I spend my days in a chemistry lab making drug-like molecules. A lot of these end up being small quantities (less than 0.1 g!) and usually have the appearance of a white/off-white powder. Occasionally I get a colour which is very exciting.
The question a non-chemist might ask is “how do you know you’ve made the product”? Lots of different analytical techniques have been developed over decades to help chemists determine the chemical structure of the products they’ve made.
In this post I will give you an introduction to the seven analytical techniques I carry out on my samples to prove I have made the right molecule. These various bits of data go in my experimental write up which will make up a large chunk of my thesis.
1) LCMS
LCMS stands for Liquid Chromatography Mass Spectrometry. This combines two techniques: liquid chromatography allows to you separate a mixture into its components while mass spectrometry will tell you the molecular weight of each of those components.
Ideally the read off of a pure sample will just show that there’s one component in your mixture and the molecular weight from the mass spec will match the calculated weight of your product (i.e. the number you get when you add up the molecular weights of the individual nitrogen, carbon, oxygen, hydrogen etc. atoms).
2) 1H NMR
NMR stands for nuclear magnetic resonance. There are many different “flavours” of NMR. Proton NMR (written as “1H”) looks at the different hydrogen atoms present in a molecule and the different environments they’re in.
For example, some protons will be attached to carbon atoms, while others will be attached to nitrogen atoms. This means they will have a different “magnetic environment” due to the different number of electrons in different atoms and how they’re distributed between different atoms and chemical bonds.
Below is an NMR spectrum for ethanol, the type of alcohol found in alcoholic beverages. There are three different proton environments in the molecule of ethanol: a hydrogen attached to an oxygen (blue); a hydrogen attached to a carbon that’s attached to a carbon and an oxygen (green); and a hydrogen that’s attached to a carbon that’s attached to another carbon (red).
The appearance of the peaks is affected by the number of nearby protons e.g. the CH3 peak is split into a triplet because there are two protons on the adjacent carbon atom. Undergraduate chemistry degrees have entire modules dedicated to interpreting NMR spectra, so I won’t go into too much detail here.
***
LCMS and 1H NMR are usually the minimum pieces of data you need to be sure you’ve made the right compound to move on with other chemistry. The two pieces of data combined give sufficient evidence that you’ve made something that weighs the same as the expected molecular weight; has the expected number of protons in the predicted magnetic environments to be the right products; and gives an indication of the purity of the compound.
For a PhD thesis you usually need to provide “full characterisation” for compounds which involves more analysis than those two techniques. I agreed with my supervisor that I would only get full analysis for final compounds that are to be tested by a biologist or compounds that don’t appear to have been made before by anyone else.
I determine if a compound is new by doing a literature database search and if zero hits come up in the search, I can assume no one has published the synthesis of this molecule before. I need to provide as much information as possible to prove it is the right compound. Below are 5 additional types of analysis I get on samples that fall into the final/unknown category.
***
3) HRMS
High resolution mass spectrometry is just a slightly higher quality version of LCMS. Having two mass spectrometry experiments that match up gives more concrete evidence about what the molecular weight of my chemical product is.
4) 13C NMR
13C NMR is like 1H NMR but it instead looks at the different environments of carbon atoms in a molecule. There are lots of different types of NMR experiment that are used depending on atoms are in your compounds e.g. fluorine, phosphorus. Going back to the ethanol example, I would expect to see two peaks in a 13C NMR spectrum of ethanol because there are two types of carbon atom present in the molecule.
There are also different types of NMR experiment that combine different types of NMR e.g. COSY, NOESY, HSQC, HMBC etc. which I may talk about in another post some time.
5) IR spectroscopy
IR spectroscopy involves firing infrared radiation (IR) at your sample to determine what types of chemical bonds are present. Different chemical bonds have different energies and absorb and emit different levels of IR. An IR spectrum (see below) will tell me if I have C-H/C=O/C-N bonds present in my product, but not how many there are.
6) Melting Point
We know that ice melts at 0 °C. Similarly, different products will have characteristic temperature at which they change phase. Recording the melting point of a sample will allow a chemist making the same compound in the future to compare their melting point with yours. The melting point also gives an idea of how volatile a compound is i.e. how easily might it boil off into the atmosphere.
7) Rf value
The final type of analysis I get is an Rf value, which stands for retention factor. You may have carried out chromatography at school or a science fair, whereby you separate a mixture into its components by running a liquid through a medium such as filter paper.
I use a slightly more sophisticated version of this technique called thin layer chromatography (TLC) where I run a spot of my product up a silica plate. The Rf value is a ratio of the distance travelled by the spot of product divided by the distance travelled by the liquid.
This gives an indication of how clean the product is (I’ll see multiple spots if there are any impurities). It also tells you how well the product dissolves in that particular liquid – it will travel further up the plate if it dissolves really well in the e.g. water that I run my TLC plate in.
There are other types of analysis beyond this set that I could also use but they are either excessive, time consuming or unnecessary for the type of molecules I make. I think of full characterisation as compiling as much evidence as possible to be sure I’ve made the right molecule, fitting various jigsaw puzzle pieces together to build up a concrete picture of what my product is.
Some of the techniques are qualitative rather than quantitative and aren’t necessarily as sophisticated as others e.g. IR just tells me I have a C-N bond present while NMR will tell me if there are protons attached to that nitrogen, how many, and what other protons/carbons they are close to.
These seven types of data I get are the standard ones expected in most theses and journals, but it wasn’t always that way: in the 1960s chemists only had IR and MP techniques available whereas now 1H NMR and mass spectrometry are the standard minimal analytical techniques, with MP and IR as added extras.
Last month a new analytical technique, called micro-electron diffraction, was published and generated a lot of excitement amongst chemists – on my Twitter feed at least! MicroED might save chemists a lot of time analysing compounds in the future by rapidly generating a 3D “X-ray skeleton” of the molecule using electron beams. This would save having to do all these other analyses.
For more information about different types of spectroscopy and analytical techniques used in chemistry, check out the resources below:
Does it surprise you how much analysis is done on one sample? What types of analyses do you use in your work to make sure a job has been done properly? Let me know in the comments below.
The Chemistry of a PhD 