The last chemical step in an Fmoc synthesis does two jobs at once, and that simultaneity is the whole problem. Trifluoroacetic acid severs the peptide–resin linkage and strips the acid-labile side-chain protecting groups in the same pot, at the same time, from the same molecule. Everything that leaves becomes a reactive electrophile dissolved in the few millilitres that also contain the chain you spent the whole synthesis assembling. A systematic comparison of scavenger mixtures put the point plainly: the majority of side reactions at this step modify susceptible residues — tryptophan, tyrosine, methionine and cysteine — with the side-chain protecting groups and linkers liberated by the acid itself [1]. The cleavage is not a workup. It is a reaction with its own kinetics, its own by-products and its own optimum, and it is routinely where an otherwise clean assembly is lost.
What is actually doing the alkylating
The tert-butyl cation that reaction schemes draw is not the only electrophile in play. Work on TFA-mediated removal of t-butyl and Boc groups from protected amino acids established that the acid converts the departing group into t-butyl trifluoroacetate, and that this ester alkylates methionine and tryptophan in trifluoroacetic acid in its own right [2]. The same study showed that isobutene reacts with TFA to regenerate the ester, with the equilibrium displaced in favour of the ester at room temperature, so the volatile alkene cannot be assumed to escape the vessel — it is recaptured as a reagent [2]. On this picture a scavenger's job is to consume the ester faster than your peptide does, and that is precisely how those authors ranked scavengers: by their reaction rates with the ester, with thiophenol showing rates comparable to the competing reaction of TFA itself with the t-butyl cations [2].
The framing makes an uncomfortable prediction, which the same work confirmed. Scavengers of sulfide structure react with t-butyl trifluoroacetate to form sulfonium compounds, and those sulfonium compounds themselves possess alkylating properties [2]. Thioanisole and dimethyl sulfide therefore do not simply destroy the electrophile; they convert it into a different one with a different selectivity. A cocktail is a managed competition between several bad outcomes, not the removal of a hazard.
Arginine contributes a chemically distinct electrophile. The aryl sulfonyl groups used on the guanidino function transfer intramolecularly to the tryptophan indole, and the extent of transfer depends on the spatial distance between the two side chains: when arginine and tryptophan are separated by a single residue the transfer of the Pmc group is most pronounced, and it cannot be completely prevented by the scavenger mixtures in current use [3]. That study also found the extent of the side reaction to depend on the identity of the intervening residue, on where tryptophan sits in the chain and on the solid support used [3]. This is the single most important fact about cleavage design: the dominant side reaction can be a property of your sequence, not of your reagent list.
The scale of it is not small. In a deliberately susceptible model, [Asn2,Trp4]dynorphin A-(1–13) synthesised with the indole unprotected, cleavage with Reagent K, with Reagent R (90% TFA / 5% thioanisole / 3% EDT / 2% anisole) and with TFA containing 20% EDT and 4% water all gave similar results: alongside the desired peptide, the crude contained 22–30% of a side product attributed to modification of tryptophan by a Pmc group [4]. The trialkylsilane-containing mixture tested in the same comparison gave the lowest recovery of the desired peptide and the highest levels of Pmc-containing peptides [4]. Reaching for a silane because it is the modern default is, for this sequence class, the wrong move.
What each component is for
Triisopropylsilane is a hindered hydrosilane long used as a cation scavenger during protecting-group removal, and it can act as a reducing agent rather than merely as a trap [5]. That reducing character is not confined to stray cations. In TFA/TIS (98/2) at 37 °C, TIS actively removes acetamidomethyl, 4-methoxybenzyl and tert-butyl protection from cysteine, with lability in the order Cys(Mob) > Cys(Acm) > Cys(But), and it promoted disulfide formation as well; thioanisole and triethylsilane were similarly effective at both [5]. If your plan depends on an Acm or Mob group surviving the global cleavage so that a second disulfide can be closed afterwards, the scavenger you chose for cation control may quietly dismantle that plan.
Water and phenol act as nucleophilic and aromatic traps for the same cations; thioanisole and 1,2-ethanedithiol act as sulfur nucleophiles, with EDT additionally holding liberated thiols in the reduced state. The classic five-component mixture assembled from all of them — 82.5% TFA, 5% phenol, 5% water, 5% thioanisole, 2.5% EDT, known as Reagent K — was the most efficient of the mixtures surveyed at suppressing a broad range of side reactions, and was validated on ten peptides of 20–50 residues [1]. The thiol-free alternative, Reagent B (TFA–phenol–water–triisopropylsilane, 88:5:5:2), was used alongside Reagent K in the synthesis of the tryptophan-rich tridecapeptide amide indolicidin, where initial purities as high as 93% were obtained immediately after cleavage [6].
The thiols are the components people most want to remove, for reasons of smell rather than chemistry. 3,6-Dioxa-1,8-octanedithiol substitutes for them: across a set of peptides cleaved with several cocktails and compared by RP-HPLC and MALDI-TOF MS, replacing the malodorous thiols with the almost odourless DODT did not degrade the purity of the end product [7]. But no scavenger is inert toward every residue. Peptides containing (2-furyl)-L-alanine are partially degraded during TFA deprotection precisely because of the dithiol scavengers: ethanedithiol and propanedithiol drive formation of a bis-dithioacetal after acidic hydrolysis of the furyl ring, and cocktails built on water with triisopropylsilane, or on DODT, were sufficient to minimise it [8]. Read that as a general warning about non-canonical side chains rather than a fact about one residue.
Choosing by sequence, not by habit
Methionine
Methionine suffers two acid-catalysed reactions during the final acidolytic cleavage, and they are different problems with different fixes: oxidation to the sulfoxide, Met(O), and S-alkylation by tert-butyl to give a sulfonium salt [9]. Quantitatively, a Met- and Cys-containing pentadecapeptide model drawn from the DsbC active site was 15–55% in the methionine sulfoxide form when cleaved with the widely used cocktails K, R and B, whereas none was detected after cleavage with Reagent H (TFA 81%, phenol 5%, thioanisole 5%, EDT 2.5%, water 3%, dimethyl sulfide 2%, ammonium iodide 1.5% w/w) [10]. Adding 1.5% w/w NH4I to K, R or B also eliminated the sulfoxide, but at a lower yield of the desired peptide than Reagent H gave [10]. A more recent pair of cleavage solutions built on TFA–anisole–trimethylsilyl chloride–Me2S–TIS with 1 mg triphenylphosphine per mL eradicated the oxidation and reduced the S-alkylation; TIS was required only for the cysteine-containing cases, and the TMSCl did not affect either cysteine or tryptophan [9]. Usefully, the sulfonium adduct is not terminal: heating the peptide at 40 °C for 24 h in 5% acetic acid reversed it to the free methionine-containing peptide [9].
Cysteine
Free cysteine thiol appears early in the cleavage and is a strong sulfur nucleophile sitting in a solution of the same t-butylating ester shown to alkylate methionine and tryptophan [2], which is why thiol scavengers survive in cysteine-rich work despite their smell: they compete for that electrophile. Where S-tert-butylation does occur, the lability ranking noted above places it at the resistant end — the least labile of the three cysteine S-protecting groups examined under TFA/TIS [5] — so it is not a species the cocktail will undo on its own. The counter-consideration is the one already noted for silanes and thioanisole: the same additives that suppress alkylation also promote disulfide formation and strip S-protection [5]. For peptides where the released thiol must stay free, work on an S-xanthenyl side-chain-anchoring strategy showed that cocktails containing 80–90% TFA plus appropriate thiols or silanes released fully deprotected free-sulfhydryl peptides in good yield from that support, while 1–5% TFA released the peptide still partially protected — two different jobs, two different acid strengths [11]. Where the disulfide is wanted immediately, adding DMSO as a mild oxidant to the cleavage cocktail achieves acidolytic release and disulfide formation in one operation [11].
The linker is a reagent
Acid-labile handles decompose under TFA into cations of their own, and they alkylate. The C-terminal amide of peptides made on Rink amide resin is alkylated by fragments of the decomposing linker, giving C-terminal N-alkylated peptide amides; including 1,3-dimethoxybenzene in the cocktail prevented their formation almost quantitatively, and in the cysteine- and methionine-containing test peptides no oxidised by-product appeared even without thiols in the mixture [12]. That is a mass shift on the C-terminus, not on a side chain, so scanning your sequence for the classic sensitive residues will not predict it — it follows from the resin you chose, a decision belonging to the resin and protecting-group guides and taken long before the cocktail is mixed.
Time and temperature buy deprotection and cost integrity
Both variables push in two directions. Acidolysis of the more resistant side-chain groups is progressive rather than instantaneous — the cysteine work above is a direct demonstration, since it recovered a lability ranking only by holding peptides in acid at elevated temperature [5] — so incomplete deprotection is cured by time, while every additional minute is also additional exposure of tryptophan, methionine and cysteine to the electrophiles already in solution [1, 2]. The trade is measurable. In a 28-residue DsbC model peptide cleaved with Reagent H, the reduced dithiol form was the major component at 51% after 3 h, whereas extending cleavage to 10 h converted the peptide to the intramolecular disulfide, isolated at 35% [10]. Nothing about that is a failure of the cocktail; it is the cocktail continuing to work.
Temperature is the same bargain, and the authors of that cysteine study close by naming scavenger identity, scavenger concentration, reaction temperature and reaction time as the four parameters deserving careful consideration [5]. Heat is also a repair tool rather than only a hazard, as the acetic-acid reversal of the methionine sulfonium salt shows [9]. The practical consequence is that a cleavage time copied from a protocol written for a different sequence carries no authority. If you have the material, run a short time course and take a crude analytical trace at two or three points.
Scale, workup and the losses that never reach the chromatogram
Cleavage cocktails are dosed per gram of resin rather than as a fixed volume, and the reason follows from the mechanism above rather than from any measurement quoted here: the alkylating species are generated from the peptide-resin itself, so the volume sets their concentration and therefore their rate of reaction with the peptide. Throughput scales accordingly. An automated deprotection and cleavage workstation processing 50–500 mg samples of peptide resin on a two-hour cycle, using continuous aqueous–ether extraction to strip scavengers and reaction by-products, returned crude peptides free of residual scavengers in 50–80% yield, and the method covered all but the most hydrophobic peptides [13]. At preparative scale in a single tryptophan-rich synthesis, 0.8 g of HPLC-purified indolicidin at >99% purity corresponded to a 39% overall yield based on the C-terminal Arg(Pmc) anchored to the PAL–PS resin [6].
The precipitation solvent is a chemical choice, not a convenience. Methyl tert-butyl ether and diethyl ether are widely treated as interchangeable for the cold-ether workup, but precipitating peptides from strong acid solution with MTBE gives rise to t-butyl alkylation by-products readily detectable by MALDI-TOF MS, and the problem can reach undesirable proportions when the peptide-resin contains a high proportion of aromatic residues — conspicuously so in peptide nucleic acid syntheses — so diethyl ether is the advisable workup in those cases, leading consistently to cleaner products [14]. If you precipitate into MTBE and then find t-butyl adducts, the workup is a live suspect alongside the cleavage.
Some losses never appear as a peak at all. In the solid-phase synthesis of two cyclic cystinyl octapeptides, recovery of peptidic material from the resin was the limiting difficulty, and the cause was irreversible reattachment of already-cleaved peptide to the support through alkylation of various side-chain functions — which those authors could not avoid even using the most efficient scavengers or their cocktails [15]. A crude that is clean by peak area but poor by recovered mass is telling you two different things, and only one of them is on the chromatogram. Dissolution behaviour compounds this: for aggregation-prone fragments of amylin, cleavage with trifluoroacetic acid, phenol, ethanedithiol, thioanisole and water proved optimal, and hexafluoroisopropanol was the best solvent for getting the crude peptides into solution at all [16]. Material that will not dissolve is not material you have quantified.
What the crude trace does and does not license
A clean analytical trace immediately after cleavage licenses one conclusion: under those cleavage and workup conditions, the assembled chain gave a product distribution dominated by one species eluting where you expected. It does not establish that the species is the target. The characteristic cleavage by-products are alkylations, and an alkylated peptide differs from its parent by a small hydrophobic increment that a gradient tuned for the target may or may not resolve; a Pmc adduct, an N-alkylated C-terminal amide and a sulfonium salt each have a defined mass, so mass confirmation is what separates them from the parent. Nor does the trace distinguish a side reaction that happened during chain assembly from one that happened in the cleavage vessel — deletions, aspartimide and racemisation present differently, and belong to the synthesis-impurity guide.
The discipline that follows is unglamorous. Decide the cocktail when you decide the protecting groups, not on the day of cleavage: an unprotected indole a residue or two from an arginine commits you to a problem no scavenger fully solves [3], and choosing protection that never generates the transferable sulfonyl electrophile is a planning decision, not a cocktail decision. Cleave a small analytical sample and look at it before committing the batch. And when a crude looks worse than the on-resin monitoring predicted, treat the hours in acid as the prime suspect rather than as the step where nothing happened.