Case studies, webinars, and articles you’ll love
Receive the latest content on a regular basis.
How do you choose the right one? Once you’ve chosen, how do you ensure consistently precise and reliable results? Measuring moisture content can be a minefield, but getting it right pays dividends.
In this webinar, Dr. Zachary Cartwright and moisture content researcher Conner Jeffries:
Dr. Zachary Cartwright: The way that this is going to work today is, we're going to be covering some research that Conner has recently worked on and collected some data. So we're going to work this through an interview format, where I'll be asking Conner a series of questions before we talk about a new technology that METER has released.
Here’s what we’ll talk about:
ZC: So to start off, why moisture content measurements can be so fickle. The first thing I want to do is simply start with a definition of what moisture content is. So Conner, how would you explain moisture content to somebody?
Conner Jeffries: Yeah. So I think it's important to start from the beginning here. So what is moisture content? It's really, we're just talking about the mass of water compared to the mass of everything else in your substance.
So, I think it can be a little bit confusing, because you have to get the water out to measure it, so there's no real way to measure it in situ. So you have to remove the water, and that's really the tricky part.
So typically, what we do is we'll heat up the sample, and we'll remove the water that way. So that's called loss on drying. The other option is to dissolve your sample along with the water into a solvent, but that's pretty specific to certain methods.
ZC: Who is measuring moisture content, or maybe who should be? Why is this important, and who is looking at this?
CJ: Moisture content really is about percentage, it's about yield, it's about mass. And so I think most people want to know how much moisture is in their sample. But really, moisture content, specifically, should be about knowing yield. I don't think it's a great measurement for quality control or quick checks.
ZC: And then, as you measure moisture content, what are some of the difficulties that you may come across?
CJ: Because it's so dependent on the process, we have to remove the water, and we have to typically put energy into the sample. So that can run into problems. So the assumption is that you're putting in heat or energy, and you're removing the water. If you remove other things, other volatile compounds, that will overestimate the amount of water. Also, if you're putting in too much heat, you start decomposing your sample, you'll lose stuff in the form of gas, and you'll overestimate the amount of moisture from that, as well.
ZC: So how do you know how long to heat something, or what temperature to set it at?
CJ: You really have to do a bit of research to get to the bottom of the best methods for your particular samples. But typically, low and slow is unfortunately the best way to go. It's the safest, which is the most accurate.
ZC: Anything else about samples? You mentioned volatiles earlier, but what might make a sample easy to measure moisture content, and what are some of the more difficult samples that you've come across?
CJ: Things that are difficult, well, I'm going to say cannabis right off the bat, because I've been doing a lot of work in cannabis recently.
But also, dried fruits are particularly tricky – dense things, things that have lots of sugar in them, things that are hard to sample. Sampling plays a big role in being able to get the moisture out.
So, those are the things that I would consider difficult samples. Easy samples are things that are typically high in moisture content that can easily evaporate.
ZC: And then the last point you had in here is, do you really need to measure moisture content? I think this goes back to your original point, moisture content is more for yield. But is there another measurement, or how else would you look at the Water?
CJ: For yield, use moisture content. There's really no other way to know the mass of water in your sample. But if you want to know how the water is held in your sample, water activity is a much better measurement. Because water activity basically combines the moisture content with the structure of your sample. It doesn't tell you the percentage of mass of water in your sample, but it tells you how that water is available in your sample.
ZC: So if you're concerned with yield, maybe use moisture content, but if you’re interested in safety and quality, that's when you want to start bringing in some water activity measurements?
CJ: Water activity is much better for that.
ZC: All right. From here, let's look at some of the methods that are used to measure moisture content. There are lots of different methods, but we're going to focus on these top three. So the first one we have here are ovens, or vacuum ovens. So what is this method? How does this work?
CJ: This is the tried and true, original method. If we go back and look at AOAC, we can see lots of oven methods, from the '20s and '30s. And so those haven't changed in 100 years. You put a sample into a really stable heat source, it heats it up. You know how much heat you've put into it, and you basically just check the mass to see if it's stopped changing. So they're very reliable. Most labs have an oven, and yeah, it's stable.
ZC: The next one we have here, I know it's one that you like, Karl Fischer titration. Well, why do you like this method, and maybe why do others dislike it?
CJ: Right. Yeah. I like Karl Fischer because I'm a chemist. So my background is chemistry, and it's fun sometimes.
That being said, I see lots of people using Karl Fischer titrators in the wrong way. I think they're drawn to the fact that it can be really precise and accurate, but there's a whole pipeline of chemistry that's involved in using it.
They were developed for measuring small amounts of water in fuels and other petroleum products. So it's really good. Basically, it doesn't matter what else is in your sample, because it's chemically selective for water.
ZC: Yeah. This is when you really want to pinpoint parts per million.
CJ: Yeah. It's really not great for lots of water. High moisture content samples aren’t ideal here. You have to really make your samples really small if they have lots of water in it. So it's best for small amounts of water.
ZC: And then the third one we have in here are moisture balances. I see these a lot, especially within the food industry. What's a moisture balance, and how's it compared to these other two?
CJ: So with an oven, you have to make the mass measurements yourself, you have to have an analytical balance. So the idea with a moisture balance is that you're combining a heat source with a balance. That's the idea. In practice, my experience with them has been less than ideal, but they can be very fast.
ZC: We'll look at some of the data later in this presentation to see why they're not really ideal. Out of these three methods here, which of these are referenced methods that you see being used in the cannabis industry or food or pharma industries?
CJ: Oven's always going to be a reference method. I think lots of pharmaceutical labs, perhaps, use Karl Fischer. I think I've seen it in some cannabis labs. I know I've used Karl Fischer for testing cannabis. But I don't think balances are known as reference methods.
ZC: Sure. Well, from here, let's look at these just a little bit more, and look at some of the pros and cons of each. So let's start with the oven here at the upper left. What are some of the pros about working with the oven? What do you like about this method?
CJ: Right. Again, they’re very stable, you can put lots of samples in it, so you can do large batches. The main problem is that they're not super fast. You have to heat them up. And then, of course, lots of people have ovens, so maybe you can just get a quick moisture content that way. And then I guess it's con if you don't have an analytical balance, buying both an oven and an analytical balance can be expensive.
ZC: And then Karl Fischer, what are some of the pros here, as well as the cons?
CJ: Again, very precise, it's compatible with things that are volatile. But at the same time, it requires that the samples be soluble in organic solvents. There's some ways around that, but it's always going to require this chemical pipeline. So if you don't have good chemistry handling capabilities, it's going to be a problem.
ZC: So maybe not the best device for a startup or a new company.
CJ: Yeah. Also, they tend to be quite expensive. If you don't have a chemist to run it, it can be a little bit daunting, I think.
ZC: And then last one here, moisture balances, some of the pros and cons for these devices.
CJ: Right. So moisture balances are fast, they combine the drying and the weighing. But in my experience, there's a whole range of quality in moisture balances. You can pay several thousand dollars for a moisture balance, which sounds like a good deal, because balances can be expensive, ovens can be expensive. But I've found that the performance of them varies widely.
The other thing is that you don't actually know the temperature of your sample, so it's just doing its best to estimate your sample temperature. Because of that, it can cause decomposition, at least in my experience, by putting too much heat into your samples.
ZC: We'll look at an example of that, as well.
ZC: There's a couple other methods that we just wanted to mention. These three methods so far, these are not the only methods, but there are a couple others that Conner has looked at. So we're just going to discuss them briefly. The first one is this microwave method. What is this, and how does this work?
CJ: Yeah. It works very similar to a moisture balance. The idea is that instead of using a halogen bulb or an IR heat source, it's using a microwave source. Works great for wet samples, samples that have lots of water in them. My experience with them is pretty similar to moisture balances, where they tend to decompose things and cause issues.
ZC: And then the next one we have here is distillation.
CJ: Right. This is a little esoteric. As a chemist, I'm like, yeah, just do a distillation. But this is a reference method. So the AOAC has a lot to say if you need to distill your sample or do a distillation with your sample. It can be a lot of work. But for something that has lots of volatiles in it, or other conflicting compounds, I might go back to distillation.
ZC: And then the last one we have here is just using desiccant chambers, kind of the same thing. Right?
CJ: Right. Yeah. So the idea here is that you have just a drying chamber, you're not introducing any heat, and there's either a desiccant or something similar, trapping the water in them. That takes a long time, right?
CJ: Maybe a week or two. Maybe it's a good reference, but it's not going to be something you do regularly.
ZC: Sure. All right. From here, let's look at some of the research that you've completed. So we're going to look at a variety of different products here.
We're looking at dried mango, and I just want to point out that a lot of these graphs are looking at the difference from the reference oven method.
Maybe you can describe why you set these graphs up that way, as well as what we're looking at here with different moisture balances. Are these the same balance? Are these replicates? Are these different instruments?
CJ: For most of this data that we're going to look at, it's a moisture content difference from a reference oven method that I found for these particular samples. These are three different moisture balances that I tested. And then in this slide, there's also a moisture analyzer.
You can see that the moisture balances are struggling with a dense fruit sample, like dried mango here. They're not able to fully remove all the water in that sample.
Conversely, you see something like the microwave analyzer, that it removes all the water, but then because it's dry, and it's got a lot of sugars in it, it starts burning it, and so you overestimate.
In this case, you're overestimating by 4% moisture content, but in reality, that's over 40% off from your sample. So it's a large margin.
ZC: Do you see this often, where those samples are getting burnt? Why does that happen in some of these instruments?
CJ: It's just too much energy being put into them. Again, a microwave and halogen bulb, they're not at a set temperature. So you're just pumping a bunch of energy into your sample, which works great for wet samples and things that you want to dry quickly. But as soon as that initial moisture is out, it leaves your samples really susceptible to burning.
ZC: I know we have another dried fruit sample here, so dried blueberries. The results here look really similar, the moisture balances are having the same issues. Right?
CJ: Right. The moisture balances didn’t burn these particular samples, but I didn't get a microwave moisture content for these because they burned, so I stopped. So the moisture balances, again, underestimated, they're not able to get all that moisture out from these dense samples.
ZC: And then let's look at some other types of products. So here we have a protein and greens powder. What's happening with this type of product?
CJ: This is particularly dry, we're talking less than 5% moisture content sample. They're overestimated because they start decomposing. So like we saw with the dried mango, these dry powdery samples are easily burned by all this infrared radiation power getting pumped into them.
ZC: The same is true for this whey protein powder, as well. So it looks like, again, the same thing is happening?
CJ: Yeah. It's just overestimating because it's decomposing the sample.
ZC: And then I believe we have one more in here for cannabis. This one is set up a little bit differently. So this is just the total moisture content using several different methods. Can you walk through this graph?
CJ: Yeah. So there's no consensus on the best method for moisture content in cannabis. I'm doing a lot of research right now, trying to get to the bottom of it.
But we can see that some reference methods, Karl Fischer, desiccant chamber, those match pretty well. But once you start getting into other methods, where you're putting a lot of heat in.
We know that cannabis has lots of volatile compounds in it. So we would expect that putting lots of heat into a sample, perhaps in a vacuum oven, would overestimate the amount of moisture. We see that in a vacuum oven method or in moisture balance at a higher temperature or using an oven. This American Herbal Pharmacopeia method is an oven at 105 degrees.
ZC: I actually just released a short article explaining the need to standardize the moisture content method being used in the cannabis industry. Because if you're producing cannabis, you might send your product to a certain lab because you're getting numbers that you like, because they're using a specific type of method, where it might make your product look different than it really is or different than if you send it to another lab.
CJ: Right. Yeah. I think all THC content is pinned to dry weight measurements, so these numbers affect your potency calculations. I think the California Cannabis Control Board has realized that this is a problem, and they need to get to the bottom of it.
ZC: From here, let's talk about a few methods that you could be doing, or if you're listening, some things that your team could be doing to reduce variation and improve your accuracy. So what could teams be doing differently, Conner?
CJ: Right. I guess, the first thing I would say is, do you really need moisture content? Do you really need these yield numbers? So if you're just doing quick checks, water activity might be a better metric for you.
ZC: The second thing you have in here is, method validation and proficiency testing. What do you mean by this?
CJ: Right. So, are you testing, with standards, your moisture balances? Are you testing an oven? Are you checking that your methods are actually valid? So I think that's something that I spend a lot of time doing.
CJ: We verify all of our instruments constantly.
ZC: But I'm always surprised that when I speak to people in the industry, how many people actually aren't taking this step, or they're not doing it very often.
ZC: The next point we have in here is many replicates. Of course, for anything, the more replicates you have the better. How many is enough? Or is more always better?
CJ: It depends. I guess I would say more is better, but to a degree. You can do statistical analysis, if you want. But this ties into method validation and proficiency testing, is that you need to basically reliably say that this is the moisture content of your sample, and you do that by taking many sub samples.
ZC: The next one we have in here is, just standardize your method and look to the regulators. So whether it's AOAC, or these other ones that we have listed, making sure you look, and that you're using the right method for your product.
ZC: And then the last one we have here is using a moisture model from an isotherm. I decided to put this one in here.
This is something that we do a lot at METER, basically taking your moisture content based on a water activity measurement, and the relationship between these two measurements, which is called a moisture sorption isotherm. So we can make a model this way, and then get a really high precision moisture content measurement.
CJ: We do this a lot. We know the moisture content relationship with water activity. You can get moisture content from water activity, but you can't get it the other way around.
ZC: So from here, let's talk about a brand new instrument that we've released here at METER, something that's going to help your team measure moisture content very quickly and precisely.
So this instrument is called the ROS 1. I took this from our marketing. They say, “Not hotter, just smarter.”
I like this, because like we've discussed for some of these other instruments, they've tried to get quicker measurements by heating up samples very quickly. Those moisture balances are using halogen bulbs, and this just leads to problems. Like Conner showed, this can lead to samples burning, or underestimating or overestimating the moisture content in those samples.
I just want to take a moment here to talk about some of the features of this new instrument, and then we'll look back at the data that Conner has collected, and compare this instrument to what we've already looked at.
So the first thing is that this instrument conforms with AOAC, ASTM and ISO. So if you need to put in a specific time and temperature, you can do so with the ROS 1.
Here, there's no method development. So this is a sample-agnostic instrument. You can put any sample in here, and it will be able to measure the moisture content. The next thing I have is that it automates busy work. So you don't have to write anything down, it's going to automatically graph the time, the temperature, and the weight change for your product. It's rapid testing and high throughput.
As you can tell from the image, you can put in nine samples at a time, and within 40 minutes, you'll have the results for these nine samples. So about four minutes per sample. It's really simple to use. This has to do with the desktop app that comes with the ROS 1, it's called Bridge. It makes it really easy to start a test, to view the data that you've collected, and then export out that data as needed.
The results are going to be really precise. This has to do with the temperature control in this unit, as well as the vapor pressure control and the scale that is in the ROS 1.
The results are highly repeatable. It has to do with having a lot of control over those previous things that I just mentioned.
Something really cool about this instrument is the auto dry detection. So as evaporation starts to slow down for the sample, the instrument can actually detect that, and it will start to take more measurements, so that it can stop the test right when that sample is dry.
I thought I would also have you talk about, Conner, how this helped to improve your workflow for using it in the lab.
CJ: Yeah, for sure. I think one thing that stands out to me is not having to use weird little buttons on a tiny little instrument. I think for a while there, that was the popular way to build instruments, to make everything standalone. But moving back to having a desktop really streamlines the data side of things.
Also, the instrument itself is just... it's a breeze to use.
ZC: Now let's look at some of the things that you collected. So looking back at dried mango now, ROS 1 is there on the far right of your screen. What is this showing us?
CJ: Right. We're seeing that the ROS 1 is really able to get really close to the reference method for dried fruit, whereas these moisture balances struggle to get all the moisture out. Because the ROS 1 is basically a reference method, it's very consistent with the oven.
ZC: And then let's look at, again, the dry blueberries.
CJ: Very similar.
ZC: So really close to that reference oven method.
Let's look at the protein and greens powder. Again, really close.
CJ: Yeah. I feel like we're negative on moisture balances, just because they really struggle with a lot of samples. But it's fair to say that they don't work very well with these samples.
ZC: Here we are again. This is presented in a slightly different way. Can you explain why you're showing it this way?
CJ: The previous slides showed an actual moisture content difference. This is the moisture content difference as a percent difference from the reference. Because those moisture contents are small for this particular sample, we can see that a small change really results in a large percent difference. Even though the moisture balance might have had somewhere around 1% difference, absolute, it's about a 15% like moisture content difference from the actual reference method.
I just want to point out we're basically shooting for about 2% precision here. So this difference of about 2% is the uncertainty of both the reference method and ROS 1.
ZC: And then next we have the protein powder.
CJ: Again, yeah, the same thing. This is the absolute difference, 0.1%. The moisture balances are overestimating considerably compared to if we look at the percent difference here, about 2% for the ROS 1, and upwards of four to 15% for the moisture balances.
ZC: That 2% for the ROS 1 is, again, within that expected range that was the target?
CJ: Right. I think it's not this slide, but there's one slide after this. We'll look at precision.
ZC: I like this slide a lot. Here, we're looking at a moisture standard. Is that correct? Why is this slide important?
CJ: This is looking at sodium tartrate, which is a pretty common moisture standard. This goes back to method validation.
I was actually shocked at how poorly these moisture balances performed. I was trying to verify them using this typical moisture content standard, and it really overestimated the amount of moisture. I still don't know what's a good moisture content standard for some of these instruments.
I know some companies make their own moisture content standards, but some of those can be very expensive.
ZC: Right. Something like this moisture standard, it should, essentially, read the same for all of these different methods, but like you said, the balances are really struggling here.
CJ: Yeah. So the actual moisture content of that particular standard is 15.6%, and they overestimated it considerably.
ZC: But ROS 1 here is pretty close to 15.4?
CJ: Yeah. That's well within the expected uncertainty.
ZC: All Right. Let's jump to the slide that you were thinking about. What does the slide show?
CJ: This is the percent relative standard deviation, which is the standard deviation expressed as a percent of the mean. We can see that the reference oven and the ROS 1 are basically exactly where we would like them to be, between one or 2%. Can't really get much smaller than that.
The moisture balances, even if they were accurate, they're not very precise, which is two strikes against them in my experience.
ZC: This is one of my favorite questions to ask our clients in the food industry — if they know the precision of their method. Most of the time, they either don't know, or once they do some investigation, I think they're shocked by the numbers that they see.
CJ: Yeah. You should be able to get 2% relative standard deviation or less.
ZC: All right. From here, let's just finish with talking about choosing the right instrument for you and your team. So what instrument makes sense? There's a lot of things that you can think about.
The first thing is that not all moisture analyzers are the same. We've gone over lots of different methods today. Anything you want to add to this?
CJ: Yeah. If you find yourself just doing quick spot checks with a moisture balance, ask yourself – do you even need moisture content? And do you need to be using a moisture balance? Could you use water activity for that metric instead?
ZC: Sure. Because water activity is something that we can measure in as fast as 60 seconds now. And so you're right, if you're doing a spot check, maybe think about what you really need.
The next point we have here is narrowing down your options can be really daunting. There's all these different options, all this different information. What can somebody do to narrow down those options?
CJ: Right. That's tricky. It's a case by case basis here. And so I have lots of suggestions, but I would have to sit down and talk with you, because maybe you’re having trouble getting the moisture out of your sample, maybe you're burning your samples, maybe something else. So I think we can have those conversations.
ZC: Yeah. I think that this webinar is a good starting point, and I hope it's helping you to already narrow down your options and see what's out there, and understand maybe what issues you may be currently struggling with, and what else is out there.
The next out option we have here is that many instruments claim to be multipurpose. What do you mean by this bullet point?
CJ: Well, you can't just put a bunch of grain in a Karl Fischer titrator. It's not going to work. Similarly, you're going to have issues putting certain samples in a microwave or in a moisture balance. So be careful with “all-purpose” claims.
ZC: I think this really highlights how helpful the ROS 1 can be, because it is really sample-agnostic. You can put just about any sample in there, and using that algorithm and that method that I mentioned earlier about tracking the evaporation rate is how we make sure we get the right measurement.
The last one here we have is quality versus quantity. What do you mean here?
CJ: This goes back to the metrics like, do you need moisture content? And this fast moisture content that you're getting, is it reliable? And is it actually helping you with yield?
ZC: Like Conner said, a lot of your products that you're measuring, or things that you're working on, it does take sitting down and speaking with us, and making sure that you get the right instrument.
So here's our contact information. You can also reach out to our sales line, here's their email, as well as their phone number.
We'd love to talk with you. If there's something that caught your attention in this webinar, we're happy to just talk it through with you and understand your products better.
ZC: Water activity is a really high precision, highly accurate measurement. It's something that we have standards for, and it really allows us to pinpoint and understand the water in your product.
So if you're concerned about reaction rates, or if you're concerned about some type of physical transition, like caking and clumping, or microbial growth, this allows you to understand where your product is, and then avoid these issues. It takes a little bit more understanding of your product, and maybe using an isotherm, like I mentioned earlier, but I think it just really helps to pinpoint where you're at for your product.
Whereas moisture content, there's always going to be lots of variance, and it's harder to understand. So like Conner mentioned, moisture content is really good for yield. Anything else that you would add Conner?
CJ: Yeah, I think pretty much sums it up. I think a lot of people are familiar with moisture content more as a concept and so that becomes the de facto measurement – “How much water is in my sample measurement?” — but you should really consider water activity in certain scenarios.
CJ: Yeah. You can't completely seal off Karl Fischer titrator, so you constantly have to keep calibrating it. Every time you take a measurement, you have to calibrate it. So it's a constant struggle to get that instrument stable, because it's always drifting, so water's always impinging on it. It can be slow, but it's something that you always have to consider.
CJ: I think there was an Ohaus one, there's a Mettler Toledo one, and... it was either a Torbal or Veritas. One of those two. I can't remember.
ZC: I think we just tried to pick the top ones that we see people using.
ZC: Yeah. So I briefly mentioned this, but the way that this auto dry detection works is, there's an algorithm working in the background. Basically, what we can detect is the rate of evaporation.
So we're drying down that sample for about four minutes, and right near the end of this test, it can actually determine how quickly the water is evaporating. As that evaporation slows down, the instrument detects this.
To make sure that it stops right at the right point, it starts taking more measurements right at the end of the test, so that it can stop it right when that sample is dry.
This is what makes our instrument really unique, and this is what takes away the need to develop any type of method, because this is already built in and it can detect the slowing of evaporation to stop the test correctly.
What have you seen, Conner? I know we looked at a lot of problematic samples, like the dried fruit and the protein powders. Did you see any type of burning or anything on the ROS 1 that that was concerning?
CJ: No. Obviously, if you crank up the temperature to something that's absurd for your particular sample, then you can get burning. But something we didn't mention was that you can't get the sample absolutely dry. There's no such thing as 100% dry, it's only as dry as the environment. So that's where the ROS 1’s algorithm is doing a really good job in saying, this is a very precise measurement of this equilibrium that we've reached. So I think we've done a good job of modeling that particular stability that lets us say, okay, this is dry, this is a stable equilibrium.
ZC: We didn't have any of those specific samples in this presentation. But in your research, did you look at these types of samples, as well?
CJ: Yeah. I would say that high-moisture samples are the easy samples. The water in them is much more available, it's not bound in any weird fibrous matrix or anything like that. So it's just going to readily remove itself when you heat it up. So those are the samples that are easy to get loss on drying data for. They just might take longer.
CJ: Yeah. This is a new thing that we've developed. I'm not sure if it's been rolled out yet. We had an instrument a while ago that would reduce the vapor pressure in the chamber. We had an instrument that did that, it was called the TrueDry. It used desiccant tubes to blow dry air over the sample, removing the vapor pressure factor from the equilibrium. Unfortunately, it was very complex from a hardware standpoint. So we took that mechanism out, but we’ve been able to use past learnings and modern AI to predict very precisely what the TrueDry used to measure directly.
ZC: So it's like a vapor pressure correction?
CJ: Yeah. So in the case of a particularly high humidity area, we could correct for that, or any variation in environment.
CJ: Well, the best way to validate a moisture method is to use a standard.
So you can find standards for lots of things, moisture standards for grains. They tend to be very expensive, which is probably why we don't see a lot of people using them. But they do exist for a lot of examples, and sometimes they don't exist, so you have to make your own internal standards, which can be a pain, certainly.
But if you really value good moisture content data, if you really value accurate yield data, it's probably worth validating your methods.
CJ: I believe we have, but I don't have that data off the top of my head.
ZC: I will add that we're constantly adding to our data set for ROS 1. So if there's something that you're interested in and you need a proof of concept, we will either conduct that testing here at METER, or we might even have you send samples in.
So we're already doing this with a lot of clients interested in this instrument, who are unsure about making the switch. We either do the research here and then show it to them, or we collect their samples and run the test, and then present it to them.
Case studies, webinars, and articles you’ll love
Receive the latest content on a regular basis.