Figure 1-3
Figure 4-6
Figure 7-9

Introduction - Sounding the Alarm
At the 1998 ASNT Fall Conference, Albert Broz, chief scientist of NDT for the US Federal Aviation Administration (FAA), presented a paper titled, "Fluorescent Penetrant Inspection Lessons Learned." The paper's published summary included these alarming statements: "Recent incidents of engine disc failures have been attributed to the inconsistent application of standard fluorescent penetrant inspection guidance" and "The National Transportation Safety Board has attributed the cause of a number of engine failures to missed fluorescent penetrant inspection opportunities."

It is true that many variables prevent the penetrant process from being foolproof, and "missed fluorescent penetrant inspection opportunities" must be anticipated. Nevertheless, the author, having 40 years of experience in the industry, cannot recall such damaging criticism as that leveled by Broz. Especially disturbing is that the "missed fluorescent penetrant testing opportunities" cited above involve critical turbine engine parts whose failure has led, or could lead, to a catastrophe.

---

When vapor degreasing was the first step in the penetrant process, parts emerged hot from the degreaser. 

---

Such statements from this authority should alert the penetrant industry that something is wrong. The industry — which includes manufacturers of penetrants, users of penetrants, and those who dictate what to use and how to use it — has been challenged to improve the sensitivity and reliability of the penetrant method for inspecting critical turbine engine parts. The industry cannot be lethargic. It must develop a process that will detect anomalies in such critical turbine engine parts as discs, fan blades, and compressor blades at a higher level of confidence.

 

The Disappearance of Vapor Degreasing
To what can we attribute these complaints? What has changed in recent years? One thing stands out: the virtual elimination of vapor degreasers due to environmental concerns. Vapor degreasing is not outlawed, per se — a solvent other than methyl chloroform may be used in special equipment designed to meet EPA/OSHA requirements.

When vapor degreasing was the first step in the penetrant process, parts emerged hot from the degreaser. And, although some cooling may have taken place, penetrant was applied to the parts while they were still at an elevated temperature.

In the days of vapor degreasing, methyl chloroform (1,1,1-trichloroethane) was the accepted solvent until it was found to deplete the ozone layer and its manufacture was prohibited. During usage, the temperature of this solvent's vapor was 75 °C (167 °F), the solvent's boiling point. Parts to be tested would be placed in the degreaser's vapor zone, and the vapor would condense on the colder part and return to the reservoir, carrying organic contaminants. When the part's temperature reached that of the vapor, condensation stopped, the part would be removed, and penetrant would be applied.

There were (and are) specifications restricting penetrant application to surfaces over 49 °C (120 °F). Even if such specifications were observed, 49 °C (120 °F) is still 22-28 °C (72-82 °F) higher than usual ambient temperatures. Consequently, where vapor degreasing was in use, heat assisted fluorescent penetrant testing was practiced, albeit unintentionally.

 

Heat May Increase Fluorescent Penetrant Testing Reliability
One available potential means of increasing confidence in the testing of critical parts is to use heat during the penetrant dwell step. Twenty years ago, in 1979, a technical paper "Heat Assisted Fluorescent Penetrant Inspection" was presented at the ASNT Spring Conference in San Diego and published in Materials Evaluation in September 1979. Considering the current FAA challenge to improve the penetrant process and considering the data presented in this 1979 paper, it is appropriate that the subject of heat assisted fluorescent penetrant processing be revisited, and that the industry be reminded of its advantages.

Heat assisted penetrant processing, as outlined in 1979 paper, calls for the test piece to be heated to a desired temperature, such as 71 °C (160 °F), and for penetrant to be applied and allowed to reach the temperature of the test piece. The objective is to elevate the temperature of both the penetrant and the part. If the penetrant is applied by immersion rather than by spraying or brushing, the test piece should be removed immediately from the penetrant tank to avoid cooling. The penetrant's temperature must increase to the predetermined level for the test to be completely successful.

Sherwin and Holden explain that it is theoretically logical that applying heat will materially improve the penetrant process. As stated in 1979, the reasons are as follows:

• Penetration is dependent on penetrant movement (molecular motion) and the kinetic molecular theory postulates that the average velocity of molecules increases as temperature increases.
• One study determined that the rate of penetration is inversely proportional to the viscosity; the lower the viscosity, the more rapid the penetration. And viscosity decreases as temperature increases.
• Heat vaporizes anomaly entrapped solvents and moisture which might otherwise interfere with penetration. It liquefies heavy oils and waxes present from previous processing, facilitating penetrant displacement" (Sherwin and Holden, 1979).

The reasons why heat improves penetrant performance still apply: heat
energizes the penetrant, lowers its viscosity, and facilitates removal or displacement of crack contaminants.

For the 1979 study, a penetrant with a viscosity of 44.9 x 10⁶ m²/s (44.9 centistrokes [CS]) at 16 °C (60 °F) had a viscosity of approximately 5.0 x 10⁶ m/s (5.0 CS) at 71 °C (160 °F). Everything else being equal, a 5 x 10⁶ m²/s (5.0 CS) penetrant is far more likely to enter
an anomaly and fill its reservoir than a 44.9 x 10⁶ m²/s (44.9 CS) penetrant. Furthermore, as the part and penetrant cool during the penetrant dwell, the penetrant's viscosity increases. A higher viscosity penetrant is more likely to resist over removal than a low viscosity penetrant. Thus, heat assistance gives the best of both worlds: low viscosity for quicker, more complete penetration, and high viscosity to resist over removal. A single penetrant provides both higher sensitivity and increased reliability when heat is introduced and followed by cooling.

The 1979 paper was motivated by results obtained with high temperature penetrants — both visible and fluorescent. It was observed that high temperature penetrants, under many conditions, revealed anomalies with more authority than when processed with conventional penetrants at room temperature; "In effect, this investigation was a continuation of work done in developing penetrant testing materials for use on elevated temperature surfaces, such as multipass weldments with 177 °C (350 °F) preheat temperatures. In this effort, it was discovered that heat, in many instances, improved penetrant performance" (Sherwin and Holden, 1979).

 

Findings from 1979
Cracked aluminum blocks, sheared into two sections, one heated and the other held at room temperature, were among the test pieces used in the 1979 experiments. These experiments were meant to confirm heat's marked effect in assisting penetration when the discontinuities are contaminated with foreign substances such as cutting compounds, ultrasonic couplers, and oils (see Figures 1-8).

Figure 1 shows a cracked aluminum block, sheared into two pieces, one heated to 93.3 °C (200 °F), and the other unheated. The same conventional visible penetrant (type II) and nonaqueous developer were applied to both sections. The heated section shows a more complete and pronounced crack pattern when compared to the unheated section.

Figure 2 depicts a cracked aluminum block, sheared into two equal sections, and contaminated with a difficult soil, in this case a wax like substance. One section was inspected at ambient temperature with a conventional visible dye penetrant (type II) system; the other section was inspected at 177 °C (350 °F) with a special high temperature visible dye system. The results speak strongly in favor of heat assistance.

Figure 3 compares the performance of a method A (water washable), level 3, fluorescent penetrant (type I) system on a similar sheared aluminum block contaminated with a visible dye penetrant. One section was inspected to 93 °C (200 °F) and the other to 20 °C (68 °F). It is readily apparent that the use of heat improves performance of a fluorescent penetrant confronted with exposing contaminated surface anomalies.

While cracked aluminum blocks are useful to illustrate the power of heat assistance, the principal test piece for the 1979 paper was a set of 12 special PSM-5 panels. The difference being, instead of five cracks in ascending gradation from 0.76-6.35 mm (0.03-0.25 in.) in diameter, as are found in conventional PSM-5 panels, the five cracks were all in the 0.76 mm (0.03 in.) range. The plating thickness was a minimum of 0.08 mm (3 x 10^-3 in.) and probably closer to 0.12 mm (5 x 10^-3 in.).

Figure 4 illustrates the difficulty of finding the cracks in panel 10 in the set when processed at ambient temperature. The penetrant used was equivalent to type I, level 4, method D, a nonwater washable fluorescent penetrant removed using a hydrophilic emulsifier. The developer was a dry powder. Figure 5 illustrates the increased response when panel 10 was processed at 71 °C (160 °F). Figure 6, a composite of Figures 4 and 5, compares processing with heat to processing without heat. The penetrant was a high boiling point penetrant; its initial boiling point was greater than 260 °C (500 °F).

Three observers were asked to rate each indication according to apparent crack size and brightness on the following scale: zero — no indication, one — dull and small, two — dull and medium, three — dull and large, four — bright and small, five — bright and medium, and six — bright and large.

Figure 7, a composite photograph of panel 3, compares heat assisted processing. The three observers rated the crack indications for the first heat assisted and first ambient temperature processing of panel 3 in Table 1. This table, typical of those generated in 1979, shows that heat assisted indications were perceived as larger and brighter. Since larger, brighter indications attract the eye, they are less likely to be missed by an inspector.

As a rough measure of how much fluorescent penetrant testing is improved by heat, the ratings were totaled and averages recorded. The ratings' averages revealed that heat assistance dramatically improves performance on some panels, but that its effect was negligible on others. For example, while heat improved performance about 370 percent on panel four and 260 percent on panel 10, improvement was only 25 percent on panel one, and a mere two percent for panel 8.

The data did show that the poorer the degree of anomaly detection by conventional ambient temperature processing, the more significant the performance improvement with heat assistance. And that is what we are seeking — improved probability of seeing microscopic anomalies on gas turbine engine rotating parts, anomalies which are more likely to be overlooked during conventional processing.

The Sherwin and Holden (1979) paper states, "In instances where conventional processing virtually fails, heat assistance gives results and the true meaning of our data is that heat assistance makes possible positive penetrant performance — positive crack identification — which is not otherwise possible."

 

Further Investigation
Additional research was conducted using contemporary AMS 2644 approved penetrant materials, and with more exacting test pieces. The principal penetrant was a type I, level 4, method D, a nonwater washable fluorescent penetrant removed by the hydrophilic emulsifier method. Two types of test pieces were used: tapered twin Ni-Cr test panels with lateral cracks in tapered plating with a depth of 0-50 mm (0-2 x 10^-3 in., and versions of twin KDS panels. Like PSM-5 panels, these panels have cracks induced in their plating in the 0.76 mm (0.03 in.) diameter range but their plating thickness is 0.03 mm (1 x 10^-3 in.) rather than 0.08 mm (3 x 10^-3 in.). Thus, cracks in this type of panel are extremely shallow.

In the case of both types of panels, one of the twins was heated to 70 °C (158 °F), and the other held at room temperature. The results indicated that for the two types of panels having relatively shallow cracks, there was no apparent difference between processing with or without heat. This contrasts with the 1979 results obtained with cracked aluminum blocks and special PSM-5 panels whose cracks had greater depth.

 

Additional Test Using Aqueous Cleaners
In order to test the theory that the switch from vapor degreasing to aqueous cleaners may have adversely influenced penetrant system performance, research was directed to level 4, method D (nonwater washable) fluorescent penetrant testing following aqueous cleaning. Cracked aluminum blocks, sheared into equivalent sections, were again used to illustrate the effect of inadequate drying before fluorescent penetrant testing and to determine if unevaporated water sufficiently deters penetrant anomaly entry so as to invite missed fluorescent penetrant testing opportunities.

Figure 8 shows the answer to be a definitive yes. The cleaner was a 10 percent alkaline solution, heated to 49 °C (120 °F). The block sections were immersed in the solution for 5 min and rinsed immediately. One section was heated to 71 °C (160 °F) following the rinse and before penetrant application; the other section was not heated and penetrant was applied immediately.

This test was repeated but with two significant changes. First, neither section was heated after removal from the aqueous cleaner and rinsed. Second, two different fluorescent penetrants were used; one was the same type I, level 4, method D (nonwater washable) described previously and the other was a type I, level 4, method A (water washable). The results indicated by Figure 9 were that the water washable penetrant performed better than the nonwater washable penetrant. The particular level 4, method A, water washable fluorescent penetrant used in this test would be classified as a water displacing penetrant. It did not contain a petroleum distillate and would be especially appropriate for these conditions. However, while tests performed with such products were cursory, as our concern is ultrahigh sensitivity testing, it was found that conventional level 2 and level 3, method A, water washable fluorescent penetrants containing a large percentage of a petroleum distillate also would function under similar conditions.

As the tests and the figures indicate, anomaly retained moisture deters nonwater washable penetrant entry. There are two ways around this problem. The first is to use a water washable (water compatible) penetrant. The second is to use heat to drive out, or evaporate, the moisture. Unfortunately, there are prohibitions against doing either.

Standing in the way of using a level 4, method A (water washable) fluorescent penetrant following aqueous cleaning on critical parts, such as gas turbine engine discs, are directives based on strong data which credit the method D penetrants with greater reliability. Standing in the way of using heat to supplement the performance of a level 4, method D (nonwater washable) penetrant are directives such as paragraph 7.2 of ASTM-E 1417 which reads "The component, penetrant, and ambient temperatures shall be in the range from 4-49 °C (40-120 °F) for type I (fluorescent) penetrants or 16-52 °C (60-125 °F) for type II (visible)."

In the 1979 paper, it was argued that the temperature restriction should not apply to high boiling point penetrants, such as those commonly found in the level 4 classification.

As part of our 1979 experiments, several stainless panels were coated with a high boiling point penetrant, placed in a preheated 104 °C (220 °F) oven for one hour. After this baking, penetrant was scraped from the panels and compared to that scraped from panels which were not put in the oven but exposed to the atmosphere for one hour. The drop in brightness was eight percent with one hour of baking in a 104 °C (220 °F) oven. Loss of brightness from dwelling on a 71 °C (160 °F) part's surface for 20 min should be relatively insignificant.

 

Conclusions
Long before the 1979 paper, heat was used successfully by experienced technicians where the testing assignment was difficult and unusually critical. Thus, the concept of using heat to improve the penetrant process is not new. If a component is at a high temperature and applied penetrant is allowed to reach the component's temperature, the penetrant will exceed its normal capability, especially in showing difficult to discover cracks. Regardless of the source of heat, the theory is the same:

• Heat speeds up the molecular movement of the penetrant fluid and lowers its viscosity, causing it to flow or migrate more vigorously.
• Heat aids in clearing anomalies of interfering contaminants by liquefying or thinning heavy greases, oil, or carbonaceous substances, making such soils more easily displaced by the flowing penetrant and by evaporating water residue.
• Heat adds to the penetrant's capacity to dissolve and clear cracks of tenacious soils which allows entry of a greater quantity of penetrant.
• Heat expands the gas (air) in the cracks, resulting in its evacuation, so the penetrant's entry is more complete.

The concluding message is as follows:

• Heat assisted fluorescent penetrant testing is a proven and demonstrable way to increase the sensitivity/reliability of the penetrant process.
• When vapor degreasing was the primary precleaning method, heat assisted fluorescent penetrant testing was practiced successfully, albeit unintentionally.
• When vapor degreasing was practiced, complaints of missed fluorescent penetrant testing opportunities were fewer.
• When replacing vapor degreasing with aqueous cleaners, extreme caution must be used to dry processed parts adequately before applying penetrants, particularly nonwater washable penetrants.
• The restriction on the use of heat on high boiling point penetrants to 49 °C (120 °F) maximum during the penetrant dwell stage is not justified.
• Cleaning is an integral part of the process and, for critical turbine engine parts, it should be as tightly specified as other processing steps, such as removal, drying, etc.
• Those who control and regulate how penetrants are to be used in the testing of critical gas turbine engine parts, especially overhaul, should and give more consideration to heat assisted fluorescent penetrant testing. It is a proven concept.

 

References
American Society for Testing and Materials, "ASTM Standard E-1417-99, Standard Practice for Liquid Penetrant Examination," 1999.

Broz, Albert, "Fluorescent Penetrant Inspection Lessons Learned," ASNT Fall Conference and Quality Testing Show, Nashville, TN, October 19-23, 1998, p. 107.

Johnson, Russell H. and Ernest Grunald, Atoms, Molecules and Chemical Change, 3rd. edition, Englewood Cliffs, NJ, Prentice-Hall.

Sherwin, A.G. and W.O. Holden, "Heat Assisted Fluorescent Penetrant Inspection," Materials Evaluation, Vol. 37, No. 10, September 1979, pp. 52-56.

 

* Sherwin, Inc., 5530 Borwick Ave., South Gate, CA 90280; (562) 861-6324; fax (562) 923-8370; e-mail info@sherwininc.com; Web site www.sherwininc.com.

 

Copyright © 2000 by the American Society for Nondestructive Testing, Inc. All rights reserved.

[ Materials Evaluation ]