Given what we know of physics and chemistry today, approximately how many molecules are possible?
Is there a specialty in physics or chemistry which studies ways to assemble new molecules?
I’m not sure how you define types but there are an infinite number of different possible molecules. Some molecules can form chains so you could theoretically always make one that’s bigger.
Possible chemical space is very large. As the author says:
It is estimated that there are about 1 billion chemically reasonable molecules with up to 13 heavy atoms (C, O, N, S, Cl, which covers a lott of the stuff you’re likely to see in a ‘typical’ organic molecule, but does neglect, for example, the other halogens, which are also perfectly reasonable in organic molecules). A 13 heavy atom molecule is really pretty small in the scheme of even small molecules, let along macromolecules.
“Possible” or “stable”?
Let’s put it another way. The human genome has about 9 billion DNA base pairs. But each base pair could be A-T, T-A, C-G or G-C. So 9 billion times 4. Except it’s perfectly chemically plausible to have all sorts of nucleotides that differ from adenine, guanine, cytosine and thymine just by a few tiny differences. The number of plausible nucleotides that could be stuffed into a DNA-like molecule is really large. And so the number of unique polynucleotide chains gets really large, really fast. And that’s just for a particular type of molecule where the monomers are arranged linearly.
Anyway, the number is, like, really really big.
How many ways can to stick two 2-4 lego bricks together? It’s a lot more than you think. Now imagine that there’s no theoretical limit to the size of the lego contraption, and there are over 100 types of bricks. Of course, most of those can’t be used in organic molecules, but even with just C H O N S and Cl you’ve got billions of billions of combinations.
9 billion times 4? Hah! That’s puny. The number you actually want is 4 to the 9 billion power. Let me put that number into perspective for you… Wait, no, I can’t. There just isn’t any way. If you tried to write that number out in standard notation, you couldn’t do it in your lifetime.
Well, you might just be able to, if you were rather long-lived. That number has about 5.4 billion digits, so if you write it out at a pace of two digits every second, without pausing to eat, sleep, or take bathroom breaks, it would take you about 85 years.
Going back to the OP, and in very simple terms …
The number of possible *atoms *is limited because of the way protons & neutrons interact. You can’t just pile more and more of them together & have them stick together. Radioactivity is the result of a bunch of them refusing to stay stuck together & then spontaneously breaking apart releasing energy as they go.
Molecules are different. They can hook together like links in a chain. One that can be practically infinitely long. So (simplifying mightily & cheating just a bit) 5000 carbons hooked end to end can be modified by hooking one more on the end to get a chain of length 5001. And that’s a different molecule. As was 4999 carbons. And 4998. And …
So when you consider there are a few dozen plausible kinds of links that all like to hook together, and you can hook them together in chains totalling hundreds of thousands of units in darn near any order, you get a stupid-large number of possible combinations.
Now add the idea that that chains can have branches, merges, loops, and tangles. It’s not like a railroad train, one long line coupled end to end. Even a simple T with 50 atoms on the stalk and 5000 across the top can be assembled 2500 different ways depending on where the stalk & top connect. etc.
It’s the ultimate tinkertoy set with an infinite supply of each of many different kinds of spools & sticks. Just keep tacking stuff onto your ever growing blob-o-sticks to make ever new and ever different blobs.
LSL,
The arbitrarily long chains you talk about, are those called polymers?
Would the properties of the 5000-carbon-atoms-molecule and 5001-carbon-atoms-molecules be liable to be different?
Do branches, merges, loops and tangles tend to have the same kind of properties across molecules?
Does each molecule have to be created and observed to know its properties or is it possible to predict with some accuracy the properties of yet-uncreated molecules by postulating the constituent atoms and their arrangement?
And then we have Photonic molecules
To take another biological example, proteins are a class of molecule which are polymers assembled from 20 possible amino acids with a typical length of around 100 - 10,000 amino acids. We really have no way of predicting the function of a protein given its amino acid sequence. Merely predicting how a single small protein folds up is a very hard problem, requiring serious computing resources, and even then our best predictions are often flawed.
That’s not even considering the myriad of chemical modifications that can be done to any single amino acid on a protein. Trying to understand a few modifications of a handful of proteins consumes half of my waking hours, and will continue to do so for a few more years…
I knew I was doing something wrong. Yes, 4^9,000,000,000. Somehow I was thinking with 9 billion base pairs you get 9 billion molecules. However, that’s the number of different ways to have a molecule of all A’s with one T mixed in there somewhere. 9 billion ways to have 8,999,999,999 A’s and one T. Another 9 billion ways to have all A’s except one G. Another 9 billion to have all A’s except one C.
Now let’s count all the ways to have all A’s except TWO T’s!
That would be all branches of chemistry except for analytical chemistry (which studies how to identify and measure substances).
Some people would say that physical chemistry is also unrelated to designing new synthetic methods, but they’re forgetting that physchem tells us which conditions will be best. “If your reaction is exothermic cool it - if it’s endothermic heat it” is physical chemistry. I once found myself surrounded by a team of Orgos who had never studied any physchem, deeming it unnecessary, and their results were amazingly poor because they didn’t know things like that. Taking one of their recipes and changing the conditions according to what physchem tells us increased yield by a factor of 500 - from “there’s barely enough dust in the vial for elementary analysis” to what articles like to describe as “beautiful, transparent yellowish crystals”.
Some properties of molecules are predictable, others not. For simpler polymers like nylon or cotton, we don’t care about the exact properties of each molecule in a jacket, just about the aggregate properties.
Missed window: If we could separate strands of cotton made of 5000 units and others made of 5001 units, both straight (no branchings), some properties other than molecular weight would be slightly different.
Add branchings and you get even more groups to separate and even more properties that would be different. The main component of “carbohydrates” is starch; the main component of “dietary fiber” is cellullose; they are built of the same monomer (glucose), but fiber has a type of branching the human body can’t break, whereas all the links in starch are digestible by humans. As my BioChem teacher said “a very important little difference!”
Chains are one type of polymers. But not all polymers are chains.
Basically, if I’m understanding it correctly, a polymer is a large molecule which is made up of repeated groups of sub-units. Think of the sub-units as legos and the polymer is what you build out of them. You can just stretch them out indefinitely in a line and make a chain. But you can also build outward in all directions. A tire, for example, is one gigantic polymer molecule.
As for “different properties”, well, that depends on what properties you’re looking at, and what differences you consider significant. Take those DNA molecules, for instance: If you compare the DNA from two different organisms, you’ll find that they have basically the same density, tensile strength, electrical permeability, water affinity, and so on over a very long list of material properties. So in that sense, no, they’re not different. On the other hand, if you put them in the nucleus of two different functioning ovum cells, and gestate them, you can get anything from a human to a sequoia tree to a paramecium, or just a dead cell. So in that sense, they’re extremely different.