Concepts for chemical reaction monitoring

[ProdigyChild]

Introduction
Since a few years electronic sensors emerge on the market that include the complicated steps of transforming the physical value in an easy to use signal.
For example: instead of using a thermocouple and having to struggle the really difficult process of amplifying a tiny signal amplitude to usable voltages, calibration, mapping to celsius, one could use an IC like LM75 ( http://www.national.com/ds/LM/LM75.pdf ). This IC is a temperature sensor that can be read out digitally and the result is the celsius temperature.
Similar sensors can be found for RGB sensors. These yield an absolute value calibrated at the factory.
These sensors are EASY TO USE even for people who are not experts in electronics. Of course there are still analog sensors that require much more knowledge in electronics (e.g. MLX93247 thermopiles for remote IR temperature measurement, see: http://www.datasheetarchive.com/search.php?q=MLX90247 ).

Interfacing to a computer opens up endless possibilities, if the computer can control the reaction, too. A separate thread on how to control chemical reactions by a computer will be the next step provided there's enough interest on this topic.
Also, the 'engineers' in this forum will have to provide interface schematics, boards and assembly instructions to let the 'pure chemists' take advantage of the electronics. Since I'm more an engineer member than a chemist member and I'm willing to do so this should not be to much a problem ;)

Challange
My question to you is:

what sensor to use
for what physical/chemical value
and how it works (theory behind and calculations required)


Example: ph-measurement using color sensors and an pH indicator
pH measurement is possible using special electrodes using a nontrivial amplifier. Too expensive to buy, too difficult to maintain, too complicated for an amateur to build himself. Why that? Every amateur chemist has determined pH quite a few times, so why not mimic this process?

For example, let's use phenolphthaleine, a preferably green LED and a photodiode. The green LED shines through the solution under observation onto the photodiode. Once the pH rises, the phenolphthaleine turns pink and absorbs nearly all of the green light and the photodiode 'sees' less intensity.
Similarly we can use an easy to use digital RGB sensor like Avago's ADJD-S371-Q999 that can detect color directly in conjunction with a more complex indicator solution.
I haven't tried these methods yet, it's just theory. But I can hardly imagine it won't work.

[megalomania]

What exactly do you want to control? Heating/cooling? Reactant addition? Stirring? When a reaction is done? Personally I was thinking last month I would love to have a webcam set up to monitor a distillation with a motion alarm that goes off when liquid starts dripping. This beats standing around like a fool literally watching the pot boil (necessary because switching flasks to catch different fractions could happen at any time, but you have to be there when it happens).

I have two temperature controllers now, one free and the other under $30. These are great for controlling other "things" (whatever you can imagine) because they send simple on off signals depending on your temperature preset. I want to use one of mine for an electric furnace to control the temperature. The temp controller controls a solenoid that in turn turns the power to the furnace on and off. Simple, easy, fun :)

I would say pH meters with computer interfaces are not dreadfully expensive (about $150 at the low end, plenty of used models too) and are well worth the investment if you really need it. There are very inexpensive digital pH meters that could possibly be modified to send their output signal to some electronic device that can read the signal. This might take some electronics know-how, but I imagine there is info available that can translate digits to computers.

Now if you are talking about building some sort of combinatorial synthesizer, that would be a wonderful project. I rather enjoy the concept of microchemical synthesizers, and I think they could be made available very inexpensively if someone bothers to design such a thing. On the small scale you can achieve very high pressures, high temperatures, use exotic catalysts (cheap when you use a tiny amount), get high intensity UV or visible light, and use inexpensive solid state microwave generators and ultrasonic emitters. Throw in some peristaltic pumps with solvents and hook up a few cylinders of gasses and one could make just about anything, albeit in small quantities.

I wrote a short story a few months ago about what the future might be like if microchemical synthesizers become cheap and ubiquitous.

My first thought when I read the title was "the cheapest way to monitor the progress of a chemical reaction is to spot some TLC plates." That's not exactly what you mean though.

I have been trying to wrap my head around how to monitor chemical reactions in a microwave. I am still looking for a cheap fiber optic temperature sensor and a microwaveable pressure transducer. The inability to monitor the temperature during a microwave reaction rather limits the range of experiments that can be adequately performed, and not knowing the pressure of a sealed vial can be dangerous. If I had 10 grand sitting around I could buy a CEM microwave reactor with all the fancy computer controls. I refuse to believe these things cost any more than a few hundred dollars to make. I will figure this out one day.

[Positron]

EDIT: Crap, I can't edit the post title? SIMPLE DATA ACQUISITION INFO ON THE POST AFTER THIS ONE.

----------------------------------------------------------------------------------------

ProdigyChild,

I commend you on the great post. Things like this are what we need to help kick off what I think is a great new section on The Forum.

OK. Other than 1). Carbon Dioxide 2). Hydrogen and 3). Carbon Monoxide sensors, I am not familiar with any sensors that directly measure other discrete chemical compounds. There ARE ways to measure anything, it's just a matter of how much time and money it takes to do it.

Here's a list of some other sensors (a general term) and transducers (a very specific term referring to the direct changing of a process into a measurable output.)

Color: As mentioned before, with a source of white-light (LED!) and a single chip, you can convert Red, Green, and Blue light to corresponding voltages. For those reactions where liquids or gases change color, a chip can tell you exactly what color, and how much of it. Could be extremely valuable. And YES, if applied correctly, it will work. 100%.

Oxidation-Reduction-Potential: You guys will know a lot more about this one than I will. I believe these "ORP" transducers (XDCR's) measure the small voltages present on specific electrodes during a chemical reaction.
See:
http://en.wikipedia.org/wiki/Oxidation_reduction_potential
http://en.wikipedia.org/wiki/Table_of_standard_electrode_potentials

Pressure Transducer: As a reaction proceeds within a vessel, the pressure builds up to "X" level at which something can be turned on or off. This also applies to a vacuum. Apply vacuum to a volatile solvent within a vessel, seal the vessel, then watch the pressure come up (indicated by a voltage) as the substance evaporates.

Optical: As previously mentioned, sensing an amount of light that is present. Shine a laser through a glass reaction vessel onto an optical sensor, to determine the opacity (clarity) of the liquid OR gas inside. If the liquid or gas is luminescent (bio or chemiluminescent) to some degree, measure it!

Temperature: Many Resistor-Temperature-Detectors (RTD's) easily achieve accuracy's of 1/100th of a degree Centigrade. Thermocouples are cheap; the accuracy isn't as good (+/- a couple degrees C), but they have huge ranges.

Gas Sensors: Oxygen and Carbon Dioxide/Nitrogen Dioxide sensors are easy to get. Commonly used in automotive exhaust systems. Also available are Carbon Monoxide and Hydrogen gas detectors, which simply give an on/off signal if the gas is present.

Other Types: Liquid-level, PH, magnetic, infrared, humidity, acceleration, force, etc.

Electrical Conductivity: It may be quite easy to determine relative concentrations of liquid and gas solutions based on how conductive (or resistive) these are to electric current. As an example, pure water (deionized) is an excellent insulator. Add a little salt, and you can turn on a light bulb through it! The current (Microamps, Milliamps, Amps) tells how much salt has been added. Keep the measurement time reasonably fast to avoid the effects of electrolysis. :cool:

[Positron]

Hi All,

For those that don't know, getting data from sources and storing this information to a computer, is referred to as Data Acquisition, or DAQ. The "data sources" can be volt-meters, current-meters, temperature probes, pressure sensors, etc.

I wanted to mention a product from a company called Dataq (dataq.com). For a grand total of 25 bucks, you can record four analog voltages that span between -10V and +10V to a file on a computer. The file can be imported to Excel if you wish, but the included software for the unit is quite excellent. $25.00 gets you the four analog inputs (at a "recording speed" of 240 data points per second), and a Serial-Port interface.

You WILL need to convert the output of the sensor to a voltage that the DAQ unit can read. The software creates a graph on the screen that is scaled with -10V at the bottom, and +10V at the top. Yes, the graph can be scaled to a smaller voltage window, but you'll lose resolution (it will get blocky).

I am not affiliated with Dataq in any way, other than that I'm an extremely pleased customer. I recently tried out their "DI-148U" USB DAQ unit for $50.00. It can record eight voltages, simultaneously, at 240Hz. My new laptop doesn't have a serial port, so I had to spend the whopping extra $25.00 to go to USB connectivity :).

Anyhow, 240 samples per second is PLENTY fast enough for most chemical reactions (ignoring detonations :D). A trick that Dataq uses to make its money, is that they have some extra software that will allow recording at data-rates of up to 14,400 samples per second :cool:

Truly great products. Just wanted to mention it.

[megalomania]

That Dataq equipment is much cheaper than the products at Omega.com, of course Omega has some very sophisticated devices. I see today they are hawking a Wireless Sensor System Web-Based Monitoring, http://www.omega.com/ppt/pptsc.asp?ref=zSeries, starting at $95. I was looking for some optical temperature measuring systems, and now I know they cost over $2000 not including accessories like the IR probes.

I saw a journal article last year about using real time NMR sampling to plot the course of a reaction. I can't even begin to fathom how they continuously and automatically run samples through a NMR and evaluate the spectra in real time. NMR accessories are expensive as it is, this setup must cost a fortune. I have seen GCMS systems that continually sample a reaction, this is much easier to do since samples can be directly injected into the GC.

This time of sampling makes me wonder if there is some sort of testing methods that could be conducted on extracts of a reaction. Perhaps chemical additives that cause color changes or alter conductivity depending on the presence or concentration of a desired compound. Such reactions would be undesirable in the reaction flask, but on samples it would be fine. Small sample sizes might make it easier to use optical measurements.

I think, at least for organic reactions, that most forms of reaction monitoring beyond temperature control is rather useless. There are very few reactions on the lab scale where pressure, pH, conductivity, etc. are useful or measurable. How useful is it really to have an optical instrument? I actually can't think of one reaction where that would be useful. There are reactions that change color, sure, but that is in no way indicative of the completion of the reaction. If you pour through the literature you will observe almost exclusively that chemical reactions are evaluated by time and temperature.

What I would like to see is a schematic of how a thermocouple based temperature sensor can control the power levels of a microwave. Microwave reactions do not work well with the domestic microwave method of pulsing the microwaves from full blast on to off. Domestic microwaves with adjustable "power levels" actually turn the magnetron rapidly on and off, and lower levels merely keep it off slightly longer, but it still outputs 100% power when on. Since the chemical reactions tend to cease completely when the microwaves cycle to off, and to react too fast when on, laboratory microwaves always stay on, but change the outputted wattage.

I am trying to figure out how a temperature controller can decrease the power to a magnetron without shutting it off. This seems a bit trickier, at least to my limited electronics experience, than simply turning the thing off completely.

[ProdigyChild]

About a year ago, I've disassambled an old microwave oven and could not resist to power the magnetron a 'little bit'.

It turned out that - YES - such a conventional magnetron can emit less than full power provided you simply don't give it full voltage/current :)

The biggest problem (to me at that time) was the heating of the cathode. It runs at 3.4V and approx 12A. The anode needs kilovolts.

I can see 2 possibilities of power control.

1. Lower anode voltage / current.
This requires modifying the high voltage supply or even build one from scratch. Probably not everybodies deepest desire.

2. Reduce cathode heating.
Simply use the HV from the original oven, but put a low resistance potentiometer in series with the heating. Be careful as the anode is grouded and the cathode is actually highly negative with respect to ground. That means the potentiometer is at HIGH VOLTAGE. Since the resistance of the cathode heating is about 0.25 ohms it won't be too difficult to improvise such a potentiometer using a steel wire.

However, Mega, do not expect the output of such a contruction to be constant power. It will be modulated by the 60Hz of the AC line because the circuit is too simple.
If constant power is required then only a good, adjustable high voltage supply (up to 5kV) will do the job. Up to 1000V is quite straight forward. DC voltages above require more uncommon diodes or cascading techniques.

Electronic Power control
Let's assume we use cathode heating power control. In order to avoid using opto-couplers (5kV ones!) we ground the cathode and put 5000V on the case of the magnetron. Don't touch any more any casing ;)
Now connect the schematic attached to be able to control the magnetrons heating with a DC input voltage. This was the difficult part.

I used to use the AD597 IC to amplify the thermocouple voltage.

Another OPV will be necessary to compare the current temperature to the set-point temperature and generate a useful power control voltage.
Also some dampening must be implemented to avoid temperature oscillations. This depends on the reaction vessel properties.

[megalomania]

Modifying the oven for constant output has already been solved by Stefan Binder at the Art ‘n Electronics website some years ago. He has a page up about converting domestic microwave ovens at
http://stefanbinder.privat.t-online.de/uW.htm

I bought this broken microwave oven cheap at a flea market (high voltage capacitor was shorted). After replacing the capacitor I put a burning toothpick into the oven to produce those cool ball lightning-like plasmoids. Didn't work too well - I rarely managed to produce any plasmoids. I think it's because of the low oven power (500W).
But putting evacuated vessels (light bulbs, fluorescent tubes, old HeNe lasers) inside the oven cavity worked nicely - except that now 500W was too much power and the vessels melted very fast. So I decided to make the oven power adjustable. The microwave oven's own power selector doesn't work as it just turns the magnetron on and off for variable amounts of time (some seconds). The food with its high thermal capacity doesn't mind, but this way of power reduction is not suitable for plasmas. The easiest thing would have been to simply connect the microwave oven transformer (MOT) to a variac. But this way I would not only have adjusted the cathode voltage but the heating voltage, thus the filament temperature, as well. So I kept the original MOT to heat the filament and used my 4 MOT power supply for the cathode voltage. This solution has another advantage - instead of the original half wave doubler the magnetron is now powered by the full wave rectified output of the four MOTs - which makes the plasma more stable. Of course when powering the magnetron with 4kV now, power is much higher than before, but because of the robust construction of magnetrons this shouldn't do much harm for short periods of time - as long as it doesn't melt.


The site does not exactly include schematics, but it has pictures of the thing. I wonder if connecting 4 transformers might be a little overkill. I am not exactly familiar with what having so much extra voltage would do to the microwave output, does the high voltage help it run constantly, or does it actually increase the wattage of the magnetron?

I think automatically controlling the power output of the magnetron might be somewhat difficult, at least from my perspective, but a simpler method would be to manually adjust the power until an equilibrium state of a chemical reaction is achieved. Like a variac control knob, the power can be dialed down while watching a temperature meter to regulate the heating of a reaction. A big problem with microwave chemical reactions is at full blast it just gets hotter and hotter, so being able to control the power to moderate a reaction is a very important feature.

The hideously expensive CEM systems use microprocessor controlled temperature measurement and power output control to sustain a reaction at a set point temperature. I imagine the microprocessor calculates some Fourier wave equation by rapidly raising and lowering the power until a steady temperature is achieved.

My unmodified domestic microwave demonstrates very clearly why these things are unsuitable for some experiments. I had a reaction constantly boil over because the wattage is just too high. Dialing down the power to a defrost setting just means the reaction goes from dead to boil over as it turns off and on. I don't want the microwave to stop emitting microwaves, and I only want full power during the initial stages of heating from room temperature where the dielectric loss factor is lower. Since the tan delta of a reaction is temperature dependent, less and less power should be applied to level off the rate of heating.

While I am sure a system could be made to regulate the power automatically, the nature of microwave chemical reactions, especially in a modified domestic microwave oven, means it should be watched like a hawk. Best to have a manual power control knob just like regulating the power of a heating mantle while watching the thermometer in an ordinary reaction.

[BlackFalcoN]

I think the key for achieving such a device is to replace the controller part of the microwave.

The way most microwave ovens work is by varying their duty cycle according to the power setting. This means that the "defrost setting" for example, will run the cavity magnetron element on full power(100% voltage) for 10 seconds, then do nothing for 10 seconds (0% voltage), then repeat the cycle, until the count down timer runs out.
This method is called Pulse Width Modulation (PWM) and works with a square (block) wave signal. The standard control circuitry works great for defrosting food, since food is not very critical about it's temperature. Therefor the basic power settings use a long period for enabling/disabling the magnetron (10 secs - 15 sec - ... )

Cavity magnetrons are not designed to run efficiently on a different voltage that they were designed for. They need a fixed high voltage to work efficiently.

So instead of changing the amplitude (voltage) of your signal, you alter the frequency of the current signal. The average output will be the smouthed out result of
a series of high and low pulses.

The duty cycle of a signal is the ratio between the 'uptime' of the signal compared to the 'total time' of 1 period of the signal.
For example, if you have a PWM with a duty cycle of 75% @ 5000 Volts, then the average voltage of 1 period of the signal = 3750 Volts.
( 75% of the time @ 5000 V, 25% of the time @ 0 V ). 0 through 5000 Volts can be achieved by just varying the duty cycle for 0 to 100% while never altering the top voltage of your square signal.

Look at the attached image for a graphical aid for understanding PWM.

What you want to do to achieve a much more precise temperature control, is replace the control circuitry that translates your selected power setting on the microwave oven, into a PWM signal. ( since even the lowest power setting changes too slow to achieve the desired level of control )

All that is really needed to design your own much more precise control unit to switch the magnetron on/off is a micro controller that will generate a new digital PWM signal, a power transistor or HV relay to switch the high voltage circuit (since you can't just hook up the 5V logic from the microcontroller to a kV appliance ;) ) and a potentiometer that will read a variable analog value selected by the user to represent a certain power setting

It's not very expensive either:
25$ for an Atmel168 Board (you don't want to design your own PCB and solder the whole thing, flash bootloaders into EEPROMs and design a programmer when you can buy a fully assembled unit that works straight out of the box via USB for under 25$.)

Look into the 'Arduino Diecimila', which is an open source project based around the powerful Atmel mega168 MCU and is designed for amateur developers with a very easy and user friendly programming language. It can even be programmed in native C for more experienced power users.
And it is a one time purchase, since one microcontroller can be easily used in other projects as well.

A salvaged power transistor and a cheap variable resistor can be literally had for pennies.

Heck, you could even hook a cheap analog thermistor to your micro controller, and let the whole thing run on automatic if you desire.
The microcontroller probes the thermistor a couple of times a seconds, and changes the output of the magnetron accordingly.

I never done this in practise, so is there anything about my design that would not work according to other members ?

EDIT:
Not really sure, but you probably also might want to add additional cooling for your cavity magnetron element if you intend to run it at a duty cycle of 100% for extended periods, since they will heat up internally too, but were never designed to run at that setting.

[ProdigyChild]

Most likely, there is a certain threshold of voltage that is required to compensate the intrinsic losses of the magnetron tube. Once the input voltage is slightly above this threshold the oscillation builds up and some RF is emitted although at a terrible efficiency close to zero. But this is not Mega's problem (the efficiency at low output powers) - he simply wants a way to reduce power. It's not that bad if the magnetron is inefficient.

On the other hand, if varying anode voltage was inefficient then all domestic microwave ovens were inefficient. Their anode voltage varies sinusoidal from 0 up to 5000V and down to zero again. Varying anode voltage is the typical usage! At least with the one diode voltage doubler supply found in the 3 ovens I've opened so far and in Stefan Binder's .

My electronic suggestion was to tune the cathode's emmision of electrons and keep the voltage high. Adjusting the cathodes heating power in the schematic is done with PWM at a frequency well above the heatings reaction time (seconds).

Stefan Binder's solution provides a more stable HV supply providing 100Hz ripple instead of 50Hz and the possibility of further smoothing using a capacitor.
This can be beneficial if oscillations of 50Hz/60Hz are not tolarable for a chemical reaction. However, I can't believe that 50Hz (=20ms) are too slow.

If someone suggest using a microcontroller for control tasks then I say YES!
As to the type of controller I'm hooked on the ARM7 LPC2000 types (NXP). Don't like the Atmel parts neither ATmega nor SAM7. Too f...ing much of code needed to enter the main() for the SAM7, really. PLL, clock distribution,....

[megalomania]

I was just reading about the Arduino and its group of open source modders yesterday at Daily DIY, what a coincidence. If these are that cheap, perhaps some automatic control could be devised after all.

Indeed modifying a microwave in certain ways does require additional cooling of the magnetron. At the very least a stronger forced air cooling fan, and at the extreme end water cooling of the fins similar to water cooling blocks used in computers.

One of the benefits of monomode microwave reactors is a much more intense microwave field with less wattage. If the field is tuned right, and the sample is precisely positioned at a node, a 100-300 watt magnetron is more than sufficient. I believe the rule of thumb is a monomode microwave is 4 times as power as a multimode microwave.

If I could figure out what the dimensions are of the CEM corp monomode research microwave reactors that use self tuning circular waveguides, I would much prefer to build something like that. Their patents include schematics, but no measurements. The design is very simple, just a central cylindrical cavity irradiated by 7 vertical slits in the inner wall of a cylindrical waveguide that makes a 360 but the end is closed off. The hard part is the height and radius of the wave guide must be exact to sustain a monomode field. The height, width, and spacing of the slits must also be very precisely placed.

I have a caliper, and I know where there is a CEM Benchtop... I just need a little quality alone time with the thing :)

[Positron]

1. If the microwave isn't powerful enough...I'd recommend buying a bigger microwave. How fast do you want to raise the temperature? Pawn shops have more of these things than they know what to do with...kilowatts too...

2. Cutting slits, tuning waveguides, making antennas...is probably more complicated than necessary for this application. Sounds like redesigning the wheel to me, but I haven't read all of the posts.

3. BlackFalcon had it right with the PWM theory, except that I think that two chips from Radio Shack could be used to generate the ON/OFF signal instead of a microcontroller. I'd use an NE555 in astable-multivibrator mode to produce a triangle wave that's about 1Hz (peaks on the waveform are one second apart) in frequency. Then, take the triangle off of pins 6 and 7, and compare this with a variable DC voltage that's created off the power supply with a small PC-board-mount potentiometer. Use a slow op-amp to do the comparing...LM741, LM358, TL082, TL072 are common part numbers. Internet search terms: 555 PWM generator, PWM signal generator, PWM circuit, triangle wave PWM. The entire circuit could literally be built with parts from Radio Shack.

---------------

I think that the solution to this entire problem would be to use an off-the-shelf temperature controller.

Type in Temperature Controller on Ebay and you'll see what I'm talking about.

Slap a thermocouple on its input, and slap a relay on it's output (it may already have one internally) to control power to the high voltage transformer that powers microwave. Inside the microwave, connect the primary-side (120 or 240VAC) to the mains voltage, running one leg THROUGH the relay on the output of the process controller. That way, the controller in the microwave is completely bypassed. If you've got it right, shorting the wires that go to the relay together will turn the magnetron on.

Once finished, you use buttons to tell the controller what temperature to hold the reaction at, and it regulates it there.

Ex. the controller says "shit...the temp's too low", so it clicks the relay on, powering up them maggie, getting the chem's hotter.

Controller says "shit, the temp's too high", it clicks the relay off, cutting power to the maggie, so the reaction cools off.

Most controllers will click the relay once per second or so, but in this case it's irrelevant.

These controllers are capable of regulating the temperature of the reaction within a couple of degrees or better.

The relay contacts need to handle the current that powers the microwave. A 240VAC 20-amp relay will do the trick.

You'll need to work out how to get the thermocouple inside the microwave cavity without bringing the radiation back out to the outside World. I'd run the thermocouple into a 100% sealed copper tube (1/4" refrigerator tubing, or small brass tubing) that dips down into the solution you're trying to heat. Solder it all the way around the wall of the oven where it enters. Use a 150W iron and paste flux (C-flux is awesome from hardware store).

[megalomania]

No, just using a temperature controller is not a solution, as I wrote earlier in this thread. These reactions need to be constantly exposed to microwaves to maintain the reaction rate. I would say the simplest way would be to manually dial down the power to achieve the right temperature.

Since the dielectric loss factor of a material increases with temperature (the hotter a material gets, the more microwaves it absorbs) some materials need to be hit with higher power at room temperature, and then gradually get less and less power.

Having a temperature controller turn a microwave on and off is no different than what a domestic microwave does now. If you set the power level of a microwave to 50% it stays on half the time, and is off half the time. Your reaction goes from frothing out of control to no reaction at all.

For better temperature control you need to be able to fine tune the power output to suit the sample and sample size. A rapid on and off cycle just spikes the temperature and plummets the temperature. Worse yet, a microwave transparent solvent might stay at a high temperature for several minutes, and the temp controller will keep it off, but during this time the reaction is not happening because it depends on microwave energy to initiate it.

The best microwaveable thermocouples are supposed to be aluminum shielded probes that make a grounding connection to the metal shell of the microwave oven. I have not yet found where to get such a probe. The important part is to ground the probe to prevent arcing, which gives false temperature readings. My parents once had a microwave back in the 80's that had a plug in temp probe for checking the internal temperature of something like a turkey or other large meat item. Who cooks a turkey in a microwave anyway? I remember wondering how could this work since it is metal. The walls of a microwave are metal too, but they are all grounded.

Building waveguides and tuning them is an extremely difficult engineering challenge made worse by the fact the companies who build such things keep their cards close to their chests concerning the details. The CEM waveguide though is pure simplicity, it is self tuning just by being round, it is just thin sheet metal spot welded together. It's just that the height, width, radius, etc. of the metal has to be exact because of the nature of microwaves at a wavelength of 2.45 GHz. I could spend years running simulations in CST Studio Suite, or I can spend 5 minutes measuring the real thing.

There is a picture in US patent 6607920 (figure 4 and 5) that shows what the waveguide looks like. There is not much to it, they just left out the dimensions. This is like having a new car without the keys...

[Positron]

No, just using a temperature controller is not a solution, as I wrote earlier in this thread. These reactions need to be constantly exposed to microwaves to maintain the reaction rate. I would say the simplest way would be to manually dial down the power to achieve the right temperature.

If I'm understanding you correctly, what you're saying is that the reaction needs the application of low-power microwaves 100% of the time, and that something other than temperature information is needed to control the reaction. Perhaps that "something" is the crazy microscopic analysis (PET) thing that you mentioned earlier.

But, you did say it. You said that the simplest way would be to manually dial down the power to achieve the right temperature. By saying that, you implied that a human would be dialing down the power until the human was satisfied with the temperature. Therefore, I believe you implied that the reaction could be controlled by the temperature.

I guess that's why I still (sort of, now?!) think that a temperature controller would be the solution, because a temperature controller is capable of maintaining the temperature of a reaction to better than .001 degree C, even if the reaction itself is exothermic or endothermic.

1. A microwave oven's internal temperature control circuit can't do that.
2. A human can't do that.
3. A specifically-designed temperature controller would be the right tool for the job.

Worse yet, a microwave transparent solvent might stay at a high temperature for several minutes, and the temp controller will keep it off, but during this time the reaction is not happening because it depends on microwave energy to initiate it.

Yep, from what you're saying, it sounds like temperature information alone is not enough to be able to properly control the reaction. If this is the case, then all bets are off for the temperature controller's capability of fully controlling the reaction. Which means that I am/was wrong about it being a solution to the problem. My apologies if this turns out to be the case.

For better temperature control you need to be able to fine tune the power output to suit the sample and sample size.

Yes, industrial temperature controllers will fine tune the power output to suit the sample and size.

Having a temperature controller turn a microwave on and off is no different than what a domestic microwave does now.

There is a huge difference between the shitty temperature control circuit in a microwave oven (that doesn't even have temperature feedback!!!) and a specifically-designed temperature controller. At least in my microwave oven, there isn't a thermocouple, RTD, or thermistor...

A rapid on and off cycle just spikes the temperature and plummets the temperature.

The rapid application of on/off cycles is a good thing. It's when the application is too slow (such as in a microwave oven) that the thermal mass of the sample can't filter or average or smooth out the temperature changes. The result is temperature ripple.

If the thermal mass of the substance to be heated is very low, then the frequency of that application of energy will need to be increased to decrease the temperature ripple.

The best microwaveable thermocouples are supposed to be aluminum shielded probes that make a grounding connection to the metal shell of the microwave oven.

Anything that isn't microwaveable, can be made microwaveable by placing it inside of a faraday cage.

A "microwaveable thermocouple" is just a normal thermocouple stuck inside of a metal tube that is connected...I mean really fucking connected on all sides with no gaps...to the enclosure of the microwave cavity where it enters.

I have not yet found where to get such a probe.

www.omega.com has some pre-shielded thermocouples, but it's a lot easier and a lot less expensive to just buy some cheap ones and run them into your own copper tube. Copper is easy to solder. The stainless steel normally used to shield thermocouples isn't so easy to solder.

The important part is to ground the probe to prevent arcing, which gives false temperature readings.

Not grounding (a better word would be shielding) any wires or probes entering a microwave cavity will do several things. It will not only add a ton of noise to the measurement, but it can (and probably will) straight up fry the inputs to the circuit from the ridiculously high voltage coupled onto the wires from the electric field inside the oven. At the same time, it will probably re-radiate the 2.4Ghz out into the free space around the wires until the oven is turned off.

[BlackFalcoN]

BlackFalcon had it right with the PWM theory, except that I think that two chips from Radio Shack could be used to generate the ON/OFF signal instead of a microcontroller.

That's very true, soldering your own dedicated controller would probably be cheaper, but you have to take into account that the average chemical engineer/scientist is not an electronics specialist, and a microcontroller allows for very easy additions of extra sensors later.

Also, for 25$, do you really want to spend hours and hours figuring out how to design, assemble, solder and make the PCB ?

The Arduino MCU I mentioned also allows for very easy communication with a PC, so external control and monitoring from a network, or extensive logging of sensory output to a PC logfile are easy to do.

Also, the MCU can be used in different generations of microwave designs, where as Radio Shack components are pretty permanent once you have soldered them.

If you are absolutely confident in your electronic skills, and have performed all the calculations in advance, then I agree, Radio Shack items is the way to go since your project will be somewhat cheaper + you don't have to program any IC's.

If you have an open-ended project, that will grow with time and you want to keep tweaking and adding additional modules to it, go with a MCU.

[megalomania]

A manually controlled microwave power output would work the same as a heating mantle. You can crank a heating mantle up to 70-90 in the beginning to heat up a flask faster, then lower it to about 40 while the reaction starts, and gradually fiddle with the power knob to maintain the desired temperature. Because many materials being microwaved tend to absorb more microwaves the hotter they get, the ability to dial down the power would be of great benefit to sustain a steady temperature.

Some, not all, reactions do require a continual application of microwave energy to sustain them. Many reactions conducted in the microwave do just depend on heat, as comparative experiments using convection heating has shown. Temperature control is still very important because some compounds decompose at elevated temperatures, and some undesirable side reactions.

There is no need for some other type of a sensor to monitor the reaction, just temperature control.

I would think with a better explanation of what kinds of chemical reactions are possible in a microwave, the problem could be solved easier.

1) Microwave reactions identical to convection heating (heating mantle). These reactions work just by applying heat. Temperature control is important to maintain a temperature within a range. This is like boiling a cup of water on the stove, or in the microwave.

2) Microwave reactions in a quenched environment. Using a nonpolar solvent that is transparent to microwaves means a reaction can happen by being zapped with the high energy of microwaves, but will immediately be cooled by the solvent. Another form of this is forcing a stream of gas into the reaction mix to keep it cool. Reactions of this type might involve metallic catalysts that superheat because of arcing, but on the microscopic scale, and the superheating lasts a very tiny fraction of a second.

3) Microwaved reactions in a sealed container. As the temperature rises, so does the pressure, and if the temperature gets too high the container could burst. This is where a pressure sensor would be of great benefit, but that is largely a safety mechanism to scram the microwave after crossing a dangerous threshold. By controlling the temperature, the pressure can stay within acceptable limits since temprature and pressure are related.

4) Microwave effect reactions. The still mysterious application of microwave energy that seems to enable some reactions to happen only when exposed to microwave radiation. These reactions are not reproducible with convection heating, nor when all other factors are equal.

One very important factor that gives microwave reactors an advantage is the decrease of time to run a reaction. This effect requires a constant application of microwave energy for some as yet undetermined reason. It is for this reason I do not want a microwave that pulses on and off. Once your microwave shuts off the reaction is little more than a convection heated reaction. The ability to reduce microwave power output to a minimal level is invaluable.

Trust me when I say this: It is unacceptable for the microwave to cycle on and off; it must decrease its power output instead.

How would you sustain water at a slow boil inside a microwave without it boiling over? You can set it on high for the first few minutes until it starts to boil, but if you reset the microwave to 50% you will notice it goes from boiling over to not boiling at all.

There is a reason chemists no longer use unmodified domestic microwave ovens. There is a reason the expensive laboratory microwave reactors all have adjustable power outputs. There is a reason these microwaves emit microwaves continually, not pulsed on and off.

A regular thermocouple, like those at Omega, can be used, but it still needs to be modified. Grounded is the proper word because that's the word used in the scientific literature that tested and evaluated the best microwaveable thermocouple. There is a big debate right now about using thermocouples, but I don't plan on doing any high level pharma research in a modified microwave to where it would matter. For whatever reason, using aluminum as the sheath gives truer temperature readings. The purpose of grounding the thermocouple to the microwave cavity is to give the electric current that builds up in the thermocouple sheath some place to go other than arcing to the thermocouple.

Evaluation of shielded thermocouples for measuring temperature of foods in a microwave oven.
RAMASWAMY H. S. (1) ; RAUBER J. M. (1) ; RAGHAVAN G. S. V. (2) ; VAN DE VOORT F. R. (1)
Journal of food science and technology, 1998, vol. 35, no 4, pp. 325-329.

Flurorptic techniques are normally used for measuring temperatures in a microwave environment because thermocouples have been generally assumed to cause electrical perturbations and arcing inside a microwave oven. However, shielding of the thermocouple junction in an aluminium tube grounded to the microwave cavity wall has been reported to prevent electrical perturbations and permit measurement of temperatures. Shield heating was reported to be the main cause of error. In this study, several shielded thermocouples were evaluated for temperature measurement inside a microwave cavity and compared with a standard fibre optic temperature sensor. Results indicate that thermocouples can be used for temperature measurement in microwave ovens and shield heating, a common problem associated with these probes, can be by proper design (shield isolation and body insulation) to yield reasonably accurate values of temperatures (errors <2°C). The probes can be effectively used to monitor/measure temperature of food products during microwave heating.

Temperature Measurements during Microwave Processing: The Significance of Thermocouple Effects.
Journal of the American Ceramic Society, Volume 84 Issue 9 Page 1981-1986, September 2001.

Reliable and accurate temperature measurement during microwave processing of ceramic bodies is controversial. Although thermocouples are routinely used in conventional thermal furnaces, their presence in microwave furnaces can locally distort the electromagnetic field, conduct heat away from the sample, induce thermal instabilities and microwave breakdown, and lead to serious measurement errors. These thermocouple effects have been studied and found to be more pronounced in low- and medium-loss ceramic materials. To decrease the thermocouple effects during the processing of advanced ceramic materials, an optical, noncontact temperature sensing system has been developed, calibrated, and incorporated into a computer-controlled microwave furnace.

Microwave irradiation of wood packing material to destroy the Asian longhorned beetle.
Fleming, Mary R. ; Hoover, Kelli ; Janowiak, John J. ; Fang, Yi ; Wang, Xin ; Liu, Wenmin ; Wang, Yuejin ; Hang, Xiaoxi ; Agrawal, Dinesh ; Mastro, Victor C. ; Lance, David R. ; Shield, Jeffrey E. ; Roy, Rustum
Forest Products Journal, 1-JAN-03

Due to equipment malfunction, two different multi-mode microwave ovens with non-uniform field distributions were used for the microwave temperature gradient experiments. The 2.45-GHz Panasonic oven (model 4 NNL728BA) is 120 V AC/60 Hz with 1580 W of input power and 1000 w of power output. The oven was modified with a Powerstat Variable Transformer (Type 3PN117C: input 120 V, output 0 to 120 V, 12A). The second oven was a 2.45-GHz Amana Radar Range (model #RR4DW) with an output power of 1600W. The oven was modified with an Adjust-A-Volt Variable Transformer (output 0 to 120 V). In both setups, three k-type thermocouple wires were threaded through small holes on the side of the microwave. A thin layer of aluminum metal tape was wrapped around each wire and around the welded tip. The wires were grounded by taping the wires to the metal inner chamber walls with aluminum tape. Each thermocouple was connected to an Omega silver-plated selector switch (model #SW142-6-B: 2 pole, 1/2 din). A K-type compensating lead connected the switch to the Omega Digital thermometer (model #2168A). Both of these systems were tested for radiation leakage with a Rahman Radiation meter Model #48lB and a Rahman Probe Model #82 manufactured by General Microwave Corp. In addition, a simple boiling water test was performed to check the accuracy of the thermocouples. The thermocouple tips were placed in a beaker of water and the microwave was turned on. A temperature reading of 98[degrees]C was recorded when the water began to boil in the Panasonic system and 99[degrees]C in the Amana System. This was deemed acceptable.

To test the potential of a loss of sensitivity with aluminum foil/tape shields on the thermocouple readings, one shielded and one bare thermocouple were placed close together in a Kanthal heating coil fumace. The oven was ramped to 400[degrees]C in 8 minutes. Readings were taken every 15 seconds. In all cases, the shielded thermocouples were the same or lower temperatures than the bare thermocouples. At temperatures lower than 100[degrees]C, the shielded thermocouple lagged by 0[degrees] to 4[degrees]C. Between 100[degrees] and 200[degrees]C, the shielded thermocouple lagged by 8[degrees] to 16[degrees]C. From the results of this experiment, we can assume that temperatures recorded during the microwave experiments are accurate at temperatures under 100[degrees]C, but are likely understated at temperatures over 100[degrees]C.

Irradiation of the ALB in China

A common kitchen multi-mode microwave oven with a non-uniform field distribution (Galanz model #WD900) was used in all experiments conducted in Hohhut, China. The input was 1400 kW and output 900 kW The oven was modified by splicing a Powerstat Variable Transformer (Type 226: input 240 V, output 0 to 240 V, 7.5 amps, 50/60 frequency) into the main power cord (Fig. 1). In addition, three type k thermocouple wires were threaded through the grating on the side of the microwave. A thin layer of aluminum metal tape was wrapped around each wire and around the welded tip. Since the inner microwave chamber was aluminum metal, the thermocouple wires were grounded with aluminum tape to the inner chamber walls. Each thermocouple was connected to an Omega silver-plated selector switch (model #SW 142-6-B: 2 pole, 1/2 din). A type k compensating lead connected the switch to the hand-held Omega thermocouple read-out (model #HH-26K: type k, -80[degrees] to 482[degrees]C). A Holladay Industries microwave leakage meter (model #HI1501: 0.1 to 100 mW/[cm.sup.2] at 2450 MHz) was used to test the microwave set-up for radiation leaks. The simple boiling water test as described for the U.S. experiments showed an acceptable temperature reading of 98[degrees]C.

The reliability of thermocouples in microwave ceramics processing.

It is not rare to hear arguments against the use of thermocouples for taking temperatures in processes that are taking place under microwave fields. However, the simplicity of this device makes it attractive to consider its use. One question that arises when thermocouples are employed is whether the electric field perturbs the measurement, and if the thermocouple affects the processing. The process that was chosen for conducting this test was the synthesis of spinel (MgAl2O4) using microwaves as a power supply and hematite (Fe2O3) as an additive for both spinel formation promotion and susceptor. The alumina-based systems are very important to study because this is one of the most common ingredients in refractory materials. There are many discussions about the improvement of the process when microwaves are used, but a kinetic comparison cannot be performed if the temperature is unknown, and that is the reason for emphasizing the measurement techniques. The analysis of the obtained samples was carried out by X-ray diffraction of powders. The results of this work show that there is no difference between the products obtained when the thermocouple was inserted in the system, compared to processing without it; hence the thermocouple is appropriate for this application.

Combined effect of microwave and activated carbon on the remediation of polychlorinated biphenyl-contaminated soil.
Xitao Liu, Gang Yu
Chemosphere, Vol: 63 Issue: 2, April, 2006, 228-235.

The temperature measurement in microwave field is regarded crucial for reactions. Some researchers ( Cuccurullo et al., 2002) believe that traditional technique of thermocouples for temperature measurements cannot be applied inside microwave ovens due to the electromagnetic field and the metal probes interactions, while there are reports that a thermocouple introduced normal to the direction of the electromagnetic field does not influence the electromagnetic field distribution (Liu et al., 1994; Roussy et al., 1995 [ Liu et al., 1994, Roussy et al., 1995]). Menéndez et al. (1999) compared an Inconel sheltered type-K thermocouple and an infrared pyrometer in measuring the temperature of carbon bed during microwave treatment and found that the obtained results for these two techniques were comparable. According to their method, the temperature profiles of soil/GAC mixture in microwave field under different microwave power, GAC amount added, soil mass and moisture content, were recorded by a sheltered type-K thermocouple.

Effect of microwave power

Besides the nature of the material (dielectric properties) being treated, the temperature attained in microwave irradiation depends as well as on the microwave power applied to the sample. The majority of the absorbed microwave power is converted to heat within the material as

(1)http://img177.imageshack.us/img177/8885/si1fh8.gif

where E is the magnitude of the internal electric field, is the relative effective dielectric factor, ε0 is the permittivity of free space, f is the microwave frequency, T is the temperature, t is the time, ρ is the density, and Cp is the heat capacity ( Clark et al., 2000).

For a given material, the temperature can in principle, be modified by adjusting the input power ( Menéndez et al., 1999). One special point to be noted is that the power levels used here are continuous, not just timed mark/space ratio control. The investigated power series was 300, 500 and 700 W. For 20 g of soil, GAC added and water content involved were 1.0 g and 0.54%, respectively.

Thermal Treatment of Active Carbons: A Comparison Between Microwave and Electrical Heating.
J. A. Menéndez, E A Menéndez, A. García, J. B. Parry and J. J. Pis
Journal of Microwave Power and Electromagnetic Energy, vol 34 iss 3, 1999 pg 137-143

Two commercial activated carbons were subjected to thermal treatment in a N2 atmosphere using a microwave multimode resonant cavity and a conventional electric tube furnace as heat sources. The temperature of the carbon bed during the microwave treatment was monitored using an infrared pyrometer and an Inconel sheltered type-K thermocouple. A comparison between both methods of measuring temperature was made. When similar treatment temperatures are used, both techniques produce similar changes in the textural and chemical properties of the activated carbons. However microwave treatment is much less time-consuming than conventional heating. Microwave treatment in an inert environment seems to be an efficient and attractive way of removing oxygenated functionalities from carbon surfaces and of increasing the hydrophobicity and basicity of carbons.

I have dozens more references on modifying domestic microwaves for laboratory use. I am not spouting theory when I make my statements on modifying a microwave, all my information comes from published sources who have actually built and used modified domestic microwaves in their research. The only problem is these people are a little thin on exactly HOW they modified their microwaves. Therefore I know what I want accomplished, it's just a matter of figuring out the specifics.

[ETCS (Ret)]

Magnetrons can be operated in the Continuous Wave Mode by applying well regulated D.C. to the Cathode (Anode is typically at Ground Potential).

An adjustable power supply capable of delivering sufficient Cathode current to the Magnetron over the voltage range of 0 Volts to approximately 3 Kilovolts is needed to find the Operational Modes of oscillation (there will be several) as the voltage is slowly increased from zero to maximum. Some means of measuring the microwave output power (a microwave thermistor or diode, or both) is used to identify the power levels attainable in the several modes, and the Cathode Voltage Range for each mode.

As the Cathode voltage is increased in each mode of oscillation, the microwave power output will increase proportionally until the voltage exceeds the maximum for each mode.

Once the mode which delivers the maximum output is identified, and the voltage range of the mode is accurately measured, then a custom power supply can be devised to concentrate efficiently on that desired voltage range. This will enable the output power of the magnetron to be continuously adjustable from minimum to maximum within the mode, by simply varying the Cathode Voltage.

The output power will be considerably less in Continuous Wave than is produced in Pulsed Drive Operation. It may be sufficient, however, to use as a microwave source for your laboratory purposes.

The microwave oven magnetron may be ideal for this purpose since it is designed to operate well over a variable input voltage range. Making your own Negative Power Supply with sufficent regulation and stability over its range of adjustment is the more difficult challenge.

More difficult, but by no means beyond your skill levels. Or, in the event that it may be, almost any electronics tech or ham radio operator would be able to provide the needed assistance.

The cost would not be high.

[megalomania]

I actually found the perfect journal article soon after I made my last post. It describes in detail the electronics required to make a modified domestic microwave oven into a laboratory oven with adjustable power control, and a temperature controller on/off system. I see now the ideal modified microwave has both a manually adjustable voltage which raises and lowers the outputted wattage, and a temperature controller that acts as a safety system to turn the microwave off when it gets too hot, and back on when it cools.

Computer controls come into play that are very similar to kiln controllers for ceramics. The ability to program the machine to ramp the output of X watts over Y time, hold for a time, then decrease power would be very useful. Ideally the controller would set his power output at such a level as to keep the temperature controller from turning off the microwave as little as possible. I gather this could be computer controlled as well, with the computer trying to achieve as little temperature fluctuation as possible by adjusting the power, and the controller shuts the whole thing off as needed if it gets too hot.

The article is:

Microwave power control strategies on the drying process I. Development and evaluation of new microwave drying system.

W.M. Cheng, G.S.V. Raghavan *, M. Ngadi, N. Wang

Journal of Food Engineering 76 (2006) 188–194

Abstract
A phase-controlled electrical power regulator was developed and connected in series with the original cycle-controlled power regulator of an existing domestic microwave oven. The microwave oven was further modified such that combined microwave and convectional drying can be accommodated. The system performance was evaluated which included calibration of the maximum microwave output power, determination of the microwave distribution in the cavity, establishment of the relations between the output power and the input voltage for the phase-controlled power regulator. It was observed that phase-controlled power regulator could be successfully used for quasi-continuous (fast-switching) power regulation with the maximization of power efficiency. The degradation of output microwave power was recorded and the non-uniform distribution of microwave field in the cavity was also verified.



I have a few other good journals articles I found, but none as informative as this one. I wonder if any of the electronics gurus will shed some light on the plausibility of what this article has to say? This article goes above and beyond with its descriptions of how to gauge the power level.

I read in one of my articles something about a magnetron acting like an induction coil if it is run continually, thus the reason they get switched off is to dissipate the buildup of current and send it in another direction thus preventing undue heating of the device. I gather the modifications detailed in this article causes the magnetron to rapidly switch on and off and 60Hz? Am I correct in my understanding of how this works or not? I should say this type of on off activity is acceptable, probably unavoidable if indeed the magnetron heats up if not shut off; it is the 10-15 seconds of being off part that is undesirable.

[Positron]

Mega and All,

The "phase-controlled electrical power regulator" is nothing more than a light dimmer. The light dimmer in the article is chopping up the incoming 120V power to the high-voltage transformer in the microwave. Each pulse (of varying width...and therefore of varying power...) occurs every 8.3 milliseconds...plenty fast enough for good temperature regulation.

So, as simple a concept as this is, there are still a few caveats.

1. A magnetron requires two things to generate microwave power. One is filament power, which is a low-voltage (around 10V) The other is is high voltage power.

Read in the article that the diac/triac circuit (light dimmer) is cutting the high voltage power on and off to regulate temperature, while the filament stays continuously connected to the power source.

So, you must have separate transformers. One for filament power, and one for high-voltage power. For the high-voltage power you can just use the transformer that comes with the microwave. For filament power, you must source a transformer (of the correct voltage and current rating) for the filament in the particular magnetron that you're using.

p. 191: "For keeping the filament functioning a new filament transformer which had its own power source was introduced to replace the original one."

Also, when a light dimmer (basically a high-speed switch) is switching a transformer, there will be huge voltage spikes formed when the light-dimmer "switch" lets-go of the winding. This spike must be suppressed with a snubber circuit...which is mentioned in the article.

Short answer, yes it will work if those two things are taken into consideration.

[ProdigyChild]

I have my doubts about this simple design. The original circuit contains a voltage doubler - a diode + a capacitor.
Applying a dimmer circuit to a (transformer plus) capacitive load is kind of a short-circuit :eek:. If their setup survived then it was only luck.

It is possible to oversize the Triac but you can't oversize the HV rectifier diode because you want to reuse it. The diode will eventually die from repetitive overcurrent spikes.

Yet another example of published scientific bullshit. Perhaps the other people cited in the paper who did not come up with the high-frequency switching idea were all smart enough to figure out its incompatibility to the voltage doubler circuit :p

[Positron]

The rectifier is already seeing high-current pulses during normal use.

The rectifiers are rated for the surges of current that it takes to bring the capacitor back up to voltage, after the load has drained the charge during the low-line condition. You are correct that yes, there will be high-current pulses, but I believe you're incorrect in that it's such a big deal.

The current ratings (both average and pulsed) of rectifiers are based on the maximum allowable die temperature. Most rectifiers can handle HUGE pulsed currents as long as they don't get too hot. Which, in the situation with the light dimmer + voltage multiplier circuit, I don't believe it will. It's a current-over-time thing.

Rectifiers are tough to kill anway. I have personally melted the solder off of the leads of microwave oven diodes....and had to use gator-clips (a mechanical connection) to hold them onto the wires so that I could continue to use them! (I have piles of them, they're sacrificial).

You were a bit quick to bash the article, in my opinion. Unless you are REALLY good, I recommend refraining from doing such things until you've tried it yourself ;)

Pretty good almost-catch though...

[megalomania]

I found a microwave yesterday referenced in another journal article as being one of the only models with a continuously variable output rather than a pulsed on/off output. The Panasonic NN-T551 has four power settings with the max being 900W. I couldn't find any on sale, and I am not so sure this is the only microwave oven to have this feature. If Panasonic made one, they may have made additional models by now.

Perhaps a maintenance guide or schematic of this model could shed some light on methods to get a variable power output. I have seen websites that offer maintenance guides, but not for free.

This Panasonic model uses four buttons to achieve lower power settings, so they must be avoiding a variac for some reason. Probably cost.

As for the article, they did actually build the system, as has at least one other researcher. It must work, but researchers only publish good news, they rarely publish things that don't work, or that fail. They will, however, publish about other peoples ideas that do fail when they tried it for themselves.

In another modification to increase the applied power of microwave energy for plasma generation...

"Microwave-induced plasma reactor based on a domestic microwave oven
for bulk solid state chemistry" by David J. Brooks and Richard E. Douthwaite; Review of Scientific Instruments, 75, 12, 2004, pg 5277-5279

... the researchers applied a brass box to the inside of the microwave oven attached to the waveguide to decrease the volume of the cavity, and thus increase the applied power of microwave energy to their samples.

I can't attach this file because it is watermarked with the downloaders access information. Basically, it looks like the scientists attached a section of waveguide that fits inside the oven chamber, with a water reservoir at the end to soak up excess microwaves. The water reservoir has cold water continuously circulating through it. They report an increase of 114K to 174K with this modification using inorganic salts with known melting points as their temperature gauge. They reported a maximum plasma temperature of 1400K at 900W with gasses other than argon.

That figure does not seem right for a plasma temperature, it is too low by half to a third I would think. I think what they mean is they were able to heat a capillary tube to a temperature of 1400K with a plasma.

Still this is rather discouraging if that is the plasma temperature. To produce NOx gas from air would require plasma temperatures near 5000K. The plasma temperature may be a function of applied wattage, and 900W is not powerful enough. I reviewed a few other references I have for microwave plasma systems and they all seem to have a power output of 2kW or more.

[ProdigyChild]

Just because one experiment is in accordance with a theory does not mean the theory is correct.

Another: If you want to drill one hole through 5cm of concrete you can use wood drills - especially if you have multiple of them. Most people won't do that.

No, this author proposes an idea each absolutely uneducated person (in terms of electrical engeneering) would come up with. That's fine as long as he does publish such on his private web-site. It is problematic if such an idea get's a scientifc touch (more serious).
That's why I was quite angry about the article and the author and still I am. And one question remains: why has no manufacturer come up with that simple design so far?

Poor designs don't necessarily show their flaws immediately. Failing capacitors on motherboards, exploding batteries in notebooks are extreme examples. Both were not discovered in the beginning. Semiconductors DO degrade. There is no sharp line between operational and broken. How long will the cheap Chinese torch light work properly, that permanently overheats the LED? Long enough to sell that cr*p for sure. But you'll never get the umpteen thousands of hours of life time.

I found on the net some information about microwave ovens and for the first time I read that a MOT has a 'magnetic shunt' built in to prevent short-circuit. I thought only NSTs and OBITs have this. That's why the author was lucky.

Consider me an a**hole but I can't judge 'torturing electronic devices' as good.

[megalomania]

Would you care to explain a little more about what you think the author is doing wrong? I can't exactly see what the problem is as I am not too familiar with the nuances of electronics.

The original circuit contains a voltage doubler - a diode + a capacitor.

What do you think a correct solution would be for this setup? How would you protect the diode so it will not fail from current spikes?

I have been reading more about the Arduino board and its various projects (http://www.arduino.cc/). It seems quite likely that some very inexpensive sensors could be incorporated into DIY scientific instruments to be controlled by a computer.

One of the fancy features of the $10,000 CEM microwave benchtop reactor is a computer programmable scheme to customize power levels and reaction times. This is rather like how kiln controllers ramp and soak pottery. It requires a very sophisticated, and very expensive, temperature controller to handle more than two temp and time presets. With a computer interface there would be no limit to what you could accomplish.

There is an Arduino project to measure VOCs (volatile organic compounds) using a VOC sensor available from Figaro Engineering Inc. (see http://www.instructables.com/id/How-To-Smell-Pollutants/). I noticed they have an ammonia sensor too. Although I didn't see any prices, I gather these sensors are extremely inexpensive.

I could possibly incorporate the ammonia sensor into my ammonia burner for the production of nitric acid. I am already building a temperature controlled electric furnace for it, and the addition of the Arduino could make it more sophisticated. Quite possibly the ammonia sensor could be used to determine the concentration of ammonia vapor from an ammonia generator. I want to use the U2A (urea to ammonia) process to generate ammonia, but the quantity of ammonia produced would vary wildly depending on a number of variables including temperature, pressure, and concentration of urea solution. With this sensor I could have the computer maintain a steady ammonia output by keeping the reaction at an optimal temperature, and it could tell me when it is time to add more urea solution because ammonia concentration would drop.

This sensor could also detect the presence of unreacted ammonia after it passes the catalyst. One of the engineering challenges for this apparatus is finding the optimal residence time of ammonia on the catalyst, to fast and you get unreacted ammonia, too slow and you get nitrogen and hydrogen... Using the Arduino, I could start with a very fast ammonia flow, detected by the sensor, and slowly decrease the velocity. When ammonia can no longer be detected, that's the optimal residence time.

I looked a little for a NOx sensor, but I did not find one (the gasoline exhaust sensor from Figaro detects CO, CO2, and H2). This would be the ideal gas to measure.

[BlackFalcoN]

I found this link while researching this topic a bit more in-depth:

http://www.geocities.com/vsurducan/electro/micro/micro.htm
or
http://rapidshare.de/files/39271986/MICROCONTROLLER_BASED_INTERFACE_UNIT_FOR_5KW_MICRO WAVE_OVEN.pdf.html


It's the design & construction for a continues wave, microcontrolled 5kW Microwave oven with variable temperature control and sensors.

It utilizes the PWM technique I described before. (Which I still think is the approach to take for this project)

It's design and functioning seem pretty solid to me.
What do you all think about the design and do you see any shortcomings ?
( just because it works does not mean I can't be improved for our purposes ;) )

The firmware part is a bit complex to understand at first because they utilize a MicroChip PIC16F877, which is a bit harder to program and flash for the amateur electronic engineer, but as I said before, that could easily be replaced by a Arduino MCU.

[BlackFalcoN]

This patent is also relevant to the topic:

http://www.patentstorm.us/patents/5354972-description.html


Field of the Invention:

The present invention relates to a power supply for a microwave range utilizing a pulse width modulation signal, and particularly to a power supply for freely varying the output voltage.



EDIT:

http://www.agrenv.mcgill.ca/agreng/theses/theses/313ZhenfengLi2004/313ZhenfengLi.pdf

'Design of a Microcontroller-based, Power ControlSystem for Microwave Drying';


Microwave drying is an energy-efficient drying method.

The output power of most commercial microwave ovens is controlled in an intermittent fashion, where the amount of microwave energy is determined by the ratio of “ON cycles” to “OFF cycles.”

To provide a more efficient and continuous power control for the magnetron, a microcontroller-based, feedback power control system was developed.

...

An alternate method of power control is pulse width modulation (PWM). It can offer better output power spectral characteristics than phase control and is increasingly being adopted in modern electronic power converters (Trzynadlowski, 1998).

[ETCS (Ret)]

Here's an approach to varying the Cathode Supply to the Magnetron that is rather simple:

[ProdigyChild]

What do you think a correct solution would be for this setup? How would you protect the diode so it will not fail from current spikes?


I would put an inductor in series with the MOT transformer secondary side, to limit the current slope. Perhaps a resistor parallel to the inductor to reduce the ringing/oscillations. I should do some measurements to determine preferable values.


I looked a little for a NOx sensor, but I did not find one (the gasoline exhaust sensor from Figaro detects CO, CO2, and H2). This would be the ideal gas to measure.

There are those electronic signs in the public that remind us of the air pollution. They show O3, dust, SO3 and NOx so there MUST be a sensor for NOx!
Please note I don't advertise these signs as a source of NOx sensors ;)

EDIT: Just found one at digikey:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=AD81-ND
See datasheet pdf for details.

[FUTI]

Guys I have a question in-line I think with topic but out of current line of speech within this topic. You guys talk about MW reactors and detectors etc. I just don't know where to find filter that cut-off visible light, since I'm making a homemade UV-lamp. All other parts are easy to find and I got them. What are they made from? Where to find them? Is there some homemade version I'm not aware? Thanks. Mods please move this question if you find it misplaced.

[megalomania]

I searched for these "high pass" filters for transmitting UV light and excluding other wavelengths not that long ago. Such filters are sold by Edmund Optics (www.edmundoptics.com). Beware that they are extremely expensive! Even the factory seconds are quite pricey.

I too have been looking into building a UV light source for photochemical reactions. I keep getting stuck on the prices of individual components. It seems to be cheaper to get an aquarium UV sterilizer as a kit rather than get each component separately.

There is a project at the Renewable and Appropriate Energy Laboratory (RAEL) at University of California at Berkeley that is building inexpensive UV sterilizers for third world countries (http://rael.berkeley.edu/old-site/uvtube/uvtubeproject.htm).

For the life of me I can't figure out if a regular ballast for fluorescent lights will work fine on a germicidal UV bulb. I can't seem to find any inexpensive lights with a high enough wattage to run a high powered bulb.

One of the great ironies of the fluorescent tube industry is that the UV bulbs are the same as regular bulbs, except they have no phosphor coating, and the UV bulbs cost 20 times as much. How can something with less material cost so much more? Granted it must be because of the quartz glass used in UV bulbs allowing maximum UV transmission, but only IF the manufacturer uses quartz glass...

Now that I am learning about microwave induction lighting, it seems microwave and UV photochemistry are a perfect match :) With electrodeless UV bulbs, you can zap them in a microwave field and they will emit UV light indefinitely. I have not found any concrete sources that give prices of these bulbs yet. Induction lighting is still a somewhat new concept for the consumer end of things, and lab or industrial equipment is still prohibitively expensive.

I have considered making my own bulb. The idea is not that terribly far fetched either. With some rudimentary glass blowing of a quartz tube you fill the bulb with 5 mg or mercury and argon gas. You can even get the gas + mercury from fluorescent bulbs. I have worked some with vacuum and gas systems in the lab, so I know the rudiments of the setup. There is a very nice setup on the MAKE blog about a guy who built a neon sign DIY system superior to commercial systems. He built it because some professionals in the field told him it was impossible to build your own... I love stories like that.

[ProdigyChild]

FUTI, instead of searching for a filter that absorbs all visible light you can use a dichroic mirror that reflects only the UV-light you need.

Dichroic filters are not that expensive, see:
http://www.edmundoptics.com/onlinecatalog/displayproduct.cfm?productid=1734

EDIT: Hahaha. Mega has overtaken me!

[FUTI]

Guys I envy you sometimes for the fact that you live in the land of plenty, which isn't my story. For example today I asked a company that sell bacterial strains how much would cost to obtain harmless test strain so that students can learn the trade on it. In USA it cost less then 80$(+transport), for me here price is 460$...let's talk about building prices.

I tried also to buy spare filter for UV lamp from the guys that sold my lab a new UV lamp. They didn't even respond to the fu*king e-mail! That is one firm I won't buy anything from ever. I have one half-usable filter from old UV-lamp (which had two of them but they are degraded badly do to I presume thermal stress) that I will use somehow I guess. I was looking about filters for UV light and found about Wood's glass and also this page (http://msp.rmit.edu.au/Article_04/06.html) . Do you think I should try to make that or just stick to what I got now?

And I'm glad that Mega and I share some common interest in chemistry...I will also try to make photo-reactor from the rest of the parts available I have.

[ETCS (Ret)]

Here are a couple of good resources for studying the UV lamp characteristics:

http://members.misty.com/don/uvbulb.html

http://www.negativeiongenerators.com/UV-C_spectrum.html

They both have additional links for in-depth study.

[fractional distiller]

With Regards to a microprocessor that can be used for monitoring and or possibly used for a control in a synthesis process one item of interest would be the microprocessors available to the public through http://www.parallax.com/.

I first encountered the parallax products in my younger years while attending a camp for robotics. The kind where the parents pay outrageous amounts of money for their children to play with electronic components, while college students tell them that they are so smart.

Besides being told that I was a unique and special individual with great creativity as I followed my predesigned diagram with a few variables for "student creativity". I did actually come to realize that the the projects that we were building were completely useless but the technology and tools that could be incorporated with it was not.

The controller I am familiar with is of the basic stamp variety. They offer packages that can be used and maybe very useful for process control. They offer thermocouples and flow meters and may other useful probes that can be used for process control.

[megalomania]

Wow, fractional distiller, that is very cynical, nevertheless absolutely true.

It takes a little digging to get to their prices, and I see you can request a catalog.

[Positron]

Sorry to post this so late. It's been awhile since I surfed through this section.

UV filters are readily obtained in the form of glass envelopes that shield High Intensity Discharge (HID) lights.

There are two bulbs in these HID lights; one is the small, active, light-emitting gas-containing bulb with electrodes inside. The other, is an oversized (to lower the energy density and temperature, I'm sure!) giant envelope that surrounds the smaller bulb in the middle. It's this larger glass envelope that you're after.

I once severely damaged my corneas by using a mercury-vapor HID light that had a busted UV shield. It took about 15 minutes of working near this light to put me out of commission. Trust me, this shield is the real deal.

I'd advise trying to look up the transmittance curves of this glass, so you know exactly what wavelengths it will or will not pass. Keywords will be wavelength, transmittance, UVA, UVB, absorption, HID, mercury vapor, sodium, metal halide.

[megalomania]

That's the WRONG kind of UV filter, Positron. We want a filter the excludes everything EXCEPT the UV light. The UV light is what performs the chemical reaction, but there are too many materials that block it, thus limiting the effectiveness of the reactor. A filter that passes UV while excluding other wavelengths is not so easy to find (on the cheap).

[tapira1]

Did not read carefully the whole tread; perhaps somebody already posted this info before. Here is the way we get UV light for chemical reactions: There are widely available mercury street lamps; they need a transformer to light up. All you have to do is to carefully remove ("destroy") the outher glass envelope which has the phosphor (this turns UV into visible light). The remainder of the lamp, once turned the power on, emits UV light (watch out!!! exposure to UV may damage your corneas and even leave you blind. Use protective glasses. Do not expose bare skin to this radiation either; you may get burned.

[megalomania]

I obtained one of those type of bulbs last year. My problem now is how to power it. I have looked high and low for an economic method of producing UV light for photochemical reactions. Without having to pony up for a real UV reactor, the cheapest I have found so far are aquarium UV sterilizers and a DIY electrodeless microwave UV lamp.

There are some specialty LED UV bulbs that have a low enough wavelength, but they have very little power output. I don't mean the kind of LED lights for UV curing and ROM erasing, the wavelength of that UV light is not short enough.