Helllo Aida, the default analysis for dynamic light scattering is the by "intensity". In DLS the scattering intensity is measured, so the contribution to the overall signal measured is related to the amount of light scattered by the different particle sizes. Once the intensity distribution is obtained, one can mathematically transform this to a volume or a number distribution. The volume distribtion describes the amount of volume (or mass, for constant density across particle sizes) of the different species, the number distribution provides the relative contribution from the different sizes in terms of how many particles there are. Feel free to check out this summary

http://www.materials-talks.com/blog/2014/01/23/intensity-volume-number-which-size-is-correct/

which desribes a technical note on the subject.

If in doubt, the intensity result is the one that comes directly from the data, volume and number are derived, and thus depend on information about the refractive index of the particles (as well as their shape). Hope this helps. Ulf

Dynamic light scattering (photon correlation spectroscopy or quasi-elastic light scattering) is a technique that is normally used to determine the size distribution in nanoparticulate colloids. Dynamic information of the particles are derived from an autocorrelation of the intensity variation that is recorded during a measurement. Then the second order autocorrelation curve is generated from the intensity trace. The most important use of the autocorrelation function is its use for size determination.

number and intensity may be good to be determined. Generally NP remains in agglomerated. Number is beneficial parameter when nano particles are in dispersed form. Intensity parameter may be better when NP shows photo-luminescence properties.

Helllo Aida, the default analysis for dynamic light scattering is the by "intensity". In DLS the scattering intensity is measured, so the contribution to the overall signal measured is related to the amount of light scattered by the different particle sizes. Once the intensity distribution is obtained, one can mathematically transform this to a volume or a number distribution. The volume distribtion describes the amount of volume (or mass, for constant density across particle sizes) of the different species, the number distribution provides the relative contribution from the different sizes in terms of how many particles there are. Feel free to check out this summary

http://www.materials-talks.com/blog/2014/01/23/intensity-volume-number-which-size-is-correct/

which desribes a technical note on the subject.

If in doubt, the intensity result is the one that comes directly from the data, volume and number are derived, and thus depend on information about the refractive index of the particles (as well as their shape). Hope this helps. Ulf

In addition to Ulf's explanation:

Intensity-weighted size distribution is what the machine yields, it detects scattered photons.

Volume-weighted size distribution is derived from the Intensity-weighted one by calculation; thus errors tend to propagate and increase the risk of artefactual peaks.

The same goes for Number-weighted size distribution, but even worse, as the latter is derived from the Volume-weighted distribution.

It mainly stems from the fact the relatioship between light scattering by particles and hydrodynamic diameter is not linear.

Intensity-weighted results are more robust, but they are not the most physically relevant

Number-weighted results are the most physically relevant, but they are biassed by artefacts.

Volume stands in the middle.

I agree with CF and Ulf's explanation, however with my personal experience on the DLS for many years, I have always found volume-weighted HDD more closer to TEM diameter for the NPs.

All three weightings are biased: intensity and number based weightings are biased to two extremes while volume based weighting runs in the middle. We described this issue in one of our recent papers (Sci Total Environ. 2014 Aug 15;490:11-8) as:  "Because notable discrepancy exists between different weightings (Nicomp, 2006; Pokhrel et al., 2013b), we chose volume-weighted HDDs for comparisons across treatments due to higher confidence in data accuracy compared to intensity weighting; the latter being directly associated with the sixth power of particle diameter yields potentially erroneous and larger HDDs (Nicomp, 2006; Stebounova et al., 2011; Silva et al., 2014)."

I hope this helps!

LP

As already mentioned in some of the previous comments representation of the DLS data as number based is a possible option but should be avoided unless you have a really narrow distribution. Having larger particles, the measurement itself is very insensitive to small particles. Number based the resulting uncertainty is unfortunately hugely magnified.

As most of answers have already stated, DLS yields a correlalogram which is numerically fitted to obtain intensity based size distribution. Further, number and volume average distributions are calculated from intensity distribution. To obtain volume average diameters one needs an exact value of refractive index (both real and imaginary components are required) and it is assumed that nanoparticles are spherical and intensity based particle distribution is correct. As at nanoscale RI values can vary greatly with size, an exact estimation of volume and/or number averages is difficult to obtain.

Another problem with intensity based distribution is that it gives false results in case of high scattering nanoparticles (mostly metallic nanoparticles). I faced this issue with gold particles with rough surface. In that case size estimated by DLS was almost 15-20 nm larger than actual size as observed by TEM. This difference was less than 10 nm for spherical gold nanoparticles (particles were not absolutely spherical).

To answer the question, intensity based distribution is least prone to error if nanoparticles are not highly scattering in nature. Converting intensity based distributions to number and volume average is tricky and should be avoided. In case one has two or three definite particle sizes in same population (as in case of proteins forming multimers) one can use number/volume average values to obtain the ratio between multimers and monomers as these distributions give number of particles within each size range. These estimates will not be quantitative though.

HI all,

I did some experiment usin Zetasizer DLS.

I get three parameters Z-average 172.9 nm Volume-mean 211.5 and Number volume 95.20

I read before that volume and number size are derivated from intensity.

I know that the real size of my NLs is 95nm .

In this case, what is the more robust parameter?

what does real size should mean if you have more than one particle, e.g. first 10 and the other 100 nm. Further you can only have information about measured or calculated size.

Seems your result of 'real' size is obtained by a method determining a number weighted size, e.g. by TEM. Therefore it is no wonder that it is close to number weighted result of DLS provided your particles have a narrow size distribution.

in this way. do you think that number volume is the most robust parameter?

There seem to be no 'robust' parameter in this regard. I read the recommendation to report the values according to method used for determination, i.e. number based for TEM and intensity for DLS. Unfortunately, for comparison between methods one has to convert. Differences in results of different methods are inherently connected to the methods used, but also enable conclusion about the sample itself.

The answer depends on your question. Are you looking at

1. the population of each type of particle in your sample? Then use Number distribution. This relationship is 1:1. So, 1 particle, 1 count.

2. the total volume? use volume distribution. This scales 1000:1 as it depends on the volume of a sphere (4/3*pi*r^3)with the assumption that each particle you measure is a sphere. This will obviously vary if the particle does not behave like a sphere in solution. TEM along with DLS might give you a slightly better idea.

3. the size? use particle distribution. This scales 10.00,000:1 with intensity and can give you the most probable size of your particle but number distribution should also be considered when calculating size. You would preferably want a monodisperse population of particles (check number distribution for this) and then look at the size.

I personally would do a size-exclusion MALS to give me more confidence about the size of my particles and then complement that with TEM. If that is not possible, then definitely TEM should give you some, if not all, answers. Good luck.

Often DLS and TEM data do not agree with each other; TEM is more reliable than DLS if you have hard particles like silver or gold. DLS tends to show larger particles compared to TEM sizing. This discrepancy is largely due to the differences underlying the principles of two methods used.