On this page
A wet print can lose fine mesh texture while retaining a broad ridge. That difference is not necessarily contradictory evidence about the paste. Surface disturbances of different lateral sizes can relax at very different rates. Track the actual profile scale and time after screen release before deciding that a longer wait or a lower viscosity will improve the finished layer.
Key design decisions
- Report the lateral wavelength and amplitude of the topography rather than one visual smoothness score.
- Use the capillary model only where its Newtonian, constant-property and small-disturbance assumptions are defensible.
- Keep surface leveling separate from boundary spreading and solvent-removal readiness.
1. Define the feature that is supposed to relax
Choose a region inside a broad printed deposit and identify the surface feature being followed. A fine periodic mesh impression, a wide squeegee ridge and the outer edge of a narrow line involve different lateral scales and boundary conditions. A single average roughness value can hide the survival of the broadest feature.
Record both amplitude and lateral spacing using the same location and measurement definition at each observation. The initial amplitude should be associated with a reproducible time after screen separation. A measurement taken after an unrecorded transfer delay may miss most of the early change. Keep the printed material state distinct from the dried or fired topography that appears later.
2. Use capillary leveling as a limited physical model
Surface curvature creates a pressure variation that can drive a liquid from high-curvature regions toward a flatter profile. Viscous resistance opposes that flow. A simple thin-film description assumes a Newtonian liquid on a rigid no-slip substrate, a shear-free upper surface, small slopes and constant surface tension and viscosity.
A filled screen-printing paste can have yield behavior, time-dependent rebuilding, wall slip and solvent loss. Those are not small details that automatically disappear after printing. Treat the ideal model as a way to test scale sensitivity, not a universal paste model. If the material stops flowing because its structure rebuilds, a constant-viscosity calculation cannot prescribe how long the real surface must be left to level.
3. Follow one small sinusoidal disturbance
Consider an otherwise uniform film of mean thickness h0 with a small sinusoidal height disturbance. Its amplitude must be much smaller than the mean thickness, and its lateral scale must support the thin-film approximation. Linearizing the capillary flow equation produces an exponential amplitude decay for that single spatial mode.
The decay rate is proportional to the fourth power of the spatial wavenumber. Wavenumber is two pi divided by wavelength; it is not simply the reciprocal wavelength when used in this expression. Missing the two-pi factor introduces a large numerical error. Keep the complete wavelength definition and units with any spreadsheet implementation.
a(t) = a0 exp(−t/τ); τ = 3η/[γ h0³ q⁴]; q = 2π/λ
- a and a0 are disturbance amplitudes in m; h0 is mean wet-film thickness in m.
- λ is lateral wavelength in m; q is wavenumber in 1/m.
- η is dynamic viscosity in Pa·s, γ is surface tension in N/m, and τ and t are in s.
Small-amplitude sinusoidal disturbance of a uniform Newtonian film; small slopes, constant properties, rigid no-slip support, no evaporation or contact-line motion, and negligible gravity, inertia and disjoining-pressure effects.
4. Calculate the penalty for a longer wavelength
Assume a viscosity of 10 pascal-seconds, surface tension of 0.030 newton per metre and mean thickness of 20 micrometres. These are illustrative model inputs, not measured properties of a company paste. For a wavelength of 0.50 millimetre, the calculated decay time is approximately 5.013 seconds.
Doubling wavelength to 1.00 millimetre increases the decay time sixteenfold, to approximately 80.203 seconds. After about five seconds, the short-wavelength amplitude is close to 37 percent of its initial value, while the long-wavelength amplitude remains close to 94 percent. A visibly smoother fine texture can therefore coexist with a broad residual ridge in the same ideal material.
The example compares small interior disturbances, not a droplet spreading over bare ceramic. It does not establish a permitted production waiting interval. Before using it to plan a trial, determine whether the real paste maintains anything resembling the assumed mobility during that interval.
5. Separate sensitivity from a process adjustment
Within the ideal equation, doubling viscosity doubles the time constant. Doubling mean thickness reduces it by a factor of eight, while doubling wavelength increases it by sixteen. These ratios explain why comparisons need thickness and topographic scale, not just a viscosity number.
They do not justify doubling a printed layer's thickness to obtain a smoother surface. Thickness also changes material volume, electrical behavior, drying, firing and step coverage. Likewise, an unauthorized solvent addition may change wetting, solids content and recovery as well as apparent viscosity. Use the sensitivity to identify which quantities must be measured, then evaluate any actual process change within the material's permitted preparation route.
6. Design an observation that can reject the simple model
Compare the same spatial band at multiple known times without mechanically disturbing the wet deposit. Confirm that the measurement method can resolve the chosen amplitude and does not flatten, drag or contaminate the surface. If repeated observation is invasive, use appropriately matched specimens and retain the additional specimen-to-specimen variation.
A constant-property exponential predicts a straight line when the logarithm of positive amplitude is plotted against time. Persistent curvature, a plateau or different apparent rates over successive intervals can reject that simple description. Do not fit only the early or late interval that produces the desired viscosity. Also track mean thickness or mass where feasible: changing deposit volume indicates that the no-evaporation assumption needs separate attention.
7. Match the observation to the next process question
The useful comparison separates interior shape evolution from material loss and changing footprint. Keep those records together without collapsing them into one smoothness acceptance number.
| Observation | Check that separates mechanisms | Interpretation to avoid |
|---|---|---|
| Fine texture reduces but a broad ridge remains | Compare the two spatial wavelengths over the same time interval | All topography must decay at one rate |
| Interior surface smooths while line width increases | Track footprint and edge position independently | Leveling automatically preserves electrical spacing |
| Amplitude change slows to a plateau | Check recovery, evolving properties and measurement resolution | A constant-viscosity time extension will necessarily remove it |
| Mean thickness falls with surface change | Track mass or volume and drying exposure | All height reduction is redistribution |
| Profiles differ only after dryer entry | Compare identical print-to-dryer timing and initial profiles | The deposited starting geometry was necessarily equal |
8. Transfer a profile requirement with its observation time
State the relevant surface-band amplitude or ridge metric, the wet-state time window and the downstream function it protects. A dried-film cosmetic check alone cannot reveal when an objectionable ridge became fixed. Connect the profile to the actual next-layer coverage, insulation or electrical requirement instead of specifying smoothness solely because it looks attractive.
Preserve the approved waiting interval, environmental condition and deposit geometry when transferring the process. A larger print, a thicker region or a different mesh pattern changes the starting topography and may change the relevant relaxation scale. Confirm the downstream dried and fired result after any adjustment. The final decision concerns a defined material-and-pattern sequence, not a universal promise that every print will level within the calculated time.
Review a post-print ridge or mesh-mark problem
Provide time-linked surface observations and the feature whose final geometry matters.
- Paste identity and preparation history, substrate and actual mean wet thickness.
- Surface profiles with spatial scale, amplitude definition, observation method and time after screen release.
- Footprint or edge-position changes, print-to-dryer interval and any simultaneous mass or thickness change.
- Screen and deposit geometry, environmental conditions and the next-layer or fired functional requirement.
The drawing-upload form loads as you reach this section.