Recently in AP Chem I did a lab that involved doing various tests (melting on hot plate, flame test, etc.) to identify a metal out of a list of metals.
This had me wondering: how do chemists identify mystery substances in the real world? What tests are used?
Let's imagine aliens bestow upon us a block of some unknown pure substance. Guess we'd better figure it out!
One of the interesting things about this Universe is that, no matter what that material is, it must be composed of some combination of the (known) chemical elements, almost certainly the first ~100 or so in the periodic table (excepting some very unusual edge cases, such as stellar material, which isn't stable in conditions even vaguely close to ambient anyway). So the first order of business is to figure out which of these elements are present, and in what proportions. There are many methods to do this, but generally it's a good idea to use a procedure which does not significantly affect a sample (so-called "non-destructive analyses").
A reasonable option with no prior knowledge is to use X-ray fluorescence (XRF) spectroscopy. Nowadays there are small handheld "guns" that blast a sample with X-rays, which are absorbed by the inner-shell electrons of the atoms in a process called electronic excitation. When these excited electrons go back to their energetic ground state, they release X-ray photons of very specific energies depending on what atom they're attached to, and these photons can be measured. Here is a very nice video from Ben Krasnow of Applied Science which shows one of these XRF guns in action. He goes into the operation principle a bit and tears down the instrument.
There is another class of options, such as inductively coupled plasma mass spectrometry (ICP-MS), which is a destructive technique where you break a portion of a sample down into a plasma of its constituent atoms, and then fling these resulting ions through a magnetic field and measure their trajectory - the exact trajectory will depend on their mass-to-charge ratio, which can be measured very accurately and compared to the known elements. For these techniques (as well as many others), you typically have to "prepare" the sample for measurement, and unfortunately this can add some confounding factors.
Great, so now you know precisely which elements are in a sample and what their proportions are. Are you done? Not even close - the way those atoms are put together matters at least as much as the identity of the atoms. The number of different ways atoms can combine very quickly leads to dizzying complexity.
With a little bit of luck, your sample is composed of crystals. If the crystals are at least a few dozen micrometres across, you have a good chance of performing X-ray crystallography. The geometric regularity of crystals allows X-rays to penetrate the sample and mutually interfere as they bounce outwards, in a way that very sensitively depends on the exact placement and identity of atoms in a sample. This generates a so-called diffraction pattern, which can be very carefully analyzed with a lot of number-crunching, allowing back-calculation of the structure of the crystal, and therefore of the substance. This is perhaps the golden standard of chemical analysis. In the past 10-15 years or so, a similar technique called cryogenic electron microscopy (cryo-EM) has risen to importance, which allows a similar measurement of crystals by shooting electrons into a sample, instead of X-rays. Though there are significant additional complications, the main benefit here is that the crystals being measure can be much smaller, on the order of nanometres across.
Of course, the unknown sample may not be crystalline at all. From this point onwards, it becomes less clear what direction should be taken, and you get into a lot of case-by-case considerations. For example, nuclear magnetic resonance (NMR) spetroscopy is another very powerful and general technique which can provide information on the connectivity of atoms in a sample, but in a more indirect way than crystallography. Typically a sample is placed in a very intense magnetic field, often 10 000 or 100 000 times greater than Earth's magnetic field (enough to levitate small animals), which aligns the atomic nuclei of a sample in the direction of the magnetic field. The sample is then pulsed with radio waves, which knocks the nuclei over, and as they attempt to reorient themselves they emit photons of very specific energies, which once again can be used to infer the environment surrounding these atoms. This technique is typically very effective for organic substances composed largely of carbon and hydrogen, and less so for inorganic substances. There are many, many, many more techniques, each with their own strengths, weaknesses and regions of applicability. It takes a skilled chemist to figure out what the next move is.
Now, after all that, here's the kicker - virtually all real-world samples are not pure substances, but a mixture of several, sometimes even thousands, with their relative proportions varying by over a dozen orders of magnitude. The mere separation of these mixtures to allow accurate characterization of the individual components is its own huge field, full of complexities. It is not inconceivable for an unknown sample (especially one coming from a completely unknown environment) to admit only a superficial chemical analysis, even with the best techniques available today. Frequently, very special samples (such as space rocks) are analyzed to the best of our current ability, and then simply carefully stored for decades in the hope that that more advanced techniques can one day be found and allow an old sample to be further studied.
