Composite materials — fiber-matrix, ply properties, laminate theory: Visual walkthrough

Before we start, three plain-word anchors we will reuse constantly:

Step 1 — Two blocks, glued, seen as one

WHAT. We shrink the whole composite down to one tiny representative cube of side $1$ : one block of fiber welded to one block of matrix. This is the smallest chunk that still "looks like" the material.

WHY. If we understand one cube, we understand the slab — the slab is just millions of copies. Choosing side $1$ is a trick: it makes "volume fraction" and "thickness fraction" the same number, so $V_{f}$ literally is the width of the fiber block.

PICTURE. In the figure the fiber block (red) sits beside the matrix block (black). Its width is $V_{f}$ ; the matrix width is $V_{m}$ ; together they fill the unit cube.

Step 2 — Pull ALONG the fibers: parallel springs

WHAT. Pull the cube left-to-right, along the fiber axis (call it direction 1). Both blocks are gripped by the same two end plates, so they are forced to stretch the same amount.

WHY. Perfect glue (we assume perfect adhesion) means the fiber and matrix cannot slide past each other. Same original length, same stretch ⇒ same strain. This is exactly what two springs mounted side by side between the same two walls do — pull the wall, both stretch equally.

$fiber strain ϵ_{f} = matrix strain ϵ_{m} = common strain ϵ$

PICTURE. The red fiber block and black matrix block are two parallel springs pinned between the same moving plate. The plate moves one distance; both springs feel that same stretch.

Each block answers with its own stress, because each has its own stiffness:

$σ_{f} = E_{f} ϵ σ_{m} = E_{m} ϵ$

$E_{f}$ = fiber modulus (huge, ~ $230$ GPa) → fiber pushes back hard.
$E_{m}$ = matrix modulus (small, ~ $3$ GPa) → matrix barely resists.

Step 3 — Add the forces, not the stresses: birth of $E_{1}$

WHAT. The total force pulling the end plate is the fiber's force plus the matrix's force. Force = stress × area, and with a unit cube the areas are just $V_{f}$ and $V_{m}$ .

WHY. Parallel springs share the load by adding their forces. We convert stresses to forces (multiply by the area each block presents), add them, then convert back to an "average stress" over the whole face.

$σ_{comp} = fiber’s share σ_{f} V_{f} + matrix’s share σ_{m} V_{m}$

Substitute $σ_{f} = E_{f} ϵ$ and $σ_{m} = E_{m} ϵ$ , then divide both sides by the common $ϵ$ (that division is the definition $E_{1} = σ_{comp} / ϵ$ ):

$E_{1} = E_{f} V_{f} + E_{m} V_{m}$

$E_{f} V_{f}$ = the stiff fiber, weighted by how much of the cube it fills.
$E_{m} V_{m}$ = the soft matrix, weighted by its share.
The result is a plain weighted average — the Rule of Mixtures. Because $E_{f}$ dwarfs $E_{m}$ , the fibers dominate: they do almost all the work.

PICTURE. The bar chart shows the two contributions stacked; the tall red slab ( $E_{f} V_{f}$ ) towers over the thin black one ( $E_{m} V_{m}$ ).

Step 4 — Now pull ACROSS the fibers: series springs

WHAT. Rotate the pull $90°$ — pull top-to-bottom, across the fibers (direction 2). Now the load must travel through the fiber block and then through the matrix block, one after the other.

WHY. Stacked blocks in the load path are series springs. The rule flips:

Same stress passes through both (the same force squeezes through each layer's face) — like a single force pushed through two springs in a line.
Strains add up (total stretch = fiber's stretch + matrix's stretch).

$σ_{f} = σ_{m} = σ ϵ_{total} = ϵ_{f} V_{f} + ϵ_{m} V_{m}$

Here $ϵ_{f}$ and $ϵ_{m}$ are weighted by $V_{f}, V_{m}$ because the fiber layer is only $V_{f}$ tall, so its contribution to the total stretch is its strain scaled by its height.

PICTURE. The two blocks are now end-to-end springs; one force line runs straight through both. The soft matrix (black) does most of the stretching because it is the weak link.

Step 5 — The inverse rule: birth of $E_{2}$

WHAT. Write each block's strain from its own stress ( $ϵ_{f} = σ / E_{f}$ , $ϵ_{m} = σ / E_{m}$ ), plug into the sum, and divide by $σ$ .

WHY. We want $E_{2} = σ / ϵ_{total}$ . Since strains add and stress is shared, we get the reciprocals adding — the tell-tale signature of series elements.

$ϵ_{total} = σ (\frac{V _{f}}{E _{f}} + \frac{V _{m}}{E _{m}}) ⟹ \frac{1}{E _{2}} = \frac{V _{f}}{E _{f}} + \frac{V _{m}}{E _{m}}$

$\frac{V _{f}}{E _{f}}$ = the fiber's floppiness (compliance) share — tiny, because $E_{f}$ is huge.
$\frac{V _{m}}{E _{m}}$ = the matrix's floppiness share — large, because $E_{m}$ is small. This term dominates the sum.

The weakest link (matrix) controls a series chain. So $E_{2}$ collapses to matrix-like values.

PICTURE. A "floppiness" bar chart — now the black matrix bar is the tall one. The material's crosswise stiffness is hostage to the soft phase.

Step 6 — The edge cases: what happens at the extremes?

WHAT. Test the two formulas at inputs where the answer is obvious, to prove they are trustworthy.

WHY. A formula you cannot break at the extremes is a formula you can trust in between. We sweep $V_{f}$ from $0$ (all matrix) to $1$ (all fiber).

$V_{f} = 0$ (pure matrix): $E_{1} = E_{m}$ and $E_{2} = E_{m}$ . Both give the matrix — correct, there are no fibers to matter.
$V_{f} = 1$ (pure fiber): $E_{1} = E_{f}$ and $E_{2} = E_{f}$ . Both give the fiber — correct.
In between: $E_{1}$ rides a straight line from $E_{m}$ up to $E_{f}$ (weighted average). $E_{2}$ hugs the bottom, only lifting off the matrix value near $V_{f} \to 1$ — because until fibers form an unbroken column across the load, the soft matrix still gates every load path.

$E_{1} (0) = E_{2} (0) = E_{m}, E_{1} (1) = E_{2} (1) = E_{f}$

PICTURE. Both curves plotted against $V_{f}$ . They meet at both ends (as they must) and spread hugely in the middle. The red $E_{1}$ line is straight and high; the black $E_{2}$ curve crawls along the floor.

The one-picture summary

The entire derivation is one fork in the road: which way do you pull? Pull along → parallel springs → forces add → arithmetic average $E_{1}$ (fiber wins). Pull across → series springs → stretches add → reciprocal average $E_{2}$ (matrix loses).

Recall Feynman retelling — say it like a story

Imagine gluing a stiff steel wire into a rubber block. Pull it end-to-end along the wire: the wire and rubber are clamped between the same two hands, so both stretch the same tiny amount — but the wire fights back enormously and the rubber shrugs. Their pushbacks add up, each counted by how much room it takes up. That sum, per unit stretch, is $E_{1}$ , and it's basically the wire — a plain average that the stiff phase wins. Now pull sideways, across the wire: the force has to go through the wire layer and then through the rubber layer, one behind the other. The same force squeezes through both, but the rubber stretches a lot and the wire barely at all — the total give is dominated by the floppy rubber. Adding floppinesses instead of stiffnesses gives the reciprocal formula for $E_{2}$ , and the soft rubber drags it down. Check the ends: no wire → it's all rubber; all wire → it's all wire; both formulas agree there. So one material, two totally different stiffnesses, all from asking which way am I pulling?

Cloze checkpoints:

Along fibers, fiber and matrix share the same

strain (they stretch equally — parallel springs).

Across fibers, fiber and matrix share the same

stress (the load passes through both — series springs).

The longitudinal modulus is a weighted average of

the stiffnesses

E

(Rule of Mixtures).

The transverse modulus is a weighted average of

the compliances

1/ E

(inverse Rule of Mixtures).

V_{f} \to 1

, both

E_{1}

and

E_{2}

approach

E_{f}

(all fiber, no matrix left).

Where this goes next: stacking these plies at angles builds the full laminate; picking fiber vs metal is material selection; a soft-core version is sandwich structures; these moduli feed finite element models and launch load sizing, while low expansion matters for thermal design and every gram counts in the mass budget.