Round Robins for Molar Mass

Topic 1: Selection of columns

Two column producers were testing the Eurolignin samples with the goal of coming up with recommendations on choice of columns: Polymer Laboratories (PL) and Polymer Standard Services (PSS). According to the many data we had at that time, there seemed to be three Eurolignin samples which represent a "good spread" of properties of possible lignins: Alcell, Curan Kraft and bagasse. The main goal was therefore in a first stage to find already columns which work at least with these 3 lignins. Recently only the Alcell lignin has been kept as reference. The other references being the ultrafiltrated fractions.

One important point is that viscometry measurements seem to be able to help here (although it is not easy to get a good signal): if the visco signal does not follow the UV or RI signal we can conclude that there must be another reason for this than lignin association or "real" high molecular weight tails or real multiple peaks.

Topic 2: Calibration

It turns out that the values of Mn and Mw seem to depend much on calibration problems. In all data I have analysed so far, it seems that using a reference lignin (Alcell) for calibration can reduce the interlaboratory variation of Mn and Mw to less than 50%. Therefore, if you do some work don't forget to introduce Alcell lignin into your samples. We can use that for further challenging that approach.

I could identify for the moment two aspects of this problem. First, the usual standards give not a good reference for the hydrodynamic volume of lignin. This has been shown in literature several times and we could confirm that (a simplified example: in DMF the Mn values are systematically 2 to 5 times higher than the ones we get with THF). Second, there seem to be important factors related to the "spread of the signal" (see also topic 3 for that): this spread seems not to be eliminated by using standard references (see graphs in my presentation mentioned above: it's striking that in this case the graphs seem to be more characteristic for every lab than for the lignin type!). However the use of a reference lignin for that seems to be able to reduce a lot of these variations.

Absolute calibration with MALDI and MALLS seems to give lots of problems also, even on fractions. But in the meantime we got reasonable results on the ultrafiltration fractions.

Topic 3: Resolution, separation, bandwith

See item 6 in the "guidelines" and the second remark under topic 2. We are not very advanced on the understanding of this point. But it seems to me an important point to be checked in upcoming work.

Topic 4: Derivatisation

One idea I follow there, is to try to avoid derivatisation, as this introduces additional incertitudes. Maybe DMF with Li salts could be one option, but we have not enough data yet about that. However, testing is going on about derivatisation also. For the moment, my impression is that acetylation and use of THF is actually a quite good method (although it does not completely eliminate the non SEC effects, this is now very clear and also verified on the Eurolignin samples). Also other alternative derivatisations have been tested: Acetobromination (or bromoacetylation as some call it). See working document about that also on the Internet Forum under "Round Robins" and "Ongoing Molecular Mass Round Robin". It seems that although the method is very practical and makes soluble some lignins that are not soluble after usual acetylation, is does not eliminate problems of acetylation: some samples that are completely soluble after acetylation, are not anymore so after bromoacetylation. My interpretation of the data available is furthermore that interaction problems with column (at least in the case of polystyrene columns) actually get worse.

Topic 5: Association with lignin molecules

Problems appear with that again: even with Light scattering it seems to be very difficult to be sure that what we measure are real molecular weights and not associations. Also on GPC in some cases, for example acetobrominated curan lignin in THF, there is clearly association shown by viscometry.

Again, the most promising method to advance on that is viscometry: if there is no viscometry signal we can be sure that there is no association. If there is a signal, we have to check the same sample in different conditions with different solvents (interest of this network) and hopefully by comparing the multiple results (and if yields and solubilities are well controlled) we can probably conclude on the nature of the peaks.

Topic 6: Losses on columns

See item 4 in the "guidelines". Some methods proposed there were questioned in a post by PSS also on the ILI Internet Forum (unfortunately no more accessible). Please pay attention to that if you are planning work for measuring that point.

Topic 7: Choice of solvent

THF, DMF, NaOH are still good options that are verified. One result: we have tried to find out advantages of DMAc over DMF, but so far nothing is really clear. Solubility seems to be very similar. So for the moment, as DMF is more current in laboratories, we recommend rather this approach. In recent work both DMF and NaOH give most reasonable results.

Topic 8: Purification

A presentation was given about this in the EWLP of Vienna (work of Hamburg: Bodo Saake). The purification of Ca, etc. salts in lignins by ion exchange resins does not seem to really improve the situation.