back to DACCO SCI News and Highlights

dacres@daccosci.com (Chester Dacres) 
davis@daccosci.com (Guy Davis)

ABSTRACT

An electrochemical impedance spectroscopy (EIS)-based in-situ corrosion sensor has been adapted and evaluated for use with steel heat exchanger tubes in boilers, coated buried steel pipes, and painted steel structures. An excellent correlation was obtained between the logarithm of the ratio of the breakpoint frequencies, as measured by the sensor, and corrosion rate for the boiler tubes. Use of this sensor and appropriate electronics would allow the corrosion of the boiler tubes to be monitored in real time and the inhibitor concentration automatically controlled to prevent excessive corrosion. The EIS sensor is also sensitive to the quality of coating of a buried steel pipe with and without the application of cathodic protection. Similar results were obtained from a sensor attached to the pipe and from a separate electrode driven into the soil. A hand-held version of the EIS in-situ sensor is suitable for inspecting painted metal structures, such as storage tanks and locks and dams, under ambient, service conditions. An excellent correlation was obtained between the sensor measurements and the amount of corrosion on test panels immersed for up to 28 years.

Keywords:

corrosion, sensor, monitoring, boilers, pipes, electrochemical impedance spectroscopy, paint, coatings, nondestructive evaluation, electrochemistry

INTRODUCTION

The cost of degradation of the nation's infrastructure is difficult to determine, but is known to be very large. Studies by the National Bureau of Standards (NBS) estimate that overall corrosion costs in the United States are 4.2% of the Gross National Product (GNP) [1] or 290 billion in 1996 - more than $1000 per capita. The nation's infrastructure is aging and maintenance budgets are inadequate to prevent deterioration. New construction and major repairs/refurbishment are being reduced or delayed so that current hardware and structures must last longer than their design lifetimes.

Pipelines, storage tanks, and boilers are all subject to corrosion. Corroding infrastructure can cause ground water contamination, explosion potential, and/or loss of service or material. Although various corrosion control procedures (cathodic protection, paints, inhibitors) are utilized, human error, design limitations, environmental changes, material degradation and other factors can cause these procedures to partially fail over time.

One means to increase the lifetime of hardware and structures and to get the most out of stretched maintenance budgets is to track structure health from the early stages of deterioration so that its progression can be monitored and predicted. Consequently, maintenance can be performed on a needs basis when it is relatively inexpensive before damage and loss of capability becomes critical and expensive remedial action is needed. The maintenance schedule would be a proactive one rather than one that reacts to failures.

DACCO SCI, INC., (DSI) has been developing an in-situ corrosion sensor for detecting corrosion under paint coatings for aging aircraft [2,3].^, This sensor uses the established electrochemical impedance spectroscopy (EIS, also known as ac impedance) to monitor corrosion in real time beginning at the initial stages of moisture ingress through the paint film and incipient attack of the underlying metal well before any visible signs of corrosion are present. Quantification of the corrosion by use of the sensors would facilitate maintenance records and decisions.

The focus of study in the effort reported here has been the corrosion that occurs in boilers, specifically, heat exchanger tubes. Corrosion control of boilers is currently limited to adjusting the chemistry of the water to inhibit corrosion. This involves manual sampling and water chemistry modification, which is an inadequate means to control the corrosivity of the water. Such procedures are manpower extensive and subject to human error and omissions. Furthermore, although the water parameters measured are related to corrosion of the tubes, the extent or rate of corrosion, itself, is not measured.

Underground pipe systems are also of concern. Underground pipeline systems are subject to soil-side corrosion and generally are protected with coatings and cathodic protection systems. Inspection is needed to assure that the corrosion protection system is working. Pipeline corrosion-induced leaks can be difficult to detect and locate when they occur and are even more difficult to predict. Leaks and other damage can result in loss of performance and readiness, environmental and structural damage, and high costs. A secondary aspect of the program involved inspection of cathodically protected, buried coated steel pipe. Such inspections would provide assurances that the corrosion protection is effective and working well.

Buried and above ground storage tanks and locks and dams pose similar problems and solutions to pipelines. Thus inspection/monitoring techniques developed for pipelines will be applicable to the soil-side of underground tanks. Above ground structures, the interior of underground structures, bridges, and locks and dams are typically painted for corrosion protection as cathodic protection is not feasible. Inspection of painted metal is also important.

TECHNIQUE DESCRIPTIONS AND EXPERIMENTAL PROCEDURES

Techniques

Electrochemical Impedance Spectroscopy

EIS uses very small excitation voltages, generally in the range of 5 to 10 mV peak-to-peak, between a specimen and a reference electrode [4,5].^, The current induced by this voltage is measured and an impedance determined as a function of frequency. EIS is based on the fact that the behavior of an electrochemical cell and that of an electronic circuit are analogous. This allows equivalent circuit modeling of a given electrochemical cell. Fundamental AC circuit theory can then be applied to the circuit model and the results accurately correlated to reveal physical and chemical properties of the electrochemical cell. In an electrochemical cell, the presence of a coating, an electrolyte, and diffusion, act to slow the flow of electrons and each can be modeled as resistors, capacitors, inductors or a combination of elements. These factors give rise to a particular circuit for a given electrochemical cell, such as the one shown in Figure 1.

The pore resistance (R_po) is the reflection of the amount of penetration by the electrolyte into the coating. The coating capacitance (C_c) is a measure of the effectiveness of the coating. The double layer capacitance C_dl and the resistance R_COR characterize the coating/metal interface.

Performing conventional EIS on a sample involves applying ac voltage of varying frequencies between a reference electrode and the sample which is immersed in a conductive electrolyte. The current or impedance (magnitude and phase) is measured between the sample and a separate counter electrode. With the in-situ sensor, a single electrode is applied to a coated metal and this electrode acts as both the reference and counter electrodes. Equivalent results are obtained for the two approaches.

Figure 2 shows examples of EIS data, expressed in the Bode magnitude and phase formats, which demonstrates how a typical coated metal degrades in moisture. Initially the coating exhibits purely capacitive behavior (slope of -1 in the magnitude plot and phase angle near 90° in the phase angle plot) with the resistances shown in the equivalent circuit of Figure 1 being very high. As moisture permeates the coating, these resistances decrease and the impedance at low frequencies drops and becomes independent of frequency, corresponding to resistive behavior (phase angle near 0°). As degradation proceeds, the frequency range over which the coating exhibits resistive behavior increases.

Most of the measurements on this program were taken with a Gamry Instruments, Inc. Potentiostat/EIS System. This system is completely portable and the EIS experiments and analysis are performed using the CMS300 Electrochemical Impedance Spectroscopy System Software. It would be suitable for field use. In other cases, data were acquired with an EG&G PAR Model 398 Electrochemical Impedance System. This is a bench-top unit that is not suitable for field use, but has greater capability than the Gamry, particularly in the maximum frequency obtainable.

Polarization Resistance and Decay Testing

The polarization resistance technique is used to determine the corrosion resistance of a sample in a given medium or electrolyte. This essentially nondestructive test involves first finding the corrosion potential, E_corr, of the sample. A controlled-potential scan is then applied over a small range, typically ± 20 mV with respect to E_corr. The slope of the potential-current function at E_i=o is the polarization resistance (R_p) and is used to determine the instantaneous corrosion rate and the corrosion current, I_corr. Since the applied voltage is never far from the corrosion potential, no polarization-induced changes in the surface occur.

Polarization decay testing is useful in determining the condition of an immersed or buried structure. This test is obtained using NACE Standard RP-0290-90. In addition, polarization decay testing is useful in determining the effectiveness of a cathodic protection system and is used as a criteria for cathodic protection as outlined in NACE Standard RP-0285-90.

Polarization decay testing consists of impressing a current on a structure for a period of time and then removing the current source to allow the structure to decay to its natural potential. The voltage is recorded over a period of time. The polarization decay curve is used to determine if the amount of cathodic protection is adequate (a -100 mV shift). Furthermore this test can be used to determine the condition of the structure since the potential of a well coated structure will generally decay at a faster rate than a corroded structure.

The EG&G PAR Model 273 Potentiostat was used to perform polarization resistance scans. The scans were performed over the range of 20mV above and below the open circuit potential of the sample. The data were acquired and analyzed using the EG&G PAR Model 352 SoftCorr II Corrosion Measurement and Analysis Software to determine the values of E_corr, the instantaneous corrosion rate and R_p.

Approach

Three different applications of the corrosion sensors were evaluated:

• Boiler tubes
• Cathodically protected buried steel pipe
• Vinyl-coated steel following long-term immersion in natural fresh water

Boiler tubes.

The in-situ EIS corrosion sensor was modified to monitor corrosion inside a boiler tube (1.25 in. diameter ASTM/ASME A/SA 178-A/214-90A ERW Grade W1010). Different solutions were used to obtain a variety of corrosion rates:

• Aggressive (5% NaCl, as received and abraded)
• Neutral (5% Na₂SO₄)
• Inhibitive (5% Na₃PO₄ + NaOH, pH 10).

The response of the sensor was correlated with corrosion rates determined by polarization resistance.

Buried steel pipe.

Three sections of 24-in long, 12-in. diameter ASTM A53 Grade B steel pipe were obtained. One received a good covering of Royston R28 Roskote Mastic protective coating. One received a poor covering of Mastic and the last one was left bare. The average coating thickness was 40 mils. Two in-situ corrosion sensors were applied to each pipe section. In the case of the bare pipe, the sensor was isolated from the pipe with spray enamel. The pipes were subsequently separately buried in containers in the laboratory. An impressed current anode was inserted into the soil of each container. The clay soil characteristics are given in Table 1. The pipe potentials were allowed to reach equilibrium in their containers for several days. Baseline EIS measurements were made throughout this period. An impressed current was then applied to each pipe to polarize it to a protection potential of -850 mV vs Cu/CuSO₄ (NACE STD RP-0169). EIS measurements using the in-situ sensor were taken as a function of time, with and without the impressed current. A limited number of measurements were also taken using the impressed current anode itself as the electrode instead of the attached sensor.

Vinyl-coated steel.

Five vinyl-coated steel panels that had been immersed in natural fresh water for up to 28 years were provided by CERL (Figure 3). Some specimens were in excellent condition while others were in poor condition. The specimens are described as follows:

• 4K194 2-yr old gray vinyl in excellent condition
• R115 young vinyl in very poor condition
• 2429 22-yr old red vinyl in good condition
• 2176 28-yr old vinyl zinc-rich with white vinyl topcoat in fair condition
• 2578 17-yr old vinyl zinc-rich with white vinyl topcoat in excellent condition.

	Figure 3. Vinyl coated steel specimens following long-term immersion in natural fresh water. Clockwise from lower left: 4K194, 2429, 2578, 2176, and R115.

RESULTS AND DISCUSSION

Boiler tubes.

An excellent correlation (Figure 4) was obtained between the logarithm of the ratio of the high and low breakpoint frequencies (frequencies at which the phase angle of the EIS spectrum is 45°, see, e.g., Figure 5) and the corrosion rates for rates less than ~8mpy. The linear trend in the figure holds for both the aggressive (NaCl, with and without abrasion), benign (Na₂SO₄), and inhibitive solutions at room temperature. At higher corrosion rates, the difference in log breakpoint frequencies saturates near 6 which coincidentally is near the upper limit of frequency measurements with the EG&G instrument. Because anticipated corrosion rates of boiler tubes are less than 1 mpy, the linear region of the correlation provides adequate range for inspection. Representative phase angle spectra are given in Figure 5 to illustrate the changes in width of the capacitive region (log of ratio of the breakpoint frequencies) that occur with time as the iron passivated and the corrosion rate decreased.

Figure 4. Difference of the logarithms of the high and low breakpoint frequencies, i.e., the ratio of the two frequencies, as a function of corrosion rate for boiler tubes in different solutions.

Figure 5. Bode phase angle spectra for boiler tubes immersed in Na₂SO₄ solution for different lengths of time. The breakpoint frequency is defined as the frequency at which the graph crosses the 45° line (shown as dotted).

Buried steel pipe.

EIS measurements from the in-situ sensor track well with the degree of coating of the buried pipe. As shown in Figure 6, the fully coated pipe has the highest impedance while the pipe with no coating has the lowest impedance. After initiation of the impressed current cathodic protection at 7 days, the differences between the low frequency impedances increase even though the measurements shown in this figure were taken with the impressed current temporarily turned off.

Introduction of impressed current cathodic protection has two effects (Figure 7):

• Increased noise at high frequencies
• Increased impedance at low frequencies.

Such behavior was observed for all three pipes. The low-frequency impedance continues to be a function of the quality of the coating (Figure 8). The magnitude of the difference is similar to that observed with the impressed current turned off (but following initiation of the impressed current, Figure 6) - the pipe with the good coating exhibits a low-frequency impedance approximately 3 orders of magnitude higher than the bare pipe.

The utility of this approach to inspect pipes would be enhanced if similar measurements could be acquired without the use of an attached sensor. Such inspection, then, would be suitable for the millions of miles of pipes currently in service. Figure 9 shows that similar results are obtained from a remote electrode (the impressed current anode) and from the attached sensor provided that the impressed current is temporarily turned off during the EIS measurement. The same ranking of the three pipes and their coatings were obtained. Differences between the two sets of spectra likely result from soil properties.

Figure 6. Low frequency impedance measured using the attached sensor as a function of time for three buried pipes: bare, partially coated, and fully coated. Impressed current is off.

Figure 7. EIS spectra of the partially coated pipe with and without impressed current cathodic protection. The attached sensor was used to acquire the measurements.

Figure 8. Low-frequency impedance measurements taken with the attached sensor with the impressed current on as a function of burial time.

Figure 9. EIS spectra of the three buried pipes comparing data acquired with the attached sensor and with the impressed current anode as the sensor. The impressed current was turned off during the measurement.

Vinyl-coated steel.

Electrochemical impedance spectra of the five vinyl-coated panels using the hand-held corrosion sensor are given in Figure 10. Excellent correlation was obtained between the visual condition of the panels and the electrochemical measurements: Panels in excellent condition exhibited very high impedance (up to 10⁹ at low frequencies) with capacitive behavior (log impedance vs log frequency slope of -1). In contrast, panels in poor condition showed relatively low impedance (~10⁴ ) with mostly resistive behavior (impedance independent of frequency). These spectra are typical of good and bad coatings, respectively.

Figure 10. Impedance spectra of vinyl-coated steel specimens following immersion in fresh water. The data were acquired by the hand-held sensor and the portable Gamry potentiostat.

For comparison, equivalent spectra were taken using a flat electrochemical cell and conventional remote electrodes. The two methods of data acquisition - hand-held in-situ sensor and conventional electrodes - give equivalent results. This agreement, along with many other comparisons, serve to validate the in-situ sensor and connect the sensor results with the historical database of EIS measurements.

It is interesting to note that one of the bad specimens (2176) had some areas in which the coating looked relatively good. When the hand-held sensor was used to separately inspect the good and bad areas, significant differences were seen in the spectra (Figure 11) with the measurements from the good area approaching that of the good specimens. In contrast, the flat cell was unable to distinguish between the good and bad areas. We attribute this difference to the conventional EIS measurements being dominated by defects or other small bad areas within the analysis area. In this usage of the hand-held sensor, the measurements are more localized and controllable so that results are indicative of small good areas. Other measurements have shown that the detection range of the hand-held sensor (and the permanent in-situ sensor) is highly dependent on the surface conductivity of the coating - a dry coating will give a very localized corrosion reading while a wet coating will give a wider ranging corrosion reading. Under carefully controlled laboratory conditions, detection of a scratch 15 feet away from the sensor has been achieved.

Figure 11. Impedance spectra for poor vinyl specimen 2176 in both "good" and "bad" areas. Measurements were taken with both the hand-held sensor and conventional EIS.

SUMMARY AND CONCLUSIONS

The EIS in situ sensor can monitor or inspect corrosion of boiler tubes, buried pipes, coated steel structures, and, potentially, composite/metal structures. Specific findings and conclusions include:

• The EIS in-situ sensor can be adapted for use in a boiler system to monitor the corrosion rate of the interior of the tube.
• An excellent correlation was obtained between the logarithm of the ratio of the two breakpoint frequencies and the corrosion rates in the range of 0-8 mpy. The specific correlation may be temperature dependent.
• This sensor and appropriate electronics can feasibly monitor the corrosion rates in real time during boiler operation and control servo valves to increase the inhibitor concentration if the corrosion rate increases.
• The EIS sensor is capable of determining the quality of a coating on a buried pipe and the presence of impressed current cathodic protection.
• Similar measurements can be obtained using remote electrodes, such as the impressed current anode itself. This capability would allow inspection of existing pipes without attached sensors.
• The hand-held version of the EIS in-situ sensor is suitable for inspection of painted steel structures, such as storage tanks, locks and dams, etc. The measurements clearly reflected the extent of corrosion on specimens immersed for up to 28 years.

ACKNOWLEDGMENTS

We are gratefully to Vince Hock of the Army Corps of Engineers Construction Engineering Research Laboratories (CERL) for providing the vinyl-coated steel specimens. This work was funded by CERL under SBIR Phase I contract DACA88-97-C-0003.

REFERENCES

1. J.A. Payer and G.M. Ugiansky, "Impact of the NBS - Batelle Cost of Corrosion Study in the United States," Proc. Corrosion, Symp. Int. Approaches to Reducing Corrosion Costs (NACE, Houston, 1986).

2. G.D. Davis, C.M. Dacres, M. Shook, and B.S. Wenner, "Electrochemical In-Situ Sensors for Detecting Corrosion on Aging Aircraft," in Proc. Workshop on Intelligent NDE Sciences for Aging and Futuristic Aircraft (El Paso, TX, 1997). P. 141.

3. G.D. Davis and C.M. Dacres, "Electrochemical Sensors for Evaluating Corrosion and Adhesion on Painted Metal Structures," patent pending.

4. J.R. Scully, "Electrochemical Methods for Laboratory Corrosion Testing," in Corrosion Testing and Evaluation: Silver Anniversary Volume, ASTM STP 1000, R. Baboian and S.W. Dean, Eds., American Society for Testing and Materials, Philadelphia, 1990, p. 351.

5. W.J. Lorentz and F. Mansfeld, "Determination of Corrosion Rates by Electrochemical DC and AC Methods," Corrosion Science 21, 647 (1981).

DACCO SCI, INC. ¤ 10260 Old Columbia Road ¤ Columbia, MD 21046
(410) 381-9475 (Baltimore) ¤ (301) 596-7019 (Washington) ¤ (410) 381-9643 FAX
dacres@daccosci.com (Chester Dacres) 
davis@daccosci.com (Guy Davis)

© 2000 DACCO SCI, INC.
Please send any comments to dacres@daccosci.com
25 April 2000