Corrosion of Different Materials Connected to Carbon Fiber

In this study, Impedance Spectroscopy (EIS), was performed. The corrosion resistance versus time of exposure of both carbon fiber and metal alloys were thoroughly examined and compared for just over a 200-day period. Collecting this data revealed drastic changes in impedance values for several of the 11-examined fastener/carbon fiber interconnections immersed in 3.5% sodium chloride (NaCl) solution, from day one of exposure to day sixty specifically. The range of impedance values directed the preselection of one stainless-steel and titanium fastener for further assessment, with the goal of recognizing that EIS could detect trends of corrosion and degradation of material. Equivalent R/C Circuit modeling was created and conducted from the impedance data obtained via potentiostat for selected stainless steel and titanium fasteners. This was done to determine how many interfaces the interconnected model contained. After trials of 1RC, 2RC, 3RC and 4RC imbedded circuit analysis, the identification of three overall interfaces was suggested. This meant the interconnected system contained three interfaces that were reacting with seawater within the replicated galvanic system. After EIS and equivalent R/C circuit analysis was complete, the identification of interactions between the interfaces and what type of surface changes had taken place was completed by a well-known electron microscopy process. Scanning electron microscopy-energy dispersive spectroscopy (SEM-EDS) was conducted on both selected fasteners and carbon fiber coupons. The primary focus was on the metal fasteners, the carbon fiber was to be studied in the future. The enhanced photos revealed corrosion of the stainless-steel fastener. Specifically, there was signs of crevice corrosion on the outer threads in between two interfaces. There was also the identification of chlorine atoms on the surface of stainless steel fasteners, recognizing the cause of corrosion was by chemical reaction and not mechanical failure.


LIST OF TABLES
The fasteners are used for the structure, vital systems and through hull penetrations which are directly exposed to the seawater environment, ultimately corroding after time and exposure. Therefore, by utilizing modern day electro-chemistry, corrosion and analysis techniques, a display of what could happen with metal selection would be investigated and proven.
There were three main investigations of carbon fiber to metal galvanic interactions in this study, by applying electron impedance spectroscopy (EIS), equivalent R/C modeling and lastly scanning electron microscopy-energy dispersive spectroscopy (SEM-EDS). These investigative techniques were selected due to having a history of use and proven to be extremely accurate within the world of corrosion evaluation. The materials used were chosen for their known properties and also their ranking on the galvanic series. [1,2] The carbon fiber was sourced from the Department of Defense and its overall lay-up and properties are proprietary information that was not disclosed for the study. However, carbon itself was designated for multiple reasons, one due to the fact that it serves as a conductor and is very noble on the galvanic series. [1,2] When connected to some metal alloys this creates an anodic reaction for the metal and a cathodic reaction for the composite, leading to corrosion of the metal, along with de-lamination, fracture and blistering of composite vessel hulls. For this study, stainless-steel and titanium metal alloys were the metals chosen for the anode of a galvanic system. Each metal alloy was in the form of fasteners to be connected to the carbon fiber via tap and threading. Titanium was utilized due to its inherently strong resistance to corrosion and high nobility. In comparison, stainless-steel is much lower on the Galvanic Series and is less noble.
The idea was to select metals that were far apart on the Galvanic Series and also

THEORETICAL BACKGROUND -ELECTROCHEMICAL PROCESS
In this study an electrochemical process was the main operating process to cause damage. This process occurs when there is a known transfer of electrons between atoms and molecules, which either go to or from them. This results in the change of oxidation state. This process transpires when a voltage is applied directly to a metal alloy for example a difference in potentials between materials. With this reaction there comes the depletion or degradation of the materials involved in the process. This progression is normal in the marine environments, especially in seawater. The ocean water is approximately 3.5% NaCl and is an extremely good environment for corrosion to take place.
[2] There are two main factors that cause this corrosion specifically to occur, one being conductivity and the other oxygen solubility.
Metal alloys react differently in this environment due to their variable compositions resulting in nobility and individualistic properties. With that said, it is important that the control of corrosion is to be consistently studied and prevented.

THEORETICAL BACKGROUND -CORROSION
There are several types of corrosion that occur on the surface of different metal alloys. Each type of corrosion is dependent on several factors. For the purpose of metals in a seawater environment the following corrosion cases can occur; crevice, pitting, stray current and galvanic corrosion. Corrosion is initiated by having four necessary elements. The interaction of these elements directly leads to corrosion.
First is an anode, which is the electrode where oxidation reactions generate electrons and positive metallic ions are formed from atoms. Normally the anode is deemed the "sacrificial anode", which is the metal that in fact corrodes. The second element of this cell is a cathode, which is the electrode that is on the receiving end of the electrons produced by the anode. In the marine environment, negative hydroxyl ions will be produced at the cathode. The cathode is usually totally protected from corrosion. However, in the case of carbon fiber composites, these hydroxyl ions can cause damage in the form of blisters or cracking as they combine with positive sodium ions and form sodium hydroxide on the carbon fiber surface. Sodium hydroxide forms an osmotic cell which increases pressure inside the composite, resulting in blisters or cracking of polymer. The third element is an electrolyte, which would be seawater.
This liquid will serve as the main conductor to which current can flow and be carried.
Lastly, the fourth element is a return current path, meaning a conductive pathway is needed to connect the anode to the cathode, which in this case is direct contact between the metals and carbon fibers in the thread region. If one of these elements is not present then it is fair to say that corrosion will not occur. When corrosion does occur on the surface of metal, it is affected in the forms of pits, crevices and overall material loss.

THEORETICAL BACKGROUND -EIS
Electrochemical Impedance Spectroscopy (EIS) has been utilized for several years to detect and monitor corrosion rates of materials over a period of time. What the process completes, is the measurement of a metals impedance to current flow, which is also directly related to the corrosion current. The impedance of a material is defined as the difference of frequency-dependent potential divided by the frequency- Depending on how many interfaces the metal-based system being measured contains, the number of parallel RC elements in the model to be applied will be increased to represent each interface. The EIS response of an equivalent circuit can therefore be calculated and compared to the actual EIS response to an electrochemical cell. This means that the measured impedance can be compared to theoretical models and the individual circuit elements values can be computed to provide the theoretical data.
These elements can be compared in order to prove how many interfaces from a is commonly used in several engineering applications. Its purpose is to obtain detailed high-resolution examination of surfaces of a material being studied, to determine if any surface dependent processes have occurred that changes the appearance and provides information towards the mechanism for the changes. This technique depends on an electron beam interacting with the surface of interest. A primary electron beam is directed onto a material surface, then electrons will be emitted from that surface, which include backscattered electrons, secondary electrons and in addition x-rays.
These secondary electrons are collected by a detector within the system and from the magnitude of the collected signal as a function of spatial position on the materials surface, an image is produced. Backscattered electron images differentiate areas by atomic number, meaning it allows the identification of heavy or light elements that are present on the surface of the material being studied. If the x-rays emitted from the surface are collected they can be analyzed by their energy, which is unique for each element, permitting detection of elements on the surface being inspected. In this study metal fasteners will serve as a WE. The application of small sinusoidal potential of a fixed frequency induces a current with the WE then the impedance, the voltage divided by the resulting current is computed at each frequency.
By using small amplitude, it does not disturb the properties of the material being evaluated, essentially making EIS a non-destructive test (NDT). This directly correlates to what is happening on a samples surface in terms of corrosion rate. A common three electrode electrochemical cell is utilized for the entire cycle of the experiment conducted on the carbon fiber and metal samples. The EIS set-up is detailed in Figure 1. [11] The carbon fiber sample is detailed in Figure 2, reconstructed using a 3D modeling program.

EXPERIMENTAL TECHNIQUES
Carbon fiber test coupons were machined to be an approximate cross section of There were long-term (LT) and short-term (ST) samples created for this experiment.
Time dependent impedance data was recorded.
In terms of the experimental set-up, the electrodes utilized were the counter electrode (CE), reference electrode (RE) and the sample, which was the working electrode (WE). The CE used was platinum, this electrode was used due to its stability [1,2], the WE were the actual carbon fiber/fastener interconnection and the RE was a saturated calomel electrode, Figure 6. The electrochemical impedance |Z| of each sample was measured at intervals to collect time dependent data. All the electrodes were electrically connected to a PC via cell cables in order to acquire proper readings.
The system was calibrated each time readings were taken in advanced to ensure accuracy using a "Dummy Cell", Figure 7. A pre-test was performed on other steel and aluminum metal alloys in order to validate the system was taking accurate readings, this delivered precise analyses for each sample connected. Respectively, each cell cable was color coordinated in order to guarantee correct experimental setup, Table 1.
After initial set-up the EIS was performed, recording data, which commenced at day 1 and continued to day 30 and ultimately until just over 200 days. The overall testing schedule for recording purposes was days 1, 3, 5, 10 then every three days until in values were recorded in order to identify future trends or stability of the material being evaluated. The potentiostat was fixed with initial parameters detailed in Figure   8, which remained consistent for entire study.
Calculations of impedance data versus area of fastener exposed inside the three-electrode electrochemical cell were derived using a dial caliper, Table 5. This determined if the amount of area exposed (fastener) associated with the impedance values obtained via potentiosat, Table 2. The main focus after obtaining the impedance data was to somehow correlate or compare the data to the amount of metal surface area exposed to the seawater solution. This would determine if the higher or lower |Z| values were surface area dependent. This would indicate that the higher the value of impedance obtained was due to the surface area exposed for each metallic alloy.
Subsequently, when the surface areas were obtained the next phase was to attempt to identify which samples were showing the most significant changes in impedance over the time exposed. This would recognize any developments of increase or decrease in values and be used as a secondary proof, isolating which metals needed to be studied further. Graphs were produces to represent at frequency of 0.01 Hz and 60 days of exposure, detailing how stable or unstable the samples actually were and if values were similar or different, Figures 10, 11, 12, 13. Also a bar graph was generated to represent all the stainless steel (SS) and titanium (Ti) impedance readings from day 1 and over day 200, Figure 9. This information was specifically utilized to isolate which samples were to be chosen for supplementary analysis.

DISCUSSION OF EXPERIMENTAL TECHNIQUES
In general, both these materials, stainless steel and titanium alloys obtain their corrosion protection by forming passivating surface layers during exposure. Titanium alloys are thought to be stronger passivators then stainless steel. This is mainly due to the stainless steels relying on chromium for passivation, which is usually only 18% by weight of the material, while titanium is usually around 90% by weight in alloys. The greater amount of passivators in the titanium alloy would suggest a better passivating system, especially as Pourbaix diagrams indicate that titanium has a much more stable passive layer over a wide pH range. Indeed, data from US Navy test on crevice corrosion indicates 304 and 316 stainless steels were not particularly resistant to crevice corrosion, and this is also supported by experimental data. On the other hand, titanium alloys are strongly resistant to crevice corrosion in a marine environment from both empirical and experimental data. One interpretation of the data would then follow along the lines that the continual decrease in long term data, along with some scatter would be indicative of an unstable passive film for stainless steel, while a very stable impedance value after a slight decrease would be indicative of a stable surface condition for the titanium.
For stainless steels, the presence of the chloride ion in solution is a destabilizing ion for the passive film that is formed. As it is present in the sodium chloride solution, then crevice corrosion would be expected. In addition, the cathodic carbon fibers would polarize the steel into its anodic region and tend to induce crevice corrosion. For titanium alloys the chloride ion at neutral pH ranges is not a destabilizing ion and crevice corrosion would not be expected.
After all the EIS data was collected and analyzed, the next phase was to determine how many actual interfaces were required to model the experimental data to provide an indication of the processes occurring. The way this was to be accomplished was by performing equivalent R/C circuit modeling. This process has been completed in the past and in previous studies for composite materials and metals containing surface layers, for example paint on the surface of a steel vessel hull. [6,7] The idea is to represent the physical sample in an analog circuit. Bode plots will be generated from analyzing the single circuit and the curves will be compared with the EIS data |Z| curve. Both curves will be evaluated to determine if they match. The notion is if they match and are a "true fit", then the exact number of interfaces will be confirmed. This modeling will be described in the next chapter.       To determine which circuit would work best and give the most accurate analog display of the interconnected model, trial and error was employed. The R/C circuits also had to be created similarly to that of what is known to be done with a close looped circuit that is in series. This way if a 2R/C circuit did not operate and did not show a good "fit" in comparison to the impedance curve selected, then a 3R/C circuit would be attempted.     The carbon fiber test coupons were similarly removed by hand from the plastic tubes with minimal heat applied and placed in a separate container and dated accordingly. The next phase was to determine how to clean the surface area on the fasteners that retained adhesive outside of the area either in the carbon fiber or in the salt water, without causing any mechanical damage or remove valuable information.
It was critical to preserve any surface coatings that were created due to the fasteners being immersed in the electrolyte for the extended period of time. A low power rotary tool was utilized, along with a rotary brass cleaning brush bit. This cleaned the area which retained adhesive, Figure 28. After the fasteners were stripped, they were then prepared by placing a piece of copper tape to delineate the exposed area of carbon fiber and salt water contact and ready for SEM analysis, Figure 30. This was extremely important for the SEM-EDS progression, and the ability to distinguish the exact area that was exposed to the electrolyte solution as well as the carbon fiber/fastener interface. The specific distances were already determined previously, which can be identified in Table 5 Table 2, Table 3 and Table 4. After about the first 10 days of exposure to seawater solution the impedance remained in the range of 2.67E+03 Ω to 2.78E+03 Ω. This was preliminarily expected to happen for the titanium sample, due to nobility and high ranking on the galvanic series [1,2].
EQUIVALENT R/C MODELING For the purposes of modeling, the potentiostat software contained pre-created models which were used initially to see how good a fit they were. There were two main equivalent circuit models used, a carbon fiber based and the other paint based.
The paint-based R/C model can be viewed in Figure 27. These models were the starting point for developing an EIS model using electrical circuit elements. This led to correlating theoretically how many interfaces the inter-connected model created contained. Therefore, the model was analyzed first for possible interfaces within the galvanic system of a metal fastener in a carbon fiber composite. A detailed 3-D model was completed to represent the hypothetical interfaces that reacted with the electrolyte solution, 3.5% NaCl, Figure 22. After trial and error, the two pre-created models were not showing accurate results via Y2 Fit Z-Curves.
This then led to attempting to re-create circuits that would represent what physically is occurring in the cell on an analog scale. Initially a 2R/C circuit was created then a 3R/C and lastly a 4R/C, Figures 23, 24 and 25. After this was completed then circuit analysis took place, specifically for SS2 and Ti2. Each Y2 Fit Z-Curve was generated by the potentiostat software and each R/C model was ranked for goodness of fit from 1 to 3, best fit to worst, or a tie in between a respective two models, Tables 6 and 7.
After review of the erratic points that corresponded to certain dates when EIS was conducted, it was suggested that overall that the 3R/C circuit was the best and true fit. What this meant was that the SS2 and Ti2 fasteners contained the same number of interfaces, which are indicated to be three overall. These interfaces are as follows; carbon fiber to fastener, fastener to electrolyte and lastly carbon fiber to electrolyte, as shown in Figure 22. This successful circuit analysis further indicated that the initial assumptions and theories at the initial set-up of the inter-connected model.
Previous research involving carbon fibers under an impressed voltage, from a potentiostat [9] modeled EIS data, and successfully used a 2R/C model. As shown here this model was not applicable when the voltage of the system was controlled by galvanic interactions. To understand this, further work should be considered as it is important to understand the exact process occurring.

SCANNING ELECTRON MICROSCOPY-ENERGY DISPERSIVE SPECTROSCOPY
After performing SEM-EDS the following results were identified for the galvanic samples. The SS2 fastener showed direct crevice corrosion in between two interfaces. These interfaces were the metal to electrolyte and the other being the carbon fiber to fastener. In Figure 36 and 38, there was clear crevice corrosion present at the exposed area of the fastener. Figure 37  When the SEM-EDS was completed for SS2 and Ti2, there was visual support for corrosion at the crest of the stainless-steel fasteners that were immersed in seawater. The form of corrosion that formed on the stainless-steel fastener, SS2, was crevice corrosion. There are mechanisms for this type of corrosion, which is considered localized [1]. Crevice corrosion usually occurs on a metal that comes into contact with another mating surface. In this case, the mating surfaces are the carbon fiber composite and metal interface. The crevice actually can take place within a crack of the metal or under a surface deposit in a form of an acid solution. This reaction that takes place and causes a depletion of oxygen and is consumed. Usually crevice corrosion takes a long period of time, however when you have a case of extremely dissimilar or less noble metals, the process can be accelerated. Titanium is significantly more resistance to crevice corrosion than stainless steels for that very reason, its nobility is much greater.
The carbon fiber was not focus of this research, however future experiments should take place to determine the adverse effects that the material withstood, if any.
It can be implied that by performing SEM-EDS after EIS data and equivalent circuit modeling, that the SEM-EDS information supported the suggestion that scatter and the large drop in the EIS data were indicative of corrosion, while a lower drop in EIS data with time and stable EIS data over a long period indicated no corrosion was occurring.
It also indicated the mechanism of corrosion was in fact crevice corrosion. As crevice corrosion occurs at an interface, it is hidden from view. This is a then an interesting situation, as the threads of the fastener in the case of stainless steel are being removed, reducing the load bearing capability of the fastener. For the case of titanium, no crevice corrosion was found and so the fastener and carbon fiber composite would be a much more stable load bearing situation. Further work on the reduction of load bearing capability by the crevice corrosion process would be interesting to continue further investigation. After performing equivalent R/C modeling a 3R/C EIS model was found to provide the best fit of EIS experimental data for a galvanic system, independent of metal alloy behavior. Then SEM-EDS process revealed that corrosion in fact took place on the stainless-steel fasteners. Although only one was identified for further investigation, it can be directly inferred that the other five fasteners exhibited similar behavior. To be exact crevice corrosion took place after the formation of sodium chloride atoms on the surface of the metal and having the other three elements needed for corrosion to exist. As for the titanium fasteners, it was originally theorized that the metal would not exhibit much if any corrosion behavior. This was due to previous studies along with the material properties of the metal itself, being extremely noble.
The carbon fiber did show signs of delamination; however, it couldn't be determined if it was caused by a chemical reaction or from mechanical failure from tapping the coupons.
Overall there is minimal or no research completed in terms of impedance within a "galvanic system". The outcome was not certain upon the start of the experimentation; however, it was proven that impedance values can be studied within a three-electrode connection. It has also been confirmed that the research has directly related corrosion to the instability and magnitude of the absolute drop of impedance.
Lastly it was proven that corrosion occurred on the stainless-steel fasteners and the identification was directly attributed to the process of EIS and equivalent R/C modeling.

MECHANICAL TESTING
The study of corrosion of metals and degradation of carbon fiber that is interconnected and immersed in seawater needs to be constant. It is vital to understand that this form of corrosion can be detrimental and ultimately lead to engineering failure at a large scale. There were certain items that were not accomplished in this research.
Mechanical evaluation needed to be completed in the form of tensile (tension) and Barcol hardness tests. Meaning, the carbon fiber test coupon in its entirety, a coupon that was tapped then ultimately a coupon tapped and subjected to the seawater environment. The tensile test of these three individual specimens would produce data relevant to the overall ultimate tensile and breaking strength along with significant elongation properties of the carbon fiber. The Barcol hardness test would be performed to gain an idea or a baseline of how the carbon fiber would react under loads and also how the fasteners when corroded were affected in terms of material properties and ability to remain whole.

FUTURE SCANNING ELECTRON MICROSCOPY EVALUATION
Although; due to the data and analysis completed by EIS and Modeling, which led to a significant corrosion evidence. There were also nine other fasteners that needed to be evaluated under SEM-EDS. Those specific fasteners were filled with epoxy resin, which can be seen in Figure 58. This was an original intended process; however due to time constraints, total cutting, surface preparation and polishing it was unable to be accomplished. The idea was to cut the samples in half creating a cross section that would then be slowly sanded down to the surface of the treaded region.
Then the samples would be polished and placed into the SEM-EDS chamber for assessment. A burn out test also was unable to be completed, to determine the official lay-up and thickness of the carbon fiber. This specific test would also reveal how the material would withstand in extreme changes in temperature, hot/cold. Since the material was proprietary and distributed by the U.S. Government, the measurements taken for the purpose of research were done by hand. The determination of these unknowns, would aid in understanding the overall behavior of metals and materials subject to this environmental condition.