Markforged Onyx is a short-carbon-fibre-filled polyamide supplied as a filament for fused filament fabrication systems in the Markforged product family. The designation identifies the material, not a single printer model. Compatible desktop and industrial platforms historically include the Onyx One, Onyx Pro, Mark Two, X3, X5, and X7, though regional availability and current production-series support vary. Onyx is formulated as a polyamide matrix containing dispersed micro-carbon fibres. It is not a continuous-fibre feedstock and is not sold as injection-moulding pellets; it is delivered on sealed spools and processed through Markforged toolheads under closed parameter control.

The following property matrix is supplier-reported for xy-oriented printed coupons and should not be treated as design allowables without orientation-specific validation.

The reported values are xy-plane measurements and do not represent isotropic mechanical performance. Short carbon fibres align along the extrusion vector, so the 2.4 GPa tensile modulus and 37 MPa tensile stress at break are toolpath-dependent. Markforged does not publish full z-direction tensile data for Onyx. In a material-extrusion part, layer-interface strength is generally lower than in-plane strength because interlayer diffusion is limited by the thermal history of the weld region. Loading through the build direction therefore requires tensile validation on the target machine and toolpath.

Onyx differs from unfilled nylon 6/6 primarily in flatness retention and dimensional stability after cooling. The dispersed carbon phase constrains gross part warpage and improves edge stability in large, flat fixtures, but the absolute stiffness increase is modest because the matrix remains a short-fibre composite. Markforged does not disclose the carbon loading, fibre aspect ratio distribution, or the specific polyamide grade used. In comparison with injection-moulded short-carbon nylon compounds, Onyx is not a pelletized material intended for twin-screw extruder compounding or conventional injection-moulding production. No melt mass-flow rate under ISO 1133-1:2022 and no viscosity curve are published for Onyx. Processing conditions are generated by the Markforged slicing environment rather than by independent screw-speed or barrel-temperature adjustment.

How Does Onyx Compare With Unfilled Nylon 6/6 in Moisture and Shrinkage Behaviour?

The polyamide phase of Onyx remains hygroscopic. In unfilled nylon 6/6, absorbed water acts as a plasticizer, lowering stiffness and shifting mechanical response. Onyx contains a carbon phase that reduces the volume fraction of moisture-absorbing polymer, but the matrix still takes up water at exposed surfaces and layer interfaces. If a spool is left above 60% RH for extended periods, absorbed moisture can reduce interlayer fusion and produce steam evolution at the extrusion nozzle. Conditioning for mechanical testing is commonly performed at 23 °C and 50% RH according to ASTM D618-21, but printed composites may not reach equilibrium on the same timescale as an unfilled resin. Published equilibrium moisture data for Onyx is limited. The 2.5 wt% to 3.0 wt% moisture uptake values typical of unfilled nylon 6/6 at 23 °C and 50% RH cannot be applied directly to the composite because the carbon phase is non-absorbing and changes the matrix fraction.

When pre-drying is necessary, the standard practice is to dry the spool in a desiccant-backed chamber at a set point below the filament softening point until the spool mass stabilizes. Markforged does not publish a loss-on-drying specification for Onyx; a fixed time-and-temperature routine without mass verification is less reliable than weight-loss monitoring. On open desktop systems without full active chamber heating, long and flat Onyx parts can still show corner lifting. Adhesion control, brim construction, and build orientation are used to recover flatness. The carbon fibre filler is abrasive. Generic brass nozzles on non-Markforged hardware are not appropriate for extended runs; hardened steel or equivalent abrasion-resistant tooling is required if open-platform processing is attempted.

When Continuous Carbon Fibre Reinforcement Becomes Necessary Under Service Loading

Onyx is also the matrix phase for Markforged continuous-fibre reinforcement. The supplier’s continuous carbon fibre data reports tensile stress at break of 800 MPa and tensile modulus of 60 GPa in fibre-dominated test coupons. Continuous fibreglass reports 590 MPa tensile stress at break and 21 GPa tensile modulus. These values do not transfer directly to a full printed part. Continuous fibre is deposited in selected layers along specified paths, while unreinforced Onyx regions still fail at the matrix limit of 37 MPa. Effective reinforcement requires principal stress trajectories to be identified and followed without interruption. A bracket loaded along the fibre path can approach fibre-dominated stiffness, but the same bracket loaded transverse to the fibre or across an unreinforced wall remains near the matrix properties. Z-direction blind features, thin ribs, and small boss intersections often cannot be continuously reinforced; they remain governed by printed weld strength. Published data for interrupted continuous-fibre geometries is limited.

A production checkpoint fixture used to locate automotive lamp housings can be printed in unreinforced Onyx when the locating force is compressive and the ambient temperature is below the material’s load-bearing deflection range. The carbon filler provides better hole-to-hole stability than an unfilled fine-extrusion polyamide under similar toolpaths, and the 75 Shore D surface resists marring of painted surfaces under modest contact pressure. If the fixture includes gas springs, high-force toggle clamps, or screw-driven clamping, continuous-fibre reinforcement or metal bushing inserts are installed in the reaction zones. The base grade is not selected for static-dissipative benches or tape-and-reel contact plates because baseline Onyx has no published controlled surface resistivity.

Onyx can be drilled, tapped, and finish-machined. The carbon filler increases tool wear relative to unfilled nylon. Carbide or polycrystalline diamond cutters reduce edge breakdown and replacement frequency. Low cutting speed and controlled chip extraction prevent local melt smearing of the polyamide matrix. Repeated threaded assembly is typically supported with press-in or heat-set threaded inserts rather than direct tapped holes because layer interfaces can crack around coarse threads under high preload. Internal channels produced in Onyx have as-printed layer-step roughness; they are not specified as leak-tight sealing surfaces without post-finishing or liner components.

Moisture Absorption, Pre-Drying, and Sealed Storage Ceilings

The hygroscopic polyamide phase absorbs atmospheric moisture in storage and service. Markforged supplies Onyx in sealed, desiccated spools to maintain dry as-packaged conditions. Once the barrier is opened, storage below 60% RH in a sealed bin with desiccant is a minimum control. Filament defects such as surface stippling, reduced part density, or steam evolution at the nozzle indicate that drying is required. Direct published water absorption values for Onyx are limited; conventional polyamide drying cycles should not be assumed without mass-loss verification. Because the matrix is polyamide-based, exposure to hot water, steam, strong mineral acids, or concentrated oxidizing agents promotes hydrolysis and degradation. Chemical compatibility should be screened under ASTM D543 or an equivalent immersion standard before the material is specified for aggressive service.

The heat deflection temperature of 145 °C does not establish a continuous service ceiling. It reflects a short-term deflection condition under 0.45 MPa flexural stress in ASTM D648-07 Method B. Sustained loads at elevated temperatures can produce creep and dimensional drift before the heat deflection temperature is reached. Users specifying under-hood tooling or process-heated fixtures should test creep and recovery on printed specimens because published creep data for Onyx is limited. For long outdoor exposure, the standard datasheet does not specify an ultraviolet stabilization package; the polyamide surface may show chalking or fibre bloom unless post-coated.

Onyx Is Not a Substitute for Onyx FR or Onyx ESD in Regulated Environments

Baseline Onyx carries a UL 94 HB horizontal burn rating and is not formulated as a flame-retardant grade. Onyx FR is a separate flame-retardant variant. Compliance for transportation interiors, electrical enclosures, or other regulated installations must be established against the end-use standard and thickness rather than assumed from the base composite. Similarly, Onyx ESD is formulated for static-dissipative tooling and packaging, whereas baseline Onyx has no published controlled volume or surface resistivity. Applications requiring controlled electrostatic discharge should follow IEC 61340-5-1 and validate the printed surface after any post-machining or coating operation because cutting can expose carbon-fibre-rich regions and change local surface resistivity.

The flame-retardant and electrostatic-dissipative variants contain different additive packages and may differ in mechanical properties, process settings, and drying behaviour. They are distinct products rather than simple performance upgrades of the same formula. In production environments with regulated chemical or electrical exposure, substituting base Onyx for a functional variant without requalification introduces compliance risk.

The melt behaviour of short-carbon-filled polyamide differs from that of neat nylon. The filler raises apparent viscosity and increases shear-thinning character, but Markforged has not published cross-model viscosities, modified Arrhenius shift factors, or Rabinowitsch-corrected flow curves for Onyx. The closed toolhead architecture removes operator-level rheology correction; however, it also means that external melt-index testing under ISO 1133-1:2022 is not a useful process-control parameter for this material. A processor comparing Onyx with a pelletized short-carbon nylon run on a twin-screw extruder with an L/D ratio above 32:1 must account for the orientation induced by extrusion deposition rather than injection-moulding flow morphology.

Batch-to-batch consistency on Markforged platforms is managed by sealed filament packaging and spool identification, with the printer applying the matched parameter set from the slicing environment. This is a process-control feature rather than a direct property certification. Aerospace or medical use still requires lot-test coupons and traceability because the public data sheet does not provide a full statistical design-allowable basis such as CMH-17 or a regulatory-approved material allowable set.

Post-machining of carbon-filled nylon generates fine composite dust, which should be controlled with local extraction. The polyamide matrix may emit decomposition vapours if machining temperatures rise above the polymer degradation threshold. Sharp cutting edges with low feed rates reduce frictional heat. No specific occupational exposure limit for Onyx dust is published by Markforged; general controls for nuisance dust and carbon-fibre particulates apply.

In robotic gripper jaw geometry printed at 100 µm layer height, Onyx provides a mass-efficient alternative to aluminium when clamping loads are low. The mechanical ceiling is defined by the 37 MPa xy tensile limit and by unreinforced layer-weld boundaries. The 25% tensile strain at break reported for xy coupons does not imply equivalent ductility in z-oriented pull-out features; such features can delaminate at lower apparent strain. Gripper fingers with snap-fit retention or threaded inserts should be tested with worst-case layer orientation and repeated clamping cycles before series deployment.

In composite tooling bodies for metrology, the short carbon fibre filler lowers thermal distortion relative to unfilled nylon, but stainless steel or ceramic inserts provide the final precision surfaces. Onyx is machined or bonded into the fixture body; the as-printed surface should not be expected to hold micron-level datum geometry without post-machining. The operational boundary for metrology fixtures is therefore set by the dimensionally stable composite body combined with hard reference elements, not by the printed surface alone.