2. Surface Topography, Metrology and Surface Integrity

Introduction and context

Tribological performance is governed not only by the nominal geometry of two contacting bodies but also by the microscopic geometry and physical condition of their surfaces. A machined component may appear smooth to the eye and satisfy its dimensional tolerances, yet its surface may contain asperities, grooves, waviness, residual stresses, altered microstructure, or embedded particles that strongly influence friction, wear, lubrication, sealing, fatigue, and thermal behaviour. Surface topography is therefore a functional engineering variable rather than a purely cosmetic attribute.

Surface metrology provides the methods by which this topography is measured, represented, and interpreted. Surface integrity extends the description beyond geometry to include near-surface material characteristics produced by manufacturing, such as hardness gradients, plastic deformation, residual stress, microcracks, phase transformation, and chemical alteration. In advanced tribology, the central question is not simply whether a surface has a low roughness value. It is whether the measured surface, under a specified operating condition, provides the required real area of contact, lubricant retention, load-carrying capacity, wear resistance, and fatigue life.

This chapter develops the language and methods needed to relate manufacturing processes and measurement uncertainty to functional performance. It introduces profile and areal parameters, filtering, bearing-area curves, contact and optical instruments, multiscale descriptions, and the interpretation of results in engineering decisions.

Learning objectives

After studying this chapter, the reader should be able to:

1. distinguish form, waviness, roughness, lay, texture, and surface integrity;

2. explain how contact and optical instruments acquire surface data and identify their principal limitations;

3. interpret common profile and areal roughness parameters, including their dependence on filtering and sampling;

4. use bearing-area curves and related parameters to assess contact, lubrication, and wear behaviour;

5. explain why surface descriptions are scale-dependent and how multiscale or fractal approaches add information;

6. relate manufacturing processes to characteristic topographies and subsurface integrity;

7. evaluate uncertainty, artefacts, and the validity of a metrology result;

8. select an appropriate measurement and interpretation strategy for a tribological application.

The language of an engineering surface

A real surface is commonly represented as a height field, z(x,y), relative to a reference plane. A line extracted from this field gives a profile z(x). The observed profile is not a single physical phenomenon. It is a superposition of nominal form, waviness, roughness, and measurement noise. Form represents large-scale deviations from the intended geometry, such as barrel shape or out-of-roundness. Waviness consists of more widely spaced undulations, often caused by machine-tool vibration, workpiece deflection, runout, or thermal instability. Roughness refers to shorter-scale irregularities generated by cutting edges, abrasive grains, material tearing, or local deformation.

The distinction is operational rather than absolute. A feature classified as waviness at one measurement scale may be classified as roughness at another. Consequently, roughness values have meaning only when the measurement direction, cutoff or filter, evaluation length, instrument, and processing method are stated.

Lay is the predominant direction of surface marks. Turning generally produces circumferential or helical lay, grinding often produces directionally organised scratches, and honing or plateau finishing produces cross-hatched textures. Texture is the broader spatial description of directional and statistical features, including isotropy, anisotropy, periodicity, skewness, and spatial correlation. Two surfaces with the same Ra may therefore behave differently if one has a strongly directional lay and the other is isotropic.

The engineering surface also has a material dimension. Surface integrity includes the condition of the surface and subsurface region after processing. Relevant attributes include microhardness, residual stress, metallurgical phase, grain alteration, plastic flow, tensile damage, burns, pores, smeared material, and foreign-particle contamination. A polished surface with tensile residual stress and a disturbed subsurface may be less reliable in fatigue or wear than a visibly rougher surface with favourable compressive stress and an undamaged microstructure.

From profile numbers to functional descriptors

For a sampled profile containing n height values z_i, the arithmetic mean roughness is

Ra = (1/n) Σ |z_i|.

Ra is easy to communicate and useful for process control, but it discards sign, spacing, peak shape, and spatial arrangement. The root-mean-square roughness is

Rq = [ (1/n) Σ z_i² ]^1/2.

Rq gives greater weight to high peaks and valleys. The ten-point height, Rz, and maximum height, Rt or Rmax depending on convention, are more sensitive to isolated features and should be used carefully when the sampled length is short or contamination is possible.

Skewness, Rsk, describes the asymmetry of the height distribution. A negative value indicates a surface with relatively deep valleys, often desirable for lubricant retention, whereas positive skewness indicates a surface dominated by peaks. Kurtosis, Rku, indicates the sharpness or tail-heaviness of the distribution. It can reveal a surface containing rare extreme features, although it does not identify their spatial arrangement.

Spacing parameters and autocorrelation describe how rapidly the profile varies along the surface. A surface with a small Ra but closely spaced sharp asperities may produce high local pressures. A surface with a similar Ra and broader, rounded features may generate different contact and lubrication behaviour. Parameters must therefore be selected according to the failure mechanism being studied.

Areal metrology evaluates the complete surface rather than a single line. Common areal parameters include Sa, the arithmetic mean height; Sq, the root-mean-square height; Sz, the maximum height range; Ssk and Sku, areal skewness and kurtosis; and parameters describing developed interfacial area, texture direction, and spatial frequencies. Areal measurement is particularly important for surfaces with non-periodic texture, crossed lay, isolated defects, or direction-dependent behaviour.

No parameter should be treated as a universal quality index. A bearing raceway, piston skirt, seal face, additively manufactured surface, and textured hydrodynamic pad require different descriptors. Functional parameters, such as material ratios, valley volumes, peak density, or feature size, are often more informative than a single average roughness value.

Bearing-area curves and contact function

The bearing-area curve, also called the Abbott–Firestone curve, is obtained by ranking surface heights and plotting the material-bearing ratio against a depth or height level. It indicates the proportion of a surface that would support contact at a particular material level. The upper portion is associated with peaks that engage during initial contact; the central region represents the load-bearing plateau; and the lower region represents valleys that may store lubricant or accommodate debris.

A surface with excessive peak content may produce high running-in wear, high asperity pressure, and scuffing. A surface with deep but narrow valleys may retain lubricant without greatly reducing its load-bearing area. Conversely, broad deep valleys may reduce effective contact area and compromise stiffness or sealing. Bearing-area analysis therefore supports decisions about plateau honing, polishing, superfinishing, and surface texturing.

The reduced peak height, core roughness depth, and reduced valley depth are often derived from the bearing-area curve using specified material-ratio limits. These quantities are especially useful for engineered surfaces whose performance depends on separating fragile peaks from a stable core and functional valleys. Their values are highly dependent on filtering, evaluation length, and the selected material-ratio limits; they are not intrinsic constants independent of measurement procedure.

In tribological design, bearing-area information should be combined with contact mechanics. Under a nominal load, only a fraction of the apparent area is initially in contact. Asperity deformation may be elastic, elastic–plastic, or plastic, depending on asperity radius, hardness, load, and material pair. A topography that appears favourable from its bearing curve may still generate damaging local stresses if its peaks are sharp, poorly supported, or chemically reactive.

Measurement technologies and their operating logic

Stylus contact profilometry

A stylus instrument drags a diamond tip across the surface while a transducer records vertical displacement. It is robust, relatively simple, and widely used for industrial roughness control. It can measure difficult engineering surfaces and is less affected by optical reflectivity than many non-contact systems.

The stylus radius and force impose a physical limit. A large tip cannot enter narrow valleys and may smooth sharp features. Excessive force can deform soft coatings or polymeric materials. The traverse direction matters: measuring perpendicular to lay generally captures greater height variation than measuring along lay. Stylus instruments also produce a two-dimensional profile and may miss isolated areal defects.

Optical profilometry

Optical methods include interferometric, focus-variation, confocal, and structured-light approaches. They acquire three-dimensional topography without mechanical contact and can provide dense areal data. Interferometry is highly sensitive for smooth surfaces and small height variations but may be affected by vibration, discontinuities, steep slopes, and limited coherence conditions. Focus-variation methods are useful for rough or steep surfaces but depend on surface texture and optical focus. Confocal methods reject out-of-focus light and can resolve complex topography, although reflectivity, transparency, and acquisition speed remain important considerations.

Optical results can contain missing points, spikes, stitching errors, shadowed regions, or false heights caused by multiple reflections. A visually attractive colour map is not evidence of measurement validity. Calibration artefacts, repeat scans, reference standards, and inspection of raw data are necessary.

Other methods

Atomic force microscopy provides nanoscale topography over small fields and is valuable for thin films, coatings, and research into nanoscale contact. It is slow and has a limited measurement area. Scanning electron microscopy gives excellent morphology but generally requires indirect methods or stereo reconstruction for quantitative height measurement. White-light and laser scanning systems offer rapid areal inspection but may be sensitive to reflectivity, slope, and surface contamination.

The instrument should be selected from the function backwards. A seal face requiring micrometre-scale plateau information may be poorly represented by a nanoscale technique with a tiny field. Conversely, a coating whose failure begins at nanometre-scale defects cannot be characterised adequately by a long-traverse stylus alone.

Filtering, sampling, and measurement uncertainty

Filtering separates scales in the measured signal. A high-pass or short-wavelength filter may isolate roughness from waviness, while a low-pass filter may remove form or roughness depending on convention. Gaussian filters are widely used, but robust filters may be preferable when large peaks or valleys would distort the mean line. The cutoff wavelength, nesting index, evaluation length, and end-effect treatment must be documented.

Sampling must satisfy the spatial content of the surface. If the sampling interval is too large, short-wavelength features are lost or aliased. If the traverse is too short, long-wavelength waviness and rare defects are not represented. Areal measurements require suitable lateral spacing in both directions and a field size capable of capturing the texture scales relevant to function.

Measurement uncertainty includes calibration uncertainty, repeatability, reproducibility, environmental effects, data processing, instrument resolution, and sampling uncertainty. Vibration, temperature drift, cleanliness, operator alignment, optical reflectance, stylus wear, and specimen fixturing can all affect results. A reported value such as Sa = 0.20 micrometres is incomplete unless the measurement conditions and uncertainty are known.

A useful uncertainty budget distinguishes:

• Type A components estimated from repeated measurements;
• Type B components estimated from calibration certificates, instrument specifications, resolution, and prior knowledge;
• sampling uncertainty caused by surface non-uniformity and limited measurement area.

For process qualification, the decision rule should account for uncertainty. If a specification limit is close to the measured value, accepting or rejecting a component solely from the central estimate may create an inappropriate risk of false acceptance or false rejection.

Multiscale and fractal descriptions

Surface features occur over many length scales. A machined surface may contain tool marks at the feed scale, vibration marks at a larger scale, burrs at a smaller scale, and material or coating features at still smaller scales. A single roughness parameter at one cutoff cannot represent this hierarchy.

Multiscale analysis examines how surface amplitude, developed area, contact geometry, or spectral energy changes with observation scale. Power spectral density describes how variance is distributed over spatial frequencies and can distinguish surfaces with similar Ra but different characteristic wavelengths. Wavelet methods provide localised scale information and can identify transient defects or changes along a profile.

Fractal approaches model certain surfaces using scale-dependent irregularity and parameters such as fractal dimension. They can be useful when apparent roughness changes with magnification and when asperity populations lack a single characteristic wavelength. However, a fractal fit is not automatically a physical explanation. Finite measurement bandwidth, filtering, digitisation, noise, and non-fractal process features may produce apparent scaling. Fractal parameters should therefore be reported with the scale range, fitting method, and evidence of a valid scaling regime.

Multiscale descriptions are valuable in contact mechanics because real contact area, asperity curvature, and local pressure depend on the scale at which the surface is observed. They are also relevant to lubricant films, where a surface feature may be large compared with the film thickness and therefore act as a hydrodynamic disturbance, while smaller features may be submerged within the film.

Manufacturing processes and the origin of surface integrity

Manufacturing leaves a characteristic topographic and metallurgical signature. Turning commonly produces feed marks, tool nose effects, built-up-edge damage, and directional anisotropy. Milling may produce periodic scallops, runout-related variation, and interrupted-cut damage. Grinding can produce fine directional grooves, wheel-loading defects, tensile thermal stresses, grinding burn, or microcracks if heat generation is excessive. Honing and plateau finishing deliberately modify the peak and valley populations, often creating a cross-hatched texture for lubricant retention.

Electrical discharge machining may create recast layers, craters, microcracks, and residual stresses. Laser processing can generate resolidified material, oxidation, and steep local gradients. Additive manufacturing often produces partially fused particles, stair-step effects, pores, unmelted regions, and anisotropic texture; post-processing may be necessary before a surface is suitable for sliding or sealing contact.

Shot peening, burnishing, rolling, and laser shock processing alter both geometry and subsurface stress. They may reduce roughness or increase it, but their principal purpose may be the production of compressive residual stress and work hardening. Coatings introduce interfaces, thickness variations, columnar structures, and defect populations. Thus, surface topography must be examined alongside hardness, microstructure, chemistry, and residual stress.

The correct manufacturing target is functional. For example, a cylinder liner may require a plateau capable of supporting load, valleys capable of retaining oil, and a subsurface resistant to scuffing. A bearing raceway may require extremely low waviness and controlled roughness because long-wavelength errors influence vibration and lubricant film formation. A seal face may require low waviness but carefully controlled microtexture to avoid leakage and excessive friction.

An applied metrology workflow

A sound investigation begins by defining the tribological failure or performance requirement. The engineer should ask whether the concern is running-in wear, abrasive wear, lubricant starvation, leakage, fatigue, friction instability, coating delamination, or process drift. This determines the relevant scales and directions.

The specimen is then cleaned using a method that does not remove loosely attached functional material or alter the surface. Measurement locations are selected to represent process variation, not merely convenient regions. The instrument is calibrated, the stylus or optical objective is checked, and the measurement orientation is recorded. Raw profiles or point clouds should be retained before filtering.

Next, form is removed appropriately, the filter and cutoff are selected, and the evaluation length is justified. Profile and areal parameters are calculated, but interpretation proceeds from the actual topography, bearing-area curve, and spatial information. Repeat measurements quantify repeatability, while measurements at different locations quantify sampling variation. Results are compared with manufacturing records and functional tests.

The final report should identify instrument, calibration status, measurement direction, field or traverse size, sampling interval, filter, cutoff, parameter definitions, number of repeats, uncertainty, and exclusion rules for defects. This level of traceability is essential when metrology results are used in supplier qualification, failure analysis, warranty decisions, or regulatory evidence.

Illustrative case study: premature wear in a plateau-honed cylinder liner

Background

An engine manufacturer introduced a revised honing process intended to reduce friction. Dimensional inspection showed that the cylinder liners met diameter and roundness specifications. Initial roughness control based on Ra also indicated compliance.

Situation and problem

Endurance testing nevertheless showed increased oil consumption and accelerated ring wear. The process team initially suspected lubricant formulation, but the problem occurred only in a subset of liners produced during a particular shift.

Analysis

A single-line measurement parallel to the dominant honing direction produced acceptable Ra values. Additional measurements perpendicular to the lay revealed variation in the cross-hatched texture. Areal maps showed that some liners contained high, sharp plateaus and insufficiently developed valleys. Bearing-area curves indicated excessive reduced peak height and a small valley component. Optical measurements also revealed local torn regions that were missed by the original traverse.

Metallographic examination found a locally smeared near-surface layer associated with an overloaded honing step. Microhardness and residual-stress measurements showed that the topographic problem was accompanied by altered surface integrity. The original inspection plan had treated Ra as a sufficient descriptor and had not controlled measurement direction, sampling location, or uncertainty.

Approach and solution

The manufacturer revised the specification to include plateau and valley parameters derived from the bearing-area curve, areal texture direction, maximum local peak limits, and a minimum number of circumferential and axial measurement locations. Honing pressure, abrasive condition, coolant filtration, and workpiece rotation were also controlled. A small set of liners was subjected to functional ring-friction and oil-retention tests before full production release.

Outcome

The revised process reduced the population of sharp plateaus and restored a more stable combination of load-bearing material and lubricant-retaining valleys. The important improvement was not the pursuit of the lowest Ra; it was the alignment of metrology with the actual contact and lubrication mechanisms.

Learning points

The case demonstrates that dimensional compliance does not guarantee tribological performance. Measurement direction and field size can conceal anisotropy and local defects. Bearing-area curves often explain functional behaviour better than Ra alone, and topography should be interpreted with surface integrity and process history.

Challenges, professional practice, and emerging directions

Common errors include comparing values obtained with different cutoffs, measuring only one direction, ignoring waviness, reporting excessive decimal places, removing genuine defects during data filtering, and treating optical maps as automatically superior to contact measurements. Another error is to optimise a surface parameter without validating friction, wear, sealing, or fatigue performance.

Professional practice requires traceability, honest reporting of uncertainty, and clear separation between measured data and inferred function. A metrologist should not alter filtering merely to achieve specification compliance. In regulated or safety-critical applications, measurement procedures, calibration, data retention, and decision rules should be documented and auditable.

Digital surface metrology is increasingly integrated with manufacturing execution systems. Automated optical scanning can map large populations, while statistical process control can detect drift in texture parameters. Machine-learning models may classify defects or predict wear from topographic features, but their predictions require representative training data, controlled measurement conditions, and physical validation. Explainability is especially important when a model influences acceptance of safety-critical components.

Emerging research combines three-dimensional topography with residual-stress mapping, chemistry, hardness, and in situ tribological measurements. Digital twins of contacts may use multiscale surface data to predict lubricant film behaviour and asperity interaction. These developments will be useful only if measurement uncertainty and scale compatibility are explicitly managed.

Best-practice principles

A reliable surface investigation follows several principles:

1. Define the functional failure mechanism before selecting a parameter.

2. Measure the scales, directions, and locations relevant to the contact.

3. Record raw data and fully specify filtering and sampling.

4. Use both profile and areal information when texture is directional or non-periodic.

5. Interpret roughness together with bearing-area behaviour and surface integrity.

6. Quantify repeatability, reproducibility, and sampling uncertainty.

7. Correlate metrology with friction, wear, sealing, fatigue, or lubricant tests.

8. Treat unusually high peaks, deep valleys, and missing optical data as features requiring investigation, not automatic processing artefacts.

Chapter Summary

Surface topography is a scale-dependent description of form, waviness, roughness, lay, and spatial texture. Surface integrity adds the near-surface mechanical, metallurgical, chemical, and residual-stress condition created by manufacturing. Contact stylus and optical instruments provide complementary measurement capabilities, but every result depends on sampling, orientation, filtering, calibration, and uncertainty. Ra and related averages are useful process indicators but cannot uniquely predict tribological performance. Bearing-area curves, areal parameters, spatial analysis, and multiscale methods provide more functional information about load-bearing plateaus, lubricant-retaining valleys, asperity interaction, and texture direction. Manufacturing processes create distinctive topographic and subsurface signatures, so metrology must be connected to process history and tribological testing. The most reliable practice is function-based, traceable, uncertainty-aware, and supported by both measured data and physical interpretation.

Key Terms

Form: Large-scale deviation of a surface from its intended nominal geometry.

Waviness: Medium-scale surface undulation between form and roughness scales.

Roughness: Short-scale irregularity of a surface after specified form and waviness treatment.

Lay: Predominant direction of surface marks produced by manufacturing.

Texture: Spatial and directional description of surface features.

Surface integrity: Geometrical, mechanical, metallurgical, chemical, and residual-stress condition of a surface and its subsurface.

Ra: Arithmetic mean of the absolute profile deviations from the reference mean line.

Rq: Root-mean-square value of profile height deviations.

Skewness: Statistical measure of asymmetry in the height distribution.

Kurtosis: Statistical measure of the sharpness or tail weight of a height distribution.

Bearing-area curve: Plot of material-bearing ratio as a function of surface level or depth.

Reduced peak height: Bearing-curve parameter describing the prominent peak region above the core.

Reduced valley depth: Bearing-curve parameter describing the valley region below the core.

Stylus profilometer: Contact instrument that traces a surface with a sensing tip to obtain a profile.

Optical profilometer: Non-contact instrument that determines surface height using optical principles.

Cutoff wavelength: Specified scale separating components such as roughness and waviness.

Sampling uncertainty: Uncertainty arising because a limited measured area may not represent the whole surface.

Power spectral density: Description of how surface variance is distributed across spatial frequencies.

Fractal dimension: Scale-related descriptor of geometric irregularity in a surface model.

Aliasing: Misrepresentation of spatial features caused by inadequate sampling frequency.

Review Questions

Short Answer Questions

1. Distinguish form, waviness, roughness, and lay.

2. Why is Ra insufficient as a complete description of a tribological surface?

3. What information does a bearing-area curve provide?

4. State two limitations of stylus profilometry.

5. What is meant by sampling uncertainty in surface metrology?

Long Answer Questions

1. Explain the principles, advantages, and limitations of contact and optical surface measurement methods.

2. Discuss how filtering, cutoff wavelength, evaluation length, and measurement direction influence reported roughness values.

3. Explain the functional significance of reduced peak height, core roughness, and reduced valley depth.

4. Relate turning, grinding, honing, electrical discharge machining, and additive manufacturing to characteristic surface integrity issues.

5. Discuss why multiscale and fractal descriptions can add value beyond conventional roughness parameters, including their limitations.

Analytical / Application Questions

1. A component meets its Ra specification but experiences lubricant starvation and high running-in wear. Design a metrology plan to investigate the problem and justify the parameters you would select.

2. Two surfaces have identical Sa values, but one has strongly directional grooves and the other has isotropic texture. Predict how their tribological behaviour might differ and identify measurements needed to support your prediction.

3. An optical scan shows missing points and unusually high peaks on a coated surface. Explain how you would determine whether these are measurement artefacts, genuine defects, or coating features before accepting or rejecting the component.

22. Multiple Choice Questions with Answers

Question 1

Which statement best explains why Ra alone is insufficient for predicting tribological performance?

A. Ra can be measured only with optical instruments

B. Ra does not describe spacing, direction, peak shape, or valley distribution

C. Ra is independent of the measurement cutoff

D. Ra directly measures subsurface residual stress

Correct Answer: B. Ra does not describe spacing, direction, peak shape, or valley distribution

Question 2

A surface has a negative skewness. Which interpretation is most appropriate?

A. It is dominated by high peaks

B. It has a perfectly periodic texture

C. It contains relatively more deep valleys than peaks

D. It must have high waviness

Correct Answer: C. It contains relatively more deep valleys than peaks

Question 3

What is the principal function of a bearing-area curve?

A. To determine material-bearing ratio at different surface levels

B. To measure residual stress directly

C. To eliminate all measurement noise

D. To determine chemical composition

Correct Answer: A. To determine material-bearing ratio at different surface levels

Question 4

Which condition most directly causes a stylus instrument to miss narrow valleys?

A. A stylus tip radius that is too large

B. A low optical reflectivity

C. A high camera frame rate

D. A negative skewness

Correct Answer: A. A stylus tip radius that is too large

Question 5

Why should measurement direction be recorded for a machined surface?

A. Surface texture is often anisotropic because of lay

B. Direction changes the material hardness

C. Direction eliminates the need for filtering

D. Direction determines the calibration certificate

Correct Answer: A. Surface texture is often anisotropic because of lay

Question 6

A profile is sampled at intervals that are too large compared with its shortest features. What is the likely result?

A. Residual stress is increased

B. Aliasing or loss of spatial information

C. Improved valley measurement

D. Automatic removal of waviness only

Correct Answer: B. Aliasing or loss of spatial information

Question 7

Which manufacturing process is particularly associated with cross-hatched texture intended to support lubrication?

A. Honing

B. Electrical discharge machining

C. Shot blasting only

D. Casting without finishing

Correct Answer: A. Honing

Question 8

Why can a surface with low average roughness still produce high local contact pressure?

A. Low average roughness guarantees broad rounded asperities

B. Sharp or closely spaced asperities may remain despite a low average value

C. Low roughness always increases residual stress

D. Average roughness measures only chemical composition

Correct Answer: B. Sharp or closely spaced asperities may remain despite a low average value

Question 9

Which statement about fractal surface analysis is correct?

A. Every engineering surface is fractal over all scales

B. A fractal fit should be assessed over a stated valid scale range

C. Fractal dimension replaces the need for calibration

D. Fractal analysis is independent of sampling and filtering

Correct Answer: B. A fractal fit should be assessed over a stated valid scale range

Question 10

In the cylinder-liner case, what was the most important improvement in the inspection strategy?

A. Replacing all measurements with diameter inspection

B. Optimising only the lowest possible Ra

C. Adding bearing-area, directional, areal, and local peak information

D. Measuring only along the honing direction

Correct Answer: C. Adding bearing-area, directional, areal, and local peak information