FTW98FA364

Shortly after the airplane took off, the crew heard a loud noise and felt a slight vibration from the left propeller. They returned and landed. It was discovered that the propeller erosion shield had debonded and a portion had penetrated the passenger cabin just aft of the airstair door, knocked ajar Seat 1A Passenger Service Unit, ricocheted off the ceiling, and penetrated the inner window pane at Seat 2C. Laboratory examination of the propeller blade revealed an area 'consistent with adhesive disbonding (debond) on about 80% of the total bond area...' Three hundred hours before the accident, the blade had been overhauled and repaired. The repaired area contained numerous voids, air bubbles, dirt and debris. Evidence of chaffing (rubbing) was found at several locations in the debonded area, indicative of movement between the debonded erosion shield and the propeller blade body. Samples of the erosion shield adhesive were tested and found to have been mixed in improper proportions, reducing bond strength.

History of flight

On August 19, 1998, at 1715 mountain daylight time, a Beech 1900D, N251ZV, owned by Mesa Airlines, Inc., of Farmington, New Mexico, and operated by Great Lakes Aviation, Ltd., of Spencer, Iowa, as United Express flight 1605, was substantially damaged when the erosion shield separated from one of the left propeller blades and punctured the pressure vessel during initial climb at Denver, Colorado. There were no injuries to the airline transport-rated captain, commercial certificated first officer, and 12 passengers, but one passenger sustained a minor injury. Visual meteorological conditions prevailed, and an IFR flight plan had been filed for the scheduled domestic passenger flight operating under Title 14 CFR Part 135.

The airplane taxied from its gate at 1710, and took off from runway 17R at 1714. The following was contained in the captain's discrepancy report: "Reaching 500 feet agl [above ground level] after takeoff, heard loud 'bang.' Slight vibration from left prop. Returned for landing. Discovered hole in fuselage from prop leading edge strip." The airplane returned to Denver and landed on runway 17L at 1719 and arrived at its gate at 1725.

Postaccident examination revealed the entire erosion shield was missing. There were three holes punctured in the fuselage. The first hole was about 12 inches aft of the airstair door and just above the cabin window. The second hole was slightly higher than the first. The third hole was at the top of the fuselage. A portion of the erosion shield entered the cabin and knocked out the overhead passenger service unit at seat 1A. Two small dents were found in the ceiling. The inner window pane at seat 2C was punctured and a piece of the erosion shield was found lodged between the inner and outer panes. Another piece was found between the ceiling and the pressure vessel. The passenger seated in row 2C sustained a small cut on his forehead when he was grazed by a piece of the shrapnel. He refused medical treatment.

Aircraft information

N251ZV (s.n. UE-251), a Beech 1900D, was manufactured by the Raytheon Aircraft Corporation in 1996. It was equipped with two Pratt and Whitney PT6A-67D engines, driving two Hartzell HC-E4A-31 composite propellers. The propeller blades are made of Kevlar, fiberglass, foam and epoxy. A metallic erosion shield is bonded to the leading edge of the propeller blades, and de-icing boots are bonded to the inboard end of the erosion shield and propeller blade body.

According to the aircraft maintenance records, the four propeller blades were last overhauled and repaired by Aircraft Propeller Service, Inc., in July 1998.

Tests and research

The four composite blades from the left propeller were removed from the airplane and sent to NTSB's Materials Laboratory in Washington, D.C., where, on August 27, 1998, they were examined. According to the Materials Laboratory's Factual Report (No. 99-23, attached), the surface of the damaged blade (s.n. 811) was smooth and shiny, "consistent with adhesive disbonding (debond) on about 80% of the total bond area...The remaining outboard separation area had a much rougher powdery white appearance typical on cohesive separation within the bonding material."

The report further stated that the propeller blade (s.n. 811) had been repaired by injecting adhesive into the joint. This repair reportedly occurred approximately 300 hours before. The repaired area was readily apparent during the examination because it contained numerous voids, air bubbles, dirt and debris. Evidence of chaffing (rubbing) was found at several locations in the debonded area, indicative of movement between the debonded erosion shield and the propeller blade body.

The other three blades (1802, 808, and 806) appeared to be undamaged when visually inspected, but tap tests indicated approximately 80 to 90% debonding of the erosion shield under the deice boots on blades 1802 and 808. Blade 806 had "spotty" debonding. When the deice boots were removed, the erosion shields on blades 808 and 806 were found to be cracked.

Samples of the erosion shield adhesive were removed and forwarded to the manufacturer for analysis. Using differential scanning calorimetery (DSC), the samples were found to be "fully cured," and contained no "environmental contaminants" such as water or solvents. Further tests of the adhesive, Hysol EA 9330, showed that the cured mixture contained approximately 120 to 140% (by weight) of the recommended amount of part B of the epoxy. A mixing ratio of 100:33 of part A (hardening components) to part B (catalyzers) is recommended by the data sheet.

In a report prepared for Hartzell Propellers by the University of Dayton Research Institute, it was found that decreased bond strength would result if Hysol EA 9330 was improperly mixed (such as a 20 to 40% excess of part B). Oil and dirt contamination detected on the blade surfaces would also reduce bonding strength of the joint. However, it could not be determined if the contamination was introduced before or after the erosion shield debonded. Finally, the report noted that exposure to high temperatures could cause a separation.

Additional information

The airplane was released to Great Lakes Aviation for repairs on August 20, 1998. After being examined at NTSB's laboratory, the composite blades from the left propeller were sent to Hartzell Propellers. Hartzell Propellers then sent four new replacement propeller blades to Great Lakes Aviation on September 1, 1998.

In addition to the Federal Aviation Administration, parties to the investigation included the Raytheon Aircraft Corporation, Great Lakes Aviation, Hartzell Propellers, and Aircraft Propeller Service, Inc.