Why water is the prime suspect
In reversed-phase HPLC, water is typically 50–95% of the mobile phase. Most of what enters the column is the water you used to prepare it. At 50 ppb TOC, every litre of mobile phase carries 50 µg of contamination — enough to leave a visible trace on a UV detector.
Symptom 1: Baseline drift
Looks like: baseline climbing or sagging through a gradient, with slow undulation even isocratically.
Mechanism: organic contaminants in the water elute progressively during the gradient, shifting background absorbance at the detector. Higher TOC, more drift.
Test: prepare a fresh mobile phase from a newly opened bottle of HPLC-grade water and run the same gradient. If drift drops markedly, the problem is water, not column or method.
Symptom 2: Ghost peaks
Looks like: peaks appear on blank injections, at inconsistent positions, changing with each mobile-phase batch.
Mechanism: plasticisers, detergent residues and microbial metabolites from water or containers retain on the column and elute unpredictably. Long-term storage in plastic carboys is a common source.
Test: run a blank gradient. If peaks persist, substitute items one at a time: water, organic phase, container, vial septa.
Symptom 3: Retention-time drift
Looks like: retention times shifting day to day, poor reproducibility.
Mechanism: metal ions in the water — iron and copper especially — interact with the silica matrix and alter column performance. A system whose resistivity only looks good right after a cartridge change may still carry elevated metal background.
Symptom 4: Abnormally short column life
Looks like: steadily rising backpressure and rapid loss of efficiency.
Mechanism: particles and bacteria plug the frits. Bacteria can keep multiplying in the mobile phase — phosphate-buffered aqueous phases are ideal growth media.
What chromatography water should meet
- Resistivity: ≥18.2 MΩ·cm at 25 °C, sustained
- TOC: <5 ppb (aim <3 ppb for trace work)
- Bacteria: <1 CFU/ml
- Particles (>0.22 µm): <1/ml
- Reactive silica: <0.1 ppb (critical for MS — silica deposits on the ion source)
Three things easily missed
- UV lamps age. Rated around 8,000 hours; as output falls, TOC creeps up while resistivity still looks perfect. Replace on schedule, not on failure.
- Dispensing practice matters more than the instrument. Open vessels, detergent-washed carboys, touching the bottle mouth — any of these ruins good water. Use dedicated containers and draw fresh.
- Degas mobile phases. Dissolved oxygen causes baseline noise, particularly at low detection wavelengths.
A pragmatic troubleshooting order
When the baseline misbehaves, work from cheapest to most expensive: water → container → organic phase → column → detector. Putting water first often costs far less than replacing a column.
