Materials Science · Composites, fibres and interfaces

Direction dependence and mixture estimates

Useful background: Units, loads, stress and strain; Atoms, bonding, structure and defects; Processing, structure and useful properties. Each worked case explains the quantities it uses.

Back to the chapter

After reading, you can

Apply a simple mixture modulus only with its strain assumptions.
Derive parallel and series estimates from force balance and compatible extensions.
Convert mass fraction to volume fraction using constituent densities.
Calculate constituent loads and distinguish directional stiffness from strength or a unique architecture diagnosis.

A parallel arrangement shares extension

For continuous aligned constituents loaded along their length with good bonding, an equal-strain approximation gives E = V₁E₁ + V₂E₂. Volume fractions sum to one. This is a model for a specified architecture and linear-elastic regime.

Imagine two strips joined to the same rigid end pieces and pulled along their length. Here guides constrain the end pieces to remain parallel without rotating and to translate only axially. The guides may supply moment reactions while the prescribed axial force remains the sum of the two strip forces. Both strips start with length L, so moving one end piece by ΔL imposes the same axial strain ε = ΔL/L on both. The stiffer strip needs more force for that same strain. Equal extension does not mean equal force.

We will compare two ideal one-dimensional arrangements of fictional constituents with axial moduli E₁ = 100 GPa and E₂ = 10 GPa. A modulus relates small axial stress and strain through σᵢ = Eᵢεᵢ, where subscript i identifies the constituent. Both are stipulated to remain linear elastic during the examples. These constants are teaching inputs, not measured properties of named fibres or resins.

In this model, connections transmit the required force without slipping, and we neglect lateral constraint and Poisson coupling, the coupling between axial and sideways deformation. The real response can depart from that model when ends, interfaces, voids, damage or three-dimensional constraint matter. The elastic-modulus, stress and strain preparation lessons supply the loading definitions.

For parallel strips, add the forces

Let A₁ and A₂ be the cross-sectional areas and A = A₁ + A₂ the complete area. With equal initial lengths, the first constituent’s volume fraction is f = A₁/A; the other fraction is 1 − f. A volume fraction is a dimensionless volume divided by complete material volume. In these two solid, void-free constituents, the two fractions sum to one. The opening formula uses V₁ and V₂ for these volume fractions: V₁ = f and V₂ = 1 − f.

The strip forces are F₁ = σ₁A₁ = E₁εA₁ and F₂ = E₂εA₂. Force balance gives F = F₁ + F₂. Dividing by complete area A gives average stress σ = [fE₁ + (1 − f)E₂]ε. Therefore the effective axial modulus is E_parallel = fE₁ + (1 − f)E₂. The weighting follows from area and common strain; it is not a rule to average every property of a mixture.

At f = 0.500, E_parallel = 0.500 × 100 + 0.500 × 10 = 55.0 GPa. Under a shared strain of 0.000600, constituent stresses are 60.0 MPa and 6.00 MPa. Their area-weighted mean is 33.0 MPa. For total A = 10.0 mm², the two 5.00 mm² strips carry 300 N and 30.0 N, which sum to 330 N. One MPa equals one N/mm².

A long, well-bonded, aligned fibre composite loaded along its fibres can sometimes be represented usefully by a shared-strain estimate away from the ends. The actual load-transfer length, fibre distribution and boundary conditions still matter. The separate load-transfer lesson explains why a short or poorly coupled fibre may not develop that common-strain response.

Other arrangements share load differently

An ideal series arrangement has equal stress and gives 1/E = V₁/E₁ + V₂/E₂. These estimates illustrate why direction and connectivity matter. They are not exact universal formulas for every transverse fibre composite or damaged laminate.

Now put the two pieces end to end, with the same cross-sectional area A. Let their initial lengths be L₁ and L₂ and their combined length L = L₁ + L₂. The first volume fraction is now f = L₁/L. With no other axial force applied at the joint, the same force F passes through both pieces. Equal areas then give the same axial stress σ = F/A. Equal stress does not mean equal extension or strain.

The local strains are ε₁ = σ/E₁ and ε₂ = σ/E₂. Since the end-to-end extensions add, ΔL = ε₁L₁ + ε₂L₂. Divide by L to obtain average strain ε = σ[f/E₁ + (1 − f)/E₂]. Thus 1/E_series = f/E₁ + (1 − f)/E₂. The reciprocal modulus is compliance: a larger compliance means a larger strain at the same stress.

At f = 0.500, the compliance is 0.500/100 + 0.500/10 = 0.0550 GPa⁻¹. Inverting gives E_series = 18.1818 GPa. The compliant piece supplies most of the total extension even though its volume equals the stiff piece’s volume. The retained graph shows the parallel estimate only; it does not display or predict this series value.

For an actual continuous-fibre material loaded across its fibres, nearby matrix regions experience different stresses and strains. Calling that direction “series” does not make its fields uniform. Use these formulas to understand two specified ideal arrangements and to test assumptions; an actual transverse modulus requires an appropriate model or measurement.

For a 100 GPa constituent and a 10 GPa constituent, the equal-strain mixture line runs from 10 GPa at fraction zero to 100 GPa at fraction one. At fraction 0.5 its value is 55 GPa.

The same overall force exposes different load sharing

Use complete length 100.0 mm, complete area 10.0 mm² and force 330 N for both ideal arrangements. Each contains equal constituent volumes. For parallel strips, ε = 33.0 MPa/55,000 MPa = 0.000600 and ΔL = 0.0600 mm. Both strips extend by that amount. Their unequal forces add to the applied force.

For series pieces, each is 50.0 mm long and both carry 330 N, so both have stress 33.0 MPa. Their strains are 33.0/100,000 = 0.000330 and 33.0/10,000 = 0.00330. Their extensions are 0.0165 mm and 0.1650 mm, summing to 0.1815 mm. Dividing total extension by 100.0 mm gives ε = 0.001815; 33.0/0.001815 = 18,181.8 MPa checks the compliance calculation.

These are deliberately different geometrical arrangements with the same constituent quantities. They are not two measured directions of one real fibre coupon. For these positive scalar moduli, their arithmetic and harmonic means give useful comparison values; no claim about universal bounds on a general three-dimensional Young’s modulus follows from this elementary calculation.

Original ideal one-dimensional comparison: equal volumes of 100/10 GPa constituents, L = 100 mm, A = 10 mm² and F = 330 N. Extensions are calculated, not measured.
Quantity	Parallel strips	Series pieces
Shared local quantity	Strain 0.000600	Stress 33.0 MPa
Constituent forces	300 N; 30.0 N	330 N through each
Local extensions	0.0600 mm for each 100 mm strip	0.0165 mm + 0.1650 mm
Complete extension	0.0600 mm	0.1815 mm
Effective modulus	55.0 GPa	18.1818 GPa

Guides keep the parallel end pieces nonrotating and parallel while they translate axially; moment reactions are allowed and the axial resultant remains the sum of the two strip forces. Original two ideal one-dimensional arrangements, not two measured directions of one coupon. Blue outlines mark the 100 GPa constituent and orange outlines the 10 GPa constituent. In parallel, two 100 mm strips each have 5 mm squared area; rigid end pieces impose the same strain 0.000600 and extension 0.0600 mm. Forces 300 and 30 N add to 330 N, giving 55 GPa effective modulus. In series, two 50 mm pieces share area 10 mm squared and 330 N tensile force, giving common stress 33 MPa. Strains 0.000330 and 0.00330 produce extensions 0.0165 and 0.1650 mm, which add to 0.1815 mm over 100 mm; effective modulus 18.1818 GPa. The opposing external tensile arrows show force direction; force-arrow lengths are schematic and do not encode magnitude. Strip or piece dimensions and exaggerated separations explain connections, not an actual fibre field, measured extension picture or strength prediction.

Mass fractions require a density conversion

The formula uses constituent volumes. A heavy constituent can provide most of the mass while occupying half the volume. To convert, let m₁ and m₂ be masses, ρ₁ and ρ₂ constituent densities, and w = m₁/(m₁ + m₂) the first mass fraction. Each volume is mᵢ/ρᵢ. Therefore f = (w/ρ₁)/[w/ρ₁ + (1 − w)/ρ₂]. Use densities in matching units so the ratio is dimensionless.

For fictional densities ρ₁ = 2400 kg/m³ and ρ₂ = 1200 kg/m³, a 1.00 cm³ specimen with f = 0.500 contains 0.500 cm³ of each constituent. Densities are 2.40 and 1.20 g/cm³, so masses are 1.20 and 0.600 g. Total mass is 1.80 g, complete density is 1800 kg/m³, and w = 1.20/1.80 = 2/3.

Substituting w = 2/3 directly into the parallel formula would give 70.0 GPa. That overestimates the correct 55.0 GPa by 15.0 GPa, or 27.27% relative to 55.0 GPa. Converting w with the two densities recovers f = 0.500 and removes the error. Equal mass and equal volume fractions coincide for equal densities, apart from the trivial pure-constituent endpoints.

For this void-free additive-volume model, complete density is ρ = f ρ₁ + (1 − f)ρ₂. Unknown void content, uncounted constituents or changed density can break an inference from measured mass to the two fractions. Density and properties relative to mass develops that account. A density conversion is bookkeeping, not a stiffness law for a defective material.

Specify the direction and ask what a measurement identifies

A directional modulus describes response to loading in a specified direction. In a real anisotropic composite, turning the load can involve shear and coupled normal strains. One cannot generally obtain the 45-degree modulus by averaging an along-fibre value and an across-fibre value. This lesson does not introduce the full elastic stiffness tensor; it identifies why two scalar averages do not replace it.

A measured or supplied effective modulus alone need not identify composition and architecture simultaneously. Within our restricted two-arrangement model, 55.0 GPa fits parallel f = 0.500. It also fits series f = 10/11: f/100 + (1 − f)/10 = 1/55, all in GPa⁻¹. The same stiffness can therefore have two different explanations if the arrangement and fraction are both unknown.

With independently known constituent densities and no voids or other phases, these two candidates would have different complete densities: 1800 kg/m³ for the half-and-half parallel case and 2290.91 kg/m³ for the series f = 10/11 case. A complementary density observation could discriminate between those specified candidates. Without those assumptions it is additional evidence, not a unique architecture detector.

Finally, stiffness is the slope of an elastic response, whereas strength concerns failure. Neither 55.0 nor 18.1818 GPa supplies constituent breaking limits, interface debonding thresholds, fatigue behaviour or fracture resistance. A high stiffness estimate cannot qualify a real component. The related composite failure lesson addresses failure paths separately.

Worked examples

Work through the case

An ideal equal-strain composite has 50% material at 100 GPa and 50% at 10 GPa. Estimate its modulus.

Use volume fractions 0.5 and 0.5.
E = 0.5 × 100 + 0.5 × 10.
The result retains GPa.
The equal-strain condition is justified here by equal-length ideal parallel strips sharing end displacement. Forces add in proportion to area and modulus.
At a shared strain 0.000600 and total area 10.0 mm², calculate F₁ = 100,000 × 0.000600 × 5.00 = 300 N and F₂ = 10,000 × 0.000600 × 5.00 = 30.0 N. Their sum 330 N gives average 33.0 MPa and 33.0/0.000600 = 55,000 MPa.

55 GPa under the stated idealisation. The stiff constituent carries ten times the force despite equal volume.

Check the series estimate by adding actual extensions

Two ideal equal-area pieces, each 50.0 mm long, have moduli 100 and 10 GPa. A 330 N force passes through their 10.0 mm² section. Find total extension and effective modulus.

The common stress is 330/10.0 = 33.0 MPa. Use 100,000 and 10,000 MPa for the constituent moduli.
Extensions are (33.0/100,000) × 50.0 = 0.0165 mm and (33.0/10,000) × 50.0 = 0.1650 mm.
Total extension 0.1815 mm over 100.0 mm gives strain 0.001815.
Divide 33.0 MPa by 0.001815 to obtain 18,181.8 MPa = 18.1818 GPa. This matches the inverse of 0.0550 GPa⁻¹.

0.1815 mm and 18.1818 GPa. The compliant half supplies ten times the extension of the stiff half.

Turn a mass specification into the correct volume input

A fictional two-constituent mixture is two-thirds of its mass at 2400 kg/m³ and one-third at 1200 kg/m³. Under the stipulated parallel conditions, does its modulus equal 70 GPa?

Choose a convenient total mass 1.80 g: constituent masses are 1.20 and 0.600 g.
Divide by 2.40 and 1.20 g/cm³ to obtain 0.500 and 0.500 cm³. The first volume fraction is 0.500.
The parallel estimate is 0.500 × 100 + 0.500 × 10 = 55.0 GPa.
The incorrect 70.0 GPa result treats a mass fraction as an area/volume fraction. Its relative error is 15.0/55.0 = 27.27%.

55.0 GPa. Density conversion is required before applying this volume-weighted relation.

One modulus can fit two model explanations

A record supplies 55.0 GPa but omits fraction and arrangement. Restrict possible arrangements to the two ideal models and constituents to 100/10 GPa. Can the record identify a half-and-half parallel arrangement uniquely?

Parallel: 55 = 100f + 10(1 − f), so f = 45/90 = 0.500.
Series: 1/55 = f/100 + (1 − f)/10 = 0.100 − 0.090f.
Solve f = (0.100 − 1/55)/0.090 = 10/11, or about 0.909091.
Both fractions lie between zero and one. The single response is compatible with two different candidate accounts; an independent composition/geometry or suitably qualified density observation is needed.

No. Parallel f = 0.500 and series f = 10/11 both fit 55.0 GPa within their stated assumptions.

Try the reasoning

Could mass fractions be substituted unchanged?
Show the answer and reasoning
Not generally. With the fictional densities 2400/1200 kg/m³, mass fraction 2/3 becomes volume fraction 1/2.
Convert using constituent densities. Convert each mass to volume using its own density before dividing by total volume.
Does 55 GPa establish composite strength?
Show the answer and reasoning
No.
Elastic response does not specify failure. Forces carried by constituents and interface or constituent failure limits would need separate evidence. The model stipulates linear response only over the example’s strains.
For a new f = 0.250 by volume, calculate both ideal moduli using the same 100/10 GPa constituents. Does the series value equal the arithmetic mean?
Show the answer and reasoning
E_parallel = 32.5 GPa; E_series = 1/(0.250/100 + 0.750/10) = 12.9032 GPa.
Parallel forces add at common strain; series extensions add at common stress. The 0.0775 GPa⁻¹ compliance must be inverted. The constituent quantities alone do not select the arrangement.
Use the f = 0.250 parallel arrangement, total area 8.00 mm² and shared strain 0.00100. Find each constituent force and the complete average stress.
Show the answer and reasoning
F₁ = 200 N, F₂ = 60.0 N; total 260 N and average 32.5 MPa.
Areas are 2.00 and 6.00 mm²; local stresses are 100 and 10 MPa. Multiply each stress by its own area. The average 32.5 MPa divided by 0.00100 is 32.5 GPa. Linear behaviour at this strain is an explicit teaching stipulation.
A new equal-area series pair has total length 80.0 mm, with 20.0 mm of 100 GPa material and 60.0 mm of 10 GPa material. Both carry stress 20.0 MPa. Find complete extension.
Show the answer and reasoning
0.00400 + 0.120 = 0.124 mm; complete strain 0.00155 and effective modulus 12.9032 GPa.
Use each piece’s own length and compliance: ΔLᵢ = (σ/Eᵢ)Lᵢ. The 0.250 volume fraction here comes from length fraction, not cross-sectional area fraction.
A new mixture has equal masses of the 2400 and 1200 kg/m³ constituents. What first volume fraction and ideal parallel modulus follow?
Show the answer and reasoning
f = 1/3 and E_parallel = 40.0 GPa.
For equal mass, the denser constituent occupies half the volume of the other. Normalising volumes 1/2400 and 1/1200 gives 1/3 and 2/3. Equal mass does not mean f = 1/2.
A report gives along-fibre 55 GPa and across-fibre 18 GPa, then claims 36.5 GPa at 45 degrees because it averaged them. Is that deduction justified?
Show the answer and reasoning
No. Those two scalar values do not determine the off-axis response.
Shear stiffness and coupling information, as well as actual architecture and boundary conditions, may enter. Neither elementary one-dimensional arrangement supplies a general 45-degree modulus.
Under the restricted void-free model and known 2400/1200 kg/m³ densities, a 55 GPa candidate also has independently established density 1800 kg/m³. Does that agree with the series f = 10/11 candidate?
Show the answer and reasoning
No. That candidate predicts 2290.91 kg/m³; 1800 kg/m³ corresponds to f = 0.500.
Use ρ = 1200 + 1200f, so f = (1800 − 1200)/1200 = 0.500. This comparison discriminates the two stated candidates; unknown porosity, other phases or uncertain densities would reopen the inference.

What to carry forward

State the shared-strain or shared-stress assumption.
Effective stiffness is not strength.
Parallel force balance gives a volume-weighted modulus when strains match; series extension balance gives a volume-weighted compliance when stresses match.
Use area fractions for equal-length strips and length fractions for equal-area pieces; explain why they equal the volume fractions.
Convert mass to volume using constituent densities. A composition percentage needs its basis.
Treat direction, architecture and boundary conditions as model inputs. One stiffness value cannot generally identify them all.

References

Teaching resources and research supporting this lesson’s explanations and examples.