Fiber optic sensors to monitor reinforced concrete corrosion

Reinforced concrete is subject to corrosion, which ultimately leads to total failure of the structure. In the first stage, elements such as moisture and chloride penetrate the concrete. Upon reaching the reinforcing steel bars, high concentrations of these elements attack the passive layer of hydrated iron oxide that protects the bars against corrosion. Rust begins to form, ferric compounds convert to ferrous, and the volume of the structure expands by five to 10 times. Immense pressure starts to build, causing more cracks. If this process continues unchecked, the structure eventually fails.

Therefore, corrosion monitoring is essential to preserve the life of civil structures, and early diagnosis and monitoring of seemingly healthy concrete enables pre-emptive corrosion control measures. However, identifying the specific causes of corrosion can be challenging, since there are several causes.¹ Deterioration and cracks can be the result of poor concrete mix, poor workmanship, inadequate design, shrinkage, chemical and environmental attack, physical or mechanical damage, and corrosion of the reinforcing steel (RS). Here, we present a road map for developing a suite of sensors and an associated system to detect corrosion in the RS and the degree of damage at a certain moment in time.² Our proposed system enables high resolution and would offer an accurate estimate of the remaining life in a concrete structure.

To develop our approach, we considered a corrosion detection technique that has been exhaustively researched over the last 20 years: fiber Bragg grating (FBG) sensors.³ These induce a periodic variation in the refractive index of the fiber core to enable selective reflecting of certain light wavelengths (and transmission of others), indicating a change in the parameter under observation (temperature or strain, for example). FBG sensing uses wavelength multiplexing technology, which combines optical carrier signals into a single fiber. However, in practical applications, FBG sensors are not sufficiently accurate. The usual method is to embed the FBG in a flexible structure and measure its wavelength shift by examining its reflection spectrum, but such methods are expensive and therefore not widely used.

By understanding the stages of corrosion (see Figure 1), we can develop special sensors to improve on the accuracy of the FBG approach. First, we can monitor moisture ingress using a fiber optics humidity sensor. Increased moisture better enables chloride to penetrate the concrete, accelerating the corrosion process and causing large volume expansion. Ultra-sensitive fiber optic pressure sensors can detect this expansion, and the resultant cracks introduce acoustic emissions, which we may detect using a high-frequency vibration sensor based on phase-shifted gratings. These are in-fiber interferometers with very narrow linewidth. They feature a very narrow transmission notch of about 5 picometers, which is created by deliberately erasing a small group of periods in the middle of the grating length. To detect the relative change in this notch, we have adapted an optical lock-in amplifier called the Pound-Drever-Hall technique, which was first used in astronomy studies to detect extremely weak signals from distant stars.

Figure 1. The various stages of concrete corrosion over time. CO₂: Carbon dioxide. Cl ⁻ : Chloride.

Using phase-shifted gratings, we can adapt the sensor to monitor humidity and to detect changes in parts per million. Such sensors would be 100 times more sensitive than normal fiber gratings, and they could also enable monitoring of pressure rises inside the concrete.

In summary, there are three sensors we aim to develop: a moisture/chloride sensor, a pressure sensor, and a crack detection acoustic emissions (AE) sensor. To obtain the best signal from inside the concrete, the phase-shifted gratings need various packaging methods that enable mounting and coupling. Some packages involve coating the fiber with a special material, which upon contact with the elements of interest demonstrates strain or relaxation. For example, the moisture sensor comprises a fiber deposited with a thick coating of polyimide that is shrunk by curing at 300°C. When the fiber comes into contact with moisture, it releases the stress. Such types of non-destructive tests primarily measure the magnitude of the energy released within a material when physically strained. We expect AEs from concrete cracks to have a high frequency signal that is very easily absorbed by the structure. Therefore, sensors need to have superior signal-to-noise performance, of around 50 decibels.

Our future work will focus on AE sensor development, and we plan to complete a field test of this method by the end of 2014.