Fragmentation / deletion Δm loss of dipeptide severity: high

DKP (2,5-diketopiperazine) formation

After residue 2 is loaded on resin and Fmoc-removed, the free α-amine cyclizes onto the C-terminal ester (the resin-anchor) — forms a 6-membered diketopiperazine that drops off the resin. Net: the entire dipeptide vanishes.

Affected residue(s): any

Neighbour(s) that trigger it: P (esp. as residue 2) G Sar N-Me-Xxx

Why it happens (mechanism)

Liberated H-Xxx-Yyy-O-resin; the α-amine of Xxx attacks the ester carbonyl tethering Yyy to the resin, in cis-amide geometry, forming a 6-membered cyclo(Xxx-Yyy) which is no longer attached. Especially favorable when Yyy = Pro (already locks cis-amide) or when either residue is N-methylated.

When it strikes (triggers)

Any time after Fmoc removal of residue 2 in SPPS. Prolonged exposure to base before coupling residue 3. Wang/HMPA resins (ester linker, more leaving-group-able). Aggravated by: H-Pro-Yyy- or H-Sar-Yyy- at the C-terminus.

How to spot it (MS signature)

Not a +Δm — the affected peptide is a truncation by 2 residues (and a separate small DKP product is detectable). Spot it by missing the dipeptide in HPLC + lower yield, plus a peak at the cyclo-dipeptide mass.

How to prevent it

Use CTC (2-chlorotrityl) resin, not Wang/HMPA — bulky linker disfavors the cyclization.
If a Wang-type resin is required, couple residue 3 *immediately* after Fmoc-removal of residue 2. Don't store the deprotected dipeptidyl-resin overnight.
Avoid cis-amide-forming pairs at residues 1-2: -Xxx-Pro-resin is the classic offender. If unavoidable, switch to CTC resin or use a hindered C-terminal protecting group.
Reduce base concentration during deprotection (2% DBU + 2% piperidine for 1 min, not 20% × 10 min).

If it already happened (salvage)

Once detached, the dipeptide is gone. The resin loading is now -2 residues short. Re-load and restart, or accept the reduced length.

Source

Yi Yang, Side Reactions in Peptide Synthesis (Elsevier, 2016), Chapter 1, §1.10.