In case anybody has any interest in how a real materials analysis laboratory works I thought I might write a short primer.

In TV and movies labs often have mystic special tools that can look at anything and tell you everything. If only life worked that way. In real life you have to use a variety of different techniques to look at different materials, different phases of matter, and to determine different kinds of information.

The list of material analysis techniques is extensive. I can't speak to all of these techniques. I have substantial knowledge of two or three that I personally do in the lab, or have done in the past. I have passing knowledge of half a dozen more that my coworkers are responsible for. Mostly I can speak to some general truths that will serve the layman well.

A great many of the techniques determine elemental, chemical, or molecular composition information: Auger(french, pronounced "oh-jay," not "Ah-ger"), XPS, EDX, SIMS, TXRF, ICP-MS, and many more. These techniques are generally some form of spectroscopy, that is, they measure discrete occurences along a continuous spectrum which might be mass or energy, as examples.

Microscopy techniques make up most of the remainder. As opposed spectroscopy microscopy involves magnifying the physical arrangement of a material. SEM, TEM, AFM (also called SPM). Are various microscopy techniques. Auger combines both spectroscopy and microscopy capabilities as it is essentially a modified SEM.

Then there are oddball measurments. XRD/XRR can be used to determine crystaline phases of solids as well as film thicknesses. TM involves a lot of measurements of strain, stress, crack propagation, adhesion strength, and so on. AFM can be combined with various electrical modules to get a host of different electrical information about a sample including magnetic domains, conductivity, and capacitance.

Spectroscopy: in and out

The simplest way to characterize a spectroscopy techniques is to ask "what goes into the sample" and "what comes out of the sample". These questions pretty much define the technique. For instance consider the following:

The "in" category in each case is what the sample is bombarded with, while the "out" category is what the detector is looking for coming off the sample. Auger then hits the target with a beam of electrons and looks at the energies of the electrons ejected (not just any electrons but "auger" electrons which are produced through a particular reaction). Secondary Ion Mass Spectroscopy (SIMS) hits the sample with ions and looks at the mass of secondary ions produced from the surface. Xray Photoelectron Spectroscopy hits the surface with xrays and examines the energy of the emitted electrons. Energy Dispersive X ray (EDX) is the inverse of an XPS- it hits the sample with an electron beam and captures xrays produced. Inductively Coupled Plasma Mas Spectroscopy (ICP-MS) uses plasma to knock ions off of a sample, which are then put through a mass spectroscope. Total Reflection Xray Fluorescence (TXRF) uses xrays to stimulate the sample to release other xrays, the energy of which is characteristic of the material.

Notice that many of these techniques appear complimentary- Auger and EDX both use electron beams, XPS and TXRF both use xray beams. It seems like you should be able to combine both functions in one, but unfortunately this is not generally the case. Consider Auger and EDX. EDX uses pretty low currents for the electron beam in general. It is usually an add on to a SEM (Secondary or Scanning Electron Microscope). If the EDX electron beam were likened to a flashlight the Auger electron beam is a steel cutting laser. It doesn't have much greater current but focuses it down to such a small area that the current flux is enormous. It cooks the hell out of the sample in no time. An electron column designed for one use is ill suited to the other (as evinced by the fact that Auger's can act as SEMs but are very bad at it).

Microscopy: contrast mechanisms

As compared to the Spectroscopy techniques the usual question for Microscopy is "what does it use for contrast mechanism?" That is when you get a picture from the microscope what is the real difference between black and white?

An Atomic Forces Microscope (AFM) uses a very sharp tip which taps along the surface developing a topography map of the surface of the sample that can resolve atomic scale differences. In that case the contrast mechanism corresponds to physical height differences.

A Scanning or Secondary Electron Microscope (SEM) uses an electron beam which is scanned over the sample while a detector picks up other eletrons. These can be backscatter electrons (the initial electron beam reflected back at the detector) or secondary electrons (electrons from the material knocked loose by the incident electron beam). Backscatter electrons show up more with heavier elements. Secondary electrons show up more with lighter elements. In either case the contrast mechanism is likelyhood of the material to put an electron into the detector.

A Transmission Electron Microscope (TEM) shines an electron beam on a very thin sample. A detector picks up the same electrons on the other side after thay have been partially blocked or more likely diffracted through the material. This gives a picture of the material. The contrast mechanism then is the ability of the material to transmit an electron, which is related to the material as well as the crystaline phase of the solid and probably a thousand other things besides (TEM gives gorgeous pictures but it's really finicky).

An optical microscope of course has a contrast mechanism based on the absorbtion or reflection of light wavelengths, although our brains are well adapted to converting that to three dimensional topography information.

Trying to compare information from one microscope technique to another is damn difficult because the different contrast mechanisms mean the same thing can look entirely different. On top of that microscopy can be a little tricky because we are an extremely visual species. It is very easy to look at a SEM picture and assume the light areas are taller (closer to the eye) and the dark areas are pits but that's entirely an artifact of how our brains interpret light images.

Two case studies

So let me give two examples of techniques I know pretty well: AFM and XPS.

AFM (Atomic Forces Microscopy, again) is a microscopy technique used to look at surface topography. You can use it to determine, on the small scale, roughness and step heights mainly. It is sensitive enough that in ideal conditions you can determine atomic scal differences (a couple angstroms). Much more easily it can be used for sub-nanometer measurements (less than 10 angstroms). For comparison the smallest virii are ~20 nanometers. It does this by tapping a tip on the surface using a feedback loop to maintain a constant amplitude as the tip moves back and forth over the surface. If the surface rises up the tip has to pull up to keep the amplitude from dropping. If the surface drops away the tip has to move down to keep the amplitude from rising.

It uses piezoelectrics to do the majority of the movement. Piezos have a characteristic by which if they are deformed they produce a voltage delta across their ends, or if a voltage delta is applied they deform. This makes them very useful, although the range of their movement is narrow (they are still solids). Since Piezos provide the x-y movement as well as the z (height) the AFM can only scan a limited area at a time. It'd take me some 200 scans to cover an entire average bacterium. Clearly if you need information about long scale issues the AFM is limited.

Other difficulties include that any sort of forces that interfere with the normal oscillation of the tip will be (incorrectly) interpreted as topography changes. Something as simple as an adsorbed fluid layer on the surface can cause a problem. A charged sample can be a nightmare as the electrostatic repulsion or attraction between sample and tip will screw with the topography reading. If the chemical nature of the sample has an affinity for the chemical make up of the tip then there can be attractive forces that cause difficulty (Silicon tips are common for AFM, and silicon tends to stick to other silicon surfaces, gosh do you think we ever use silicon in our semiconductors? Yeah...).

AFM really can't give you any sort of chemical, elemental, or molecular information, although it can sometimes pick up changes in material. If the tip interacts with one material one way and another differently you can sometimes pick this up with phase imaging, although all you know is there is a difference you can't identify what the difference is.

So AFM has some great strengths but also significant limitations, such that there is definite call for other microscopy techniques. SEM for instance can look at an area it would take me years to scan through AFM.

XPS is a totally different bird then. Xray Photoelectron Spectroscopy. Photoelectrons are electrons ejected from a surface when it is hit with elecrtromagnetic radiation. In this case the EM radiation is an xray beam. The electrons that come off are gathered by a collector which determines their energy. Why?

Well in the first place we monochromate the xray source, this means all the xrays reaching the sample are the same frequency. The energy carried by an EM wave is proportional to its frequency. Consequently each photon striking the surface has the same energy. When a photon ejects an electron energy is conserved:

E(gamma) = E(e-) + E(binding)

The energy of the incoming xray beam (gamma) has to equal the kinetic energy of Electron plus the binding energy (how strongly held the electron was in the atom before the xray came along). Each element has a distinct set of binding energies that characterize it based on it's atomic number (number of protons) and the arrangement of its electron orbitals. By looking for which electron energies reach the detector you can determine which elements are present on the surface. It is slightly more complicated than this (you have to include a work function for the tool in the energy equation above) but it gives you the idea.

Not only can the XPS determine elemental information it can often give chemical state information. What's the difference?

Elemental information: 33% Si, 66% O
Chemical information: 100% SiO2

Elemental information only tells you which elements are present, it doesn't tell you what they are doing. Chemical information gives you a sense of which elements are bound together. In this case the binding energy of Oxygen is a little bit different depending on if the Oxygen is itself associated with Nitrogen, or Silicon, or Chromium, or whatever. XPS can *sometimes* distinguish these differences, which is a major advantage over many other spectroscopy techniques.

XPS has to take place in a ultrahigh vacuum environment which is a pain in the ass, frankly. It has to be this way because you can't have the photoelectrons running into any atmosphere between the sample and the detector, if they did your energy equation above loses all meaning. Making such an ultrahigh vacuum environment isn't that hard, making it accessable is. By adding a loadlock you make the whole thing very hard to keep at the pressures you need. A hell of a lot of work goes into monitoring and maintaining the vacuum, and very bad things happen to the tool if the pressure spikes. You've got roughing pumps, turbo pumps, sublimation pumps, ion pumps, You've got like three different types of vacuum gauge each sensitive to a different range of pressures. Every one of those components is subject to breaking (frequently sometimes).

AFM can be done in vacuum but usually isn't, XPS has to be done in vacuum. In fact most of the materials analysis techniques require at least mid if not high or ultrahigh vacuum.

XPS is also a slow technique. Incoming xrays can cause the sample to produce secondary xrays, electrons, Auger electrons. Of these it is only the second category that the XPS looks at. Consequently it has to spend a fair amount of time just collecting in order to get enough data to produce usable results. It is not infrequent to set up a run and come back to it a day later. The vacuum system alone means it take at least 30 minutes to get a sample in the machine at all.

XPS can't provide molecular data. If elemental is the simplest composition data, and chemical is the next, then molecular is the third. Imagine even a simple organic molecule (say C4H4O4) XPS could give you some sense of which elements were bonded to which but the overall molecule would be very hard to determine (we have some carbon bonded to hydrogen, some oxygen bonded to carbon, some oxygen bonded to hydrogen, but does it all fit together, and if so how?).

Actually that raises another limitaion of XPS- it can't detect hydrogen at all. Depending on the source of the xrays (and their resultant energy) it will be blind to some other materials at the other end of the periodic table, but these areusually of limited interest anyway.

Other spectroscopy techniques have different blindspots, and offer different information. Unfortunately there is no One Technique to rull them all. You have to use an array of techniques to really understand what is going on.