Seized Drugs & Forensic Toxicology | Session 3 from the 2026 NIJ Forensic Science Research and Development Webinar Series

Megan Chambers: Good afternoon, and welcome to the 2026 NIJ Forensic Science Research and Development Webinar Series. We're excited for you to join us. My name is Megan Chambers. I am a program manager at the National Institute of Justice, and I will be moderating today's session. And NIJ is the research, development and technology agency of the US Department of Justice and is dedicated to improving knowledge and understanding of crime and justice issues through science.

We're excited to share with you some of the research our programs support. Today's webinar, Seized Drugs and Forensic Toxicology will showcase NIJ supported research addressing the rapidly evolving landscape of controlled substances and toxicological samples. Presentations will focus on emerging synthetic drugs, novel psychoactive substances, analytical challenges and toxicology, and innovative detection methods responding to the changing drug threat landscape.

Next slide, please. Before we begin, we'll review some important information. To adjust your audio settings, click on the down arrow button next to the grayed out mic icon to adjust as needed. Questions can be submitted at any time, and there will be a Q&A after each presentation. To submit your questions, click the Q&A icon at the top of your screen. Please note we may not be able to address all questions.

Closed captioning is available. To access, click the more button with three dots. Select Language and Speech and then select Show Live Captions. Recordings of the presentations will be posted on NIJ’s website at a later date. Certificates are not being offered for this webinar series. Today's presenters, in the following order, are first, Dr. Tyler Davidson, Associate professor and graduate program director in the Department of Forensic Science at Sam Houston State University.

Second, we have Dr. Marta Concheiro, associate professor of forensic sciences, master's program at John Jay College of Criminal Justice, and Dr. Ana Pego, assistant professor of forensic toxicology at John Jay College of Criminal Justice. Third, we'll have Dr. Rajesh Sardar, professor of chemistry and chemical biology at Indiana University, Indianapolis. And lastly, we'll have Dr. Ashley Fulton, a research chemist at the US Naval Research Laboratory.

Please note that there will be a Q&A session after each presentation. The presenters may not have time to address all questions, and we appreciate your understanding, and we'll do our best to respond to as many as possible. Next slide, please. And with that, let's hear from our first speaker, Dr. Tyler Davidson. Tyler, the floor is yours.  

Dr. Tyler Davidson: All right Megan, thank you very much. Greatly appreciate that. As Megan said, my name is Tyler Davidson. I'm an associate professor and the graduate program director here for the department of forensic Science at Sam Houston State University. And today, I would like to talk about copper phosphine ion complexation for the analysis of Cannabis plant material.

Next slide. Before I begin, I do need to acknowledge that the authors of this presentation have no real or apparent conflicts of interest or any financial relationships. This project is obviously funded by the National Institute of Justice. However, the opinions, findings and conclusions expressed today are those of the authors and do not necessarily reflect the review of NIJ.

Next slide. So we have to start by talking about Cannabis. So Cannabis is a dioecious flowering plant with several variants, including those shown on screen here, Cannabis sativa, Cannabis indica and Cannabis ruderalis. This is one of the oldest and most widely used plants throughout the world. However, the two subspecies that we're going to focus on today are marijuana, which is a Schedule I controlled substance as defined by the Controlled Substances Act at the federal level, and hemp, which was legalized in the 2018 Agriculture Improvement Act.

Now, in that act, the legal threshold here to differentiate hemp and marijuana was a 0.3% delta-9 THC threshold, based on a dry weight basis. Now we know there was further clarification provided in 2019, which told us that this 0.3% threshold was actually the total THC, which includes the acidic precursor THCA. Now, of course, there's ongoing legislation, including some that we're going to see here very shortly this fall, H.R. 5371.

But regardless of what comes out of the final determination here, we know that we can no longer deal with this question from a qualitative perspective to differentiate hemp and marijuana. We now need to be looking from either quantitative or at least semi quantitative analyses. Next slide, please.

So within Cannabis we have a complicated set of cannabinoids and terpenes and other chemicals as well. But I want to focus today on our cannabinoids. I've shown on screen here several of the common cannabinoids including delta-9 THC, which is the main psychoactive component of marijuana, and CBD, which is the main component that we see present in hemp.

I haven't shown the acidic precursors, but that's these two structures plus the carboxylic acid functionality there. And then on the bottom of the screen here, I wanted to highlight a couple of the specific delta-9 THC isomers, including delta-8, delta-10 and Exo. And what you can see here is that the only difference in these structures is the double bond location.

And for the sake of today's presentation, the presence of a double bond or absence in the location is going to be very relevant. Next slide, please. So in terms of traditional analyses, gas chromatography mass spectrometry is our standard sort of gold standard technique. However there are concerns with cannabinoid conversion as you have heating in the GC inlet. Particularly we're worried about the conversion of CBD to delta-9 THC, which will impact the accurate reporting for the delta-9 THC content. If we look at techniques that do not involve the, you know, elevated temperatures going to LC, for example, depending on the detector, if we use a UV detector, now we have issues with the UV spectra being not overly distinguishable.

And likewise, even looking at product ion spectra, we still have issues differentiating, for example, delta-9 THC and CBD. Next slide. So this is where ion complexation enters the conversation. Ion complexation is the approach where we introduce a transition metal.

Today I'm going to take a look at silver and copper. And we know that these transition metals will have affinity for the alkene functionality. So they basically will have reversible binding with carbon-carbon double bonds or pi bonds.

And what we’ll observe here is a metal adduct. But the formation of this adduct is going to be dependent on the steric factor. So for example that alkene position within the molecule. So the presence or absence and then the actual position. So if we look at the two compounds that I started with today, delta-9 and CBD, you'll see that in delta-9 THC we see an internal secondary alkene.

However, in CBD we actually see an internal secondary alkene and we see an additional terminal alkene. So when we interact these cannabinoid structures with any of these transition metals, we're going to have subtle differences in the binding affinity associated with each metal. Next slide, please. So there are some previous examples of this. The first example I want to show here is by Huang et al., that looked at paper spray ionization tandem mass spectrometry for CBD oil.

In this paper they use simple silver salts, and they basically showed that they could get distinguishable MS/MS fragmentation based on the presence of unique MS/MS product ions as seen on screen. This work was then continued by a different research group highlighted here by Ieritano et al., that combines this concept with ion mobility and showed that basically these silver adducts could be separated with IMS separation, and they still saw distinguishable MS/MS product ions. Next slide.

More recently, Sboto et al. also looked at paper spray tandem mass spectrometry. In this case they actually studied a bunch of different transition metals. Highlighted on screen here is their work with copper, which is going to align most closely with what we're going to talk about today. But what we can essentially see here is that they formed a unique precursor, in this case copper cannabinoid minus a hydrogen, that upon MS/MS activation they were able to see distinguishable product ions.

So a way to differentiate between delta-9 THC and CBD. And the most recent work that I'd like to talk about that sort of laid the foundation for the topic of today's presentation was by my former Ph.D. student,  Alleigh Couch, where she demonstrated the presence of multiple cationic species, shown here on screen in the red, the blue and the green, where we see hydrogen extraction in the red and oxygen insertion in the blue, and then the loss of a proton in charge neutralization with a dye copper species in green.

So all of these works have laid the foundation, but all four of them have the same sort of limitation, one being selectivity. So any double bond that's present, as we know there's lots of double bonds in our forensically relevant compounds. Also from a solubility perspective, a lot of these salts are not overly soluble in a highly organic solvent.

So we need to consider that. And then finally stability as well, trying to create something that is going to have long term stability. Next slide, please. So that's where we began with this idea of copper phosphine ion complexation. If we use a support phosphine ligand, we should be able to modulate the binding to improve the selectivity, improve the solubility and improve the stability.

So in this study, we looked at 11 copper phosphine complexes and 20 common cannabinoids. Our optimized ratios here ended up being about 50 micrograms per mL to 15 micromolar of the copper phosphine complex. Of course these are different complexes. So sticking with a general description here. We then went through a stepwise evaluation for all of these complexes to try to figure out what is going to give us the best performance, the best selectivity for our target cannabinoids.

This included looking at these solvents. So for example, use a protic solvent like methanol, how does that impact our resulting data versus an aprotic solvent such as acetonitrile. And all of the characterization work was completed on an Agilent Technologies 6530 Q-TOF. So we had good, high resolution, accurate mass measurements. Next slide, please.

So this slide works a little bit better with a bunch of transitions. But for the sake of progressing through this, I pulled out the transition. So what I want to highlight here is that we went through a sequential step. So all of the steps in blue here, we looked at solubility in the protic and the aprotic solvents. We looked at the ability to form a complex for delta-9 THC.

We looked at the ability to avoid sort of solvent effects or analytical splitting of signal, as well as the ability to differentiate between our delta-9 THC isomers. And finally, the abundance of the sort of characteristic ions of interest. And at each one of these steps, we were able to eliminate some of our copper phosphine complexes, so that ultimately we ended up with three optimal copper phosphine complexes, one that was mono dentate, which was our (PPh3)2, and two that were bidentate, which was our (R)-BINAP and our dppf.

Next slide. So let's look at some of the data here. So in the top row we have delta-9 THC. In the bottom row we have CBD. What I would like to highlight here is that the CBD in all cases looks just like the copper phosphine complex because the CBD does not bind to the complex, meaning we get no analytical signal from CBD.

But in comparison for delta-9 THC, we do see a precursor ion that's going to shift depending on the complex, but in all cases it is formed through adding the complex and the addition of that oxygen through the proposed oxygen insertion. And so we now have these three complexes that just based on full scan data, we can now readily differentiate between delta-9 THC and CBD, keeping in mind that CBD is our main sort of interference with our quantitation of delta-9 THC. Next slide, please.

So to summarize what we did, and to lay the foundation for the rest of today's talk, we took our list of 11 copper phosphine complexes, and we narrowed this down to three complexes. We explored the solvents and we saw that solvent does matter. So ideally we use aprotic solvents, such as acetonitrile. And the reason for this is we don't have competition for charge carriers.

When we have methanol with the extra hydrogens, we do have the ability to still form some protonated precursor, which again is going to cause analytical splitting and hurt our sensitivity. Now the advantages of this copper phosphine ion complexation technique, the primary one is we can now differentiate delta-9 THC and CBD in full scan. So basically turning a blind eye to the presence of CBD which again is our main interference.

Also this works for differentiating the non isobaric cannabinoids. So for example, CBN, which is shown in spectrum F here on the bottom right, this is shifted by four daltons relative to delta-9 THC which is shown in spectrum C here on the top left. So those are the advantages. That works really great, but there are still some limitations.

The primary limitation here is that if you look at spectrum C versus spectrum E, you'll see that we get the same precursor. Now we don't get the same abundance, and we don't get the same binding affinity between these two. Because again it depends on the location of the double bond. However, we do unfortunately see the same precursor ion. So this is an issue that needs to be addressed.

Likewise, although this is all very interesting and very exciting qualitative work, it doesn't address the legal definition. And so what we had to do in stage one was characterize these complexes. And then in sort of stage two here we want to expand this into semi quantitative and quantitative studies, which is what I'm going to focus on for the rest of today's talk.

Next slide, please. So why semi-quantitative. Because I said at the beginning that we need a quantitative value to address the legal question. Well perhaps that's not necessarily true because we could use an administrative threshold. Because we could use this to triage suspected marijuana casework. And we know that most of the marijuana casework has well above this 0.3%.

We can actually set an administrative threshold or a decision point. And anything that is above that will be marijuana. And anything that is below that will require additional testing. But that may be a small percentage of our samples. Primarily because we know that we can eliminate the interference from CBD, this works really great because we don't need to, we don't have to concern ourselves with the CBD.

Of course, we do need to worry about the other cannabinoids that may be present. However, if you're going to use this type of yardstick comparison, we can't do the total THC math and have a single point of comparison. So we need to convert our THCA to delta-9 THC through a forced decarboxylation protocol ahead of time. Again, this does introduce some of the same challenges we talked about from a GC perspective.

But the benefit here is we're taking a more conservative, semi quantitative approach. Next slide, please. So with this we chose a 1% threshold. Again this is three times that legal definition to give us quite an amount of buffer between.

What we know from authentic casework samples is many times we're well above this threshold. So that's not a huge concern. And anything that's less than this, we're going to follow up with more targeted quantitative testing.

The idea behind this to try to present this somewhat simply, is that we essentially fortify our sample with delta-9 THC-d9, or an internal standard at 1%. And then we use that as a yardstick. Is our signal greater or below that threshold? As I mentioned previously, we do have to go through forced decarboxylation. In this study, we did this at 150°C for 25 minutes to maximize the conversion of THCA into delta-9 THC, while attempting to minimize any further conversion to CBN, for example.

We had to optimize which complex was best, which I'll show you here in a second, as well as our ion complexation conditions and of course the chromatography to address our delta-9 THC isomer differentiation. Next slide, please. So determine the best complex I want to show you guys the product ion spectra associated with these three copper phosphine complexes. On the far left here we see copper (PPh3)2 . And what you'll see is we get two product ion species shown here in green that contain the cannabinoid. This is essential because we want to use these product ions for semi quantitating in this case.

And unfortunately for our BINAP and our dppf, we don't see that. We actually lose the cannabinoid based on the differences in affinity. So we quickly eliminated them as our potential candidates and focused on a copper PPh3 moving forward. Next slide. So to optimize our ion complexation approach we looked at post column infusion. The idea behind this is we have our chromatographic separation.

We introduce our copper phosphine complex post column. We do this with a T-junction. And we optimize our flow rate to about 500 microliters per hour, and our concentration to ten micromolar of that copper phosphine complex. Now, the idea behind this was to maximize the precursor intensity, which in this case was m/z 915, which is that that complex that contains copper, the two PPh3 groups, the cannabinoid and that oxygen insertion.

Now, I don't have the precursor information here on screen, and this is just the product ion information. But what I want to show here is the optimized MRM transitions that we developed where we used our quantifier based on the lower energy conditions, so m/z 653 and we used our qualifier at m/z 391 here based on the higher energy conditions. Of course, we replicated this with our internal standard, but I don't have that data shown here. Next slide, please.

Finally chromatography. So we wanted to separate out 18 cannabinoids. We originally characterized with 20. We removed two.

We removed the THCA-B and removed the second internal standard once we had focused down on our scope of study here. Our chromatography was done with a Poroshell EC-C18 150 millimeter column. And the important part here was we had added the free mobile phase again avoiding as much as possible the source of protons.

Although of course we do have protons from water. The total chromatography here was a 17-minute method. And this included changing the flow rate as well throughout the gradient. Next slide, please.

So in terms of the validation, I'm going to go through this quickly here. But obviously you can go back and take a look at the presentation for more details. One thing I do want to note real quick, had a little bit of brain fog here. The within and between should be flipped. So my apologies for that.

In terms of our limited detection, we can get down reliably to one microgram per mil or lower, depending on the matrix. You'll see we looked at three matrices pure solvent in this case as acetonitrile. We use ZeroC, which is a plant matrix, that Cannabis plant matrix that is cannabinoid free. And we also use fortified hemp samples, which are going to be the best representation of the real world samples.

But of course, also are going to elevate some of our background noise as well. Our within-run, between-run precision in all cases were less than 7.6%. So we had very good reproducibility there. Matrix effects, we did have some issues. You'll see that we had quite a bit of enhancement of our Delta-9 THC, relative to our internal standard. However, which I'll talk about here on the next slide,

we normalized these results to a 1% QC in our fortified hemp matrix. So even though we do see this enhancement, it's also going to be enhanced in our QC. So when we normalize this will help account for some of that. We also looked at our 95% confidence interval to try a bracket where our confidence would be around our 1%.

And again we looked at carryover and didn't have any issues there. Next slide, please. So in terms of method validation I wanted to show two examples. This first example here highlights that all of our potential isobars, are chromatographically separated with the one potential issue being our delta-8 THC. Those of you that work in this realm, you know the challenges associated with separating out these isomers, but all other isomers were baseline on resolved.

And I've also shown the presence of our internal standard, which is our delta-9 THC-d9, which is shifted. So of course we can differentiate that with mass spectrometry. We also did have ion ratios for distinguishing delta-8 and delta-9. But as I'll show you on the next slide here, there could still be some challenges with that.

Next slide, please. So when we went to authentic samples for our accuracy study, we chose six authentic samples that were near our 1% threshold. We ran our QC at the beginning and then every five samples. And the reason for that was so that we could normalize to our highest QC, which is going to be our most conservative approach.

With five of the six samples we nailed, it worked great. One of our samples, we struggled. When we went back and took a look, this sample had a ton of Delta-8 THC. Clearly a sample that would be characteristic of something that was spiked, not natural origin. And so I wanted to show the data from this.

And what you can see is our delta-8 is completely morphing our internal standard and our delta-9 THC. And so by having this issue, we realized that we needed to make sure we assessed our authentic samples that contain delta-8 and delta-9 THC to make sure this is going to work in other circumstances. Next slide, please.

So for our authentic samples, we had 20 samples that we extracted and analyzed in triplicate. 50% of these samples contain delta-8 THC. So we could assess this. Again, this is all going to be normalized to that 1% QC to help take care of the fact that we do have some of these matrix effects. We got 20 of the 19.

Let's try again. We got 19 of the 20 samples correct. The one sample we did miss, we believe this is tied back to the matrix effects and some of the inherent variability between the Cannabis matrix. This sample is S11, which you can see on screen between the delta-9 and the delta-8, it does have more than 1%.

However, we're focused on the delta-9, not the delta-8, and we should have measured less than that ratio of one. And unfortunately we measured more than one, more than that ratio of one. So we were incorrect in this case. However, we were able to get 95% of our samples correct. Next slide, please.

And the last thing that I want to touch on very briefly as I use up my 20 minutes here is quantitative. So obviously we did this semi qualitative thing. Now we need to do the quantitative thing. To do this we're transitioning to a PDA-MS instrument. The benefit of this is we can run external calibration curves. We don't need to worry about matching our internal standards between our acids and our neutrals.

There's also some benefits in terms of cost savings and availability as well. And the idea behind this was, take everything we learned from our characterization, everything we learned from the semi quant and apply it to and apply it to the quant. Next slide, please. And of course, as all good research projects go, we saw unexpected results. So our complex behavior we saw previously on the different instrumentation suddenly shifted.

We weren't seeing our complexes for acetonitrile or methanol. Interestingly enough, in methanol we did see the formation of molecular ions. We believe that this is tied to source configuration and some of the differences between these instruments, which led us into exploring the redox chemistry. This exploration is ongoing. But essentially what this breaks down to is we know that our phosphine ligand oxidizes.

And so we have started exploring the bare copper salts, looking at copper one and copper two, and looking at stability and different solvents and trying to understand this chemistry. And interestingly, we have seen this with copper one before, where with methanol, we observe the presence of molecular ions, and a previous publication that came out of my group. And so this might potentially open up a route for differentiation. As we know from a GCMS perspective, collisional activation of those molecular ions should result in characteristic, product ions or characteristic fragment ions that we could differentiate these cannabinoids.

Next slide, please. So with that I will go ahead and conclude today. Hopefully I've helped convince you that copper phosphine ion complexation can be a viable alternative method for the analysis of Cannabis.

We know that we can increase the selectivity based on the alkene structure. And cannabinoids are a great target molecule for that because the double bond keeps moving. While delta-9 and CBD were our primary focus, of course there are many delta-9 THC isomers that are available, and we identified that we had one optimal complex in this case our copper (PPh3)2 .

But of course we can continue to modify this complexation behavior as necessary. And we're working on developing both semi quantitative and quantitative approaches. And we think that this opens up a potential avenue for cannabinoid isomer differentiation, regardless of the impending legal steps here. Next slide, please. Of course, I would like to acknowledge this project was funded by NIJ and of course, my program manager, Megan, and her mentor Frances, for the opportunity to present today.

I'd also like to thank Brent Wilson at NIST for providing the authentic samples, as well as the Co-PIs on this project Christopher Zall and Britni Skillman. And all of the students that have contributed over the last year and a half here, including Alleigh Couch, Emma Fitzgibbons, Emma Hardwick, Christany Liggins and Brianna Neilan. Next slide, please. And finally, as I open up for questions with only a minute long, I just wanted to highlight I do have the QR codes here seen on screen, so if there's anything that caught your attention throughout today's presentation, feel free to check out those QR codes and check out those previous publications.

And of course, thank the Department of Forensic Science, for their continued support. With that, I would love to take questions. I don't think I cut too, too much into my five minutes. Thank you.  

Megan Chambers: Thank you, Tyler. We will now hear from our next. No. We're moving into the Q&A session, so please remember to put your Q&A, all your questions into the chat. So we do have a few. So, Tyler, given that the matrix effects were pretty large, about 145%, and seemed to have caused some issues, do you have any plans to try to further optimize to address these effects?  

Dr. Tyler Davidson: Yeah. So that's sort of a two parter here. So we took what we learned from the characterization and applied to semi quant. In the semi quant perspective,

since we're using that administrative threshold and we're going quite conservative, we're trying to avoid some of those you know potentially sticky questions. If you have a sample that's 6%, that's going to be well above that regardless of those matrix interferences. Also keep in mind because we have that normalization to our 1% QC, while we do have some of those matrix enhancement effects, we are also normalizing those results to the 1%.

So it's helping to kind of mitigate some of those. Now from a quantitative perspective of course, yes, we need to address that because we're looking at quantitative from a PDA perspective. We're now looking at the UV-Vis absorption. And we're using the ion complexation results for sort of a confirmation associated with that. We anticipate that we will be able to mitigate those issues. Of course, we're not done yet, but that's our expectation.

Megan Chambers: Thank you. Going to move right on to the next question. There are four different isomeric standards available to purchase, but only one peak is shown in your chromatogram. Did you include all four of them or only one isomer?

Dr. Tyler Davidson: I'm going to assume that this is tied to one of the, like the, Delta-10, for example. And yeah, so we looked at one of the isomers. We didn't look at all four of the isomers. From a characterization perspective, we're really interested in understanding the double bond location before getting into some of the isomers associated with that. This is an area that we, we have some interest in. And this point has been raised in the past by other presentations, other reviewers.

The challenge with that, of course, is always tying back into, as comprehensive as you make the method. Then you're building out your chromatography as well and trying to address all of this. This point has also been raised that if you think about what you actually see in Cannabis plant material, many of the delta-9 THC isomers we're exploring, you don't see unless they are being adultered or spiked onto the plant material. So we're sort of trying to strike a balance between what is naturally present in the plant material versus maximizing what we can derive from a selectivity perspective. And it's certainly an interesting balance.

Megan Chambers: Thank you. Okay. Next question. Can I ask were your source for your matrix match QCs from plant material.  

Dr. Tyler Davidson: Yes. So, we have a series of authentic samples. Some of those came from, many of them came from NIST, I should say, shout out to Brent as well. We appreciate that. So we used hemp samples that were known hemp samples that were previously characterized, either tied through authentic samples from previous work that Brent has done or tied to the CannaQAP program.

I'd have to go back and look specifically for which hemp samples we picked in this case. But the idea was those were previously characterized, and we also characterized them with an in-house validated, PDA method without the ion complexation, just the regular analysis. And the idea behind that was, we know that there is no one standardized plant matrix.

In fact, there's lots of literature out there that has shown this, you know, Cayman has developed the ZeroC as a way to – actually it’s not developed by Cayman, my apologies, but it's commercially available through Cayman – has developed the ZeroC as a way to try to address this as a blank Cannabis matrix. We've tried hops. We have seen, the CannaQAP program also had a placebo sample in there as well. So this constant challenge of trying to figure out what is, you know, sort of the standard matrix is something that that we have struggled with.

And what we found to be our best approach is to take sort of a variety of a couple different hemp samples and mix those together, trying to represent sort of the, the best of, of all worlds. But of course, it's not a perfect solution. And you see that in the matrix effects, I believe.

Megan Chambers: Thank you. We're going to try and get through a couple more. TPP is a very bulky ligand. How did you mitigate the solubility issue?  

Dr. Tyler Davidson: Yes. So, I'm not going to click back through the slides because I actually don't control the slides right now. And I don't want to put AJ through that. I believe TPP was one of the ones that we eliminated relatively early on in this process because of issues like that.

So we started with 11 copper complexes. And as we narrowed down whether solubility or whether it was in terms of the actual formation of the delta-9 THC complex, or not being able to differentiate between, you know, some of the common cannabinoids. We quickly eliminated those down to the final three complexes, which ultimately we've now narrowed down to one based on the MS/MS as I showed in the semi quant work.

Megan Chambers: Okay. Thank you. Are you using a Cannabis plant dominant in indica or sativa? Or does using either or affect your analysis?  

Dr. Tyler Davidson: Yeah. So that's a great question. And I don't I, I know what we have, but I don't know enough to intelligently comment on across the board.

So we're primarily working with sativa because we haven't explored some of these other variants. This opens up to, I'd be hesitant to give a formal definition on that. My hypothesis, if you will here, is based on what we've seen here, that we would expect similar results, because, remember, we're extracting out the cannabinoids. We're focusing on those targeted cannabinoids.

There's going to be differences in terms of what cannabinoids are present and their, you know, their contribution. Of course that's also tied to life cycle, growth cycle, maturation, things like that as well. So you know, that's, that's another variable to consider. But since we're looking at the post extract, I would suspect that the cannabinoids we're dealing with are going to be, quite similar, although the concentrations and abundances could, could vary a little.

Megan Chambers: Okay. Thank you so much, Tyler. And thank you to everyone who's putting their questions in the chat. But we now have to move on to our next set of speakers.

Dr. Ana Pego and Dr. Marta Concheiro. Ana, Ana and Marta, please take it away. All right.

Dr. Marta Concheir-Guisan: Thank you so much, Megan. So we're going to present our research that is titled, Toward Standardized Cocaine Hair Testing: Analytical Advances and Decontamination Findings. Next slide, please.

As disclaimer of course, the opinion, findings, conclusions or recommendations that we are presenting today are only ours and do not necessarily reflect those of the DOJ. Next slide, please. Next slide, please. All right. So the first question we have is why hair testing matters, right. So hair testing is a very interesting approach in forensic toxicology because it provides us a very wide window of detection.

Hair can preserve evidence of drug exposure over weeks and even months. And obviously, this is going to have this window to be able to detect exposure much longer than classical biological matrices, such as blood, that is long hours to few days, and urine that normally is about few days. So the drugs get incorporated into the hair from the blood.

The drugs are present in the blood, and then they cross, to get into the hair matrix where they get incorporated. And the most important aspect of that is that once the drug is incorporated into the hair, they stay there. And as the hair moves along, the drugs move as well. So assuming a growth rate in the head hair of about one centimeter per month, we can establish the history of drug exposure.

So then we can do a retrospective assessment of drug consumption. Because of that, hair testing is important in forensic and clinical investigations. And it has applications in cases such as in child custody, drug workplace drug testing, drug facilitated sexual assault and other context. But drugs can also be incorporated into hair due to sweat and sebum, and also due to the environment.

There's contamination because the drugs are present in the smoke or in powder and dust. Next slide, please. And this will produce a challenge in the interpretation because a positive drug testing could be because the drug was actually incorporated into the hair or because it was present due to external contamination. And the key question is to understand how hard we can wash the hair sample to remove just external and preserve the markers of true exposure.

Among the different drugs in our research, we focus on cocaine due to the high prevalence this drug and its tendency to, to have issues of external contamination. Next slide, please. So the current knowledge is that the detection of various specific metabolites of cocaine and very minor one, the hydroxy cocaine and the hydroxy benzoyl, has improved the specificity of drug consumption.

And also it’s recommended sequential washes to remove the contaminations. However there still are remaining limitations. The washing procedures across the laboratories vary widely, and also there is a lack of a systematic approach to study the effect of the different solvents, was number, duration and agitation. So this grant wants to address both analytical specificity and contamination optimization.

Next slide, please. So the goals of this grant, the first one, it was optimize the decontamination procedure to improve the cocaine hair testing. Second goal is to assess adulteration, cosmetic damage. So we want to develop a tool to grade the hair damage and quantify how much is going to be the decrease of cocaine and metabolites within the hair.

And finally, we want to evaluate the effect of different cosmetic treatments, how these cosmetic  treatments actually remove the external contamination or how they can facilitate the uptake of external cocaine. So the outcome that we want to achieve is standardized, evidence-based protocols for reliable and impartial interpretation. Next slide, please.

So in this presentation we're going to focus first on the analytical method we would develop. And then, in the in the first experiments, addressing the removal of the external contamination that will be done by Dr. Pego. Next slide, please. Next slide. So in the analytical method the first part is to develop the LC-MS/MS parameters. We use a triple quadrupole 8050 from Shimadzu, an instrument using electrospray in positive mode.

And we monitor two transitions, MRM transitions per compound. The chromatography column we use it was the Kinetex EVO column that is especially important for the retention of very polar compounds, such as some of the metabolites of cocaine. Flow rate is 0.2 mL/min, temperature 45 degrees, and in gradient mode, as you can see on the slide, the ramp. The mobile phase was 10mM of ammonium formate with 0.1% formic acid in water, and B, 0.1% formic acid in acetonitrile.

Next slide, please. Here you can see the different MRM transitions for all the compounds included in the method. I just want to highlight that in the case of the hydroxy benzoyl metabolite, the meta and the para, the transitions are exactly the same. So we had to invest quite a lot of effort to completely resolve those compounds. And the same happened with the hydroxy cocaine, the meta, the para and the ortho.

Also there are two metabolites of cocaine, the AEME and the EME. They're highly, highly, highly polar. So it was a challenge also to retain them on the column. Next slide, please. As internal standard, as you can see we included quite a lot of deuterated analogs to compensate for any kind of issues and improve the accuracy of our analysis.

Next slide, please. Here we have the chromatogram for the rotation of AEME and EME. Both of them were retained at least one minute. Next slide.

And here we can see the rest of the compounds just highlighting the separation of the hydroxy metabolites with the different forms. The para, meta and ortho. Next slide, please.

So for the hair analysis that we’re recommending, it’s a very interesting matrix. However, it implies a lot of work in the laboratory. The very first thing that we need to do is to create hair segments. If you remember, I just mentioned at the beginning of the presentation that we assume a hair growth of one centimeter per month on the head.

So like that, we can establish the historical consumption. After the segmentation, we do the washing. That is the big focus of our study. Then homogenization, typically pulverization.

After that incubation to really extract the drug from the hair matrix. Many times after that step, cleanup is necessary to remove interferences. And finally, the instrumental analysis. Next slide, please.

So in our project we started with our initial method that we've been using for cocaine hair testing, and following recommendations of previous publications in the Society of Hair Testing. So our first approach was to wash the hair with sequential washes of water, methanol and dichloromethane. After drying the sample, it was pulverized. And after the pulverization, we just got ten milligrams of the hair, added 1 ml of methanol, applied ultrasound for one hour, and after centrifugation and evaporation, we finally reconstituted to inject in the instrument.

So we have quite differing steps here. Next slide. So what we wanted to do was actually optimize the procedure. And the best approach in the case of hair, since hair fortification is kind of difficult to achieve, we used authentic hair from cocaine users that was donated by the Institute of Hair Testing in Zurich in Switzerland. From those pools, we were able to apply different procedures.

We wanted to study a simultaneous pulverization and digestion using a bead mill at a speed of six meters per second and one minute per cycle. And we were trying different numbers from 3 to 10 cycles, but also to investigate the approach that we had at the beginning with the hair. After being pulverized, apply the ultrasound with different time and temperature.

Next slide, please. So here we have the results of this simultaneous digestion and pulverization, comparing the different number of cycles from 3 to 10. As you can see on this slide here, we represent the average of the peak areas. And the part that is in orange is the highest one for most of the compounds.

And this is the one that we correspond with five cycles. Next slide. Here we are presenting the results. When we did further investigation with the ultrasound, changing the time and changing the temperature.

I know that the bars are kind of crowded, but you can see there is a tendency of going up. So what is that? The very first bars are at room temperature.

The others are at 50 degrees. So we can see that more temperature, we get a better digestion. And also they are chronologically, it's showing on this slide. So we go from one hour to four hours.

So more time also implies a better extraction. Next slide, please. Therefore, we try to then to compare the five cycles of the pulverization and digestion, simultaneous with the ultrasound and the time.

And as you can see on the slide we have different, a variety of results. However, five cycles showed the best compromise, especially it was a very efficient process for key minor metabolites. So we decided to pursue that one. Next slide, please.

So our optimized procedure was with five milligrams of hair and adding 1ML of methanol along with internal standard. And doing this simultaneous pulverization and digestion, it did simplify enormously the procedure. And at the very end, after the evaporation, we reconstituted in nano filter vials to avoid any particles getting in the instrument. Next slide, please.

So, we're going to talk about validation. In validation, we follow the guidelines of a in forensic toxicology ANSI/ASB 036 standards. And we’re just going to go through all this slide very quickly because I, I don't want to take a lot of time from Dr. Pego. Next slide. For linearity, just to highlight it was 0.01 to 1 nanograms per milligram for all compounds, but cocaine from 0.05 to 5.

Next slide. Bias and precision limit of quantification. It was okay for all compounds. It’s slightly above 20% at 22.3 for AEME in the intraday precision. Next slide. Bias and precision at low, mid and high accuracy where we see this start with criteria. So it was first successful for our compounds. Next slide. Matrix effect. It was detected.

Keep in mind it’s a very simple procedure. We were expecting that. We got ion enhancement for a couple of compounds and ion enhancement, suppression sorry, for the other two. For several of them the variability among ten different sources was greater than 20%. That's why we use all these deuterated analogs to compensate for these effects. Next slide, please. Extraction efficiency was above 70% across the board.

And the process efficiency was above 54% for all compounds. Next slide. And when we compare our method with previous publications targeting the hydroxy metabolites, we have to highlight that our panel was one of the most extensive ones. And our procedure was very rapid, using a minimum amount of sample, five milligrams very low compared to the other publications.

Keeping the sensitivity where it was in these other reports, and importantly, we developed all these methods using authentic hair samples from cocaine users. Next slide. And with this, I’ll switch with Dr. Pego.

Dr. Ana Pego: Thank you. So, now that Dr. Conchiero mentioned all the validation and everything that went through for us to achieve this method, I will delve a little bit into the actual beginning of the decontamination protocol optimization. So before we want to actually understand how to optimize this protocol for real hair samples, we must know whether contaminating these hairs with cocaine or the internal standard cocaine-d3, is going to produce the same results.

Because eventually we're going to be using real hair samples from cocaine users. And therefore we cannot contaminate it with cocaine itself, but rather with the internal standards. So the first thing was to understand whether these two behave the same. Also to understand in terms of the contamination solution, whether to use 0.1, 1 or 10 micrograms per ML, to incubate it 30 minutes or 1 hour, to dry the hairs inside the tube where the contamination takes place versus a paper towel, to have hairpieces or hair strands, and brown hair versus blonde hair.

Because we know that sometimes the hair color has a significance in the uptake of drugs. So the first thing that we understood was that indeed contaminating it with cocaine or cocaine-d3 produced very similar results. So we could go ahead with the cocaine-d3 idea. And we decided to then use a solution of 0.1 micrograms per ML, since that was sufficient, doing an incubation of 30 minutes, drying our hair with paper towel because that proved to be faster but also more efficient, since inside the tube we would still get some moisture.

The hair strands also proved more efficient, and that's actually more closely related to when we would have a real case, it will also be, provided to us as hair strands most of the time. So that was ideal. And in terms of brown hair and blonde hair, we did not find any difference whatsoever. So we went along with the brown hair, provided we had more of that drug free hair in the laboratory to use.

Next slide, please. So knowing this, we knew that we wanted to then take these hair strands, and we wanted to contaminate them with these conditions that I just mentioned and then divide them into 2 to 3 centimeter segments. And then each segment was submitted to a specific decontamination procedure. We wanted to try several things, which I will later explain to you what I mean by that.

So for each segment we first obtained a non-wash five milligram aliquot and then the rest of the aliquots that were washed differently. And this whole procedure was done in triplicate. Next slide, please.

So what we performed was this statistical test called design of experiments. And this is very useful because it allows us to systematically evaluate the influence of different parameters, because the truth is, we wanted to test various different parameters that we believe play a role in this decontamination. And this design of experiment approach really allows us to look at it from a robust and unbiased evaluation because it provides us with statistical significance of parameters, and it gives us, it's actually designs for us the different experiments, for us to understand later with the results which ones are being the optimal ones.

Right. So what parameters we wanted to evaluate. First of all, solvent type. But each time you can only evaluate three at a time. So we did two different experiments. The first one we tested ethyl acetate, methanol and water. And in the second experiment, in terms of solvents, we tested acetone, hexane and isopropanol. For the number of washes we tested one, five and nine washes.

And for, in terms of the wash duration, so for how long would that wash be performed, we tested two minutes, ten minutes and 30 minutes. And for agitation, we used none, ultrasound and vortex. Next slide, please. So here you can see the actual output that you get from design of experiments. So it tells you if you want to test all these parameters each run is going to be different.

So it gives you kind of the layout of the experiments that you have to perform. In this case, we had to do 27 different experiments. And if you see for example on standard run number one, we would wash the hair with ethyl acetate, wash it once for ten minutes using no agitation at all. If then you go, for example, for run number 12, we would wash it with ethyl acetate five times, each time ten minutes, using ultrasound as the agitation.

So these are the different combinations that we then, based on this, could input the results into the same software and get the responses. Next slide, please. So first before I go into the actual surface graphs provided to us by the design of experiments software, here you can see the absolute numbers for percent removal. So how much was actually removed, compared to that non wash segment.

And you can see already that judging by this and all the 27 experiments that you can see on the left water was the solvent that removed seemed to remove the most. Next slide, please. And as for experiment two. So we tested different solvents here. It looks like acetone is the one that is, looks like it's removing the most. We do see some outliers where actually we are seeing an enhancement in signal rather than actual removal.

Next slide, please. So the first thing that we get as a result from this, statistical software is that we see which parameters are having a significance. So anything to the right of the red line seems to be significant. And the software tries it on the linear and quadratic model as well. So judging by this graph, we can already see that the parameters that are really having an effect is, the solvent, the actual solvent, and the number of washes.

So how many sequential washes you do. Next slide, please. So knowing this what you can then do in the software is of course compare all of the parameters to each other. So we decided to compare solvents to all the rest of the parameters we tested. Since solvent seemed to be, for sure, playing a role in these scenarios.

So when we compare solvent with time, and here we are looking for the response is an optical observation, for the response which gives us the darkest red color. That is showing us the highest percent removal. So for instance, when we compare solvents of ethyl acetate, water and methanol against time, what you see is that we are getting the best response between water and methanol.

So both and somewhat close to 30 minutes, but in the range of 10 to 30 minutes is where we see the best response. And on the right, we have the experiment two. So the other three solvents, and we can see that the three perform very similar, but again towards the 30 minute mark. Next slide, please. And then when we compare again the solvents to the number of washes, what we see is again in the first experiment, water or methanol are performing well, and they are performing well towards the highest number of washes.

So the more time we wash it, the better the percent removal. And the same goes for isopropanol and acetone. Also towards the nine washes mark. Next slide, please. In terms of agitation, so for the first experiment, again, the same two solvents remain. And it did see we did seem to achieve some of the best results using no agitation. But for when using isopropanol, acetone or hexane, you can see that acetone performed better using, agitation, either ultrasound or vortexing. So we're getting that Y- or U-shaped type of graph, where you see the edges is where the dark red colors are present.

Next slide, please. So judging by what we have seen so far, the most efficient solvents they do seem to be water and methanol. The most efficient agitation being ultrasound. Because between none and ultrasound, we were still getting very good responses. And time between 10 or 30 minutes. And the most efficient repetitions nine times. And this is important also for the future of this grant because the next step we're going to do is use again a software called simplex centroids where we're going to actually input different mixtures of these solvents and see what performs best. Because perhaps it's not that solvent alone but rather a mixture of the solvents together at different percentages of each.

Next slide, please. So one of the things we did see to have the bigger effect, and Dr. Concheiro already mentioned that, is indeed the sequential washing. And this is very important because until now most of the research groups what we do is a sequential wash of once or twice of different solvents. But what we were testing here is using the same solvent.

So on the top you see water and the bottom methanol. And you can see that indeed the percent removal is much greater as we continue to wash sequentially. So now we're going to delve more into these to understand whether at some point we are reaching a plateau or not. And of course, this is bound to change when we will eventually perform also real case hair samples where we will need to be careful to not start removing the drugs that are actually inside of the hair.

So it will be a compromise study here. Next slide, please. So to conclude, we have developed a fast, sensitive and specific method to determine cocaine and these 10 very important metabolites as they allow us to really distinguish between external contamination and ingestion of the drug. We used only five milligrams of hair, which is really very little amount. But we still kept those limits of quantitation that we need also according to SoHT guidelines.

Water and methanol seemed to be the most efficient. And sequential washes are indeed a critical factor. Next slide, please. So yeah this concludes our presentation. We would just like to thank NIJ and Megan for inviting us here today and for the grant, obviously, and John Jay College. And thank you all for the attention.

Megan Chambers: Thank you, Ana and Marta. We're going to move into the Q&A part of our presentation. And just a reminder to put any questions that you have in the Q&A box. All right. So first question for you guys. A 2016 study in Analytical Chemistry found that washing can push external cocaine deeper into hair. How does your method distinguish this from cocaine incorporated into the hair and reduce false positives?

Dr. Marta Concheir-Guisan: Yeah. No that's a great question. And actually that's the thing. This will be the second part of our research. Now we are taking hair that has no cocaine. And we are doing the contamination and trying all these different steps. If you remember we did the contamination with the deuterated cocaine because later we are going to contaminate hair from cocaine users with deuterated to be able to tell that difference.

Ana, I don't know if you want to add anything. No. Exactly. That's exactly the point. When we have real hairs, and that's the that's our next step, we will then be able to see whether we start removing some of that cocaine which is present, but also whether we are inserting some of that cocaine-d3.  

I believe that paper you're mentioning is a MALDI study. So they actually saw, image wise, what people have seen that it migrates inside the hair. We will be able to see whether we also find some cocaine-d3 on our regular procedure.  

Megan Chambers: Thank you very much. Next question. Do you plan to also investigate the efficiency of phosphate buffer as a wash solvent? Dr. Marta Concheir-Guisan: All right. Yeah. That's, we discuss about that actually long time ago. We may.

At the end, we just decided to go with water. In terms of the pH and also because in part of this investigation, we actually are injecting the washes. The phosphate buffer, even if to evaporate that, they would prove a little bit problematic injecting in the instrumentation. So for the moment we just keep with water. But definitely it's something that we may explore in the future.  

Megan Chambers: Okay. Next is, this method was optimized for cocaine and its metabolites. But how do you think it would perform with other analytes of interest?

Dr. Marta Concheir-Guisa: The method is, absolutely, the cocaine metabolites are the target compounds. It's a target analysis. We give those transitions that I was showing to the instrument. So it is looking for those specific transitions for cocaine metabolites. When we need to analyze or add other analytes to the method definitely will need to add those methods and check that the chromatography work or adapt that accordingly. I don't know Ana if you want to add anything.  

Dr. Ana Pego:  Yeah, yeah. If I may add in terms of like the extraction itself, we actually in our group, we do perform that simultaneous extraction and pulverization for other drugs, and it works as well.  

Megan Chambers: Okay. Next question. Are there plans to determine how the extraction method performs on treated or dyed hair? Definitely. Yeah. So we were highlighting the outline. Definitely. Because all these treatments for sure are going to have an effect on the matrix. Ana please, if you want to add something.  

Ana Pego: Yeah. No. Exactly. And it was part of our preliminary studies before this grant. We actually tested this in hairs that we treated ourselves in the lab.

So we have, we do have some preliminary data for this type of study, using hairs that have been dyed, bleached, straightened and some cosmetic products like oil, gel and so on. And we plan on continue to do that in the next phase. Yes.  

Megan Chambers: Okay. And let's do one more question. When validating your method and experimental design, will you be testing the quantification of cocaine concentrations at different time periods or different distances of hair segments as part of your study, such as the identification of cocaine at different segments on the hair?

Dr. Marta Concheir-Guisan: I'm not sure I really understood the question.

Megan Chambers: It looks like they're trying to ask, are you going to look at, like, different how you had segmented the hair, are you going to look at the amounts of cocaine in different segments of the hair?

Marta Concheir-Guisan: I mean, in our study, we are contaminating the whole strand with the same concentration and then creating the segments. All segments have the same contamination. And then these different segments are washed differently to be able to compare. But we always take aliquots of that and wash to understand that the contamination is the same across the board. On our authenticates, it depends how much information you want to get from the hair strand. Each segment will represent a different time period. One centimeter a month ago, two centimeters two months ago, and so forth. So definitely segmentation is a part, a crucial part, of hair testing. So yeah, it will be done.  

Megan Chambers: Okay. That makes sense. Thank you both so much. We're going to move on to our next presentation by Dr. Rajesh Sardar. Rajesh, go ahead and take it away.  

Dr. Rajesh Sardar: Good afternoon, everyone. First of all, Megan, thank you very much for the invitation. It's been a pleasure to give a talk. And I should also thank Tyler for giving such an outstanding talk before to set up the stage that why mass spec is so important in drug detection. So today I'm going to talk about a multimodal detection system, which we developed already. And we are in the process of testing the samples, real world samples. Next slide, please. So, before I go, so this is the disclaimer for this one. So this research was funded by the NIJ.

And so any opinions, findings and expression is all my responsibility, not the NIJ’s or Department of Justice. So please, next slide. Before I move, sorry to bring this slide. And, this is one of the heartbreaking things I ever encountered in my life, that most of the work I'm going to present here is designed by my graduate student, Sumon Hati.

He’s an outstandingly genius student. And unfortunately, some minor illness during his visit to India, he passed away. So he was very young, 29. And it was very unfortunate this happened. So I thought that I will dedicate this talk to his memory. And he was a very good friend of mine. So please, next slide. So, why we need to detect the drug in the drug overdose patients. Mostly the focus of our research is forensic toxicology.

So, you know, there are some improvement. I saw that in the past. This year for sure. But over the years, we have seen that there is a serious rise of synthetic opioids and other drug consumption. And then corresponding to the drug related death rate. That is, and then one of the part is, I live in Indiana, and if you look at this map on the left side, left top corner, it doesn't look good for Indiana for sure as I can tell you.

And then if you see there are many, many more opioids could be synthesized from the known substance because chemists are smart, and this type of research can be done in someone's garage. So to overcome the challenges. So what is our plan. Please, next slide.

So if you look at also the other statistics, which I have been, federal drug overdose after Covid-19, United States, that also increased because of people who are staying at home. There are many other challenges, but opioids is not that easy to, you know, detect because one of the challenges, small amount of, either fentanyl or carfentanil could be very lethal. So the concentration of opioids in blood itself is very low. On top of it, during the metabolic process, the opioids undergo decomposition. So intact drug detection is quite challenging. So most of the detection method, I will talk a little bit more, heavily relies on the detection of what is called the metabolites.

So our goal here is to detect the drug in intact form. Right. So we need to extract drugs. Or we need to design the platform which can detect opioids or many other drugs like you know, benzos or any type of, other type of drugs directly in plasma. So please, next slide. So what we are talking about is the current gold standard analytical method. And the gold standard analytical method includes gas chromatography mass spectrometry, which is basically very. But, the challenge is it requires large sample volume. Not only that, and obviously, it has to be done by a specialized chemist or analytical chemist because of the sample purification.

And other than that, liquid chromatography is another, is a gold standard technique for forensic tox analysis, specifically that drug from by human bio fluid. But we also know that there are other challenges associated with it, specifically low concentration drugs. It can be lost during the purification process. And sometimes the technique suffers also low sensitivity in a complex acid. Another assay is called ELISA, which is enzyme linked immunosorbent assay. This is heavily used for fentanyl analog and synthetic cannabinoid detection. But there is a big issues with this ELISA technique that it has a very severe matrix effect. ELISA technique heavily relied on the fluorescence based or colorimetric detection.

So if the drug has a very, you know, or the blood is coagulated or it is kind of dried, it is very difficult to even get the signal from the drug because of the other matrix effect. So there is definitely an unmet need to develop a widely applicable technique. And I'm talking about the technique, which is a quantitative analytical technique. It's not like we want to detect it, but we want to quantify from a diverse range of bio fluid. So please, next slide. So our research addresses a critical question, how do chemists, forensic science or healthcare professionals or law enforcement community can find the technique which can provide a highly reliable quantitative tox sample analysis. And then the data we’ll generate can be useful for the code, in the code. So to do that, we propose to test a central hypothesis the nanoparticle.

So we are talking about surface-enhanced Raman spectroscopy (SERS) technique development. The reason is SERS is a very sensitive technique. And we are talking about a sensitivity which is parts per trillion. Yeah, that is very standard sensitivity we can achieve because our, the way we designed our sensors basically. And then from there we can also determine the sensitivity and selectivity using the chemometric analysis and so on and so forth. But at the end, what we really want, we want to design a sensor which can provide the concentration and also the specificity with a single atomic level. So our goal is to combine both the surface and Raman scattering along with the mass spectrometry together to determine the drug concentration. On top of it, I'm going to show you a system where we can do a reverse phase chromatographic separation to extract the drugs on column.

Here, on column means if you think about this microneedle these needles, we are going to use on column synthesis and do the detection. And finally, our goal is that we can design these sensors in an array. Basically, this is a 96 well plate array. So in this, instrument which has automation, so we can use our sensors in an automated stage. So we can reduce even time, you know, time lag for the sample analysis. So with that, please, next slide. So here we are using the SERS. So I've talked about the SERS is a unique technique because it gives the molecular fingerprint.

And it is a nondestructive technique. But there are other challenges with the Raman technique because one of the challenges is that if the drug is almost, you know, regioisomer, or, they have a diasteromer. It is very difficult to distinguish between these drugs. Or different fentanyl analogs could be challenging until unless we perform any type of chemometric analysis.

So next slide, please. So, we are going to use a nanoparticle-based system. Here we are. We are going to use nanoparticles, which provide a property called localized surface plasmon resonance. And this unique property creates a very strong electromagnetic field around nanoparticles. And because of that electromagnetic field where these nanoparticles are very responsive to Raman scattering intensity.

Now regular Raman, which is called just the plain Raman scattering is very weak because one of a million photon will be your anti-Stokes photon that can be used for the Raman scattering. But we are using a technique called surface-enhanced Raman scattering that will amplify that signal. Sometimes you know million volts, million is a very easy way to amplify the signal.

So we are going to use that one. So let's move to the next slide too. So the design of our sensors. So our sensor construct design. A mold which, is a showed in A. So we poured the PDMS inside that mold. And then we let the PDMS to cure. Once the PDMS is cured, we remove that PDMS and then get a kind of a well plate microneedle.

You can see that, on the top right, you can see that there is my student is holding. So this is the 96 array of these microneedles. But once we generate these microneedles then there are a lot of designer tools we have to take. One of the designer tools is that we want the surface of the micro needles and the top surface is not atomically smooth. Do you want the rough surface, the roughness increases the Raman activity. And then we modify these microneedles with nanoparticles. So here we use gold nanorods. So we functionalized the gold nanorods with polyethylene glycol. One of the reasons is polyethylene glycol removes any type of biomolecules which are not required.

So they basically remove the fouling effect. So now all these data you can see that the top right one, you have C,D,E, which are the different scanning electron microscope image of the microneedles we fabricated there. So their top is somewhere around couple of hundred nanometers with, so that would be the perfect area for our Raman analysis.

So I'm not going to go into detail of this one. So one of the things is you can see the bottom right. These microneedles are quite highly reproducible. And the SERS signal is quite homogeneous throughout it. And if you look at the bottom left you can see that the, it’s very hard to see here, but the way we add these nanorods, which is made with gold basically, they kind of stand up. And once they stand up you will definitely, they improve, it enhances the Raman sensitivity.

That is the whole goal. And you also can see that once we coat the needles, you can see that there is a kind of blueish gray color. So that blueish gray is telling that is our nanoparticle absorbed onto the microneedle surface. Next slide, please. So, with that, we analyzed, we have a collaboration with, Indiana University Emergency Department to get the samples. And we also have a collaboration with Indiana State Department of Tox here. We get drug overdose based on samples.

And here is the goal here to detect the drugs, unpurified samples, means directly from plasma. So please, next slide. So here we, so this is the IRB approval for this project. We analyzed two different drugs here. First we analyzed the drugs, and then I'm going to mention about, what are the drug composition basically. So the left one is your Raman signal from, I think, how many, ten. I think ten samples we analyzed initially, and then each drug, we calculated the corresponding, what is called corresponding concentration over there. Okay, so each drug we calculated. So one concentration was determined from the Raman and one corresponding determined by the mass spec.

So I collaborate with professor Nick Manicke, my colleague. He’s an expert in the mass spectrometry. So the samples we analyzed, it is, you know, the same samples were given to him but it is a blinded study, so his group doesn't know what we are looking. Once the analysis is done, and then, I look at the samples and then we meet in a month, every month, to discuss the data. So, as you can see, this is an excellent fit. And based on this, we have the concentration from both the samples. And you can see the concentration is well matched.

So it tells that our Raman technique is quite, you know, good. Not only to detect specific type of blood, drug in a blood samples, but also we can determine the precise concentration. And we also correlated this data with a healthy control. Healthy control means these are the patients who have no history of any type of drug taking, even prescription medicine.

So we know. And the three star means we can definitely distinguish between the drug overdose patient from the healthy control and for the fentanyl sample. So next slide, please. So moving forward, this is the kind of, this is the raw mass spec data where we look at, there are samples which has just pure fentanyl, there are samples which has also alprazolam. And there are two samples. There are I think there are three samples, which have both alprazolam and fentanyl. So using Raman we can identify both type of drugs in the same samples or individual drugs in other samples. And then the mass spec data not only confirms the concentration. The mass spec data we also used as a confirmatory test because of its, you know, proven efficiency.

So next and next slide, please. So with that, I'm going to move the last five minutes about talking about that technique. So now whatever I discussed so far, what happened is that when we have the blood samples, we use ten microliters of pure blood or plasma and drop just on top of the sensors, which is the needles.

And now another challenge happened because we can do the Raman technique very easily, but when we try to do the mass spec of the same needles, we have a very, we struggle because mass spec, the sensitivity is low, first of all. And second thing is, in order to do that, this paper spray ionization mass spec, which my colleague Nick has developed, so that technique requires to apply the high voltage. So we cannot apply that voltage with the needles because the, we call it a micropillar, they are non-conductive. So our goal is to make it as a conductive substrate so that we can apply the potential there. But now once we make the conductive substrate, another challenge happened how we are going to concentrate the drug because the drug can go in any direction.

So we came up with a novel idea. So here we are using a micropillar, which have the three different molds at once. First mold is using the pores of the micropillars the way we want to functionalize the pores that can act as a silica modification in a long hydrophobic chain of modified silica as the column for reverse phase chromatography technique. That is what we want to do. The second one is once we extract the drugs in column extraction, we are going to use the Raman. And we are going to use the mass spec to analyze the drug from the same pillar without doing any surface modification. So the fabrication strategy is a little bit unique. And this is what I was telling.

This is where my student, Sumon who designed, developed this technique, countless hours of literature and practice and testing. So we use PDMS, we pour the PDMS with the acid called citric acid monohydrate, we call it CAM. So once you mix this one, we pour inside the mold and then this mold we then cure it in the oven at a particular temperature. And then the whole mold just easily can be peeled off. And then while we peel off the mold, we add this whole entire mold in ethanol and sonicate for two hours. So what happened. The citric acid monohydrate is soluble in ethanol.

So once it is solublized in ethanol, it just, wherever it accumulates it creates the pore. So now you can say okay, how reproducible it would be. The reason is we use very fine powders of this one. We have our own method. And you know, there's definitely going to be reported in the literature when we publish, but it is a highly reproducible data.

So this is the typical scanning electron microscope. You can see that the pores is literally extended from top of the pillars. And you can see there are tiny, tiny pores there. All these pores are all interconnected. Just thinking about this is a micropillar of a chromatography column. Next slide, please. So using this pore, we modify our surface with this called long chain.

So this was the initial study we started with because the modification of the surface of these PDMs is very essential. Because without modification, we cannot extract the drugs, because for reverse phase, we need a hydrophobic stationary phase and then polar mobile phase. So our hydrophobic stationary phase, we want a polyethylene glycol with a double bond. Because if it is too hydrophobic, then what happened? The drug is not, the plasma is not going to go through, or the blood is not going to go through. So we tried a technique called DCC coupling to see it. But when we try to do the mass spec of that sample, the mass spec data was not very clean.

And then there are a lot of you can see that barely there is any fentanyl signals, but there are all other junk present there in the mass spec. So what we decided we are going to modify this technique, please. Next slide. So the modification comes with a unique approach. So what we do is, we take this porous PDMS micropillar and then we perform the plasma cleaning. Plasma cleaning produces hydroxyl group on the surface, which can be further treated with water, hydrogen peroxide, HCl with a particular concentration. Once we have this hydroxyl group on the surface, then the salination chemistry can be performed to attach it. Now, you can see that there are a lot of CH2 linkage on there.

And at end we have the OH group. So it is a very fine balance between the hydrophobic tail and the hydrophilicity, so that the drug, so that the plasma can soak through the entire pillar. We try to define the weight percentage of Silwet. So we call this one Silwet. This is the commercial name. So we determined that for 25 weight percent, this is the best. And that six hours of salinization is good. And, once we perform this one, and then we also determine what percentage of drug we, drug loss happen while we are adding this plasma to one. We use 200 microliters of plasma to soak through these needles.

And we observe that, okay you know, as we increase the Silwet percentage, we get more and more drug loss, which we do not want that. And using this one, we use acetonitrile and formic acid, our spray solvent, which reduces the plasma matrix effect. And we are able to distinguish between fentanyl and alprazolam very easily. Next slide, please. So now we have, our microneedle is completely optimized. Sorry I just put a timer for myself. So optimized. Now we have developed, we first we started with the fentanyl and three different benzos, alprazolam, etizolam, and bromazolam to determine what is the limit of detection, linear range and limit of quantification.

As you can see, this is in parts per trillion level. What I see the quadrillion level. This is the Raman data which is in femtomolar concentration range and it has a very large, dynamic range as well. So this is the Raman data we generated for all the drugs to quantify. Next slide, please. And then we have our mass spec-based data.

So the mass spec-based data here, I just presented two of them. Our data. One is Carfentanyl and one is Etizolam. As you can see, the left two column is our PDMS micropillar-based mass spec, and the right one is our standard paper spray mass spec. And evidently you can see the limit of detection of our PDMS micropillar is much better compared to your paper spray mass spectrometry. So it is consistently provide the better limit of detection, limit of quantification and linear dynamic range. So that tells that, okay, you know, our whole effort towards this, constructing this unique micropillar is what, to move forward. So please, next slide.

So, with that, our next plan is we started analyzing. We do not have, all the data yet. So, I would like to conclude the presentation is that we have fabricated this new porous micropillar construct that can act as a stationary phase for reverse phase chromatography technique to pre-concentrate the drugs from the human plasma. It serves outstanding sensitivity and then linear dynamic range. And we are now looking to analyze different types of regioisomers and diastereomers. So that is our target because even mass spec struggles to distinguish between these two. But we believe using the chemometrics, we should be able to distinguish between the regioisomers and diastereomers.

And we have started analyzing the 18 driving under the influence blood plasma samples for Department of Toxicology. And then, we have few analysis we haven't done the comparison with the ISD scientists because it will it is a blinded study. So once we are done, we will figure out how our test went. And with that, I would like to acknowledge, my next slide, please. Acknowledge my group. First of all, Sumon’s unconditional, hard working for this project to bring it up to this phase. And then, Ashlynn Meadows, she’s a master’s student in Manicke’s group she’s ready to graduate. And to one, this is, you know, this incident happened in the lab, my postdoc and my other grad student, they stepped up to continue performing this project.

And then, definitely my other collaborator Nick and John to help us with this project. And finally, I would like to thank Megan and Dr. Scott for helping, you know, with the funding to carry out this research and obviously NIJ for the financial support to conduct this toxicology research. And I would like to thank all of you for your attention. Thank you, Rajesh.

We're going to now move into the Q&A part. So first question for you. Also just a reminder, please put your questions into the Q&A box. Okay. First question, how do you use the micropillars for mass spec analysis? Do you extract substance from the pillars or are you using a direct analysis such as DART-MS?

Very good. So, think about, this is the micropillar, right? The micro pillar is fully coated with this,  time instead of gold, we coated the whole pillar with silver. So once we coat the whole pillar with the silver, it is fully conductive. So we applied, we use the alligator clip to apply the potential at the bottom of the pillar to let the solvent to spread. And it goes inside the nozzle, basically, mass spec nozzle. Does that make sense? Yes, that makes sense. Okay. Next question is, I noticed you're using a log function for your calibration curve. Can you give the reasoning? That's a little unusual for tox work. Right. So just why we are doing this. The idea is if we are using the small amount of some,

so if you look at the data we used for mass spec, it is a very limited range. But if you consider the data for the Raman, it is quite a larger range. One of the reasons we took such a large range of Raman, because if the drug somehow decomposed, and if you have a very minute amount of drugs in the tox samples, can we even detect that minute amount rather than detecting metabolites? Because that is a very standard approach for LC-MS-based method that detecting the metabolites. We want to detect the intact and decomposed drug. So why we took it, depending on when the samples was collected or not, the decomposition could be different. So we have a lot. And we want to see whether our technique can be useful for a large, dynamic range.

So that's what we chose. I understand nanograms, you know, still nano, which is ppb level concentration, could sometimes be lethal for very particular. But like when we are designing the calibration, we say, okay, if alprazolam we are designing in ppb to other range, why not do everything for rest of the drugs. So that was the reason. There are multiple.  

Megan Chambers: Thank you very much. Since your method can analyze multiple drugs in the same plasma sample, how do you make sure one drug does not interfere with the detection of another?  

Dr. Rajesh Sardar: Okay. That's a very important point basically. So, you know, as I mentioned, that even the fentanyl analogs if you have it, sometimes Raman could be very difficult. But what we do is, in the case of the Raman, one drug doesn't interfere with other one because we are looking for the molecular signature fingerprint for each drug. So their signature will come differently. It might be a little complex to know precise concentration. So that is why we are going to also do, when there is an unknown sample analysis, we are also going to perform the chemometrics to distinguish drug.

And then go back and look at the Raman data again to look at the signal, exact signal, to see how much there. I cannot give the right answer until, unless we have the analysis of real world samples, quite a large cohort like, you know, 50, 60 to 200 cohorts. Then we can determine whether it is influencing the one drug to other one or not. But I guess this is definitely, we're going to discuss in the next meeting.  

Megan Chambers: No, that makes a lot of sense. All right, another question. Can the surface be amenable to thermal desorption coupled with GCMS?

Dr. Rajesh Sardar:  You know, that is one thing. Even Nick asked that question, right? Is can we make the surface stable? I don't know what would be the temperature. But if anything goes below 70 degrees, I would say the answer would be no. The reason is that silver is going to oxidize at that stage. If it is very close to high, you know, high, very high temperature, you know, 70-80 is the max temperature. So the simple answer would be maybe not.  

Megan Chambers: Okay. And one more question. Do you have plans to use these micropillar designs with other drugs of interest/analogs. Oh yes. Yes. Yeah. That is true. So we have narrowed down almost 21 drugs. That is, Megan knows the names of all the drugs. But we started with these seven to demonstrate our proof of concept and, then, you know, we want to file the patent and publish a paper before we get scooped. So then we are going to do a larger cohort study. Our target is to almost have 200 plasma and blood samples to analyze and then multiple drugs in each cases. So, yes, this is the first stage, stage one. I would say in our projects that we are moving into stage two, which is mostly the blood and plasma sample analysis.

Megan Chambers: All right. Thank you so much, Rajesh. We're going to now move on to our final presentation of the day. Dr. Ashley Fulton. The floor is yours.

Dr. Ashley Fulton: Thank you, Megan, and thank you for inviting me to this seminar. Today, I will be discussing our research. I know with a lot of letters on the on the, title, but basically it's our method for fentanyl detection in vapor mode. This project was conducted in collaboration with Florida International University, under the co-pi of Lauryn DeGreeff. Next slide. Again, I have to say the disclaimer that has been said all day. This project was supported by NIJ, and the opinions, findings and conclusions or recommendations expressed are those of the authors and not of the DOJ. Next slide.

So fentanyl. We have been hearing about fentanyl for a while now. It is a very highly potent schedule two synthetic opioid that is 50 times stronger than morphine and heroin. Small amounts can cause overdose, although it is used medically as a sedative and for pain management for those who suffer from cancer. Illicit use has increased exponentially. As you can see from the graphic to the left, between 2013 and 2022, the deaths fentanyl has caused has increased substantially.

In fact, in 2023, there were 77 million pills containing fentanyl with about 55 million pills containing a fatal amount of fentanyl. This is also increased through the advent of fentanyl analogs or fentalogs that range in potency, with carfentanil being the highest potent, which is about 10,000 times more potent than morphine. Next slide. So let's get to our motivation and our approach to this problem. So with the increased illicit use there is increased contact by first responders. That's police officers, medical personnel and firefighters. For naval purposes, our naval fleet could come in contact with that as well through unknown substances that are in crime scenes or at the scene of an emergency event.

The DEA recommends that the unknown substances be handled by trained professionals. This, of course, is timely and costly. So a solution to this is to develop a non-contact method of detection. Next. Current methods all require well, most of them require bulk handling of the sample material. This increased handling of the material would of course increase the chances of an inadvertent overdose. Next. Our method would use the vapor of the unknown substance to identify if fentanyl is involved. So we have been focusing on the use of a handheld ion mobility spectrometer because of its commercial availability. And it's quite easy to use no matter your background, if you have a scientific or nonscientific background, and its ability to detect trace narcotics and explosives.

The device can also be used in particle and vapor mode sampling. So how do we detect fentanyl? Next. Fentanyl has an extremely low vapor pressure as you can see. So we cannot directly detect fentanyl in the vapor mode. We will have to identify a surrogate. Next slide. This brings us to our project design where we developed a generalized approach to identify and detect low volatility substances. For the initial step, we must elucidate the vapor signatures of fentanyl and fentalogs to identify a surrogate.

Then, using that surrogate develop, excuse me, develop an alarm for a handheld instrument. Next, we generate the vapor to mimic the drug's vapor without using actual drugs. And next, we test the capabilities through limited detection and false alarms and mixtures and also street samples as well. Next.

So first, the vapor signature was determined using solid phase microextraction, with analysis using GCMS. Optimized method SPME-GC-MS was developed and tested on fentanyl, fentalogs and adulterated fentanyl. Through that, we did comparative analysis of all of those headspace analysis and determined that N-phenylpropanamide, or NPPA, was the ideal target analyte for fentanyl. NPPA is a known degradant and chemical marker that is formed through hydrolysis. Next slide. As you can see in the molecule of fentanyl layer. Next slide. So next I was able to incorporate a signal for NPPA within the mobile trace,IMS, where the retention time for NPPA was identified at 5.89 milliseconds. Next slide.

This alarm was then tested for a limit of detection through the vapor generation of NPPA at different concentrations. The instrument was allowed to be tested within five rounds. And based on that, we determined that the limit of detection was five nanograms. Next slide. Next, we wanted to see if the alarm would sound from other substances that could be identified at the scene of the crime. But none of them were identified as NPPA using the IMS alarm system. Next slide. You can do all of them next. Next. Yeah. Next, I sampled adulterated, pharm-grade fentanyl in different ratios.

1 to 1 equal amount and, 1 to 10 fentanyl to adulterant. And with all of them NPPA was detected and all mixtures. This is based on weight, solid amount, where the solid amount was then tested through vapor detection. And looking at even more adulterated systems NPPA was still identified in the bottom left. Next slide. And then, in collaboration with the DEA Special Testing and Research Laboratory, I was able to take the instrument to identify if NPPA would be in those systems. And for the three samples that we were allowed to test, they were identified in all three samples. Next slide. Our next collaborators, Maryland State Police Forensic Sciences Division Laboratory, where the samples were more adulterated.

The system was only able to identify NPPA in one sample. That is with increasing the sampling time. So what it told us is that the dilution significantly hindered the detection of NPPA. So next slide. That meant we needed to add one more step to our design. And next. That is pre-concentration of the vapor. If we can concentrate the NPPA analyte and then detect, it would increase our capabilities of identifying highly mixed and adulterated samples. Next slide. So to do that our pre-concentration method is the silicon nanowire array that has the ability to collect analyte onto silicon nanowires. Where a current can then be used to desorb the wires and allow for a bolus of analyte to then go to the detector for analysis. I think this is the animation for this as well.

Next slide. So oh sorry. Can you go back? The wires can also be functionalized even further through a polymer coating. The polymer coating can be used as a means of collecting analytes  more specifically. Next slide. To identify the proper polymer coating for fentanyl, a quartz crystal resonator was coated with five different acrylate based polymers. And then the QCM was used to identify the amount of analyte that is collected and desorbed to determine the optimal polymer coating. That polymer coating was then deposited onto the silicon nanowire array for analysis. Next.

Of the five acrylate base coatings, the top coating was determined to be the EGMEA acrylate coating that was then deposited onto a silicon nanowire array. Next slide. And used in our initial study where the silicon nanowire array was placed into a headspace chamber in the presence of NPPA for overnight and then placed into the silicon nanowire desorber onto a GCMS. And it was shown that there was an eight fold increase in NPPA detection while a benzyl acrylate coated silicon nanowire array had no affinity for NPPA.

Next slide. From there. We wanted to see if NPPA in the presence of other adulterants and cutting agents would still allow for a detection of NPPA. For this we use the TVGen. It’s a mobile vapor generator that was developed here at NRL. With this, we were able to generate vapors of NPPA with the cutting agents listed here.

The vapor was then introduced to the silicon wire array for a certain amount of time and desorbed onto the APCI-LIT mass spec for analysis. With this required, of course, parameter optimization and we were able to achieve that as well. Next slide.

So here I wanted to depict a complete run of adsorption, equilibration and desorption. That five parts per billion vapor of NPPA and analyte. The chosen one for this is methyl benzoate, which is the target analyte for cocaine.

As you can see, the top one is the total ion chromatogram. And for the following two, I was able to isolate the NPPA major ion and the methyl benzoate major ion. From here you can see that for NPPA during adsorption, there is no, there's no sign of NPPA going to the mass spec.  

It is assumed that it's being adsorbed onto the silicon nanowire array at that time, but with methyl benzoate, you can actually see it. And then when it's desorbed, you can see the NPPA, but you do not see the methyl benzoate. So with that, we can see that the affinity that NPPA has for the polymer is very significant

and it is capable of pre-concentration. Next slide. So I also looked at other cutting agents as we saw in the previous table. And with all of the cutting agents, I was still able to identify NPPA in the desorption. They had no adverse effects when pre-concentrating, using the silicon nanowire array.

Next slide. And with that, I wanted to summarize everything. We were able to identify a polymer coating for pre-concentration and use the functionalized silicon nanowire array to sensitively identify NPPA in the analysis of APCI-LIT mass spec in the presence of common adulterants. This is significant because we were able to detect dilute samples of fentanyl and decreasing the sampling time.

Going forward, I would like to use analysis using the IMS and test the pre-concentrating capabilities. Next slide. Again this is funded by NIJ. I wanted to thank Megan and Frances for their continued support. And my collaborators at Florida International University as well as our collaboration with the Maryland State Crime Lab and the DEA Special Testing Research Laboratory. And I'll take any questions from here.

Megan Chambers: All right. Thank you, Ashley. We're now going to move into the Q&A part. So the first question, also reminder, please put your questions in the Q&A box. All right. So the first question is if the nanowire array can pre-concentrate the vapor signature of the target analytes, could it also concentrate the vapor signature of the drugs themselves. Or is the vapor signature of the drugs too low to allow for concentration?

Dr. Ashley Fulton: So you're asking if it can concentrate all of the ions within the signature.  

Megan Chambers: Correct. Okay.  

Dr. Ashley Fulton: So I haven't been able to see that because based on selecting the polymers affinity for the target analyte. And with that we've only used the target analyte in analysis. I have not used a complex or the parent compound to see if other ions are going to pre-concentrate as well, using the same polymer.

Megan Chambers: Okay. Thank you.  

Dr. Ashley Fulton:  That is a, that's a good question, but we haven't looked at it yet.  

Megan Chambers: Okay. Next question. If NPPA is the surrogate for fentanyl due to its vapor pressure, is there any risk for false positives if NPPA is representing a drug other than fentanyl?

Dr. Ashley Fulton: So we did test. No, we haven't. That's a good question. NPPA is unique to fentanyl. So it would not be used for any other, used as a target compound for any other drug. But because it is a degradative compound and a marker for fentanyl, I can't see it being used for anything else other than fentanyl.  

Megan Chambers: Okay. Thank you. The next question is, you showed the LOD was five nanograms but it was in some container. Do you think this will translate to reasonable scale to detect from outside a container for safety?

Dr. Ashley Fulton: That's a very good question. And that was a concern of mine as well. We haven't tested it outside of containment due to safety reasons. So, but it will increase dilution. That's why pre-concentration is so, so needed. Yeah. So that is the reason for the concentration because it will cause dilution.

Megan Chambers: Okay. The next question on the IMS, you showed that cocaine and oxycodone had a negative effect on peak height. Given that, do you plan to test the nanowires against those compounds?  

Dr. Ashley Fulton: Yes. Actually our, our full project does include these other target drugs and finding polymers that have affinity to those as well. And, yeah, we are looking into it.

Megan Chambers: Okay. There are no more questions in the chat. And so with that, thank you, Ashley, very much. And so we're now going to wrap up our webinar for today. This ends the Seized Drugs and Forensic Toxicology session. Thank you to all the attendees for spending time with us today. And thank you to all five of our presenters.

If you enjoyed this and would like to hear more NIJ-supported researchers, we have sessions every day this week. Tomorrow's session is on Forensic Biology and DNA. If you are interested, make sure to register separately for each day's session at NIJ.gov. For any follow up questions, please contact [email protected]. We look forward to your participation in future events and hope you enjoy the rest of your day.