27. QCM-D technology – a microbalance to scrutinize molecule surface interaction

What is Quartz Crystal microbalance with Dissipation monitoring? And what’s the deal with all the harmonics?

In this episode, we talk to Fredrik Pettersson and Erik Nilebäck, both Senior Application Scientists at Biolin Scientific, about the QCM-D technology. Erik has a MSc in Engineering Biology and Devices and Materials in Medicine and a Ph.D. in Bioscience, and Fredrik has a MSc in Biophysical engineering.

Both Erik and Fredrik have extensive experience working with the QCM-D technology and in this episode, they share lots of useful information and insights that they have gathered over the years. The conversation starts with the basics, and we talk about what QCM-D technology is, what information it provides, and when it is typically used. We then move on to talk about different versions of QCM:s and their respective strengths and weaknesses. Finally, we go into how a QCM-D measurement is run in practice, and Erik and Fredrik share some useful tips and tricks on how to get the most out of this surface-sensitive technology.

Thanks for listening! If you are interested in surface and interface science and related topics, you should check out our blog -  the Surface Science blog

Transcript

QCM-D technology - a microbalance to scrutinize molecule surface interaction
[00:00:00] Science on surfaces.

Malin Edvardsson: Hello there and welcome to this podcast. Science on surfaces, tips, tricks, and tools with me Marlin Edvardsson. So today we will talk about QCM-D technology. We will talk about what it is and a bit about how it works. When and how it's used and some pros and cons. So here with me in this call, I have two of my colleagues Erik Nilebäck and Fredrik Pettersson, both senior application scientists at Biolin Scientific. Welcome. Thank you. So Eric, your background is in surface science by your sensing and project management. Yes. You have a master of science in engineering, biology, and devices and materials in medicine from Linköping University and a PhD in bioscience from Chalmers university of technology and your PhD project [00:01:00] was on developing new biosensing application areas for QCM-D.

And you've been working at Biolin Scientific for 8 years. Yeah. And Fredrik, you have a master of science in biophysical engineering, also from Linköping, and you've been working with application and technology development for analytical instruments. So even before your master's graduation and you worked with the multi-variate guests sensing at Nordic sensor technologies and cell based assays and electrophysiology at Symmetricom.

And now you've been working with QSense QCM-D at Biolin Scientific for nine years. So these are great backgrounds for today's discussion on this. Uh, so let's just dive in and start with the basics. I mean, QCM-D, how would you describe this technology to someone who's [00:02:00] never used it and who's not familiar with it.

Fredrik. You go!

Fredrik Pettersson: Yeah. To start with, it depends a little bit on the background of the person. If you come from a bio background, biochemistry or biology, You would for one thing say that it's label-free, that it measures molecular interactions and molecule surface interactions without any labeling. Um, if you come from other areas, maybe it's more appropriate to say that it is a mass sensor or measuring mass bound to a surface, a specific surface.

And, uh, that you also can measure properties of the material bound or the binding. 

Malin Edvardsson: Yeah. So surface sensitive time resolved method. 

Fredrik Pettersson: Yeah. Time-resolved I forgot. But yes, that's true. Of course. That's one of the great [00:03:00] advantages of QCM-D that you can follow the entire process and with quite detailed time resolution too, 

Malin Edvardsson: so yeah, 

Erik Nilebäck: please go.

Yeah. Just also. Similar tracks that are usually used to describe it, then depending on the level of the person, I mean, technical level of the person you're talking to, it could also be just a device to weigh molecules more or less. I mean, if you really want to be more kind of fundamental in your description.

So, um, both mass and structure are usually added. 

Malin Edvardsson: Exactly. And we actually didn't mention, uh, the abbreviation QCM-D is short for quartz, crystal micro balance with the dissipation monitoring. So it's, it's a balance for small masses. So, so what, uh, mm. Um, how would you say, like what mass [00:04:00] resolution are we talking about in terms of what we can measure.

Erik Nilebäck: I mean it's, it's in the nanogram range. Yeah. So molecular layers, single molecular layers and larger structures on top of that, uh, up to a few micros, I would say. 

Fredrik Pettersson: Yeah. So some monolayers, I mean, that means that we can really measure the single, fairly small molecules also on single layers of those. So it's, uh, pretty much, uh, uh, a smallest you can get without going sub atom size.

Malin Edvardsson: And if you would compare it to similar technologies where you characterize surfaces or molecule surface interactions or, uh, layers at surfaces. How would you compare QCM-D to, for example, [00:05:00] I don't know, AFM ellipsometry, SPR.

Erik Nilebäck: Yeah, 

Fredrik Pettersson: maybe you have more experience about that, Eric. I mean, for one thing, we actually measure a mass, so use a sound wave to probe, and that's quite different from most of those methods, for instance. SPR and the ellipsommetry on the optical methods because they don't measure the mass, they measure change in refraction index.

So this would be a more direct measurement AFM. I haven't used at all 

Erik. 

Erik Nilebäck: Oh yeah. So, I mean, I basically do my PhD. I use those techniques and QCM-D for. So if this chemical systems that we started, uh, like hydrated, polymer, layers, and biomolecules, um, and, uh, I would say, as you actually wet you actually using a scale in QCM-D with that acoustic sensing..

And then you can compliment that with optical techniques, so [00:06:00] you can get complimentary data about the mass, for example. So I helped to get techniques to not show the couple, uh, It's usually water down by molecules, but then, uh, and there are, so the structure, I don't you, I guess you mentioned that fairly short memory, but uh, I mean, you, you, you're not measured as structure rearrangement in the optical techniques.

You should be, you can see as a change in the refractive index, but it's hard to quantify. 

Malin Edvardsson: So both of you have mentioned, um, acoustic technology. So what is an acoustic technology.

Erik Nilebäck: Yeah. So, I mean, actually you're, you're, you're measuring vibrations. You're not measuring. Uh, so you actually measuring the, the, the, uh, uh, waves that propagates into the, the, uh, movement of, of the crystal more [00:07:00] than, uh, changes in, for example, for something light. You can compare the vibrations of the crystal to the vibrations of a guitar string or something similar, the actually measuring those kinds of frequency changes.

Malin Edvardsson: Um, yeah. So let's, let's talk a bit more about, uh, the working principles. I know Fredrik, you've talked a lot about and explained a lot, or you often explain, uh, how this technology works. And so maybe you have a straightforward way to describe to the listeners. 

Fredrik Pettersson: Yes. So that was my reaction. When I started as an application scientist here that I thought we had quite complex and very much mathematics in our, our explanations about the fundamentals of QCM-D.

So I started thinking a lot about it and figure out, yes, it is an acoustic method is a vibration, uh, being transmitted through a layer on the liquid. It's actually, it is a [00:08:00] soundwave. So some ways are what we are measuring and what we are interacting with the QCM-D sensor. Uh, so in that they can say that the quartz crystal itself, it gives a tone.

Our QSense sensor has five megahertz as the fundamental frequency that their basic tones, of course, it's ultrasound them very, very high ultrasound. And in difference to kind of the medical ultrasonic measurements you're used to, we're not measuring time to the reflection, but you're actually measuring how a layer on top of the quartz sensor itself affects the sound, the tone level. If you have thick layer, you'll get lower tone, just like a longer guitar string would give a lower tone. So there are going to make sure frequencies change that's actually directly related to mass. And then also if you measure how we're, uh, [00:09:00] the sensor gets dampened, how those oscillations dampens out and dies out after you stop stimulating it, that gives you the material properties, but all in, all it is, has really to do with how sound is propagated through the film you're measuring and out into the liquid around it.

Malin Edvardsson: So we measured two parameters. One is a sound and how sound the sound changes when something, yeah, the frequency and the tone. Yes, exactly. When, um, molecules absorb or bind to the surface. And then we also measure, um, the decay time of the tone and that's assess something about the softness or the structure of what's on the surface.

Erik Nilebäck: Yeah, what I mean, or if you, in other words, how much energy that is lost during each or oscillation of the crystal. 

Malin Edvardsson: So we get the information on mass changes and [00:10:00] also structure, structure, and structural changes with QCM-D. 

Fredrik Pettersson: Yes. So it's just like, I mean, take a tuning fork.

If you strike it to something and it will ring for very long time when you hold it up in air. But if you can wrap the tuning fork in a blanket or something soft, then the tone will die out quickly so that we can see the dissipation measures. How, how soft, uh, the, a layer around on the sensor or, or, or liquid surrounding it is.

Malin Edvardsson: Yeah, it's very straightforward to imagine that it would be more difficult for the tuning fork or whatever that oscillates it's harder to move when you put something heavy on top. So the heavier, the, whatever you put on top, the harder it is to move, which then 

Fredrik Pettersson: yeah, harder. I mean, that's a, that's a bit of a misconception, a high mass load.

Won't read. Uh, in the sense [00:11:00] of a vibrating, it will just vibrate slower. Uh, so, um, um, so you don't really get affected, uh, in the amount of crystal, uh, Vibrates by the mass itself. So thickly is not the problem. The, what gives dissipation and this dampening material effect is when it's soft, the course, then you will actually have this interaction that the vibration dies off more quickly.

And that's how you kind of measure the softness. 

Malin Edvardsson: Hmm. Great. Um, so I mean, both of you have experienced using the technology. Um, so. I mean, how have you used it, Eric? You started already during your PhD. 

Erik Nilebäck: Yeah. Or even before that? During my master thesis. So, uh, yeah, both my master thesis. And then in, in, uh, after that, uh, Following up that actually [00:12:00] my PhD period.

It was a collaboration between the Biolin Scientific and Chalmers University of Technology, industry, PhD students. I used to QCM-D as my main technique and then complimentary techniques to kind of study our system. So I used it for, uh, yeah, all kinds of biomolecular systems. So going from. Mainly using biotinulated systems that is a coupling chemistry to be able to have very specific binding of biomolecules on a surface.

So we have, uh, a yeah, I mean, in our product portfolio, we have a Biotin sensor that was part of my project to develop that and then to use that on different types of systems. So that was both, uh, biotinulated antibodies, biotinulated to kind of, um, Yeah, carbohydrate layers. And then, uh, also looking at kind of repeated antibody, antigen interactions, uh, also to the carbohydrate layers, we also do the cellular interaction.

So they would [00:13:00] combine it with micro light microscopy to study that as well. So during that time there was a lot of focus on, on biomolecular systems and then kind of the bio side of, of QCM-D. 

Malin Edvardsson: So it questions. Who are you asking? That the QCM-D could help you answer. 

Erik Nilebäck: Yeah. So, uh, there was a few, but mainly it kind of working with specificity, so that could measure specific interactions.

So we looked at both, uh, specific information or changes in proteins, uh, and they could actually measure these confirmational changes and verify that with a dissipation shifts. So they got elongated when they changed confirmation. And then we could see the dissipation shift that was specifically triggered with known drug molecules.

So it was novel way of looking at confirmation changes for that system. And then similar way. And in, uh, in the carbohydrate systems, we could [00:14:00] use enzymes that we know degraded these systems and see the specificity and compare that to in different systems. So, uh, Verifying the biological activity of, of the layers on top of the QCM status.

So like I said, it was a lot of focus on PhD work is kind of developing new, uh, strategies or protocols for using the techniques. 

Malin Edvardsson: And what would you say Fredrik are questions that are asked? What other kinds of questions have you seen be asked 

Fredrik Pettersson: for? Yeah. I mean, to start with my, my first experiences of QCM, of course, crystal micro balance has been with the film thin film deposition.

When you have the vacuum systems and use Proctor or thermally evaporate, uh, metals on top of the Silicon wafers. And for that purpose, you use the QCM calibrated, basically just for the density of the metal. And then you can get to the [00:15:00] position speed without quite easily. Uh, but I mean, compared to. This instrument compared to QSense are very crude.

I mean, they have lower resolution than that. Uh, you don't really mind so much about the linearity and of course you measure, uh, nanometers size there of very rigid layer. Uh, so as an analytical technique, Of course, the sensitivity is higher. And, and, and when I started here at Biolin, I started working with development of the instruments and we were on the verge to launch our automated instrument queues of QSense Pro.

And then we had to work a lot with details and reproducibility and things like how, how does the instrument hold the sensor? How can we increase the reproducibility in sensor clamping sensor holding in the instrument? How can we make a flow cell that fills reproducibly [00:16:00] every time without an air bubble formation.

And also how can we keep temperature control stable enough when this robot coming in pipe fitting liquid, basically not onto, but very close to the sensor. So there is lots of questions there. You really needed to work on to be able to take advantage of, uh, of the precision and, and the kind of. Extreme, uh, resolution in the QCM D method.

Uh, so that does kind of been my way into the company and also kind of my bridge over to application development, where I'm working now to work with those questions issues. How can we prove, uh, how can we improve and how can we prove the reproducibilities I've been working with. Some, uh, demo experiments, fruit up.

Now one important issue there is when you get out doing a demo to a customer, can you [00:17:00] have a clean enough sensor working in someone else's lab, then you need to clean your sensors to make a good demo, but that might be difficult and very stressful in someone else's lab. We don't know exactly equipment they have, et cetera.

So, uh, so that, that has been very close to my heart. Kind of. See where the limits is in reproducibility. Can we see that a certain protein always bind certain amounts. I've been working a lot with albumin as a modal protein, which is very predictable. We know the size of it. We can anticipate what the monolayer albumin, a whole lot of signal that would give, and also.

How can you work with coatings on a sensor to actually reduce variation from, uh, the cleaning of the sensor, et cetera, and also little bit in, uh, how can we really, at the demo highlight this, uh, uh, quality of the QCM-D data that you see, uh, uh, [00:18:00] changing the material properties, uh, together with a mass change.

So, I mean, is a pretty cool experiment to do in the QCM-D sensor, because you really see how the layers softens up and that you get fragments, kind of, uh, um, slightly going out to, into the, into the box solution that kind of waving around increasing the dissipation. And then we start loosening the mass.

So that, that is really interesting phenomenon to study with the QCM because we see so much things happen. I mean, having an albumin in on the surface and you bind in the trips in the see a little bit shifted from the trips in bind scene, but then quite rapidly, it starts cutting the protein change and you get the increase in dissipation.

And then after that comes the most change. So there's so much information in this, uh, uh, just seeing the picture. When you, when you change one protein to another, you see different phases of this. And the time resolution is really so quick, so we can [00:19:00] see the trips in binding, and then you can see it start cutting.

So that's it's but yeah, it's fascinating. It's it's um, um, it gives you so much detail on what happens on the, on the surface. 

Malin Edvardsson: Yeah. Yeah. So we have, uh, you've mentioned a few examples now, but how, how would you guide someone in general to when this technology could be used or could be helpful to use? 

Erik Nilebäck: Um, I would say in general, when you're interested in looking at the molecular scale, uh, and you're, especially if you're talking about interactions, so if you have.

Uh, a material that, you know, interacts with another material, you're known to know how that happens. So yeah, I want to actually look at, especially if you want to look at how and when it happens, so time resolved and , and if you know that it's a kind of a [00:20:00] soft or a could have visco-elastic properties as we call that with another name, uh, commonly, I will say it's a real jackpot because then you can really see things.

So, and especially if you, for example, you know, Be interested in, I don't know, methods of polymers or it, or it will be by by molecules. Uh, I would say that you're very close or, or polymer layers, but it, it could also be, if you were interested in seeing how bulk materials interact with the surface, for example, you can also have these, these effects, even that can be, uh, it's another type of measurement, but there, I mean, there's yeah.

I don't know. Uh, F for me, the possibilities are very, I mean, are endless more or less, uh, w when, when you're actually interested in the details and you need to be interested in what happens at the molecular level. Um, so I would say that then it's a very good technique to use in [00:21:00] different circumstances.

Fredrik Pettersson: Yes. Yeah. I agree with you there. I mean, you take the biomolecular perspective quite a lot around, uh, I would say out of that, or even within that scope, also, if you look at things like protein aggregation, for instance, one of the advantages is that you can follow. We have resolution enough to follow the single molecule interactions with the surface, but.

I mean, we can really work to one or 10,000 times thick layers. You can go all the way to micrometres and that's pretty outstanding that you can follow this length really from a good location on set. You can see the. Um, monolayer of interaction with the protein, then it can follow the dimerization and ultimately formation and eventually the aggregation process that eventually also leads to some structural change in the protein itself when it goes through really thick layers.

So, so that range with that precision, that's [00:22:00] pretty unique here that they can follow the entire, uh, scale of aggregation processes. The fouling processes. Yeah. Uh, so that's an advantage in, in many, many applications. Uh, and of course, in difference, many other surface sensitive techniques it's that we actually studied the exact material surface.

So, uh, the only method I know that really measures the same thing is the very simple method, uh, contact angle. We're actually measure the first interaction layer, but then contact. That method gets messed up when you have adsorption phenomenon going on. So that's more after and before, and you don't get the exact way, but you get the change of the surface property, but QSense also measures that kind of interaction between the liquid, the surface and the salute being solved in the liquid are, uh, so you can really study all that, [00:23:00] that.

Three part system of your molecule, your surface and your solvent, uh, and with a huge range of different materials. And that's also pretty unique that you can do that. 

Erik Nilebäck: I would, uh, I would second that, that you're at you're very free in the type of materials used in distance. So that is something that is really, that kind of drives the applicability of, of, of QCM-D.

Fredrik Pettersson: Yes, definitely. And that of course means we can start a really detailed phenomena. It can study even dif different formulation of a, of a polymer like nylon or something. You can actually look at one formulation with some softener added to the nylon or, another one with longer or shorter chains of the polymer.

And you can see details in how they get swelled by a solvent. Uh, maybe how, how. The polymer gets removed by solvent or how something absorbs to it. So you can really see those [00:24:00] differences in, in the actual material, on the level of, of differentiating different formulations of a polymer. So in that sense, it's a very real, very realistic method.

You work with a real surface with advantages and disadvantages, of course. So when you look up at level I'm in contamination and you see differences there. 

Malin Edvardsson: So, uh, Fredrik, you mentioned, uh, uh, briefly before the, uh, like the standard route, the classical QCM, plain QCM. Um, so if you would compare QCM-Ds to other QCMs, I mean, how would you compare it?

Because there are lots of other QCMsout there. Also, in addition to the standard one, QCM-I, QCM-A

Fredrik Pettersson: yeah, it depends on the measurement method. So all are based on the quartz crystal and, and the, the piezoelectric, uh, uh, [00:25:00] properties of the quartz that itself kind of translates a vibration, an acoustic vibration into an electrical signal.

So that's what all of these instruments are using. And then the main difference is in what electronics you use to record that the vibration or the changes the most simplistic one is it's called a resonance circuit. So that actually uses electrical circuit circuits that puts the crystal and the electrical circuit together in self oscillation.

And that means they will kind of track a change in the resonance frequency only by. Measuring out an output voltage, uh, from that circuit. And they have, for one thing, they are there. That's very simple circuit, quite cheap, and you can get the voltage out of it. So very easy to sample the computer, but eh, they cannot be too fast because then they will kind of get [00:26:00] unstable.

If you try to probe that a resonance. Um, kind of higher. And then since the output voltage is proportional to the mass to it, you get some limitations in, uh, things like the ADC measuring that voltage, et cetera. So they have a very, they have a kind of larger steps in the resolution of the mass bound to the sensor.

Uh, so good enough for, for cheaper applications. And then you have the intermediate ther and that is using. Uh, sir, kind of these, uh, circuit analyzers. So that is a, an electronic piece that can apply a frequency. So voltage sinus wave, but different frequencies. And while it does that, it can measure the amplitude of the current going through the quartz sensor and also the phase shift.

And the phase shift is kind of related to, uh, capacity, [00:27:00] uh, And property, so of the sensor, and, uh, Uh, surface a surface by material on that, that the entire system. Uh, but then you can sweep this and really measure like, um, frequency scanning or impedance VictorOps topical light can measure the peak of the residents and you can see the width of the peak.

And that means you can get information on the dissipation of the system. Uh, those, those be. Basically approach the sensitivity as QCM-D, but they are much slower since you need to scan many frequencies. Um, uh, but, um, but the advantage also there based on the kind of standard piece of electronics, you can buy one of those impedance analysis instrument, uh, circuit centralizers.

So they will become a little bit cheaper. And then you have the QSense where we measure with a ring dumb method where we [00:28:00] stimulate. At the latest known frequency and that we measure the entire DK curve. So that means for every time we stimulate the sense with QSense,since we get one frequency and and one dissipation, value.

So that makes this a much, much faster. And also then of course, since you get, if you don't need to measure fast, it means you get more data points to average to reduce noise. 

Malin Edvardsson: So the first category you mentioned, measure. One parameter, a single frequency and the second category measures, frequency and bandwidth, or impedance or dissipation, some sort of energy loss.

Yes. And then QCM-D also measures two parameters or the QSense QCM-D also measures, two parameters. And you also have the, uh, the multiple harmonics. 

Fredrik Pettersson: Eric, do you want to elaborate a little cannot measure harmonics at all. Then with impedance [00:29:00] spectroscopy, you can do it with the superintendent license.

You can do it, but the higher range, third, the more expensive that part will be, of course a QSense can do that by default if different harmonics, but the advantage of that laboratory. Right? 

Erik Nilebäck: Right. W w w with having several. Yes. I mean, I mean, it's basically for, for the identical system, you get a larger set of data.

Uh, so you can enable, uh, modeling of the applying different kinds of, uh, physical models where based on the theory. So for, for, for viscoelastic properties of the day or so, uh, both, uh, storage and loss, modulus of your layers, and then also thickness mass. Um, not being a mistake, the two for four, um, for rigid layers, you have a, more of a linear relation through the, when [00:30:00] Sauerbrey equation, basically constant times the frequency you get the mass estimated mass value, but you can use it basically as the models, you can use a soft layer as it can get also additional information.

And by having the largest set of data from the overtones. Or harmonics, uh, you, you will, it will be able to model cause they need input data to be able to have a high quality modeling. 

Fredrik Pettersson: Yes. So, I mean, you needed to use the property of sound on how the different frequencies when you go to from five to 15, 25, 35 megahertz and how those soundwaves actually behave a little bit differently.

Yes. Depending on the layer properties. 

Erik Nilebäck: Yeah, yeah, yeah, yeah. You can actually find information within the overtones or the harmonics as well. Uh, so they have a little bit different. So the higher overtones, propagated it a bit shorter into the, uh, sample forks. And then [00:31:00] for thick layers. For example, when I started salient cells attaching to the sensors, we typically looked at the fundamental or the, or the, uh, third harmonic, uh, because they propagate longer into the, into the system.

Yeah. So you can also have those advantages. You can actually discriminate, uh, things happening on the sensor between the harmonics, but I mean, I would say the major advantages that enabled modeling and using that, but then also you can get this added info. Yes. Because they behave differently. 

Fredrik Pettersson: But also, I mean, from the user perspective, one really important aspect is that if you look at the fundamental, oh, We know for sure that the fundamental frequency is definitely a lot more sensitive to, to all kinds of disturbances and that's, I mean, QCM is a very sensitive technique.

So of course it's sensitive to all kinds of stuff that you don't want to measure and that you need to keep down. So the fundamental frequency. [00:32:00] It's more sensitive to mechanical stress. Uh, it senses all over the sensor of surface, which makes it more sensitive to, uh, variations in the clamping, how the sensor is held.

Um, it is a kind of sense. It penetrates deeper into liquid. It's more sensitive to air bubbles flowing through the flow cell and, um, uh, Yeah, temperature sensitivity is kind of the same for all of the harmonics, but, uh, it's, it's much easier to keep things in check. If you use the. higher, harmonics and not the fundamental because the fundamental one is very sensitive to disturbances.

And my recommendation for any user is to always record the fundamental, because that's a good quality indicator. And then if you can actually use that signal for your measurement, then you're lucky.

keep control over your experiment. You can do it, but, but usually just keep it as a, as a quality profile and see, okay. Uh, got some strange noise in the editor now. [00:33:00] And then you can see F1 had a big jump into some disturbance from when the air bubbles or something. Okay. Okay. This was not the measurement signal.

This was just some disturbance. 

Erik Nilebäck: Yeah, exactly. You should always record the fundamental . It's good to use. 

Malin Edvardsson: Okay. Now we've touched upon a pros and cons and you're describing this high sensitivity, both as a pro and a con actually, right? Because I mean, high sensitive is great, but it also makes it a bit more tricky maybe to handle or to run measurements, reproducibly.

Erik Nilebäck: Yes, it might then one thing that it's important, but, but that, but that is for all surface sensitive techniques that have, uh, have, uh, that are that sensitive in that signal you need. Uh, a reproducible way of preparing your sensors. You need to know that you have clean sensors and that they have clean solutions, uh, clean tweezers.

I mean, EV everything, the process needs to be controlled. So, [00:34:00] uh, but that goes for all techniques, I would say that was similar and it applies also to QCM-D. So it is a sensitive technique and can really measure. Details of the layer interactions and how, what happens at the molecular level. But that also means that you need to have the clean illness at the molecular level in check.

So 

Malin Edvardsson: it's a measured dirt. 

Erik Nilebäck: Yes. Yeah, exactly. And of course, this might vary from system to system. How thorough you need to be. I mean, if you, if you measure, uh, interaction or thick layer, so you have a lot of, for example, real big aggregate being formed, are you measuring complex fluids then? Of course it might be a little bit different.

But, uh, I mean, yeah, I'm a little bit biased coming from the surface chemical side where we are very picky about these things. So preparing solutions, looking at single molecular interactions, but I think in, in a, in, in ways, it's, it's something that can be an advantage when you have everything in place and you know what to do, and you know what [00:35:00] you're looking at and you can be kind of confident in that you have actually looking at what you hypothesized to.

Uh, but yeah, it might be a drawback. They can be time-consuming sometimes with sample preparation. 

Malin Edvardsson: So could you just, you know, briefly describe an experiment. I mean, to someone who's never used instrument, how would, how would you actually run an experiment or a measurement. 

Fredrik Pettersson: Yes. I mean to start with, of course you need to go down into lab with some kind of plan for what you want to do, but the first thing I do, that's always prepare my sensor.

So clean it to the level of cleanliness I need. And of course that depends a little bit on how thick layers I want to measure. Uh, and also, I mean, how life like surface do I need? Am I a surface scientist? And I may be needed a molecularly clean gold surface or, or silica surface, just to be sure that I can bind my next layers in a perfect way.[00:36:00] 

But if you're interested in a more realistic situation, then you might actually want to have a slightly. Contaminated, uh, stainless steel surface that corresponds more to a life-like situation. Yes. But anyway, and they'll either of those approaches, of course, you must make sure that your pre-cleaning program.

Uh, insurance, the reproducibility so that you have the same surface every time you repeat the instrument, uh, and also that you get the low drift. So for instance, if you have some, some contamination of a surface, it would be a bit tricky to get to a good stable baseline. If you, if you start cleaning it out by the experiment buffer you're using.

Uh, so that's the first thing, clean the sensor in a reproducible way, and then really mount the sensor in the instrument and set the instrument to your correct temperature, uh, which is very important. Yes, it's super important. And that's the most common reason for drift [00:37:00] in QSense measurements is, is that you haven't stabilized the temperature for long enough.

I can also recommend the sometimes, sometimes it can be quite good to do some of the cleaning inside the system also. Uh, if you want to have really, really good stability, it can be quite good to actually do the cleaning in the flow cell. In the instrument, the cause flowing the sample in a low height flow cell can actually speed up chemical processes like cleaning.

So next step after that, then I would go out to the lab and prepare my liquid samples for the experiment. If it is a protein or a surfactant or a, or whatever samples there are, because yeah, why not take advantage of this time? I want my sensor to get stable at the measurement temperature. Um, when preparing the sample, I will be very, um, I will be very careful with [00:38:00] contamination risks.

And, uh, I mean, if, if a contaminant has higher surface affinity than the molecule, you think you're measuring, then you actually might ruin your experiment completely or draw your own conclusions. So it's quite good to be aware of the quality and the purity of your molecules. Again, here, that might be something you want to reproduce.

If you want to look into mining process and they use industrial quality surfactants, then of course you should use the, the, uh, industry quality factors, maybe 75 or 80% purity. Uh, but maybe then also compare them to real pure analytical 99.5% pure molecule. And. Of course, let things take time. Uh, surface active, uh, molecules usually have tendencies to also aggregate.[00:39:00] 

So be careful with, uh, solving your molecules, use ultrasonication, for instance, to get rid of both air in the liquid and also make sure you, you break apart any aggregates. Um, What's next off to that directly. 

Erik Nilebäck: I forget anything. Yeah. So you have your, you have your molecules dissolved in the buffer that you're going to intend to use, and then you, yeah.

So you have both kind of background buffer and you have your samples that you did in the concentrations that you'd want to use. So, uh, and then after that, so you have stabilized your sensor in this instrument and the temperature, and then you add a background buffer and stabilizing that, and then you add your analytes.

And be sure I, uh, I prefer to use, uh, for all my experiences in, in with QSense, uh, instruments. I prefer to use the, in the notes within the software, [00:40:00] always, uh, to, I mean, if you use the pro you're gonna get it automatically from the samples, you have noted that they're going to use some of the times, et cetera, but for the, and the Analyzer and Explorer systems, I, I encourage people to use the notes because it's really good way of keeping track of what you do.

Yeah. And adding times, et cetera, because it's it, it's not that easy always to remember exactly what you did. So that's a good idea to, to do. Um, and then, I mean, uh, if you talk in this simplest biomolecular adsorption, that will be buffer, adding your analyte and then rinsing with the buffer again, once it.

I know this is somethin that Fredrik likes to talk about your, that you have stabilized your signal. So you have a kind of saturation in the signal, so you don't see the drift. Hopefully it will be flat. So you have really come to come to, uh, uh, equilibration state states. So 

Malin Edvardsson: this can take five minutes [00:41:00] or three hours or even longer.

Yeah. 

Fredrik Pettersson: Yeah. I mean, that's super, I've seen so many measurements even from, from publication in journals. We have seen that they claim we are approaching equilibrium and that they stopped the measurement and maybe even just stop it and don't measure the desorption phase and how stable the surface was in buffer.

So that's, I mean, always start and end in the same liquid situation. That's crucial. Uh, yes. And also baseline. Make sure you take some minutes, five, 10 minutes of baseline, to be sure that you really have low drift compared to the size of your signal. So if you're measuring for one hour and you're measuring.

Uh, 50 or 100 Hertz of signal that then you wouldn't need to care so much if you have one Hertz per hour drift. But if you're measuring five Hertz [00:42:00] signal, like a really thin, I mean, I acid layer that then, then of course one Hertz per hour is kind of borderline and you need to see if you can improve it to lead a little further, which you can.

Uh, so that's quite important. Uh, Hmm, setting your, uh, to the right flow. Speed is important also. I mean the, the standard flow module for the analyzer and Explorer is about 40 microliters. So that's means having 100 micro liter per minute, would exchange have complete liquid exchange in the flow cell in about one minute and the same for the, uh, for the Pro system where it's only 17 microliter, then.

Uh, 30 or 40 microliters per minute would give complete liquidity change in one minute. Uh, so compare that to your experiment length, because of course, if you run higher flow speeds, you will have foster liquid in change and more sharp onsets of your [00:43:00] kinetics. But on the other hand, if you were to run really long experiments, you would consume a lot of liquid.

Um, so that's it. So that's one thing that's important to keep. And then of course, there's this run to equilibrium and be sure it is equilibrium. If you're studying protein absorption, for instance, you will get a quick, you're often a very quick signal from the first layer. Yes. A monolayer of proteins absorbing, but then if they actually have some confirmation change in the protein, that will start promoting aggregation, that can take time.

And means that you will have a much, much slower kinetics in the following stages of aggregation on the following multilayers. So that means you cannot really only approach a plateau. You need to be sure. Is this really at all? Do you have some aggregations? And this is that. So even if the protein proteins option, the first monolayer, maybe it [00:44:00] takes 10 minutes, then I will run for two hours just to be, if I'm interested in the multi-layer formation, very different 

Erik Nilebäck: kinetics.

Exactly. I mean, you can see a whole new faces in this being formed, so it can be very interesting data. And 

Malin Edvardsson: why is it, why is it important then to end with the buffer, the same buffer as you started with, or the solution that. 

Erik Nilebäck: Yeah. So, so you, you, you need to be able to compare, uh, so you need the high, uh, basically you need to have the same, uh, viscosity and density in the, uh, in the, uh, background buffer.

When V if you want to model your date, that would be the short answer. Or do you. What would you say? 

Fredrik Pettersson: That's one thing, because I mean, QSense being very sensitivity sensitive? We also measure of course, differences in the liquid itself with different density viscosity. We'll actually see a little bit change on signal.

Usually not that much problem. If you are [00:45:00] below a say one milligram per milliliter, that's true. That's true. If you're at that scale, you will usually not have much viscosity shift. Uh, But if you go higher than that, those concentrations you would actually need to measure or verify your viscosity of your sample liquid, because that one might be different from the buffer itself.

And that's of course, good to know when you model your data and, uh, and the only way to be sure that you actually have the exact same situation of towards going back to the same buffer, but then also off kinetics is quite interesting to see how does my sample release from the surface. Will that tell me anything about the interactions.

Yeah, of course it does. So it's always good to do that. Um, 

Malin Edvardsson: okay. So we talked about, uh, you know, a lot of situations where QCM-D technology is very useful. So what are the limitations, or are there any [00:46:00] situations where you wouldn't recommend using it? 

Erik Nilebäck: I mean, w we, we touched up on it now, but that is if you're going to use it very viscous liquids.

They're highly viscous. Cause that will dampen the signal. So you can't actually measure it. But then we're talking about very, very high viscosities. Uh, I don't know these numbers per heartbeat, perhaps, you know that Frederick yeah. 

Fredrik Pettersson: You start losing the highest harmonics really around tens, 10 to 15 centipoise yeah.

Then I think you can mesh around 50, 60, at max. So. Uh, that's about where it goes. Uh, of course you can increase the temperature to reduce the viscosity and a measure of not viscous, uh, solvents in low temperature. Uh, but yeah, that's a real limitation. Uh, I don't think the situation is a much better with other surface methods.

Like[00:47:00] 

Erik Nilebäck: so there might be no, no other techniques that can be used. 

Fredrik Pettersson: So maybe it's an advantage. Yeah. One thing I could mention is, I mean, if, if you caught, if your surface isn't stable in the buffer you want to measure, that might be a limitation. I know people, I know customers using that situation. So they see something washing water.

Let's see the takeoff of a few, nanometers per per hour. Or some and per minute, and then they see if I add this surfactant, or if I change the pH in this way, then you will have a difference. Of course you can, you can do a slope analysis and the, uh, look at, uh, desorption rate or cleaning rates or adsorption rates in different situations, even if you can't get a stable baseline.

Uh, but it would be very difficult to kind of measure absolute thicknesses there. Uh, but you could still measure [00:48:00] differences in. desorption and adsorption speeds. 

Malin Edvardsson: So-so talking about surfaces then. What real life surfaces can be used. Are there any limitations? Of course there are, but what would you, how would you describe those?

Erik Nilebäck: But, I mean, if we talk about the, I mean, this is a real advantage with QSense, as we told before, but since you're not fixed, if a competitor. Uh, not the technique like SPR for example, then you need to have a surface pump that monitors activity. So first, or if it's, if someone do you need to have a reflective surface, so you're not limited to those kinds of predictions.

Of course, you need to be able to coat it on top of the sensor and it needs to be, uh, stable. Uh, as I say, predict in this whole bit that they're going to analyze it, but by you mean you can use metals polymers or. Uh, in different ways, [00:49:00] as long as, as long as you can coat them on a sensor surface. 

Fredrik Pettersson: Yes.

Yeah. Yes. Really? Anything of course must bind tightly to the surface, but yeah, Palmer, ceramics and nitrides, and as we mentioned earlier, you can actually look at real formulations. We can see difference between soda-lime glass orborosilicate glass, uh, and. And, uh, yeah, different steel formulations, different steel types, different types of stainless steel, for instance.

Uh, so that's possible to do, um, one restriction is homogeneity you, it, it's a big advantage having a homogeneous layer, you can measure porous layers, but that's, it's a lot more difficult to analyze the data in that. Uh, and of course, small differences in pore sizes might change your measurement situation a lot in that case.

Uh, so that might, might [00:50:00] make things a little bit messy. Um, and of course, the top layer on top of the quartzsensor. So the quartz sensor is actually built up by actually single Crystalyn slab of quartz. Uh, the thickness of it is, is what defines the. Fundamental frequency, the resonance frequency. So our sensor are 0.3 millimeters thick about.

So that's your quartz to be able to stimulate that electrically and read the signal. We actually have a gold layer on top of it, but gold doesn't stick very well to quartz. So in between there is it's an additional layer of chromium or at some occasions titanium. So quartz, the it's chromium and gold and then on top of that, you need to put your sensor material and, uh, but you can really use any technique there.

You can use sputtering or, uh, or, uh, spin coating or, uh, or some kind of inline flow [00:51:00] methods, something so very flexible there. But what we can't do is put thick things on there. You can't put like a piece of fabric or, or, or. 

Erik Nilebäck: Or like a real roll paint, for example, it's not really possible when you get almost millimeter thick, very thick.

That's not possible. And 

Fredrik Pettersson: I guess not even those kind of 20 micrometer thick polymer films, I don't think 

Erik Nilebäck: either no, like taping him. So that kindness also, uh, that then to getting to too thick. Uh, yeah, so you need to have a thinner model system. 

Malin Edvardsson: When you say thin, I mean, what thicknesses are you talking about? 

Erik Nilebäck: I mean, it depends a little bit on how well they stick to the surface, but, uh, I mean, if it's a, uh, soft layer, you can't have as thick as you would, if it's a rigid later.[00:52:00] 

But 

Fredrik Pettersson: then when I got meters, but on the other hand, on the other hand, if you look at the interaction with such a surface, if you're interested in material that is actually 20 micron thick or something, is that really the same thing as surface interaction? How would you characterize, if you have 20 micrometres thick layer, then maybe you have a inhomogeneous in the micro meter scale.

So how relevant is it? Study normal sites, some that kind of surface and that's very relevant. I mean, you need to have the right model system and surface science on the nanometers angstrom size. That is, that is, it goes crosswise, our intuition sometimes. Yeah. Uh, we have, we have sometimes questions about, uh, Cotton fibers for detergent measurements and how can we get cotton to surface?

And then you're really thinking the wrong way because the cotton fiber is 10 to 20 micrometers. [00:53:00] Uh, so just having that on the surface, we'll be over the thickness limit of the QCM-D technique, but then the fiber itself, you don't really interact with the fiber. Yes. You adsorb into it into it. But the fiber is built up of thin, thin layers of kind of well-structured, uh, cellulose fibrils.

And those are kind of the scales. If you take cellulose as in the fibril and then put down them down on the sensor, then you can get the flat layer. Uh, and then you might maybe add some fatty acids or some proteins, other proteins as in a real, uh, cotton fiber or, or realistic textile fiber or wood. So it is those things that give the difference between the different types of, uh, textile fibers or, or wood fibers.

But that is really the scale where we look with QCM-D. If you interact with. monolayers or, or multilayers, uh, cellulose [00:54:00] microfibrils. And so that would be kind of taking a cotton fiber, take the top layer of that one, flatten it out and put it out on the sensor. So. And that's also the relevance gain for the, for the molecule interactions with the cellulose.

So it makes sense, not only from kind of what the restrictions, but also from the kind of interaction perspective that is the model scaling. It need to 

Erik Nilebäck: be off. Yeah. Because to me what you would, if it would be possible now it's not, might not be technically possible, but, but if you would, would coat with cotton fibers, what you would actually look at is the interface between the cultural fibers and the sensor that the meat that is.

Area, mainly for the, for the QCM. So you would look at the surface interactions mainly in any case, but if you have a model system that is tailored to work well with QCM-D, then you will get a much better and more reproducible scene. 

Fredrik Pettersson: Or, or even if you, if you could make the cotton fibers touch [00:55:00] really tied to a sensor surface, sort of with would vibrate with it, then you would measure the liquid between the fibers.

So that's yeah, 

Erik Nilebäck: yeah, exactly. 

Malin Edvardsson: So, is there any other key information related to this technology that we didn't cover so far that you would like to add? 

Erik Nilebäck: I mean, uh, I would say what I, I have something like a reflection I may, when I started using QCM-D a long time ago, that it was even though it might sound, you said other things to think about with cleaning, et cetera, but it's kind of straight-forward.

Uh, how to use it, then you start to get the data, say, okay, frequency, decreased mass. Something is bound to the sensor and it's, you really not need more info than that. To be able to start analyzing, start to doing measurements. So it's, it's, it's quick to get up and running with the technique then to get into the finer details.

Like all the techniques you need to time to, to get into that. But then to start [00:56:00] doing measurements. Straight forward. And that is a real advantage. You can try new things. Can use it as a workhorse technique through try different systems. So that I think is. Yeah, I 

Fredrik Pettersson: agree. I mean, it is a good technique since you can do the application development on the instrument itself, because it can start playing around with the measurement situation.

See kind of a, how thick do my layers get just at a high concentration, just to see it, uh, compare that to actually take the dimension of the molecule, uh, and the density of the molecule there, you have all the information to actually calculate a frequency response. So, but just that the molecule, uh, the length of a amino acid for instance, can, can be around three, four nanometers, and then it would say, okay, that would be.

Yeah, some number of Hertz, like 10 Hertz, and then you can actually compare the measurement and see, did I get anywhere close to mono layer of that? Or am I getting a much larger signal that would signify multilayers [00:57:00] so that's one thing I always do. I tried to predict what the amount of absorption should I expect with this molecule.

If I get the mono layer, that's a good comparison number to have there. Uh, and then you can start playing around. Take take time. So, so you're sure what, what happens with the kinetics there and, um, 

Erik Nilebäck: yeah, I would say that that's a good comment to have good knowledge of your system, as you said, you know, what kind of seen as good I expect.

That's a really good thing to keep in mind as well. 

Malin Edvardsson: Yeah. Great. So I think we will end this episode with this, so thank you. Both of you for sharing all this information. And, uh, thanks to everyone, listening. And I also would like to take the opportunity to mention to those of you who are listening or watching that if you're interested in surface science or related topics, you should also check out our blog, the surface science blog.[00:58:00] 

Thank you.