Mr. Ralph Hindle: Thank you very much. I wanted to especially thank Joe Weitzel for inviting me to speak. If any of this talk doesn't interest you, it's all Joe's fault

And I do want to thank Frank for describing retrospective analysis. I don't say the word very well. Therefore, I call it swimming with sharks

It's about the same kind of thing, as you'll see as we go through the talk here a little bit. And I want to thank Matt and Don out of my lab for helping me with this work

So, this is the other side of the Rocky Mountains looking west. So, on the other side of the mountains, they have eclipsed Vancouver. So, it's about an 11- or 12-hour drive and beautiful downtown Calgary there, which is where I'm from

And I have to step out here so I can see what I'm supposed to be talking about. Oh, here we go

So, yes, the thing with analysis these days in your environmental analysis or otherwise is that your customer wants everything. And they want it in a hurry. And you have to have high-throughput tools to do that. And you have to be able to confirm the identity

There's so many compounds. Even if you know the accurate mass or even if you know the formula, there's so many compounds that it can be

So, there's a number of tools, and I'm going to talk about some of them in here for confirming the identity, including metabolite ID and phase I and phase II metabolites

And we're looking for typically really low levels of compounds because these things can act not so much in an acute--causing acute problems but also some chronic problems and development issues

So, in environmental analysis, it's really low levels acting over a long period of time that can cause trouble

And there's also the parent compound versus the metabolites. So, we have to be able to relate them and try to estimate what the levels might be

So, I'm not going to show a whole lot of specific data, you know, about wastewater and what we're finding and on a big survey, but I am going to step back a little bit and talk about if you're going to use online or thinking of using online SPE as a technique, you know, some of the initial hurdles you're going to find, and how it can help you out if you get through those hurdles

So, here's a look into my lab. And on the left is a TripleQuad. On the right is my 6520 Q-TOF. And it is nice that it's open like that because I can now take my online SPE system, and I can roll it up to whichever instrument that I want because I only have the one online SPE

So, typically, we're going to use it--use the Q-TOF to determine what compounds might be there, and then as we go through and start doing production samples, run them on the TripleQuad

And so, what I did is I went down to Home Depot, and I bought a rolling tool cart. And it fits very nicely on the top. So, you know, what we have is the drawers with all of the consumables in there

And then we have a large-volume autosampler with a 900-microliter loop in there and a--we happen to have a quaternary pump that's actually over on the system, but--on the rest of the system. But, a quaternary or binary pump, a second pump is what you need in order to be able to run the online SPE and do cleanup in between the sample injections

And then we have a couple of valves here. And down the bottom is a six-pored valve that you often find just in the column of an--of your LC system. And that's quite usable. We have a second one here

But, the second one is usable as long as you have a second pump. But, the top one's kind of interesting because it has a 12-port--I put this slide in because it's nice and neat. All the plumbing isn't there

This one here, we have six different small columns, 30-millimeter columns that are in there. And the idea in a production environment is that all of these columns are going to be the same

And all you're going to do is switch from one column to the next from one injection to the next

So, if you can get about 200 injections of fairly dirty samples onto these columns, if you can get 200 per column, then without changing anything here, you can get about 1,200 analyses done. So, that's kind of a convenience from a production point of view

But, the other interesting thing is you could put a different column chemistry into each one of these cartridges. And so, if you're doing method development, it's quite easy to change it over. You just change the position of the valve through the mass sensor software

So, it is a little bit like a spaghetti show there, but it's quite effective

We label all our lines because, once you take it off--well, taking it off isn't the problem. Taking two of them off, that's a bit of a problem

So, when you initially start, I started by looking just at the small PLRP, which is a polymer-based resin that's in the cartridge. And this is what I get when I run a normal HPLC-type analysis, where I loaded on the front of the column, and then I start to elute it off

So, I wasn't too impressed at this point. But, the idea is not to load it onto the column and then elute it off. The idea is to load it all on the column and then wash off all the nonretaining compounds to waste, then backflush the column and onto an analytical column, where we perform the separation

So, as soon as we do that, then you can see here that I'm using an Eclipse Plus column as an analytical column. And this is the same volume injection

So, these area counts, they're about 2 million, 2.5 million. So, they're fairly comparable, given the change in chromatography

So, this is actually a very impressive peak for that amount of time and, you know, I think a two-minute load, where I'm just washing with aqueous on the SPE column

So, the next thing is the SPE column selection. You know, what are you going to pick, and how does it relate to a typical chromatogram with just the analytical column

So, this is just the analytical column, where I'm injecting 10 microliters, so a fairly typical injection. And that is estradiol. So, this is probably from Gus's water I understand, from the previous talk

And this--just to test the effects of the large-volume injection, this, again, is only on the analytical column. But, I diluted by a factor of 20, and then I injected 20 times more. So, there's a 200-microliter injection

And there's a pretty much perfect overlay of the peak on peak width and peak shape and just a slight difference in the response. So, that was very encouraging

So, then I took a PLRP column and added that into the mix with the back flushing. And I realized that I had used this one for a fair bit of analysis with direct--with nondiluted uncleaned-up urine samples

And so, I grabbed the wrong column. This was not a very good peak. I took a beta column, which is a porous shell AQ column, which is able to withstand lots of water flowing through it, you know, 0 percent B. And I used that as a--as my SPE column

And so, my chromatography got a lot better. And my response came back. And so, I wasn't quite sure what was going on there. So, typically, you'd go back and say, "Well, if I want to use the PLRP," which is very short, I went and used the new one, and it improved quite a bit

So, it is important that, you know, to realize that you are going to expire these columns at some point, depending on what you're running on them

And depending on the compounds and the mobile phase, there is going to be a difference in the phase that you select in the SPE versus the phase that you select in the analytical separation

So, these are all with the backflush on the SPE system. And what I wanted to see was the linearity, is I injected 500 microliters of different concentrations

And so, you can see here--if you can't see on the Y axis, we're way up on the 10 to the seven, you know, tens of millions of counts. So, we're either getting some saturation of the detector on the Q-TOF--all this work was on the Q-TOF--or insufficient ionization in the source

But, you can see here that the next three orders of magnitude, they're all around, you know, 60,000 counts, 600,000 counts, 6 million counts

So, the linearity is very good. The peak shape really doesn't change. It's quite excellent

So, again, that's all with estradiol

And so, one of the big tests with wastewater effluence or any other matrix is the ion suppression because the more interferences that we have on the water at the time our compounds are coming out, we're going to get suppression

If we don't account for it with internal standards, then we're going to have quant issues. And there's only so many internal standards you can put in when you're looking for hundreds of compounds

So, here's about a dozen internal standards. And the black or the gray is normalized response for the internal standard in LC-grade water, and the red bars are the response for the same amount injected but in the wastewater effluent

So, in the worst case, I think we're down around 60 percent ion suppression--or sorry, 60 percent loss of signal. And in one case, we have quite a bit of ion enhancement. But, generally, we're in the 75 to 80 percent for signal, even with a 500-microliter injection of wastewater

So, it looks like a good technique. And this is with a three-minute wash on the SPE column, where those nonretained salts and other things that are in the water are going to go to waste while we enrich the sample on the SPE column

And so, it kind of brings us to the swimming with sharks. I have no data on sharks. And no sharks were hurt in the making of these Powerpoints

But, this is us as analysts. We're down there in the water. And we're typically trying to determine, you know, how much of the nasty things are in the water. How much of these things are going to hurt? You know, and we're counting them

So, this is targeted analysis. We got a diver down there. He wants to count the piranha and see what kind of problems we're going to have. But, there are other things that are in the water

And so, there's--you know, one of the problems, if we're not paying attention, we can see all these things in our target analysis by TripleQuad or by SingleQuad, by SIM

But, when we start doing TOF analysis and Q-TOF analysis, we get to do a retrospective look at the data, and then we can start to see that there are other problems. And sometimes, they're bigger than what we were looking for in the first place

So, this is a chromatogram, 500 microliters, again with a three-minute wash of wastewater effluent

And I have a custom 130-compound database that I've set up. And this is what's returned. So, without revealing the identity of all of them, you know, you can see that the front of the chromatogram or the SPE is taking place is relatively quiet

Three minutes, we switch over. And so, we're getting a fair number of compounds that are present in the wastewater. And as we open up and look at more and more compounds, bigger databases, we're going to start to see more peaks

So, at some point, I wanted to know if any of the glucuronides or sulfates, any of the conjugated drugs might be in the water. And I started looking

And there was a weak hit for pentazocine-glucoronide, and so, I--like I say, it was weak. So, I just went to see if the parent drug was there

And so, you can see that it was not only there, but it was quite high. So, there's the shark. You know, there's many other drugs

So, even if we're using a 2,000- or a 6,000-compound database, there's always other compounds that we can look at that we don't know about

So, when you're looking at emerging contaminants and somebody says, "Oh, maybe you have pentazocine in your sample," your sample has long been disposed of. And you're able to go back just with the chemical formula and look at the TOF data and see if it was there

So, as long as you keep this over months and over years, much longer than the samples in storage, you would be able to go back and look historically to see if it's been a problem. How long has it been a problem

And you know, one of the things that struck me, reading in the newspaper a number of months ago in Calgary is the finding of PMMA as a contaminant in ecstasy. And it's quite toxic, and it's killing people

And so, I had some urine samples from the Center of Toxicology in Alberta that we're collaborating with and just went back and looked at PMMA

So, we didn't really find any over the couple of years before when we looked. It was a newly identified compound. But, at least we were able to go and look

So, we could say that, you know, in the samples that we looked at, it wasn't a problem then. So, it's a new problem

So, that may help, you know, depending on your application, law enforcement start to pin down timeframes and so what's changed in the last three months

So, Agilent has a lot of tools to work with the data. And we use some of these. You know, now that we start to look at lower and lower levels with the online SPE and save some time and look at very low levels, the PCDL databases and libraries, they have forensic and toxicology databases and libraries, the libraries being different as they have MSMS spectra in them instead of just the accurate mass info from the databases and also personal databases

So, if you're working on proprietary compounds, you can start to add your own to the library and to the database

And so, out of that mix of compounds that I showed you, there was--there's one that came up. Diltiazem came up as a potential hit. And it was one of the internal standards as well

So, I ran MSMS on it. And there's the spectrum that I got. But, diltiazem has a nominal mass of 415. And I'm seeing predominantly 418 here

But, actually--so, if you look at the forward search in--against the PCDL, so you can search, just like on a NIS [sp] library. You've got a forward search. I think it's 51. But, a reverse search, where the spectrum from the library is compared against the spectrum of your unknown, it's actually 81 because it finds the 415 without too much trouble

So, the 418 is actually from the D3 of the internal standard

And the library contains as many collision energies as you want for your spectra. So, by default, 10, 20, 40 volts in the collision cell, and knowing that, when I ran my MSMS and I only wanted to pick one collision energy, then I just picked the midpoint for this analysis

So, another tool that they have is the molecular structural--molecular structure correlator. And in there, I can create the compound exchange format, the CEF file, from that previous run with the MSMS data

And then I import it into MSC, which is much easier to say than molecular structure correlator

And what it gives me is a list over here of potential formulas. And it scores them based on accurate mass and some other criteria. And then each block--so, mass 178, there's four possible structure--or four possible chemical formulas that would fit. And it gives you the PPM difference of the accurate mass of the 178

So, the 178 could be one of these four in red and then, you know, 175. There's a couple of them in green. There's a couple in black

So, it does block them together by color to help you determine, you know, which ion in the spectrum are you--that you're looking at

So, there's the parent compound in the center. And you know, for whichever one that you're on, it gives a list of chemical formulas that the final structure could be for the product ion. And it shows you where the bond is breaking to get that

It assigns penalty points. This was all published in an article. And Agilent has used that penalty point system to try to determine what the structure might be

So, basically, the lower the number of penalty points, the easier it is to break the bond

So, all of these are tools to help you identify or confirm that this really could be the compound, even if you don't have the actual standard

So, again, in, you know, the type of work that we're doing, environmental analysis, it's all--it often comes down to, you know, you read the spec sheets, the ultimate detection limits. How much can you see of reserpine

Well, how many people here are analyzing reserpine? One? One hand? Yeah. So, in environmental analysis, we're not looking at that very much. But, we're trying to get to lower and lower detection limits all the time

So, common detection limits here, milligram per liter, the PPM, micrograms per liter. We get down to picograms per liter for the parts per quadrillion

But, the real ultimate detection limit is one molecule per planet. We're not there yet. And one of the problems is there's going to be storage issues

Once you get it to the lab, where you going to put it, right

But, I did have the opportunity to work on a 6550 Q-TOF, which Agilent introduced last year. And on my 6520, the level that we have at the top, 10 picos, so about five picograms is typically where my detection limits are on my 6520

So, it's kind of interesting working on this instrument because we were optimizing at that level and diluting down from there. Oh, we don't want to put that much in there. That's--you know, that's five picograms. That's a little bit too much

So, we did serial dilutions once we had optimized a little bit. And so, there's one--10 picograms, one picogram, 100 femtograms, and 10 femtograms

And so, if you use that 10 femtograms and if you were able to put it on column with online SPE in 1.8 mils, you know, which is the multidraw kit, the 900-microliter loop that I have, then we're looking at, you know, five parts per quadrillion in the vial

So, you start to combine, you know, some offline cleanup with online cleanup. Where do you draw the line? This can save a lot of time in sample prep by having very sensitive instruments and being able to inject large volumes with online SPE

And then we get the nice accurate mass information

So, again, thanks very much to Joe Weitzel and to my team in Calgary. And be happy to answer any questions