Sandwich Panels: How a Weightless Core Makes a Skin 37× Stiffer

The I-beam idea, flattened into a plate

A sandwich panel is a structural I-beam turned inside-out and spread across a surface. Two thin, stiff faces (facesheets or skins) — aluminum, GFRP/CFRP laminate, steel, or plywood — are bonded to a thick, lightweight core — rigid polymer foam (PVC, PU, PET, PMI/Rohacell), balsa, or an aluminum/aramid honeycomb. The faces act as the flanges, carrying the bending stresses as an in-plane tension–compression couple; the core acts as the web, carrying the transverse shear and, critically, holding the faces apart at a fixed separation d.

The whole payoff lives in the second moment of area. For a symmetric sandwich of unit width b, face thickness t, and centroid-to-centroid spacing d = c + t (c = core thickness), the parallel-axis term dominates:

• Flexural rigidity: D = E_f·b·t·d²/2 + E_f·b·t³/6 + E_c·b·c³/12
• The first term (faces about the neutral axis) usually swamps the other two. The face's own bending (t³/6) and the soft core (E_c ≈ 20–200 MPa) contribute little.

Because D grows with d² while mass grows only linearly with core thickness (foam density ρ_c ≈ 40–200 kg/m³), doubling the core thickness roughly quadruples bending stiffness for a few percent more weight. That is the entire value proposition.

Faces take bending, the core takes shear

Sandwich mechanics splits the load path cleanly. Treat the faces as membranes and the core as a shear-only medium — the classical Allen/Plantema thin-face theory. The bending moment M is resisted by an equal-and-opposite force couple in the two faces separated by d:

• Face stress: σ_f = ± M / (b·t·d). A face is a flange at lever arm d/2 from the neutral axis.
• Core shear stress: τ_c = V / (b·d), essentially uniform through the thin core because the faces carry almost no shear.

This is why the core can be weak in-plane yet still work: it never sees the big bending stress, only the modest shear flow V/(b·d). A structural PVC foam with core shear strength τ_c ≈ 0.8–1.5 MPa and shear modulus G_c ≈ 20–40 MPa is entirely adequate under a facesheet seeing σ_f = 100–300 MPa.

The price of a soft core is shear deflection. Unlike a solid beam, a sandwich beam's tip or mid-span deflection has two additive parts:

• δ = δ_bending + δ_shear = (k_b·P·L³)/D + (k_s·P·L)/S, with shear rigidity S ≈ G_c·b·d²/c ≈ G_c·b·d.
• For short, deep panels the shear term can dominate — a sandwich is not an Euler–Bernoulli beam, and ignoring the G_c term under-predicts deflection by 20–50%.

Sizing a panel: the design procedure

A practical panel is sized against three limits simultaneously — stiffness, face strength, and core shear — then checked for local instabilities. A clean workflow:

• 1. Set the load and span. E.g. a simply-supported floor panel, span L = 1.5 m, uniform w = 5 kPa, width b = 1 m → M_max = wL²/8, V_max = wL/2.
• 2. Pick faces for strength. Solve σ_f = M/(b·t·d) ≤ σ_allow / n. For a 6061-T6 face (σ_y ≈ 275 MPa) with safety factor n = 1.5, choose t and target d.
• 3. Pick core thickness for stiffness. Compute D from the target deflection limit (often L/240 or L/360) using the two-term δ formula, iterating d.
• 4. Check core shear. τ_c = V/(b·d) ≤ τ_core / n. If it fails, use a denser core, not thicker faces.
• 5. Check the local modes below. Face wrinkling, dimpling, and indentation frequently govern before global strength does.

A representative aerospace floor panel: two 0.4 mm CFRP faces (E_f ≈ 70 GPa), 12 mm Nomex honeycomb (ρ_c ≈ 48 kg/m³, G_c ≈ 40 MPa), areal mass ≈ 2.3 kg/m², carries a 1.3 kN wheel-cart load over a 500 mm bay with L/300 deflection. The equivalent solid plate would be several times heavier.

Face wrinkling and the local failure zoo

The failure mode that separates sandwich design from ordinary plate design is face wrinkling: the compression face, elastically supported by the core, buckles into short wavelengths like a beam on an elastic foundation. It can occur at a face stress far below the material's yield. The Hoff–Mautner result gives a compact estimate:

• σ_wr ≈ 0.5·(E_f · E_c · G_c)^⅓ (isotropic core, thin face). The cube-root coupling means a soft core (low E_c, G_c) drops the wrinkling stress fast.
• Example: E_f = 70 GPa, E_c = 90 MPa, G_c = 35 MPa → σ_wr ≈ 0.5·(70,000·90·35)^⅓ ≈ 0.5·(2.2×10⁸)^⅓ ≈ 0.5·605 ≈ 300 MPa. A softer 40 kg/m³ foam might drop this below the working stress.

Other local modes to design against:

• Intracell dimpling (honeycomb only): the face buckles into each unbonded cell; critical stress ∝ E_f·(t/s)² where s is cell size — small cells resist it.
• Core shear crimping: a short-wavelength global shear buckle when the core shear modulus is too low; σ_crimp ≈ G_c·d/c.
• Local indentation: a concentrated or point load crushes the core beneath one face — the reason panels get local doublers, potted inserts, or edge close-outs at fasteners.
• Debonding / disbond: adhesive failure at the face–core interface, the dominant in-service degradation; propagates under peel and fatigue and is why NDT (tap test, thermography, ultrasound) matters.

Choosing cores and faces: the material trade space

The core sets almost everything except in-plane strength. Designers trade shear properties, density, cost, temperature, and moisture resistance:

• Rigid PVC/SAN foam (Divinycell, Airex): ρ ≈ 40–200 kg/m³, τ ≈ 0.5–3.5 MPa, G_c ≈ 13–75 MPa. Isotropic, thermoformable, closed-cell — the marine and wind-blade workhorse.
• PET foam: recyclable, good temperature stability, slightly lower properties per density; rising in wind blades.
• PMI foam (Rohacell): high performance, survives autoclave cure at 180 °C and pressure — aerospace interiors and radomes.
• Aluminum / Nomex honeycomb: the best stiffness-to-weight of all (ρ ≈ 30–130 kg/m³) but anisotropic, prone to moisture ingress and dimpling, and needs film adhesive and a good bond line.
• End-grain balsa: very high compressive/shear strength per cost, natural, but heavier and moisture-sensitive.

Faces are chosen for in-plane stiffness and environment: GFRP (cheap, marine), CFRP (aerospace, E ≈ 70–140 GPa), aluminum (dentable but tough, easy to inspect), and steel (structural insulated panels, refrigerated trailers). A governing rule: keep face thickness ratios modest — the classic guidance is d/t > ~5.77 and core-to-face stiffness such that the thin-face assumptions hold; extremely thin faces on soft cores invite wrinkling.

Where sandwich panels earn their keep

Anywhere bending stiffness per kilogram is the currency:

• Aircraft: cabin floors, galley/lavatory monuments, control surfaces (rudders, flaps, spoilers), radomes, and engine cowls — mostly CFRP/Nomex honeycomb, built to MIL-HDBK-23 and OEM specs, inspected relentlessly for disbonds and water ingress.
• Marine: hulls, decks, and bulkheads of racing yachts and patrol craft — GFRP or CFRP faces on PVC/PET foam or balsa, vacuum-infused; the whole IMOCA/America's-Cup structural philosophy is sandwich.
• Wind energy: the shear webs and shells of 80 m+ blades are foam- or balsa-cored sandwich, sized against buckling of the huge low-pressure-side panels.
• Building & transport: Structural Insulated Panels (SIPs) — OSB faces on EPS/PU foam — and metal-faced insulated panels for cold stores and cladding, doing double duty as structure and thermal barrier (foam gives k ≈ 0.02–0.04 W/m·K).
• Rail & road: refrigerated trailer bodies, high-speed-train floors, and RV walls.

The common thread: a flat or gently curved surface that must be stiff, light, and often thermally or acoustically insulating at once — a combination a monolithic plate cannot touch.

Limits, best practice, and what bites you

Sandwich construction is unforgiving of detailing and moisture. The best-practice checklist that separates a durable panel from a warranty claim:

• Design the edges and joints, not just the field. Loads enter through fasteners and inserts, which crush the core; use potted inserts, ramped core-to-solid transitions, and edge close-outs. A bare fastener through a foam core is a failure waiting to happen.
• Respect the shear term. Always size deflection with δ = bending + shear; for short-span, deep panels the G_c term dominates and Euler theory lies.
• Keep water out. Honeycomb and open-cell foam wick water; freeze–thaw then blows the bond. Closed-cell foam, sealed edges, and drainage are non-negotiable in marine/aero service.
• Verify by test. Qualify cores/adhesives per ASTM C393 (flexure), C273 (core shear), C365 (flatwise compression), C297 (flatwise tension/bond), and reduce data per D7250. Wrinkling and indentation are hard to predict — test them.
• Mind fire, temperature, and creep. Polymer foams soften and creep near 60–80 °C and off-gas in fire; SIPs and interiors carry fire and smoke ratings for exactly this reason.
• Fatigue lives in the interface. Cyclic peel and core shear drive disbond growth; design bond-line stresses low and inspect on a schedule.

Get the geometry (d/t), the core shear check, and the wrinkling margin right — and detail the edges — and a sandwich panel delivers stiffness a solid plate can only dream of at a fraction of the mass.

A monolithic aluminum plate vs a foam-core sandwich of equal mass per unit area (same alloy skins, structural PVC foam core)

Property	Solid plate	Sandwich panel	Ratio / note
Total thickness	1.0 mm	4 mm (0.5+3+0.5)	core does the leveraging
Areal mass	~2.7 kg/m²	~2.7 kg/m² + ~0.2 kg core	core adds little
Flexural rigidity D	1× (baseline)	≈ 37×	scales with d²/t²
Bending strength	1×	≈ 9×	face stress = M/(t·d·b)
Governing failure	Face yield	Core shear / face wrinkle	new modes appear
Cost & inspection	Low, easy NDT	Higher, debond-prone	trade-off