Effect of humidity on fiber-optic temperature sensing

Contents lists available at ScienceDirect Measurement journal homepage: www.elsevier.com/locate/measurement Effect of humidity on fiber-optic temperature sensing Kun Wang ∗, Marcus Schukar , Konstantin Hicke , Xin Lu Bundesanstalt für Materialforschung und -prüfung (BAM), Unter den Eichen 87, 12205, Berlin, Germany A R T I C L E I N F O Keywords: Distributed sensing Humidity effect Temperature sensing Optical fiber coatings Temperature uncertainty A B S T R A C T This study investigates the influence of ambient humidity on the temperature sensitivity, measurement accuracy, and uncertainty of optical fibers with different coatings, using a Rayleigh-based distributed sensing technique. Polymer-coated fibers (acrylate and polyimide) and metal-coated fibers (copper and gold) were tested under controlled humidity (30%–90% RH) and temperature (20–60 ◦C) conditions. Rayleigh-based measurements revealed a slight but consistent decrease in temperature sensitivity with increasing humidity for all polymer-coated fibers, attributed to humidity-induced changes in coating stiffness and strain transfer. In contrast, metal-coated fibers exhibited humidity-independent behavior and superior stability. These findings highlight the non-negligible role of ambient humidity in fiber-optic temperature sensing, particularly in Rayleigh-based systems using hygroscopic coatings. The results provide practical guidance for fiber and coating selection in humid environments and offer broader insight into humidity–strain–temperature coupling mechanisms relevant to other fiber-optic sensing mechanisms. 1. Introduction Fiber-optic temperature sensors are extensively utilized in harsh or inaccessible environments, including nuclear reactors, geothermal wells, subsea pipelines, wind turbines, and structural health monitoring [1–6], due to their inherent immunity to electromagnetic interference, compact size, and suitability for harsh environments [7,8]. Based on their operating principle, fiber-optic temperature sensors are generally categorized into point sensors, such as fiber Bragg gratings (FBGs), which measure temperature at discrete locations, and distributed temperature sensing (DTS) systems that provide continuous measurements along the entire fiber length [9,10]. DTS systems typically employ Rayleigh, Brillouin, or Raman scattering, each offering distinct trade-offs in spatial resolution, measurement range, and system complexity [11–15]. Both pointand distributed-based fiber-optic sensors have experienced significant advancements in sensitivity, spatial resolution, and interrogator range. However, achieving reliable long-term temperature monitoring remains challenging due to environmental fluctuations [16], where strain and humidity are the two most common and influential factors in practical applications. The influence of strain primarily manifests as temperature–strain cross-sensitivity, significantly affecting measurement accuracy. This issue has been extensively studied, and various hybrid sensing schemes have been proposed to mitigate it [17,18]. As compensation techniques for strain effects have matured, humidity has become the dominant environmental factor limiting the performance of ∗Corresponding author. E-mail address: [email protected] (K. Wang). fiber-optic temperature sensors. Several studies have investigated the influence of humidity on distributed fiber-optic sensing, particularly in polymer-coated fibers. For example, polyimide-coated fibers have been shown to exhibit humidity-induced strain transfer variations due to their hygroscopic properties [19,20]. Others have proposed strategies to extract humidity-insensitive temperature measurements, either through coating design [21] or machine learning-based decoupling methods [22,23]. While these studies provide valuable insights, most are limited to specific fiber types or narrow environmental conditions. Despite these efforts, the lack of a comprehensive investigation into humidity-induced effects still significantly limits the achievement of highly accurate and precise temperature measurements. In this study, we present a comprehensive investigation into how ambient humidity influences the temperature sensitivity and measurement uncertainty of optical fibers with different coatings, specifically polymer and metal. Using the Rayleigh-based distributed sensing technique, we evaluate fiber performance under controlled temperature (20–60 ◦C) and humidity (30%–90%) conditions. While the observed absolute changes in sensitivity are relatively small, typically within 2%, they become non-negligible in high-precision applications, where environmental stability and uncertainty must be rigorously considered. Furthermore, we quantify the measurement uncertainties under varying humidity levels to assess the robustness of different fiber coatings. These results provide insight into the reliability of different coatings and guide the design of DTS systems for long-term use in humid environments. https://doi.org/10.1016/j.measurement.2025.119905 Received 2 September 2025; Received in revised form 8 November 2025; Accepted 27 November 2025 Measurement 261 (2026) 119905 Available online 28 November 2025 0263-2241/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ). K. Wang et al. 2. Theoretical background 2.1. Physical principles of fiber-optic temperature sensing Fiber-optic temperature sensors are broadly classified into point sensors and distributed sensors. Point sensors, such as FBGs, measure temperature at discrete locations along the fiber. FBG sensors operate by detecting shifts in Bragg wavelength, which are induced by temperature-induced changes in the refractive index and fiber length. They offer high resolution and low uncertainty but require multiple gratings for spatially resolved monitoring, which limits scalability and increase system complexity. In contrast, DTS provides continuous temperature measurement along the entire fiber length. DTS systems typically exploit one of three scattering mechanisms: Rayleigh, Brillouin, and Raman scattering. Among these, Rayleigh-based distributed sensing, particularly using optical frequency domain reflectometry (OFDR), tracks spectral shifts in Rayleigh backscatter pattern (with sub-picometer resolution), a unique ‘‘fingerprint’’ resulting from random refractive index fluctuations inherent to the fiber. This interferometric nature makes OFDR highly sensitive to local strain transfer, especially in fibers with environmentally responsive coatings. OFDR systems can achieve submillimeter spatial resolution and sub-picometer spectral sensitivity [24, 25]. Brillouin-based distributed sensing, such as Brillouin optical timedomain analysis (BOTDA), relies on the interaction between incident light and thermally excited acoustic phonons in the fiber, generating a frequency-shifted Brillouin backscatter signal. The Brillouin frequency shift (BFS) exhibits linear dependence on temperature and strain, with typical sensitivities of approximately 1 MHz/◦C and 0.05 MHz/μ𝜀 in standard single-mode fibers [26]. Raman-based DTS, by contrast, is intrinsically strain-insensitive and derives temperature information from the anti-Stokes to Stokes intensity ratio. However, Raman sensing suffers from lower spatial resolution and limited sensitivity, especially under conditions with high attenuation due to environmental factors. Considering that the Raman effect is strain-insensitive and Brillouin scattering exhibits significantly lower sensitivity and limited spatial resolution, this work is based on Rayleigh backscattering. OFDR is conceptually similar to interferometric point sensors such as FBGs, as both detect spectral shifts induced by changes in optical path length or refractive index, linking the principles of point and distributed temperature sensing. 2.2. Temperature–strain cross-sensitivity and environmental interference In optical fiber sensors, both temperature and axial strain influence the refractive index and physical length of the fiber, resulting in spectral shifts of the backscattered or reflected signal. This leads to cross-sensitivity, where the measured signal contains influence from both temperature and strain. The spectral shift (𝛥𝜆) in Rayleigh-based systems can be approximated using a model similar to that of a weak random FBG response, and its relation to external perturbations can be described as follows [27]: 𝛥𝜆 𝜆=𝐾𝑇⋅𝛥𝑇 +𝐾𝜀⋅𝜀 . Here, 𝐾𝑇 is the temperature coefficient, accounting for both the thermal expansion of the fiber and the thermo-optic effect, and 𝐾𝜀 is the strain coefficient, describing the photoelastic response of the silica matrix. As Rayleigh scattering is sensitive to both temperature and strain, practical implementations often require compensation strategies such as: (1) physical strain isolation [28], (2) dual-parameter algorithms including machine learning [29], or (3) hybrid sensing using complementary scattering mechanisms such as Raman-assisted temperature compensation [30,31]. FBG sensors, operating on a similar principle to OFDR, measure temperature and/or strain by detecting shifts in the Bragg wavelength: 𝜆𝐵= 2 𝑛eff 𝛬, where 𝑛eff is the effective refractive index of the fiber core and 𝛬 is the grating period. Temperature changes influence both 𝑛eff and 𝛬, producing a measurable wavelength shift. Like OFDR, FBGs detect spectral shifts induced by changes in optical path length or refractive index, highlighting the conceptual similarity between point and distributed sensing approaches. Nevertheless, in practical sensing scenarios, temperature variations can induce internal strain within the fiber due to mismatched thermal expansion between the fiber and its coating. This thermally induced strain is particularly relevant in polymer-coated fibers, where compliant mechanical properties and humidity-dependent behavior result in non-negligible strain transfer to the core. Among polymer coatings, polyimide and acrylate are two commonly used polymer coatings for optical fibers. Polyimide offers excellent thermal stability but absorbs moisture in humidity conditions, causing swelling and mechanical softening. Acrylate coatings, though widely used in telecom applications, absorb even more water and exhibit significant softening at elevated humidity, mainly compromising their mechanical stability. In contrast, metal-coated fibers provide a more rigid mechanical interface and are impermeable to water. These coatings maintain consistent strain transfer efficiency and are largely unaffected by environmental humidity. However, they may still suffer from long-term degradation, such as surface oxidation or delamination in chemically aggressive environments. Overall, the choice of coating material significantly influences the temperature sensitivity of the sensing system, particularly under dynamic environmental conditions. These effects are especially critical for high-precision distributed sensing applications, where humiditydependent coating behavior introduces subtle yet measurable variations in sensitivity and uncertainty. The influence of water uptake on coating mechanics and interfacial strain transfer dynamics is further examined in the next section. 2.3. Humidity-induced effects on fiber and coating Ambient humidity can influence the performance of fiber-optic temperature sensors through a variety of indirect mechanisms, primarily affecting the silica fiber core, the coating material, and the core– coating interface. These effects are especially relevant in long-term deployments under high-temperature and high-humidity conditions. In the silica matrix, water ingress leads to hydrolytic breakdown of Si–O–Si bonds and the formation of hydroxyl (OH−) groups, which introduce characteristic absorption peaks at wavelengths around 950, 1244, and 1383 nm [32]. As a result, the optical attenuation increases, which reduces the signal-to-noise ratio (SNR) of the sensing system and undermines its performance. Over time, continued humidity-induced degradation may further lead to microcracks, surface roughening, refractive index changes, and stress corrosion, thereby compromising fiber’s tensile strength and long-term reliability [33]. In addition to its chemical effects on the silica core, ambient humidity also significantly influences the mechanical properties of fiber coatings, which are critical for transferring strain from the environment to the fiber core. This effect is particularly evident in polymer coating such as polyimide. Despite their thermal robustness, these coatings are hygroscopic and tend to swell and soften upon water absorption. The resulting reduction in Young’s modulus and changes in thermal expansion compromise the coating’s ability to efficiently transfer strain to the fiber core. Consequently, optical fiber sensors that are strainsensitive, including but not limited to Rayleigh-based systems, may exhibit decreased temperature sensitivity and increased measurement uncertainty under humid conditions. Although short-term humidity exposure may cause reversible changes [34], repeated humidity cycling or long-term exposure can lead to cumulative calibration deviations and signal drift. Among polymer coatings, acrylate coatings are widely used in telecommunications but provide only limited thermal and environmental stability. While acceptable for low-precision applications, these Measurement 261 (2026) 119905 2 K. Wang et al. Table 1 Geometric and coating specifications of the tested optical fibers. Fiber Core/Cladding (μm) Coating Type Coating Diameter (μm) Coating Thickness (μm) Fiber 1 9/125 Polyimide 155 ± 5∼15 Fiber 2 6.4/125 Polyimide 155 ± 5∼15 Fiber 3 6.4/80 Polyimide 102 ± 5∼11 Fiber 4 4.2/125 Polyimide 155 ± 5∼15 Fiber Au 9/125 Gold 155 ± 10 ∼15 Fiber Cu 9/125 Copper 160 ± 8∼17.5 SMF-28 9/125 Acrylate 250 ± 0.7 ∼62.5 characteristics render them unsuitable for high-resolution distributed sensing in humid or thermally unstable environments. These vulnerabilities underscore the limitations of polymer-coated fibers in long-term high-precision sensing applications under varying humidity conditions. To address these limitations, alternative coating strategies, particularly the use of metal coatings, have been explored for their potential to enhance environmental robustness. Metal-coated fibers, such as those with gold or copper layers, have been developed to offer enhanced resistance to environmental influences, particularly humidity and temperature fluctuations. Their metallic coatings are impermeable to water and maintain mechanical stability under humid or thermally aggressive conditions, thus preserving consistent strain transfer and stabilizing temperature sensitivity. In addition, unlike polymer coatings that are limited to moderate temperature ranges, metal coatings enable operation at substantially higher temperatures, making them attractive for harsh-environment applications. However, long-term use may still lead to surface oxidation or delamination. Additionally, thin water films adsorbed onto metallic surfaces can increase surface conductivity, potentially introducing noise in densely packed optoelectronic assemblies [35]. In summary, humidity-induced mechanisms acting on both the fiber and its coating can significantly alter temperature sensitivity, system stability, and measurement uncertainty. A clear understanding of these physical interactions is crucial for selecting appropriate fiber coatings and for interpreting distributed sensing data in high-humidity environments. 3. Experimental methods 3.1. Fiber types and coating configurations To systematically investigate the influence of environmental humidity on temperature sensing performance, we selected a set of optical fibers that represent the most commonly used coating types and structural configurations in fiber-optic sensing applications. These fibers exhibit distinct mechanical behaviors and humidity responses, enabling comparative analysis of sensitivity, stability, and measurement uncertainty under controlled environmental conditions. The tested samples include four polyimide-coated single-mode fibers (Fibers 1–4), two metal-coated fibers (gold and copper, denoted as Fiber Au and Fiber Cu), and one standard acrylate-coated fiber (SMF-28), as summarized in Table 1. The polyimide-coated fibers differ in their core/cladding diameters and coating thicknesses. While polyimide and acrylate-coated fibers have been widely studied in fiber-optic sensing, the temperature sensing performance of metal-coated fibers under varying humidity conditions remains largely unexplored. Their inclusion in this study enables a direct evaluation of their robustness and comparability against polymer-coated fibers in humid environments. The polyimide-coated fibers were selected due to their working temperature up to 300 ◦C, which may induce changes in coating stiffness and strain transfer upon water absorption. Conversely, the metal-coated fibers are expected to be water-impermeable and mechanically stable under varying humidity. The acrylate-coated SMF-28, commonly used in telecommunications, serves as a comparison. Although not optimized for harsh conditions, its response provides a useful baseline against Fig. 1. Schematic diagram of the experimental setup for combined humidity and temperature characterization of fiber-optic sensors. Fibers were loosely coiled on a stainless-steel mesh tray inside a climate chamber. A Rayleighbased interrogator was connected externally for distributed sensing. specialized coatings. By studying these fiber types under controlled humidity and temperature conditions, this work aims to quantify and explain the humidity effects on temperature sensing. 3.2. Measurement systems and experimental setup Distributed temperature measurements were conducted using a Rayleigh-based OFDR. This system provided a spatial resolution of approximately 10 μm, with a wavelength scanning range of 1525– 1610 nm and a sampling interval of 2 mm. The high resolution enabled precise localization of spectral shifts along the fiber length, allowing accurate analysis of temperature sensitivity and measurement uncertainty under varying environmental conditions. All environmental conditions were regulated using a programmable climate chamber capable of independently controlling temperature and relative humidity (RH). Experiments were performed at RH levels of 30%, 50%, 70%, and 90%, with each RH condition tested at three temperatures: 20 ◦C, 40 ◦C, and 60 ◦C. The four RH points were chosen to represent typical lowmedium-, and high-humidity conditions commonly used in optical fiber humidity experiments. A humidity range between 30–90% RH is frequently adopted in optical fiber humidity sensing studies [36] and has also been used in investigation of acrylateand polyimide-coated fibers [37]. This choice ensures clear differentiation between humidity states while maintaining stable control with the climate chamber. Each temperature level was held constant for two hours at a fixed RH level to ensure thermal and humidity stabilization before data acquisition. The four polyimide-coated fibers (each 2.5 m) were fusion-spliced in series and measured using the Rayleigh system. The two metal-coated fibers (each 2 m) were also spliced together and measured under identical environmental conditions using the same setup. All fusion splices were protected using splice protection sleeves, and the sleeve ends were additionally sealed to prevent humidity ingress. The entire sensing fiber was placed inside the climatic chamber to ensure that the humidity exposure corresponded only to the coated fiber surface. The acrylatecoated SMF-28 was included and measured alongside polyimide-coated samples in the same configuration. Measurement 261 (2026) 119905 3 K. Wang et al. (a) Spectral shift at 30% RH (Rayleigh) (b) Spectral shift at 90% RH (Rayleigh) Fig. 2. Distributed Rayleigh spectral shift profiles along Fiber 1 at three temperatures (20 ◦C, 40 ◦C, and 60 ◦C), under low (30%) and high (90%) relative humidity. To ensure uniform environmental exposure, all fibers were loosely coiled around a vertical stainless-steel mesh tray placed inside the chamber. Fig. 1 shows a schematic of the experimental configuration. The metal-coated fibers were tested separately under identical environmental conditions using the same setup. 4. Experimental results and discussion 4.1. Rayleigh-based temperature sensitivity under humidity To evaluate how environmental humidity influences Rayleigh-based distributed temperature sensing, we first examine the spectral shift profiles obtained from Fiber 1 under different conditions. Fig. 2 shows the Rayleigh spectral shift along the fiber length at three temperatures (20 ◦C, 40 ◦C, and 60 ◦C) for low (30%) and high (90%) RH levels. The overall spectral shift arises from both the thermo-optic responses of silica and from humidity-induced strain transfer caused by swelling of the hygroscopic coating, particularly for polyimide-coated fibers. Similar coupling between thermal and hygroscopic responses has been analyzed in previous studies on polyimide-coated fibers. Humidity absorption in the coating can induce a small strain, which adds to the intrinsic thermo-optic response and leads to the overall wavelength shift observed here. The approximately 100 pm spectral shift between 30% and 90% RH lies within the range reported in earlier work [20, 38,39]. No abrupt variations or local anomalies are observed, indicating uniform strain transfer and stable coating behavior. However, as raw spectral profiles do not clearly show sensitivity changes, further quantitative analysis is required. Fig. 3 presents the Rayleigh-based temperature sensitivity of four polyimide-coated fibers and one acrylate-coated SMF-28, derived from linear fits of spectral shift data at 20 ◦C, 40 ◦C, and 60 ◦C under RH levels from 30% to 90%. The slopes of each fit represent the temperature sensitivity in pm/◦C. All polyimide-coated fibers exhibited a consistent trend: the temperature sensitivity slightly decreased with increasing humidity. For instance, Fiber 1 showed a reduction from 11.05 to 11.01 pm/◦C, and Fiber 4 decreased from 10.73 to 10.62 pm/◦C, between 30% and 90% RH. Although these changes are modest (typically within 1%–2%), the trend is consistent across fiber geometries. The observed reduction in sensitivity is attributed to the humidity-induced softening of the polymer coating, which reduces its Young’s modulus and weakens the mechanical coupling to the fiber. This softening effect is enhanced at elevated temperatures, leading to a more pronounced decrease in strain transfer efficiency and, consequently, a greater reduction in temperature sensitivity. It should be noted that the small 1%–2% variation in temperature sensitivity refers Fig. 3. Rayleigh-based temperature sensitivity of four polyimide-coated fibers and one acrylate-coated SMF-28 as a function of relative humidity. For each humidity level (30%, 50%, 70%, 90%), the sensitivity was obtained from a linear fit of the spectral shift measured at 20 ◦C, 40 ◦C, and 60 ◦C. Each point therefore represents the humidity-dependent temperature sensitivity. to the polyimide-coated fibers, whereas the acrylate-coated SMF-28 shows a larger decrease (from 11.31 pm/◦C to 10.65 pm/◦C, ≈5.8%). These observations are attributed to humidity-induced mechanical effects in polymer coatings. Increased moisture softens the coating and reduces its Young’s modulus, weakening the mechanical coupling between the coating and the cladding, and thereby diminishing the efficiency of strain transfer into the fiber core. The reduced transfer of thermally induced strain results in a lower Rayleigh shift per degree of temperature change. While polyimide coatings offer improved thermal robustness, their hygroscopic nature still leads to measurable sensitivity degradation under high humidity. In contrast, acrylate coatings absorb even more water and thus suffer from larger performance losses. Although the observed changes appear modest, they can introduce non-negligible cumulative errors in high-precision applications. To better quantify these effects, we calculated the resulting temperature deviation (𝛥𝑇 ) caused by sensitivity reduction between 30% and 90% RH. For each fiber, the Rayleigh spectral shift between 20 ◦C and 60 ◦C was obtained from repeated measurements. The inferred temperature was then calculated by dividing the spectral shift by corresponding sensitivity at each RH level. The difference in inferred temperature reflects the error caused by humidity-induced sensitivity drift. Fig. 4 shows the calculated 𝛥𝑇 for each fiber, which represents humidity-induced calibration errors obtained from spectral-shift data Measurement 261 (2026) 119905 4 K. Wang et al. Fig. 4. Humidity-induced temperature error (𝛥𝑇 ) resulting from Rayleigh sensitivity degradation for each fiber. 𝛥𝑇 was calculated based on spectral shift measurements at 20 ◦C and 60 ◦C under 30% and 90% RH, using the respective sensitivities at each humidity level. at 20 ◦C and 60 ◦C under 30% and 90% RH, rather than a direct temperature dependence. Fiber 2 shows negligible deviation (0.0002 ◦C), while Fiber 1 and Fiber 4 exhibit modest errors below 0.5 ◦C. Fiber 3 and SMF-28 show the most significant deviations, with SMF-28 reaching 1.1702 ◦C. These results demonstrate that even subtle sensitivity changes can lead to non-negligible temperature errors in high-precision applications, especially if uncompensated. 4.2. Temperature measurement uncertainty To evaluate the reliability of Rayleigh-based distributed temperature sensing under varying environmental conditions, we calculated the temperature uncertainty using Type A evaluation, i.e., based on the statistical standard deviation (STD) of repeated measurements along the fiber. Specifically, uncertainty was obtained by dividing the STD of Rayleigh spectral shift (calculated over the full fiber length for each condition) by the corresponding temperature sensitivity at each humidity level. The resulting values, expressed in ◦C, represent the minimum resolvable temperature deviation under each condition and serve as a direct indicator of sensing performance. In the OFDR system, each scan records the full Rayleigh backscatter profile along the entire fiber at that moment. Under steady temperature and humidity, the spatial STD of the Rayleigh spectral shift thus represents the random measurement noise along the fiber. Several consecutive scans were taken under the same conditions to confirm repeatability, and no systematic drift was observed between scans. Small environmental fluctuations during acquisition had no measurable influence on the results, confirming that temporal variations were negligible compared with spatial variation. Accordingly, the spatial STD provides a reliable estimate of the Type-A uncertainty along the fiber under steady-state conditions. Fig. 5 presents the Rayleigh-based temperature uncertainty of four polyimide-coated and one acrylate-coated SMF-28 under RH levels from 30% to 90%, evaluated at 20 ◦C, 40 ◦C, and 60 ◦C. In general, all the fibers maintain low uncertainty across all conditions. At low temperature (20 ◦C), high humidity (90% RH) degrades stability for all fibers, most strongly for Fiber 4. At intermediate temperature (40 ◦C), uncertainties are low and may even decrease at high RH levels for some fiber (e.g., Fiber 1). At elevated temperature (60 ◦C), humidity effects are small and fiber-dependent. These observations indicate that humidity–temperature coupling is temperature-dependent and non-monotonic, and that coating formulation/thickness together with fiber geometry may modulate the response. Beyond RH and temperature, spatial non-uniformities in coating thickness or modulus (e.g., local softening, microcracks, or partial debonding) can alter strain transfer and increase the along-fiber STD used in our Type-A evaluation; this may contribute to the fiber-dependent trends in Fig. 5. 4.3. Performance of metal-coated fibers In contrast to polymer coatings, metal-coated fibers offer fundamentally different mechanical and moisture barrier properties. To evaluate the environmental robustness of alternative coating materials, we investigated the performance of two metal-coated single-mode fibers, one copper-coated and one gold-coated, under varying humidity conditions. These fibers are specifically designed for deployment in chemically or thermally harsh environments where mechanical stability is critical. Distributed temperature measurements were performed using Rayleighbased sensing at three temperatures (20 ◦C, 40 ◦C, 60 ◦C) and four relative humidity (RH) levels (30%, 50%, 70%, 90%). 4.3.1. Humidity-robust temperature response of metal-coated fibers Fig. 6 presents the Rayleigh temperature sensitivity of the metalcoated fibers as a function of RH. The copper-coated fiber exhibits a higher baseline sensitivity of about 17 pm/◦C, whereas the goldcoated fiber remains around 14.4 pm/◦C. Both fibers maintained nearly constant sensitivity across all humidity levels, with deviations below 0.2 pm/◦C. These results confirm that metal-coated fibers are largely immune to the humidity-dependent strain transfer degradation observed in polymer-coated fibers (see Section 4.1), due to the non-hygroscopic and rigid nature of the metallic coating. The Rayleigh sensitivities of both metal-coated fibers are significantly higher than that of the polyimide-coated fibers and the SMF-28, which all exhibited sensitivities around 11 pm/◦C. The higher temperature sensitivity of the metal-coated fibers can be attributed to the much larger thermal expansion coefficients and higher Young’s moduli of copper and gold compared with fused silica and polymer coatings. The metallic coatings expand more strongly with temperature and, due to their high stiffness, transfer thermally induced strain more effectively to the fiber core, resulting in the observed higher temperature sensitivity [40,41]. 4.3.2. Temperature uncertainty analysis To further evaluate humidity resilience, Rayleigh-based temperature uncertainty was calculated for the two metal-coated fibers using the Type A method described in Section 4.2. Fig. 7 shows the results at 20 ◦C, 40 ◦C, and 60 ◦C. In all cases and across all tested temperatures, uncertainties remain low, within approximately 0.07 ◦C. At 20 ◦C, both fibers show a slight increase towards high RH, though the magnitude remains small. At 40 ◦C, uncertainties are nearly constant across all humidity levels. At 60 ◦C, both fibers exhibit a shallow variation with RH, but without a consistent trend between them. Such minor variations could arise from factors such as small nonuniformities in coating thickness or modulus, residual stress distribution from the coating process, or localized surface effects. While their magnitude is negligible for practical applications, they may slightly influence the spatial STD used in the Type-A evaluation. Overall, the two metal-coated fibers exhibit comparable performance, confirming that humidity has a negligible influence on their measurement stability. 5. Humidity-induced measurement challenges and pathways forward Our findings clearly demonstrate that humidity has a non-negligible impact on temperature-sensing performance, particularly for fibers with polymer coatings. This effect becomes more pronounced at elevated temperatures, where measurement uncertainty increases with rising humidity. The degradation in sensing performance is primarily attributed to moisture absorption within the coating, which induces micro-strain in the fiber core and slightly modifies the local Measurement 261 (2026) 119905 5 K. Wang et al. (a) Temperature uncertainty at 20 ◦C (b) Temperature uncertainty at 40 ◦C (c) Temperature uncertainty at 60 ◦C Fig. 5. Rayleigh-based temperature uncertainty of four polyimide-coated fibers and one SMF-28 at (a) 20 ◦C, (b) 40 ◦C, and (c) 60 ◦C, as a function of relative humidity. Fig. 6. Rayleigh-based temperature sensitivity of copperand gold-coated optical fibers as a function of relative humidity (30%–90% RH). Each point represents the slope from linear fitting of Rayleigh spectral shift measurements versus temperature. refractive-index distribution, thereby affecting both sensitivity and stability. An effective strategy for mitigating these effects lies in the selection of appropriate coating materials. Comparative analysis of the tested fibers showed that metal-coated fibers, especially those with copper or gold layers, exhibited the most stable temperature sensitivity and the lowest measurement uncertainty across all humidity levels. The metallic coating serves as an impermeable barrier against water diffusion, thereby maintaining a stable thermal response of the fiber core even under high-humidity conditions. In contrast, polymer coatings such as polyimide are hygroscopic rather than strongly permeable. Water absorption within the polymer coating can slightly modify strain transfer to the fiber core but does not lead to water ingress within the timescale of this study. Previous investigations have reported that hydrolytic degradation or OH− diffusion through polyimide becomes significant only after weeks to years of continuous high-humidity exposure [42, 43]. Regarding metallic coatings, copper is known to oxidize under humid or elevated-temperature conditions, forming Cu2O/CuO layers whose growth rate increases exponentially with temperature [44]. Gold coatings, in contrast, are chemically inert and provide superior resistance to both corrosion and water penetration. Experiments have shown that gold-coated silica fibers can remain intact for several days in molten-salt and steam environments, where uncoated fibers fail within only a few hours [45]. Although quantitative comparisons among polymer-, copper-, and gold-coated fibers under controlled humidity are still limited, available data consistently indicate the trend Au > Cu > polyimide/acrylate in environmental stability. The present results add complementary experimental evidence under moderate temperature and relative-humidity conditions, demonstrating that humidity effects in polymer coatings mainly originate from hygroscopic strain transfer, while metal-coated fibers maintain stable sensitivity within the tested time scale. While no degradation was observed within the duration of the present measurements, long-term exposure may eventually lead to oxidation or delamination depending on the coating material. Future work will therefore include extended humidity-cycling tests to assess the long-term durability of metallic coatings and to quantify their performance under repeated environmental conditions. In addition to coating selection, system design also plays an essential role in measurement stability. While Rayleigh-based OFDR offers Measurement 261 (2026) 119905 6 K. Wang et al. (a) Temperature uncertainty at 20 ◦C (b) Temperature uncertainty at 40 ◦C (c) Temperature uncertainty at 60 ◦C Fig. 7. Rayleigh-based temperature uncertainty for copperand gold-coated fibers at (a) 20 ◦C, (b) 40 ◦C, and (c) 60 ◦C, under varying humidity levels. high sensitivity to micro-strain, this characteristic also makes it more vulnerable to effects induced by humidity. Techniques such as incorporating reference system, increasing spatial averaging, or selecting fibers with inherently lower strain sensitivity may help suppress these humidity-related effects. Although this study focuses specifically on Rayleigh-based DTS, the humidity-induced effects observed here are broadly relevant to other fiber-optic sensing configurations. Sensing mechanisms that depend on strain or refractive index, such as fiber Bragg gratings, Fabry–Pérot interferometers, and microstructured fibers, are likely to exhibit similar instability when coated with hygroscopic materials. These insights underscore the importance of considering environmental cross-sensitivity in both sensor calibration and system design. Future work could explore the development of advanced fiber architectures, humidity-insensitive coatings, or integrated humidity– temperature co-measurement schemes for in-situ compensation. Ultimately, the methodologies and findings presented here provide a foundation for improving the long-term stability and accuracy of fiberoptic sensing technologies across diverse environmental conditions, not only for temperature monitoring but for broader multiparameter sensing applications. 6. Conclusion This study systematically evaluated the impact of ambient humidity on the temperature sensitivity and measurement stability of optical fibers with different coatings, using a Rayleigh-based distributed sensing technique. Experimental results showed that polyimide-coated fibers exhibit a slight but consistent reduction in temperature sensitivity as relative humidity increases, with the magnitude of the effect depending on temperature and fiber geometry. This degradation is attributed to water absorption in the coating, which softens the material and reduces strain transfer efficiency to the fiber core. Although the sensitivity changes are typically within 1%–2%, they can introduce measurable temperature deviations, exceeding 1 ◦C in the most affected case. In contrast, metal-coated fibers (copper and gold) demonstrated excellent stability across all humidity and temperature conditions. Their Rayleigh-based sensitivity remained nearly constant (variation ≤ 0.2 pm/◦C), and the resulting temperature uncertainty stayed low across 30%–90% RH, confirming the effectiveness of metallic coatings in eliminating humidity-induced mechanical effects. Meanwhile, the acrylate-coated SMF-28 showed larger variations in both sensitivity and uncertainty, consistent with its known hygroscopic behavior. These comparisons highlight the critical role of coating permeability and mechanical coupling in maintaining sensing accuracy under varying environmental conditions. Overall, these findings underscore the importance of accounting for humidity effects in fiber-optic temperature sensing, particularly in Rayleigh-based systems using polymer-coated fibers. They also emphasize coating material selection as a key design parameter for ensuring long-term performance in humid environments. Beyond practical guidance for fiber selection, this work offers broader insight into the coupled effects of humidity, coating mechanics, and strain transfer dynamics. The conclusions are relevant not only to DTS systems but also to other humidity-sensitive fiber-optic sensing platforms, such as FBGs, Fabry–Pérot interferometers, or microstructured fibers. By combining high-resolution Rayleigh-based DTS, uncertainty analysis, quantitative error estimation, and comparative coating evaluation, this study provides a framework for designing humidity-resilient, high-accuracy fiber-optic sensors for diverse environmental monitoring applications. Measurement 261 (2026) 119905 7 K. Wang et al. CRediT authorship contribution statement Kun Wang: Writing – original draft, Validation, Methodology, Investigation, Formal analysis, Conceptualization. Marcus Schukar: Software, Methodology, Investigation. Konstantin Hicke: Writing – review & editing, Supervision. Xin Lu: Writing – review & editing, Validation, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments The project 22IEM07 INFOTherm has received funding from the European Partnership on Metrology, co-financed by the European Union’s Horizon Europe Research and Innovation Programme and by the Participating States. Data availability Data will be made available on request. References [1] A. Ukil, e. Braendle, P. Krippner, Distributed temperature sensing: Review of technology and applications, IEEE Sensors J. 12 (5) (2012) 885–892, http: //dx.doi.org/10.1109/JSEN.2011.2162060. [2] H.-N. Li, D.-S. Li, G.-B. Song, Recent applications of fiber optic sensors to health monitoring in civil engineering, Eng. Struct. 26 (11) (2004) 1647–1657, http: //dx.doi.org/10.1016/j.engstruct.2004.05.018, https://www.sciencedirect.com/ science/article/pii/S014102960400183X. [3] Y. Liu, Y. Huang, Y. Bao, Machine learning-empowered automatic analysis of distributed fiber optic sensor data for monitoring coincident corrosion and cracks in pipelines, Measurement 247 (2025) 116805, http://dx.doi.org/10.1016/ j.measurement.2025.116805, https://www.sciencedirect.com/science/article/pii/ S0263224125001642. [4] S. Zhang, B. Liu, J. He, Pipeline deformation monitoring using distributed fiber optical sensor, Measurement 133 (2019) 208–213, http://dx.doi.org/10.1016/ j.measurement.2018.10.021, https://www.sciencedirect.com/science/article/pii/ S0263224118309400. [5] H.C. Hyer, D.R. Giuliano, C.M. Petrie, Toward local core outlet temperature monitoring in gas-cooled nuclear reactors using distributed fiber-optic temperature sensors, Appl. Therm. Eng. 230 (2023) 120847, http://dx.doi.org/10.1016/ j.applthermaleng.2023.120847, https://www.sciencedirect.com/science/article/ pii/S1359431123008761. [6] A. Minardo, G. Persichetti, G. Testa, L. Zeni, R. Bernini, Long term structural health monitoring by brillouin fibre-optic sensing: a real case, J. Geophys. Eng. 9 (4) (2012) S64–S69, http://dx.doi.org/10.1088/1742-2132/9/4/S64, arXiv:https: //academic.oup.com/jge/article-pdf/9/4/S64/26800634/jge12_4_s64.pdf. [7] P. Lu, N. Lalam, M. Badar, B. Liu, B.T. Chorpening, M.P. Buric, P.R. Ohodnicki, Distributed optical fiber sensing: Review and perspective, Appl. Phys. Rev. 6 (4) (2019) 041302, http://dx.doi.org/10.1063/1.5113955, arXiv:https://pubs.aip.org/aip/apr/article-pdf/doi/10.1063/1.5113955/ 14575982/041302_1_online.pdf. [8] I. Ashry, Y. Mao, B. Wang, F. Hveding, A.Y. Bukhamsin, T.K. Ng, B.S. Ooi, A review of distributed fiber–optic sensing in the oil and gas industry, J. Lightwave Technol. 40 (5) (2022) 1407–1431, http://dx.doi.org/10.1109/JLT. 2021.3135653. [9] M. Johnson Singh, S. Choudhary, W.-B. Chen, P.-C. Wu, M. Kumar Goyal, A. Rajput, L. Borana, Applications of fibre bragg grating sensors for monitoring geotechnical structures: A comprehensive review, Measurement 218 (2023) 113171, http://dx.doi.org/10.1016/j.measurement.2023.113171, https://www. sciencedirect.com/science/article/pii/S0263224123007352. [10] J. Zhao, T. Ma, F. Zhang, Distributed optical fiber sensors for pavement engineering: A-state-of-art review, Measurement 246 (2025) 116732, http://dx.doi.org/10.1016/j.measurement.2025.116732, https://www. sciencedirect.com/science/article/pii/S0263224125000910. [11] J. Li, M. Zhang, Physics and applications of raman distributed optical fiber sensing, Light, Sci. Appl. 11 (2022) https://api.semanticscholar.org/CorpusID: 248557566. [12] S. Xu, F. Xing, H. Yang, G. Li, Multiple compensation and linear fitting demodulation for raman distributed temperature sensors, IEEE Sensors J. 24 (23) (2024) 39002–39010, http://dx.doi.org/10.1109/JSEN.2024.3474972. [13] L. Qiu, Z. Zhu, T. Li, D. Zhou, Y. Dong, High-sensitivity distributed temperature sensor based on brillouin scattering with double-coated single-mode fibers, IEEE Sensors J. 21 (5) (2021) 6209–6216, http://dx.doi.org/10.1109/JSEN.2020. 3043251. [14] S. Ochi, K. Kikuchi, S. Tsurugai, H. Lee, Y. Mizuno, High-resolution distributed temperature sensing along polymer optical fiber using brillouin optical correlation-domain reflectometry, Opt. Fiber Technol. 90 (2025) 104144, http://dx.doi.org/10.1016/j.yofte.2025.104144, https://www.sciencedirect.com/ science/article/pii/S1068520025000197. [15] C. Chen, L. Chen, X. Bao, Distributed temperature profile in hydrogen flame measured by telecom fiber and its durability under flame by ofdr, Opt. Express 30 (11) (2022) 19390–19401, http://dx.doi.org/10.1364/OE.455640. [16] Z. Sun, X. Xiao, W. Zhao, K. Ai, Y. Lv, H. Liu, Q. Sun, Z. Yan, High-precision distributed temperature sensor based on phase-shifted fbgs interrogated by ofdr, J. Lightwave Technol. 43 (2025) 2357–2362, https://api.semanticscholar.org/ CorpusID:273682198. [17] K. Wang, Z. Ma, J. Zhang, W. Shuai, B. Sun, M. Zhang, High spatial resolution sensing of temperature and strain without cross sensitivity based on chaotic brillouin dynamic grating, IEEE Sensors J. 23 (2023) 9317–9322, https://api. semanticscholar.org/CorpusID:257745279. [18] P. Xu, Y. Peng, K. Wen, Y. Sun, X. Dong, J. Yang, Y. Qin, Bending-lossresistant distributed temperature and strain discriminative brillouin sensor based on 98 mol% germania-doped few-mode fiber, J. Lightwave Technol. 41 (2023) 4854–4861, https://api.semanticscholar.org/CorpusID:256982023. [19] P. Stajanca, K. Hicke, K. Krebber, Distributed fiberoptic sensor for simultaneous humidity and temperature monitoring based on polyimide-coated optical fibers, Sensors 19 (23) (2019) http://dx.doi.org/10.3390/s19235279, https://www. mdpi.com/1424-8220/19/23/5279. [20] X. Lu, M. Schukar, Humidity response of optical fibres with hygroscopic coatings and its temperature dependence, J. Phys.: Photonics 7 (2) (2025) 025026, http://dx.doi.org/10.1088/2515-7647/adc85f. [21] T.F.P. Neves, L. Scherino, R. Bernard, M. Bouet, A. Pastre, R.M. aes, S. MartinLopez, H.F. Martins, P. Petagna, L. Thévenaz, Humidity-insensitive optical fibers for distributed sensing applications, Appl. Opt. 62 (15) (2023) 4017–4029, http://dx.doi.org/10.1364/AO.487264. [22] C. Karapanagiotis, K. Hicke, A. Wosniok, K. Krebber, Distributed humidity fiberoptic sensor based on bofda using a simple machine learning approach, Opt. Express 30 (8) (2022) 12484–12494, http://dx.doi.org/10.1364/OE.453906. [23] M. Mądry, B. Szczupak, M. Śmigielski, B. Matysiak, Simultaneous temperature and relative humidity measurement using machine learning in rayleigh-based optical frequency domain reflectometry, Sensors 24 (24) (2024) http://dx.doi. org/10.3390/s24247913, https://www.mdpi.com/1424-8220/24/24/7913. [24] M. Froggatt, J. Moore, High-spatial-resolution distributed strain measurement in optical fiber with rayleigh scatter, Appl. Opt. 37 (10) (1998) 1735–1740, http://dx.doi.org/10.1364/AO.37.001735. [25] B.J. Soller, D.K. Gifford, M.S. Wolfe, M.E. Froggatt, High resolution optical frequency domain reflectometry for characterization of components and assemblies, Opt. Express 13 (2) (2005) 666–674, http://dx.doi.org/10.1364/OPEX.13. 000666. [26] M. Niklès, L. Thévenaz, P.A. Robert, Simple distributed fiber sensor based on brillouin gain spectrum analysis., Opt. Lett. 21 (10) (1996) 758, https://api. semanticscholar.org/CorpusID:14277078. [27] S.T. Kreger, A.K. Sang, D.K. Gifford, M.E. Froggatt, Distributed strain and temperature sensing in plastic optical fiber using Rayleigh scatter, in: E. Udd, H.H. Du, A. Wang (Eds.), Fiber Optic Sensors and Applications VI, in: International Society for Optics and Photonics, vol. 7316, SPIE, 2009, 73160A., http://dx.doi.org/10.1117/12.821353. [28] J. Han, T. Yang, X. Wang, Advanced fiber optic sensing for cryogenic simultaneous temperature and strain measurement, IEEE Trans. Instrum. Meas. 73 (2024) 1–14, https://api.semanticscholar.org/CorpusID:271819447. [29] H. Gemeinhardt, J. Sharma, Machine-learning-assisted leak detection using distributed temperature and acoustic sensors, IEEE Sensors J. 24 (2024) 1520–1531, https://api.semanticscholar.org/CorpusID:265766800. [30] A. Coscetta, E. Catalano, E. Cerri, N. Cennamo, L. Zeni, A. Minardo, Hybrid brillouin/rayleigh sensor for multiparameter measurements in optical fibers, Opt. Express 29 (15) (2021) 24025–24031, http://dx.doi.org/10.1364/OE.426427. [31] Y. Mizuno, K. Nakazawa, H. Javid, K. Noda, K. Nakamura, H. Lee, Fiberoptic temperature sensing using raman spectrum near rayleigh peak, Opt. Fiber Technol. (2024) https://api.semanticscholar.org/CorpusID:269233248. [32] O. Humbach, H. Fabian, U. Grzesik, U. Haken, W. Heitmann, Analysis of oh absorption bands in synthetic silica, J. Non-Cryst. Solids 203 (1996) 19–26, http://dx.doi.org/10.1016/0022-3093(96)00329-8, https://www. sciencedirect.com/science/article/pii/0022309396003298. optical and Electrical Propertias of Glasses. [33] A. Rose, T. Bruno, The observation of oh in annealed optical fiber, J. NonCryst. Solids 231 (3) (1998) 280–285, http://dx.doi.org/10.1016/S00223093(98)00676-0, https://www.sciencedirect.com/science/article/pii/ S0022309398006760. Measurement 261 (2026) 119905 8 K. Wang et al. [34] R. Janani, D. Majumder, A. Scrimshire, A. Stone, E. Wakelin, A. Jones, N. Wheeler, W. Brooks, P. Bingham, From acrylates to silicones: A review of common optical fibre coatings used for normal to harsh environments, Prog. Org. Coatings 180 (2023) 107557, http://dx.doi.org/10.1016/j. porgcoat.2023.107557, https://www.sciencedirect.com/science/article/pii/ S0300944023001534]. [35] S.N. Mustaffa, A.R.A. Rashid, T.S. Jin, P.S. Menon, M.Z.A. Razak, A.M. Markom, H. Haris, I. Saad, A.R. Muhammad, Metal oxide coated optical fiber for humidity sensing application: A review, IEEE Access 11 (2023) 126568–126600, http: //dx.doi.org/10.1109/ACCESS.2023.3331365. [36] X. Rao, L. Zhao, L. Xu, Y. Wang, K. Liu, Y. Wang, G.Y. Chen, T. Liu, Y. Wang, Review of optical humidity sensors, Sensors 21 (23) (2021) 8049. [37] T. Neves, R. Magalhães, L. Scherino, S. Martin-Lopez, H.F. Martins, P. Petagna, L. Thévenaz, Humidity Effect on Acrylate-and Polyimide-Coated Fibres for Distributed Sensing Applications, Optical Fiber Sensors, Optica Publishing Group, 2020, pp. T3–73. [38] A. Swanson, S. Raymond, S. Janssens, R. Breukers, M. Bhuiyan, J. Lovell-Smith, M. Waterland, Investigation of polyimide coated fibre bragg gratings for relative humidity sensing, Meas. Sci. Technol. 26 (12) (2015) 125101. [39] N.A. David, P.M. Wild, N. Djilali, Parametric study of a polymer-coated fibre-optic humidity sensor, Meas. Sci. Technol. 23 (3) (2012) 035103. [40] S.-W. Kim, Characteristics of strain transfer and the reflected spectrum of a metal-coated fiber bragg grating sensor, Opt. Lasers Eng. 96 (2017) 83–93, http: //dx.doi.org/10.1016/j.optlaseng.2017.04.012, https://www.sciencedirect.com/ science/article/pii/S0143816616304328. [41] X. Wang, X. Sun, Y. Hu, L. Zeng, Q. Liu, J. Duan, Highly-sensitive fiber bragg grating temperature sensors with metallic coatings, Optik 262 (2022) 169337, http://dx.doi.org/10.1016/j.ijleo.2022.169337, https://www.sciencedirect.com/ science/article/pii/S0030402622006635. [42] K. Sohma, N. Iwaguchi, T. Kawano, T. Fujii, Y. Koyano, Estimation of longterm change in physical property of optical fiber coating considering effect of humidity, SEI Tech. Rev. (81) (2015) 17. [43] A.A. Stolov, J.A. Wrubel, D.A. Simoff, R.J. Lago, Acrylate-based specialty optical fiber coatings for harsh environments, IWCS 2016 (2016) 27. [44] Y. Wan, X. Wang, H. Sun, Y. Li, K. Zhang, Y. Wu, Corrosion behavior of copper at elevated temperature, Int. J. Electrochem. Sci. 7 (9) (2012) 7902–7914. [45] A. Leong, S.D. Rountree, J. Zhang, Corrosion of silica-based optical fibers in various environments, Corros. Mater. Degrad. 4 (3) (2023) 445–465. Measurement 261 (2026) 119905 9