- ▸Bolts and seals are not two independent items: they sit in series on one bolt-flange-seal-flange-nut load path, so getting one end right does not save the joint if the other end is wrong
- ▸Three things must align first: medium and galvanic pairing, preload versus the seal's compression window, and continuous versus peak temperature
- ▸The compression window comes from the seal supplier; applying a bolt torque table straight to a graphite ring usually cracks it during assembly
- ▸Temperature is the only parameter that can disqualify both parts at once - settle it before strength class and coating
- ▸The value of a single purchase order is not unit price but interface ownership: missing items, wrong sizes and split shipments get reconciled by someone
Why These Two Should Be Specified Together
On a pump, a valve or a run of pipe, bolts and seals look like independent items. In fact they share one load path: tightening the bolt creates preload, preload passes through the flange face into the seal, the compressed seal rebounds, and that rebound locks the preload into the joint.
The trap is that both ends can be individually correct and the combination still wrong. A typical case: the bolts are taken all the way to class 10.9 for strength, the seal is picked on temperature, and only at assembly does it emerge that the clamping force a 10.9 joint needs to reach design preload exceeds what the seal can take in compression. The bolt is not wrong. The seal is not wrong. Nobody checked the pair.
Bought separately, that gap is structural: the fastener supplier answers for mechanical properties, the seal supplier answers for its own compression window, and the interface falls exactly between the two contracts.
Check One: Medium Decides Material and Galvanic Couple
| Service | Bolt direction | Seal direction |
|---|---|---|
| Indoor, dry, non-corrosive | Zinc-plated carbon steel 8.8 | Resin-impregnated carbon graphite |
| Wet, marine, de-icing salt | A4-80 (316) or stainless | Resin impregnation, not in direct contact with a carbon steel flange |
| Chemical, acid or alkali | A193 B7 / B16 or duplex, per the medium table | Impregnant selected for the medium; "graphite is corrosion resistant" is not a specification |
| Hot flue gas, hot oil | B7 and above | Metal impregnation or high-purity graphite |
Two points that are easy to miss:
First, carbon graphite's corrosion resistance comes mainly from the impregnant, not the graphite. The graphite matrix is inert to most media but it is porous - if the pores are not sealed, the medium travels through them. The same ring in resin impregnation and in metal impregnation can behave very differently in acid or alkali, so quoting just "carbon graphite" says almost nothing.
Second, graphite is the cathode. In contact with stainless or carbon steel it accelerates corrosion on the other side. The ring survives; the flange face does not. This shows up most in seawater, salt spray and chloride service - either insulate the joint or upgrade the flange material with it.
Check Two: Preload Versus a Brittle Part's Compression Window
A seal's usable compression is a window, not a number: too little and it leaks, too much and it cracks. Both ends of the window come from the seal supplier; the buyer's job is turning that window into a workable tightening procedure.
A common misconception here is treating the torque value as a property of the bolt. For the same bolt, dry assembly, oiled, zinc-plated and molybdenum-disulphide coated convert torque into preload at noticeably different efficiency, because friction consumes a different share. Torque therefore has to be stated together with surface finish and lubrication - a number on its own means little.
A graphite ring is brittle and particularly sensitive to uneven loading. In practice that means three rules:
- Tighten diagonally, in stages, stepping up to the target - never straight to final torque
- Use a torque wrench and record the values instead of working by feel
- Control flange face parallelism and roughness to the seal's requirement - an uneven face distributes pressure unevenly, and the crack always starts locally
Check Three: Settle Temperature First, Then the Class
Temperature is the one parameter that can disqualify bolt and seal together, which is why leaving it to the end is the easiest way to fail. Reverse the order:
1.Write down the continuous operating temperature and the peak, with peak duration and frequency - a short excursion and sustained over-temperature affect materials differently
2.Check the seal: resin impregnation has a defined ceiling, and past it the resin carbonises, stops sealing the pores and performance collapses quickly; metal impregnation goes higher, but not indefinitely
3.Then the bolt: class 8.8 covers conventional service, B7 covers medium and high temperature, stainless solves corrosion and low temperature rather than heat
4.Finally thermal cycling: repeated heating and cooling spends the seal's rebound budget, and the number of reassemblies counts too - a graphite ring loses rebound with each compression, so the third assembly no longer has the margin of the first
What a Combined Purchase Order Should State
| Item | How to write it | Cost of leaving it out |
|---|---|---|
| Bolts | Standard + size d×L + property class + finish, e.g. DIN 933 M16×80 8.8 hot-dip galvanised | Wrong class or coating arrives; whole lot rejected |
| Nuts / washers | Same standard family, class matched to the bolt (class 10 nut with a 10.9 bolt) | The nut strips first and the bolt's strength is wasted |
| Seal | ID × OD × thickness + material and impregnant + medium and temperature | "Carbon graphite" alone arrives unsuitable for the duty |
| Quantity | Units × parts per unit + spares, with the spare ratio in the contract | One or two missing pieces stop a whole unit |
| Packing and marking | Same lot in the same case, part numbers matched to the drawing | Mixed parts sorted by eye |
| Documents | Bolt mill certificate and test reports, seal material data | Cannot be produced at customer acceptance or third-party inspection |
Three Recurring Mistakes
One: torquing a graphite ring with a bolt table. Torque figures in bolt handbooks assume steel parts that can take the load; a graphite ring is not one. Work backwards from the seal's compression window to the required clamping force, then to torque.
Two: writing "per sample" for the seal. A sample proves the shape, not the impregnant or the temperature class. After sample approval, confirm material and duty in writing.
Three: splitting the shipment between two suppliers. Bolts arrive first, the seal follows; to keep schedule the bolts go in, then come out again when the seal lands. The bolts have been tightened twice, and preload and threads both pay for it - a few days of schedule bought with one assembly's worth of life.
Summary
Matching fasteners with seals means putting three parameters on one sheet: medium drives material and galvanic couple, preload drives the tightening procedure, temperature decides whether the combination is usable at all. Once all three pass, price becomes worth comparing.
The other half of matching is sourcing. If you also need impregnated carbon graphite sealing rings - resin or metal impregnated, for mechanical seals, pumps, valves, gearboxes and ice machines - the product notes and size ranges are at hhseal.cn. Send the fastener sizes and classes as described above, and the two can ship on one booking.