Materials Science · Processing, structure and useful properties

Shape-making and internal structure

Follow rolling and polymer moulding through both external dimensions and internal arrangement. Use controlled bending and tensile evidence to separate geometric flexibility from modulus, plastic-flow resistance and directional material response.

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

Back to the chapter

After reading, you can

Explain how a shaping operation can change both dimensions and internal structure.
Distinguish elongated grains from crystal texture, and chain orientation from crystallinity.
Separate a thickness effect in elastic bending from a changed modulus or proof stress.
Design and interpret a controlled comparison that retains material direction, local state and test conditions.

A forming operation does two jobs

Rolling changes sheet thickness and may elongate grains, introduce defects and create texture. Moulding a polymer shapes a part while flow and cooling can orient chains and establish crystallinity. External dimensions and internal state evolve together.

A thin metal strip bends more readily after rolling. Did the material become less stiff, did its dimensions change, or did both happen? Shape is visible, but a shaping process also subjects the material to forces, flow and temperature histories. Its final state must be investigated as well as its outline.

Keep three descriptions separate: dimensions such as thickness; internal arrangement such as grains, defects and chains; and response under a specified test. They connect, but none substitutes for the others. A drawing that satisfies a thickness tolerance does not certify a modulus or a uniform microstructure.

Rolling changes grains, defects and sometimes their orientation distribution

In rolling, rotating rolls compress a workpiece through a gap and produce permanent deformation. In a simple dense-metal model, assume negligible volume change, no material loss and negligible width spread. Reducing thickness then requires an increase in length. This is a statement about the whole workpiece; it does not say that every grain deforms identically.

Within a crystalline metal, slip involves dislocation motion on particular crystal planes and directions. Neighboring grains constrain one another as the workpiece changes shape. Grains can become elongated, while dislocations accumulate and rearrange. In a cold-worked state that is not substantially restored during the operation, interacting dislocations can increase resistance to further plastic flow. This need not increase the small elastic stress–strain slope by the same factor.

Grain shape and crystal orientation answer different questions. A grain outline describes the region occupied by one crystal. Its lattice orientation describes how the atomic crystal directions lie relative to the specimen. Plastic deformation can rotate these orientations. A preferred, non-random distribution of crystal orientations across many grains is called crystallographic texture. It is not the roughness that a finger feels on the surface.

Elongated outlines alone do not measure texture. A grain’s long axis need not be one particular crystal direction; an orientation-sensitive diffraction or electron-diffraction measurement is needed to establish the distribution. The first sketch marks these two descriptions separately. The drawn orientations are illustrative, not predicted rolling angles or a measured texture.

The starting stock, reduction, temperature, direction and later operations matter. Recovery or recrystallisation during hot deformation can change the developing state; later annealing can change it again. The preceding history and temperature lessons explain those paths. Here, do not assign a universal defect state or modulus from the word “rolled”.

Original schematic side-view comparison of a dense metal before and after an ideal rolling shape change. The initial block is drawn 130 pixels long and 80 pixels deep; the final block is 260 pixels long and 40 pixels deep, preserving drawn area and representing constant width out of the page and constant volume. Four grain regions are drawn in each block. Grain outlines describe shape. Separate small crossed strokes indicate illustrative lattice orientations and do not measure a texture distribution. The final outlines are elongated and the chosen orientation symbols differ; neither drawn angles nor four grains constitute a prediction of actual rolling texture. The specimen long direction and thickness direction remain distinct. Grain shape, crystal orientation and dislocation structure require appropriate evidence; no pixel dimension is a real microstructural measurement.

Flowing a polymer can orient chains while cooling changes what persists

For a thermoplastic route, softened or molten polymer is driven into a mould cavity, then cooled into a removable part. A polymer consists of long connected molecular chains. Flow can deform their coil-like arrangements and create a preference in the directions of chain segments. After flow slows, molecular motion can relax some of that orientation; cooling can make relaxation slow enough that an oriented state remains.

Crystallinity is a different description: the amount of material in regions with regular crystalline packing. A semicrystalline polymer contains both crystalline and noncrystalline regions. Having chain segments preferentially directed one way does not, by itself, establish that their positions form a crystal. The second figure contrasts aligned but disordered segments with a region of regular packing.

A polymer capable of crystallising can develop crystalline regions during suitable cooling. An amorphous thermoplastic can retain orientation without doing so. Which structure persists depends on the polymer and local flow and cooling histories. Thus a mould cavity sets the intended shape but does not prescribe one unique chain orientation or crystalline fraction.

Nor does one observation settle the whole part. Flow and cooling can vary with location. Bur and Thomas’s optical-sensor study monitored filling, cooling and shrinkage during moulding of several resins; its sensing model explicitly considered crystallisation and glass formation. That is an example of investigating state during processing, not evidence that every moulding has the same internal structure.

Orientation can make response directional, meaning it differs with the direction of testing. A visible flow mark is not a direct measurement of chain orientation or crystallinity. Suitable structural evidence and tests along identified material directions are needed. No universal “along the flow is always stronger” rule follows from the process name.

Three original qualitative polymer arrangement sketches, not a processing prediction or a microscopy image. A shows curved chain-segment symbols with varied directions. B shows preferentially horizontal but irregularly placed curved segments: orientation without a periodic crystal arrangement. C shows a small shaded crystalline region containing regularly spaced parallel segment symbols, with curved noncrystalline segments outside. These are separate examples, not three compulsory stages, complete molecules or measured crystalline fractions. The drawing distinguishes a directional preference from regular packing. Flow and cooling can influence which arrangement persists, but neither the mould outline nor an apparent flow line measures chain orientation or crystallinity.

Separate their contributions

A thinner strip may bend more because of geometry even if its material becomes harder to deform plastically. To isolate a material response, use appropriately defined stresses, strains and test geometries. To evaluate the part, retain both the new geometry and the new material state.

Use a controlled elastic bending comparison. For a long, uniform rectangular strip, its second moment of area in the bending direction is I = bh³/12, with width b and depth h. In the elementary beam model its bending rigidity is EI, where E is the effective Young’s modulus in the relevant direction. The cubed thickness already explains why small dimensional changes can matter strongly.

For a simply supported strip with a centred transverse force F, the same model gives midpoint deflection δ = FL³/(48EI), where L is the support span. Here assume small linear-elastic bending, a uniform effective modulus, negligible shear deflection, no support indentation or slip, and no important residual curvature. The formula is supplied for this test geometry; it is not a prediction for arbitrary hand bending.

Compare coupons cut from initial and rolled stock, each with b = 10.0 mm and tested over L = 100 mm. Their depths are 2.00 and 1.00 mm. Independent tensile evidence below gives E = 100 GPa in the tested direction for both chosen states. At the same F = 3.20 N, the beam model gives 0.100 and 0.800 mm deflection. All values are original invented teaching data, not a named alloy’s properties or an actual rolling experiment.

The rolled coupon is eight times more compliant in this bending test because its I is one eighth as large. Compliance means displacement per applied force; stiffness is its reciprocal in this model. A separately supplied 0.2% proof stress rises from 150 to 240 MPa in the chosen cold-worked state. Greater resistance to a specified permanent-deformation endpoint can coexist with easier elastic bending of a thinner strip. Neither number is a universal cold-rolling result.

Chosen matched-width and matched-span bending comparison. The supplied E applies in the tested material direction. Both loads remain in the stipulated small elastic range.
Quantity	Initial-state coupon	Rolled-state coupon
b; h	10.0 mm; 2.00 mm	10.0 mm; 1.00 mm
Span L; force F	100 mm; 3.20 N	100 mm; 3.20 N
E	100 GPa	100 GPa
I = bh³/12	6.666… mm⁴	0.8333… mm⁴
Predicted midpoint δ	0.100 mm	0.800 mm
Separately supplied 0.2% proof stress	150 MPa	240 MPa

Use stress and strain to compare material response

A force–extension slope also includes geometry. To compare elastic material response, use engineering stress σ = F/A₀ and gauge strain ε = ΔL/L₀, with original cross-sectional area A₀ and original marked gauge length L₀. The initial linear slope E = Δσ/Δε is the modulus for the tested direction and conditions. Measure gauge extension rather than silently including machine or grip motion.

The table supplies two small elastic readings from separate representative tensile coupons, both with L₀ = 50.0 mm. Width is 5.00 mm; the initial and rolled thicknesses are 2.00 and 1.00 mm, so A₀ is 10.0 and 5.00 mm². Coupon preparation is assumed not to introduce a consequential state difference. Temperature, direction, loading rate and measurement method are matched. The readings are invented and are not the bending coupon being destructively reused.

Between 100 and 200 N, the initial coupon has Δσ = 10 MPa and Δε = 0.0001; the rolled coupon has Δσ = 20 MPa and Δε = 0.0002. Both give 100 000 MPa = 100 GPa. The rolled specimen’s greater extension at a given force is explained by its smaller area here, not by a lower modulus. The independently supplied proof stresses describe a different part of the response.

This normalisation does not erase all processing differences. Keep direction, location, thermal history and test timescale. If the state varies through thickness, a uniform scalar E may not describe every test; bending weights the outer regions differently from a uniform axial test. Do not machine away an oriented surface layer and then claim the remaining coupon represents the original whole thickness without checking.

Original invented small elastic tensile readings. All extensions are measured between the original 50.0 mm gauge marks; original area differs between the two states.
State	A₀ (mm²)	ΔL at 100 N (mm)	ΔL at 200 N (mm)
Initial	10.0	0.00500	0.0100
Rolled	5.00	0.0100	0.0200

Separate the question before choosing the control

To ask whether shaping changed a material response, compare suitable normalised measurements or prepared coupons with matched geometry, direction, location and conditions. To ask whether the finished part works, retain its actual dimensions and internal state. Making all test pieces the same shape can aid one comparison while removing a feature important to the other question.

For a moulded plate, compare equal-geometry coupons cut parallel and transverse to a recorded flow direction at comparable locations and depth. A difference in their stress–strain response establishes direction dependence for those conditions. It does not by itself measure a crystalline fraction or prove that orientation alone caused the difference; a structural measurement must test that explanation.

A good investigation can change one relevant input while holding the others fixed, then add a second controlled comparison. If both thickness and cooling history changed, two raw bending readings cannot assign a separate contribution to each. Record the competing explanations and choose an observation that can distinguish them. That turns “processing changed something” into a testable processing–structure–property argument.

Worked examples

A thinner strip bends more, although its proof stress rises

A rolled strip is thinner and bends more easily by hand. Does that establish a lower elastic modulus? Work the controlled 3.20 N bending comparison, then compare the separately supplied proof stresses.

The controlled test holds width, support span, force and the independently assessed E fixed. I_initial = 10 × 2³/12 = 20/3 mm⁴. I_rolled = 10 × 1³/12 = 5/6 mm⁴. The ratio I_rolled/I_initial is 1/8.
Use E = 100 000 N/mm². For the initial coupon, δ = (3.20 × 100³)/(48 × 100 000 × 20/3) = 0.100 mm. For the rolled coupon, replacing I by 5/6 gives δ = 0.800 mm. Thus the deflection ratio is 8 at the same load.
The separately supplied proof stress increases from 150 to 240 MPa, a ratio of 1.6. It concerns the defined 0.2% permanent-deformation endpoint, not the elastic slope. Increased dislocation interactions can raise this resistance in the chosen cold-worked state.
The controlled case demonstrates how changed geometry alone can explain easier elastic bending while proof stress rises. An uncontrolled hand test may also introduce changing span, force, plastic deformation or direction. It cannot identify E without resolving those differences.

No. Here E stays at 100 GPa while halving thickness increases elastic bending deflection eightfold. The chosen proof stress rises by 60%; stiffness, geometry and the permanent-deformation endpoint are distinct.

A raw extension difference disappears after normalisation

Use the two tensile readings from each state. Do they support a lower elastic modulus in the rolled state?

For the initial coupon, the force increment is 100 N and A₀ = 10.0 mm², so Δσ = 100/10.0 = 10 MPa. Gauge extension increases by 0.00500 mm over 50.0 mm: Δε = 0.00500/50.0 = 0.000100.
For the rolled coupon, Δσ = 100/5.00 = 20 MPa. Extension increases by 0.0100 mm, so Δε = 0.0100/50.0 = 0.000200.
E_initial = 10/0.000100 = 100 000 MPa; E_rolled = 20/0.000200 = 100 000 MPa. The force–extension slopes differ, but the normalised initial elastic slopes agree.
This is a result for the supplied measurements in one direction, not a rule that shaping never changes modulus. Texture, chain alignment, porosity or another state change can matter in other comparisons. Actual slope estimates also need adequate resolution, repeatability and uncertainty assessment.

Both chosen records give E = 100 GPa. The raw force–extension comparison alone would confuse a smaller area with a lower modulus.

Use matched geometry to identify directional response

Two representative moulded-polymer coupons have A₀ = 4.00 mm² and L₀ = 25.0 mm. At 24.0 N within their stipulated initial linear range, extension is 0.0500 mm parallel to the recorded flow and 0.0750 mm transverse to it. Conditions and sampling are otherwise matched. What can be inferred?

Both stresses are 24.0/4.00 = 6.00 MPa. Strains are 0.0500/25.0 = 0.00200 and 0.0750/25.0 = 0.00300.
With zeroed origins and the stipulated linear response, effective tensile moduli are 6.00/0.00200 = 3000 MPa and 6.00/0.00300 = 2000 MPa. The parallel value is 1.5 times the transverse value in these invented data.
The result identifies direction dependence for these specimens and conditions. It is consistent with an oriented structure but does not prove that orientation is the only cause, or determine crystallinity. Confirm the material arrangement with a suitable measurement.
Do not transfer the ratio to an arbitrary moulded polymer, another location, a long-duration load or a failure-strength comparison. These numbers describe a selected small elastic response, not all useful properties.

The chosen parallel and transverse moduli are 3.00 and 2.00 GPa. A directional material response remains after matching geometry; its microscopic cause and crystalline fraction require more evidence.

Try the reasoning

Can chain orientation matter in a moulded part? Two equal-geometry coupons show different initial slopes parallel and transverse to flow. Does this mean every chain became a crystal?
Show the answer and reasoning
Orientation can create directional response. Different slopes do not establish that every chain crystallised, or measure a crystalline fraction.
Direction and regular packing are different structural descriptions. Aligned noncrystalline segments are possible, and a semicrystalline polymer retains noncrystalline regions. Compare suitable structural evidence with the matched mechanical result instead of turning a modulus ratio into a crystalline percentage.
Which assessment retains geometry: material or component performance? A process gives the same tested E but halves a part’s rectangular thickness at unchanged width. Can its bending performance be assigned the old stiffness?
Show the answer and reasoning
Component performance retains the actual geometry. The old bending stiffness cannot be retained merely because E is unchanged.
In the stated small-beam model I changes by (1/2)³ = 1/8, so EI changes by the same factor. A normalised material comparison answers a different question from the finished part’s deformation under its actual supports and load.
In an ideal dense-metal rolling model, a marked piece starts 120 mm long, 20.0 mm wide and 2.00 mm thick. It ends 1.50 mm thick with the same width. Assume constant volume and no trimming. Find its final length and distinguish its percentage thickness reduction from its percentage length increase.
Show the answer and reasoning
Final length is 160 mm; thickness reduction is 25%, while length increase is 33⅓%.
Initial volume is 120 × 20 × 2 = 4800 mm³. Final length is 4800/(20 × 1.5) = 160 mm. Thickness reduction uses the original 2.00 mm denominator: (2 − 1.5)/2 = 0.25. Length increase uses 120 mm: (160 − 120)/120 = 1/3. The whole-piece balance assumes negligible width spread and volume change; it does not predict individual grain shapes or texture.
Another shaping route reduces strip thickness to 80% of its old value and independent matched testing gives E_new = 1.25 E_old. Keep width, span and transverse load fixed in the same beam model. Find the bending-rigidity and deflection ratios. Does increased E guarantee less bending?
Show the answer and reasoning
EI becomes 0.64 of the old value; deflection becomes 1/0.64 = 1.5625 times as large. The greater E does not overcome the chosen thickness reduction.
I scales as h³, so I_new/I_old = 0.8³ = 0.512. Multiply by the measured E ratio: 1.25 × 0.512 = 0.64. Deflection scales inversely with EI under the stated controls. This transfers both measured material and geometric changes; it does not attribute the E change to one unmeasured mechanism.
A polished rolled-metal section shows long grains. A report calls their long axes the measured crystal directions and concludes that all grains share one lattice orientation. What evidence is missing?
Show the answer and reasoning
An orientation-sensitive measurement is missing; grain outlines alone establish neither lattice directions nor a common orientation.
A grain’s occupied shape and the orientation of its atomic lattice are different descriptions. Crystallographic texture is a distribution over many grain orientations. Diffraction or a suitable orientation map can test it; the elongated outlines and the schematic crosses in the figure must not be treated as measured orientation data.
Two moulded plates have different thicknesses and cooling schedules; the thinner one bends farther under the same nominal force. A report assigns the difference entirely to lower crystallinity. Give a more informative comparison.
Show the answer and reasoning
First establish the actual geometry, supports, loading and material response; then compare suitably matched specimens while changing cooling history explicitly and independently assess structure.
Thickness can change bending compliance strongly even at fixed E. Cooling can also change the retained state, and the nominal force alone does not establish matching span or support. Equal-geometry, same-direction/location tests can help separate material response; a suitable crystallinity measurement is needed before naming its change as the cause. Machining must not silently remove a distinctive layer.

What to carry forward

A shaping process sets dimensions while also changing the state on which later operations act.
Grain shape is not crystal texture; chain orientation is not the same quantity as crystallinity.
Compare normalised material responses under stated conditions, then restore actual geometry when assessing the part.
Use controlled comparisons and structural evidence to distinguish causes; an easier bend does not identify a lower modulus.

References

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