Find the question you need
Choose a route or read from the beginning. All teaching chapters stay open; the grouped index below is expandable.
Derive timing bias
Chapters 3–5. Convert timing offsets into phase and measured components.
Test a false crossover
Chapters 6–7. Separate frequency effects from true transitions.
Check the instrument
Chapters 8–10. Check linear range, geometry and phase calibration.
Full chapter index · Practice and worked checks · Rheology subject root
In oscillatory rheometry, a small timing error between measured stress and strain can change the apparent storage modulus G′ and loss modulus G″ without changing the measured stress amplitude. If the timing error moves the reported phase angle across 45°, a material may appear to cross from storage-dominated to loss-dominated response even though its true response at that frequency has not changed.
For a fictional material with complex modulus magnitude 1,000 Pa and true phase angle 42°, G′ is approximately 743 Pa and G″ about 669 Pa. At angular frequency 50 rad/s, a 2-millisecond positive relative timing offset adds about 5.73° to the measured phase, producing approximately 48° and reversing the apparent ordering. The calculation is a measurement-system warning, not proof that any particular instrument has this error. This guide derives the phase relation, checks a false crossover and links it to the Gold Standard of Rheology.
Chapter index
Read the phase signal · Chapters 1–2
1. Define what the two oscillatory moduli mean
2. Derive the in-phase and quadrature components
Derive timing bias · Chapters 3–5
3. Translate relative timing error into phase error
4. Calculate a false crossover from a small delay
5. Look for the bias inside a frequency sweep
Test a false crossover · Chapters 6–7
6. Separate signal synchronization from sample physics
7. Check whether a crossover belongs to the linear regime
Check the instrument · Chapters 8–10
8. Design a phase-calibration evidence trail
9. Propagate phase-angle uncertainty into both moduli
10. Answer the questions that cause false transitions
Practise and continue · Chapters 11–12
11. Practise a new frequency and inspect the margin
12. Continue from observed phase to trustworthy material evidence
CHAPTER 1 OF 12 · Read the phase signal
1. Define what the two oscillatory moduli mean
In a small-amplitude oscillatory shear experiment, the instrument imposes or observes a sinusoidal strain and a sinusoidal stress response. The relative phase between them describes how much response is in phase with deformation and how much is in phase with deformation rate. In linear viscoelastic analysis, those components are represented by storage modulus G′ and loss modulus G″.
For a simple sign convention, let strain be γ(t) = γ0 sin(ωt) and stress be τ(t) = τ0 sin(ωt + δ), where δ is the stress lead relative to strain. Then complex modulus magnitude |G*| = τ0/γ0, G′ = |G*| cosδ and G″ = |G*| sinδ. The loss tangent tanδ = G″/G′ whenever the denominator is nonzero.
The measured phase is not a direct photograph of molecular behavior. It is inferred from sampled, filtered and synchronized instrument signals. Timing conventions, mechanical compliance, inertia and software corrections enter its practical determination. A phase calibration problem can therefore change the inferred division of modulus components while the underlying sample is unchanged.
This article concerns timing-driven phase bias in the linear oscillatory regime. It does not repeat the bulk stress-relaxation or creep-recovery derivations, and it does not claim that all apparent crossovers are artifacts. Genuine frequency-dependent changes can also produce a crossover.
CHAPTER 2 OF 12 · Read the phase signal
2. Derive the in-phase and quadrature components
Use the trigonometric identity sin(ωt+δ) = sin(ωt)cosδ + cos(ωt)sinδ. Thus τ(t) = τ0 cosδ sin(ωt) + τ0 sinδ cos(ωt). Dividing the two response coefficients by strain amplitude γ0 yields G′ = (τ0/γ0)cosδ and G″ = (τ0/γ0)sinδ.
At δ = 0°, the response lies entirely in phase with strain in this ideal convention, giving G′ = |G*| and G″ = 0. At δ = 90°, the response is entirely in quadrature, giving G′ = 0 and G″ = |G*|. Real measurements may produce small negative or beyond-range values from conventions, instrument artifacts or nonlinear conditions, so phase interpretation must include quality checks.
A phase of 45° makes cosδ = sinδ, hence G′ = G″. That mathematical equality is often called a crossover. It marks equal storage and loss components at a particular frequency and test state. It does not by itself prove the material has a unique, universal gel point or relaxation time under every condition.
Amplitude ratio and phase answer separate questions. If the phase shifts while |G*| remains at 1,000 Pa, the two components rotate around a constant magnitude in the G′–G″ plane. Recognizing this signature helps diagnose a pure timing error before attributing the change to materials chemistry.
CHAPTER 3 OF 12 · Derive timing bias
3. Translate relative timing error into phase error
Suppose the instrument’s stress and strain channels are misregistered in relative time by Δt, with the sign defined so that measured stress lead is greater than the true lead. Then δ_meas = δ_true + ωΔt when angles are expressed in radians. If a different wiring or processing delay reverses the sign, the phase bias is negative instead.
The same fixed time offset produces a larger angular bias at higher frequency because one oscillation period becomes shorter. In degrees, Δδ_deg = (180/π) ωΔt. For ω = 50 rad/s and Δt = 0.002 s, the phase bias is 0.1 rad ≈ 5.73°. For ω = 5 rad/s with the same offset, it is 0.01 rad ≈ 0.573°.
This simple expression models one constant relative timing offset. Real phase calibration can be frequency-dependent, with mechanical compliance, digital filters, inertia and feedback dynamics creating additional behavior. A calibration measured at one frequency does not always correct a whole frequency sweep.
The model’s usefulness is that it makes a falsifiable prediction: if a constant delay dominates, phase error measured in radians should increase approximately linearly with angular frequency. A reference measurement across frequencies can test that prediction.
CHAPTER 4 OF 12 · Derive timing bias
4. Calculate a false crossover from a small delay
Take |G*| = 1,000 Pa and true phase δ_true = 42°. Then G′_true = 1,000 cos42° ≈ 743 Pa, while G″_true = 1,000 sin42° ≈ 669 Pa. Storage modulus is larger than loss modulus under the chosen true test state.
At 50 rad/s with the defined +2-millisecond relative timing bias, δ_meas = 42°+5.73° = 47.73°. Thus G′_meas ≈ 1,000 cos47.73° ≈ 673 Pa and G″_meas ≈ 1,000 sin47.73° ≈ 740 Pa. The reported component ordering reverses. A naive interpretation might claim a material transition at this frequency.
At 5 rad/s, the same delay adds only 0.573°, yielding δ_meas ≈ 42.57°, with G′ still larger than G″. A frequency plot could therefore exhibit a crossover created by the timing system even though the constructed material had a constant true phase of 42° throughout this demonstration.
| Condition | Phase angle δ | G′ for | G* | = 1,000 Pa | G″ for | G* | = 1,000 Pa |
|---|---|---|---|---|---|---|---|
| True material response | 42.00° | ≈ 743 Pa | ≈ 669 Pa | ||||
| Recorded at 5 rad/s with +2 ms | 42.57° | ≈ 736 Pa | ≈ 677 Pa | ||||
| Recorded at 50 rad/s with +2 ms | 47.73° | ≈ 673 Pa | ≈ 740 Pa | ||||
| Equality threshold | 45.00° | ≈ 707 Pa | ≈ 707 Pa |
Every value is a fictional calculation based on a pure phase bias and fixed modulus magnitude. The numbers are not rheometer calibration data or a description of a real sample. A genuine material can also change |G*| and δ with frequency; the example deliberately isolates the instrumental mechanism.
An ordinary oscillatory frequency sweep changes ω while trying to keep the material within the chosen linear viscoelastic strain or stress amplitude range. A real sample’s G′ and G″ often depend on ω, because molecular and structural relaxation occurs on different timescales. Any suspected instrument phase bias must be evaluated alongside that genuine behavior.
A fixed time offset contributes phase error proportional to ω. Near δ = 45°, even a few degrees may swap the ordering of G′ and G″. Far from 45°, the same offset may change their numerical values without producing a visible crossover. Both types of error can matter in comparative formulation work.
The sine and cosine geometry also tells us that a pure phase rotation leaves (G′)² + (G″)² = |G*|² unchanged. If measured modulus magnitude changes strongly at the same time, a simple constant-delay model may be insufficient. Signal attenuation, compliance correction, fixture changes and sample evolution should also be investigated.
An oscillatory crossover can be discussed as an operational feature of the specified frequency, temperature and loading amplitude. Calling it a universal ‘liquid-to-solid moment’ is often too broad, particularly for materials whose response depends on temperature, aging or imposed strain amplitude.
CHAPTER 6 OF 12 · Test a false crossover
6. Separate signal synchronization from sample physics
Stress and strain are often inferred from torque and angular displacement. Their channels may pass through different electronic filters, digital acquisition paths and mechanical response functions. An apparent phase lead or lag can therefore reflect the measurement chain, not just the sample’s response.
Mechanical compliance is a related but distinct issue: some measured motion belongs to the instrument and fixtures instead of the specimen. Inertia may also shift the phase between commanded motion, actual motion and torque. These effects can vary with frequency and geometry. A single fitted delay should not be used as a universal correction when the actual error has multiple mechanisms.
To investigate, measure a suitable reference material or instrument verification state under the relevant frequency and amplitude conditions. Compare phase, modulus amplitude and geometry dependence with expected behavior and documented instrument performance. If a manufacturer provides validated phase, inertia or compliance corrections, understand their domains before altering them manually.
A small phase error can cause a large relative error in the smaller modulus component when δ is near 0° or 90°. This makes instrument calibration especially important when reporting low G″ for nearly elastic specimens or low G′ for nearly viscous ones. The error’s practical significance depends on the quantity being claimed.
CHAPTER 7 OF 12 · Test a false crossover
7. Check whether a crossover belongs to the linear regime
A frequency or amplitude sweep only supports linear viscoelastic interpretations when the measurement remains within a range where the response is approximately proportional to the imposed signal. If an amplitude sweep drives the sample into nonlinear yielding or structural breakdown, G′ and G″ derived by linear sinusoidal decomposition may not retain the same simple interpretation.
A reported crossover from a high-strain amplitude sweep may therefore reflect nonlinear material response rather than a characteristic linear relaxation timescale. Distinguish that problem from phase bias: one changes the material state and harmonics, while the other changes how a measurement is timed or corrected.
Check harmonic distortion, waveform quality, signal-to-noise ratio, wall slip and test history. The separate Wall Slip and Gap-Series Identifiability guide describes how boundary movement can distort bulk interpretation. Those effects can coexist with phase errors and should not be lumped into one generic ‘calibration correction’.
A credible crossover claim names its temperature, frequency, amplitude, geometry and protocol. If the result is central to a material choice, repeat near the crossover with independent settings or instrumentation where feasible.
CHAPTER 8 OF 12 · Check the instrument
8. Design a phase-calibration evidence trail
A useful rheometry report records the commanded and measured waveforms, frequency, amplitude, sample geometry, gap, temperature, torque range, tool compliance settings, phase correction, waveform quality and replicate results. It should distinguish raw measured phase from values that already include manufacturer corrections.
| Check | Helps separate | Does not establish alone |
|---|---|---|
| Reference-material phase across frequency | Instrument delay or phase bias | Full sample behavior for every geometry |
| Modulus magnitude with phase | Pure rotation versus amplitude change | The physical cause of all errors |
| Linear-range amplitude scan | Linear response from yielding or harmonics | Correct interchannel synchronization |
| Gap and fixture variation | Boundary and compliance sensitivity | Unique material relaxation spectrum |
In a clean validation experiment, repeat the same material in more than one relevant test configuration and ask whether the inferred crossover is stable within uncertainty. A timing offset that scales with ω may move the apparent crossing between configurations or correction settings, while a robust physical feature should behave according to a defensible constitutive model.
Instrument corrections can also introduce uncertainty. An article that reports a corrected phase to several decimal places but ignores uncertainty in the correction has merely moved the uncertainty out of sight. Include a sensitivity test around the estimated delay or phase offset.
CHAPTER 9 OF 12 · Check the instrument
9. Propagate phase-angle uncertainty into both moduli
A phase correction also has uncertainty. At fixed complex modulus magnitude M, small phase uncertainty uδ expressed in radians contributes approximately u(G′) ≈ M|sinδ|uδ and u(G″) ≈ M|cosδ|uδ. Their errors are correlated because both depend on the same uncertain angle.
Near 45° with M = 1,000 Pa, a 1° phase uncertainty equals about 0.01745 radians and contributes roughly 12.3 Pa of first-order uncertainty to each modulus. The two components move in opposite directions as phase changes, so treating these contributions as independent would lose that relationship.
If the corrected phase uncertainty interval includes 45°, the measured crossover may not be sharply resolved. This local sensitivity check does not replace correction for tool compliance, inertia, amplitude errors, sample slip or nonlinearity, but it connects the reported dominance to the measurement’s discriminatory power.
CHAPTER 10 OF 12 · Check the instrument
10. Answer the questions that cause false transitions
Does G′ = G″ prove a material has changed from solid to liquid?
It indicates equal measured storage and loss components at the stated test point under the linear convention. Interpretation depends on frequency, temperature, amplitude, sample history and measurement validity. It is not a universal phase-transition certificate.
Can a delay change G′ and G″ without changing |G*|?
Yes. A pure phase offset rotates the two components while preserving their squared-sum magnitude. This is exactly the signature used in the constructed example; real instruments may show additional amplitude errors.
Can I correct every frequency using one constant phase shift?
Not unless evidence supports a constant angular offset across that range. A fixed time delay produces phase error proportional to angular frequency, and other instrument dynamics may have more complicated frequency dependence.
Does a smooth crossover curve prove the values are accurate?
No. Smoothness can result from a systematic calibration bias. Accuracy requires appropriate reference tests, signal quality checks and uncertainty analysis, not just a visually attractive fit.
CHAPTER 11 OF 12 · Practise and continue
11. Practise a new frequency and inspect the margin
Take a fictional true phase δ_true = 43°, complex modulus magnitude 2,000 Pa, angular frequency 20 rad/s and a positive relative timing error Δt = 0.0015 s. The phase error is ωΔt = 0.030 rad ≈ 1.72°, so the measured phase is about 44.72°. It approaches the 45° equality threshold without quite crossing it.
Double the frequency to 40 rad/s while keeping the same toy true phase and delay. The phase error becomes 0.060 rad ≈ 3.44°, giving δ_meas ≈ 46.44°. The apparent ordering reverses. For this fictional model the true storage component remains larger because δ_true stays below 45°.
Now suppose an independent reference test constrains the relative timing offset to 1.5 ± 0.5 ms. At 40 rad/s, the phase-offset uncertainty corresponding to 0.5 ms is 0.020 rad ≈ 1.15°. A crossover near 45° should not be reported as a sharply determined material change without considering that margin.
For a final exercise, state the sign convention, calculate ωΔt in radians before converting to degrees, and describe one reference measurement that could distinguish a material relaxation crossover from an instrument-timing effect.
CHAPTER 12 OF 12 · Practise and continue
12. Continue from observed phase to trustworthy material evidence
The gold standard of oscillatory phase interpretation is to connect signal timing, phase, modulus components and physical material behavior with visible assumptions. A crossover is informative only when the phase and amplitude channels have been checked over the relevant frequency and strain regime, and when instrument artifacts are not silently being treated as material properties.
Return to the Gold Standard of Rheology for oscillatory shear and rheometry foundations. Continue to Parallel-Plate Torque and the Non-Newtonian Rim-Stress Correction for a different geometry-dependent inference issue, and to Creep Recovery and Delayed versus Permanent Strain for a time-domain approach to material response. Each method needs its own instrument and constitutive limits.
Sources and further reading
FHWA: Four-mm Dynamic Shear Rheometry describes instrument compliance and correction issues in oscillatory rheometry. NETZSCH: Crossover Point explains the operational G′–G″ equality criterion. Rheological characterization of basaltic melts by oscillatory shear gives material-focused background on phase and the complex moduli. The phase, delay and modulus values above are original educational calculations, not empirical instrument findings.
How X Works and The Gold Standard of Education connect this problem with a broader practice: verify the measurement clock before turning an observed change into a claim about what a material has become.
Sources checked 21 October 2026. Worked examples are educational models with the assumptions stated above.