Review—Electrochemical Probes and Sensors Designed for Time-Dependent Atmospheric Corrosion Monitoring: Fundamentals, Progress, and Challenges

Electrochemical probesandsensorshavebeendevelopedtodetectandmonitoratmosphericcorrosionofmetallicmaterialsinthepast 40 decades. Depending on the measurement methods, the electrodes and structures of probes and sensors can be different. Various mathematicalmethodsandmodelshavebeendevelopedtodeterminethetime-dependentcorrosionrateofmetalunderthinelectrolyteﬁlm.Polarizationtechniquessuchaselectrochemicalimpedancespectroscopy(EIS)andlinearpolarizationresistance(LRP)havetheadvantageofeasydatainterpretationbuthaveatendencytointerferewiththecorrosionsystemunderinvestigation.Nonpolarizedtechniquessuchaselectrochemicalnoise(EN)donotdisturbthecorrosionsystembutdatainterpretationcanbeproblematic.Toachievelongtermandreliablecorrosionmonitoring,optimizedelectrodedesignandamultichannelelectrochemicalinstrumentarerequired.Newcorrosionmodelsandnoveldatainterpretationmethodsareneededinfuturework.©TheAuthor(s)2019.PublishedbyECS.ThisisanopenaccessarticledistributedunderthetermsoftheCreativeCommonsAttribution4.0License(CCBY,http://creativecommons.org/licenses/by/4.0/),whichpermitsunrestrictedreuseoftheworkinany medium,

Metals, even metals protected by organic coatings, are prone to atmospheric corrosion. The costs of repairing metal corrosion induced failures in military equipment (submarine, ships), bridges, railways, and subway systems are substantial. Corrosion under atmospheric conditions is greatly affected by the time of wetness (TOW) which is defined as the period in which a metallic surface is covered by adsorptive and/or liquid films that are capable of causing corrosion (International Standardization Organization 9225 standard). Corrosion in a marine atmosphere is extremely severe, due to a high humidity enriched with chloride ions (Cl ─ ). In China, with the construction of "The Belt and Road" and development of maritime silk route, the field of metal corrosion detection has recently advanced in the following areas: (1) big corrosion data are of great importance and should be accumulated for fundamental materials in typical environments especially sites near the maritime silk route; (2) service life prediction for metallic infrastructure (e.g., bridges, aircraft) and electric appliances (e.g., printed circuit boards); (3) when the corrosion mechanism is understood, methods for controlling the corrosion can be explored.
Methods to evaluate the atmospheric corrosion rate of metallic structural components are in constant demand. State-of-art methods for monitoring atmospheric corrosion and comparison of them are summarized in Table I, which basically classified as two groups: physical and electrochemical methods. Radio frequency identification device (RFDI) [3][4][5][6][7][8][9] has successfully developed for corrosion monitoring in recently years, with the advantages of battery-free operation and wireless sensing. However, RFDI analysis results were qualitative and less linked to corrosion rate. Other physical methods like K-band sweep frequency microwave imaging, 1 eddy current pulsed thermography, 2 electromagnetic field gradient, 3 though these results are visualized and useful in engineering, also give qualitative or semiquantitative results. Modern technology enables corrosion extents to be measured with advanced physical methods like time-lapse X-ray computed tomography, 4 synchrotron Radiation X-Ray Tomography 5 scanning Kelvin probe, and scanning acoustic microscopy, 6,7 but they are not applicable in field corrosion test. Traditional methods using weight loss or weight gain require long test time and do not give information about the instant corrosion rate. Another nondestructive method is based on the measurement of hydrogen gas evolving from a metal alloy, 8 but is not applicable in field atmospheric corrosion monitoring. Several image analysis methods such as gray value, wavelet z E-mail: dahaixia@tju.edu.cn, qinzhb@tju.edu.cn analysis and fuzzy KolmogorovSinai entropy has been developed to study the initial atmospheric corrosion of metals, the parameters derived from this methods are lack of physical significance therefore poorly related to corrosion rate. [101][102][103][104][105][106] Because corrosion is an electrochemical process, electrochemical probes/sensors a can be used to measure it. Changes in electrochemical activity caused by anodic dissolution, passivity breakdown, or the formation of a corrosion product can be easily detected by using electrochemical impedance spectroscopy (EIS), [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24] electrochemical noise (EN) 25-28 among other electrochemical methods. Atmospheric corrosion monitor (ACM), which consists of two metal couple, is based on the galvanic current measurement and was proposed by Mansfeld and Kenkel in 1976. 107 ACM has been extensively used to monitor TOW and the corrosivity of atmospheric environment, 108,109 but is not feasible in corrosion rate measurement. Sensors designed for corrosion detection in atmospheric conditions are used under a thin electrolyte. Factors that influence the measurement include relative humidity, the size of working electrode (WE) and the counter electrode (CE), and the distance between the adjacent electrodes. This review describes the types of electrochemical probes/sensors used for corrosion detection and corrosion monitoring, explains the challenges involved in such measurements, and suggests future work that can improve this field.

Electrochemical Fundamentals of Probes and Sensors for Atmospheric Corrosion Monitoring
For atmospheric corrosion monitoring, one can use either an electrochemical probe or a sensor. An electrochemical probe is a kind of integrated electrode system in which working electrode (WE), counter electrode (CE) or reference electrode (RE) are embedded and fixed in insulating material such as epoxy resin, with the working surfaces of all electrodes are exposed to the corrosive environment; the electrochemical sensor is also an integrated electrode system but does not include a WE, that is a major difference between a probe and a sensor. Basically, a probe or a sensor incorporates one, two or three electrodes, which is introduced below: (1) One-electrode system. 43,44 A metallic wire was usually used in this situation, one end of the wire was connected as WE, and the other end was connected as CE and RE. Zou et al. 44 recently developed a Fe wire probe to simultaneously and rapidly detect a The probe usually contains the working electrode (WE) while sensor is not. the polarization resistance, capacitance and totally accumulated change of metal resistance by EIS technique. One advantage of using this method is being able to monitor the instantaneous corrosion rate and accumulated corrosion damage without using a standard RE or CE. (2) Two-electrode system: WE, CE/RE (CE and RE are the same electrode). In these situations, the materials of CE and WE can be the same as WE, therefore it is a pseudo RE or CE. Because traditional standards REs (such as saturated calomel electrode, SCE) are not convenient to be embedded in a probe; in addition, the probes usually undergo polishing procedure for several times to produce a fresh surface, and standard RE can bot be polished. (3) Three-electrode system: WE, CE, RE. In this situation, CE and RE are two different electrodes, and they can be pseudo ones (similar to two-electrode system).
During the measurement, the probe or sensor can work under potentiostatic control, galvanostatic control, or under free corrosion potential. In the potentiostatic mode, a constant potential or varying potential is applied between the RE and the WE, and the polarization current flowing through the WE is measured. In the galvanostatic control mode, a constant or varying current flows through the WE, and the potential difference between the WE and the RE is recorded. It should be noted that the RE and the CE can be the same electrode. If no input signals are exerted, the corrosion potential and current fluctuations under the corrosion potential can be recorded successively or synchronously, that is, by measuring electrochemical potential noise (EPN) or electrochemical current noise (ECN).

Electrochemical Probes
Electrodes in the probes and the corresponding measurement methods.-Probes used for atmospheric corrosion monitoring can either be one, two, three, or multiple electrode systems ( Figure 1). The electrode materials in each prove can vary from one to another, depending on the research objective and the atmospheric condition. Shitanda et al. 43 developed a one-electrode system by using electrochemical impedance spectroscopy (EIS) to study the atmospheric corrosion of a screen-printed Ag circuit board ( Figure 1, probe #1). The charge transfer resistance of Ag wire can be obtaind by establishing an electrochemical equavalent curcuit. 43 The screen-printed approach is suitable for minitoring the atmohpheric corrosion of circuit board materials such as Ag, Sn, and Cu, but is not appicable to enginnering materials such as carbon steel and stainless steel which cannot be screen-printed. Other electrochemical methods-electrochemical noise (EN), linear polarization, for examples-are not suitable for this probe due to a difficulty in data interpretation.
Two-electrode probes with the same size and made by same materials are frequently adopted, as shown in Table II and Figure 1 (probes #2, #3 #4). One of the electrodes is used as the WE, and the other electrode is used as the RE and the CE. The two electrodes are embedded in an insulating epoxy resin, and the gap between the two electrodes is 100∼200 μm. The gap plays a crucial role during the electrochemical measurement, as discussed in Critical aspects of electrochemical probe design section. An electrochemical impedance spectroscopy (EIS) measurement is preferred for this kind of probe. EPN or ECN can be measured separately but their values cannot be obtained synchronously. Nishikata's group has done significant work using a Figure 1. Diagrams of electrochemical probes designed for atmospheric corrosion monitoring (yellowish-brown color corresponds to the insulating and structural material employed in the probes, usually using epoxy resin or polyurethane): #1 probe has one electrode; #2 probe has two square electrodes; #3 probe has two electrodes with comb shapes; #4 probe has two electrodes that are composed of multielectrodes; #5 probe consists of three nominally identical electrodes; #6 probe has three electrodes; #7 probe has multiple electrodes, each with a CE and an RE; #8 probe consists of an RE, a WE, and a CE. 2-electrode probe to study the corrosion rate of stainless steel 11,14 and weathering steels 22,45 (including carbon steel 9,10,46 ) and copper 9 using the EIS technique; they also investigated the influence of electrolyte thickness and pH on the rate of atmospheric corrosion of metals. 16 In these situations, the CE, RE, and WE are made of the same material. But WE and RE can also be two different materials. The WE is usually made with inexpensive materials but the use of inert materials (e.g., platinum) in the CE is costly. The three-electrode setup is more frequently adopted than the two electrode system (Figure 1, probes #5-#8), because it allows almost all of the electrochemical methods to be adopted. A CE is used in EIS and linear polarization measurements; the CE is usually composed of the same materials as the WE to reduce the cost of probe fabrication. To measure electrochemical current noise (ECN) at corrosion potential, the following setups can used: (a) three electrodes composed of identical material (RE, WE1, WE2); (b) two working electrodes made of identical material plus one RE made of different material (WE1, WE2, RE); (c) two electrodes (WE, RE) made of identical material plus a CE made of a different material (d) three different electrodes (WE, RE, CE). If an inert CE is used in electrochemical noise (EN) measurement, then its exposed area should be small compared to the area of the WE, to mitigate polarization to the WE (the theoretical background is discussed in Critical aspects of electrochemical probe design section). If the WE and the CE are made of the same material, polarization is very slight. The RE can be made of Ag/AgCl or high purity Zn. Use of a pseudo RE or a pseudo CE can produce electrochemical noise.

Critical aspects of electrochemical probe design.-Electrodes in electrochemical probes and the effective measurable area of the WE.-
Exposed areas of the RE, the CE, and the WE are vital factors in atmospheric corrosion monitoring. The exposed area of an RE should be small, to reduce the potential noise level. However, the noise level of the RE is usually neglected during EIS or polarization measurements, because the amplitude of the applied potential signal is much higher than the noise level of the RE. Although commonly used REs are stable in aqueous solutions, when they are utilized in atmospheric conditions, especially when the electrolyte film is not continuous, at low relative humidity an equilibrium potential cannot always be formed. To reduce the noise level of an RE, it is placed in the electrolyte solution (For example, NaCl solution) while the WE and the CE are covered with a thin electrolyte film. [54][55][56][57] The exposed area of a CE can be large or small. A large CE area guarantees a larger measurable current, which is an advantage when the corrosion current density is extremely low (pA level) in dry conditions. A small area of a CE can be used in localized electrochemical measurements. In aqueous electrochemical corrosion tests, the CE is placed directly facing the WE to obtain a uniform electric field between them. All of the probes listed in Table II satisfy this criterion. A low relative humidity can influence the results, leading to a measured value lower than the real corrosion rate.
The exposed area of a WE should not be very small in EIS and polarization tests, because a small area leads to a low measured current, sometimes too low to be resolved. The area of the WE in a probe should be several cm 2 (Table II). If passive metals are used, the WE should be passivated by embedding it in epoxy resin, to avoid corrosion at the metal/resin crevice. If electrochemical noise (EN) measurement is adopted, the area ratio of WE2 to WE1 should be high, that is, the area of WE2 should be larger than that of WE1 to enhance the signalto-noise ratio. 58 The electrode configuration is asymmetrical because WE1 are WE2 are composed of nominally identical materials. The area of the WE should be much larger than the area of the CE, in order to mitigate polarization of the WE by the CE. The corrosion potential of stainless steel after being coupled with five types of CE has been investigated; 59 the results showed that the polarization to the stainless steel (SS) was negligible if the area ratio of WE to CE was higher than 100 (see Figure 2). In the case of activation polarization controlled electrochemical reactions, the coupling current flow from the WE to   Figure 2. Correlation between the coupling potential Eg and the ratio of the WE area (S W ) to the CE area (S C ) of five types of CE (Pt-Ir alloy, Pt plated stainless steel, Pt plated Ti, Pt plated Nb, and Pt wire); as long as the ratio is larger than 100, the coupling potential is close to the corrosion potential of the WE. (Note that the corrosion potential of the WE, 304 SS, is 290 ± 10mV; the area of the WE is 1.34 cm 2 ). 59,61 the CE can be written as: 60 In Equation 1, I g are the coupling current flows through the WE and E g are coupling potential. S CE is the surface area of CE, i corr,CE is the exchange current density of the CE, E corr,CE is the corrosion potential of CE, and β a and β c are the Tafel slopes of the anodic and cathodic reactions occurring at the CE. To lower I g or to mitigate the polarization of the WE, one simple method is to diminish the S CE .
Gap between adjacent electrodes and its influence on electrolyte resistance.-Early work used a large gap of 1.5 mm between adjacent electrodes, 62-64 which led to a large electrolyte resistance. As shown in Table II, the gaps between adjacent electrodes are around 200 μm. A gap either too thick or too thin can affect the electrolyte resistance of the corrosion system. A thick gap results in a very large electrolyte resistance, therefore it is hard to detect corrosion. If the gap is very thin, as the corrosion proceeds, the corrosion product may extend over the gap and cause the adjacent electrodes to interact electronically.
Measurement methods.-The measurement method defines the shapes and types of electrodes in a probe. Polarization methods, including electrochemical impedance spectroscopy (EIS) and linear polarization, require two or three electrodes. For field corrosion monitoring using EIS, it is recommended to measure the impedance at only one or two frequency points, because testing the impedance over the whole frequency range (usually 10 5 ∼10 −2 Hz) wastes time, and measuring the impedance at one or two points requires a much simpler instrumentation. Shi et al. 23 used probe #2 listed in Table II to study the corrosion rate of carbon steel or 316L steel under ultrathin electrolyte films of 5-25 μm by EIS. They 23 found that impedance modulus at 10mHz (Z 10mHz ) produced values similar to the charge transfer resistance R ct . The group further monitored the atmospheric corrosion of weathering steel over time using probe #3 in Figure 1 by measuring the impedance at 10 kHz and 10 mHz every hour, 21 and found that (Z 10mHz ) −1 correlated well with the average corrosion rate (I corr ) determined from the mass loss due to corrosion. In later work, this group derived an empirical equation for weathering steel and for nickel-containing weathering steel under natural environmental con-  [2] Though polarization techniques can conveniently monitor atmospheric corrosion, they have some drawbacks: (1) The average electrochemical activity on the WE surface is not measured, therefore, the extent of localized corrosion cannot be identified. (2) The applied external potential perturbations slightly alter the electrochemical state of the WE, especially during long-term corrosion monitoring. (3) Corrosion form and early-stage corrosion initiation cannot be identified.
A major advantage of using EN to monitor metal corrosion is its in-situ feature; it does not disturb the corrosion system investigated. Moreover, it can continuously monitor the corrosion process, and early-stage corrosion events like stress corrosion cracking (SCC) initiation and metastable pitting can also be detected. The corrosion form can be determined from EN data using mathematical methods; for examples: the continuous wavelet transform (CWT), 65 the discrete wavelet transform (DWT), fast Fourier transform (FFT), chaos analysis, and more recently, the Hilbert-Huang transform 66-69 and recurrence quantification analysis. [70][71][72][73][74][75] Noise resistance and spectral noise resistance are two most frequently used methods to quantify corrosion rate in cases of both uniform corrosion and localized corrosion. A comprehensive review paper theoretical and mathematical models toward quantitative analysis of EN data has been published in this journal. 98 Noise resistance is the result of the standard deviation (SD) of EPN divided by that of the ECN. Spectral noise resistance is defined as the power spectral density (PSD) of the EPN divided by the PSD of ECN. Noise resistance is equal to polarization resistance for the activated dissolution controlled corrosion when using a symmetric electrode system. Spectral noise resistance can be equal to impedance modulus as estimated from EIS when using a symmetric electrode system contains a noiseless RE. 98 However, EN is still a qualitative method, especially for asymmetrical electrode systems. [76][77][78][79][80] Theoretical models and experimental tests for EN are needed to achieve a quantitative determination of the metal corrosion rate. 81,82

Electrochemical Sensors
Electrodes in the sensors and corresponding measurement methods.-A sensor used for atmospheric corrosion detection usually contains two electrodes, a CE and an RE. An electrical contact can be incorporated into the sensor to connect a WE with the electrochemical instrumentation. During corrosion detection, the electrodes in the sensor are placed perpendicular to the WE surface without direct contact with the WE. Table III lists examples of electrodes used to measure atmospheric corrosion. REs mainly include Ag/AgCl electrodes, high purity Zn wire, high purity Sb wire, and saturated calomel elec-trodes (SCE) with a lugging capillary, depending on the environment. Ag/AgCl, Zn, and SCE can be used in Cl ─ -containing environments, as shown below: Sb electrodes can be used in acidic environments, but cannot be used in neutral or alkaline solutions due to potential noise. 83 The CE and the RE can be a single electrode. Khullar et al. 84 reported that a sintered Ag/AgCl electrode can be used as both RE and CE for electrochemical measurements in thin film electrolytes. Many materials can be used in the CE, as shown in Table III. The role of the CE in a polarization test is to afford a current to flow through the WE. As the CE is needed to conduct the current, it should not be prone to corrosion. During the EN measurement, the noise level of the CE should be considered, and this is discussed in Critical aspects of sensor design section.
Critical aspects of sensor design.-Electrode area.-The exposed area of an RE should be small to decrease its noise. The area of the CE depends on the measurement methods. In potential-controlled techniques such as EIS and linear resistance, the area of the CE can be large or small. If a large area is exposed, the measured range of the WE is wider; a small CE area can detect corrosion only at localized sites. If microcapillaries are used (Figure 3e), corrosion detection is very localized.
Contact mode during measurement.-Five major contact modes are shown in Figure 3: (a) the RE and the CE are very close to the WE surface, and there is nothing visible in between; (b) a wetted filter paper is placed in between the WE surface and the RE and the CE; (c) a gelled electrolyte is placed between the WE surface and the RE and the CE; (d) a porous plastic net is placed between the WE surface and the RE and the CE; (e) a micro-capillary is placed between the WE surface and the RE and the CE. The advantages and drawbacks of each configuration are discussed below. The setup in Figure 3a is not easily achieved, especially in a field corrosion test, because the distance between the CE/RE and the metal surface is hard to precisely control. Under relatively low humidity, the corrosion current is very limited due to large electrolyte resistance and is therefore difficult to detect. 92 One promising feature of the setup in Figure 3a is that the exposed atmospheric condition is not greatly disturbed, and hence the corrosion proceeds under real atmospheric conditions. In Figure 3b a wetted filter paper (usually soaked in NaCl solution) is used as the electrolyte, which makes the electrochemical measurement effortless Figure 3. Diagrams of the contact modes when using an electrochemical sensor: (a) RE and CE are very close to the WE surface, and there is nothing in between; (b) a wetted filter paper is placed between the RE/CE and the WE surface; (c) a gelled electrolyte is placed between the RE/CE and the WE surface; (d) a porous plastic net is placed between the RE/CE and the WE surface; (e) a microcapillary is placed between the RE/CE and the WE surface.
due to a significant decrease in electrolyte resistance. But the introduction of such a wetted paper leads to a change in the corrosion environment, which might accelerate the metal corrosion. In the case of a gelled electrolyte, as schematically shown in Figure 3c, although the electrochemical test is easy, issues similar to those in Figure 3b are faced. Recently, our group used a thin porous plastic net to isolate CE and WE, 86 see Figure 3d. As the plastic net does not need to be wetted prior to electrochemical measurement, the impact on the corrosion environment is low. Lens paper may be another candidate for the separator, because it is thin (only several μm), porous, and economical. However, it is still not suitable for long-term corrosion monitoring, because a crevice is formed if a porous plastic net or lens paper are used. To resolve this problem, the shape of the CE can be changed from cylindrical to a thin porous metal net, to mitigate its impact on the atmospheric environment.
With the development of 3D printing and screen-printing techniques, the electrodes and the major structure materials can be fabricated rapidly, offering the advantages of customization and reproducibility. For instance, Rohaizad et al. 93 fabricated Ag/AgCl pseudoreference electrodes by 3D printing. Shitanda et al. 94 fabricated a threeelectrode type micro-electrochemical cell by screen-printing. Komoda et al. 95 screen-printed a silver/silver sulfate reference electrode with long-term stability.
Location of reference and counter electrodes.-The RE should be as close as possible to the CE, so that the RE and the CE detect at the same place on the WE. Alternatively, the CE can be designed as a wire circle around the RE, guaranteeing that they are facing the same detection area on the WE surface.
The fixation of the electrodes is also important. For WEs that are magnetic, three or more magnets can be attached to each leg of the sensor (Figure 3d), so it can sense the objective under study. For metals that are nonmagnetic, binding materials can be used to fix the sensor above the WE.
Measurement methods.-In field corrosion monitoring, stable and simple methods and electrochemical instruments that can work in harsh environments for a long time are highly appreciated. In our previous work, 96,97 a Compact (cRIO) module 9263 (National Instruments, USA) plus a self-made ZRA circuit was successfully operated in a Zhoushan marine atmospheric condition for a half year of corrosion monitoring. 96,97 Corrosion data were recorded at 15 minute intervals automatically and data were saved on a flash disk. 96,97 Methods that do not disturb the corrosion system (in-situ) are preferred. EIS and linear resistance measurements are robust tools to determine corrosion rates of metallic materials, but are slightly destructive to the corrosion system under investigation. Electrochemical noise (EN) is an in-situ technique that is suitable to identify localized corrosion on passive metals, but quantifying the atmospheric corrosion rate with theoretical and mathematical models is challenging. Another question that significantly affects the quantitative analysis of EN data is that the noise level of the pseudo CE (e.g. if a Pt electrode is used). As the measured ECN is a reflection of current transients produced on both WE and CE, therefore, if the noise level on CE is significantly lower than the WE, the measured EN data are a reflection of the corroding electrode-WE. Overall, the noise level of CE should be evaluated prior to EN measurement, to make sure the EN data are valid. 98

Comparison of Probes and Sensors
A probe has a flat and open surface so the electrodes are well exposed to the atmospheric conditions. The sensor must be put above the WE surface and the RE and the CE must be close to the WE surface, therefore the sensor is partially covered by the WE surface, reducing its exposure to the environment. A sensor is not fit for long-term corrosion monitoring, but is adept at short-time corrosion detection.
Since atmospheric corrosion proceeds under dry-wet cycles, pitting corrosion can be initiated in wet conditions. The use of wetted filter paper and other damp materials during corrosion detection can accel-erate corrosion of the metal under investigation. Also, the corrosion environment should not be disturbed during testing.

Challenges and Future Work
Electrochemical measurements at low relative humidity.-Atmospheric corrosion is highly dependent on the material exposed and the relative atmospheric humidity. In relatively dry conditions when the electrolyte film is very thin and discontinuous, the corrosion rate is low and is difficult to measure with electrochemical methods. When the corrosion current is lower than pA in magnitude, it may inaccurately indicate the corrosion status of metal-based structures. In order to build process-based atmospheric corrosion models, further work is needed to (1) understand the wetting behavior of corroded and contaminated surfaces and (2) to detect when electrolyte is present. 99 Since low corrosion rates are a problem, future work should develop electrochemical instrumentation with high potential and fine current resolution.

Electrochemical instrumentation development aims at portable, intelligent, wireless corrosion detection.-As the cost of corrosion in
China was approximately 310 billion USD in 2014 (3.34% of the gross domestic product), 100 it is of great importance to detect corrosion and control it. Existing electrochemical instruments are single channel or multichannel instruments that can measure the corrosion at different sites on metallic structures. The instruments are portable and intelligent, and can work in harsh environments for a long time, with the data obtained being stored on a flash disk. Wireless detection enables indoor control of outdoor measurements. Often, the sensor or probe applied in field corrosion monitoring contains a tiny electrochemical circuit and is self-powered.
Big data processing and interpretation.-Long time corrosion monitoring generates a mass of corrosion data. Data processing software such as LabVIEW can serve this purpose. We developed an electrochemical noise (EN) data batch processing software based on Lab-VIEW 2010, which can output statistical results rapidly. 96 EN data are not always interpreted successfully, therefore new models based on EN sources need to be developed. The aim is to obtain an accurate corrosion rate by in-situ electrochemical methods.
Electrochemical probes/sensors.-Electrochemical probes are successfully used in atmospheric corrosion monitoring, although they cannot be used for metallic components or structures. Problems with field corrosion detection by sensors include: (1) determining the correlation between the area of a CE and the measured area of a WE; (2) development of sensors that do not greatly disturb the corrosion environment for long-term corrosion monitoring; (3) discovery of novel electrode materials that enhance the signal-to-noise level during electrochemical measurements.
Assisted by corrosion simulation software such as Comsol, the electric field distribution between the CE and the WE can be defined, which can help to solve problem (1) mentioned above.

Conclusions
Electrochemical probes and sensors used to detect and monitor atmospheric corrosion were reviewed, leading to the following conclusions: (1) An electrochemical probe can be comprised of one, two, or three electrodes, two or three being preferred. Two major factors that affect the electrochemical measurement results are the gap between adjacent electrodes and electrode size. (2) Electrochemical sensors usually have two electrodes, a reference electrode (RE) and a counter electrode (CE). Gaps between the CE and the WE and the size of the CE are two important considerations, because the gap hugely affect the electrolyte resistance that measured, and the size of the CE mainly affect the magni-tude of the measured current. Various methods are used to isolate the CE from the WE to reduce the electrolyte resistance. But such isolations using wetted filter paper can affect the corrosion environment. (3) The measurement methods should be determined prior to probe and sensor design, because the methods used will define the type and size of the RE and the CE. (4) To determine atmospheric corrosion rates more accurately, future work will emphasize: (a) novel sensor and probe design that can measure corrosion rates more accurately; (b) data processing methods that enable a rapid output of results; (c) the development of corrosion models that reflect corrosion rates; (d) more realistic corrosion simulation to assist sensor and probe design.