Ok I will devide this in two sections:

Possible issues with the DLS:

First of all which DLS peaks are you talking about? You can do volume distributions (volume of particles at certain size) number distributions (number of particles at a certain size) and intensity distribution (intensity scattering from different particles sizes) for more infos on what is what see:

http://nanocomposix.com/sites/default/files/handbooks/introduction.pdf

(a useful img is on page 6)

E.g. number distributions give higher peaks for small particles than volume distribution (you need a lot more small particles to contribute to a certain volume than you need from big particles) So depending on which you are looking at the peak heights differ.

(you can change this in the settings when you do analysis of your DLS data)

Another possibility for not seeing it although its there is sedimentation. If your vesicles sediment before you can measure them you might not see them with DLS. (this is more likely for bigger particles) If you have high enough quantities you might see that though. (you can also test that by letting the solution sit and then TEM the supernatant maybe)

Possible issues with TEM:

Second: A big difference between DLS and TEM is that the first is in solution while the second is dry. So there might be sthg happening while the drying process (or staining in case you do that)

Your TEMs look a bit like aggregates. So what you see might be the aggregates from the small particles that form during sample prep. Another artifact you might see is buffer crystals or impurities that were in your solution (depending on your sample prep again, but that you can rule out by looking at the solution without your particles). Another possibility for TEM artifacts is radiation damage. (I once observed structures that looked a bit like your last img after too long measuring times)

Ok thats all that comes to my mind, hope it helps :-) have fun, Romana

more TEM

Tem

TEM again

Ok I will devide this in two sections:

Possible issues with the DLS:

First of all which DLS peaks are you talking about? You can do volume distributions (volume of particles at certain size) number distributions (number of particles at a certain size) and intensity distribution (intensity scattering from different particles sizes) for more infos on what is what see:

http://nanocomposix.com/sites/default/files/handbooks/introduction.pdf

(a useful img is on page 6)

E.g. number distributions give higher peaks for small particles than volume distribution (you need a lot more small particles to contribute to a certain volume than you need from big particles) So depending on which you are looking at the peak heights differ.

(you can change this in the settings when you do analysis of your DLS data)

Another possibility for not seeing it although its there is sedimentation. If your vesicles sediment before you can measure them you might not see them with DLS. (this is more likely for bigger particles) If you have high enough quantities you might see that though. (you can also test that by letting the solution sit and then TEM the supernatant maybe)

Possible issues with TEM:

Second: A big difference between DLS and TEM is that the first is in solution while the second is dry. So there might be sthg happening while the drying process (or staining in case you do that)

Your TEMs look a bit like aggregates. So what you see might be the aggregates from the small particles that form during sample prep. Another artifact you might see is buffer crystals or impurities that were in your solution (depending on your sample prep again, but that you can rule out by looking at the solution without your particles). Another possibility for TEM artifacts is radiation damage. (I once observed structures that looked a bit like your last img after too long measuring times)

Ok thats all that comes to my mind, hope it helps :-) have fun, Romana

I agree with Romana; the lower 2 images look like beam damage to me. I would advice to do a cryo-TEM study to indentify the actual morphology of the particles

Hi ZZ,

the last two examples are almost certainly artefacts caused by beam damage. If you ever record similar images, if you leave the beam focused on an area and watch it over course of tens of seconds, then you'll be able to see bubbling and boiling caused by beam sensitivity.

Can you post a higher resolution of the first image? Is it negatively stained? If so what with? You can tell if these are lipid vesicles because they will have a collapsed appearance in projection, sometimes a bit like red blood cells. This is caused by the 3-D spherical vesicles flattening down and sometimes popping as they adhere to the TEM grid surface and then dehydrate.

I have seen with certain peptides structures similar to your and they are not vesicles but rather unstructured gel like micelles. It's possible your lipid is forming these type of structures as the sample air dries and rapidly up-concentrates. Hence you don't detect them in solution.

I use DLS on my nanoparticles and prepare the sample in the cuvette at 0.0125 mg/mL-- this works for getting strong detection of 150 d.nm particles. You may want to consider whether you have the UV absorption value and refractive index of your primary polymer component entered correctly into the DLS software. Try shaking your stock of particles before diluting them in the cuvette because if they have set for 3 years then they may have sedimented by now.

I think Romana gave an good answer. In my opinion, you should be more precise in the description of DLS results. Different size distribution may results from the statistical analysis (intensity, number and volume). The intensity distribution is more sensitive to larger particles compare with the others. You should compare them all. If you are quite sure of the spherical shape, the z-average value should be a good compromise for size identification. Usually, it is not so easy well observed 5nm particles with common DLS instrumentation. Maybe, a careful analysis of correlation curves are needed. However, I think the main problems arise with the drying process for TEM sample preparation.

If you put the specimen into a Zeiss Libra, which you have at Case, you can confirm that the vesicles have the expected elements, C, O, and whatever metal you have used. The smaller ones look like vesicles to me, but the larger look as though they may have multiple layers. On the other hand, maybe they just collapsed.

for your 2nd question, you may use the cryo-TEM to visualize the original (or fresh) vesicle polymer.

Hi ZZ,

I agree with Schirhagl and Richard Collins.

1) The DLS results maybe induced by the aggregation of big particles, so that only small colloids or polymers can be detected. And the Simulation curves in DLS is not so confident especially when the size distribution is very broad. You can have a look at the Report Quality in your DLS result. Generally speaking, Rh(intensity) bigger than Rh(volume) bigger than Rh (number).

2) There's possible that when you prepared the samples of TEM, the polymers are second-assembly during the drying period. DLS is tested in solution and TEM in dry Situation. Another possible thing is that there are some impurities in the solvent or the organic salt in buffer solution.

3) Some of Peptide, Glucose or similar Things will have beam damage when you focus on them for a Long time during TEM observing.

Your TEM images look just like the ones we got on dense liquid-like protein clusters using a cryo-TEM (see attache image). Although, we can characterize these clusters very nicely with DLS, I still think these particles seen in (cryo)-TEM images are some kind of artifact coming from sample preparation. So, I am quite curious the know if these are indeed artifacts and how they are formed.

I agree with the above comments and would like to add one very simple argument. You say that the polymers are hydrophilic and that they are linked by complexing metal ions. The fact that by DLS you detect only the parent polymer in solution but not the species liked together by complexation indicates to me that the concentration of metal ions and/or polymer is too low for this cross-linking to happen. As the sample dries on the TEM support film, these concentrations continue to increase and eventually the rings will form. I would thus suggest you repeat the DLS experiments with much higher metal ion concentration. Maybe this will solve the problem.

The only reliable information from DLS must derived from the intensity plot, except if you correct for the fact that you are sensing hollow spheres.

See for example:

http://www.sciencedirect.com/science/article/pii/0005273683901542

Nevertheless, the scattering from hollow spheres is not too different from spheres, e.g. according to H.C. van de Hulst's monograph, in comparison to a full sphere the polarizability of a hollow sphere with a shell thickness of 1/5 the radius of the full sphere only changes by 5 %.

Therefore, I doubt it is a fundamental problem with DLS. Although your volume/ or number data may not be 100% correct, the DLS should be able to "see" the particles moving, if they are present in solution, and thus determine their size by the Einstein-Stokes equation and give you the corresponding intensity data.

Thanks for all your answers. I knew much more about DLS and TEM. I would like to add more information about the experiment. The metal I used is cisplatin and we want to make nanoparticles for cancer therapy. 1. All the samples were made in water and filtered through 0.4um filter; 2.No negative stain was made for TEM because the particle already had cisplatin. 3. For the DLS experiment, I checked both intensity and volume distribution and I didn't find any peak around 100 nm.

According to all your answers, I think these 'particle-like' structures are mostly not vesicle and they were formed during drying step. I have no idea about beam step, will this make these structure fused as show in Figure D-E?

I agree with R. Schirhagl and Mathias Brust, although I'm less inclined to think it's beam damage. If you are able to detect monomer but not complexes by DLS then the large "cross-linked" coordinated complexes you sought to make did not form. However, in contrast with Matthias, I think it's more likely that your metal ion concentration may in fact be too high, effectively saturating the negative charges on your polymers and preventing complex formation (analogous to the prozone/postzone phenomenon well-known in immunology). Only upon drying and rapid increases in concentration of solutes do weak interactions between the molecules favor aggregation over exposure to the gaseous phase.

Zhuxian- In response to your question whether the beam can cause the fusion of materials the answer is that it's not likely. The beam can cause "boiling" as one of the other contributors mentioned, but when that occurs there is generally visible damage to the support medium, which is not apparent in your samples.

Especially for the vesicles with a lot of "structure" it might be helpful to tilt your sample stage to see if the sample has three-dimensionality or is just 2-D flat on the sample holder.

In general, the larger the tilting difference between to images, the easier the three-dimensionality can be determined. A sketch usually helps: Normally if there is any feature which is on the left hand side of the "sphere" (the plane that divides left from right is parallel to the rotating axis and perpendicular to the sample holder) on the image with no tilt and moves to the right hand side after tilting, it tells you that the sample is not just flat. (assuming you do not tilt more than a total of 90°)

Dynamic light scattering like all photon scattering depends on refractive index differences between the media and the particles. Your visicles probably do not enough Delta eta differences to be detectable on your light scattering system.

You can add a heavier metal to the fluid on the inside of your vesicles or add a soluble heavy metal to your media to get greater delta eta between your vesicles and the media, The downside to this approach is that heavy metals are generally toxic and this strategy will ruin the vesicles for biological applications.

on the other hand, there is pretty good contrast between your vesicles and whatever liquid the vesicles are in the, presumably, optical photomicrograph above. Therefore, dilute this suspension of vesickes sightly to eliminate the overlap seen above and use image analysis to determine the particles size distribution of your vesicles. ImageJ is a free downloadable program on the USA NIH site. Contact me directly for some notes that I provide in my characterization class at Penn State to get our procedure for particle size analyses by image analysis.

In the number distribution you have biggest chance of seeing smaller particles. To exclude beam damage you can try measuring TEM at lower keV (in case you dont have cryo at hand). Also when I saw this on my samples you could really see the change. So you can watch out for changes in appearence while scanning.

I appreciate all above comments related to the above question. I belief that due to high energy electron beam irradiation of TEM causes the deformation in the structure of your polymer. This may finally provide you the distinct particles as you have seen. The same incident I have observed during the TEM experiment of my bismuth containing glasses. You can read the following related article.

http://link.springer.com/article/10.1007/s11051-011-0468-y

Just I have an idea related to your DLS experiment (it may not be a justified argument), as your polymer is hydrophilic and cross linked by water soluble metal ion, so the brownian motion of the complex decreases as the size of complex increases. This may be a cause for which you have not determined the particle size using the DLS techniques.

Zhuxian, if you cannot confirm the vesicles you saw under TEM are your samples, then the vesicle size you obtained from TEM may be wrong. So the actual size of vesicle may be out of the detection limit of DLS. You probably need another microsizer to see if you can find the size distribution at micro-meter scale.

For the fused vesicle, I saw similar phenomenon. I used a surfactant to synthesize nanoparticles and sometimes the surfactant was not removed completely. When I used copper grid with carbon film to observe the morphology of nanoparticles under TEM, similar vesicle like structure could be seen. They seemed to be exploded under the electron beam that always contains heat to make the organic molecules unstable. Thus the structure of your vesicles should be changed when you were doing TEM, which means TEM results are not reliable. I am not sure that what accelerate voltage you used for TEM, but it is better to use low kV and Cryo TEM must be the best for you.

Hope this will help.

Cryo-TEM would be the best option. One of my friend recently obtained nice images of lipid vesicles through cryo-TEM.

Concerning damage in the TEM, it is not totally correct, that less beam energy yields less damage, although it might be a valueable rule of thumb. For carbon this rule is true below 100kV, but 300kV electrons "interact" less with the carbon atoms than 100kV electrons do.

For beam damage in the TEM look here:

http://www.sciencedirect.com/science/article/pii/S0968432804000381

In regard to TEM experiments, it might be worth testing different kV settings in order to see if you get repeteable results. As in, if your vesicle-like structures are formed during the beam-sample interactions, which is quite possible according to my own experience and the comments of other fellows, you could obtain different sizes depending on the beam current.

You have a lot of good sugestions here. I read that when you make your suspension you use water. Perhaps you may try a neutral buffer? the polar nature of the water molecules may be interacting with your complexing. HBS or PBS might give you better results. Cheers!

I think before discussing the size you should be sure about the chemical composition of the observed particles in TEM images So my advice is you may try to get a EDX if the instrument includes the apparatus. But if not: there are some type of EDX detector attached Field emission SEMs which can detect this particles and within the same time getting a EDX data from the sample. Because with the true set of the DLS data it is a very low possibility to ignore the that sizes. A final note the sample characteristics should be clearly stated in DLS measurments the dielectric and refractive index is the first comes to the mind. But the concentration parameter should be precise to be sure about the viscosity parameter and the particle stability. To be sure about the stability of the particle size you should work with gradually evaporated samples.

Good luck

DLS gives particle size in terms of number and volume distribution graphs. So these graphs itself gives vesicle size. and also it gives poly dispersivity index (PDI) that gives the distribution.

there are many vesicles articles like liposomes and niosomes that can be detected by DLS. so this vesicualr system must be detected by the DLS but it gives particle size in terms of number dsitribution and volume distribution graphs.

The most important thing is that you have to give accurate refractive index in the DLS instrument as refractive index has important in particle size measurement.

Dalwaldi, the refractive index is important for volume conversions, if you just want an intensity or number-based measurement then R.I. is not critical for DLS.

Dynamic light scattering (quasielastic light scattering) gives the dynamic structure factor (S(q,t) and the intermediate structure factor g^(1)(q,t). Proceeding from these to particle sizes faces the serious challenge that the transformation is not very stable. Given that you have the image above, you know the approximate size of these particles, and hence can calculate the diffusion coefficient and the corresponding relaxation time 1/D q^2. The first step is to look at S(q,t) over the relevant time scales and determine if you see any relaxation. Your correlator should someplace report a file or output to computer with the actual spectrum.

There are now several issues: (1) did the correlator settings cover the relevant time scale? I have no idea what correlator you are using; timescale may not be an issue. (2) What was the vesicle concentration in solution? I infer the image is TEM after drying, in which case you may have in solution actually a very low vesicle concentration, which I realize is not what you expected.. (3) As correctly noted above, the vesicles and solvent could be index matched. You might try running up the integration time and laser power and see if a signal finally emerges from the noise.

Having been doing light scattering spectroscopy since 1969, I can assure you that refractive index increment dn/dc is absolutely critical for light scattering. If you have somehow managed to index match the solvent and the vesicles, the vesicles will scatter no light and not contribute to the scattering spectrum. Arranging to have index matching, when you are trying to do it deliberately, is difficult, so I would rate this as a low probability event.

Obscure technical possibility: Are the TEM images from liquid on which you first tried to do DLS? If not, it might be the case that when you filtered the vesicle solution to remove dust, the vesicles were absorbed by your filter and did not reach the scattering cell. There was once upon a time an issue with DNA being absorbed by one of the standard nanoporous materials of 45 years ago; the solution was to pass a bovine serum albumin solution through the filter firs,t and precoat the filter (yes this does sound like a stone age approach, but we are discussing many many years ago).

As you are at Case Western, you might usefully contact my former student Kiril Streletzky, who is at Cleveland State.

Dear Professor Phillies,

Thanks for your explanation and suggestion about measuring the particle size by DLS. I'm now out of the sample. I will follow your suggestion to running up the integration time and laser power when I try this experiment again. I only used the standard sample to check the setting. And now I realized the setting may work for the standard sample but it may not good for my vesicle. The solution for DLS was filtered. As you mentioned, the vesicle might be absorbed on the filter. I have never thought about this issue and I would like to try measure my sample before filter them. Thanks again for your time.

The vesicles -- I found your note above -- are a significant part of the size of the filter pores. They should have fit, but perhaps as another potential problem could have been disrupted by the filtration step. What standard sample did you use?

The particle size are much less than the filter pores and it usually won't be removed. I still think the most possibility is the vesicle formed during drying step if there is nothing wrong with my DLS setting. The standard samples are coming with the DLS instruments. It's some kind of nanoparticles. I forgot the materials.

I would suggest you to go through this paper.

http://link.springer.com/article/10.1007%2Fs11051-013-1576-7

hope it helps?

With your sample being hydrophilic polymers in aqueous solution (conditions?), you might consider trying other sizing technologies than ensemble (DLS), but based on separation and individual counting such as shown here:

http://www.mrzgroup.umd.edu/pdfs/2012_Tends_biotech.pdf. Applicability of separation techniques depends on sample stability. These are good contacts to talk to and figure out what's doable with your specific case. When you have made a new batch, you might also contact me and we can look into a quick measurement.

Try with freeze-fracture replicas. It si no easy to do but the results (images) are very convincent. It could help you to have a very close image to the real aspect of your sample. I think it could be a good experience. Bonne chance!