Dr. Joe Machamer: Thank you for coming. And what I want to do is start off with a little bit of an overview slide. Basically, what I want to do is talk very quickly reviewing the basics of GPC calibration and show how there are some common techniques to improve the molecular weight range that this calibration curve can be applied to and how technology from Tosoh Bioscience is improved upon what's been used conventionally and then get into some more details regarding columns made with this type of technology and then show how the capabilities of this technology is really--comes to a full fulfillment when they're run on the EcoSEC GPC system

And then I'm going to conclude with an update on some work I've been doing with Excel and coming up with methods to use spreadsheets to model and do different kinds of thought experiments with how different kinds of columns can be used in combination

So, this is a typical GPC calibration curve. And what we have on the X axis is the retention volume. For the separation, on the Y axis, we've got the molecular weight. And for very, very large molecules, they're going to be the first type of material to elute from the column. So, they're eluting in the excluded volume

Very, very small molecules are coming out at the end of the chromatography in the included volume. And then in between, we have a calibration curve where the selective permeation takes place. And what happens here is the polymer elutes as a function of its molecular weight

And when the chromatography is seen, which you see on here, the large material's coming out first in the excluded volume, followed by smaller material that's in the selected permeation range, and then the included volume

And if you look at calibration curves in catalogs from Tosoh or any other company that makes GPC columns, typically, they'll have, you know, eight or 10 or 12 different types of columns and their calibration curves laid on top of each other

And kind of a pattern that you can see when you look at that is there tends to be a tradeoff between the resolution that particular column can provide and the linear range across which that resolution can be used

And so, that's what I'm trying to show here with these couple of curves. Basically, if you have a column with a very wide molecular weight range separation capability, typically, it's going to have less resolution in terms of the shape of the slope compared to other columns where these type of material in the column's giving a very shallow slope, which means great resolution. But, then typically, there's a tradeoff in that the molecular weight dynamic range is smaller

So, what you can do if you want to increase the resolution, of course, there's a couple things that are done. Typically, you could add more columns of the same type. So, take two columns that are--have identical material in them. Hook them together. You're increasing the amount of pore volume in the column. And you increase the resolution. Or, the other thing one can do, of course, is to use a smaller particle size in the column

And then this is kind of the main part of what I want to talk about today, which is how to increase the linear range across which you can do the chromatography

So, there's basically two approaches that have been used. One is to use a series of columns where each individual column has a different pore size. So, you would have, you know, a series of columns with different exclusion limits, if you will. That's one approach

Another approach has been to take the material that would be in those individual columns and mix them together, creating a mixed-bed column, and then connect a series of those types of materials to each other for a mixed bed

And the idea is that you would take these individual calibration curves. And by either hooking them together in series or mixing the beads together and then packing the column, you would create a calibration curve that had a wider linear range

And then this is a schematic kind of showing what this looks like. The idea here is, in this conventional approach, you might have different types of particles. So, you might have particles that have large pores, medium pores, and another type of particle that has small pores

So, the idea is you can pack one column with large pores, second column with medium pores, and the final column with small pores, attach them to one another, and you'll get a wider linear range

Or, you can take these materials, these individual types of materials with the different pore sizes, mix them together, and then pack a series of columns creating a mixed-bed type. So, this has been the traditional approach. And there's a problem, though, with that

If the individual columns that you're attaching to each other or the beads that you're mixing together into a mixed-bed column don't--aren't on the same slope in terms of their linear calibration ranges, you can end up getting artifacts in the chromatography that will look like peaks but really are just the fact that these linear ranges are not overlaying smoothly enough to give you decent data

So, because of that problem, Tosoh developed a multipore type of packing. So, the idea here is, instead of having individual particles that are either large pore, medium pore, or small pore, what we do is we take the particles, and we have a distribution of pore sizes within each particle. And then you would pack a series of columns with that type of material

And what that does is that smoothes out any errors or incompatibility between the different individual shapes of these individual calibration curves

And here's an example where, in the top chromatogram, we have phenol resin separation, where we're using the sequential column approach. We have a G4000, 3000, 2500, and 2000XL

And you can see there's a little bit of distortion in the chromatogram. And basically, that's due to the fact that, for this--these sets of columns that were used, the linear calibration curves weren't ideally optimized with each other

However, if this is run on a multipore type of column, then you get the smooth curve, which is more accurately representing the distribution of the polymer within the sample

So, we have this multipore technology in, you know, conventional large columns and larger beads, you know, 10-micron-type beads. But, also, we have what are called super multipore columns, SuperMultiporeHZ columns. And these are designed to have ultra-low polymer absorption

They still have this multipore technology to improve the linearity without getting the chromatogram distortion. And they're also packed into semi-micro columns

So, what this means is we're using a smaller column. It's a shorter column and smaller diameter. And what that gives you is run times that are 50 percent faster. They're twice as fast basically because the column is half as long. And then the solvent consumption is 85 percent less because the column is half as long. And then we're running it at a third of the flow rate. So, that's saving us 85 percent on the solvent costs

And what this graph is showing here is calibration curves for polystyrene, where we're comparing this more conventional approach with the series of columns, where each type of column has a particular pore size range in it--that's what we're showing in the purple--to conventional mixed-bed type, which is in green

And then the blue here is the SuperMultiporeHZ. So, this is the material where the particles have a distribution of pores within each particle. And you can see that that's giving the best resolution and linearity of those--of the three methods

And what we have now--showing now is electromicrographs of the individual particles from three different types of SuperMultipore columns that we offer. And what you can see here is that the columns are available in 3-micron, 4-micron, and 6-microne particles, which are small particles for GPC use

But, because the column is half the length and we're running at a third of the flow rate, it works fine for polymers. There's no issues with shearing or anything like that

And also, you can see that the beads are very uniform. So, they're--it's a mono-dispersed type of technology as well

And if you do pore size measurements of the particles and you compare the distribution of pore sizes that are in the SuperMultipore type of column with a conventional type of column, you can see how in a conventional column, basically, the pore size distribution is fairly narrow, where it's very, very broad on the SuperMultipore columns. So, that's where we're getting that excellent linearity and wider range

This is three calibration curves showing the three different materials that we have. So, the material here, the HZ-N material is the 3 micron. And then we have the 4 micron, the M, and the 6 micron is the H. So, we're getting very wide useful range. And if you're looking for the oligomer region or smaller polymers, then the end material is giving very good resolution

So, kind of a summary here of the SuperMultiporeHZ columns is these packing materials have a broad pore size distribution. And what that gives is the ability to not have any inflection point in the calibration or distortions in the calibration curve or the chromatogram due to an imbalance between the different types of pore sizes being overlaid

So, you're getting a really very linear calibration curve so there's less air and more accurate molecular weight measurement

The porosity range gives you high resolution in the oligomeric region. And also, because of the concept here, there's less lot-to-lot variation because we're not having to try to match different batches of beads to keep the calibration curves overlaying in a linear fashion and, also, small particle size, providing for high speed and high resolution with low solvent consumption

And this is just a chart showing how a conventional column will compare to a semi-micro column. And basically, you know, conventional column is going to be 7.5 or 7.8 millimeters in diameter. Semi-micro column is smaller or also half the length. And so, the benefits once again are saving on the run time as well as the sample consumption or the solvent consumption

And then this chart shows for these individual types of SuperMultipore columns the separation range so that HZ-N has got up to 50,000, HZ-M up to a million, and then the SuperHZ up to 10 million. So, there's quite a range of products you can choose from there

So, when you're using these smaller columns, what you need to really think about is how the volume of the system that you're running these separate--these columns on might affect the separation

And what we're talking about here essentially is the band spreading that happens within an HPLC system due to aspects that are not related to the column, things like eddy diffusion and longitudinal diffusions. And there's things that--within the column that happen as well

So, the idea is that, if you run a small column on an instrument with the very large dead volume, you're going to lose resolution just by in essence Brownian motion and the randomness that happens as molecules move through a column

And to show that, what we have here is the chromatogram using a SuperH column. Actually, we're using four of them, using THF as the mobile phase at 0.35 mLs per minute. And we are running the same column and samples in mobile phase at same flow rate on two different systems

So, one system is our EcoSEC GPC system, which is designed to have very, very low dead volumes, compared to a conventional GPC system

So, in purple or red here, this is the resolution from the EcoSEC instrument. And essentially, we're getting many more peaks with baseline resolution. And we're able to see many more peaks, compared to running that same column and samples on a conventional system. And the reason for that is the lower dead volume of the system

And so, here's a picture of the instrument. And part of the reason that we get this low dead volume is the design of the system is where all of the components are essentially in one box, if you will. And so, that lets us place the individual components very, very close to each other so we can use very, very short lengths of tubing. We also use a very low volume in our flow cells

And the combination of all that is, when you run very small columns on this instrument, you're able to maintain good resolution

And here's some chromatography simply showing how a chromatogram with--and the conventional column versus the SuperMultipore--or, sorry, the SuperHZ column when run on the EcoSEC, we're able to maintain excellent resolution. But, once again, this is on a smaller column, so very fast run times

And if you're using expensive solvents, you know, something like HFIP, the ability to run at a third of the flow rate and still have really good resolution, you know, can be a significant cost savings

So, we have a customer who--that was his main attraction to the instrument. He was using HFIP. Of course, it's very expensive. And the fact that he could reduce his solvent usage by 85 percent by simply choosing an EcoSEC GPC system made all the difference to him

Another feature of the instrument is the dual-flow refractive index detector. And what that provides is superior performance in terms of baseline stability

So, what we're showing here are chromatograms where we're running the same column samples, mobile phase, etc., but on two different systems. We're overlaying five chromatograms on the EcoSEC GPC system. And you can see that there's virtually complete overlap

There's really no drifting or no difference between any of these individual injections. However, when that same sample is run on a conventional refractive index detector, you can see how the peaks are offset from one another. And that's going to lend to inaccurate data

And the way this dual-flow system works is, in a conventional refractive index detector, the material that's in the reference cell of the detector is not flowing. It's a stagnant pool of THF typically

And as you know, THF will degrade over time. So, if you're running samples over a weekend or overnight, if the material and your reference cell is degrading over time but you're having fresh solvent flowing from the column for your samples, there's going to be a continual change in the difference in refractive index between those two materials, which you will see as drift in your chromatography

With the dual-flow design of the EcoSEC GPC system, we are continuously flowing solvent through the reference side of the cell as well as the solvent side. So, we're always able to account for any degradation that's taking place in the mobile phase

Another feature, of course, is when you're working with GPC, the precision of the pump flow is very, very important, even more important in GPC compared to HPLC. And that's because--and what we're really after, of course, is the molecular weight of the polymer. And that's the log of the molecular weight that's related to the elution volumes

So, it's very, very critical to have extremely precise flow rates

And one of the aspects of a lab that can affect the precision of an HPLC system is the room temperature. And what happens here is, if the room temperature is changing, the temperature of the solvent will change over time, which changes its compressibility, which then in effect changes its flow rate

So, to illustrate this, we've got two chromatogram--or, a chromatogram overlaid here with an intentional change that we've made in the room temperature

So, if we change the room temperature, we cycle it from 19.9 up to 19.95 and back down and up and down again, and then we plot the elution time for a given peak, on a conventional system, you can see how the change in the temperature of the room is affecting the elution time of the molecules. They're completely in synch with one another

However, if you do this same experiment with the EcoSEC GPC system, the--of course, the temperature in the lab is changing because we're intentionally doing this. But, the retention time and elution volume for the sample remains constant. And that's because the instrument has an oven system that keeps the pumps at a constant temperature

And so, if you have, you know, any swings of temperature in your lab, they're not going to have any effect on the precision of your GPC analysis

And here's some chromatogram or some charts showing reproducibility data. And this is done on polystyrene standards. And basically, what we're seeing with these temperature control pumps is we're getting excellent retention time precision

So, CVs for retention time are on the order of less than 0.04 percent a day. And then when you translate that into the molecular weight precision, molecular weight CVs can be less than 0.2 percent or less on a day

Also, the instrument is--and the software is all designed around being optimized for GPC-type analysis. So, it has all the calculations needed for MW-, MZ-, and MN-type calculations and then, also, very nice interface where any changes in valve positions or injections or autosampler are reflected in the screen that's available

And a very nice feature also is that the column oven is unique in that it can hold up to eight 30-centimeter-long columns. So, if you are in a lab where you're using lots of different columns, you can keep them all at the same temperature very conveniently with this column oven

And then this is a chromatogram showing a whole series of different polymers. We've got epoxy resin, polycharbonate, polyvinylchloride, polystyrene, etc., run on the supergel--I'm sorry, the TSKgel SuperMultiporeHZ showing very nice separation

And you know, we're running four columns here, and we're getting this analysis done in 26 minutes or so

Okay? So, are there any questions on the SuperMultipore columns or the GPC system

Okay. So, the next thing I'm going to do is switch gears completely. And I'm going to talk about something I've been working on a little bit. And this is this Excel-based column selection tool

And the motivation for this was twofold. One of them is I would get frequent requests to make a recommendation to a customer about which Tosoh columns to use versus the type that they were using currently from another vendor

And simply looking at charts from different vendors and columns with different dimensions and different particle sizes, it was challenging in my head to try to keep together and figure out what's the best thing to recommend

So, that was something I was running into. And what happened is I came across a separation report that Tosoh Bioscience has on our Website. And this is Separation Report 28

And what this shows is something that I found very interesting. And it's basically--the idea is, if you run a series of molecular weight standards on an individual column with a given pore size and a different--a second column with a different pore size and you just add up the numbers that you got when you ran them separately, you just do a calculation, you just do math, you can get a calibration curve that matches almost exactly what's observed when you do the experiment

So, the idea here is, basically, the separations with GPC, it's very--I don't know mathematical's the right word. But, essentially, you can look at a calibration curve. And if you can come up with a scheme to mathematically add those calibration curves to each other, that's going a long way towards predicting how it's going to behave when you actually run the sample

So, the combination of me coming across this chart and having a need to look at how different calibration curves--you know, comparing different types of columns to each other, that's what the motivation was

So, let the chart--this shows here the status. This is a work in progress. Of course, I had great ideas how far this was going to go. And I'm partway along the path. But, I just want to share with you what--where I've gotten so far, just show you a little bit about how I'm doing--how I've been doing this

So, the first step is we want to take the calibration curve. And we want to take it from being something printed on a piece of paper into numbers that we can use with Excel

So, we'd scan this into our computer. And this is where I spent a lot of time was finding a software package that would let me easily convert, you know, this 2D plot into XY coordinates

So, I came across this software package called Graph Digitizer. It works very, very well. It handles the fact that the Y axis is logarithmic. And you can do multiple curves with one setup. So, that's worked very, very nicely

And so, what I'm able to do now is I can take any set of calibration curves that I get off the Internet or from a brochure or something, scan it into the system, and then I can convert it into an Excel type of format

And then I can simply do math by adding whatever curves I'm interested in comparing. And I would get a result which theoretically should be how those columns would behave if you actually ran them

So, in this example, I thought, "Let's do something really, really crazy. Let's take the column that we offer with the smallest pore and the column with the largest pore and hook them together theoretically in the model and see what happens.

And so, this is something that you would probably never do in reality because, typically, you would want to have, you know, different materials that have a more continuous relationship to each other

But, essentially, what we do is we take the G1000 calibration curve. So, basically, this is small pores. So, it's, you know, showing a good separation, you know, up to 10,000 or so. But, then everything larger than that's coming out in excluded volume

And then we've got the G7000, which of course is able to give good resolution across a broad range of materials. I'm sorry. That's what this one is

Simply add them together, and this is the composite that you would get

And then the other thing I thought about, well, what if we took the same column? And what the heck? We just--what if we hook 10 of these together, something you would never do, once again, in a real lab. But, I wanted to see how that would affect the resolution

And indeed, the more columns we're hooking together, the larger pore volume we're seeing and the more shallow the slope is, so the better the resolution that we're obtaining

So, this is kind of pretty much where I am now. I've got the tools in place where I can play with these things and make the spreadsheets make sense. And so, now, the next step for me is to figure out how to compare columns that maybe have different geometries or different particle sizes or even using different molecular weight standards for the calibration to begin to expand how this tool can be used

So, that's what I've been doing with the Excel technique. And--oh, here's what I did. Here's where I did it with--I'm sorry, with 10 of the same columns

So, if you have one G7000 column, you get not much resolution compared to if you were to hook 10 of them together. But, it's probably not something you would do in reality

Another thing, of course, I can do with this, I can compare how mixed-bed resins are comparing with SuperMultipore-type resins and how that's comparing with using a series of different columns

Okay? So, my conclusion is TSKgel SuperMultiporeHZ columns provide accurate wide linear range separations. They're designed around the semi-micro concept. So, these smaller columns provide a time savings and money savings because of the low flow rate and the quick runs

And the EcoSEC GPC system, when combined with these semi-micro columns, offers superior performance and versatility because you're able to really get a lot of resolution out of small columns when they're run on a system that has very, very low dead volume

So, thank you for coming