Multi-material, functionally graded and 4D printing
📍 Where we are: Part IV · Materials and New Systems — Chapter 9 of 19. We have looked at single alloys and their powder; now we ask what happens when one part must contain several materials, or change shape over time.
Gluing two different materials together is easy; making the join as strong as either side is hard. Think of a steel bolt cast into concrete: the bolt is strong, the concrete is strong, and the failures happen where they meet. Multi-material printing tries to place different materials voxel by voxel, ideally with a smooth hand-over between them. "4D printing" adds time: the part is printed flat and later bends or unfolds when heated, wetted or magnetised.
What this chapter covers
We consider three families: multi-material metals (graded or joined), multi-material soft matter and electronics, and shape-morphing "4D" structures. For each we separate what has been demonstrated at lab scale, what is in production, and what is contested. The question that runs through all three is the same: what happens at the interface?
Problem
Real parts want different properties in different places: a hard, hot-resistant surface on a tough core, copper where heat must be conducted away and steel where load is carried, a rigid body with embedded sensors and wires, a soft joint beside a stiff link. Conventional manufacturing builds these by joining separate parts. AM could place the materials directly, but each added material adds an interface, and interfaces are where mismatches in melting point, thermal expansion and chemistry concentrate. Reviews frame multi-material AM as placing properties such as hardness or corrosion resistance exactly where they are needed [1], and compositionally graded metals as a frontier of AM [2].
State of the art
Multi-material soft matter (lab-scale)
Skylar-Scott et al. introduced multimaterial multinozzle 3D printing, exploiting the diode-like behaviour of viscoelastic inks meeting at a junction to switch at high frequency between up to eight materials, with voxel volumes approaching the nozzle diameter cubed; demonstrations included a Miura origami pattern and a millipede-like soft robot using epoxy and silicone inks whose stiffness differs by several orders of magnitude [3]. This is a lab-scale demonstration of voxel-level programming, not a production process.
Multi-material metals
Metal interfaces fall into two kinds. In laser directed energy deposition (DED), the longer time at temperature tends to give diffuse, well-bonded interfaces but also increases solute segregation, intermetallic formation and grain coarsening; laser powder bed fusion (PBF-LB/M) tends toward sharp interfaces with low segregation but high local stress [4]. Common failure drivers are brittle intermetallics (Fe-Ti, Ti-Ni, Al-Ti), thermal-expansion mismatch and residual stress.
Examples from the literature, all lab-scale:
- In a DED graded Ti-6Al-4V to Invar 36 study cited in an LLNL review, stresses between FeTi and Fe2Ti exceeded 4 GPa and led to cracking in the as-built specimens [5].
- A sharp Fe-to-316 transition showed many pores and cracks, while a graded sample showed no cracking [6].
- But grading is not automatically safe. In multi-material PBF-LB/M, an Al2219/Ti-6Al-4V graded material showed high quantities of brittle Al3Ti, with cracks [7]. A graded path can itself cross intermetallic-forming compositions.
- Interlayers can help: depositing a nickel-based intermediary between copper and steel in DED gave a gradual transition that prevented crack-prone binary phases [8].
Reviews list the challenges for functionally graded materials (FGMs) as intermetallic generation, mismatched thermophysical parameters and varying dilution rates across gradient layers [9], and note that computational phase-diagram approaches are the most efficient way to anticipate such problems when designing the gradient [6]. The closest public example of dissimilar-metal DED at component level is NASA-led work on bimetallic liquid-rocket combustion chambers: a powder-bed copper-alloy liner combined with Inconel 625 deposited by hybrid DED, hot-fire tested with an industry partner [10], and later work describing bimetallic AM as being matured for propulsion hardware [11]. These are development programmes with hot-fire tests, not evidence of routine serial production (the two NASA-led conference papers are the only public component-level sources we found, so they are cited despite low citation counts); we found no verified report of series production of dissimilar-metal DED parts, and a 2025 CIRP Annals paper concludes that scaling from lab to industry remains the challenge [12].
Embedded and printed electronics
Printing structure and electronics together is a different problem: conductors need low resistance and fine features, structures need stiffness and accuracy, and the processes rarely overlap. A review of 3D printing and advanced manufacturing for electronics surveys the routes to printed electronics [13], and a review of conductive nanomaterials covers inks based on metal nanoparticles, nanowires, carbon nanotubes and graphene for 2D and 3D printed flexible electronics [14]. A 2023 review of printed electronics with nanomaterials discusses multimaterial and hybrid techniques and integration with discrete components [15]. Evidence level: mostly lab-scale, with niche production for printed circuit-board-like products.
4D printing
4D printing combines shape-memory or stimulus-responsive materials with AM so that a printed structure changes shape afterwards. For example, shape-memory polymer composites have been printed flat and self-morph into preprogrammed 3D shapes when heated [16], and multimaterial shape-memory polymer architectures with a resolution of a few micrometres have been printed by projection microstereolithography [17]. Early biomimetic 4D printing programmed swelling-driven shape change into printed hydrogel composites [18]. A 2025 outlook names material limitations, scalability and computational complexity as barriers to industrial adoption [19], and a 2025 article on shape-memory implants notes that (meth)acrylate resins shrink significantly on curing, lowering dimensional fidelity and shape-memory performance [20]. Evidence level: lab-scale.
Why it is hard
- Incompatible processing windows. Two metals may need different laser power, preheat and atmosphere; changing material mid-layer in a powder bed means cross-contamination, so most multi-metal work uses DED or layer-by-layer powder switching.
- Property mismatches create stress. Differences in thermal expansion, elastic modulus and crystal structure drive residual stress at the interface, which is already large in AM (Defects, residual stress and distortion).
- Phase diagrams hide traps. The designer's smooth gradient crosses compositions where brittle intermetallics form. Avoiding them needs thermodynamic calculation, not intuition (Simulation and ICME).
- Characterisation and qualification. An interface is thin, buried and variable; testing it is slow, and a certified property for a graded region barely exists (Qualification and certification).
- Resolution versus function. Conductors, dielectrics and structure each want a different process, and 4D materials trade resolution against response (resin shrinkage, for example).
Sharp, graded and interlayer interfaces compared: grading removes the step but may cross brittle intermetallic compositions; the interface is the site to watch.
Original diagram by the authors, created with AI assistance.
Open questions
- Can gradient paths be planned automatically so that they avoid intermetallic-forming regions, by combining CALPHAD calculations with printing constraints? Computational design is promising but not yet routine.
- Is a sharp interface (PBF-LB/M) or a diffuse one (DED) better for fatigue? The literature compares the microstructures, but we found no authoritative verdict, so the question is contested.
- How should an interface be inspected non-destructively, and what is its allowable?
- Can electronics survive thermal cycling when embedded in a printed structure? A 2023 study of hybridised printed electronics stresses that long-term reliability evaluation is needed, and that adhesion depended on substrate material and ink deposition angle [21].
- Which 4D materials can go beyond a laboratory demonstration to repeatable, durable actuation?
Industry impact
Multi-material AM is a demonstration-rich, production-poor area. Where it is used in practice, it tends to be where the interface problem has been designed around, for example with interlayers, or where the materials are similar enough to avoid brittle phases. For most manufacturers the pragmatic route is still to print one material and join or assemble the rest. The commercial promise of integrated electronics and graded parts depends on interface qualification, which is the same bottleneck that Qualification and certification addresses.
Why it matters
Parts that integrate functions would be a real advantage over conventional manufacture, and they are one of the few things AM can do that nothing else can. They are also where the field's optimism most outruns its data on durability.
In the real world
Harvard's multimaterial multinozzle printing is a public benchmark for soft matter. For metals, DED papers on graded Fe/Ti and Cu/steel components illustrate both the benefit of interlayers and the cracking that follows when the gradient crosses a brittle phase.
Key terms
- Interface — the boundary region where two materials meet and whose properties govern the joint.
- Intermetallic — an ordered compound of two metals, often hard and brittle, such as FeTi or Al3Ti.
- Functionally graded material (FGM) — a material whose composition changes gradually through the part.
- Interlayer — a third material placed between two incompatible ones to avoid brittle phases.
- Voxel — a three-dimensional pixel; a unit volume of printed material.
- Dilution — mixing of the deposited material with the substrate or the previous layer.
- Shape-memory polymer — a polymer that recovers a programmed shape when triggered, for example by heat.
- 4D printing — printing structures designed to change shape or function after printing.
- Structural electronics — electronics embedded in or printed on the load-bearing structure.
Where this leads
Soft, living and microscopic materials push the same problem further. Bioprinting and micro/nano-scale AM examines what is real and what is hope.