How to Develop a HILIC Method: Manufacturer Guide

A successful HILIC method is not simply a reversed-phase gradient run backwards. Retention depends on a water-enriched layer at the stationary phase, but hydrogen bonding and electrostatic interactions can be equally important. Consequently, an amide, bare-silica and zwitterionic HILIC column may give completely different selectivity for the same polar analytes. This guide converts the official recommendations of Waters, Agilent, Merck/SeQuant, Thermo Fisher Scientific and Phenomenex into a practical method-development workflow.

The most useful starting rule: select the stationary-phase chemistry before optimizing the gradient. Begin around 85–90% acetonitrile, use a volatile 5–20 mM ammonium buffer when LC-MS compatibility is required, dissolve the sample in the initial mobile phase whenever possible, inject a small volume and allow substantially more equilibration than in reversed-phase LC.

When HILIC is the right separation mode

HILIC is particularly valuable when a compound is poorly retained on C18 even with a highly aqueous mobile phase. Typical candidates include amino acids, nucleosides and nucleotides, small organic acids, sugar phosphates, glycans, polar pharmaceuticals, ionic metabolites and other hydrophilic substances.

According to the Thermo Fisher Scientific HILIC overview, the mobile phase normally contains more than 60–70% organic solvent, usually acetonitrile. The small aqueous fraction forms a water-rich layer on the polar stationary phase. Analytes partition into this layer and may additionally interact through hydrogen bonding, adsorption and ion exchange.

Do not assume that higher polarity always means stronger retention. The charge of the analyte and the charge of the stationary phase may reinforce, weaken or even reverse the expected retention order. Record analyte pK_a, expected charge at the working pH and stationary-phase chemistry before screening conditions.

Choose the column chemistry first

Column family	Stationary phase	Useful first application	Manufacturer-specific starting advice	Important limitation
Waters Atlantis Premier BEH Z-HILIC	Zwitterionic phase on hybrid particles with MaxPeak hardware	Acidic, basic and zwitterionic polar small molecules; metal-sensitive analytes	Use at least 10 mM buffer on-column, equilibrate a new column for at least 50 column volumes and keep the sample solvent no stronger than the initial mobile phase.	Recommended pH range is 2–10; buffer precipitation must be avoided during solvent changes.
Waters ACQUITY BEH Amide / Glycan BEH Amide	Neutral amide bonded to a hybrid support	Carbohydrates, glycans and other strongly hydrophilic neutral or charged compounds	Condition thoroughly. The Glycan BEH Amide guide specifies 50 column volumes for first use, 20 column volumes before the first injection and 8–10 column volumes between injections.	Published glycan conditions such as 50 mM ammonium formate at pH 4.4 and 60°C are application-specific, not universal small-molecule settings.
Agilent Poroshell 120 HILIC-Z	Zwitterionic superficially porous phase	Polar acidic compounds and metabolites; fast LC-MS screening	Use acetonitrile as the weak solvent and screen buffer concentration and pH as selectivity variables.	Do not transfer a bare-silica method without re-optimization; the electrostatic selectivity is different.
Agilent ZORBAX HILIC Plus	Bare silica	Polar bases and mixed polar test sets where silanol interactions may add selectivity	Agilent gives 5–10 mM acetate or formate as a typical mobile-phase starting level and warns that equilibration may require 20–50 column volumes.	Ionized silanols can strongly retain cations and cause tailing or overload if pH and ionic strength are unsuitable.
Merck SeQuant ZIC-HILIC	Sulfobetaine zwitterion on silica	Broad polar and ionic compound screening	Use 5–20 mM formate or acetate, retain at least 3% water, equilibrate for 7–8 column volumes and keep the sample at 60–100% organic or at the initial composition.	Specified pH range is 3–8. Strong alkaline cleaning and poorly soluble phosphate buffers should be avoided.
Merck SeQuant ZIC-pHILIC	Sulfobetaine zwitterion on a polymer support	Polar compounds requiring alkaline or wider-pH screening	The dedicated user guide permits operation from pH 2–10 and recommends equilibration with 80% acetonitrile.	Polymer and silica versions are not interchangeable; their surface, efficiency and selectivity differ.
Thermo Scientific Accucore 150-Amide-HILIC	Amide phase, 150 Å solid-core particles	Glycans, peptides and hydrophilic biomolecules	Condition first with 50:50 acetonitrile/buffer, then 90:10, then the initial method conditions. Thermo notes that HILIC often needs two to three times the equilibration used for reversed phase.	High-water cleaning increases pressure on an amide HILIC column; reduce the flow if water exceeds approximately 50%.
Phenomenex Luna HILIC	Cross-linked diol	Very polar neutral and ionizable small molecules	Phenomenex recommends 10–30 column volumes for equilibration. Its column-care guide uses 80:20 acetonitrile/water as the HILIC storage condition.	For contaminated columns, the cleaning procedure is chemistry-specific; do not improvise a strong acid/base wash.

The entries above summarize official care-and-use documents. Always check the current certificate, pressure limit, pH range and storage solvent for the exact particle size and hardware before use.

A practical generic HILIC starting method

Scouting method for a 2.1 × 100 mm small-molecule column

Mobile phase A: 10 mM ammonium formate or ammonium acetate in LC-MS-grade water, adjusted to the selected aqueous pH.
Mobile phase B: acetonitrile, LC-MS grade.
Initial condition: 10% A / 90% B.
Gradient: hold 90% B for 1 minute, move to 60% B over 12 minutes, hold 2 minutes, return to 90% B in 0.5 minute.
Re-equilibration: begin with at least 10–20 column volumes; use the larger manufacturer requirement when specified.
Flow: 0.30–0.40 mL/min, subject to column pressure and particle-size limits.
Temperature: 30°C as a neutral starting point.
Injection: 0.5–1.0 µL initially, with the sample prepared in 80–90% acetonitrile if solubility allows.

This is a screening method, not a validated universal method. If the least retained analyte elutes at or near the void, increase the initial acetonitrile content or test a more retentive chemistry. If strongly retained compounds do not elute, extend the gradient toward 40–50% acetonitrile or add a sufficiently strong aqueous wash that remains compatible with the column.

Concrete HILIC gradient programs

All programs below use the same convention so they can be entered directly into most LC data systems:

Eluent A: aqueous volatile buffer at the stated concentration and aqueous pH.
Eluent B: acetonitrile.
% B: percentage acetonitrile delivered by the binary pump.
Default hardware: 2.1 × 100 mm column, 0.35 mL/min, 30°C, 0.5–1.0 µL injection.

Buffer-concentration note: with 10 mM salt only in eluent A, the delivered salt concentration changes during the gradient. At 90% B it is 1 mM; at 60% B it is 4 mM. This common binary setup is useful for initial scouting, but it does not satisfy a manufacturer’s requirement for a constant or minimum on-column salt concentration. If ionic strength materially changes selectivity, use separately premixed eluents with equal final salt concentration, a ternary pump or the exact preparation specified in the column manual—and confirm salt solubility before connecting the column.

Gradient 1: broad small-molecule screen

Use this first for an unknown mixture of polar pharmaceuticals, metabolites or contaminants on a zwitterionic, bare-silica or diol column.

Time (min)	% B: acetonitrile	Function
0.00	90	Initial HILIC retention
1.00	90	Initial hold and focusing
13.00	60	Linear analytical gradient; slope −2.5% B/min
15.00	60	Elute strongly retained polar compounds
15.50	90	Return to initial composition
25.50	90	Ten-minute re-equilibration; extend if the column manual requires more volume

Recommended buffer screen: first run 10 mM ammonium formate at aqueous pH 3.0; repeat with 10 mM ammonium acetate at aqueous pH 6.8. Use pH 9 only on a column explicitly rated for alkaline operation.

Gradient 2: SeQuant ZIC-HILIC broad ionic screen

This program follows the generic gradient direction and slope given in the official ZIC-HILIC guide: 90% to 40% acetonitrile in 20 minutes. It is suitable when highly polar ionic analytes remain too strongly retained in Gradient 1.

Time (min)	% B: acetonitrile	Function
0.00	90	Initial condition
2.00	90	Initial hold
22.00	40	Linear gradient; −2.5% B/min
25.00	40	Strong aqueous-rich elution
25.50	90	Return
33.50	90	Eight-minute re-equilibration, corresponding to approximately eight column volumes at the stated dimensions and flow

Starting mobile phase: 10–20 mM ammonium acetate at aqueous pH 6.8. For standard silica-supported ZIC-HILIC, remain within pH 3–8. Use the ZIC-pHILIC polymer phase when deliberate alkaline screening is required.

Gradient 3: amide phase for sugars and neutral highly polar compounds

Amide phases often need a less extreme initial composition than zwitterionic phases for carbohydrates. This scouting program is appropriate for Waters BEH Amide, Thermo Accucore 150-Amide-HILIC or comparable amide columns; it is not a substitute for a manufacturer-validated glycan application.

Time (min)	% B: acetonitrile	Function
0.00	80	Initial condition
1.00	80	Initial hold
13.00	55	Linear separation of increasingly hydrophilic compounds
16.00	55	Late elution
16.50	80	Return
26.50	80	Re-equilibration; use 20 column volumes before the first injection on Waters Glycan BEH Amide

Small sugars: begin with 10 mM ammonium formate or acetate. Labelled N-glycans: Waters specifically presents 50 mM ammonium formate at aqueous pH 4.4 and 60°C as a starting condition; use the application-specific gradient and label chemistry for that workflow.

Gradient 4: acidic polar metabolites on a zwitterionic phase

Use this narrower gradient when acids elute too early in the broad 90–60% screen. It retains the high-organic region longer and avoids an unnecessarily strong aqueous endpoint.

Time (min)	% B: acetonitrile	Function
0.00	95	Strong focusing and retention
2.00	95	Initial hold
12.00	70	Analytical gradient
15.00	60	Late-compound elution
17.00	60	Hold
17.50	95	Return
29.50	95	Long re-equilibration for retention stability

Column choices: Agilent Poroshell 120 HILIC-Z, Waters Atlantis Premier BEH Z-HILIC or SeQuant ZIC-HILIC. Begin at aqueous pH 3 and compare pH 6–7 because electrostatic selectivity can dominate the change in retention.

Gradient 5: complex-matrix method with a strong wash

Use this program for plasma, urine, wastewater, food or plant extracts when late matrix components accumulate. Confirm that the exact column tolerates 20% acetonitrile and reduce the flow during a high-water segment if the manufacturer warns of excessive pressure.

Time (min)	% B: acetonitrile	Flow (mL/min)	Function
0.00	90	0.35	Initial condition
1.00	90	0.35	Initial hold
13.00	60	0.35	Analytical gradient
15.00	60	0.35	Hold
16.00	20	0.20	Move to aqueous-rich matrix wash at reduced flow
20.00	20	0.20	Wash retained salts and polar matrix
21.00	60	0.25	Intermediate composition to avoid an abrupt solvent change
22.00	90	0.35	Return to initial composition
34.00	90	0.35	Re-equilibration

Do not use the strong wash automatically on every amide column. Thermo Fisher notes a significant pressure rise on amide HILIC phases above approximately 50% water. Waters and other manufacturers also specify solvent-transition procedures to prevent buffer precipitation. Run the wash at reduced flow and only after checking the exact care-and-use guide.

Fast 50 mm screening gradient

For a 2.1 × 50 mm column at 0.40 mL/min, use the following rapid screen before transferring the method to a longer column:

Time (min)	% B: acetonitrile
0.00	90
0.50	90
6.50	60
7.50	50
8.00	90
13.00	90

Do not shorten re-equilibration merely because the analytical gradient is short. Verify stability with at least five consecutive standard injections and require an acceptable retention-time RSD before comparing selectivity.

How to adjust a gradient when the first run fails

First-run result	Change to the next gradient	Reason
All peaks at the void	Start at 95% B, hold 2–3 min and reduce the injection volume.	More acetonitrile increases HILIC retention; a smaller aqueous plug improves focusing.
Early peaks are resolved, late peaks are too broad	Keep the initial condition but steepen the late segment, for example 75% to 50% B in 3 min.	Late compounds spend less time migrating under strongly retained conditions.
Critical pair at 75% B	Insert a shallow segment from 80% to 70% B over 10–15 min.	Resolution is gained where the pair actually elutes rather than by lengthening the entire run.
Nothing elutes by 60% B	Extend to 40% B and add a hold; also increase ionic strength if ionic adsorption is suspected.	More water increases elution strength, while buffer screens excessive electrostatic interaction.
Retention changes after matrix samples	Add a validated aqueous-rich wash and increase re-equilibration by 5–10 column volumes.	Matrix removal and reformation of the water layer are both required.

Gradient scaling to another column

Keep the number of column volumes during the gradient approximately constant when changing column dimensions or flow:

tG,new = tG,ref × (Vm,new / Fnew) ÷ (Vm,ref / Fref)

Here, t_G is the gradient time, V_m is the column mobile-phase volume and F is flow. Also scale injection volume with column volume and correct for the instrument’s gradient delay volume. On narrow-bore HILIC methods, a large dwell volume can materially change the effective initial hold and the composition at which each peak elutes.

How to choose pH and buffer

In HILIC, pH changes both analyte charge and stationary-phase charge. It can therefore produce a much larger selectivity change than a small adjustment to gradient slope. A useful development screen includes one acidic, one near-neutral and—only when permitted by the column—one alkaline condition.

Condition	Reason to test it	Possible starting additive	Main risk
Acidic, approximately pH 3	Suppresses silanol ionization and can improve peak shape for bases; alters retention of acidic analytes.	5–10 mM ammonium formate/formic acid	Some acids become neutral and may lose HILIC retention; positive-ion MS may improve while negative-ion response may deteriorate.
Near-neutral, approximately pH 6–7	Useful general screen for zwitterionic columns and mixed metabolite panels.	5–20 mM ammonium acetate	Charged silanols can cause additional cation-exchange retention.
Alkaline, approximately pH 9	Changes the charge of weak acids and bases and may reveal otherwise unavailable selectivity.	5–10 mM ammonium bicarbonate or ammonium hydroxide where the manufacturer permits it	Not suitable for conventional silica outside its stated pH range; some analytes degrade and CO₂-based buffers drift.

“10 mM buffer” is ambiguous in a gradient. If only the aqueous solvent contains 10 mM salt, the salt concentration delivered to the column is 1 mM at 90% acetonitrile and 4 mM at 60% acetonitrile. Waters explicitly recommends at least 10 mM on-column for Atlantis Premier BEH Z-HILIC. Meeting that specification requires a mobile-phase preparation strategy validated for salt solubility in the high-organic solvent. Never add an aqueous salt stock to acetonitrile without confirming that it remains completely dissolved across the entire gradient.

Example buffer calculations

For a buffer prepared from a solid salt:

mass (g) = concentration (mol/L) × volume (L) × molar mass (g/mol)

One litre of 10 mM ammonium acetate
0.010 mol/L × 1.000 L × 77.08 g/mol = 0.7708 g

One litre of 10 mM ammonium formate
0.010 mol/L × 1.000 L × 63.06 g/mol = 0.6306 g

Suitable examples include Sigma-Aldrich/Merck ammonium acetate for LC-MS, ammonium formate for LC-MS and acetonitrile hypergrade for LC-MS. Product links are examples rather than endorsements; the required grade should be chosen for the detector and validation requirements.

Measure and adjust the pH in the aqueous buffer before adding large amounts of acetonitrile. A conventional glass-electrode reading in a water/acetonitrile mixture is an apparent pH and cannot be compared directly with an aqueous pH value unless a defined measurement procedure is used.

The five variables that most often decide whether HILIC works

1. Acetonitrile content

Acetonitrile is the weak solvent in HILIC. Thermo Fisher recommends working above approximately 60–70% acetonitrile and notes that retention normally increases as acetonitrile rises from 60% toward 90%. Merck/SeQuant gives the solvent-strength order acetone < acetonitrile < isopropanol < ethanol < methanol < water for its ZIC-HILIC guidance. Methanol is therefore not a drop-in replacement for acetonitrile.

Do not run a completely dry organic mobile phase. SeQuant specifies at least 3% water to hydrate the particles. In practice, many methods start at 80–95% acetonitrile, depending on analyte polarity, solubility and the selected stationary phase.

2. Ionic strength

Ammonium salts do more than control pH. They screen electrostatic interactions and can change retention, peak symmetry and selectivity. Start with 5–10 mM for LC-MS, then compare 20 mM if tailing or excessive ionic retention remains. SeQuant recommends 5–20 mM for routine ZIC-HILIC work, while Waters specifies at least 10 mM on-column for its Atlantis Premier Z-HILIC.

Use volatile formate, acetate or—when chemically appropriate—bicarbonate for LC-MS. Phosphate can be valuable for UV methods but is poorly soluble in high acetonitrile and is nonvolatile. SeQuant explicitly advises against phosphate and other low-solubility salts in its ZIC-HILIC guide.

3. Sample solvent and injection volume

A water-rich sample plug is a strong solvent plug in HILIC. It can cause fronting, splitting, broadening and early breakthrough even when the mobile-phase method is correct. Waters recommends a sample solvent identical to or weaker than the initial mobile phase. SeQuant recommends 60–100% organic solvent or the initial composition and an injection volume of about 1% of the column volume as a starting point.

If an aqueous biological extract cannot be evaporated and reconstituted, dilute it with acetonitrile, reduce the injection volume or use on-line focusing with deliberate validation. Check for precipitation after adding acetonitrile: proteins, salts and some analytes can leave solution.

4. Equilibration

HILIC retention depends on formation of a reproducible water layer. A stable pressure and baseline do not prove that retention equilibrium has been reached. Manufacturer recommendations vary substantially:

Waters Atlantis Premier BEH Z-HILIC: at least 50 column volumes for initial equilibration.
Waters Glycan BEH Amide: 50 column volumes for first use, 20 before the first injection and 8–10 between injections.
Agilent HILIC columns: equilibration may require 20–50 column volumes.
SeQuant ZIC-HILIC: 7–8 column volumes.
Phenomenex: approximately 10–30 column volumes.
Thermo Fisher: expect roughly two to three times the equilibration of reversed-phase LC.

Example: Waters lists an approximate column volume of 0.35 mL for a 2.1 × 100 mm column. Fifty column volumes equal 17.5 mL. At 0.40 mL/min, initial equilibration therefore requires about 43.8 minutes, excluding system dwell volume. Eight column volumes would require approximately 7 minutes at the same flow.

5. Temperature

Temperature changes viscosity, mass transfer, hydrogen bonding and ionization equilibria. Begin at 25–35°C for small molecules and optimize only after stationary phase, pH and gradient have been screened. Higher temperatures may be useful for glycans and biopolymers, but the exact manufacturer limit must be respected. A temperature change can also alter selectivity, so it should not be used merely as a pressure-control setting.

Column-by-column manufacturer guidance

Waters Atlantis Premier BEH Z-HILIC

The official care-and-use manual specifies pH 2–10, at least 10 mM buffer on-column and at least 50 column volumes of equilibration. New columns are shipped in 100% acetonitrile. Waters recommends starting the initial flush with organic solvent at low flow and then increasing the flow gradually.

Before introducing a buffered mobile phase, flush with five column volumes of a water/organic mixture whose organic content is the same as or lower than that of the intended buffered phase. This reduces the risk of salt precipitation. For storage, remove the buffer, flush with ten column volumes of HPLC-grade water followed by ten column volumes of 100% acetonitrile and seal the column.

Waters BEH Amide and Glycan BEH Amide

Amide phases are useful when hydrogen-bonding selectivity and carbohydrate retention are required. The Glycan BEH Amide guide calls for extensive first-use conditioning and documents 8–10 column volumes between injections. Inadequate re-equilibration is specifically associated with retention-time drift.

For labelled N-glycans, Waters presents a universal starting method using 50 mM ammonium formate at pH 4.4 and 60°C. These settings are relevant evidence for glycan analysis, but they should not be copied blindly to a low-molecular-weight metabolite method.

Agilent Poroshell 120 HILIC-Z and ZORBAX HILIC Plus

The Agilent HILIC method-development overview treats pH, buffer concentration and stationary-phase chemistry as major selectivity variables. HILIC-Z is zwitterionic, whereas ZORBAX HILIC Plus is bare silica; therefore, charged analytes can behave very differently on the two phases.

The Agilent normal-phase and HILIC column guide gives acetonitrile/water with 5–10 mM acetate or formate as a typical mobile phase and states that 20–50 column volumes may be required for equilibration. During a column screen, use the same test mixture and identical mobile phases on both chemistries so that selectivity—not a simultaneous method change—is being compared.

Merck SeQuant ZIC-HILIC and ZIC-pHILIC

The concise ZIC-HILIC user guide is unusually specific: it recommends 5–20 mM formate or acetate, at least 3% water, 7–8 column volumes of equilibration and a sample solvent containing 60–100% organic solvent. It also provides a useful generic gradient from 90% to 40% acetonitrile over 20 minutes and an isocratic starting point of 80:20 acetonitrile/buffer.

Standard ZIC-HILIC is specified for pH 3–8. The polymer-supported ZIC-pHILIC guide permits pH 2–10, making it more appropriate for deliberate alkaline screening. For high-salt matrices such as urine or plasma extracts, SeQuant suggests a near-aqueous end-of-run wash while maintaining sufficient ionic strength.

Thermo Scientific Accucore HILIC phases

Thermo offers silica, zwitterionic, urea and amide HILIC selectivities. The manufacturer’s method-development page recommends volatile ammonium formate or acetate for charged compounds and MS detection and emphasizes that little HILIC retention is normally observed below 60% acetonitrile.

The Thermo HILIC column-care guide recommends conditioning first with 50:50 acetonitrile/buffer, then 90:10 and finally the initial method conditions. It also warns against topping up old buffer and recommends 0.2 µm filtration, or 0.1 µm for UHPLC. Accucore 150-Amide-HILIC, with 150 Å pores and amide hydrogen-bonding interactions, is positioned for glycans and hydrophilic biomolecules.

Phenomenex Luna HILIC

Phenomenex describes acetonitrile as the weak HILIC solvent and the ammonium-buffered aqueous phase as the strong solvent in its HILIC technical tips. Its column care and use guide recommends approximately 10–30 column volumes for equilibration and lists 80:20 acetonitrile/water for Luna HILIC storage.

For a contaminated HILIC column, Phenomenex gives a staged ten-column-volume cleaning sequence: 95:5 water/acetonitrile, then 95% of 100 mM ammonium acetate at pH 5.8 with 5% acetonitrile, then 95:5 water/acetonitrile and finally the mobile phase. Confirm compatibility with the exact column before applying this procedure.

HILIC in complex matrices

Plasma, urine, wastewater, food and plant extracts contain salts and polar endogenous compounds that may be retained more strongly than the target analyte. The result can be ion suppression, overload, memory effects and a gradual change in stationary-phase behavior. A neat-standard chromatogram is therefore insufficient evidence that a HILIC method is suitable for the sample.

Sample preparation

Remove proteins and particles before injection.
Use protein precipitation, SPE or dilution appropriate to the analyte.
Check analyte recovery after converting the extract to high acetonitrile.
Use an inline filter or guard column for dirty matrices.

Matrix controls

Run extracted blanks and post-extraction spikes.
Measure matrix factors in multiple matrix lots for quantitative LC-MS.
Include pooled QC samples through the sequence.
Monitor retention drift, pressure and internal-standard response.

For salt-rich samples, do not simply increase the injection volume to gain sensitivity. A salt-rich aqueous plug is simultaneously a strong injection solvent and a source of ion suppression. Lower injection volume, higher-organic reconstitution and improved cleanup often give a higher effective signal-to-noise ratio than injecting more extract.

A controlled development experiment

Rather than changing one factor randomly after every failed chromatogram, use a compact, staged screen:

Prepare a representative test mixture. Include the earliest and latest expected analytes, internal standards and at least one post-extraction matrix sample.
Screen three different phase types. A practical set is bare silica, amide and zwitterionic. Run identical 90–60% acetonitrile gradients.
Select the two best phases. Judge retention factor, critical-pair selectivity, peak shape and matrix separation—not total peak count alone.
Screen pH. Compare approximately pH 3, 6–7 and, only on a compatible phase, pH 9.
Screen ionic strength. Compare 5 or 10 mM with 20 mM buffer while keeping pH and gradient constant.
Optimize the gradient. Adjust initial acetonitrile, final acetonitrile and slope around the critical pair.
Challenge the injection solvent. Compare the intended sample solvent and volume with a high-acetonitrile reference preparation.
Determine the actual equilibration requirement. Shorten the re-equilibration in steps until retention shifts, then add a safety margin.
Test real matrix and sequence robustness. Alternate blanks, QCs and high-concentration matrix samples to reveal carryover and column conditioning effects.

Troubleshooting HILIC methods

Observation	Likely cause	First experiment
No or very low retention	Too much water, sample plug too aqueous, analyte not sufficiently hydrophilic or unsuitable phase chemistry	Increase initial acetonitrile by 5–10 percentage points, reduce injection volume and compare a different stationary phase.
Retention time drifts later or earlier	Insufficient re-equilibration, changing buffer composition, evaporation of acetonitrile or column conditioning by matrix	Double the equilibration volume, prepare fresh solvents gravimetrically and run repeated standards until retention stabilizes.
Split, broad or fronting peaks	Strong aqueous sample solvent, volume overload, mass overload or precipitation	Reduce injection fivefold and reconstitute in the initial mobile phase.
Severe tailing of ionic analytes	Electrostatic secondary interaction, insufficient ionic strength or active metal/silanol sites	Compare 10 and 20 mM buffer, change pH and screen zwitterionic versus amide or inert-hardware chemistry.
Sudden pressure increase	Salt precipitation, matrix particulates, protein precipitation or high-viscosity aqueous wash	Stop the gradient, identify where pressure increased, verify salt solubility and inspect the inlet frit/guard column.
Good standard but poor matrix response	Coeluting polar matrix, ion suppression, adsorption or recovery loss during high-organic reconstitution	Run post-extraction spikes, a dilution series and a matrix-factor experiment.
Carryover after matrix samples	Strong ionic adsorption, inadequate needle wash or insufficient aqueous column wash	Use a mixed aqueous/organic needle wash and add a manufacturer-compatible strong wash at the end of the gradient.

Final recommendation

The most reliable HILIC development strategy is to screen stationary-phase chemistry before fine-tuning gradient time. Use at least one neutral amide, one bare-silica or diol phase and one zwitterionic phase when the sample contains mixed polar analytes. Begin with high acetonitrile, a low-to-moderate concentration of volatile ammonium buffer, a small injection prepared close to the initial mobile phase and the full equilibration specified by the column manufacturer.

For complex matrices, the winning method is not necessarily the one with the greatest retention. It is the method that separates the analyte from the matrix, maintains stable retention across a long sequence, tolerates the real sample solvent and gives acceptable recovery, peak shape and detector response. Document column volumes, preparation order, aqueous pH, solvent lots, conditioning history and storage conditions; HILIC is highly reproducible only when these details are controlled.