C18 vs C8 vs Phenyl vs PFP vs HILIC Selectivity

C18 and C8 primarily retain analytes through hydrophobic interactions, while phenyl and pentafluorophenyl (PFP) phases can alter selectivity through aromatic, dipolar and other interactions. HILIC instead favors retention of many polar solutes under high-organic conditions. Choose the phase by the critical analyte pair, not by a stationary-phase label alone.

How does stationary-phase chemistry change HPLC selectivity?

Selectivity describes how differently two compounds are retained, not merely how long they remain on the column. For two retained peaks, the separation factor is the ratio of their retention factors, with the later-eluting peak in the numerator. A change of bonded ligand, accessible silanol population or mobile-phase composition may change relative retention even when total retention changes little. In reversed-phase liquid chromatography, hydrophobic partitioning is often dominant, but steric accessibility, hydrogen bonding, electrostatic interactions and aromatic interactions can matter for particular analytes.¹

α = k₂ / k₁ (k₂ ≥ k₁ > 0)

Here k is the retention factor and α is the separation factor. Because α is defined using the elution order, it cannot by itself reveal an elution-order reversal: compare named analytes and their individual k values across phases. For a worked discussion of the metric, see HPLC separation factor.

Diagnose this automatically

Chromatography Troubleshooting Decision Engine

Any HPLC symptom, one starting point — the engine narrows hundreds of failure modes to the few that fit your evidence.

Start a diagnosis

No account needed

How do C18, C8, phenyl, PFP and HILIC differ?

C18 (octadecyl) and C8 (octyl) are alkyl-bonded reversed-phase chemistries. C18 commonly offers strong hydrophobic retention, whereas C8 often provides less retention under comparable conditions; actual differences depend on coverage, surface area and ligand architecture.¹ Phenyl-bonded phases add π–π interactions with aromatic and unsaturated analytes, although hydrophobic interaction often still dominates; the π–π contribution is most noticeable with short spacers, while on a phenyl-hexyl phase the hexyl spacer’s hydrophobic interaction dominates.² The organic modifier matters too: acetonitrile can suppress the π–π contribution, so a phenyl phase run in acetonitrile may give much the same selectivity as C18, whereas methanol is more likely to change the peak order.² PFP phases add a highly fluorinated aromatic surface with dipolar, π–π, charge-transfer and ion-exchange interactions; they often give selectivity distinct from conventional alkyl phases and generally retain basic analytes more strongly than their alkyl counterparts.³ HILIC is a retention mode, not a single bonded ligand: bare silica, amide, zwitterionic and other polar surfaces can operate in HILIC conditions, with substantial mechanistic differences.⁴

Table 1 pairs each chemistry with its USP packing class⁵ and with the usable pH range and carbon load that manufacturers state for two example products of each type.^6–9 These ranges are typical, not family-wide limits: the datasheet of the specific column decides, and other particle types can extend the range (one hybrid-particle amide phase is rated pH 2–11).⁹ A shared L-code places two columns in the same category; it does not show that they have the same selectivity.

**Table 1.** Stationary-phase chemistry, USP packing class and typical specifications. pH and carbon load are manufacturer-stated values for two example products per type, in the order Phenomenex Luna / Agilent (ZORBAX Eclipse Plus; Poroshell 120 PFP; HILIC Plus); typical, not limits.
Phase	Principal retention/selectivity features	USP packing	Usable pH (examples)	Carbon load (examples)	Choose it when
C18	Hydrophobic interactions; coverage and silanols affect selectivity	L1	1.5–9* / 2–9	17.5% / 9%	General reversed-phase screen for moderately hydrophobic analytes
C8	Hydrophobic interactions with shorter bonded chains	L7	1.5–9* / 2–9	13.5% / 7%	C18 retains too strongly or a different alkyl selectivity is useful
Phenyl (phenyl-hexyl examples)	Hydrophobic plus π–π and steric (shape) interactions; spacer and organic modifier matter	L11	1.5–9* / 2–8	17.5% / 9%	Aromatic critical pairs need an orthogonal screen
PFP	Fluorinated aromatic surface; dipolar, π–π, charge-transfer and ion-exchange interactions	L43	1.5–8 / 2–9	11.5% / 5.1%	Conventional C18/phenyl phases give inadequate critical-pair selectivity, especially for aromatic or basic analytes
HILIC	Water-rich interfacial partitioning plus surface-specific interactions	Depends on the surface: L20 (diol), L3 (bare silica)	1.5–8 / 1–8	5.7% / 0% (bare silica)	Very polar analytes show inadequate reversed-phase retention

*Luna: pH 1.5–9 under gradient conditions, 1.5–10 isocratic.⁶ Carbon load is listed for orientation only; as the FAQ below explains, it is not a stand-alone selectivity metric, and a bare-silica HILIC phase has none.

Figure 1 sets the five chemistries side by side: the bonded ligand, when to screen it and which interactions it brings.

Five columns for C18, C8, phenyl, PFP and HILIC. Top: a silica particle with a schematic ligand (long alkyl chain, short alkyl chain, phenyl ring on a spacer, pentafluorophenyl ring on a spacer, polar OH group with water) and when to screen that phase. Bottom: a matrix of six interactions (hydrophobic, π–π, dipolar, ionic, steric, water-layer partitioning) with filled dots for principal and rings for contributing interactions. — **Figure 1. Five stationary-phase choices.** Each column pairs a schematic ligand with the situation in which the article suggests screening it and the interactions it brings. C18 and C8 are alkyl phases that differ in bonded-chain length; phenyl adds π–π and steric (shape) interactions through a spacer; PFP adds a fluorinated aromatic surface with dipolar, π–π, charge-transfer and ion-exchange interactions; HILIC is a retention mode on polar surfaces, driven by partitioning into a water-rich layer. Residual silanols contribute ionic interactions on silica-based reversed phases; PFP’s hydrophobic contribution is inferred from its reversed-phase use. Filled dots mark principal interactions and rings contributing ones; real columns vary by product and mobile phase. Schematic ligand sketches, not molecular-scale drawings or data.

Which stationary phase should you choose for your analytes?

Start with the analyte structures, pKa values, expected ionization at the proposed mobile-phase pH and the identity of the critical pair. If both analytes are moderately hydrophobic, screen a representative C18 and evaluate retention, peak shape and separation. If retention is excessive, C8 is one possible alternative, although changing organic composition may be more efficient. If the critical pair consists of aromatic isomers, a phenyl or PFP chemistry may provide a more informative selectivity change than simply changing C18 brand. For highly polar solutes that elute near the void under reversed-phase conditions, screen a chemically appropriate HILIC column with a compatible injection solvent and equilibration procedure. These are starting hypotheses, not predictions of elution order.

Figure 2 shows why the change of chemistry matters: under the same conditions, a second phase can reverse the order of a pair, so the comparison has to be made on named analytes.

Two schematic chromatograms. On a C18 phase, analyte A (taller peak) elutes before analyte B (shorter, highlighted). On a phenyl phase the same two peaks appear in reverse order, B before A. Below, a note explains that alpha uses elution order, and three checks in order: confirm peak identity, record each analyte's k, measure the critical pair. — **Figure 2. Stationary phase can reverse elution order.** Schematic traces, not data: the same two analytes under the same conditions elute A before B on a C18 phase and B before A on a phenyl phase, where added aromatic interactions can change relative retention; here A is taken, for illustration, as the analyte that gains more aromatic retention. Peaks are equal-width Gaussians and each analyte keeps its own height on both phases. Because the separation factor α = k₂/k₁ is defined with the later peak in the numerator, it cannot by itself reveal a reversal; compare each named analyte’s k. Before calling an apparent reversal real, confirm peak identity, record named-analyte retention factors and measure critical-pair resolution and peak shape. A reversal is possible, not guaranteed.

Table 2 translates these hypotheses into a screening plan. The phase is a candidate to test, not a guaranteed separation.

**Table 2.** Analyte-driven first and alternative column screens.
Observed method problem	First screen	Orthogonal screen	Verification
Moderately hydrophobic pair, little prior information	C18	Phenyl or PFP	Measure named-analyte retention and critical-pair resolution
Strong retention on C18	Adjust eluent strength	C8	Confirm retention window and selectivity
Aromatic isomers coelute on C18	Phenyl	PFP	Compare named-analyte elution order and peak identity
Polar compounds poorly retained in RP	Suitable HILIC chemistry	Different HILIC ligand	Verify equilibration, retention reproducibility and sample-solvent compatibility

For broader method development, see How to develop an HPLC method and the retention-factor guide.

Why can two columns with the same USP L-designation separate differently?

A USP L-designation describes a category of packing chemistry; it is not a quantitative certificate of selectivity equivalence. Two nominally L1 columns can differ in ligand density, endcapping, silica purity, pore architecture, residual silanol activity, particle design and bonding chemistry. These differences can affect basic, acidic and sterically constrained analytes disproportionately. A change in peak order or a loss of critical-pair resolution must be evaluated experimentally, even when both products carry the same designation.¹

Table 3 identifies diagnostic observations that can help separate chemistry-related differences from generic efficiency problems.

**Table 3.** Column-property differences and discriminating observations.
Column difference	Potential observation	Useful comparison
Bonded-phase coverage	Different hydrophobic retention	Neutral probe retention under matched conditions
Residual silanol activity	Basic analytes shift or tail selectively	Basic versus neutral probe behavior
Aromatic ligand or spacer	Changed selectivity for aromatic pair	Named-analyte retention on alkyl and aromatic phases
Surface chemistry and water layer in HILIC	Different polar-analyte retention or equilibration	Retention repeatability and mobile-phase sensitivity

For validated or compendial methods, the applicable procedure and regulatory framework take precedence over the general rules of thumb given here. A permitted adjustment under a compendial framework is not permission to substitute arbitrary chemistry without evaluating the method and its regulatory context.

How does the hydrophobic-subtraction model compare reversed-phase columns?

The Snyder–Dolan–Carr hydrophobic-subtraction model describes reversed-phase column selectivity with five measured column descriptors: hydrophobicity (H), steric resistance (S*), hydrogen-bond acidity (A), hydrogen-bond basicity (B) and cation-exchange activity (C). In the original framework, the relative retention of probe solutes is used to estimate how columns differ across those interactions. The model offers a more defensible comparison than relying on C18 or L1 labels alone. Its C term is especially sensitive to measurement pH, and published descriptors apply to their characterization conditions.^1,10 USP’s PQRI procedure for choosing a replacement reversed-phase column compares columns on these same five parameters.¹¹

The model is not a universal mechanistic equation for every stationary phase: it was developed for reversed-phase columns, not as a direct classification of all HILIC systems. Nor does a small descriptor distance automatically prove a compendial substitution is acceptable. Review the underlying model in Snyder, Dolan and Carr. For a separate treatment of chromatographic performance, see HPLC resolution.

How should you screen a second column when the first fails?

Hold the analyte mixture and a documented set of method conditions constant for the initial comparison wherever chemically feasible. Confirm peak identity with reference materials or suitable detection evidence before interpreting an apparent elution-order reversal. Record retention factors for each named analyte, critical-pair resolution, peak shape and pressure. Select a second phase because it changes a plausible interaction—such as alkyl to aromatic chemistry—rather than simply because it is another product. When moving to HILIC, redesign mobile-phase and injection-solvent conditions appropriately instead of assuming a direct plug-in swap. Finally, verify repeatability and the method’s required system-suitability attributes.^1,4

Figure 3 summarizes the path: assess the analytes, pick a first and a second screen by the observed problem, then verify each screen.

Three-band workflow. Band A lists what to assess first: structures, pKa values, ionization at the mobile-phase pH and the critical pair. Band B shows three alternative branches, each with a schematic chromatogram: a hydrophobic pair (first C18, then phenyl or PFP), co-eluting aromatic isomers (phenyl, then PFP) and polar analytes near the void (HILIC, then a different HILIC ligand). Band C shows four sequential verification steps. — **Figure 3. Choose a first and a second column screen.** Start from the analyte structures, pKa values, expected ionization at the mobile-phase pH and the identity of the critical pair. A moderately hydrophobic pair with little prior information starts on C18 with phenyl or PFP as the orthogonal screen; if retention is excessive, adjust eluent strength before trying C8. Aromatic isomers that co-elute on C18 go to phenyl, then PFP. Polar analytes that elute near the void go to a suitable HILIC chemistry, then a different HILIC ligand, with a compatible injection solvent and equilibration. Every screen is verified: fix comparison conditions where feasible (for HILIC, redesign them), confirm peak identity, record each analyte’s k, critical-pair resolution, peak shape and pressure, then verify repeatability and system suitability. Schematic traces, not data; starting hypotheses, not predictions of elution order. For validated or compendial methods the applicable procedure takes precedence.

Frequently asked questions

Is C18 always more retentive than C8?

No. C18 often provides stronger hydrophobic retention than a comparable C8 under similar conditions, but different manufacturers’ materials are not controlled comparisons. Bonded-phase density, silica surface area, pore structure, endcapping and mobile-phase conditions can all affect observed retention. A C8 from one family can behave differently from what a simple chain-length comparison suggests. Measure the retention factors of representative neutral and target analytes before concluding which column is stronger under the method conditions.

Does phenyl always retain aromatic compounds more than C18?

No. Aromatic interaction opportunities do not guarantee greater absolute retention. Phenyl phases can differ in ligand attachment and spacer chemistry, and the mobile phase strongly affects the balance of interactions; with acetonitrile in particular, a phenyl phase may give much the same selectivity as C18. Their practical value is often a change in relative retention for a critical aromatic pair, not a universally longer retention time. Compare named analytes and confirm peak identities before claiming a reversal or an improvement in selectivity.

Is PFP the best column for positional isomers?

PFP is a useful orthogonal screening candidate when conventional alkyl phases provide inadequate selectivity, because its fluorinated ring adds dipolar, π–π, charge-transfer and ion-exchange interactions. Phenyl phases are also well suited to aromatic positional isomers. However, its performance depends on analyte electronics, substitution pattern, ionization and mobile-phase conditions. A PFP phase can fail where a phenyl, polar-embedded or other chemistry succeeds. Evaluate the actual critical pair experimentally rather than using a broad claim that PFP always separates isomers.

Can HILIC replace C18 without changing the mobile phase?

Usually not. HILIC commonly uses a high proportion of organic solvent and retention can depend strongly on the water-rich interfacial environment, buffer composition and equilibration. Injection-solvent mismatch may distort peaks, particularly for polar analytes. Moving from reversed phase to HILIC should be treated as a method-development change with new solvent, equilibration and robustness checks, not as a simple stationary-phase swap.

Does the same USP L-code guarantee equivalent columns?

No. A USP L-code groups packing materials by specified chemistry but does not demonstrate matching selectivity for the analytes in a particular procedure. Differences in ligand coverage, silica and secondary interactions may change resolution or elution order. Any replacement should be evaluated against the applicable procedure and regulatory requirements. A method’s system-suitability result is necessary evidence where required, but should not be treated as automatic permission for every chemistry change.

What does carbon load tell me about selectivity?

Carbon load describes the mass fraction of carbon in a packing material, but it is not a stand-alone selectivity metric. Different ligand chemistries, pore structures and surface areas complicate direct comparisons. A higher carbon percentage does not necessarily imply stronger retention for every analyte, and it cannot predict aromatic or ionic interactions reliably. Use product-specific data alongside retention measurements and, when available, quantitative column-selectivity descriptors.

The takeaway

Choose stationary-phase chemistry to test a specific selectivity hypothesis: alkyl phases for a general reversed-phase baseline, phenyl or PFP for a different aromatic-interaction profile and a suitable HILIC chemistry for polar analytes with poor reversed-phase retention. Confirm the result with named-analyte retention, critical-pair resolution and method-specific requirements; neither a USP L-code nor a ligand label guarantees equivalence.

References

Snyder LR, Dolan JW, Carr PW. The hydrophobic-subtraction model of reversed-phase column selectivity. J Chromatogr A. 2004;1060:77–116. https://doi.org/10.1016/j.chroma.2004.08.121 — five column parameters (H, S*, A, B, C) and their use in comparing reversed-phase columns.
Taylor T. Getting the most from phenyl stationary phases for HPLC. LCGC International, The LCGC Blog, 13 September 2016. https://www.chromatographyonline.com/view/getting-most-phenyl-stationary-phases-hplc — π–π, hydrophobic and steric interactions; spacer effects; acetonitrile versus methanol; selectivity rather than retention.
Santasania CT, Bell DS. Mechanisms of interaction responsible for alternative selectivity of fluorinated stationary phases. LCGC North America. 2016;34(2) (Column Watch). https://www.chromatographyonline.com/view/mechanisms-interaction-responsible-alternative-selectivity-fluorinated-stationary-phases — dipole, π–π, charge-transfer and ion-exchange interactions; greater retention of bases than alkyl phases.
Alpert AJ. Hydrophilic-interaction chromatography for the separation of peptides, nucleic acids and other polar compounds. J Chromatogr. 1990;499:177–196. https://doi.org/10.1016/S0021-9673(00)96972-3 — the paper that introduced the term HILIC (bibliographic record verified via Crossref 2026-10-08; full text not opened).
United States Pharmacopeia. Chromatographic Columns database: packing definitions L1, L3, L7, L11, L20 and L43, with listed brands. https://www.usp.org/resources/chromatographic-columns (accessed 2026-10-08) — L1 octadecyl silane; L7 octylsilane; L11 phenyl groups; L43 pentafluorophenyl groups bonded by a propyl spacer; L3 porous silica; L20 dihydroxypropane (diol) groups. Listed brands include Luna C18(2) (L1), Luna Phenyl-Hexyl (L11), Luna PFP(2) (L43), Luna HILIC (L20) and ZORBAX Eclipse Plus C18 (L1).
Phenomenex. Luna HPLC columns: specifications (C18(2), C8(2), Phenyl-Hexyl, PFP(2), HILIC). https://www.phenomenex.com/products/luna-hplc-column (accessed 2026-10-08) — manufacturer-stated carbon load, pH stability and USP classification.
Agilent Technologies. ZORBAX HPLC columns poster, publication 5994-2212EN, 20 July 2020. https://www.agilent.com/cs/library/posters/public/poster-hplc-zorbax-xolumns-5994-2212EN-agilent.pdf — Eclipse Plus C18, C8, Phenyl-Hexyl and HILIC Plus carbon load, pH range and USP code.
Agilent Technologies. InfinityLab Poroshell 120 PFP, part 693975-408, specifications. https://www.agilent.com/store/en_US/Prod-693975-408/693975-408 (accessed 2026-10-08) — carbon load 5.1%, pH 2–9.
Waters. WKB32668: What is the pH range of BEH Amide columns? https://support.waters.com/KB_Chem/Columns/WKB32668_What_is_the_pH_of_BEH_Amide_columns (accessed 2026-10-08) — pH 2–11.
Zhang Y, Carr PW. A visual approach to stationary phase selectivity classification based on the Snyder–Dolan hydrophobic-subtraction model. J Chromatogr A. 2009;1216(39):6685–6694. https://doi.org/10.1016/j.chroma.2009.06.048 — independent explanation of the steric, hydrogen-bond and cation-exchange contributions.
United States Pharmacopeia. USP-PQRI procedure for choosing a replacement reversed-phase column. https://usp.org/node/12526 (accessed 2026-10-08) — five column properties, database comparison, cites¹.

C18 vs C8 vs Phenyl vs PFP vs HILIC: How Stationary-Phase Chemistry Changes Selectivity