Nano- and Micromechanics (Kathy Walsh)
Sep 5, 2023
For more information, visit https://nanohub.org/resources/22623
Kathy Walsh
6/3/15
"Nano- and Micromechanics"
Well, good day to you all.
This is a slight difference from the previous talks.
Instead of talking about what your material is made of, we're mostly going to focus now
on what your material actually does.
At the most basic level, mechanical testing tells you about how your sample responds when
you do something to it, when you apply a force to it, when you squeeze it or compress it
with a really sharp tip, or you stretch it or shear it.
Today we're going to focus on compression in a really small geometry, which is basically
just poking your sample with a really sharp stick.
You need to choose how to address the mechanical characterization of your sample.
Sometimes this is a choice that needs to be inspired by at least a vague idea of what
your sample is made of.
If you're testing really big concrete pillars that people are going to use to build a building,
it's going to be different from if you're testing a hydrogel, so you have to have an
idea sort of of what sample properties you expect, so you can get force resolution.
The reason we're focusing on nano and micromechanics is because of the lateral resolution issue.
Your samples may be inhomogeneous, and so you want to know whether you're interested
in the overall properties, in which case you'd want a larger scale testing technique, or
whether you care about the viscoelastic properties of chocolate chips.
The range of mechanical testing techniques I'm focusing on here, because that's what
MRL specialties are, range all the way from the nanoscale, where you have nanometer scale
indents into your sample, and you're probing nanonewtons to micronewtons, up through the
microscale, which is many tens of nanometers into your sample, and several micronewtons
up to millinewtons in force, all the way up to a more macroscale.
You can compress your samples or twist them, a lot of people do shear testing, we don't
do that so much here, or you can stretch your sample to get stress-strain curves.
First off, I'll justify why mechanical testing is useful.
It's mainly useful for applications.
You've got a material, you want to figure out how your material interacts with the other
stuff that the industrial engineers are going to be using to build your sample into the
next awesome product.
Things like, for example, false teeth.
You want to make sure that a tooth that your dentist implants is not going to wear your
real teeth as you chew over the course of 20 or 30 or 50 years.
This is why nano-indentation is big in the dentistry field.
The other reason we're focusing on nano- and micro-mechanics here is, again, your samples
may be spatially inhomogeneous.
You may not have a perfect material that is identical throughout its entire bulk, or your
samples may be just plain small.
It's very hard to test a nanopillar if you've got two plates that are the size of your hands.
What sort of properties people measure mechanically can be batched into about two different categories.
Here I'm saying quasi-static, it's not perfectly static because you have to move in order to
apply a force to your sample, so it's quasi-static.
These are things like elastic modulus, Young's modulus, hardness, fracture toughness, that
type of one-shot material testing.
A lot of samples have time-dependent mechanical properties, so we do dynamic testing.
This will tell you storage and loss modulus and the relationship between them.
This is what Ted was talking about yesterday with AMFM.
The techniques that we use, often people will acquire stress versus strain curves and use
this to get the Young's modulus, or load displacement curves.
In AFM, these would be force curves.
I'm going to focus primarily on this today.
Dynamic testing will tell you properties as a function of time or as a function of frequency,
because if you take something that goes for a long time, you can Fourier transform that
into a high-frequency measurement.
You can do long-time measurements like creep or stress relaxation.
The scale from nanoscale, AFM, through nano-indentation to more macro-scale, that's the DMA I'm showing
you there, can be both quasi-static and dynamic testing can be done in these regimes.
In AFM, we saw this yesterday, Scott talked about force curves, Ted talked about AMFM.
There are a variety of nanomechanical techniques for AFM which do truly nano-nanomechanics.
The nano- and micro-indentation do nanoscale to micro-scale testing, and then, of course,
there's the large-scale testing if your sample is a couple millimeters wide.
Today, I'm going to focus mainly on nano-indentation and mostly on finding the Young's modulus,
which is related to the stiffness of your sample, and the hardness, which is related
to the amount of plastic deformation.
This is elastic, this is plastic.
Elastic returns to how it started.
Plastic is just you punch a hole in something, and the hole's still there when you stop punching
it.
You might say, why are you focusing on nano-indentation instead of micro-indentation?
Is it just because nano is a cooler prefix?
Well, yeah, it is, but sometimes micro-indentation is actually more useful because if you're
going really deep into your sample, then small irregularities on the surface don't matter.
Then little surface stickiness doesn't really matter because you're probing more of the
bulk properties of your sample.
Many nano-indenters also have micro-indenter attachments.
This is an example of one that I have in my lab.
Both nano-indentation and micro-indentation are examples of instrumented or depth-sensing
indentation.
This is in contrast to the mechanical testing method since ancient times where you would
do hardness testing by just stabbing something diamond into something else and seeing the
size of the residual imprint.
Here we're recording the force that we're applying and we're recording the depth to
which you're lowering your tip.
You're poking a sample and you're recording its response as you lower and raise the tip.
This is a picture of the nano-indenter I have in my lab.
At its most fundamental level, a nano-indenter is about a tip and a sample and what's applying
and measuring the force and the displacement.
This is usually mounted on some sort of nano-positioning system to allow you to get the lateral resolution.
Think of chocolate chip cookies again, but you don't need nano-positioning for chocolate
chip cookies.
This all is mounted on a very stiff frame.
You might think this granite frame is overkill, but the point of having a stiff frame is so
that all the bending that takes place in the sample, all the deformation, all the mechanical
changes are due to your sample and not to the instrument.
You want to measure your sample, you don't want to measure the properties of the instrument.
It's my job to measure the properties of my instrument.
It's your job to measure the properties of your sample.
Since it's basically about the tip and the sample, that's what we'll focus on now.
The nano-indenter tips are almost always made of diamond or sapphire.
These are things that are hard, so they're not going to be softer than your sample.
These are things that are stiff, so they're not going to be springier than your sample,
and you're measuring your sample, not your tip properties.
Nano-indenter tips come in a variety of shapes depending on what you want to do with them.
If you want to induce a lot of plastic deformation, you want a very sharp high aspect ratio tip
so it can just dig right in.
If you want to look at elastic deformation, a broader tip that can apply the force over
a wider range.
If you want to look at, for example, a soft material, a biological material, you may not
want a very sharp tip because it might cut your sample.
I'm talking about the tips here.
What do they actually look like?
This is a residual imprint left in a piece of aluminum foil by a nano-indenter tip.
This went in a few microns.
The most popular shape of tip for a nano-indenter is a three-sided pyramid, Berkovitsch tip,
named after the person who designed it.
The typical radius of the very end of the tip is something like 100, 150 nanometers.
Now micro-indenter tips are larger.
They're in fact so large that the residual imprint is usually visible in an optical microscope.
Here I have a bright field and dark field optical microscope image of a Vickers four-sided
pyramid, which is a typical tip shape for a micro-indenter.
If you remember yesterday, Julio, talking about the optical characterization, bright
field is what you think of, you just shine the light on your sample.
Dark field is illuminated from the side so you can see nice little, you can see the edges
of the indent here.
Now if you're doing micro-indentation on soft materials, often you'll have a sphere.
Sometimes people use ball bearings.
The contact area between the tip and the sample is the most important calibration for you
to do.
You do it every time you're taking a measurement.
This is critical.
I'll show you some equations later for the specific purpose of pointing out that it's
critical because you know this must be serious if I'm going to put equations up about it.
This contact area between the tip and the sample depends on the depth to which you're
indenting in the sample.
Now I'm not talking about this outside sort of perimeter.
I'm talking about the physical area, how much of your tip that's buried into the sample
is actually touching the sample.
Of course, unless you have a perfectly, say, cylindrical stick, if you have a pyramid that's
going into your sample, the deeper you go, the more of your tip is going to touch more
of your sample.
This also depends on the surface roughness.
You can see this indent is very deep.
This is an AFM image, brighter is taller, darker is deeper, and there's some small surface
roughness here.
You can see there's some irregularities due to this tip.
For the Berkovich tip that I'm showing you here, you can see that the surface roughness
is kind of dominating.
The shape of the indent is a little vague here.
Let me try to give a more familiar example because I'm sure you don't do nano-indentation
every day.
You might end up poking stuff with pens every day a little bit more often.
We're talking about the contact area between this pen and a piece of foam.
If you've got nice, smooth sample, nice, smooth foam, then it's very predictable as you poke
the pen into the foam that you know how much of the foam is going to touch your pen.
If you have a very rough surface, even if it's made out of the same material, if you
come down to the same depth and you move the tip sideways, you're going to have more or
less contact between the tip and the foam.
You won't know what your actual contact area is.
This contact area is calibrated and then you apply that calibration to the data you get.
In order to smooth out your samples, many people will polish them, but you can't polish
a hydrogel, so sometimes this doesn't work for people's samples.
Also, polishing tends to leave little bits of the polishing material behind, like Rick's
really cool way of getting the silicon carbide thing.
That would be an artifact in this case that you wouldn't want.
This leads into one of the two rules of thumb of nano-indentation I'm going to focus on.
This is the 5% rule, and 5% is the same as 1 20th, and to me it makes more sense to think
of it the other way.
I'll talk about it in terms of 20 times instead of 5%.
You want your indents to be 20 times deeper than your surface roughness.
This is not deep enough.
The reason is because if you've got a little surface roughness and you go in 20 times deeper
than that, then the surface roughness comparatively doesn't matter very much.
This can get tricky though because of the other rule of thumb I'm going to show you,
which is particularly important for thin samples.
This is the 10% rule, also known as the substrate effect.
This is a rule of thumb to try to get rid of the substrate effect.
The technical explanation is that there are strain fields bouncing off the substrate here.
I like analogies.
Let's use the pen poking into a foam, or if you think of just kind of poking yourself
in the arm.
If you poke yourself in the arm gently, it feels like the deeper in you go, the more
properties of what is underneath you're starting to feel.
The typical rule of thumb people give is you want to keep your indentation depth 1 10th
of the total sample volume, or total sample depth I should say.
This is tricky if you're growing thin films that are very, very, very, very thin because
you've also got the surface roughness from the 5% rule.
You need a perfectly flat, very thick film in principle.
That never happens.
Now, I should say, these are rules of thumb.
They're not to be taken as, oh no, I can't get my 5% rule and 10% rule to work.
I shouldn't even try.
This depends a lot on your sample material.
If you think of the foam on a board, if you're pushing into the foam on a board, you can
go pretty far down and it still feels like foam, but then say you're poking with a pen
on a board on a piece of foam.
If you poke the pen on the board, you're still feeling the foam.
You're not deforming the clipboard here with the pen.
You're deforming the foam that's underneath.
On this case, if you have a stiff sample on a compliant substrate, you'll probably
see the substrate effect even earlier.
This could be a problem for people who are trying to build devices on soft materials.
If you're trying to measure the modulus, you might have to think of an alternate way to
do this.
I'm showing you a much more large scale, non-instrumented micro indenter here just because it's easier
to see.
I've got my sample here.
There's a little reflection.
The tip's going to come down.
As you approach the surface, you sense where the surface is based on the load that's detected.
Of course, here it's non-instrumented, so we just drop it from the sky.
When you're on the sample, this is where you start collecting your data.
You push in and collect data.
You can hold for a while to check and see if there's a time-dependent change.
Then you pull out, still collecting data.
When you're off the substrate, you can also image and see if there's a residual imprint.
Looking at the size of the residual imprint is a traditional hardness test.
This is what the data look like.
This is a load displacement curve for nano indentation.
The loads are in the micronewton to millinewton range, and the displacements are in the tens
of nanometers usually to microns range.
As you load, if you were to have a perfectly elastic material, this relationship between
force and displacement here would be the same as you go in and out because everything
perfectly returns to the way it was before.
Nobody has perfectly elastic materials.
Once you've loaded to whatever peak load you want, or peak displacement depending on how
you want to do this, you can hold or do other things to measure time-dependent properties.
Then the unloading curve is what most people use for at least their initial analysis, especially
for elastoplastic materials.
Here are the equations that I was showing you just to make a point.
Don't bother writing this down.
Just look up fundamental equation of nano indentation, or look up last year's talks PDF where I went
into detail on it.
If somebody tells you do nano indentation, this is probably what they want you to find.
They probably want you to find the reduced modulus, which is related to the sample stiffness,
which comes off the slope of the unloading curve, and the hardness, which is just related
to the peak load up here.
The reason I show this is because this A here, this is the tip area function that I was talking
about before.
The contact area between the tip and the sample matters a lot, especially if you've got hardness.
If you want to measure the hardness for a sample, make sure your sample is smooth enough
that you get a really credible contact area calibration.
Once you've got that value, you can turn this into the Young's modulus if that's what you
want to publish, but for the Young's modulus, we already know this, it's a diamond tip,
you need to know the Poisson ratio of your sample.
The Poisson ratio is, if you think of a rubber band and you stretch it, it's the relationship
between the stretching and the amount the other dimensions shrink.
This is kind of hard to measure, and so many people don't know the Poisson ratio for their
sample, especially if it's a new material.
So many people just quote the indentation modulus.
Indentation modulus, reduced modulus, effective modulus, these are all the same thing.
I've talked a little bit about doing creep testing and stress relaxation tests.
This is an example of the difference between them.
Many or I probably would even say most materials are somewhat time dependent in their mechanical
behavior.
To do a creep test or relaxation test, you load to a certain rate or displacement and
then just hold for a while and see what happens.
See what changes over time, but here's the caveat.
If you're doing this, beware of drift, particularly thermal drift, particularly in the middle
of winter when it's really cold outside and you've brought your sample from another lab
across campus and it's been outside for however long you had to walk over here, then you sit
it in a nice 70 degree Fahrenheit instrument enclosure.
It takes a while to warm up completely, so your sample's going to be expanding as it's
warming up while you're trying to do your measurement.
If you do a proper creep test, which should be several minutes long preferably, you may
see this thermal expansion just completely dominate over the mechanical properties.
Something that people do, and I won't really go into this very much because I think it's
going to be touched on in the next talk, is doing dynamic testing where instead of holding
for a long time, you can do things at a high frequency because it's a Fourier transform
relationship.
Mechanical properties can be obtained as a function of depth in nano-indentation.
This is where nano-indentation might sometimes be a more useful technique than AFM, which
is something I don't usually say because really I started off by doing AFM.
Nano-indentation might be a little better if you need deep depths here.
This is looking at the sample properties hardness and reduced modulus as a function
of depth over several hundred nanometers.
You can focus on smaller scales.
You can focus on larger scales.
This is just an indent I did, well, series of indents.
For shallower indentation, surface effects like roughness matter more, and the quality
of your tip calibration and how sharp your tip is.
This is the 5% rule I was talking about.
Make your sample smooth.
On the other side of things is the 10% rule.
You might start to measure the substrate properties.
In this case, I certainly didn't.
It flattened out with time, but you'll sometimes see the properties go up or down, and then
you can start to identify what the substrate effect, where it matters for your material.
If your samples are really thick, it's never going to matter at all unless you have really
thick samples on PDMS or something.
I should also mention that I took these indents at a bunch of different spots on the sample.
These are them overlaid at different loads in different locations, and it's uniform.
Nano indentation will give you nice uniform testing at different locations.
Now, if you're doing soft materials, everything is harder because nano indenters are traditionally
designed for the engineering materials community.
This is traditional material science.
If you're doing soft and sticky stuff, this might not be the best designed instrument
for you.
You might want to try micro indenters or large-scale testing like DMA or AFM-based techniques.
If you're doing biomaterials, like Luann said yesterday, fixative changes mechanical properties.
If you're fixing your sample, it is important to do controls, do comparisons, although I
should also mention that your sample dying and completely crumbling will change its mechanical
properties too.
Designing your sample preparation for biosamples is always a little bit tricky.
You might also need to work in fluid.
If you're doing samples, especially things with hydro in the name, like hydrogel, you
want to keep them wet.
Plenty of samples will crack when you dry them out, so you can't dry them out and then
wet them later.
You want to keep your sample wet, and you want to keep the instrument dry because electronics
do not like being soaked in buffer solution.
There are often special nano indenter tips that are extra long so that they can fit down
into a thin fluid layer without the fluid wicking up into the transducer.
If you look closely at the end of this, this is my largest diamond tip.
You can sort of see a little curvature there.
That's a hundred micron radius conspherical diamond tip.
There, you've seen a nano indenter tip.
For sample preparation, my first advice is know what your sample looks like on a larger
scale so you can figure out whether you're likely to get decent results or whether you
need to do some more preparation at the beginning again.
When you're mounting the sample, be careful to stick it down in a good way.
You can't just sit your sample in some fluid and hope it doesn't float, and never, ever
use carbon tape for nano indentation.
It is way too compliant.
All you'll measure is the glue, the carbon tape there, so don't use carbon tape.
Measure your sample, not your sample mounting.
Most people use super glue, and this seems to work very well for most samples.
If you've got porous samples, beware that the super glue will wick its way up there.
If you've got incredibly thick porous samples, well, this is not to scale.
If you've got incredibly thick porous samples, then it doesn't matter so much.
If you're doing hydrogels, one thing you can do is cast gels directly onto whatever
you're going to be measuring them in, or you can try to clamp down your samples.
Be careful that when you clamp down your samples, they don't bow up, or you don't clamp down
and distort the area you want to measure.
Once you've got your data, most people start with this model, which is beautifully successful
for elastoplastic materials, which is kind of many of the classic engineering materials.
I like this.
I like this piece of aluminum here.
But your samples may not be like that.
If you've got sticky samples, compliant samples, this is an alternate model, JKR.
It has nothing to do with Harry Potter.
You can have poroelastic samples, which are porous, and you may have fluid flow through
them that you need to consider.
Viscoelastic samples need time-dependent testing, because if you do your indents at five seconds
to load into your sample, and somebody else does the indents at two seconds to load into
the sample, you're going to get different results.
It's not because either of you have done bad indents.
It's just because your sample has different properties at different loading rates.
Poro-viscoelasticity gets complicated.
Thin films is the topic of the next talk.
There are several different models that people have come up with to try to fix this problem
of it being really difficult to satisfy the 10% rule for something which is only 10 nanometers thick.
This is a series of pictures of the toys I have to play with, which means the toys you
have to play with.
I have a nano-indenter with a micro-indenter attachment, which is where I took all the
data from today.
DMA, usually set up to do tensile testing, but you can also do dual-cantilever bending
or single-cantilever bending.
Two AFMs that are like this and one AFM like this, they're all wonderful and easy to use.
These are the ones that can do what Scott and Ted were talking about yesterday.
If you're interested in nano-indentation, I recommend both of these books.
If you're interested in soft materials, even if they're not biological, I highly recommend
this book.
It is a genuinely sincere description of all the difficulties that you're going to run
into so that you know that other people have suffered through this before and come up with
clever solutions that will hopefully make your life easier.
Both of these books are available because we subscribe to them online through the library.
If you're a U of I student, just go and download them.
Thank you very much.
I'm happy to take questions.

