Every glass-lined reactor drawing carries a line that reads "cold shock ≤ 110 °C · heat shock ≤ 120 °C". Most buyers copy it into the data sheet and never look at it again. Then, two years later, a batch turns brown and a patch of glass the size of a coin is missing from the bottom head. We are KAVORTE, an independent sourcing and engineering office in Zibo, Shandong. We source glass-lined, stainless steel and PTFE-lined reactors from 50 to 30,000 L, and we do not own a factory. This is what those two numbers mean on the plant floor, and where they get broken.
1. The two numbers are temperature differences, not temperatures
Cold shock ≤ 110 °C is the maximum permitted difference between a hot glass-lined surface and the cold medium that suddenly contacts it. Heat shock ≤ 120 °C is the same figure in the other direction, when a hot medium meets a cold vessel.
Both are ΔT values. A vessel sitting at 25 °C can take 140 °C hot oil at a ΔT of 115 °C. The same vessel at 180 °C cannot take 30 °C tower water, because that is a 150 °C cold shock — 40 °C past the limit, and into the range where spalling begins.
The limits apply to the wetted glass surface and the fluid touching it. They do not describe the jacket fluid directly. Jacket heating and cooling change the wall temperature through the steel shell, which spreads and buffers the heat, so the lining follows more slowly.
2. Why cold shock has the tighter limit
Glass is strong in compression and weak in tension. Send a hot medium against the lining and the surface layer tries to expand against colder steel underneath, so it ends up in compression — which the glass tolerates well. Send a cold medium against it and the surface contracts while the bulk stays hot, putting the surface in tension. Tension is what cracks glass.
That is the whole reason 110 °C sits below 120 °C, and why the more frequent failure in service is cold shock. On the Chinese shop floor the failure has a name: 炸瓷, literally 'burst porcelain'. It is one of the most common preventable failures on a glass-lined vessel.
A crack or a chip is not cosmetic. Once the glass is gone, the steel beneath it is exposed to the same aggressive media the lining was specified to resist, and the batch is contaminated with iron at the same time. The high-voltage test figure in kV on the works certificate exists to find pinholes and thin spots before the vessel ships, not after.
Treat the two figures as a hard process constraint. Once a lining has been chipped or locally repaired, it is no longer the factory lining the figure was measured against.
3. Six places where plants break the limit
- Cold solvent or cold water charged into a hot batch. The classic case. A 160 °C reaction mass meeting a 20 °C solvent drum is 140 °C of cold shock, delivered exactly where the glass is.
- Steam or hot water cleaning of a cold vessel. Saturated steam at 0.6 MPa gauge sits near 165 °C. Against a vessel standing at 25 °C that is roughly 140 °C — over the heat shock limit inside the first minute.
- Rapid vacuum or flash cooling. Pulling hard vacuum on a hot solvent drops the wall temperature fast, and the bulk thermometer reading lags the surface.
- Winter feedstock. Material stored in an outdoor tank at 0–5 °C charged into a vessel at 120 °C is a 115–120 °C cold shock.
- Hard jacket dumps. Opening a cooling-water valve fully on a hot vessel drives the wall down quickly through the steel. Slower than a direct charge, but not free.
- Heating an empty vessel. Running the jacket hard with no liquid inside heats the shell with nothing to absorb the energy, and the differential across the lining grows.
4. Charge and cool in steps, not in one move
The working rule: never let a fluid more than 110 °C colder, or 120 °C hotter, than the wetted surface reach the glass in a single step.
- Add cold solvent in increments with the agitator running, letting each addition equalise before the next.
- Pre-warm or pre-cool the charge through a shell-and-tube exchanger instead of dumping it in raw. That is what external exchangers are for.
- Put steam in the jacket, not into the vessel. Jacket transfer is indirect and therefore gentler on the lining.
- Open cooling water slowly, and know the water temperature you actually have. 7 °C chilled water and 30 °C summer tower water are very different events on a 180 °C vessel.
- Cool below roughly 80 °C before cold wash water goes in, or warm the wash water first.
- Never add ice or cold water directly to a hot, solvent-laden mass.
Measure the surface, not only the bulk. A thermowell sitting in the middle of 3,000 L of liquid tells you nothing about the temperature at the glass 200 mm away. A wall-mounted sensor, or a jacket-side reading used as a proxy during ramp changes, is worth the cost.
5. What belongs on the data sheet and in the SOP
Confirm the thermal shock line on the drawing before you sign, and make sure it appears on the nameplate alongside design pressure in MPa, design temperature in °C, volume and the high-voltage test figure. For the lining, confirm the thickness in mm and the test value in kV — both are stated on the works certificate.
Check the media list at the same time, because thermal and chemical limits fail in the same place. Glass-lined vessels are not suitable for hydrofluoric acid, fluoride-bearing media, phosphoric acid ≥ 30 % above 180 °C, or strong alkali above pH 12 / 100 °C. A hot, strong cleaning caustic is a corrosion problem rather than a thermal shock problem, but it will find the same chip in the lining.
Configuration matters too. Type K open vessels with a removable cover and Type F closed vessels with an integral head see different temperature gradients, and an inner coil or immersion tube creates local hot spots that a plain jacket does not. State the agitator type as well: anchor and frame agitators move viscous mass slowly, so the bulk reading can sit far from the wall temperature during a ramp.
6. Takeaway
Write both numbers into the batch record and the cleaning procedure: cold shock ≤ 110 °C, heat shock ≤ 120 °C. Train operators on the difference between the two directions, and on the fact that the limit is a ΔT between the glass surface and the incoming fluid, not a bulk temperature. Log the largest temperature step of every batch, CIP included.
When a vessel arrives, verify the thermal shock figure and the media list against the drawing and the works certificate before commissioning, not after the first batch. If the process genuinely needs a step larger than 110 °C in either direction, the answer is a different heat-transfer route — a jacket, an external exchanger, a PTFE-lined or stainless vessel — not a glass-lined vessel pushed past its lining.