Skin friction vs end bearing: how piles carry axial load

Piles carry axial load through two mechanisms: skin friction — shear resistance mobilized along the shaft–soil interface — and end bearing — compressive resistance at the tip. Total capacity is the sum, Qᵤ = Qₛ + Qₜ. A friction pile draws most of its capacity from the shaft; an end-bearing pile from the tip. Most real piles rely on both, and the split shifts with settlement.

The two mechanisms

Push a pile downward and two things resist. Along the shaft, the soil grips the pile surface and develops shear stress — skin friction (also called side resistance or shaft resistance). At the tip, the soil below the base is loaded in compression like a deep, buried footing — end bearing. In the classical static formula, the side resistance is the unit friction fₛ integrated over the shaft surface, and the tip resistance is a unit bearing pressure q over the tip area:

Qᵤ = Qₛ + Qₜ = Σ fₛ · (perimeter) · Δz  +  q · Aₜᵢₚ

Everything else in axial pile design is about how fₛ and q are evaluated — and, just as important, how much movement each one needs before it actually shows up. PileCalc's axial capacity module follows the NAVFAC DM-7.02 procedure: it integrates fₛ over 100 segments along the embedded length and adds the tip term, layer by layer.

How side friction is computed

In clay (the α-method). Under undrained conditions the shaft resistance is an adhesion — a fraction of the undrained shear strength cᵤ:

fₛ = α · cᵤ

The adhesion factor α reflects that a pile cannot grip stiff clay as efficiently as soft clay: on the Tomlinson-style curve PileCalc uses, α ≈ 1.0 for soft clay (cᵤ up to roughly 25 kPa / 500 psf) and falls to about 0.5 for very stiff clay (cᵤ near 200 kPa / 4,000 psf). Soft clay sticks; stiff clay slips.

In sand (the β / K·tanδ approach). Shaft resistance is frictional, so it scales with the effective normal stress pressing the soil against the shaft:

fₛ = K · σ′ᵥ · tan δ

where σ′ᵥ is the vertical effective stress, K is the lateral earth-pressure ratio (how much of that vertical stress acts horizontally on the shaft — near 1.0 for driven displacement piles, lower for drilled shafts), and δ is the pile–soil interface friction angle, a material-dependent fraction of φ. Grouping K·tanδ into a single coefficient is the familiar β-method: fₛ = β·σ′ᵥ.

How end bearing is computed

In clay, tip resistance is a bearing-capacity problem in undrained soil:

q = Nc · cᵤ,   Nc → 9 for embedment deeper than ~4 diameters

Nc climbs from about 6.3 at the surface to its deep-foundation limit of 9 once the tip is more than about four diameters down — which is virtually always, so "9c" is the working rule.

In sand, tip resistance scales with the effective overburden at the tip:

q = Nq · σ′ᵥ

with Nq a strong function of friction angle and installation method. From the NAVFAC table PileCalc implements, a driven (displacement) pile in φ = 30° sand gets Nq = 21; the same sand around a drilled (non-displacement) shaft gets Nq ≈ 10, because drilling relieves the stress state instead of densifying it. At φ = 36° the pair is roughly 61 vs 30. This is why an end-bearing-dominated design usually wants a driven pile.

The limiting-depth concept in sand

Taken literally, fₛ = K·σ′ᵥ·tanδ and q = Nq·σ′ᵥ would grow without bound as the pile gets deeper, since σ′ᵥ keeps increasing. Load tests say otherwise: below a critical depth, unit resistances stop growing. The NAVFAC-style fix is a limiting depth — the effective stress used in both formulas is capped at its value at a penetration of about 20 pile diameters (PileCalc's default; adjustable, with optional absolute caps on fₛ and q as well). Beyond z = 20·B, extra depth still adds shaft area, but no longer adds unit stress. Skipping this cap is one of the classic ways hand calculations overpredict the capacity of long piles in sand.

Friction pile vs end-bearing pile — and why most piles are both

The classification describes where the capacity predominantly comes from:

• Friction pile: no competent bearing stratum within reach, so the pile "floats" in soil and accumulates capacity along its shaft — e.g. a long pile in deep clay or loose-to-medium sand.
• End-bearing pile: the tip is seated on or socketed into a hard stratum — dense sand, till, or rock — and the tip term dominates. In rock, PileCalc switches to socket correlations (side resistance 0.65·√(pₐ·qᵤ) after Horvath & Kenney 1979, as adopted in FHWA-IF-99-025).

But the labels are about proportion, not mechanism. Any embedded pile develops both resistances; a "friction pile" still has a tip, and an "end-bearing pile" still drags its shaft through soil. Rough proportions for common profiles:

These are tendencies, not rules — run the numbers for your profile rather than trusting the label.

Load transfer: friction mobilizes first, the tip needs movement

The ultimate values above are not mobilized at the same settlement, and this is the single most useful thing to understand about axial pile behavior. In the t-z / q-w load-transfer framework (the Reese & O'Neill 1988 normalized curves PileCalc uses for its load–settlement analysis):

• Side resistance mobilizes at small movement — a fraction of a percent of the pile diameter. On the normalized curves, clay shaft friction is ~85% mobilized at a relative movement of 0.4% of the diameter and peaks near 0.8%; sand is fully mobilized by about 1–1.5%. For a 600 mm pile that is millimetres of settlement.
• End bearing needs an order of magnitude more. The base curves reach only ~50% of ultimate at a base movement of 2–4% of the diameter and full mobilization around 10% (clay) to 12% (sand) — tens of millimetres for that same 600 mm pile.

The consequence: at working load, most piles are friction piles, whatever their ultimate-capacity label says. The shaft picks up load first; the tip only works hard after the shaft is nearly exhausted and the pile has settled visibly. A load–settlement curve built by sweeping tip movement through the transfer curves — as in PileCalc's pile settlement calculator — makes the handoff from shaft to tip explicit. This is also why side and tip resistances traditionally carry different safety factors (2 on side, 3 on tip in the NAVFAC allowable-stress scheme): the tip value is both less certain and further from being mobilized at service settlement.

Negative skin friction: when the shaft loads the pile instead

Skin friction is a two-way street. The formulas assume the pile moves down relative to the soil; if the soil moves down relative to the pile — consolidating fill, a lowered water table, soft clay under a new embankment — the same interface shear reverses sign and drags the pile downward. That downdrag force subtracts from the useful capacity and adds to the axial load in the pile. PileCalc models this with explicit negative-friction zones along the shaft (with a reduction factor on the computed fₛ), plus zero-resistance zones for scour or seasonal effects. It is common enough, and misunderstood enough, that we cover it separately in our guide to negative skin friction and downdrag.

How capacity is verified

Static formulas of the kind above are estimates built on correlations, so practice wraps them in verification. Traditional allowable-stress design divides the computed ultimate by factors of safety (as above); LRFD codes apply resistance factors that grow with the quality of verification. On the ground, that verification is a static load test (loading a pile to plunge or to a settlement criterion and comparing against the prediction), or — hedged appropriately — dynamic monitoring during driving with signal matching. A load test that separates shaft and tip response (instrumented, or bi-directional) is the direct check on the skin-friction/end-bearing split, and it is the reason the t-z curves exist at all: they were fitted to instrumented tests. For uplift, only the shaft term (plus pile weight) is available — end bearing cannot act in tension — which is why uplift capacity is treated as its own check.

Common questions

What is the difference between a friction pile and an end-bearing pile?

A friction pile carries most of its axial load through shear along the shaft–soil interface, typical where no hard stratum is within reach. An end-bearing pile carries most of its load in compression at the tip, seated on dense soil or rock. The distinction is one of proportion — every embedded pile develops both resistances.

Why does skin friction mobilize before end bearing?

Shaft shear needs only a few millimetres of relative slip — about 0.5–1% of the pile diameter — to reach its peak, while the tip must compress the soil beneath it like a footing, requiring movement of roughly 10% of the diameter to mobilize fully. At working settlements the shaft therefore carries most of the load, regardless of the pile's ultimate-capacity classification.

What is the limiting depth for pile capacity in sand?

Unit side friction and unit end bearing in sand are computed from vertical effective stress, which increases with depth. Load tests show these unit resistances stop growing below a critical depth, so NAVFAC-style methods cap the effective stress at its value at a penetration of about 20 pile diameters. Deeper embedment adds shaft area but no further unit stress.

What is negative skin friction?

Negative skin friction, or downdrag, occurs when soil settles relative to the pile — typically consolidating fill or soft clay under new loading. The interface shear that normally resists the pile reverses direction and pulls it downward, adding axial load and subtracting from usable capacity. It is handled by excluding or reversing the affected shaft zones in the capacity calculation.

To see the split for your own profile, run the free pile capacity calculator — it reports side and tip resistance separately, layer by layer, with the limiting-depth cap and safety factors applied, in any browser.