdraft of neutral air beneath the EHD device. The fact that most ions are neutralised at the collector explains why the reading we get from ion measuring meters setup below such devices does not account for the measured thrust. In fact for a good EHD thruster, such a reading should be close to zero. If however, one accurately measures the force exerted by the air exiting the collector side over a flat surface, it is found that this force is equal and opposite to the thrust of the device.
The EHD thruster rides on this shaft of air, getting it's lift just like a helicopter - by sucking air down from the top and pushing itself up against it. Aerodynamically, it works just like a chopper, but instead of using a rotor and blades, we create the downward airflow electrically by means of ions. Using these concepts, I have optimised the EHD thruster design to maximise the ion flow and its resulting thrust, with the result of a device that not only just lifts its own weight (as in the case of Brown's & De Seversky's units), but to lift themselves together with external payloads which are heavier than their own weight. Shown below is my Spiral Hex Thruster V3.0, which carries payloads of over 100 g.
Click here to view video of Blaze Labs Spiral hex V3 thruster with 100g of payload
(File size 1284kb Format AVI - DivX coding)
Hints for a good EHD thruster:
Description Improvement factor Related link(s) Use lightweight foil - cheapest kitchen foil Payload=Thrust-foil weight No related link Experiment with best foil depth, you should find an optimum depth for each particular design Payload=Thrust-extra foil depth weight EMK Analyser Round edges on top and lower foil edges Required! See calculation on this page Eliminate corners in structure - use straight element panels instead of triangular structure Less corners=less field distortion Coated lifter Insulate any plasma glow foil edges with hot glue or araldite The less glow the better No related link Use fine bare wire for corona wire +10% See Corona losses & efficiency below Electrically heat top wire example ex Tungsten/Nichrome +10% Heated cathode thruster Stack vertically xn for improved performance +100n% Stacked thrusters Apply Bernouilli's effect +50% Autonomous thruster project Use as high voltage as possible +V% Evgenij's maths Dry out balsa supports in humid conditions. Hair dryers work fine. +% No related link Light coat of Teflon on lower foil +10% Coated thruster Use hollow carbon fibre rods for bigger thruster +% No related link Minimise ion impact area (horizontal face area of foil) +100% Horizontal plane foil device Isolate corona wire from balsa to minimise leakage in balsa +50% Use small glass tubes Modify air pressure within thruster +n x 100%, n=pressure ratio Modified pressure thruster Use a pulsing signal instead of dc 0% , no improvement despite rumors Other websites
Note balsa is highly flammable, and sparks may trigger balls of fire reducing your EHD thruster to ashes!
Corona inception voltage CIV
Corona generation is a requirement for a functional EHD thruster. Experimenters know that below a certain operating voltage, EHD forces diminish abruptly to zero. No corona equates to no air ionisation, and no ions equate to zero thrust. The most important voltage parameter we are concerned with is known as the Corona Inception Voltage CIV, the voltage at which the corona is incepted and becomes visible to the naked eye (in the dark) as a thin glowing layer around the wire. Once CIV is reached, the electric field at the surface of the wire is equal to Ei and the corona can be seen and remains visible up to a radial distance at which the voltage gradient goes down to the air breakdown field gradient E0, a constant equal to 3MV/m. When the inception voltage CIV is reached, the air surrounding the conductor will start ionising the surrounding air, and ions start their journey towards the collector, thus initiating the EHD mechanism. The CIV can be calculated from Peek's equation:
CIV = m0 E0 δ ( 1 + 0.0301 /√(δ*rw) ) rw ln (d/rw) = m0 Ei*rw*ln (d/rw) Volts.
- E0 = 3E6 V/m = Air breakdown field gradient at STP
- Ei = E0 δ ( 1 + 0.0301 / √(δ*rw) ) = Corona Inception Voltage Gradient V/m.
- m0 = The wire roughness factor
= 1 for polished wires
= 0.98 to 0.93 for roughened, dirty or weathered wires
= 0.87 to 0.83 for cables
- δ = air density factor
= 3.92 * (barometric pressure in cm) / (273 + temperature in C)
= 1 at STP (76 cm pressure, 25 degrees C temperature)
- d = distance between surfaces in m.
- rw = radius of wire in m.
So, at E0= 3E6V/m, STP (δ=1) and smooth wires (m0=1), Peek's equations can be simplified to:
CIV = 3E6 ( 1 + 0.0301 / √rw ) rw ln (d/rw) = Ei*rw*ln (d/rw) in Volts.
Ei = 3E6 (1 + 0.0301 / √rw) ... hence Ei depends ONLY on physical wire radius.
It is important to note that at this point, the wire's effective radius ro is now effectively greater than the physical wire radius rw since the glowing air surrounding it, is now acting like a radial extension to the wire.
ro = Ei/E0 * rw = rw * (1 + 0.0301 / √rw) ... depends ONLY on physical wire radius
For example, with an air gap of 30mm and a wire radius of 0.1mm, we have:
CIV=Ei*r*ln(d/r) where Ei=3E6*(1+0.0301/√r)
Ei=12.03E6 V/m and CIV=6.86 kV
Effective wire radius ro = rw * (1 + 0.0301 / √rw)
ro = 0.0001*(1+0.0301/√0.0001) = 0.4mm
Near sharp points the electric field may become very high, as indicated in the figure below. Cosmic rays and sunlight from above and radioactivity from the earth below produce free electrons in the air, also near the sharp point. There are two situations to be distinguished:
- a sharp electrode at negative potential, and
- a sharp electrode at positive potential.
(1) Point electrode at negative potential
In the strong electric field near the sharp electrode a force, F = eE, acts on the free electrons, which are thus accelerated away from the sharp electrode. When these electrons collide with nitrogen or oxygen molecules in air, they may ionize these molecules, resulting in a new free electron and a positive ion. These new free electrons can also contribute to the ionization process. The result is an electron avalanche directed away from the point electrode. Because the electric field decreases fast as a function of the distance to the sharp electrode (see the dotted curve in the figure below), the foremost electrons soon arrive in a region with an electric field that is too low to gain enough energy for ionization. The electrons drift with a relatively low speed and attach easily to neutral oxygen molecules, resulting in slow moving negative ions. A negative space charge is formed that decreases the electric field just in front of the sharp electrode. The positive ions increase the field very close to the point, but the ionization region is drastically reduced and the ionization process stops. The negative space charge now drifts away to the positive electrode, the electric field in front of the point electrode recovers, and the ionization process restarts. Because this recovery is only dependent on the time the negative ions need to reach the positive electrode, corona is observed as bursts of ionisation that are equidistant in time.
Sharp edge at negative potential
(2) Point electrode at positive potential
If the point electrode is at positive potential, free electrons are accelerated towards the point and cause ionization. Now it is positive space charge in front of the point electrode that decreases the electric field directly in front of the point, see figure below.
Sharp edge at positive potential
Further away from the sharp edge, however, the electric field is increased, extending the region of ionisation towards the other electrode. Thus, in contrast to the sharp edge at negative potential, the ionisation is enhanced by the space charge, and not decreased. This results in large discharges towards the other electrode. It is further observed that the breakdown voltage for a positive sharp electrode is lower than for a negative point electrode. Apart from ionisation of the gas, oxygen and nitrogen molecules will be raised to higher energy states through excitation by collisions with free electrons. On returning to lower energy levels, visible light is emitted, producing the coronal glow. Ultraviolet light is also emitted. The ionisation is also the source for shock waves in the air that produce the characteristic hissing sound.
Photos showing difference between corona set around a +70 kV wire (left) and a -70 kV wire (right), with respect to ground zero potential. The photo on the right is a close up of how the corona streamers actually build up around each wire.
Corona losses & efficiency
One of the important issues in the EHD thruster is the power required to generate the corona at its ion generating mechanism. Corona discharge is a low energy electrical discharge with non-thermal ionisation that takes place in the pressure close to the atmospheric, if at least one of the electrodes is of low radius of curvature, such as a wire of small diameter. This kind of discharge is self-sustained and no external energy, other than the electrical, is needed to sustain the ionisation processes and to maintain the current flow. Two discharge regions can be distinguished in the interelectrode space: the ionisation region which is confined to a small cylindrical volume around the thin wire, and the charge accumulation region, in the remainder of the interelectrode space, in which the ions drift due to the electric field without additional ionisation. No thrust occurs within the ionisation region. The electrical power needed for corona generation depends much on the voltage drop within the corona glow and this must be minimised in order to keep the ion generation mechanism as efficient as possible.
The corona voltage drop Vd= Ei * rw * Ln(ro / rw )
Maximum corona power loss = Vd * IMAX ... where IMAX= 50uA/cm for wire to trough
Eo= Electric field strength at breakdown voltage = 30kV/cm at STP
Ei= Peek's Corona Onset electric field strength
rw= Corona wire radius
ro= External plasma boundary radius (effective wire radius)
Vd= Voltage drop in corona
Hence for wire radius of for instance 0.1mm, 0.01mm, the voltage drops are 1664V, 739V respectively. For an EHD device powered by a 50kV supply, these voltage drops equate to 3.3% and 1.5% respectively, which are equivalent to the power loss percentage in each case. For negative corona, the effective external plasma radius can be approximately twice as much as that for positive corona, thus resulting in higher power losses. This is one of the reasons why changing from negative to positive corona gives slightly better performance. The change is most noticable with bigger wire diameters.
Plasma boundary thickness for different polarities
It is thus clear that positive corona is preferable, and that although thinner wires give less voltage drops, and less power losses, the advantage of decreasing the wire diameter will be less effective for wire radius below 0.1mm. For this reason, investing in carbon nanotubes for corona wires may not be worth the cost and trouble. One effect that might however be worth investing in, is the effect of temperature on the critical voltage required for corona generation. In Experiment 07 we investigated the influence of the corona wire temperature on the EHD thrust. It was found that the critical voltage Vc required for the corona-discharge breakdown is inversely proportional to the corona wire temperature T, and that the current variation is proportional to voltage variation. Thus the electrical energy required for corona generation is inversely proportional to the square of the wire temperature T. Thus, the electrical energy consumption for the corona-discharge mechanism decreases significantly as the temperature increases. So, the corona generation by means of a heated wire is much more efficient than that in a non heated wire.
Calculating collector radius of curvature
The collector must have a rounded edge with radius of curvature big enough to avoid corona inception, otherwise, ions of opposite polarity will be generated at the collector and will eventually result in thrust cancellation. The following is a small formula that gives you the minimum radius of curvature required for good operation, based on Peek's corona onset voltage:
Collector radius of curvature in mm = V/[3(1+0.03 rw-1/2)]
where V is the voltage across the electrodes in kV, and rw is wire radius in metres.
For example, for a 30kV powered thruster, with corona wire radius of 100um
Collector radius of curvature in mm = V/[3(1+0.03 rw-1/2)]
So, the rounded edge of a proper collector should have a minimum of 2.5mm radius of curvature.
Corona Streamer breakdown & its effect on maximum pressure and current
One common mistake most people make when designing an EHD thruster is to assume a breakdown electric field strength of 30kV/cm and use this value to calculate the spacing between the electrodes. In practice we find that the leakage current begins to dissipate a good percentage of the electrical input at a field strength as low as 6kV/cm, and it that spark overs (short circuit) occur at 10kV/cm, both values depending on relative humidity. So what is the 'classical' 30kV/cm value? This value is the electron avalanche breakdown process as described by J. Townsend in 1910, which is the well accepted breakdown electric field strength for dry non ionised air in a uniform field. If one however analyses the EHD thruster case closely, it can be understood why this value cannot be assumed. Since the air within EHD thrusters contains a high percentage of ionised air molecules in a highly non uniform field, it is obvious that it is now more conductive than dry non ionised air, and the electron avalanche mechanism no longer dominates the breakdown, leading to a much lower average breakdown electric field strength. In fact, the breakdown in ionised air occurs by means of a totally different mechanism than the electron avalanche, namely by corona streamers. Streamers are ionisation waves which can propagate as narrow channels through regions where E < E0.
Different types of breakdown mechanisms and their respective breakdown curves
In the above plot, you can see the conventional Townsend avalanche breakdown voltage E0 at sea level is approximately 30kV/cm, but once we have corona streamers in action, this drops to either 11kV/cm for negative streamers, or even worse, about 5kV/cm for positive streamers. The photos below show streamer formation from a positive tip to ground plane at atmospheric pressure within a 25mm gap at voltages of 12.5kV and 25kV respectively.
Streamer formation at Average E=5kV/cm and 10kV/cm for positive tip
Note that although we are applying an average of 5kV/cm, the electric field gradient at the corona wire is much higher than 30kV/cm. This self-propagation as we know, is due to highly nonuniform electric fields which result from significant gradient in current density, or space charge. If we keep on increasing the voltage, the streamer will be long enough to propagate across the whole air gap, and result in a spark, or short circuit, nulling all effective EHD thrust. Unless the power supply limits the output current, the arcing will be sustained.
Improving thrust by modifying ion drift velocity
The ion cloud accelerates and acts against the neutral air molecules, in a similar fashion as a parachute acts against air whilst going down. The time required for a parachute to reach the ground depends on its cross sectional area, gravity, density of air, shape of the parachute and the weight of the person. All these simple principles can be applied to EHD thrusters, with a analogies to each parameter. In the case of EHD thrusters, the time required for the ions to reach the collector depends upon the ion cloud's collision cross section (averaged over all possible orientations of the ion cloud), ion mobility, charge state, and the EHD thruster operating parameters (electric field strength, field shape, air gap, air pressure and temperature, and mass). Using our parachute analogy, it is easy to visualise why pulsed dc would be more efficient than pure dc at the ionisation mechanism. Note that the ionisation mechanism here is no longer assumed to be the same as the accelerating mechanism, as with the simple EHD thrusters shown in my EHD thrusters collection. The width of the ionisation pulse is equivalent to the thickness of the parachute material. We know that the effective force acts upon the surface, and that depth of the material adds only to weight. Same reasoning applies to EHD thrusters, where most of the thrust is generated at the cross sectional interface between the ion cloud and neutral air molecules. Having a thicker ion cloud won't help. The thrust relationship, given by F=id/k, shows that for the same current and airgap size, the thrust increases as mobility decreases. Also, from Power=I*V, we get a Thrust to Power efficiency relation of F/P= d/v * 1/k. A simple yet very important relation, showing that the only means of improving efficiency is either by somehow modifying the electric field gradient or/and the effective ion mobility. Here we will discuss the parameters which effect the ion mobility k.
For a given mass and charge state, the more compact ion clouds traverse the gap more quickly than elongated ones. It is sort of jumping down with a closed parachute! For maximum thrust we therefore require elongated type clouds (an open parachute). The reduced drift time of an ion as it drifts through the air gap can be described by the following equation:
where Ko is the mobility at 760 torr, 273 K, tD is the reduced drift time of the ion, E is the electric field strength, L is the air gap distance, P is the buffer gas pressure, and T is the buffer gas temperature. Thus, an increase in drift time will mean a decrease in mobility, which results in an increase in thrust. As shown clearly by this equation, higher thrust can therefore be accomplished by several external conditions, namely by higher operating pressure, lower temperature and bigger air gaps.
The time a parachute takes to reach the ground depends on its cross sectional area - the bigger it is, the longer it will take. Small, compact ions with small collision cross sections drift more quickly than large, extended ions with large collision cross sections. This is similar to the effect that causes an extended paper towel to drift to the ground much more slowly under the influence of gravity and air resistance than a crushed towel of the same mass.
In EHD thrusters, the drift time is also a function of the collision cross section Ω and obeys the following relationship:
where ze is the ion's charge, kb is Boltzmann's constant, mI and mB are the masses of the ion and buffer gas respectively, which for normal operation in air can be assumed equal, and N is the number density. This equation shows us further ways how to improve thrust (by aiming at the lowest tD. Highest density gases work best, the more massive the molecules (and ions) the better, and most important, that for a given electric field strength E, thrust increases linearly with cross sectional area of the ion cloud Ω and air gap size L, both of which can be modified by the actual design of the EHD thruster.
A word about health hazards
The air we breathe is made up of mostly oxygen and nitrogen. Ozone can be made from common oxygen in high electric field gradients or electric discharges (as in a thunderstorm). The high voltage gradient which normally generates a corona glow around the sharpest conductors breaks the two oxygen (O2) atoms apart. These oxygen atoms are extremely reactive and they recombine in groups of three with the resulting molecule being called Ozone (O3), or trivalent oxygen. Ozone is a pungent, colourless, toxic, unstable form of oxygen. Normally, it is present in low concentrations wherever there is oxygen. The ozone layer in the stratosphere, at an altitude of about 35 to 40 miles, has concentrations of 10 to 20 parts of ozone per million (PPM) of air. Normal down drafts and other atmospheric disturbances bring some of this ozone down to the surface of the earth, where concentrations seldom exceed a few parts per million.
EHD thrusters (lifters, ionocrafts..), Electrostatic precipitators, copy machines and arc welders are devices that produce some ozone as a by-product of their intended function. The curve below shows the mole fraction distribution of ozone around a positive corona wire at currents of 2.55uA/cm with input wind velocity of 0.5m/s. Ozone generation for negative corona wires is ten times greater than for positive corona, with a wider distribution curve. Gas generation also increases monotonically with corona wire diameter. In both polarity cases, ozone is primarily generated in the corona plasma region. It is convected by the airflow in the x-direction and primarily by diffusion in the y-direction. Little ozone is transported in the y direction becasue convection prevails over the diffusion of the ozone in that region.
Courtesy of Junhong Chen
Ozone can be injurious to health when reaching certain levels. It can have undesirable physiological effects on the central nervous system, heart and vision. The predominant physiological effect is that of irritation to the lungs, resulting in pulmonary edema. The normal dc corona operating in air generates copious toxic ozone, oxozone and the even more toxic oxides of nitrogen more copiously than the desired ions. Furthermore, unlike the ions, these toxic compounds do not decay. Incidentally, it should be noted that ozone in concentrations of 8 ppm, within 5 to 10 minutes produces as many chromosome breakages as 200 roentgen dose of X-rays. So, even if we have shown that EHD thrusters do not emit X-rays, they can still promote lung cancer with inhalation of the gases they produce.
Like many substances, ozone has advantages and disadvantages depending on its intended use and concentrations. Concentrations of ozone, as related to adverse health effects, influence allowable exposure time. High levels of ozone can be tolerated for a short period of time, or low levels of ozone for a long period of time, however one must keep in mind that the health hazard is cumulative.
OSHA (Occupational Safety and Health Administration) has established 0.1 parts per million (PPM) by volume of air, 0.2 mg/m3, as the maximum allowable safe concentration of ozone for an 8 hour industrial exposure. The curve shown below shows the ozone production rates for different corona currents. EHD thrusters maximum current configuration (wire to half cylinder) limits the current to about 50uA/cm (just below the health hazard limit) for 40kV thrusters, but cellular configuration or higher operating voltages may easily exceed the established safety value to hazardous levels. Ion imbalance is another issue, however, a properly designed EHD thruster will neutralise most of the positive ions generated within its accelerating field upon their exhaust from the thruster, usually by allowing enough length of collector material, and/or utilising a sharp lower edge to generate ions of the opposite polarity which eventually neutralise the charge of the accelerated ions.
Ozone, being a form of oxygen, mixes with and decomposes in air. Decomposition is more rapid in higher humidity. The amount of ozone, therefore, that will be in the air depends upon the rate of generation versus the rate of neutralisation and decomposition. Another factor affecting the amount of ozone is the dilution of the air. The above statement applies only to an airtight room. For this reason, EHD experiments should be carried out in a ventilated room or premises. For powerful EHD thrusters, ozone production may become an issue even when operated in an open space. In such a case, several mechanisms can be incorporated to reduce the ozone production to a safe level, details can be found in the following patents:
6,603,268 Method and apparatus for reducing ozone output from ion wind devices.
5,010,869 Air ionization system for internal combustion engines.
4,789,801 Electrokinetic transducing methods and apparatus and systems comprising or utilising the same.