Dielectric Properties of Water Emulsions

Dielectric constant of water for RF oven cooking can be found in many research papers.

http://www.jmpee.org/JMPEE_PDFs/09-4_bl/JMPEE-Vol9-Pg303-To.pdf

http://microwaveheating.wsu.edu/publications/dielectric%20property-03.pdf

Thank you for the reply. Though I'm not measuring at these high frequencies (sorry, I didn't state what range I'm using), the term "dielectric loss" seems to be important in the links you gave instead of "dissipation factor" that I was searching with. This will probably lead to better search results.

Dielectric constant is a measure of the ability of a substance to store energy when imbedded in an electric field. Water has a high dielectric constant because the positive (hydrogen) and negative (oxygen) align with the electric field. On the other hand, dissipation factor is the ratio of a substance's resistance to its capacitive reactance, a measure of how much energy is dissipated.

"Dielectric Constant is used to determine the ability of an insulator to store electrical energy. The dielectric constant is the ratio of the capacitance induced by two metallic plates with an insulator between them to the capacitance of the same plates with air or a vacuum between them. Dissipation factor is defined as the reciprocal of the ratio between the insulating materials capacitive reactance to its resistance at a specified frequency. It measures the inefficiency of an insulating material. If a material were to be used for strictly insulating purposes, it would be better to have a lower dielectric constant."

http://www.intertek.com/polymers/testlopedia/dielectric-constant-and-dissipation-factor-astm-d150/

Thank you, however the question was more about why the dissipation factor goes up with homogenized mixtures instead of a single drop of water even when the total water content is the same. In my research paper, I have to attempt to explain this but I have a hard time coming up with scholarly explanations.

I would think it comes down to the area of exposure, several small droplets would have have much more surface area than a single large droplet...

https://piers.org/piersproceedings/download.php?file=cGllcnMyMDEzU3RvY2tob2xtfDNQNV8xMzM3LnBkZnwxMzAzMjAxOTM3MjU=

Hi SolarEagle:

It is a very good research paper for reference and there is theoretical prediction as well experimental results for comparative verification.

Exactly, I think that's what it is as well. That is a better article than I've found so far, but it still doesn't attempt to explain what I've observed while measuring with "low frequencies":

This is just a generalization of what I see happening. The top line is the well homogenized mixture. The bottom flat line is simply a water drop inserted with a microsyringe. The measured water content is based on an estimation on the bottom line (based on the volumes) and with a karl-fischer titration instrument for the top line (very accurate).

So something is happening. But I don't know what! It makes sense that if you have individually dispersed water molecules surrounded by dissimilar molecules, it might be harder for them to change orientation, thus increasing the dissipation factor. But as to the physical reasons for this, I have no idea.

I will admit that I am not familiar with this research and have not taken the time to read the linked articles...that being said, here are my to cents.

water is polar and also conductive (when in solution with minerals). If you are dealing with a water solution the more homogenized the solution the more conductive path for charge dissipation you have. My experience with electric charges in fluids are from static buildup in pipeline fuel transfer. The more sediments in the fuel the more static is built up; but if there is entrained water in the fuel it would allow the static to dissipate to the sides of the tank faster...but we are not in the habit of introducing water to fuel intentionally.

Drew K

Well maybe it has something to do with the pattern of the field, or the density....You state "low frequencies", as compared to higher? What effect might the frequency have on either of these? You just need to run different varied tests to see a pattern emerge, which will then lead to an answer.....hopefully anyway.....

My guess is that up to about 2% water content you have more and more small drops. Above that, they maybe coalesce into larger drops lowering the total surface area. Water molecules on the surface of a drop behave differently from molecules inside the drop that are completely surrounded by other water molecules.

A microscopic examination of the emulsion might verify if this is the case and you could calculate the total surface area exposed and see if it correlates to the measured dissipation factor.

Just a guess, not my area of expertise.

Exactly :) I think that's where it goes from "dissolved" water to emulsified water. Corresponding with the plot, that ~2% figure happens to be where the mixture becomes milky! Interesting, isn't it?

So we observe these measurements roughly correlating to visible changes. I actually tried to have someone at the university use this micro x-ray CT scanner, but he has yet to actually do it. We also have an NMR imager at the university but it's too expensive to use. So unfortunately we just haven't had a chance to do anything yet. Using a microscope might work if we can squeeze the sample thin enough.

Yes, smaller droplets mean both higher surface to volume ratio and smaller radius of curvature of the water droplets.

But can you think of any reasons why this might change the dissipation factor? It sounds like we've come up with different ideas, but nothing that is "scholarly" enough to put in an article.

For dielectric constant in pure water dispersion, the issue is active surface area. When you emulsify water in to a bulk liquid, the surface area to volume of water increases and it is the surface area that mainly affects the capacitance. There will be coupling and blanketing effects as the concentration of microdroplets increase and geometry will be the major effect. If the droplets end up contaminated with something that affects polarity, the contamination level also needs to be considered.

I got into this with capacitive wood moisture content measurement. You end up in a particular situation with building an empirical chart or table that goes out the window if the geometry and dispersion of water diverges from the development samples, as in the moisture content is high enough that settling or floating of the moisture causes uneven distribution of the water. Leached minerals, such as rust, zinc and once I ran into an issue with leached copper and lead from a bullet that went through also have effects.

I'll try to write a collective reply.

Drew K, that is insightful. The dissipation factor could be higher due to a better conduction path with many more tiny water drops. Maybe that's the simplest explanation as to why the DF is higher when it's well homogenized? It allows for a better conductive path, which results in a higher DF while the actual capacitive value remains the same.

.

Solar Eagle, sorry for not stating specifics. We've had problems in our research group at my university where people get unhappy when specifics are disclosed so I'm extra careful... Anyways, I have run enough tests to come up with the plot I gave in a previous post (DF vs. %water). I used two characterizing test cells (coaxial and parallel plate) and a sensor prototype and observed a similar effect in all of them. The application was intended to be as simple as possible, which meant that using MHz or GHz to measure wasn't desirable. Using very low frequencies had a number of advantages due to the unique properties of water in the lower range. I've read there are something like 8 different effects which rise and diminish at different frequencies which all contribute to the dielectric properties. It separates into dipole pairs due to the increased amount of time between cycles and allows for a measurement we used called "dielectric thermoscopy." Here is a plot of the contributions to the dissipation factor I found in some research I came across (unfortunately it refers to a textbook from 1973):

I thought that was interesting, but unfortunately doesn't really explain the properties in my first plot that I shared. Also, I can't find the original book so I don't know what all the lines represent.

.

Rixter, I think that the surface/volume ratio is the key, but I'm not sure if there is more to it than simply what Drew K stated above.

.

Jpfalt, it's interesting because what I've noticed is that the dielectric constant doesn't seem to change with the emulsification state at the frequency that I'm using. It stays proportional to the amount of water. What changes is the dissipation factor with the emulsification state.

That is very interesting about the minerals that you found in wood. Do you mind if you share what type of measurement system/conditions/parameters you used? It's interesting because this could be very similar in terms of problems encountered.

I had thought the case was for a DC situation rather than at different frequencies. With two plates and an insulator in between, the dielectric constant determines the size of charge, or the value of the capacitor. For instance, in a DC circuit, if you take two plates with dry air in between, the capacitor will be of a certain value and hold a specific charge. If you take the same plates and put oil or plastic in between, the capacitor will have a different value, determined by the dielectric constant of the material put between the plates and will hold a different specific charge.

Dissipation in the DC circuit doesn't come into play. The capacitor in the DC circuit takes the charge and holds it.

In the DC circuit, if the water emulsion in the oil has a certain surface area, then each droplet forms it's own capacitor and the overall capacitor becomes a network of microcapacitors in series and in parallel based on the shape and location of each water droplet. This is why in the original post, the dielectric constant measured was vastly different for oil with emulsified water and for oil with a single drop hanging in the oil.

When you throw AC at various frequencies, the capacitor network response will be dependent on the moisture geometry and the resulting capacitance of each droplet with the additional resistance and heating response of the current running from end to end in each droplet with the additional effect of ionic contamination of each water droplet on the resistance value of the droplet.

Maybe with very low frequencies, it's close enough to DC that the same/similar effects apply to what you have described. But in the original description, I intended to describe the relationship with the dissipation factor, but I didn't make it very clear.

I like your description of the network of microcapacitors.

However, I come up with this question in my mind: Will this have something to do with the spaces in between the droplets, or rather the interface between the droplets and the surrounding material?

I think that the total capacitance ended up being similar with the same average water content, whether it was largely a droplet or a homogenized mixture. It was the dissipation factor which changed. (unfortunately this is hard to test, because injecting a known and reproducible amount of water in the capacitor is difficult).

But the way you describe the possible reason why the dissipation factor has increased with a homogenized mixture might not completely make sense, because the DC impedance is still in the gigaohm range, even with a few % of water and uncoated, fairly large, and close together metal electrodes. The DF (and thus the resistive component of the impedance) goes down as the frequency goes above 0 Hz (my LCR bridge is only good from 20Hz to 200 KHz).

I'm still conflicted on how to imagine this DF effect is happening. On one hand, I can imagine the theory that you described with the microcapacitors in series/parallel with each other. However, they are in the EM field and would all be affected the same and independent of each other, right? I just imagine the water molecules changing orientation with respect to the AC field. Since the quantity of dielectric material is the same, the capacitance will be the same. But with a homogenized mixture with all the microdroplets, for some reason the molecules have a harder time changing orientation? So, droplet interface or space in between droplets... hmm...

With the microcapacitors, not only do the metal plates affect the microcapacitors, but the microcapacitors affect each other.

An analog would be to take two opposite magnetic poles and position them some distance apart and then dump in a bunch of dust sized magnets. Every particle is affected by the larger magnetic field, but also by each other. You start reversing the larger magnetic poles and the dust has to rotate to conform to the larger poles and generates heat.

The distance between conductive surfaces also affects the capacitor value, so if two droplets are close together, they form a capacitor of higher value than two droplets that are far apart. With the three factors, surface area, distance apart and dielectric constant of the separating insulator, if you look at a single drop at one end of the spectrum and a very finely divided emulsion, the emulsion will have more surface area and be closer together, so I would expect more charge to be stored in the system. The only fly in the ointment is that with a very fine emulsion, the chains of microcapacitors will be very long with many, many capacitors in series between the two plates. I suppose the thing to do would be to take a specific plate geometry and set up a model of uniform droplets in small, medium and large droplet models and run the numbers to see how much it changes.

Interesting thoughts! So I guess I need to check things more thoroughly with my next set of measurements because I initially thought that the capacitance worked quite well to estimate water content (regardless of other factors). Although this was what led to my work on the dielectric thermoscopy (capacitance temperature slope) measurement to estimate water content. This was far more reliable to estimate the water content than measuring the dielectric constant alone, possibly because of the things you outlined.

What you said could explain two problems I noticed in my measurements:

1) I did notice that the capacitance of my old samples tended to differ from newly prepared (newly homogenized) samples. I initially considered this to be water loss, but they were in sealed syringes so this is probably unlikely.

2) I also noticed that peak curve in the the DF plot is shifted dramatically to the left with newly prepared samples. After they are "aged" for a while, you end up with the plot that I gave. I imagine that upon homogenizing, the droplets are very uniform in size between the samples. Then with low water contents, it becomes dissolved over time. With greater water contents, the droplets (which could have once been similarly sized to the other lower water content samples) coalesce and form bigger droplets causing the DF vs. water content curve to shift left.

You may want to contact Doble Engineering (www.doble.com). They are world leaders in testing and evaluation of electrical equipment insulating materials, and have done mountains of research into dielectrics and the sorts of effects you describe. They may have some white papers available, or other research findings that they could share with you. Their work with dissipation factor (they call the results of their field test equipment's output "Doble Insulation Power Factor") is widely used in evaluating the condition of insulatinf fluids used in power transformers, oil circuit breakers, and other high-voltage electrical equipment. Most of their work is therefore focused on low frequency (50 or 60Hz power frequency) testing. Hopefully this will give you some additional leads to explore.

Thanks for the tip! Low frequency is key here, and 50-60Hz is closer to the order of magnitude that we chose to measure at (for various reasons). I've actually found a lot of information on transformer oil sensors and measurements, but unfortunately I haven't found anything about how the dissipation factor changes with different states of water. But if they specialize in this, they might know why this is already.