Linear and Convergent Synthesis
CSIR NET organic synthesis: linear vs convergent routes, longest linear sequence, overall yield, material calculations, divergent synthesis, atom economy.
- 15 min
On this page · 14 sections
Prerequisites
- Retrosynthesis and the Disconnection Approach: “disconnect near the middle” is where convergence starts.
- Percentages, Ratios and Averages: overall yield is a product of fractions, the same arithmetic as successive percentage change.
- Olefination, Redox, Aromatic and Coupling Named Reactions: the fragment-coupling reactions (Wittig, HWE, Suzuki) that make convergence possible.
The Hook: Bricks or Prefabricated Walls
Definitions
- Linear synthesis: the target is built by a single sequence of steps, each acting on the product of the previous one (A → B → C → … → TM). Small reagents may be added at each step, but there is one main chain.
- Convergent synthesis: two or more major fragments are prepared independently (in parallel) and joined at a late stage. Each fragment has its own sequence; the branches converge on the target.
- Longest linear sequence (LLS): the largest number of consecutive steps from any starting material to the target. For a linear synthesis it equals the total number of steps; for a convergent synthesis it is shorter than the total.
- Total step count: all steps in all branches. It measures the labour; the LLS measures the time and yield.
- Divergent synthesis: the reverse idea: one common intermediate is elaborated into many different targets (analogue series, natural-product families, compound libraries).
graph LR
subgraph Linear
A1["A"] --> B1["AB"] --> C1["ABC"] --> D1["ABCD"] --> E1["ABCDE"] --> F1["ABCDEF"] --> G1["ABCDEFG"] --> H1["ABCDEFGH"]
endgraph LR
A["A"] --> AB["AB"]
B["B"] --> AB
C["C"] --> CD["CD"]
D["D"] --> CD
E["E"] --> EF["EF"]
F["F"] --> EF
G["G"] --> GH["GH"]
H["H"] --> GH
AB --> ABCD["ABCD"]
CD --> ABCD
EF --> EFGH["EFGH"]
GH --> EFGH
ABCD --> T["ABCDEFGH"]
EFGH --> TBoth routes join eight fragments in seven coupling steps. The linear route has an LLS of 7; the convergent route an LLS of 3.
Yield Arithmetic
For a sequence of n steps with fractional yields \( y_1, y_2, \dots, y_n \):
$$Y_{\text{overall}} = y_1 \times y_2 \times \dots \times y_n = \prod_{i=1}^{n} y_i$$If every step has the same yield y: \( Y = y^n \).
For a convergent synthesis, the overall yield with respect to a given starting material is the product of the yields along that material’s path to the target. The figure normally quoted is the yield along the longest linear sequence.
| Steps n | y = 95% | y = 90% | y = 80% |
|---|---|---|---|
| 5 | 77.4% | 59.0% | 32.8% |
| 10 | 59.9% | 34.9% | 10.7% |
| 15 | 46.3% | 20.6% | 3.5% |
| 20 | 35.8% | 12.2% | 1.2% |
Three lessons:
- Yields multiply, not add. Ten steps at 90% give 35%, not 90 − 10 × (something).
- Length is expensive. At 80% per step, a 20-step linear synthesis delivers about 1% of the material.
- Shortening the LLS helps more than improving one step. Raising one step of a 10-step, 90% route to 99% raises the overall yield from 34.9% to 38.4%; cutting the LLS from 10 to 6 at 90% raises it to 53.1%.
The eight-fragment example in numbers
| Strategy | LLS | Overall yield (90% per step) | Moles of fragment A needed per mole of target |
|---|---|---|---|
| Linear | 7 | \( 0.9^7 = 47.8\% \) | \( 1/0.478 = 2.09 \) |
| Convergent | 3 | \( 0.9^3 = 72.9\% \) | \( 1/0.729 = 1.37 \) |
Material requirement
Linear versus Convergent
| Feature | Linear | Convergent |
|---|---|---|
| Overall yield | lower (\( y^n \), n = total steps) | higher (governed by the shorter LLS) |
| Time | every step in series | branches can run in parallel |
| Losses of valuable material | late failures waste everything made so far | each branch carries only its own investment |
| Scale | early steps must be run on a large scale | early material in each branch is smaller |
| Planning | simple; each step follows from the last | needs a reliable late-stage coupling between two complex fragments |
| Risk | many steps, each can fail | the fragment coupling can be low-yielding, sterically hindered or unselective |
| Typical use | short targets; iterative syntheses on solid support (peptides, oligonucleotides) | large targets: macrolides, polyketides, peptides by fragment condensation, dendrimers |
Good fragment-coupling reactions are high-yielding, chemoselective (tolerant of other functional groups) and stereochemically predictable: amide and ester formation, Pd-catalysed cross-couplings (Suzuki, Stille, Sonogashira, Negishi), Wittig and HWE olefinations, Julia olefination, aldol reactions, Diels–Alder cycloadditions, alkylation of stabilised anions, glycosylation, and olefin metathesis.
Examples from Biopolymer Synthesis
Peptides
- Stepwise solid-phase synthesis (Merrifield) is linear: one protected amino acid is added per cycle (deprotect, couple, wash). Because the chain stays on a resin, excess reagents can be used and washed away, so each coupling can be pushed very close to completion. Even so, small per-cycle losses compound over long sequences.
- Fragment condensation is convergent: protected peptide segments are made separately and coupled. The risk is racemisation at the activated C-terminal residue of the acyl fragment (via an oxazolone), which is why segments are often chosen to end in glycine (achiral) or proline. Native chemical ligation joins unprotected segments chemoselectively (C-terminal thioester + N-terminal cysteine).
Oligonucleotides
Automated phosphoramidite synthesis is linear, so coupling efficiency dominates. A 20-mer needs 19 couplings:
| Coupling efficiency | Full-length 20-mer (\( y^{19} \)) |
|---|---|
| 99% | 82.6% |
| 98% | 68.1% |
A 1% drop in coupling efficiency costs about 14 percentage points of full-length product.
Dendrimers
- Divergent dendrimer growth starts from a core and adds generations outward; the number of reactions per molecule doubles each generation, and defects accumulate.
- Convergent growth builds the branches (dendrons) first and attaches them to the core at the end: fewer reactions per step and easier purification, at the cost of slower final couplings for large dendrons. See Polymer Chemistry for polymer architectures.
Other Measures of Efficiency
Atom economy
It asks how much of the mass of the reactants ends up in the product, independent of yield.
| Reaction | Atom economy |
|---|---|
| Diels–Alder, catalytic hydrogenation, many rearrangements | 100% (all atoms incorporated) |
| CH₃CH₂Br + NaOH → CH₃CH₂OH + NaBr | 46.07/(108.97 + 40.00) = 30.9% |
| Cyclohexanone + Ph₃P=CH₂ → methylenecyclohexane + Ph₃P=O | 96.17/(98.15 + 276.32) = 25.7% |
Addition reactions and catalytic processes score high; substitutions and eliminations lose the leaving group; stoichiometric reagents with heavy by-products (Ph₃PO, Bu₃SnX) score low. Protecting groups always lower atom economy, because they are added and removed.
Step, redox and protecting-group economy
- Step economy: fewer steps (especially a shorter LLS) usually matters more than any single yield. Cascade (tandem, domino) reactions form several bonds in one operation; the Robinson annulation (Michael + aldol) is the classic textbook example.
- Redox economy: avoid oxidation-level changes that are later reversed (oxidise, then reduce back). Plan so that each step moves the oxidation level towards that of the target.
- Protecting-group economy: every protection costs two steps (on and off). Chemoselective reagents, a change in the order of events, or functional groups introduced late can remove the need. See Protecting Groups.
- One-pot sequences (several steps without isolating intermediates) reduce losses on work-up and purification.
Memory Tricks
Don’t Confuse
| This | Is not this |
|---|---|
| Linear: one main chain; LLS = total steps | Convergent: parallel branches; LLS < total steps |
| Convergent (many pieces → one target) | Divergent (one intermediate → many targets) |
| Longest linear sequence | Total step count |
| Overall yield: product of step yields | Average yield: arithmetic mean of step yields (not used to compute output) |
| Yield: fraction of theoretical product obtained | Atom economy: fraction of reactant mass that can end up in the product |
| Stepwise SPPS (linear, on resin) | Fragment condensation (convergent; racemisation risk at the activated segment) |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not taken from any CSIR NET paper.
Level 1: Foundation
- A linear synthesis has 5 steps, each 90%. What is the overall yield?
- Sixteen fragments are joined pairwise in a fully convergent manner. What is the longest linear sequence of coupling steps?
- What is the atom economy of a Diels–Alder reaction?
Answers
- \( 0.9^5 = 0.590 \): 59.0%.
- 4 (16 → 8 → 4 → 2 → 1). The total number of couplings is still 15.
- 100%: every atom of diene and dienophile ends up in the cycloadduct.
Level 2: Part B Style
Practice Q1. A target can be made by a 10-step linear route, or by preparing two fragments in 5 steps each and coupling them. All steps, including the coupling, proceed in 85% yield. The overall yields (based on the longest linear sequence) are approximately
(a) 19.7% (linear) and 37.7% (convergent) (b) 19.7% and 44.4% (c) 15.0% and 37.7% (d) 37.7% and 19.7%
Answer: (a)
Linear: \( 0.85^{10} = 0.197 \). Convergent: LLS = 5 + 1 = 6 steps, \( 0.85^6 = 0.377 \). (b) forgets the coupling step (\( 0.85^5 = 0.444 \)).
Practice Q2. Which statement about convergent synthesis is incorrect?
(a) Its longest linear sequence is shorter than its total step count. (b) Fragments can be synthesised in parallel. (c) It reduces the total number of coupling steps needed to join n fragments below n − 1. (d) It depends on an efficient late-stage fragment coupling.
Answer: (c)
Joining n fragments always needs n − 1 bond-forming couplings. Convergence reduces the number of steps in series (LLS), not the number of couplings.
Practice Q3. The atom economy for converting bromoethane into ethanol with NaOH (CH₃CH₂Br + NaOH → CH₃CH₂OH + NaBr) is closest to (M: C₂H₅Br 108.97, NaOH 40.00, C₂H₅OH 46.07)
(a) 100% (b) 42% (c) 31% (d) 23%
Answer: (c)
46.07/(108.97 + 40.00) = 46.07/148.97 = 0.309, i.e. 30.9%. The NaBr by-product carries most of the mass.
Practice Q4. In automated oligonucleotide synthesis, the coupling efficiency is 98% per cycle. The fraction of full-length 20-mer (19 couplings) is closest to
(a) 98% (b) 82% (c) 68% (d) 60%
Answer: (c)
\( 0.98^{19} = 0.681 \). (b) corresponds to 99% coupling efficiency (\( 0.99^{19} = 0.826 \)).
Level 3: Part C Style
Practice Q5. Fragment A is made in 3 steps (90%, 80%, 95%) and fragment B in 4 steps (90% each). A and B are coupled in 70% yield and a final deprotection proceeds in 95% yield.
Which statements are correct?
- The longest linear sequence is 6 steps.
- The overall yield based on B’s starting material is about 43.6%.
- To make 0.100 mol of target, about 0.22 mol of A’s starting material is needed.
- The overall yield is 0.90 × 0.80 × 0.95 × 0.90⁴ × 0.70 × 0.95 ≈ 29.8%.
(a) 1, 2 and 3 only (b) 1 and 2 only (c) 3 and 4 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 3 only
- Correct. 4 (branch B) + coupling + deprotection = 6.
- Correct. \( 0.9^4 \times 0.70 \times 0.95 = 0.6561 \times 0.665 = 0.436 \).
- Correct. Yield along A: 0.90 × 0.80 × 0.95 × 0.70 × 0.95 = 0.455; 0.100/0.455 = 0.220 mol.
- Wrong. That product multiplies across both branches, which treats the route as linear. The two branches run in parallel; each starting material sees only its own branch.
Practice Q6. A chemist needs 10.0 g of a target (M = 300 g mol⁻¹). Route P is linear, 6 steps at 80% each, from a starting material of M = 150 g mol⁻¹.
Which statements are correct?
- The overall yield of route P is about 26.2%.
- About 19.1 g of starting material is needed.
- If the first two steps of route P are replaced by a single step of 80%, the starting material needed falls to about 15.3 g (same starting material).
- If every step is improved to 90%, about 6.3 g of starting material is needed.
(a) 1, 2 and 3 only (b) 1, 2 and 4 only (c) 2 and 3 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 3 only
- Correct. \( 0.8^6 = 0.262 \).
- Correct. 0.0333 mol / 0.262 = 0.127 mol × 150 g mol⁻¹ = 19.1 g.
- Correct. Five steps: \( 0.8^5 = 0.328 \); 0.0333/0.328 = 0.102 mol × 150 = 15.3 g.
- Wrong. \( 0.9^6 = 0.531 \); 0.0333/0.531 = 0.0627 mol × 150 = 9.4 g, not 6.3 g.
Practice Q7. Which statements about synthetic strategy are correct?
- In convergent peptide fragment condensation, coupling a segment that ends in glycine avoids racemisation at the activated C-terminus.
- Divergent synthesis is used to prepare many analogues from one advanced intermediate.
- A route with fewer steps always has a higher atom economy.
- Introducing a protecting group adds two steps to the sequence and lowers atom economy.
(a) 1, 2 and 4 only (b) 1 and 2 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 4 only
- Correct. Glycine has no stereocentre; activated C-terminal residues of other amino acids can racemise via an oxazolone.
- Correct. That is the definition of divergent synthesis.
- Wrong. Atom economy depends on the balanced equations, not the number of steps; one Wittig step has a lower atom economy than two addition steps.
- Correct. Protection and deprotection are two extra steps, and the protecting group’s atoms are discarded.
Quick Revision Table
| Idea | Fact |
|---|---|
| Linear | one chain; LLS = total steps; \( Y = y^n \) |
| Convergent | parallel fragments, late coupling; Y set by the LLS |
| Divergent | one intermediate → many targets |
| n fragments | n − 1 couplings; fully convergent LLS = log₂ n couplings (n a power of 2) |
| 8 fragments at 90% | linear 47.8%; convergent 72.9% |
| Material | mol SM = mol target / Y; convert mass to moles first |
| 90% per step | 5 → 59%, 10 → 35%, 20 → 12% |
| Oligonucleotide 20-mer | 99% → 82.6%; 98% → 68.1% |
| Atom economy | M(product)/ΣM(reactants); addition 100%, Wittig low |
| Economies | step, redox, protecting-group; cascades and one-pot |
| Peptides | SPPS linear; fragment condensation convergent (racemisation risk) |
Key Takeaways
What’s Next
- Protecting Groups: the price of chemoselectivity, and how orthogonal sets make long syntheses possible.
- Retrosynthesis and the Disconnection Approach: where a convergent plan is designed.
- Protein Structure and Stability: the peptide bond that solid-phase synthesis builds.
- Back to the chapter overview.