1) You may check the stability of your laser using a photodiode, if the laser oscillates you'll see it in your a.c.f.

2) You may check the correlator by correlating your dark counts.

3) Correlate a simple filtered toluene sample to see if it is your sample.

Showing a plot may help others to think about your problem. These are generic suggestions.

I have observed something similar before. In my case I am sure it was related to the samples and not the instrumental setup (I checked this). I am not sure what caused the oscillations, but I suspect incipient sedimentation. Could your gels be separating into a high concentration and a low concentration region?

Since you are working with gels the time scale of your decay is probably relatively large. What time frame / frequency do the sinusoidal oscillations correspond to?

An oscillation essentially means that at certain times photons are more likely detected than random noise.This could be due to misaligned optics/flare, vibrations, sedimentation, convection, or even laser beam shape. Temperature control can be quite tricky to achieve, especially when operating close to the surrounding room temperature. There is a publication titled "Oscillating time correlation functions from dynamic light scattering of gold-labeled tracers" in J. Chem. Phys. 108, 9594 (1998) which may provide some additional clues. If you solve the mystery, let us know. Good luck!

I've uploaded an example of a DLS run. To clarify, I'm monitoring the 'spinodal' gelation of protein solutions. Immediately after sample preparation, the intensity correlation curve can be fit with a single stretched exponential. However, after a given time (10min), strong and clear oscillations arise (my hypothesis is that the entire sample is very close to complete gelation). These g2(tau)-1 curves can be fit with a sinusoidal exponential decay. Anyone any thoughts on a physical interpretation?

Kind regards,

Mike

Thank you for showing the correlation functions (they are really interesting). These actually looks quite different from what I thought when you explained it the first time. Are your samples optically clear at this stage? How long are the accumulation time of the data? Since your sample changes a lot in only 10 min, the oscillations might be due to a significant change in your sample structure during the accumulation of the correlstion functions. Is it possible to slow down the gelation process to check this? Do you have the possibility of measuring at several scattering angles to check the q-dependency?

I am working with a CGS3/ALV SLS/DLS unit. I cannot do simultaneous multiple angle measurements. I am not 100 percent certain that the samples are still optically clear at this stage (I cannot inspect the sample during data collection). But I can tell you that the sample starts out as a clear solution and after a matter of minutes it becomes completely turbid. Slowing down the gelation process might be possible, however, I must take care not to induce nucleation of protein crystals. This is why I chose such high driving forces, to push the system to gelation rapidly, thus making sure that there are no other competing solidification pathways. There is no pronounced q-dependency of the "steady-state" gel (i.e. sample at long times). My guess is that I need to go to far smaller angles for that (my range is 12.5 - 155 degrees). Also, after the "transition" regime as shown in the screenshot, the amplitude of the oscillations decrease significantly. The "steady-state" gel still displays oscillations, but far less pronounced.

Your ALV unit also gives the count rate as a function of time during the measurement. Could you show us what this looks like for the same measurement as the other figure?

Certainly,

clearly, the intensity drops during the data collection. It would seem I need to opt for shorter collection times. What do you make of it?

(my apologies for the quick and dirty Excell figure, I'm a bit lazy today)

Since the intensity drops during the measurement, your sample is clearly changing quite a bit during the measurement.This could be the reason for the oscillations you observe. Trying shorter collecting times is probably a good idea.

Furukawa and Hirotsu wrote a detailed paper on DLS/SLS on gelling systems, see

http://dx.doi.org/10.1143/JPSJ.71.2873 They write "On theoretical side, the work by Pusey and van Megen may be the first extensive study on this subject.

They showed that the conventional treatment of light scattering, which assumes ideally homogeneous media, yields erroneous results when applied to analyze the light scattering results from real inhomogeneous systems. This is because the total scattered field measured at a single point is not a zero-mean Gaussian variable, which contradicts the basic assumption of the conventional theory."

This is in my opinion the key point as shown by the example of Mike: the observed scattering is clearly not constant on a relatively long time scale, which gives problems in the normalisation of the a.c.f. In the past several methods have been tried to speed up the averaging for instance by rotating the sample to average of inhomogeneities in the gel.

You'll find more information by searching using non-ergodic as search term.

Ok,

many thanks for the excellent feedback. I appreciate it !

Mike

Dear Mike,  These  oscillations look very interesting -   the data is published ?

I would be  interested  to know  more about the  experiment. PK.

Dear Mike,

Furukawa and Hirotsu are absolutely right: gel is the system of twisted and tightly bounded fibers and other inhomogenities at a great distance (1-10 mm). Therefore usual DLS theory is inapplicable, because it implies the presence a lot of scatterers, which move independently in accordance with Brownian law. 

When the scattering occurs in the gel, in the scattering volume phases of all scatterers change almost identically in al scattering volume. When the liquid or gel oscillates in the cuvette whole fiber network moves as a whole, and a dark field of speckle pattern of scattered light changes to light one. You really select one speckle spot when selecting coherence area for DLS. As result you see the frequency of gel oscillation. Usually it is characteristic frequency of your optical table.

We saw the same picture in the DNA and collagen gels, and in the system of cottonwool pieces in glycerol-water solution.

The question of study of inherent gel frequencies is more complicated. If you would like to discuss it, please, connect with us through e-mails: [email protected] and [email protected] .