Robotic-based terahertz spectroscopy imaging for non-destructive monitoring of water loss-induced damage in timber materials under low temperature stress

. Timber materials are widely used in various industries, from construction to furniture manufacturing. One of the critical factors affecting the structural integrity and longevity of timber is water loss-induced damage, which often occurs under low-temperature stress conditions. This study explores the application of robotic-based terahertz spectroscopy imaging as a non-destructive and efficient technique for assessing the impact of water loss on timber materials under low-temperature stress. Moisture level evolution in various timber species was experimentally investigated by performing frequency modulated continuous wave terahertz spectroscopy and using a six-DOF (degree-of-freedom) robotic arm. The reflected signals were analyzed to compute the corresponding dielectric property reference value at different moisture content levels through signal processing in the time and frequency domains. The effect of fibre direction and moisture content on THz waves was performed and the validity of frequency modulated continuous wave terahertz spectroscopy for effective, was discussed.


Introduction
Timber has long been esteemed in industrial structures for its environmental sustainability, aesthetic versatility, and structural capabilities, making it a preferred material for a wide array of building applications.Monitoring the mechanisms behind the drying process is crucial due to its potential to cause various durability issues, with excessive water content leading to fungal decay, swelling, and a general compromise in structural strength [1][2][3][4].Dietsch et al. explore the critical relationship between wood moisture content and its physical and mechanical properties, emphasizing the importance of accurate moisture monitoring in preventing decay [5].Similarly, Schmidt and Riggio investigate moisture performance in cross-laminated timber (CLT) elements during construction, highlighting the risks posed by moisture exposure even before a structure is fully operational [6].
Several common methods like weight measurement and holographic interferometry have been employed to observe the drying process effects [7][8][9].However, weight measurement, gives only an average moisture content across the material's surface area, and holographic interferometry is time-intensive and demands a strictly controlled experimental environment [10,11].
To overcome these limitations, recent advancements in monitoring water-induced damage in timber materials have been significantly shaped by the development and application of non-destructive testing (NDT) methods [12][13][14].Terahertz spectroscopy (THz-Spec) imaging has recently emerged as a promising technique for investigating moisture content in non-transparent materials [15].Operating within the electromagnetic spectrum from 100 GHz to 30 THz, THz-Spec serves as a critical link between microwave and farinfrared technologies.Its unique electromagnetic characteristics facilitate the penetration of common non-metallic materials without necessitating a coupling medium, thereby providing high spatial resolution while ensuring safety using energy levels that are benign to human tissue [16][17][18].A distinctive feature of THz radiation is its pronounced sensitivity to water, attributed to the dielectric relaxation response exhibited by to intermolecular fluctuations of polar liquids such as water within the terahertz electromagnetic domain [19].This sensitivity enables a highly accurate assessment of moisture content, as the degree of THz radiation absorption by water directly correlates with the moisture content, enabling a detailed analysis of moisture levels and distribution [19].The unique properties of THz radiation enable the accurate observation of moisture dynamics within materials.Leveraging these strengths, the THz inspection method has been adeptly used to assess moisture levels and their spatial distribution across a diverse range of materials, including biological tissues, polymers, food products, and wood [20].This broad applicability underscores the versatility and precision of THz technology in capturing moisture-related phenomena.Despite its widespread use in measuring moisture concentration, the application of THz technology as a dedicated analytical instrument for investigating the process of moisture evolution during the drying process remains relatively untapped.
In this paper, THz-Spec imaging was deployed to investigate the moisture content dynamics within European spruce and maritime pine wood samples, a commonly used materials in diverse structural applications [21].By conducting a detailed analysis of THz signal fluctuations amidst the drying process of moisture-saturated wood samples, a clear correlation between the intensity of THz signals and the moisture content of the wood samples, is established.

Point by point-scan Frequency-Modulated Continuous-Wave wave terahertz imaging
The experimental setup for the Frequency-Modulated Continuous-Wave terahertz (FMCW THz) system is depicted in Figure 1.The FMCW THz imaging system (TeraScan Cobot Edition), is composed of an interchangeable Sub THz imaging head, and a 6-axis robotic arm.The THz heads consists of an emitter, a detector, and optical components including beams and lenses optics to adjust the working distance and the resolution of the system at a given frequency of 50 mm, 75 mm, and 100 mm.For this study, the FMCW THz imaging system operated within a frequency spectrum of 0.12-2.4THz, offering a lateral resolutions less than 2 mm, and 1 mm, respectively.The system ensured a signal-to-noise ratio (SNR) of 60 dB.Due to the broadband nature of the wavelengths, the focused THz beam's spot size varied between 50 to 100 mm.The robotic arm has a reach of at least 400 mm from its main axis.It allows for lateral imaging surfaces larger than 300 x 300 mm² when positioned optimally.Its positioning repeatability is around 100 microns.The minimum imaging step is 500 m, and the image pixel size ranged from 0.9 mm to 1.8 mm.To reduce atmospheric moisture absorption of the THz waves, the experiments were conducted at 5% relative humidity.In this paper, the FMCW THz imaging system was employed in reflection operational mode with a 90° angle of incidence.

Sample preparation and moisture absorption assessment procedure
Overall, two widely known types of coniferous wood were studied, which are European spruce (Picea abies) and maritime pine (Pinus pinaster).To consider wood variability, three samples were prepared.A total of six samples were tested.Each sample had a size of 100 × 200 mm with 20 mm thickness.All test specimens were prepared out of dried wood at 12% of relative moisture content.The three Spruce samples are noted S1, S2, S3 and the three Pine samples are noted P1, P2, P3.The samples were analyzed with FMCW THz imaging system in reflection configuration.
Wood is an anisotropic and fibrous material.It is therefore important to carry out measurements in all main directions to study the effect of fiber direction on the THz signal as well as on the dielectric constant of the material.Therefore, the THz signals in the wood samples were recorded under transverse and longitudinal configurations, as showed in Figure 1: • Transverse configuration: the measurement is carried out on the top surface on which the THz wave field displacement is perpendicular to the wood fibers and tangential to the growth rings (see Fig. 1(a)).• Longitudinal configuration: the measurement is carried out on the bottom surface on which the THz wave field displacement is parallel to the fibers (see Fig. 1(b)).
Measurements were made against the top surface of the wood samples sawn into the radial direction so that the growth rings were perpendicular to the THz wave source, as shown in Figure 1.During measurement, the samples were placed on a metal plate for easy identification of reflections on the underside (see Fig. 1).Subsequently, the index of refraction and the absorption coefficient as a function of frequency were evaluated for each type of wood with the TeraScan software.The results for the two different configurations in each sample were averaged.Since the measured THz data is influenced by the density and the water content of a sample, both metrics are determined.For this purpose, the bone-dry weight was first determined.All samples were weighed at room temperature, then placed in an oven, and dried at 100 °C for 15 h.Then, they were cooled in the same oven for 10 h.Afterward, the bone-dry weight was determined to an accuracy of 0.001 g.Based on the European standards for the determination of moisture of sawn timber by oven dry method BS EN 13183-1, the samples moisture content was determined according to the following formula: where  1 is the mass of the specimen before drying;  0 is the mass of the samples at the point of bone-dry weight;  is the moisture content in percent.Based on the bone-dry weight and the determined volume of the individual samples, the respective density of the measured wood could be determined.

Effect of fibre direction on THz signal
The analysis of THz signal attenuation in European Spruce and Maritime Pine reveals a nuanced interplay between the direction of wood fibers, species-specific wood properties, and the THz signal's frequency.The obtained results for the measured attenuation in THz signal for both longitudinal and transversal direction are presented in Figure 2(a), and (b), respectively.
Maritime Pine exhibits a notable fluctuation in transversal attenuation, initially decreasing from 67.5 dB at a frequency of 0.02 THz to a significant low of approximately -14 dB, then variably increasing and decreasing across the spectrum.This behavior suggests a sensitivity to THz frequency, with certain frequencies being more effectively transmitted or absorbed.In contrast, longitudinal attenuation for Maritime Pine shows a generally decreasing trend from -36.5 dB, dipping into more significant negative values in higher frequencies, indicative of stronger signal attenuation.
European Spruce presents a different pattern, with transversal attenuation starting at 72 dB and displaying fluctuations but generally trending upwards in the mid-frequency range, reaching values above 100 dB, such as 126 dB at a frequency around 0.03 THz, before slightly decreasing.This suggests a high frequency sensitivity, particularly in the transversal direction.Longitudinal attenuation for European Spruce starts higher, at 174 dB, and decreases across the frequency range, with periods of intense attenuation, showcasing the substantial impact of fiber orientation on THz signal propagation.
Comparing the two species, European Spruce generally exhibits higher longitudinal attenuation values than Maritime Pine, indicating material-specific differences in how THz signals interact with wood.For instance, at a frequency of approximately 0.03 THz, the longitudinal attenuation for European Spruce is notably higher than the corresponding value for Maritime Pine, illustrating the influence of wood species on THz signal attenuation.

Effect of moisture on THz signal
The analysis of THz signal levels measured as root-mean-square voltage (VRMS) in European spruce under two distinct moisture conditions, is presented in This sharp increase in signal levels with higher moisture content vividly illustrates how moisture serves not just as a medium that amplifies the THz signals but also introduces greater variability in their propagation.This variability and increase in signal intensity with moisture are indicative of moisture's capability to modify the wood's dielectric properties significantly, which in turn affects how THz signals interact with the material.The variation in VRMS values, especially noticeable at the higher moisture level, points to a nuanced sensitivity of the THz signals to frequency changes or scanning conditions, suggesting that the presence of water within the wood's structure can influence the propagation of different frequencies of THz waves differently.
Analyzing the scans specifically in the longitudinal direction of the wood, which aligns with its natural fibers, we can discern the impact of fiber orientation on THz signal transmission.The longitudinal scans show a clear enhancement in signal propagation at higher moisture levels (e.g., a leap from 2 mV at 9.4% moisture to 7 mV at 57.4% moisture in certain scans), highlighting how water presence in the fibrous structure likely enhances the THz signals due to altered dielectric properties.Compared to the results obtained in Fig. 2, different wood species may exhibit unique responses to THz signals based on inherent material properties such as density and porosity, suggested by the significant differences in THz signal behavior with varying moisture contents.

FMCW THz waves for wood drying monitoring
For evaluating the effectiveness of FMCW THz waves in monitoring the drying process of wood, THz reconstructed scans of spruce wood, each corresponding to different moisture content levels: 57.4%, 27.3%, 15.8%, and 9.4%, are presented in Fig. 4. The analysis of the obtained scans the moisture variation effects on the THz signal propagation through the wood's structure.At 57.4% moisture content, which represents fiber saturation, the THz image shows high THz signal reflection levels (Fig. 4a).The uniformity and intensity of reflected THz signals indicate that the presence of water significantly impacts the THz signal, likely due to the high dielectric constant of water at these frequencies.This can lead to a pronounced reflection of the THz signal at the wood's surface, resulting in less penetration depth and stronger signal return at the surface.At 27.3%, a decrease in moisture content is observable by the transition from red to lighter orange and yellow areas (Fig. 4a).There are variations in the reflected THz signals intensities, indicating that the THz signal interacts with the wood differently across the scan.This may correspond to non-uniform drying due to variations in wood density, affecting how the signal is absorbed.With further reduction in moisture to 15.8%, the color palette shifts predominantly to yellow and green (Fig. 4c).These colors signify a weaker interaction of the THz signal with moisture compared to Fig. (a) and (b), as the green and yellow areas imply lower signal reflection and potentially deeper penetration into the wood.This scan suggests a significant moisture reduction that allows the THz signal to traverse the wood structure with less attenuation from water molecules.At 9.4%, indicative of bone-dry conditions, the image predominantly shows blue and light blue areas, marking the least interaction between the THz signal and moisture (Fig. 4d).Here, the THz signal likely penetrates the wood more deeply with minimal reflection, as the wood has low water content that does not impede the signal.This scan level appears more heterogeneous, potentially due to the natural variations in wood density becoming more discernible as the masking effect of water diminishes.

Conclusion
In this study, a robotic FMCW THz wave technique was used for evaluating the effect of wood type and fibre direction on THz signal.The analysis of THz signal attenuation in European Spruce and Maritime Pine reveals distinct patterns influenced by the direction of wood fibers and the specific type of wood.For both species, the attenuation of THz signals varies significantly across the frequency range, highlighting a complex interaction between the THz waves and the wood's internal structure.Specifically, the longitudinal direction tends to exhibit stronger signal attenuation compared to the transversal direction, suggesting that the alignment of wood fibers plays a crucial role in the propagation of THz signals.Additionally, differences between the two types of wood indicate that species-specific properties also affect how THz signals are absorbed or transmitted.The analysis of VRMS values for THz signal levels in European spruce wood at different moisture contents reveals critical insights into the moisture-dependent behavior of THz signals in wood.The significant impact of moisture on signal intensity and variability underscores the potential of THz technology for non-invasive moisture content analysis in wood and other porous materials.Moreover, the observed sensitivity to frequency and the influence of fiber orientation highlights the complex dynamics of THz signal propagation in wood, necessitating a nuanced approach to the development and application of THz-based diagnostic tools.Future research could further elucidate these interactions, potentially leveraging THz technology for a broader range of applications in wood science and beyond.

Fig. 1 .
Fig. 1.The schematic of prepared wood samples and the scanning direction for longitudinal (a); and tangential configurations (b).

Fig 3 .
At a moisture content of 9.4% (Fig 3b), the VRMS values observed span from a minimal 0.9 mV to a peak of 2 mV across various scans, indicating a relatively stable and subdued response of the THz signals.In stark contrast, when the moisture content is elevated to 57.4%, the VRMS values surge dramatically, ranging from 2 mV to 7 mv, as depicted in Fig 3a.