technical resources — fastening & adhesives
engineering guide — assembly

Attaching Magnets: Adhesives, Retention, and the Mistakes In Between

Magnets are brittle, plated with a low-energy metal surface, sensitive to heat, and constantly trying to pull themselves somewhere else. That combination breaks the standard fastening playbook. Here's what actually works — from adhesive chemistry to mechanical backup features.

why magnets break the normal rules

brittle substrate — sintered NdFeB has ceramic-like fracture behavior. no tapped holes, no press fits without careful stress analysis, no clamping in bending
low-energy surface — nickel plating is smooth and passive. adhesives that grab steel enthusiastically only develop a fraction of their strength on bright Ni without prep
thermal ceilings — cure schedules and service heat must stay below the magnet's rated temperature. a 150 °C epoxy bake will partially demagnetize an N-grade magnet rated for 80 °C
self-loading — the magnet applies force to its own joint 24/7, plus shock loads when external steel snaps into place

adhesive families that work

familyservice tempstrengthswatchouts
two-part structural epoxy−40 to 120 °C (std), 180 °C+ (high-temp)highest strength, gap-filling, chemical resistance — the default for magnet bondingfixture time 30 min–24 h; rigid grades crack under CTE mismatch — prefer toughened versions
heat-cure one-part epoxyto 180 °C+production-friendly, very strong, long open timecure temp vs. magnet grade — verify the oven won't demagnetize the part
toughened acrylic (mma)−40 to 120 °Cfast fixture, bonds through light oils, excellent on metals, peel-tolerantodor/flammability in production; exotherm on thick sections
cyanoacrylate (ca)−40 to 80 °Cseconds to fixture — great for prototypes and small magnetsbrittle, poor peel and shock, moisture-sensitive long-term. not for production structural joints
silicone / flexible−60 to 200 °Cabsorbs cte mismatch and vibration, wide temp rangelow strength — use for potting/damping alongside mechanical retention, not as the primary joint
anaerobic retaining compoundto 150 °Cideal for cylindrical magnet-in-bore joints with tight gapsneeds close-fitting joint (<0.25 mm gap) and active metal surfaces — primer required on plating
rule of thumb: a properly prepped toughened epoxy joint on Ni-plated NdFeB develops 10–20 MPa in shear. size the bond area so the working stress — including the snap-together shock load — stays under 10% of that. adhesion is a process result, not a datasheet number.

surface prep: where 80% of failures start

Almost every "the adhesive failed" return we've helped diagnose was actually a preparation failure — the bond let go adhesively, leaving one surface clean. The plating survived; the interface never formed.

degrease — wipe with isopropyl alcohol or acetone until a clean white cloth stays clean. magnets arrive with handling oils and sometimes residual polishing compound
abrade — light scuff of the bond face (fine scotch-brite or blast) breaks up the passive nickel surface and doubles typical bond strength. re-clean after abrading
prime where specified — anaerobics and some acrylics need an activator/primer on plated surfaces. epoxies generally don't, but adhesion promoters help on epoxy-coated magnets
mind the mating surface too — anodized aluminum, powder coat, and oily stamped steel each have their own prep requirements. the joint is only as good as its worse surface
bond within hours of prep — cleaned surfaces re-contaminate. build prep into the assembly cell, not a day ahead

design the joint for shear, not peel

shear — load spreads across full bond area strong ✓ peel — entire load rides one edge weak ✗ — add a pocket wall or lip
the same joint is 5–10× stronger in shear than in peel or cleavage. geometry decides which one it sees.
control the bondline — 0.05–0.20 mm is the sweet spot for most structural adhesives. use glass-bead-filled adhesive or pocket depth to set it — metal-to-metal contact (zero bondline) is a weak joint
recess into a pocket — a pocket 0.5–1× the magnet thickness converts peel loads into shear on the walls, protects edges from impact, and self-fixtures during cure
respect cte mismatch — NdFeB ≈ 5–7 ppm/°C along the axis; aluminum ≈ 23. over a wide temp range, rigid adhesive + large magnet + aluminum housing cracks something. toughened adhesives and smaller segments solve it
test the real stack — qualify with lap-shear coupons made from your plating, your adhesive, your prep, your cure — then verify with pull-off tests on actual assemblies, hot and cold

mechanical retention — with and without adhesive

countersunk magnets — the cleanest bolt-on solution: the countersink puts the screw load into compression on the magnet. use shoulder washers; torque to spec, not to feel
pockets & covers — magnet drops into a machined or molded pocket, retained by a welded/bolted non-magnetic cover or thin stainless shim. standard for high-vibration and safety-critical assemblies
retaining sleeves — rotor magnets banded with carbon fiber or non-magnetic alloy sleeves against centrifugal load; adhesive locates, sleeve retains
encapsulation / overmolding — potting the magnet in epoxy or overmolding in glass-filled nylon seals against corrosion and retains mechanically in one step. verify molding temps against the magnet grade
press fit — carefully — possible with bonded magnets and generous chamfers; risky with sintered material. if you must, use light interference, a compliant adhesive as insurance, and check hoop stress on rings
belt and suspenders: for anything safety-critical or high-vibration, the standard is adhesive plus a mechanical feature that would retain the magnet even at zero bond strength. adhesive locates and seals; geometry guarantees.

pre-production checklist

#verify
01adhesive cure temp and service temp both below magnet's rated max at its actual permeance coefficient
02bond stressed primarily in shear; peel/cleavage paths blocked by geometry
03surface prep procedure written into the work instruction, with a re-contamination time limit
04bondline thickness controlled by design, not by operator feel
05lap-shear qualification data on the real material stack, tested at temperature extremes
06mechanical backup feature present for safety-critical or high-vibration duty
07assembly fixtures prevent magnet slam-together and control placement against magnetic pull
08adhesive compatibility with magnet coating confirmed (no acidic attack on plating)
designing a magnet assembly?

We supply countersunk, chamfered, and custom-geometry magnets ready for your retention scheme — and we're happy to review your bonding approach before you cut tooling.

request a design review →