Open a packet of aspirin and read the name of the active ingredient: acetylsalicylic acid. The bark of the willow has contained a relative of it for as long as there have been willows, and people chewed the bark for pain long before anyone drew a molecule. The tablet in your hand is something else. It was assembled, in a plant, from cheap starting materials, to a purity that bark could never offer, and it costs less than the paper in the packet.
The same is true of the banana flavour in a sweet, the nylon in a rucksack strap, the dye in a T-shirt and most of the medicines you have ever swallowed. Each is a particular arrangement of atoms that a person decided to make. A chemist who has taken a course in organic synthesis looks at any of them and does not ask "what is it?" but "what would I snap it together from?" That reflex, and the bookkeeping that goes with it, is what this piece is about.
- Who takes it
- Chemistry majors, plus students heading for pharmaceutical science, medicinal chemistry and chemical engineering. Programmes that follow the American Chemical Society's guidelines or the Royal Society of Chemistry's accreditation treat organic chemistry as a foundation; the synthesis-focused second course typically comes in year 3 (at MIT it sits in Course 5, after the first organic course).
- What it is for
- Planning how to build a specific molecule from simple, available pieces, and judging whether the plan is any good.
- Comes after
- Organic Chemistry: Structure & Mechanism (35/38).
- Leads to
- Biochemistry (35/43), where enzymes are the most elegant synthetic chemists there are, and Materials Science and Manufacturing (34/13), where polymers are synthesis at scale.
- What this piece gives you
- You will not be able to plan a drug synthesis. You will see how a chemist thinks about one: backwards, with a budget, and with a conscience about waste. It describes ideas and benign examples only, never procedures. For the basics of the curved-arrow logic underneath, start with 35/38.
Work backwards: retrosynthesis
If you must build a house you do not start by hauling bricks and hoping. You start from the finished house and ask what the last thing you would do is. Put the roof on. Before that? Walls. Organic chemists plan the same way, and the habit has a name: retrosynthetic analysis. E. J. Corey of Harvard developed it into a formal method and received the 1990 Nobel Prize in Chemistry for the theory and methodology of organic synthesis.
The target molecule is drawn, and then you disconnect a bond: imagine breaking it, and ask what two simpler pieces you would need to put it back. Each piece is simpler, so you disconnect again, until every piece is something you can buy. Then you turn the whole plan around and run it forward. The rule a graduate remembers is the choice of which bond to cut:
Worked example. Aspirin contains an ester, a carbonyl carbon joined to an oxygen that is joined to a ring. Cut the bond between the carbonyl carbon and that oxygen and you get a ring carrying an OH and a CO₂H group (salicylic acid) plus an acetyl group (the CH₃C=O unit, supplied in practice by acetic anhydride). Both pieces are cheap. Esters are the single most obliging bond in organic chemistry: cut an ester and the plan usually writes itself. Try it.
The fragments are drawn as groups, not full structures. Each bond button is a possible disconnection; the readout says what you would need if you cut there.
A modern twist: software now proposes routes by searching enormous reaction databases, and some of it uses machine learning. The chemist still decides, because "reliable" depends on the other groups present and on what is actually in the stockroom.
Every step pays a tax
Textbook reactions are drawn with a clean arrow. Real ones return less than 100% of what the arithmetic promises: some of the material goes down a side road, some sticks to the glassware, some is lost in cleaning it up. Call the fraction that survives a step its yield. Steps in a row multiply:
That is the quiet brutality of long syntheses. Ten steps at 95% still lose 40%. The fix a planner reaches for is convergence: instead of one long chain, build two halves separately and join them near the end. What counts is the longest run of steps any one batch of material goes through. Five steps on each half plus one joining step has a longest sequence of 6, so at 90% a step you keep 0.9⁶ = 53.1%, against 34.9% for ten in a line, even though you carried out eleven reactions in all. Slide the numbers and see.
Convergent route: two branches of about half the steps each (rounded up), plus one joining step; the longest sequence is that branch plus the join. Every step is assumed to have the same yield; real ones do not.
This is why the steps you can skip matter as much as the ones you add, and it sets up the next idea.
Selectivity and protecting groups
Real molecules carry several reactive groups, and a reagent does not read your plan. A chemist needs three kinds of control. Chemoselectivity: of two different groups, which reacts? Regioselectivity: of two similar positions, which one? Stereoselectivity: of two mirror-image or otherwise different 3D outcomes, which one? Much of a synthesis course is learning to predict, and then to engineer, these choices.
Often the molecule already helps. Paracetamol (acetaminophen) is 4-acetamidophenol. Start from 4-aminophenol, which has both an NH₂ and an OH, and a mild acetylating agent goes for the nitrogen, because the NH₂ is the more nucleophilic of the two. Nobody has to hide anything. But sometimes nothing in the molecule tells two groups apart, and then you do it by force: cover one group with a temporary cap, a protecting group, do the reaction you want elsewhere, then take the cap off.
A sugar such as glucose is the classic case: five OH groups of nearly the same character. To change just one, the traditional route is to cap the others. The cost shows up on the yield slider, which is why chemists admire a route that is "protecting-group-free". Predict first, then see.
Carbon–carbon bonds: building the skeleton
Almost every organic molecule is a carbon skeleton with groups hung on it. Attaching groups is comparatively easy. Making new carbon–carbon bonds is the hard, valuable part, because it changes the skeleton itself. The history of synthesis is partly a list of reactions that learned to do it, and many of them won Nobel prizes: the Grignard reaction (Victor Grignard, 1912), the Diels–Alder reaction that closes a ring in one move (Otto Diels and Kurt Alder, 1950), the Wittig reaction (Georg Wittig, 1979), olefin metathesis (Chauvin, Grubbs and Schrock, 2005).
The most widely used modern family is cross-coupling, recognised in 2010 for Richard Heck, Ei-ichi Negishi and Akira Suzuki. The problem it solves: two ordinary carbon pieces both want to be left alone. A palladium atom introduces them. It is not used up. It runs a three-step loop and starts over:
One palladium atom can make a huge number of bonds, so the metal is added in small amounts. A "1 mol%" loading, one palladium per hundred molecules of reactant, means at most 100 turnovers per atom even at perfect yield. Cut the loading to 0.01 mol% and that becomes 10,000. Palladium is expensive and the product must be cleaned of it, so every extra turnover matters. Step through one loop.
Catalysis: lowering the hill, not the destination
A catalyst gives the reaction a lower route between the same start and finish. It does not change how far the reaction goes at equilibrium, and it comes out unchanged. The speed-up is exponential in how far the hill drops, because reaction rates follow the Arrhenius form e−Ea/RT:
At 25 °C (298 K), RT is about 2.48 kJ/mol. A catalyst that lowers the barrier by 20 kJ/mol, about the strength of a weak hydrogen bond or a fraction of one C–C bond, speeds the reaction up roughly 3,200 times. Forty kJ/mol is a factor of about ten million.
Illustrative barrier of 100 kJ/mol uncatalysed; the pre-exponential factor is assumed unchanged. R = 8.314 J/mol·K (CODATA).
Catalysts are also how chemists get handedness. Many molecules come in mirror-image pairs that react identically with ordinary reagents and differently in a body: the two forms of carvone smell of spearmint and of caraway. In 2001 the Nobel Prize went to William Knowles and Ryōji Noyori for catalysts that make one hand of a molecule in hydrogenation, and to Barry Sharpless for the same in oxidation. In 2021 Benjamin List and David MacMillan were recognised for organocatalysis: using small, metal-free organic molecules, such as the amino acid proline, as handed catalysts. That is the same job that enzymes do, with a molecule you can hold in your hand. And in 2022 Carolyn Bertozzi, Morten Meldal and Sharpless (his second prize) were recognised for click and bioorthogonal chemistry: reactions so selective they run inside living cells without disturbing them.
Green chemistry: atom economy and the E-factor
Plans were once judged almost only on yield. Atom economy, proposed by Barry Trost in 1991, asks a different question: of all the atoms you put in, what fraction end up in the product? A reaction can have a 100% yield and still throw most of its atoms away.
Making aspirin from salicylic acid (138.12) and acetic anhydride (102.09) gives aspirin (180.16) plus acetic acid, which is a by-product: 180.16 ÷ 240.21 = 75.0%. A Diels–Alder ring closure has 100%, because the two pieces simply join. Roger Sheldon's E-factor counts the other side: kilograms of waste per kilogram of product, solvents and all. Sheldon's survey found it grows as you move from oil refining and bulk chemicals (little waste per kilogram) to fine chemicals and pharmaceuticals, where waste per kilogram of product is often tens of kilograms and sometimes more than a hundred, because there are more steps and each carries solvents and clean-up.
The poster child is ibuprofen. The original route, used at first by the Boots company, took six steps. A later three-step route developed in the 1980s by the BHC Company, a joint venture of Boots and Hoechst Celanese, throws away far less: its atom economy is about 77%, against about 40% for the original. Anastas and Warner's twelve principles of green chemistry (1998) turned this kind of thinking into a checklist. Chemistry's environmental bill is real: waste, solvents, energy, toxic by-products, and the history of the field includes Fritz Haber, who gave the world fertiliser and also led the German chemical-weapons effort. Green chemistry does not ask for guilt. It asks that waste be counted like cost.
Masses are computed from IUPAC standard atomic weights (rounded). The ibuprofen figures are published approximate atom economies, not computed here. The solvent slider is a made-up input so you can see the E-factor move; it is not a measured value for any process.
What a graduate just sees
- Any molecule is a puzzle to disconnect, not a thing to stare at. The first question is "which bond would I make last?"
- An ester, an amide or a ring with a double bond hints at where the seams are.
- Yields multiply. A long chain at 90% a step is a bad bet, so look for a convergent route.
- Selectivity is the whole game: which group, which position, which hand.
- A protecting group costs two steps. The best route often has none.
- Making a carbon–carbon bond is the hard step, and a catalyst is how it is done today.
- A catalyst lowers the barrier, never moves the equilibrium, and a small drop is an exponential speed-up.
- Yield is not the same as atom economy, and neither counts the solvent.
Where it shows up next
- Biochemistry (35/43): enzymes are catalysts that choose one hand and one position among thousands, in water, at body temperature. Chemists borrow the idea with organocatalysts.
- Materials Science and Manufacturing (34/13): a polymer is a synthesis repeated thousands of times in one chain, and nylon is two bond-makings away from its starting chemicals.
- In the world: the pharmaceutical industry, flavours and fragrances, agrochemicals, display materials, and every process-chemistry job where "can we make a kilogram of this?" becomes "can we do it in four steps with less solvent?"
- Easier first read: the library's Food & Cooking (category 14) shows the banana-flavour ester on a plate. For the same reaction ideas inside a body, see Biochemistry (35/43).
If you want to go deeper
- 📘 Textbook. Jonathan Clayden, Nick Greeves and Stuart Warren, Organic Chemistry (Oxford University Press, 2nd edition 2012; check whether a newer edition is out). Widely assigned, and unusually good at showing why reactions happen.
- 📘 Textbook. Francis Carey and Richard Sundberg, Advanced Organic Chemistry, Part B: Reactions and Synthesis (Springer). The standard step up for synthesis planning.
- 📘 Gentler. David Klein, Organic Chemistry as a Second Language (Wiley). A study companion that trains the pattern-spotting this subject needs.
- 🌐 Free course. MIT OpenCourseWare 5.13 Organic Chemistry II (Fall 2003), whose description includes methods for the synthesis of organic compounds, and 5.12 Organic Chemistry I (Spring 2005) before it.
Organic synthesis is planning backwards. Cut the bond a reliable reaction can remake, repeat until the pieces are cheap, then count the cost: yields multiply, protecting groups add steps, and catalysts make the hard bonds and make them selectively. Good plans are short, convergent and light on waste.
Try this: read the ingredients on a sweet or soft drink and find an ester (names ending "-yl … -ate", such as "isoamyl acetate"). Then ask: which two simpler things would I join to make it? You are doing retrosynthesis. Nothing to mix or heat.
Sources. Nobel Prize in Chemistry years and citations (Grignard 1912, Diels and Alder 1950, Woodward 1965, Wittig 1979, Corey 1990, Knowles, Noyori and Sharpless 2001, Chauvin, Grubbs and Schrock 2005, Heck, Negishi and Suzuki 2010, List and MacMillan 2021, Bertozzi, Meldal and Sharpless 2022) from the Wikipedia list of Nobel laureates in chemistry, which follows nobelprize.org. Ibuprofen route steps and waste figures from Wikipedia's Ibuprofen article. Atom economy: B. M. Trost, Science 254 (1991). E-factor: R. A. Sheldon, Green Chemistry 9 (2007), "The E Factor: fifteen years on". Twelve principles: Anastas and Warner, Green Chemistry: Theory and Practice (Oxford, 1998). Atomic weights: IUPAC. All numbers in the interactives were computed and checked in node. This piece describes how chemistry works, not how to make anything.
- 🌐 MIT OCW 5.13 Organic Chemistry II
- 📘 Clayden, Greeves & Warren, Organic Chemistry
- 📘 Klein, Organic Chemistry as a Second Language