This isn't so much an issue about conductivity but instead the low polarity of the ethanol. It has two consequences: less charge and lower electrophoretic mobility.

The problem then becomes applying a high enough electric field to get a measurable velocity. When I invented PALS, it was for measuring in non-polar liquids. The choice of dip cells was essential. Using a capillary tube requires thousands of volts! The dip cell allows voltages less than 100 to be used.

John Francis Miller

It’s the field that’s important, not the absolute voltage. The conductivity of pure ethanol and methanol is adequate in the capillary cell for electrophoretic mobility. The voltage and hence field can be varied in both systems. It is often identical. 150 V (not ‘thousands of volts’) in the capillary cell over 1 cm is identical to 25V over 0.16 mm in the dip cell. Ideally one would have both.

Thanks John and Alan for your answers. That clarifies my misunderstanding regarding the conductivity. So is the limiting factor the dielectric permittivity of the medium? If so, could you estimate a threshold where a folded cell won't work anymore and a dip cell will be required? Since I have not pure alcohols but binary mixtures with water, maybe measurements might work up to a certain alcohol fraction, right? Thanks a lot for your help.

Alan F Rawle Yes, the field is important but to achieve a given field requires a higher voltage for the capillary cell vs. dip. For aqueous samples, the Zetasizer may employ up to 150V for a folded capillary cell. In 90% EtOH, expect mobilities of the order of 10^-10 instead of 10^-8 m2s-1V-1. 150V will be insufficient. You will need what I used to call "killer volts" :)

@Andreas Gerstmayr - you should make the measurements at a range of applied voltages and look for a linear relationship so that you can be sure you are resolving electrophoretic motion appropriately.

Thanks John Francis Miller for this very helpful information.

I saw you answered related questions here on researchgate and I got some "aha"-moments out of it :D One point still confuses me a little.

Is it correct, that in low-polarity solvent the double layer will extend much further into the bulk since there are less polar molecules present that could balance the surface charges and this leads to much higher drag and thus reduced electrophoretic mobility? And if so, wouldn't addition of a background electrolyte have a positive influence on the drag and thus lead to higher mobilities?

In my case the samples have very low ionic strength and I was wondering if there will be an actual double layer in case of low polarity solvent. This idea is supported by a study by Shukla et al., where the zeta potential was not a good descriptor of particle stability in low polarity solvent and the authors argued that particle interactions are essentially coulombic interactions rather than double layer interactions if I understood it correctly. But thats another topic ;)

Article Experimental and Theoretical Analysis of Ink Dispersion Stab...

Again, thanks a lot for the discussion. It already helped a lot!

John Francis Miller

Field is voltage/distance (as you know!) and lower voltages (e.g. 25 or even 1V; both may have issues related to significant Brownian motion contribution broadening the peak) can be used in the capillary cell easily. It's not true that higher voltages are needed in the capillary cell for the same field as the dip cell. Mobilities of at least 2 orders of magnitude better than you state can be measured in the system (i.e. 10E-12 m2s-1V-1) - capillary or dip cell. I have carried out many measurements in pure EtOH in the capillary cell without problem. Lower conductivities than this would require the dip cell, I agree. Pure EtOH represents a defining lower working conductivity point, IMHO. However, I do agree that a voltage 'titration' is always a sensible course of action.

Andreas Gerstmayr If you have a problem converting to zeta potential then just quote the unambiguous mobility values. There's only 50% difference anyway between the interpretations of a tightly or loosely bound double-layer.

Alan F Rawle You might be able to answer something that's been on my mind for quite some time. Malvern's sales literature happily states that phase analysis can measure mobilities down to 10E-12 m2s-1V-1. Of course, that was the statement I first made in my PhD thesis and subsequently demonstrated in published research on engine oils. But I have yet to find a published paper of any commercial implementation of PALS demonstrating measurement of mobilities that low. Can you provide some references?

BTW, the other challenge with the greater electrode separation is Joule heating due to the higher voltage required to achieve the same field strength.

Regarding conductivities, the requirement for an appropriate background conductivity is due to the Zetasizer's implementation of the technique. You can measure samples that are essentially insulating quite readily if you apply a voltage to the electrodes. However, if, as in one of the Zetasizer's operating modes (in the Pro, at least), you attempt to keep the current constant, then you require sufficient conductivity to generate a measurable current. But the alleged benefits of using a constant current instead of constant voltage are dubious.

John Francis Miller

Yes, I agree that Joule heating can be an issue. And I guess you know that part of trying to solve this issue involves pausing the voltage cycle to allow the system to 'relax'. Both the capillary and dip cells can be used under a variety of conditions (e.g. voltage) and many people such as yourself have preferences - no problem, all flavors can be accommodated... There is always a balance to be achieved between result quality and any untoward effects. Excessive electrolyte concentration exacerbates such issues. Mobility can be measured with modern electronics to an incredible small degree - always the judgement call of 'fit for purpose' needs to be made. Sample preparation is also a vitally underestimated procedure and invariably alters the original sample state.

Alan F Rawle Even 35+ year-old electronics can :) My record to date was set in 1991. Not for electrophoretic mobility as such but rather the absolute displacement of the particles. Approx. 300nm diameter PSL particles in aqueous electrolyte in an electric field at 1kHz. Visualize that - essentially random diffusion of a vibrating particle. The amplitude of the vibration (due to the electric field) was about 0.5nm. The calculated mobility agreed with other measurements at various frequencies, ranging from 1Hz up to 1kHz. Under these high field frequency conditions, perturbation of the double layer is minimal whereas at low frequencies the particle oscillation may be many double layer thickness equivalents.

Andreas Gerstmayr

One other thought that occurred to me and I keep on forgetting to write it down is that, in 'thinner' liquids such as EtOH, then the density and size of the particles can be more important than the zeta potential in terms of propensity to settle. I assume you've carried out a Stokes' Law calculation to estimate what is likely to remain in suspension? Such a calculation is shown in Table 1 of ASTM E2490-09(2021) Standard Guide for Measurement of Particle Size Distribution of Nanomaterials in Suspension by Photon Correlation Spectroscopy (PCS).

Thanks a lot Alan F Rawle and John Francis Miller for this informative discussion and please excuse my late response. I haven't done a Stokes' Law calculation yet, but thanks for the reference. I completed some measurements in the meantime and so far in 50%nPA and 50%EtOH hardly any issues. I found that for my carbon black particles, good quality data is achieved when going as low as 0.1mg/mL.

Some results were strange though. One sample gave me about +30mV for 5 consecutive measurements with very good quality factors >1, although it was expected to be negative in sign. When repeating the measurement about 20 minutes later, a more reasonable value of about -20mV was reached, although quality factor was considerably lower (

Andreas Gerstmayr

My usual thought, with results such as you indicate in your second paragraph, is that there are a few large lumps or aggregates that settle in the system or perhaps some thermal equilibration needed. If you're simultaneously looking at size or the count rate then there may be indications that this is happening. Alternatively there may need to be diffusion to take place into the pores of the carbon black. If there are air bubbles in the carbon black (or the surface is somewhat hydrophobic) that need to be displaced by the fluid then perhaps this would alter the mobility. Try equilibrating the C (not Cilia) Black with your EtOH/H2O mix for an hour or so before measurement and extract the supernatant for mobility measurement.

Alan F Rawle

Thanks for these hints. My first guess was aggregation, too. But size distribution showed a monomodal distribution around 280nm. The sample was already a couple of days old and sonicated when initially prepared and sonicated again immediately before measurement.

Hi Andreas Gerstmayr , to narrow down what might be the cause of your observation, I would do size - zeta - size repeatedly as a function of time. You can use the SOPlayer for that. Things to monitor in addition to z-average, PDI and zeta are conductivity and derived count rate. Also carefully lift the cuvette out of the sample holder at the end and check whether there are air bubbles at the electrodes, and the sample still is covering the electrodes. Ideally the general state of your sample should be similar to before the run.

https://www.materials-talks.com/blog/2016/04/21/aggregation-kinetics-with-dls/

Dear Ulf Nobbmann , thanks a lot for the advice and please excuse my very late response. I took your advice to heart and now always do a size measurment before and after each zeta potential measurement. I ran into another issue now, which is related to size measurements rather than zeta potential measurements.

I currently try to investigate the dynamics of ionomer solutions in ethanol-water binary solvents by multi-angle dynamic light scattering, and investigate the influence of ionomer concentration, addition of salt and solvent polarity on the fast and slow mode diffusion coefficients of the ionomers.

However, my measurements almost always abort due to "an unexpected error". I have attached a screenshot. This never happens when performing MADLS for my catalyst particles, only for the ionomer solutions.

Also, from what I understand, multi-angle dynamic light scattering requires knowledge of the refractive index of the ionomers, since otherwise no adequate weighting of the intensity distributions from the different angles is possible, am I right with this? Since I currently cannot determine refractive indices for these ionomers this method might not be suitable anyway, or would it?

I might open a new topic on this, since it is a different subject, but maybe you, Alan F Rawle or John Francis Miller might know what the problem here is.:)

Thanks to all of you for your valuable advice and pointers!

Cheers

Hello Andreas Gerstmayr , great to hear that you were able to take some data. Thank you for sending along a screen snippet of the ZS Xplorer software from your Zetasizer Ultra.

Yes, MADLS requires the optical properties for correct weighting of the different angles. You can always edit this afterwards and see the influence.

The screen shows that your run aborted at subrun 19. With luck there may be a way to trick the software out by performing manual runs and limiting those to 15 (in the Advanced Settings). However, the issue seems to be that the scattering intensity in your sample is changing dramatically during the measurement (over a factor ~5 in under 2 seconds). This was likely from the forward angle. If you manage to get any data from this saved, please feel free to email or attach to this thread. [ description of how to do that at https://www.materials-talks.com/blog/2018/10/15/zs-xplorer-how-to-import-and-export-data/ ]

There is a very slight chance that a software update to v2.1 might improve this. However, a better chance might be if you can try the manual mode, 15 subruns of long duration, to try to average out the slow intensity fluctuations.

If you do just single angle repeated runs in automatic mode, and provided they do not abort, does the intensity also vary this much?

Thanks a lot Ulf Nobbmann for your fast feedback. You are right, the abort always happens during the forward scattering.

I just looked into the advanced settings and I guess I'll try to do separate DLS measurements at different angles. The angular dependance of the different modes of diffusion is of interest to me anyways. I will upload data as soon as I have time to do the measurements.

In the literature I found that absolute values for the scattering amplitudes have to be extracted in order to allow for the measured diffusion modes to be adequately interpreted in terms of their significance with respect to polymer structure and dynamics. The following passage was taken from Marián Sedlák's review article in Langmuir, 1999, titled "What Can Be Seen by Static and Dynamic Light Scattering in Polyelectrolyte Solutions and Mixtures?".

"It was concluded that these structures are due to electrostatic interactions, because the slow mode appeared to be very sensitive to charge interaction parameters such as ionic strength and polyion charge. On the other hand, it appeared to be not very responsive to changes in the backbone-solvent compatibility (hydrophobic-type interactions).18,20

However, it was noted quite recently19 that the response as seen in the dynamic light scattering spectrum corresponds only to relative changes between the two modes, and it is absolutely necessary to extract from the data the scattering amplitudes of the respective modes in absolute units and to discuss them separately."

Is there a way to extract absolute values of the scattering amplitudes from the Zetasizer measurements?

Thanks again for your valuable and highly appreciated support!

Cheers

The "default" distribution output is a normalized version, where all amplitudes sum up to 100%. It is relatively straight forward to get "absolute" amplitudes by multiplying with the derived count rate. (The derived count rate is a normalized count rate, see https://www.materials-talks.com/blog/2015/06/11/derived-count-rate-what-is-it/ , written for the classic Zetasizer software, and similar holds for the later version)

This would give you the "absolute" amplitudes at one angle. However, if you need the amplitudes for different angles, then you need to normalize with respect to a standard. This may be described in the review article you cite. Typically, you would have to measure a standard (often toluene, but that may not be ideal for your case) at different angles, and then ratio the derived count rate to that intensity of the standard at that angle . Technically, you should also subtract the scattering intensity from just your background/solvent to find the scattering of your sample... So, yes there may be a way to get some sort of absolute scattering amplitude of the respective modes, yet this may get somewhat complicated if you also need these at different angles.

Thanks Ulf Nobbmann for explaining. I‘ll hopefully get some more data next week and will upload and keep you posted. Have a nice weekend.

Dear Ulf Nobbmann

I havent had time to get more data, but I have a question regarding static light scattering measurements using the Zetasizer Ultra. From my understanding, the device principally is able to perform such measurements but the software has no data processing included yet. It would be wonderful if you could share your opinion on performing such measurements with the device. Is there a way to manually perform such measurements and post-process the data externally?

Thanks a lot for your great support!