NEWS EVENTS

Viscosity and Cure Shrinkage: How Two Resin-Level Constraints Write Themselves into Final Dimensions

2026-11-22


When a vat-photopolymerized part comes out oversize, the search for a cause usually starts with the machine, the light engine, or the slicing profile. Yet a significant share of that error is already committed the moment resin is poured into the vat. The liquid has to flow into a film of even thickness, then polymerize into a solid that occupies slightly less volume than it did as a liquid. Flow and contraction are two displacements performed by the same material at two different stages. Where either lacks margin, the final dimension records it.

This article draws on seven experimental studies published between 2024 and 2026 to address four questions: how viscosity and thixotropic recovery set the morphology of a single layer; how cure shrinkage becomes measurable dimensional error; how far formulation and exposure strategy can suppress warpage; and where the evidence stops once the part is headed for jewelry investment casting.

1. Three Terms Worth Defining First

Viscosity measures a fluid's resistance to shear deformation. Photopolymer resins are almost all non-Newtonian: apparent viscosity falls as shear rate rises, a behaviour called shear thinning. That property decides whether resin spreads cleanly under a doctor blade or a release film.

Thixotropy describes time dependence. Shear breaks down the internal network; at rest, the network rebuilds. For printing, the number that matters is not the resting viscosity but the structural state the resin reaches in the seconds between being spread into a film and being exposed. Recover too little and the film flows back; recover too much and the film sets with stress locked in.

Cure shrinkage is the volume lost when liquid monomers become a crosslinked network. In radical acrylate polymerization, intermolecular van der Waals distances collapse into covalent bond lengths as bonds form, and the part gets smaller. The loss arrives in two instalments — chemical shrinkage at exposure, and further shrinkage during post-curing. This article deals mainly with the first.

One note on evidence. All seven studies cited here are peer-reviewed journal papers with volume, issue and DOI given, and each conclusion holds only under its own experimental conditions. Manufacturer technical literature is second-tier evidence, used only to describe engineering practice; product specification sheets are third-tier and apply to specific models alone. No product specification number is used to support any claim below, and no single study is treated as a general law.

2. The Levelling Window: Waiting Beats Spreading Faster

Before every layer, resin must be redistributed into an even film over the previously cured surface. Whether it levels depends on how quickly it rebuilds structure after being sheared.

Hoque and colleagues (Additive Manufacturing, 2026, 127: 105298) loaded a commercial acrylate resin with hydrophobic fumed silica (Aerosil R972) to produce high-viscosity suspensions with tunable viscosity, yield stress and recovery. Rheology showed strongly non-Newtonian behaviour: the network breaks down reversibly under shear and rebuilds with time once shear stops. Photo-rheology indicated that filler loading up to 10 wt% did not significantly change bulk curing kinetics under the tested conditions, while ultra-small-angle X-ray scattering confirmed post-shear re-agglomeration and network rebuilding.

The printing trials produced a comparison that is easy to overlook. A 4 wt% formulation had a zero-shear viscosity above the processing limit usually quoted for vat photopolymerization, yet printed reliably because shear thinning brought it into range. An 8 wt% formulation, by contrast, showed bottom-layer overgrowth, layer thinning and slanting. Judging printability from a single viscosity value misreads both ends of the range.

Yan and colleagues (Additive Manufacturing, 2026, 130: 105410) reached a consistent conclusion for doctor-blade recoating of ceramic slurries: recoating quality varies non-monotonically with rheological state, and the best results sit in an intermediate condition rather than at either extreme. Under their conditions they identified an empirical screening window of η30 = 1.36–3.12 Pa·s (apparent viscosity at 30 s⁻¹) and G′rec = 361.6–1588.6 Pa (mean storage modulus during recovery). Their composition–structure–rheology models reached leave-one-out cross-validated R² of 0.978 and 0.757 respectively — the lower second figure suggesting that recovery behaviour is harder to invert from formulation than flow behaviour is.

The most directly useful production finding comes from a second observation in the Hoque study: within the tested range, interlayer wait time influenced layer morphology, dimensional accuracy and mechanical properties more strongly than build-head speed did. A 60 s wait supplied the thixotropic recovery needed, at the cost of ultimate mechanical strength. Speed and accuracy therefore sit at opposite ends of one parameter set and cannot both be maximized. Setting a production tempo is, in effect, choosing a point on that curve.

3. Cure Shrinkage: A Measurable Link to Dimensional Error

Shrinkage is not a defect; it is what polymerization does. What can be managed is its magnitude and the way it is constrained.

Ling and colleagues (Polymers, 2025, 17(5): 610) measured volumetric shrinkage by AcuVol video imaging across one in-house model resin and eight commercial model resins, printed full-arch tooth models by DLP, and quantified accuracy as the 3D mean distance between the scanned model and the design CAD file, tracked from day one to week four. The in-house resin showed 7.28% volumetric shrinkage against 7.66–11.2% for the commercial group, and dimensional deviation of 11.66–13.77 μm against 14.03–41.14 μm. Volumetric shrinkage correlated with dimensional accuracy at a Pearson coefficient of 0.7485.

That 0.7485 deserves to be read both ways. It says volumetric shrinkage explains a substantial portion of dimensional error. It also says the remainder cannot be fixed by reformulating the resin — optical distortion, layer uniformity, washing and post-curing, and the measurement method itself all sit in what is left.

The boundary has to be stated: this is a dental model resin, a DLP process, and full-arch geometry. Treating 7.28% as an industry threshold for "low shrinkage", or 11.66 μm as a generally attainable accuracy, is extrapolation. These numbers are comparable only under the same test method and comparable geometry.

4. Shrinkage Stress and Warpage: The Part Formulation Can Move

Shrinkage cannot be driven to zero, but the stress field it generates can be reshaped. That is where formulation design actually operates.

Soon and colleagues (Journal of the American Ceramic Society, 2025, 109(1)) added 35 wt% isobornyl acrylate (IBOA) — a monofunctional reactive monomer carrying a bulky bicyclic group — to a conventional acrylate resin. Cure shrinkage fell by up to 22% even though the degree of polymerization increased; the authors attribute the effect to steric hindrance from the bicyclic group restraining network contraction. The same formulation markedly lowered the viscosity of an alumina slurry and allowed long bars with an area-to-thickness ratio of up to 350 mm to print with almost no warpage. After sintering, relative density reached about 95.0% and flexural strength about 283.2 MPa.

There is a second route on the exposure side. Pruksawan and colleagues (Advanced Engineering Materials, 2024, 26(24)) split each layer image into a sequence of sub-images using a breadth-first search, so the curing front advances like a ripple. On a 32 mm diameter disc, sub-images 10 pixels wide cut deviation tenfold; long-wave infrared monitoring showed the thermal effect falling in step; reduced residual stress improved mechanical performance by 75%. The authors also note that introducing intervals inside solid layers — a grid instead of solid fill, for instance — suppresses curl and heat build-up.

Both routes leave the material constant alone and instead change the order in which contraction happens and the constraints around it. They differ in ownership. IBOA belongs to the formulation side: once the resin is locked, the benefit comes free in production. Sub-image exposure belongs to the machine and slicing side: adjustable part by part, at the cost of splitting each layer's exposure time.

5. From Chemical Shrinkage to Macroscopic Distortion: Can It Be Predicted

What can be predicted can be compensated. Yang and colleagues (The International Journal of Advanced Manufacturing Technology, 2025, 137: 4109–4128) built a thermal simulation approximate prediction (TSAP) model that uses thermo-mechanical coupling to approximate distortion and internal stress during layer-by-layer printing, mapping microscopic reaction shrinkage at exposure onto macroscopic cold shrinkage through mathematical modelling. Input parameters and their trends were fitted from material experiments, and the model was validated on different part geometries.

The value of such work is less in predictive accuracy than in the direction it points: pre-distort the sliced model by a known amount to cancel the predictable share of shrinkage. The precondition is stable material calibration — change resin and you recalibrate. That also explains why a compensation set so often fails to transfer between materials.

6. Where the Evidence Stops: Investment Casting

Everything above stops at the printing stage. For investment casting, the dimensional chain extends one more link: thermal expansion of the wax pattern, or castable resin pattern, as the shell is heated for dewaxing.

Mukhtarkhanov and colleagues (International Journal of Lightweight Materials and Manufacture, 2025, 8(6): 716–725) proposed measuring the normal forces generated as wax samples solidify and heat up in a rotational rheometer, as a way to evaluate expansion forces. They tested three commercial waxes and one 3D-printable wax, reporting peak normal loads of 86 N for machinable wax, 18.8 N for investment casting wax, and 36.3 N for the 3D-printing wax. The paper identifies thermal expansion of the wax pattern as the main cause of ceramic shell failure during dewaxing, and argues the method suits low-viscosity waxes in particular.

This article cites the method and the order of magnitude only, not the paper's ranking. The abstract is internally inconsistent: it also states that additively manufactured patterns perform considerably better during dewaxing, while the three reported loads do not support that ordering. Until the full text can be checked, no claim about which pattern material is better should be treated as established. This is precisely where an incomplete abstract should stop an argument.

What transfers is the method, not the verdict: expansion force is a quantity that can be measured and compared. Benchmarking pattern materials and batches under one measurement protocol is far more reliable than inferring from casting outcomes after the fact.

7. Conclusion: What the Material Side Can and Cannot Fix

What it can fix is the levelling window. Given a resin's η30 and G′rec, a lower bound on interlayer wait time follows; below it, layer morphology and dimensional accuracy degrade together. Two independent studies point the same way, so this can be treated as engineering practice.

What it can partly fix is volumetric shrinkage. Formulation can move shrinkage from the 11% band into the 7% band, and sterically hindered monomers can push shrinkage stress lower still. Neither reaches zero. The balance is carried by exposure strategy, geometry and post-processing.

What it cannot fix alone is final dimension. Shrinkage is one link in the chain; optical distortion, layer uniformity, post-processing and — in casting — dewaxing expansion each hold a share. Charging all final deviation to the resin, or expecting a resin change to solve it, are the same mistake in opposite directions.

The Totus UR and URC (Ultra-Rapid Continuous) series offer three layer thicknesses: 10, 25 and 50 μm. A thinner layer means less shrinkage volume per layer and denser interlayer constraint, but the logic above does not change: the resin's levelling window and shrinkage behaviour determine which of those thicknesses the machine runs most stably at. Once a thickness is chosen, wait time, exposure and formulation should be calibrated as one set — not each optimized in isolation.

References

[1] Hoque E, Klear A, Marzana M, Golsorkhi S A, Yavitt B M. Rheological insights into vat photopolymerization 3D printing of colloidal suspensions. Additive Manufacturing, 2026, 127: 105298. DOI: 10.1016/j.addma.2026.105298. Indexed in SCI, EI, Scopus, Web of Science.

[2] Yan Y, Huang C, Shi Z, et al. Rheology-guided slurry formulation for doctor-blade recoating in ceramic vat photopolymerization: a flow–recovery framework. Additive Manufacturing, 2026, 130: 105410. DOI: 10.1016/j.addma.2026.105410. Indexed in SCI, EI, Scopus, Web of Science.

[3] Ling L, Lai T, Chung P T, Sabet S, Tran V, Malyala R. A Novel 3D-Printing Model Resin with Low Volumetric Shrinkage and High Accuracy. Polymers, 2025, 17(5): 610. DOI: 10.3390/polym17050610. Indexed in SCI, EI, Scopus, PubMed, Medline, Web of Science.

[4] Soon C W, Leong C K J, Ho T Y K, Du Z, Gan C L. Mitigation of warpage in stereolithography 3D printed ceramics via addition of sterically bulky acrylate monomer. Journal of the American Ceramic Society, 2025, 109(1). DOI: 10.1111/jace.70266. Indexed in SCI, EI, Scopus, Web of Science.

[5] Pruksawan S, Chong Y T, Zhao Y, Sivaraja V K, Ngo A C Y, Jin P, Wang F. Minimizing Polymer Curl Distortion and Heat Impact to Improve Digital Light Processing Printing Accuracy via Subdivision Method. Advanced Engineering Materials, 2024, 26(24). DOI: 10.1002/adem.202401674. Indexed in EI, Scopus, Web of Science.

[6] Yang Y, Liu B, Liu X, Li T, Wang G. Prediction of photopolymerization deformation for additive manufacturing based on thermal simulation approximation. The International Journal of Advanced Manufacturing Technology, 2025, 137(7–8): 4109–4128. DOI: 10.1007/s00170-025-15400-6. Indexed in SCI, EI, Scopus, Web of Science.

[7] Mukhtarkhanov M, Akayev S, Araby S, Shehab E, Ali M H. A novel method for evaluating thermal expansion forces during dewaxing of investment casting and 3D-printing waxes. International Journal of Lightweight Materials and Manufacture, 2025, 8(6): 716–725. DOI: 10.1016/j.ijlmm.2024.05.004. Indexed in ESCI, EI, Scopus, Web of Science.

Keywords: vat photopolymerization,resin viscosity,thixotropy,cure shrinkage,warpage,DLP,investment casting,dimensional accuracy