Microstrip Antenna - Design for optimal field emission

Hi guys,

I have to design a microstrip antenna that has to be integrated in a microfluidic device.
The idea is that the antenna will emit a field that will heat particles by induction heating.
Could you suggest me which is the best design and configuration for an such antenna that will operate between 2-4 GHz (i mean, i can operate in this range of frequency, but the antenna can also be very narrow band).

Thank you very much.
Paolo

Hello Paolo,

You definitely need to supply more info to get a useful answer. Think of: dielectric of fluid to be heated (Er', Er''), fluid volume, distance towards antenna, antenna input power (peak and average), special field requirements (think of uniformity within certain volume), etc.

As long as it is a relative homogeneous fluid with particle lengths that not is resonating with actual frequency (with wavelength compensation due to dielectric properties for the fluid), do I not see any special reason using a high frequency power source for generating EM waves for heating. If your particles are shorter then typical 1 mm can you just as well use a DC source, as heating effect and distribution will be very similar.
Guess you have a signal generator with output impedance 50 Ohm?
A very simple antenna that is having almost 100% efficiency for producing heat is then to use a 50 Ohm resistor. Heat distribution will be more omnidirectional then a dipole or loop antenna. It will radiate EM waves with wavelengths around 20 GHz (IR) instead of 2 GHz.
A 100% well tuned short loop antenna (VSWR=1) at 2 GHz can also be used. Heating effect will be almost the same but 20% of produced heat will emit at 2 GHz and 80% will emit at a bit undefined IR wavelengths, depending on antenna temperature.
An small loop antenna with as much as 20% efficiency can be complicated to build, a 50 Ohm resistor is much simpler.
It is also possible to generate heat indirect by emit sound waves but it is unpractical, if the goal is to produce heat (IR-waves).

I did once in a very serious lab use a 1GHz signal generator, a 50 HZ source and a DC source, each were limited to 1kW.
These power sources were each connected to two nails. Each pair of nails was inserted in each end of a sausage. One sausage for each power source.
How long time it did take to reach 100 degree C in the middle of the sausage were measured. It was about same time.
A bit surprising was the main difference between these methods the taste. DC sausage tasted metallic even if stainless nails was used. 50 Hz was better but 1 GHz had the best taste.

Hi, so here are more info:
- the fluid in which the particles are dispersed is water
- the particles are gold nanoparticles.
- the chamber in which the fluid is contain is a cylinder of 0,5mm (radius) 1 mm (heigth).
- the distance from the antenna is...as close as possible (range of mm)
- i would like the field to be as uniform as possible in this volume.
- frequency range that i can access with my instruments is between 2 and 4 GHz

I was thinking about a microstrip antenna because it is easier to be integrated in my setup...

I didn´t explain the problem correctly, Kafeman, the heat will be generated by induction heating IN the particle in the solution, and not by emission of IR...the idea is that JUST the metallic particle will be heated. DC source will then not work..

Thank you very much for the reply

Hello Paolo,

Assuming non-conducting cylinder containing the sample material and E-field parallel to length, you can have reasonable uniformity when placed between two plates carrying the RF voltage.

The "antenna" will be more like a resonator than an actual antenna. The sample can be placed between two plates with about 1mm separation (to get good field uniformity). Position of sample will be in or close to the point of highest E-field. Matching and tuning doesn't look difficult.

So This challenge doesn't look impossible. Of course there can be difficulties because of accessibility for other instrumentation, very high power or special environmental conditions.

In my opinion, it is that water that is heated directly, not the gold particles (assuming relative low concentration and small aspect ratio of particles). If you want to heat them directly, you may need a wavelength where water is transparent and gold has good absorption.

Let´s say the the bottom part of the sample must be free, because there will be detection with the microscope.
You can gen an idea from the picture, the 1th layer (from the bottom) is a thins glass cover (the microscope will be under this one) the middle layer is the reservoir layer and the last one is PMMA and contains the fluidics. The antenna will be placed ideally on top of this third layer.
My idea was to use a PCB and some sort of patch antenna to be placed on top of the reservoires.
What do you think?

I basically agree with E Kafeman's assumption, that there will be no noticable induction heating effect. Having an aqueous fluid, considerably dielectric heating is much more likely.

But I may be wrong. However before designing any antenna, did you calculate the required AC magnetic field to achieve the intended induction heating? After you determined the numbers, we can discuss about antennas.

In some papers that i found they stated that the absorption will be present but negligeable...i´m also skeptical about that...we will see.
It is very hard to calculate the necessary field because I will use nanoparticles and the interaction at that length scale are rather tricky to guess.
Anyway my instrumentation can go in range 2-4 GHz and the range of power is max 10 W (output of the amplifier)..
I would probably say that...the higher the better for the heating purpose... am I wrong?

PaoloDellaVedova said:

In some papers that i found they stated that the absorption will be present but negligeable...i´m also skeptical about that...we will see.
It is very hard to calculate the necessary field because I will use nanoparticles and the interaction at that length scale are rather tricky to guess.
Anyway my instrumentation can go in range 2-4 GHz and the range of power is max 10 W (output of the amplifier)..
I would probably say that...the higher the better for the heating purpose... am I wrong?

Click to expand...

Do a search on dielectric properties of water and you will see that water has high dielectric loss at 2..4 GHz. If there is substantial amount of gold particles, they will increase the E-field in the water. One will notice this via an increase in effective eps' and eps".

Regarding the setup, I can't match your cylindrical volume with the picture and text you provided.

As you have a small sample volume, you need an electrically small antenna/transduceer to avoid that your power is radiated. Using antenna concepts designed for communication purposes will likely not behave as you want. If you need significant part of your 10W as heat into the sample, you need a complete other approach.

If you look in the middle section of a chip there is a circular hole, then, being the chip a 3D object, the resulting volume is a cylinder.
Water as for sure some absorption but, this is not the point here and doesn´t change my question either: which is the confiduration that will give rise to maximum field?
Anyway..what does it mean a complete new approach? if you could be a little more pagmatic...

Water as for sure some absorption but, this is not the point here and doesn´t change my question either: which is the confiduration that will give rise to maximum field?

Click to expand...

As a first point should be considered, that we are talking about wavelengths in a 10 cm range. Geometries that are so much smaller than wavelength, that the field can be analyzed as pure electrostatical respectively electromagnetical phenomenon. This means that it would be reasonable to generate a magnetical "induction" field by a coil rather than a microstrip. That's my comment about "optimal field emission" for mm dimensions.

Secondly, I'm missing considerations about intended dynamical behaviour. Thermal time constants particle-fluid and fluid-container can be expected in µs up to a few 100 ms order of magnitude. So if you intend anything other than very short transient heating, you can heat the container as well.

Hi, thanks for the reply. So:
-First point, can you explain me why the coil would work better then the microstrip?
-Second point, the timescale is exactly what u said, and I´m interested in such phenomenon, so the idea is a very short and localized heating just around the nanoparticles.

why the coil would work better then the microstrip

Click to expand...

I'm applying common knowledge about AC magnetic fields and treat the problem at an intuitive level. It's an assumption or hypothesis so far, it should be founded by a quantitative analysis.

It should be obvious, that a tuned microstrip antenna is radiating it's energy to the free space, but generates only moderate near field E and H strength. A coil can be designed to generate a local H field without radiating much energy. A similar solution exists for generating small electrostatic fields between two electrodes.

As the volume is very small (<< wavelength), you can't heat it effectively with a communication antenna (that will have a size of > cm). You need to get the highest field at your sample.

A small resonator designed to couple efficiently with your sample will give the best results given your available input power. As coupling to the sample may not be very good, the Q factor of the resonator will be relatively high (resulting in the high E - field to get the heating). The above is what I had in mind w.r.t. "complete other approach"

The water IS of importance, as that will absorb the RF power. If the water had loss in the range of PTFE, PE, PP, etc, it would be practically impossible to get your RF power into the sample.

I assume you have in mind how much energy you want to supply to the sample in certain time. You may calculate the E-field inside the sample that is required (you need Eps" for that). This you can convert to E-field in air (as D is continuous across an interface).

As you are talking of nano particles with good conductivity in a lossy medium with size >> "nano", I don't think you will be able to heat the particles by magnetic induction heating without heating the lossy medium. To benefit from the E-field induced by a time varying magnetic field, you need area (Vinduced = A*dB/dt).

You can make some approximate calculation. Other method is to disperse your particles in a low loss medium. Put this dispersion in the H-field maximum of a resonator and determine the reduction in Q factor. Next disperse same amount in water and do the measurement again. Based on the drop in resonator Q-factor you can assess the dissipation due to the nano particles and due to the water.

As already stated by others, @ 2 GHz is water a much better absorber then gold. Gold is in the other end of the scale, a low loss metal and a good reflector. Nano scale does not change that to the better from your view.
Thin gold foil is used as heat reflector, for example is such foil used as blankets to keep people warm at accidents (reduce body temperature loss) so gold does not absorb heat well in IR/Red wavelengths. Bad news that it is hard to heat gold, if the idea was to heat gold in water, without heating any water.

By using a very strong magnetic or electric field for induced heating in near field, yes that works for any metal, it is commonly used for melting metal at foundry's. Limitation factor is that the metal melting pot must have reasonable size or enough electric loss compared to frequency wave length. At 2 GHz is it however not really nano size. A lot of bad news now.
It is also possible to make an electrolytic solution if ratio of gold ions relative fluid solution is big enough. It will result in a kind of heating due to chemical movements of these ions. Result is still radiated heating in IR range due to friction and gold will probably have a slower temperature raising then the fluid around it.
In short, your requirements are killed by rather basic knowledge.

Any method, it is a good idea to use deionized water and very pure gold surface.
A classical test to prove this is to heat two glasses of water in a microwave oven, one half filled with regular drinking water and one glass half filled with deionized, desalinated water. Place them both at good distance from each other in the oven as they else will contaminate each other with vapor. Now start the oven. Deionized water will not boil visible even if the regular water boils until the glass becomes dry. Be careful as overheated water can explode if it becomes contaminated!!
This experiment shows you practically about your skepticism if water is an absorbent at microwave oven frequencies. Water is not ideal but without doubt an absorbent. It also gives some important information about vapor, maybe needed in future experiments, explained a bit down.

Now the gold.
Place also a few of your gold particles in between these glasses.
Five minutes in the oven and the only heating that will occur is not due to 2 GHz waves from the oven, it is due to IR radiation from boiling water and that will be a very moderate heating.
You can easy try this in any microwave oven if you want to convince your self.
With this knowledge, what do you think happens if you place gold particles directly in water in a microwave oven? Correct, particles will be heated by water.
Similar thing as when a I leave the spoon in my coffee mug, when heating coffee in the microwave oven. The spoon will be warm, and even warmer a short period afterwards.

Do not know why it is important to first heat the gold part or why water must be used as solution or what size these particles have or dispersion thickness which can make my assumptions wrong, but with rather basic thinking, I had used a frequency where gold with this size works as a good absorbent, and water as a less good absorbent. A too obvious idea maybe? Lets check anyway...

Assume a nano-particle length of 100 nm. A dipole antenna with 100 nm length is in the range of UV light. A gold particle with this size will therefore be a good UV absorbent. Water can by good reasons be assumed to be a less good absorbent at this frequency, as it is very transparent at visible frequencies and if you have been diving in a clear sea you had probably noticed that blue tint seems less attenuated then IR/Red therefore seems UV to be a good frequency that is meeting your requirements. But no sunshine without sunspots. Direct heating of gold in water will cause water vapor rater quick as now is it the gold that is heating water. Vapor is rater different compared to water and acts as UV isolator, causing radiation chaos and blocks direct heating of the particle and instead is more water heated. Even at room-temperature will a thin vapor layer cover the gold surface. There is however a thermic transportation delay until too much vapor is produced which gives you a short period when it is possible to heat the gold particle more then surrounding water.

If you want to reach temperatures above 100 C or even 1000 C and have some control about what you is doing, must needed amount of power over time now be calculated. Useable pulse length can be in range of maybe max 0.1 nS depending on how high temperature you want to reach and how pure your water and gold particles are, so this solution is maybe not possible if it requires too much power during a too short period. It is relative simple to calculate power requirements if you know weight and size of your particles for a given particle-area were heating should occur during a short time frame. Knowing this can we now sort out if radiation pattern and intensity of the beam is in a reasonable power-range for for example common available UV-lasers. Time-frame can be increased by avoiding vapor.

According to this proposal, generated heat have not reached the particle by IR-radiation, instead is UV-radiation used for transferring of energy which makes is possible to heat the particle more then surrounding water.
If a short enough time frame is used for the UV pulse, is amount of direct heated water very low compared to heat generated in the gold particle.

DC current and two etched probes at 100 nm distance in your water solution will also work without directly heating water, but with less good taste as my previous experiment showed.

Ok, got it! it looks pretty impossible to heat the particles without heating the medium @ GHz range..
So it look to me that what I wanted to build is a microwave oven, right?
Anyway, let´s say that I don´t care about that right now...doesn´t matter what I will use it for, how do I build a device (not an "antenna" then) that will create an high EM-field in a localized area close to the device itself?

If RF heating is considered, it sounds more promising to go for dielectric than inductive heating, according to the mostly congruent analysis that have been presented yet. Of course, resistive heating would be a option too.

Homogenous dielectric heating can be best achieved with plane parallel TEM cells, I think. They can be scaled to any size according to your needs.

No antenna => no radiation.

Amount of magnetic energy that is possible to pick up is related to size and impedance of receiving absorber/antenna.
Water absorbs magnetic fields gradually. Estimate that about 50% of total energy can be absorbed in 50 mm thickness (deep) of pure water. If your "water antenna" is less deep than this will higher amount of energy travel through this medium without producing any energy in the water.

Magnetic coupling between TX source and water is not what I want to call effective in this case. TX antenna will also be fare from effective so total system efficiency is very low. Dominating heat factor will be due to resistive heating in any case and even that part will be relative ineffective.

We do yet not know much about circumstances around this experiment but you want expertise answer?
My impression is that similarities with a microwave oven and heating water in microwave oven and basic physic knowledge about material properties for this experiment was beyond your research. My question is then if that part not is important, how did you do find out which frequency and dispersion to use?
Maybe a bit rude but I can not see the full picture. Required type of heating system/antenna is a bit complicated and require some experience and knowledge and knowing how to handle some tools such as vector network analyzers. So you need to do far lot more research and it will take very long time if you is waiting for someone delivering optimized answers around unclear questions.

Anyway, lets go on with the microwave oven and learn something how it maybe can be (mis)used.
If you place a glass of water in the oven you now know what happens? As you probably also know is water no absolute absorber. I did in text above estimate that a signal will attenuate with about 3 dB if it travels through 50 mm of pure water and absorbed part will heat the water. What happens with remain not absorbed signal?
Most of the signals will reflect in walls of the oven and sooner or later hit the glass of water again. Each time will the wave become weaker and the water warmer. It will be some losses also in the walls but they are relative low. However will also some signals reflect back to TX antenna (magnetron) and produce heat in it. It is a part of total reflection in the TX antenna and similar problem discussed about optimizing an conventional antenna. To much heat in the magnetron can destroy it, so that is why it not is recommended to run a microwave oven without something absorbing the energy. TX power can be typical 700 W so if all power is reflected back can it cause a lot of heat at wrong place.
The nice thing however is that even if your dispersion contains relative very low amount of absorbing material, can it be compensated by a lot of re-reflections, were each reflection heat it a bit.
Your amount of water is to small as RX load to run the oven within its safe limits but it is nothing that says that you not also can have the glass of water as main absorber. Most of the absorption will then occur in the glass but it will anyway be a much better result than what you can achieve with a 10 W signal generator and a loop antenna in free space. If you is brave and want some excitement, omit the glass with water leaving only your dispersion in the oven. It can destroy the magnetron but it is also rather common that it survives a lot of bad handling. All your water will be vapor in very short time and gold particles spread around the oven. Burning CD's in an oven is popular at youtube for example and the oven seems to survive in most cases. If you do this experiment, do it on your own risk. Check some videos at youtube to understand what to expect. One more thing to think about, very close to the walls are electric fields weak, due to that walls reflect waves but also causes polarization shift which cancels out the fields near walls. This effect is due to that metallic walls are seen as mirrors by the waves. Therefore, for better heating, place the sample at least 10 mm away from any wall or metallic conductor.

You did also wanted to document something with a microscope.
Do not know, but if it is a simple optical microscope, it will maybe not fit inside oven (not joking) and it can contain parts that absorbs energy, but remove plastic decoration in the top of the oven, drill a 3 mm hole in the top, and place microscope above the hole focusing 10 mm below. Place the sample at right position inside oven with aid of parts of removed plastic decoration as that plastic probably have low carbon content and then is less affected by microwave heating. Moderate amount of heating in you sample by adjusting amount of water in the glass.

A 3 mm hole in oven is it dangerous due to radiation? Yes, you can get burnt. As the hole is small compared to wavelength is it however very low amount of radiation that radiate in a such small hole. If unsure, do a practical experiment. Hold a sausage against the hole and measure time until it becomes too warm to hold. It will not happen. I personally would not try to place my eye close to the hole but it is no risk looking with a microscope due to distance. There is however a very small risk that if the first lens of microscope is placed very close to the hole or maybe even partly inside oven, and if this lens is connected to a bigger metallic structure of microscope, can metallic surfaces of the lens be cooked. Most microscopes do not have any metallic surface for anti-glare or color-correction at this end but it is possible. The lens is normally glued in a plastic seat but small iron spring can also be a part of fixation or whole metallic tubes which can be a bit heated but it is not likely.
Start with a filled glass of water and you is safe.

Hello Paolo,

Do you check your PM folder time to time?

Reason: I sent you a PM (19 sept) in response to your request, but didn't get a response on that.