Clays are usually thixotropic; which would explain this behaviour.

Hi Damien,

Microscopically speaking it is to do with the internal  bonding, hydrogen or van de Waal.  Most likely the water is  binding to hydration centers, acting like a glue between crystalline structures. In this case the work being done is against the Hydrogen bond around these centers.  That is why it is important to investigate these materials as a function of temperature and to see if the peak frequency you observe has Arrehnius temperature dependence, f=f0exp(-Eb/kT), where k is Boltzman's constant and Eb is the bond energy (as suggested by Hiep Nguyen above).  The same behavior can be observed in Dielectric studies of Clays, where the same bonds are involved.  The file is attached.

Article Dielectric Relaxation of Water in Clay Minerals

Hi Damian,

Given the relatively low concentration you are using (10g/l), it looks to me that you are below or close to the sol-gel transition. In other words, your material is still somewhat liquid .

You can lookup the litterature for the work of Michot, Levitz and Baravian on the phase transition in clays topic.

Dear Damian,

Maybe this article can help you.

http://www.sciencedirect.com/science/article/pii/S0955221913002823

Dear Damian:

This is probably happening because your sample is viscoelastic. The storage (G') and the loss (G'') modulus are frequency dependant, even in the linear viscoelastic range. This is shown in several Rheology books, but you can check it on the book 'Rheology: principles, measurements and applications', by Christopher Macosko (one of the best rheology books in my opinion). On page 124, a brief discussion is made on the frequency dependence of the dynamic moduli.

For polymeric samples, the expected behaviour is an increase in both moduli when you increase the frequency. Depending on the material composition, the loss modulus may reach a maximum value at a certain frequency, while the storage modulus may reach a short plateau.

I hope this can help you to analyze your data.

Best regards,

Rubens.

Dear Damian

there are some features of frequency sweep graphs we should know for characterization. first of all we should know that a typical fluid would show slope of 2 for G' and 1 for G" (G' depends on square of frequency in a log-log plot, and G' depends on frequency) at low frequency regions, based on mathematical definitions describing small amplitude oscillatory shear tests. this behavior, increase in moduli, is an indication of relaxation process. 

However, there are some cases that show different behaviors like plateau regions. this plateau region is an indication of entanglements or network structures (gel-like) in suspensions and nanocomposites which prevents the relaxation mechanism. the latter case should show up at very low frequency regions to prove the existence of network structures.  at higher frequencies (shorter times) movements are quite fast and there is only very local relaxations which are not so sensitive to material structures. that's maybe why in a series of  suspensions with different particle concentrations all graphs corresponding to different particle concentrations converge to roughly the same magnitude at very high frequency regions.   

therefore, it's better to  characterize your suspensions based on their behavior at low frequency regions to see if there's a gel-like behavior, network or just an increase in moduli( terminal behavior). usually increasing the clay concentrations should lead to a plateau behavior in low frequency regions. so if your material still shows sharp increase in modoli probably it means it didnt make a network or at least a weak network is obtained or it's still in a sol state.

all the information and papers advised by other friends here are also useful and appreciated.

Best

Reza 

the typical behaviour o the suspensions is the following:

from low to high frequency, G´ and G´´ increases with frequency, with a certain slope depending on the material. In this case, if we have a low concentrated suspension below the percolation concentration we will have a slope higher than 1 for G´´ and higher than 2 for G´. allways in log log scale of course. the slopes of 1 and 1 means that our clay concdentration is so small than really there is no any interaction between particles. When I talk about interacctions I mean that they are separated enough to avoit any interaction, no posibility of colision by brownian themal movement, no repulsion or atraccion potential (london van der walls...).  In this situation, although we reach the maximum frequency of comercial rheometers (628 rad/s = 100Hz typically) without inertia issues (what is really difficult unless we know what we are doing what typically is not the case, we will never see a crossover between G´ and G´´ becouse the relaxation time is lower than 1 / 628 rad/s.

Now, let´s increase a little the concentration. Now, we are increasing the posibility of physical interations between particles, so the slope of G´ and G´´ versus frequency will decrease. the higher the concentration the lower the slopes. To increase the concentration of particles will give to the solution a more solid behaviour and solids are less dependent to frequency (in fact to the observation time). So at a certain concentration we will start seeing the crossover point between G´ and G´´. The inverse of the crossover in rad/s units will give us the relaxation time of the suspension.

If we increase more the concentration, the crossover freqency will move to lower frequencies as the relaxation time moves to higher values. Of course the slopes also increases. At a certain point. we don´t see anymore the crossover point although we go to very low frequencies. This means that G´ is allways higher than G´´. At this point we have reached a gel, solid like behaviour. The interactions between particles are so high that the particles forms a network with physical forces acting as bonds. But this gel is not a perfect solid. Still, there is a slope of G´ versus frequency. becouse we still have a material that his viscoelasticity depends on the observation time. The relaxation time is very high but not infinite and we can vary a little his liquid like or solid like behaviour changinf the observation time (frequency). G´´ typically has a u-shape in the frequency range studied. Becouse it cames from to relaxation times (this is dificult to explain without a drow). at very high concentrations limiting the maximum fraction of packaching we will get a colloid glass that I think is far from the scope of this question.

Please note, that this explanation is a general explanation regarding suspensions. Particle size, particle shape, the chemistry of the particles have a great influence in the dependence of the viscoelastic properties on frequency. This is the reason why rheology is a fantastic tool for the study of all these situations.

please don´t hesitate to ask if you need fhurther explanation.

regards

Carlos, thank you for your perfect post! I will show it to my students every time I come into explanation of dispersions. I would never have found such simple yet exact words!

Thanks a lot profesor Karpushkin. I am glad to help. best regards

Dear Carlos

May you explain more the relationship of particle size, particle shape, ,the chemistry of the particles influence in the dependence of the viscoelastic properties of frequency?is it high particle fractions affect the storage modulus at low frequency?