One possible AC corrosion mechanism- the interface local thermal effect under AC signal
In this paper one possible AC corrosion mechanism was studied by equivalent circuit method. After analyzing the thermal effect of interface CPE, author found under high AC frequency and high current condition the interface owned a higher local temperature than that of environment. This local thermal effect was thought from the dielectric loss of CPE, if CPE was thought as owning capacitance and resistance character at the same time. Some mathematical estimations were provided and it was found under general AC corrosion condition the interface local temperature was increased obviously. And it was well known local high temperature will accelerate steel corrosion in wet soil. Also some condition as SCC which was similar in some extent with AC corrosion was discussed. The result showed that at least under some conditions these two phenomenon had some inner connection.
Key words: AC corrosion, CPE, local thermal effect, frequency, dielectric loss
Corrosion caused by the discharge of 60 Hz AC current from a pipeline in a high voltage AC (HVAC) corridor has been discussed and studied over the past 20 or more years. More recent studies have specifically addressed these corrosion issues following several failures attributed to the presence of AC discharge from the . However, the mechanisms of AC corrosion are still not completely understood1,[ L.V. Nielsen, P. Cohn, ―AC corrosion and electrical equivalent diagrams,‖ CEOCOR, Committee on the Study of Pipe Corrosion and Protection, 5th International Congress, Bruxelles, Belgium, 2000.]. The body of recent (post-1980) literature indicates that AC corrosion or AC-enhanced corrosion (ACEC) is a bona fide effect; there appears to be a tacit agreement that at prevailing commercial current frequencies (such as 50 or 60 Hz) corrosion is possible, even on cathodically protected pipelines.
AC corrosion was not well understood for three reasons: (1) the electrochemical phenomenon of corrosion is normally attributed to DC; and (2) the instruments normally used to measure the electric parameters in direct currents cannot correctly detect the presence of AC with frequencies between 50 and 100 Hz. (3) present electrochemical theory was far from a perfect one which can unveil all mechanism of interface phenomenon.
An example of the last reason was that until now the mechanism of CPE was not certain. Ironically still CPE was used to explain AC impedance spectrum by us everyday. One common character of AC corrosion and CPE was that they were all about AC signal. This fact meant that maybe our knowledge and method to deal with those AC condition was poor.
In this study, traditional equivalent circuit was used to discuss the mechanism of AC corrosion and some theory estimation to this phenomenon was discussed. The CPE in equivalent circuit, which both owned capacitance and resistance character under AC signal, was thought as a dielectric loss element from interface. So under AC signal condition, the power of CPE will be a function of frequency and amplitude of AC signal. If the power was large enough, the temperature of interface will increase obviously because of interface irradiation. Also if interface temperature increased, the metal corrosion potential and rate would be changed.
2 The interface local thermal effect under AC signal
For a general electrochemical reaction interface, its equivalent circuit can be expressed as fig1.
Fig.1 A typical equivalent circuit of metal/solution interface
The CPE in circuit own the resistance expression as follows:
Where ZCPE was impedance of CPE,
Q0 was a constant,
ω was angular frequency of AC signal,
n was a constant,
i was (-1)0.5 .
CPE in above circuit owned characters of capacitance and resistance at same time. And its resistance part was a function of frequency. It meant if the applied AC signal frequency was increased, the resistance from CPE would decrease with the frequency. If the applied voltage amplitude was a constant, with the frequency increase, the whole system current would keep increase. An interesting problem is when current value is enough high, interface itself will own a high local power density which will change the local temperature of interface. It means the electrochemical reaction speed will be changed. Under this condition effects of AC signal will be completely different with that of a DC signal.
Before the interface local thermal effect was discussed, the following imaginations were given:
Value of interface reaction resistance (RF) was much greater than that of ZCPE under high frequency range. So thermal effect of RF was omit.
The solution and electrode thermal effects was omit because of high speed of liquid convection and heat conduction. Namely their temperature was thought as environment temperature.
The interface can be thought as a grey body, Stefan-Boltzmann law would govern relationship between its temperature and power density. The heat transport by convection, conduction and radiation. Interface was covered by diffusion layer. So convection can not be a main heat transport way. Conduction way needs material exchange. It means there must be enough free electrons and ions, which can enter metal side or solution diffusion layer. In fact most of the heat created by reaction resistance (RF) can be transported by this way because of electron and ions exchange. But for water dipoles adhered on double layer, there are little free electrons and ions. In a AC cycle, the dipoles adjusts their angle periodically. And if the heat from friction in this process was great enough, radiation will be the main way to transport it.
CPE parameters did not change greatly with the AC fluctuation.
The following part was an estimation of CPE thermal effects.
where I was current passing CPE,
U=U0 sinωt was applied AC signal and U0 was amplitude of AC signal,
α was phase difference between current and voltage.
where P was average power of CPE.
where PR was real part of CPE power and PI was image part of CPE power.
Obviously it was unnecessary to discuss the imagine part because it did not create real heat. The PR was a function of ω and this function owned a peak value.
When was satisfied, the PR own greatest value.
If the interface was thought as a grey body, Stefan-Boltzmann law would govern relationship between its temperature and power density.
Where A was reaction interface area, ε was grey degree (it was a constant for a special material), σ=5.67*10-8 JS-1m-2K-4 , T was interface temperature, T0 was environment temperature.
Imagine for a typical room temperature (300K) EIS testing condition, U0=10mV, n=0.8, RS=1Ohm/cm2, ε=0.2, the interface greatest temperature in test would be 300.04825K.
And if the Q0=10-3F/cm2 , it happened under ω= 103.75. Generally when testing frequency is so high the testing error can easily veil so small temperature increase. So generally in EIS test, we can omit the interface thermal effect.
Imagine under some extreme condition, a high amplitude AC signal was applied on reaction interface and at the same time the RS was very small that was always satisfied in testing frequency arrange.
If U0=1V, n=0.8, Q0= 10-3F/cm2 , ωn = 10 3.5 ε=0.2 , the interface temperature would be T=814K. It meant the interface temperature increase greatly. Obviously the electrochemistry reaction and chemical reaction speed will increase with the temperature.
For a wet salt oil pipeline condition (300K), if the U0=0.3V, RS =5Ohm/cm2, n=0.8, the possible greatest interface temperature would be T= 308.33K. if Q0= 10-3F/cm2 , it happened when frequency was 119.8Hz. In fact this frequency was in the range of AC corrosion frequency from 10-200Hz.
Another element accelerated the effects of AC corrosion was the uneven distributing of current. The high current part on pipeline would own higher temperature. Obviously because of local anode effect this part would be corroded firstly.
Maybe AC corrosion was not only met by us in condition as pipelines. In many stress corrosion crack and corrosion fatigue crack case, crack tip owned very unstable corrosion voltage because of the uncontinuous broken and recreation process of oxidation film[ The estimation of the threshold state of stress corrosion cracking based on the distribution of the electricity quantities of local anodic currents estimated from potential fluctuations - Basic investigation on the use of potential fluctuation measurement for SCC . A simple estimation was as follows. If the metal dislocation thickness was 0.3 nm and the crack tip local strain rate was 10-9m/s, the frequency of crack tip voltage was near 33 Hz. Also it was in the AC corrosion frequency range. In fact , under this condition the voltage fluctuation amplitude will be at least 0.5V because of the great voltage difference between fresh metal and its oxide. Of course above estimation was too simple to cover all SCC problems. But it gave us another view to understand some SCC phenomenon.
AC corrosion was a phenomenon related to applied voltage frequency. CPE was a special dielectric loss phenomenon on corrosion interface. It was possible that the thermal effect resulted by CPE was an important reason of AC corrosion. The estimation in paper showed it was possible the interface local temperature may be 10 degree higher than that of environment. High temperature and uneven corrosion current distribution result accelerated corrosion on metal surface. This mechanism may be an important reason of AC corrosion.