×
A picture reached me this week. A technician opens the milling chamber on a finished job: a molar crown, still violet in its pre-crystallized state, low-translucency A2. The crown is perfect. It is also lying loose in the chamber. The steel mandrel sits beside it, bare.
Nothing broke. The restoration simply stopped being attached to anything.
Every lab has a version of this story. The cost is rarely just the block — it is forty minutes of machine time, a re-nest, and occasionally a patient appointment that has to be moved. What frustrates people most is that it feels random.
It isn't. Whole-body detachment has a small number of causes, and the failed part will tell you which one you're looking at if you know where to look.
The load path, and its three weak points
A milled restoration is held in the machine by a chain: collet → mandrel → adhesive joint → block body → residual connector → crown. From the outside, a failure anywhere in that chain looks identical. Underneath, it is not.
Read the fracture surface before you blame the mill
Ninety seconds with a loupe settles most arguments.
Rows one and two belong to whoever bonded the block. Row four is where the block itself enters the investigation — but connector geometry and bur condition still need to be ruled out first. Row three is shared, and it is the interesting one.
Why glass-ceramic is less forgiving than zirconia
We build zirconia green bodies, so the contrast is one we live with daily. A pre-sintered zirconia blank is a compacted powder body. It cuts by microfracture and grain pull-out, it is porous, it damps vibration, and it forgives a surprising amount of abuse.
Pre-crystallized lithium disilicate behaves nothing like that:
• It is dense glass with dispersed lithium metasilicate crystals. Flexural strength in the machinable state sits in the region of 130–150 MPa, rising to roughly 360–500 MPa only after the crystallization firing. Everything that happens in the mill happens at the low number.
• It is elastic, and it stores energy. Where zirconia absorbs a chatter event by crushing locally, glass-ceramic springs — and delivers that energy straight into the connector and the bond line.
• It is a silicate glass being cut under water. Silicate glasses undergo subcritical crack growth when water reaches a loaded crack tip. A flaw that survives the cut can keep extending quietly afterwards.
That last point explains something technicians report constantly and rarely believe: the part was fine when milling ended, and came off on the bench four minutes later. Delayed failure is a signature, not a mystery. It tells you a crack was already there.
What has to be controlled upstream
If the fracture runs through the ceramic, the conversation moves to the block maker. Four things decide whether a block ships with a strength-limiting flaw inside it:
Melt homogeneity. Cord, striae, seed bubbles and refractory inclusions picked up from furnace lining are all Griffith flaws waiting for a tensile field. A block is only as strong as its largest defect, not its average quality.
Crystallization uniformity. The nucleation and growth schedule has to arrive evenly at the centre and the corners of every block in the load. Where it doesn't, you get zones of differing hardness — and cutting force that changes mid-pass, which is exactly what excites chatter.
Residual stress. Ingots cool from the outside in. Annealing that is shortened to gain throughput leaves locked-in tension. Milling removes material asymmetrically and releases it, which is why some blocks crack with no obvious provocation.
The bond, treated as manufacturing rather than assembly. Stub degreasing and roughening, primer, adhesive chemistry, bond-line thickness, cure verification, and destructive pull-off testing on every lot — with the results kept and traceable to the batch.
Where ICERA sits
We came to glass-ceramic from powder metallurgy and zirconia green-body manufacturing, and we brought the same manufacturing discipline with us: batch traceability, controlled heat treatment, and systematic quality verification throughout production.
Our ICERA lithium disilicate CAD/CAM blocks achieve a post-crystallization three-point flexural strength of 420 ± 60 MPa, with a fracture toughness of 2.8–3.8 MPa·m¹ᐟ² and an elastic modulus of 50 ± 10 GPa. Linear shrinkage during crystallization is approximately 0.2%.
Those numbers matter because material consistency is not an abstract specification.
Flexural strength tells you how much stress the material can withstand.
Fracture toughness tells you something different — how resistant it is to an existing crack continuing to grow.
And in a material that is being milled in its pre-crystallized state, both matter.
A block is not only as good as its average strength.
It is only as strong as its largest defect.
That is why we focus not only on the final strength figure, but on the manufacturing variables behind it: melt homogeneity, crystallization consistency, residual stress, and batch traceability.
Those are the parts of this problem a material supplier can actually own.
Five things worth checking today
1. Bur wear. A dull diamond doesn't cut — it pushes. Track cycle counts rather than judging by eye.
2. Coolant. Flow, filtration, and additive concentration. Starved coolant means heat at the joint.
3. Collet condition. Runout and grip. Vibration you can hear is vibration the connector is absorbing.
4. Storage. Heat and humidity age adhesive joints long before they age the ceramic.
5. Keep the failed part. Both halves, dry, in a bag. It is evidence.
If you have a detachment you can't explain, send us the crown and the mandrel. We'll put them under the microscope and send you the fractography, whoever made the block.
EN
