Why Can Over-Torquing the PFA Heater's Mounting Flange Lead to Premature Stress Cracking at the Thread Root?

PFA-encapsulated immersion heaters affix to tank walls or lids using mounting flanges-typically constructed of PVDF, polypropylene or stainless steel with a PFA overmold. When the flange is tightened it compresses a gasket against the tank surface to make a liquid tight seal. The installation instructions provide a torque range for the mounting bolts, however field failures commonly occur at the root of the threads in the bolt holes of the flange or the heater's own threaded fitting. These fissures occur weeks or months after installation, sometimes before the heater has run many running hours. The cause is neither a chemical attack or a heat degradation but a mechanical over straining during the installation of the . PFA, unlike metals, shows strong viscoelastic creep and no well-defined yield point. If an installation uses a torque greater than the specified range, the PFA surrounding the thread root is subjected to localised cold flow and tiny crazing occurs. Under the combined effects of residual installation stress and subsequent temperature cycling these crazes develop into complete cracks. A better understanding of this failure process involves investigation of the stress concentration at the thread roots and time-dependent failure behaviour of fluoropolymers.

Stress concentration at thread roots under compression
A threaded junction in a PFA flange or fitting presents a natural stress raiser at the root of each thread, at the acute radius where the thread profile meets the bolt shank. The stress concentration factor (Kt) at the sharp thread root for metal threads is 2.5 to 4.0, so that the local stress is 2.5 to 4.0 times the nominal stress derived from the bolt load. PFA, having lower modulus and better ductility, has even higher effective concentration since the polymer cannot load share through localised yielding as effectively as metal. Finite element study of an M12 thread in a PFA flange with a bolt torque of 15 N.m (typical for a 50 mm flange) indicates a peak main stress at the root of the thread of 22-28 MPa compared with a nominal compressive stress on the face of the flange of only 3-4 MPa. This stress concentration of 6–8 times the nominal occurs as PFA deforms elastically under the bolt head and preferentially transfers force to the first engaged thread.

If an installer over-torques-25 N·m instead of 15 N·m, the peak tension at the root of the thread increases non-linearly. The PFA starts to locally yield at roughly 12 to 14 MPa (its compressive yield strength at room temperature). After yielding, the stress-strain curve hits a plateau where further strain takes place without much increase in stress. But the shape of the thread root does not permit yielding uniformly. First the material at the root gives way, and then the yielding zone spreads out. The plastic zone spans roughly 1.5–2.0 mm from the thread root at 25 N·m. Upon removal of the torque, the elastically deformed zones try to recover, but the plastically deformed zone does not return to its previous shape. This mismatch leaves a residual tensile tension at the root - the polymer is effectively stretched over an undersized root radius. The residual stress level, evaluated by photoelastic methods on transparent PFA samples, is 4–8 MPa after one over-torque event. This residual stress alone is not sufficient to produce breaking at once but acts as a pre-load which adds to operating stresses.

Transition from Crazing to Cracking with Thermal Cycling
As with many semicrystalline polymers, PFA develops crazes instead of traditional cracks under modest tensile stress. A craze is a thin band of strongly orientated polymer fibrils extending over a voided region, typically 1–5 µm in width and 50–200 µm in length. Crazes are energy absorbing and can stabilise a fault for thousands of load cycles. When the fibrils in the frenzy break, usually under some extra force or interaction with the environment, they form real cracks. In a correctly torqued flange (torque is within the prescribed range), any crazes formed at the roots of the threads will remain stable for the projected life of the heater, as the residual tension from installation plus the thermal cycle stress is below the craze-to-crack transition threshold . The two elements that facilitate transition to breaking of an over-torqued flange are: First, the residual stress is larger (4-8 MPa vs. 1-2 MPa for optimal torque) allowing less tolerance before the crazing stress is exceeded. Second, the larger plastic zone results in a longer defect that initiates the frenzy at the thread root. The stress intensity factor at the craze tip is proportional to the square root of the craze length, therefore a longer initial craze requires less extra force to propagate.

The increased tension due to thermal cycling turns stable crazes into propagating cracks. Each time the heater cycles from ambient to operational temperature (e.g., 25°C to 90°C for an acid etching bath), the PFA flange expands more than the metal bolt. This is because PFA's coefficient of thermal expansion (100–120 ppm/°C) is about six times that of stainless steel (17 ppm/°C). This differential expansion induces an extra tensile stress in the PFA surrounding the bolt, in addition to the residual installation stress. For 50 mm flange and 4 x M8 bolts, a temperature increase of 70°C results in an additional 5-7 MPa tensile stress in the PFA around the bolt hole. In a well torqued flange the overall stress (residual 1 - 2 MPa + thermal 5 - 7 MPa) is 6 - 9 MPa, below the 10 - 12 MPa threshold for craze-to-crack transition for most PFA grades. Total stress for an over-torqued flange is 9–15 MPa (residual 4–8 MPa plus thermal 5–7 MPa), which is over the transition threshold. After 100-500 thermal cycles cracks develop from the thread root and progress across the flange thickness in the next cycles. Typical through-wall cracking needs 1,000-3,000 cycles from the beginning of the craze-to-crack transition, depending on the severity of over-torquing and operating temperature.

Detecting Over-Torque Damage Before Crack Growth
Damage caused by over-torquing is not obvious on the surface of the outer flange immediately after installation. The crazes and plastic deformation are limited to the bolt hole at the thread root and are not detectable by visual inspection. There are however, three non-destructive approaches that can detect over-torqued PFA flanges before the cracks reach the surface. The first technique is the measurement of the bolt torque retention. Install flange to specified torque, mark position of bolt head relative to flange, then re-check torque after 24 hours at room temperature. Over-torqued and excessively-creeped PFA will lose 30–50% of the initial torque within 24 hours when the polymer cold flows under sustained compression. Properly torqued flange loses just 5-15% of original torque. If there is no apparent leakage and the torque loss is greater than 25%, the flange has probably been over-torqued in this installation or before. The second option is to place a bore scope or fibre optic probe into the bolt hole after the bolt has been removed. Under magnification, the thread roots show whitened patches, a process called stress whitening, which are the result of micro-void development. Over-torqued thread roots exhibit a continuous white band along the whole root of the first two threads. Roots Torqued properly have no Whitening or isolated tiny white patches.

The third technique is electrical continuity testing when the flange has an embedded ground plane. This is a common configuration for PFA heaters for semiconductor applications. Over-torquing causes cracks or severe crazing through the flange thickness allowing moisture access to the ground plane. Measure the resistance between the ground plane and a reference electrode in the tank liquid. Properly sealed flange will indicate resistance of > 1,000 MΩ. A flange with a fracture beginning at a thread root exhibits 100–500 MΩ. The partial through wall cracking flange is 1-100 MΩ. This test only applies to flanges with a continuous conductive layer and not all-PFA flanges. The torque retention method is the most reliable field diagnostic method for all-PFA flanges.

Torque Limit Recommendations and Prevention of Failure Guide
The table below provides suggested maximum torque values for typical PFA flange and thread sizes, and the effects of exceeding these limitations. Values are for a clean, dry thread without lubrication (lubricated threads reduce friction and increase preload for the same torque, therefore a 20 - 30 % torque reduction is needed). All recommendations are for PFA at room temperature during installation. Torque values should be lowered by 15% if the flange temperature is above 40°C during installation.

Flange or Fitting Type Thread SizeRecommended Torque Range (N·m) Torque Limit (N·m) Consequence of Exceeding Limit Expected Cycles to Cracking at Limit PFA-encapsulated flange, 50 mm diameter M6 4 to 6 10 Moderate over-torque: stress whitening at thread root, 25 to 35% loss of torque in 24 hrs 500-1,000 cycles ΔT=70°C
PFA encapsulated flange, 80 mm diameter M8 8-12 18 Moderate to severe Visible crazing at root Torque loss >40% 200-500 cycles
PFA encased flange, diameter 100 mm M10 12–16 22 Severe: plastic deformation of flange face, cracks develop at first thread 50–200 cycles
Solid PFA threaded fitting (NPT or BSP) 1/2 inch (DN15) 3–5 8 Thread stripping risk, stress cracking at thread root in days 10–50 cycles (immediate failure common)
Solid PFA threaded fitting 3/4 inch (DN20) 5-8 12 Cracking of flange, fitting may seize on bolt20-100 cycles

Preventing over-torque damage involves tool selection and training.