William C. Fitzgerald, Michael N. Davis, James L. Blackshire, John F. Maguire, and David B. Mast
Keywords:
Abstract:
Because you are not logged-in to the journal, it is now our policy to display a 'text-only' version of the paper. This version is obtained by extracting the text from the PDF or HTML file, and it is not guaranteed that the text will be a true image of the text of the paper. The text-only version is intended to act as a reference for search engines when they index the site, and it is not designed to be read by humans!
If you wish to view the human-readable version of the paper, then please Register (if you have not already done so) and Login. Registration is completely free.
JCSE Volume 3, Paper 15 Submitted 27th November 2001 Evanescent Microwave Sensor Scanning for Detection of Sub-Coating Corrosion William C. Fitzgerald, Michael N. Davis, James L. Blackshire, John F. Maguire, and David B. Mast* Air Force Research Laboratory (AFRL), Materials and Manufacturing Directorate (ML), Bldg. 653, Rm 14, 2977 P Street, Wright-Patterson Air Force Base, 45433-7746, 937-255-8787, mailto2('william.fitzgerald','ml.afrl.af.mil') *Department of Physics, University of Cincinnati, Cincinnati, Ohio� 45221-0011 §1 Abstract Evanescent Microwave Probe (EMP) scanning spectroscopy has the capability to image sub-surface features through poorly conducting or dielectric materials, such as aircraft paint.� This makes EMP scanning a valuable option for detecting corrosion of aircraft primary structural members without removing paint from the area to be inspected. Evanescent microwave probes measure resonance frequency shift and power loss as functions of the conductivity, dielectric response, and topography of a material. Because each of these�� materials� properties can be changed locally by surface and near-surface defects, a change in measured parameters occurs when the probe passes over an area of corrosion.� This new nondestructive evaluation technique produces sensitive and highly resolved spatial measurements of corrosion on aircraft structures through layers of paint, thus saving time and money on aircraft maintenance. Keywords: Non-destructive evaluation, Microwave imaging, Near-field imaging. Introduction §2 Corrosion of aging aircraft has recently become an expensive and safety-critical problem. [crossref(1),crossref(2),crossref(3)].� The annual direct cost of aircraft corrosion in the U.S. alone has been estimated to be in excess of $13 billion dollars [crossref(2)].� These high costs are, in part, due to the current inability of inspection techniques to detect, characterize, and classify corrosion processes before significant damage has occurred.� Although military and civilian agencies have slightly different classification systems [crossref(4),crossref(5)], both agree that an effective corrosion maintenance program is one that detects and controls corrosion as close to the initiation stage as possible.� This would require a very sensitive detection system, with high spatial resolutions, allowing for the detection of corrosion defects with less than 5% material loss [crossref(6)].� Such a system would lead to minimized downtime and repair costs, and a reduced risk of incurring additional damage to essential aircraft structures [crossref(7)].� §3 Dramatic improvements in corrosion nondestructive evaluation (NDE) system capabilities have been made over the last 10 years [crossref(8),crossref(9),crossref(10)].� Some of the most notable NDE techniques include advanced ultrasonic, eddy current, and thermo graphic techniques. For example, ultrasonic systems have recently been developed that are very sensitive, and are capable of detecting corrosion sheet thinning at levels approaching 5% [crossref(8)].� Pulsed eddy current systems have also been developed with similar sensitivity levels[crossref(9)].� Both of these techniques, however, are still limited in lateral spatial resolution to approximately 1 cm. Although thermography systems provide increased spatial resolutions of 10 to 100 microns, environmental noise and difficulties in interpreting results currently restrict widespread use of� this� technique. [crossref(10)] §4 A relatively new NDE technique that shows great promise for detecting corrosion at initiation involves evanescent microwave probes (EMP).� The unique feature of the evanescent microwave probe is its ability to sense the complex local microwave permittivity of a material [crossref(11),crossref(12)].� Because the permittivity at microwave frequency of a metallic substrate and an oxidized or corroded area can be quite different, the local corrosion level can be detected and mapped with high spatial resolution and sensitivity.� The corrosion can be probed at the surface, in the sub-surface, and under coatings, primers, and topcoats used on typical aircraft structures. Additionally, the material can be probed with very high spatial resolutions.� Since the evanescent or near-field, EM fields of the probe provide typical spatial resolutions of <10 microns. Therefore, the evanescent microwave probe is well suited for detecting and assessing very low levels of hidden corrosion and material loss. §5 The primary objective of the work reported here is to describe a microwave probe resonant structure that has been used to non-destructively detect sub-surface corrosion in metallic samples. This paper describes the relatively simple resonant probe along with some preliminary results obtained from painted samples. Discussion Principles of Operation §6 In applying evanescent field imaging, the fields are intentionally limited to regions that are significantly smaller than the wavelength of the incident electromagnetic radiation. The fields in these regions are� unable to propagate freely, so they evanesce, or attenuate, exponentially. However, these evanescent fields can couple to, material features much smaller than the wavelength of the propagating wave, and are thus capable of affecting the resonance conditions of the EMP probe. In effect, this methodology enables detection of features smaller than the classical limit (Abbe barrier). This limit on the spatial resolution, or minimum resolvable size is 1/2 of the wavelength, λ, of the electromagnetic excitation fields. The evanescent wave will decay exponentially and contain spatial frequencies above 1/λ. To recover a signal, one must recover all the spatial frequency components contained within that signal, so the Abbe barrier is a spatial interpretation of the sampling theorem.� The fundamental concept is to closely scan a point like field source over a material object so that the evanescent field is still powerful enough to interact with the properties of the material. This will result in an image with resolved features smaller than the Abbe limit. §7 The types of evanescent field probes which have traditionally been used for near-field imaging similar to that described in this article are: coaxial transmission line resonator with aperture [crossref(13)], the center conductor of coaxial transmission line [crossref(14)], [crossref(15)], rectangular wave-guide with end-plane aperture [crossref(16)], and micro-strip resonators utilizing wire probes or loops [crossref(11)].� The open-circuit coaxial probe with a sharpened center conductor protruding slightly beyond its outer shield length can be treated as an electrically short and insulated dipole antenna. The length of the dipole is much shorter than the exciting microwave wavelength and the insulator is supplied from the free-space gap between sample and probe tip.� When the resonator tip is placed in close proximity to an object to be imaged, the resonator�s reflection coefficient S11 at resonant frequency will shift as depicted in Figure 1. The amount of shift in the resonant frequency Fr and the quality factor of the resonator, Q, are determined by the free-space gap between sample and probe tip and the microwave properties of the sample. The effective area of the probe tip is fixed once the resonator is fabricated. Holding constant the air gap distance from tip to sample, and the resonant frequency, Fr, variations in the samples� microwave properties can be mapped as the probe tip is scanned over the sample. §8 The microwave response of a material is a function of local permittivity, εr, permeability μr, and free carrier concentration [crossref(11)]. The evanescent fields decay exponentially from the probe tip, and as a material enters this near field region, it will perturb these evanescent fields, resulting in the loading of the resonator. A loss-less dielectric or a good conductor will increase the probe�s effective capacitance resulting in a lower resonant frequency with the same Q as the unperturbed probe. Sensing these variations in the resonant frequency and the magnitude of reflection coefficient of the resonant probe allows an extensive range of material parameters to be measured or inferred. §9 The easiest method for measuring the microwave resistivity of a sample is to measure the change in the reflection coefficient of the resonator at fixed operating frequency. Figure 1 shows the resonant plots of the reflection coefficient for the sample perturbing the evanescent field and the resonant plot with the sample at a distance from the probe tip in the far field. If the disturbance of the probe�s resonance is small, the maximum frequency sensitivity can be obtained by the difference of the resonant spectra, with and without the sample, and by observing the frequency at which the difference is the largest. This will allow the highest possible sensitivity and best detection resolution in scanning materials with small resistivity changes. §10 When microwave properties of the samples cause large resonant shifts, the reflection amplitude at a set frequency of operation is not linear. Therefore, operation at fixed excitation amplitude, but variable frequency is favored.� This variable frequency method requires a feedback loop to track and shift the resonant frequency. However, due to the morphology of the sample, the resistivity and the dielectric response of the sample may change simultaneously with the distance of the probe over the sample. It is, therefore, necessary to differentiate between sample sample-probe-tip distance variations and non-uniformity of the sample�s microwave properties. In both of the fixed frequency methods, when the sample to probe distance varies with the sample topography, a measurement of this distance variation is required to differentiate between the sample parameter variations and topography variations. §11 In this work, the changes in the probe�s resonant frequency Fr and quality factor Q are tracked by a frequency modulation technique. Figure 2 depicts the evanescent microwave probe resonant curve in air. The microwave excitation frequency of the resonant probe can be in the bandwidth of 1 to 40 GHz and is the center frequency, Fc, of the resonant curve, or Fc = Fr. The excitation signal of the probe is modulated at F1=100 Hz and the width of the resonant curve, or quality factor, is determined by the amplitude of the detector signal at twice the modulation frequency F2 = 200 Hz. Figure 3 shows the resonant curve shift due to the probe moving to a new position above the sample. The 100 Hz modulated signal now sweeps further up one side of the broadened resonant curve, so that the amplitude of the detector signal now contains frequency components at F1 and F2.� The frequency of the microwave excitation center frequency, Fr, is then changed so that the magnitude of the F1 component in the detected signal is reduced to zero; the excitation frequency is then shifted back to the center frequency of the resonant response of the probe tip.� The DC component of the detected signal can be determined by the average of the lowest 10 points of the modulated response.� The F1, F2, and DC data are determined at each position over the sample, generating the frequency shift (εr or μr changes), energy loss� (change in resistance) and DC (complex combination of change in resistance and εr or �r changes) image plots, respectively. §12 There are a variety of advantages, characteristics, and applications, which demonstrate the versatility of the evanescent microwave imaging. The technique is non-contact, and therefore is not invasive on the samples it characterizes. There is no sample preparation required, and testing can be accomplished in air, vacuum or submersed in a liquid media. There are no temperature limitations associated with microwaves, so samples at elevated temperatures as well as objects moving at a moderate rate of velocity can be imaged. The image resolution is easily adjusted by changing the sample to probe tip distance. There are reports of image studies showing the versatility of the EMP and its potential application in many apparently unrelated areas such as imaging defects in composite materials, imaging moisture variations in botanical samples, and imaging impurities and residual stress in semiconductors [crossref(12)], [crossref(17)], [crossref(18)]. In this paper, we report on using EMP imaging in order to investigate the presence of corrosion under painted surfaces. Experimental Procedure §13 Before measurements were made, the X � Y stage was plane-leveled with a standard dial indicator to remove any tip to sample distance variation during scanning.� Figure 1 shows the shift in the experimental resonance frequency with an aluminum plate placed near the probe. The presence of the aluminum plate close to the probe tip versus the probe tip in air shifts the resonant frequency. §14 The actual painted aluminum sample suspected of corrosion was then mounted on the X � Y stage and scanned with a probe tip to sample standoff distance of 20 μm.� The probe was operated at a fixed frequency of 6.835 GHz and a set of raster-scan images was obtained for the resonant frequency shift and resonant frequency loss.� The area suspected of corrosion was then chemically stripped of its paint and primer coating to verify the existence of corrosion. Finally, the sample area was inspected visually with a digital microscope to verify the existence of corrosion. §15 Since the thickness of the original paint layers could not be determined prior to stripping, the actual depth of penetration of the evanescent fields was not known.� In order to gauge the probing depth, the corroded sample was recoated with standard military specification primer and paint to a total coating thickness of approximately 4.7 mils (119 μm). The recoated sample area was scanned again with the same microwave parameters to verify sensitive corrosion measurement capabilities through a known paint and primer thickness. Description of Results §16 The initial sample was Al7075-T6 aluminum with primer and topcoat paint, taken from a C-130 aircraft.� Although no visible signs of corrosion were apparent, it was suspected that sub-surface (below paint) corrosion was present somewhere on the sample surface.� Figure 5 shows an evanescent microwave probe power loss plot of the 7075-T6 sample, scanned in an area of the sample that was suspected of having corrosion. The area indicated in red-yellow is where the sub-surface corrosion was detected.� Similar results are shown in Figure 6, which depicts the evanescent microwave probe frequency shift plot of the suspected area of corrosion §17 Figure 7 depicts a photomicrograph of the paint stripped Al7075-T6 sample at the location where subsurface corrosion was indicated by the evanescent microwave probe scans. The photo was taken, using dark field illumination, and clearly shows the area of pitted corrosion that existed below the paint coating prior to stripping.� The corroded area was approximately 100 microns x 100microns in size and was likely in the initiation stage.� The EMP data provided in Figures 5 and 6 clearly showed sensitive and highly resolved measurement capabilities for this small, paint-covered corrosion pit. §18 Figures 8 and 9 show evanescent microwave probe scans depicting power loss and frequency shift plots respectively. This is the identical area previously imaged, but re-coated with 4.7 mils (119 μm) of primer and topcoat. The images appear as before, and give an indication of penetration depth through the primer/topcoat in excess of 125 microns. Conclusion §19 It has been shown that a near-field coupled, coaxial, microwave resonator can be constructed and used to measure hidden corrosion under paint in aluminum with high sensitivity and spatial resolutions. �Changes in microwave resonance frequency shift, and power loss both showed evidence of the corrosion.� The evanescent microwave imaging technique was able to measure a small, 100 microns x 100 microns, corrosion pit in a sample of Al7075-T6 aluminum that was hidden under 4.7 mils (119 μm) of primer/topcoat and not visible to the naked eye.� The system has the potential for measuring hidden corrosion close to the initiation stage, with high spatial resolutions, and with sensitivities greater than that needed to detect defects smaller than the threshold limit of 5% in material loss. The use of this new and novel corrosion detection technique should lead to reduced downtime and repair costs, and a reduced risk of incurring additional damage to essential aircraft structures. References defref(1) �Corrosion Detection Technologies, Sector Study�, Final Report, Prepared for the North American technology and Industrial Base Organization (NATIBO), March 1998. defref(2) Reference deleted as it is no longer accesible. defref(3) �U.S. Combat Air Power: Aging Refueling Aircraft are Costly to Maintain and Operate�. �General Accounting Office, Washington, D.C. National Security and International Affairs Div.; Report GAO/NSIAD-96-1608, August 1996. defref(4) D. Hoeppner, L. Grimes, A. Hoeppner, J. Ledesma, T. Mills, and A. Shah, �Corrosion and Fretting as Critical Aviation Safety Issues: Case Studies, Facts, and Figures From US Aircraft Accidents and Incidents,� Estimation, Enhancement and Control of Aircraft Fatigue Performance, Volume I, Editors: J. Grandage, G. Jost, Proceedings of the 18th Symposium of the International Committee on Aeronautical Fatigue, 3-5 May, 1995, Melbourne, Australia. defref(5) D. Curtis, �Corrosion and Fatigue in the Aging Fleet,� Federal Aviation Administration (FAA), North West Mountain Region, 1992. defref(6) J. Komorowski, D. Forsyth, D. Simpson, and R. Gould, �Probability of Detection of Corrosion in Aircraft Structures�, defref(7) R. Bailey, R. Pidaparti, S. Jayanti, and M. Palakal, �Corrosion Prediction in Aging Aircraft materials using Neural Networks,� AIAA �2000-1762. defref(8) S. Hansen, B. Mossawir, A. Ergun, F. Degertekin, B. Khuri-Yakub, �Air-Coupled Nondestructive Evaluation using Micromachined Ultrasonic Transducers,� IEEE Ultrasonics Symposium, 1999. defref(9) Y. Sun, T. Ouyang, and S. Udpa, �Multi-Layer Aircraft Structure Inspection using Super-Sensitive Remote-Field Eddy-Current System,� Review of Progress in Quantitative Nondestructive Evaluation, Editors: D. Thompson, and D. Chimenti, Ames, Iowa, 16-20 July, 2000. defref(10) X. Han, L. Favro, L. Li, Z. Ouyang, G. Sun, R. Thomas, and D. Ashbaugh, �Quantitative Thermal Wave Corrosion Measurements on a DC-9 Belly Skin in the irregular paint Thickness variations,� Review of Progress in Quantitative Nondestructive Evaluation, Editors: D. Thompson, and D. Chimenti, Ames, Iowa, 16-20 July, 2000. defref(11) M. Tabib-Azar, N.S. Shoemaker, and S. Harris, �Non-destructive characterization of materials by evanescent microwaves,� Meas. Sci. Technol., vol. 4 , Jan. 1993,� pp. 583-590. defref(12) M. Tabib-Azar, P.S. Pathak, G. Ponchak, and S.R. LeClair, �Nondestructive super-resolution imaging of defects and non-uniformity�s in metals, semiconductors, dielectrics, composites, and plants using evanescent microwaves,� Review of Scientific Instruments, 70 (6), pp. 2783-2792. defref(13) T. Wei, X. Xiang, and P.G. Shultz, �Scanning Tip Microwave Near-Field Microscope,� Appl. Phys. Lett., 68(24), June 10, 1996, pp.3506-3508. defref(14) C. P. Vlahacos, R.C. Black, S. M. Anlage, A. Amar, and F.C. Wellstood, �Near- field scanning microwave microscope with 100 mm resolution,� Appl. Phys. Lett., 69(21), Nov. 18 1996, pp. 3272-3274. defref(15) B. J. Feenstra, C. P. Vlahacos, A.S. Thanawalla, D.E. Steinhauer, S. K. Dutta, and S. M. Anlage, �Near- field scanning microwave microscopy: measuring local microwave properties and electric field distributions,� 1998 IEEE� MTTS Int. Microwave Symp. Dig., Baltimore, MD., June 7-12, 1998, pp. 965-968. defref(16) A. F. Lann, M. Golosovsky, D. Davidov, and A. Frankel, �Mapping the thickness of conducting layers by a mm-wave near-field microscope,� 1998 IEEE� MTTS Int. Microwave Symp. Dig., Baltimore, MD, June 7-12, 1998, pp. 1337-1340. defref(17) M. Tabib-Azar, S. Bunrerraj, J.L. Katz, and S.H. You, �Evanescent electromagnetics: a novel, super-resolution, non-intrusive imaging technique for biological applicatons,� Biomedical Microdevices, 2 (1), 1999, pp. 73-80. defref(18) M. Tabib-Azar, D. Akinwande, G.E. Ponchak, and S.R. LeClair, �Evanescent microwave probes on high-resistivity silicon and its application in characterization of semiconductors,� Review of Scientific Instruments, 1999, 70 (7), pp. 3083-3086. §17 Fig 1. The S11 resonance curve and its modification due to the presence of aluminum sample near probe tip. §17 Fig 2.� Resonant curve showing F1= 100 Hz modulating signal and F2 = 200 Hz detector signal. §17 Fig 3. Resonant curve showing F1 = 100 Hz modulation signal with detector signals F1 = 100Hz and F2 = 200 Hz §17 Fig 4. System setup §17 Fig 5. Loss scan of potential corrosion area (click for a full-size image) §17 Fig 6. Shift scan of potential corrosion area (click for a full-size image) §17 Fig 7. Photomicrograph (original magnification 100x) of 7075-T6 aluminum stripped of paint §17 Fig 8. Loss scan of corrosion area after 4.7 mils of paint (click for a full-size image) §17 Fig 9. Shift scan of corrosion area after 4.7 mils of paint (click for a full-size image)