Khondalite: 2026

Dark Fiber Seismology: Turning Telecommunication Cables into Giant Earthquake Sensors

One of the most innovative developments in modern seismology is the use of existing telecommunication infrastructure as a large-scale seismic sensing network. Telecommunication systems often contain unused optical fibers, commonly known as "dark fibers", that are not carrying internet traffic. These spare fibers may extend for tens to hundreds of kilometers beneath roads, railways, pipelines, or across the seafloor. Instead of installing thousands of conventional seismometers, researchers can transform these unused optical fibers into dense arrays of virtual seismic sensors using a technique known as "Distributed Acoustic Sensing (DAS)".

The principle behind DAS is both elegant and remarkably sensitive. At one end of the optical fiber, a device called a "laser interrogator" repeatedly injects extremely short, coherent laser pulses into the fiber. Although optical fibers are manufactured to be highly transparent, they contain microscopic variations in density and composition. These naturally occurring imperfections scatter a tiny fraction of the incident light back toward the interrogator through a phenomenon known as "Rayleigh scattering". Under normal conditions, this backscattered light is simply part of the fiber's optical background. However, when the fiber experiences mechanical deformation, the scattered light carries information about that disturbance.

During an earthquake, seismic waves—including primary (P) waves, secondary (S) waves, and surface waves—propagate through the ground and produce extremely small strains, typically on the order of microstrain or even nanostrain. As these waves pass beneath a buried or submarine fiber-optic cable, they cause the fiber to stretch and compress by minute amounts. These tiny deformations alter both the physical length of the fiber and its refractive index through the "photoelastic effect". Together, these changes modify the phase of the Rayleigh backscattered light. By comparing the phase of successive laser pulses using interferometric techniques, the interrogator can detect these changes with extraordinary precision. Because the wavelength of the laser light is only about 1550 nm, even displacements far smaller than the diameter of an atom can produce measurable phase shifts.

The location of each disturbance is determined from the travel time of the backscattered light. Since the speed of light inside the optical fiber is known, the interrogator can accurately calculate where along the cable the strain occurred. This measurement process is repeated thousands of times every second, allowing continuous monitoring of dynamic ground motion in real time.

A key feature of DAS is that the optical fiber itself serves as the sensing element. No electronic sensors, batteries, or geophones are installed along the cable. Instead, the interrogator effectively divides the fiber into thousands of closely spaced sensing intervals, often only a few meters long. Each interval behaves like an individual seismic sensor, collectively forming a dense "virtual seismic array". Unlike conventional seismic networks, where measurements are available only at discrete station locations, DAS provides continuous spatial measurements along the entire length of the fiber. Moreover, only the interrogator requires electrical power, whereas traditional seismic arrays require every individual station to be powered and maintained.

This capability is particularly valuable for submarine seismology. Thousands of kilometers of fiber-optic telecommunication cables already traverse the ocean floor, yet large regions of the oceans remain poorly instrumented because deploying and maintaining seafloor seismometers is technically challenging and prohibitively expensive. Many submarine communication cables contain unused fibers that can be repurposed for DAS measurements, offering an opportunity to dramatically improve seismic monitoring in areas where conventional instrumentation is sparse or absent. This approach has already demonstrated the ability to detect earthquakes, ocean-generated microseisms, volcanic activity, and even passing ocean waves using existing telecommunications infrastructure.

Distributed Acoustic Sensing belongs to the broader family of "Distributed Fiber Optic Sensing (DFOS)" technologies, which exploit the interaction between light and the optical fiber to measure physical quantities continuously along its length. Different DFOS techniques rely on different scattering mechanisms depending on the parameter being measured. "Distributed Temperature Sensing (DTS)" typically uses Raman or Brillouin scattering to measure temperature, while "Distributed Static Strain Sensing (DSS)" commonly relies on Brillouin scattering to monitor long-term deformation. DAS, in contrast, utilizes coherent Rayleigh backscattering to measure rapidly varying dynamic strain, making it particularly well suited for seismic monitoring and vibration sensing. The technology was initially developed by the oil and gas industry for geophysical monitoring before being adapted for regional and global seismic observation.

Despite its considerable advantages, dark-fiber seismology also has important limitations. Unlike conventional seismometers, DAS measures "strain" along the optical fiber rather than the three-dimensional ground displacement or acceleration. Consequently, the recorded signals require different interpretation methods and cannot always be compared directly with conventional seismic records. The measurements are also highly directional, with maximum sensitivity to deformation occurring along the axis of the fiber, making cable orientation an important factor in data quality. Signal strength further depends on how well the cable is mechanically coupled to the surrounding ground; fibers directly buried in soil or attached firmly to the seafloor generally produce stronger signals than fibers loosely installed inside protective ducts. In addition, DAS systems generate enormous volumes of data. A 50–100 km fiber monitored at kilohertz sampling rates can produce several terabytes of data every day, demanding substantial storage capacity and computational resources for processing. Finally, optical attenuation limits the maximum distance that can be monitored from a single interrogator, requiring multiple interrogation units for very long fiber links.

Despite these challenges, dark-fiber seismology represents a major shift in earthquake monitoring. By transforming existing telecommunication infrastructure into continuous, high-density sensing networks, DAS provides an economical and scalable means of observing seismic activity over unprecedented spatial scales. As interrogation technology, signal processing algorithms, and computational capabilities continue to improve, dark fibers are expected to play an increasingly important role in both terrestrial and submarine seismic monitoring, complementing rather than replacing conventional seismometer networks.

Bibliography:

Marra et al. (2018) Ultrastable laser interferometry for earthquake detection with terrestrial and submarine cables. Science. https://doi.org/10.1126/science.aat4458.
Zhan Z. (2019) Distributed Acoustic Sensing Turns Fiber Optic Cables into Sensitive Seismic Antennas, Seismological Research Letters, DOI: 10.1785/0220190112.
Williams et al. (2019) Distributed sensing of microseisms and teleseisms with submarine dark fibers. Nature Communications. https://doi.org/10.1038/s41467-019-13262-7
Zhu et al. (2023) Seismic arrival-time picking on distributed acoustic sensing data using semi-supervised learning. Nature Communications. https://doi.org/10.1038/s41467-023-43355-3
Li et al. (2023) The break of earthquake asperities imaged by distributed acoustic sensing. Nature. https://doi.org/10.1038/s41586-023-06227-w

Pressure-Sensitive Adhesives: The Materials Science Behind Everyday Stickiness

Pressure-sensitive adhesives (PSAs) are among the most widely used engineered materials encountered in daily life. They are found in adhesive tapes, labels, packaging systems, protective films, sticky notes, medical dressings, and numerous industrial products. Despite their ubiquity, PSAs represent a remarkably sophisticated class of soft materials whose performance emerges from a delicate balance between surface interactions, polymer physics, rheology, and fracture mechanics.

Unlike structural adhesives such as epoxies or polyurethanes, PSAs are designed to form an immediate bond upon application of relatively low pressure. No thermal activation, solvent evaporation, or chemical curing is required. Instead, adhesion arises from the intrinsic viscoelastic properties of the material.

According to definitions adopted by the Pressure Sensitive Tape Council, the Adhesive and Sealant Council, and the scientific literature, a pressure-sensitive adhesive is a viscoelastic material that remains permanently tacky in its dry form and forms an immediate bond when light pressure is applied. The adhesive requires no activation by water, heat, or solvent and is capable of bonding to a wide variety of dissimilar surfaces.

The apparent simplicity of this definition conceals a challenging materials design problem. A PSA must flow sufficiently during application to establish intimate contact with a substrate, resist debonding when subjected to peeling forces, and simultaneously maintain dimensional stability under long-term loading. These requirements are often contradictory and must be balanced through careful control of the adhesive's viscoelastic response.

The performance of a pressure-sensitive adhesive is typically characterized through three primary metrics: tack, peel adhesion, and shear resistance.

Tack describes the ability of an adhesive to form a bond of measurable strength under conditions of low contact pressure and short contact time. From a physical perspective, tack reflects how rapidly an adhesive can wet a surface and generate interfacial interactions before significant relaxation or restructuring occurs. Measurements are typically conducted over timescales ranging from milliseconds to seconds.
Peel adhesion quantifies the resistance of an adhesive bond to progressive separation by peeling. It is commonly reported as the force per unit width required to peel one adherend from another under specified environmental conditions, peel angle, and separation rate. Peel adhesion is often regarded as a practical measure of how strongly an adhesive is attached to a surface, although the underlying mechanics are considerably more complex than a simple measure of bond strength.
Shear resistance, often referred to as holding power, measures the ability of an adhesive bond to resist creep or sliding under a sustained shear load. Unlike tack and peel adhesion, which involve relatively rapid deformation processes, shear performance is governed by long-term viscoelastic deformation and is often reported as time to failure under a constant load.

A simple example illustrates the distinction among these properties. Consider a tape used to support a picture frame on a wall. Tack determines how quickly the tape bonds when first applied. Peel adhesion determines the force required to remove the tape from the wall. Shear resistance determines whether the tape can support the weight of the frame for months without gradually creeping downward. A tape may perform exceptionally well in one of these categories while performing poorly in another, highlighting the need to evaluate all three metrics when assessing PSA performance.

Because both tack and peel measurements involve separating an adhesive from a substrate, they are frequently assumed to characterize the same property. In reality, they probe fundamentally different stages of bond development. Tack measures the formation of an initial bond. Contact pressures are low, contact times are short, and the adhesive has limited opportunity to relax or establish extensive interfacial contact. The measurement therefore emphasizes rapid wetting and short-time viscoelastic response. Peel adhesion, by contrast, evaluates a fully developed bond. Prior to testing, the adhesive is typically subjected to a standardized application pressure and allowed to remain in contact with the substrate for a defined dwell time that may range from minutes to days. The resulting measurement reflects the behavior of a mature adhesive interface.

An equally common misconception is the assumption that peel adhesion is synonymous with bond strength. In practice, different adhesion tests probe different failure mechanisms. Tensile adhesion tests evaluate resistance to normal separation, lap shear tests evaluate resistance to sliding under shear loading, and peel tests evaluate resistance to crack propagation during progressive debonding. The distinction is important because peeling localizes stress at a moving crack tip. In tensile and shear loading, stresses are distributed over a relatively large bonded area. During peeling, however, deformation is concentrated within a narrow process zone surrounding the advancing crack front. Consequently, peel tests are often substantially more severe than conventional tensile or shear tests and are strongly influenced by energy dissipation mechanisms within the adhesive.

The behavior of pressure-sensitive adhesives is governed primarily by their rheological response. As viscoelastic materials, PSAs exhibit both elastic and viscous characteristics, and their performance can be described using three fundamental parameters.

The storage modulus, G′, represents the elastic or energy-storing component of the response and is associated with solid-like behavior.
The loss modulus, G″, represents the viscous or energy-dissipating component and is associated with liquid-like behavior.
The Loss Tangent, the ratio of these quantities, expressed as tanδ = G″/G′, provides a measure of the relative contributions of viscous and elastic deformation.

The significance of these parameters becomes evident when considering the timescales over which adhesive deformation occurs. Short-duration events such as tack formation and peeling correspond to high-frequency deformation. Long-duration events such as creep and holding power correspond to low-frequency deformation.

As a result, PSA performance cannot be characterized by a single modulus value. Instead, the entire frequency-dependent viscoelastic spectrum must be considered.

This is what a good PSA does: it shows a strong frequency dependence in its viscoelastic behavior, with a markedly higher storage modulus at high frequencies than at low, and a higher loss tangent under rapid deformation than under slow. Therefore, a high-performance PSA exhibits markedly different behavior across different timescales. During application, the adhesive should possess a sufficiently low modulus to deform, wet the substrate, and establish intimate interfacial contact. During peeling, the adhesive should become stiffer and dissipate significant mechanical energy, thereby increasing resistance to crack propagation. Under sustained loading, it should retain sufficient elasticity to suppress creep and maintain dimensional stability.

In practical terms, an effective PSA must simultaneously behave as a compliant material during bonding, a tough material during debonding, and an elastic material during long-term loading. Achieving these seemingly contradictory requirements is the central challenge of PSA formulation.

When an adhesive bond fails, the mode of failure often provides important information regarding the limiting performance mechanism.

Adhesive failure occurs when separation takes place at the interface between the adhesive and the substrate. Little or no adhesive residue remains on the substrate after debonding, indicating that interfacial adhesion is weaker than the internal strength of the adhesive layer. A label that peels cleanly from a plastic surface without leaving residue is a typical example of adhesive failure.

Cohesive failure occurs when rupture takes place within the adhesive layer itself. Adhesive residue remains on both separated surfaces, indicating that the interfacial bond is stronger than the internal cohesive strength of the material. The sticky residue often left behind after removing duct tape is a familiar example of cohesive failure.

The distinction between these failure modes is more than a matter of observation. It provides direct insight into whether performance limitations originate from insufficient substrate adhesion or inadequate internal mechanical integrity of the adhesive.

From a materials science perspective, pressure-sensitive adhesives are a striking example of how macroscopic performance emerges from time-dependent material behavior. Their effectiveness is not determined solely by how strongly they stick, but by how they manage the competing demands of wetting, energy dissipation, elasticity, and fracture resistance across a wide range of deformation rates and loading conditions.

Bibliography

Sun et al. (2013) A review on mechanical properties of pressure sensitive adhesives. International Journal of Adhesion and Adhesives. https://doi.org/10.1016/j.ijadhadh.2012.10.011
Creton C (2003) Pressure-Sensitive Adhesives: An Introductory Course. MRS Bulletin. https://doi.org/10.1557/mrs2003.124
Mapari et al (2021) Developments in pressure-sensitive adhesives: a review. Polymer Bulletin. https://doi.org/10.1007/s00289-020-03305-1

Ship Speed Optimization and Energy Efficiency in Modern Maritime Transport

The shipping industry carries nearly 80% of global cargo transportation by volume and remains a fundamental pillar of international trade. According to the International Institute of Marine Surveying (IIMS), the industry recorded an average annual growth rate of 2.9% between 2003 and 2023, including a 3–5% rise in global container demand. At the same time, the sector faces increasing operational pressures arising from geopolitical instability, rising fuel and rerouting costs, port congestion, and logistical inefficiencies. Fuel expenses alone account for nearly 50–60% of a ship’s total operating cost. A 23% increase in fuel prices can raise average import prices by approximately 1%, while a 130% increase may increase import prices by nearly 6.5%. In addition to economic concerns, the shipping sector contributes close to 3% of global greenhouse gas (GHG) emissions. The International Maritime Organization (IMO) has therefore established ambitious targets to reduce GHG emissions by 70% by 2040 and achieve net-zero emissions by 2050. Achieving these targets will require substantial improvements in both technological systems and operational management practices. However, strategies aimed at reducing emissions often conflict directly with the industry’s objective of minimizing operational costs. Consequently, shipping operators frequently prefer operational optimization measures, particularly ship speed optimization, over expensive technological interventions. Today, controlled reduction and optimization of ship speed has emerged as one of the most effective approaches for reducing fuel consumption, operational costs, and emissions simultaneously.

Ship speed optimization strategies must balance economic performance with environmental objectives while maintaining strict compliance with operational and safety constraints. In practice, however, sailing conditions are influenced by several dynamic factors, including weather and sea state, waterway congestion, port berthing efficiency, operational delays, fuel price fluctuations, and unexpected disruptions. These factors significantly affect voyage duration, fuel consumption, and the relationship between speed and energy efficiency. Furthermore, different categories of ships operate under very different commercial and operational constraints. For example, container ships prioritize schedule reliability because commercial competitiveness depends strongly on timely delivery. As a result, maintaining fixed arrival times often takes precedence over energy efficiency. In contrast, bulk carriers operate within highly volatile spot markets where freight rates fluctuate continuously, making voyage time comparatively less critical. Tankers exhibit greater operational flexibility and are influenced by a combination of long-term contractual obligations and spot-market conditions. Inland waterway vessels operate under an entirely different set of constraints, where navigational regulations, lock schedules, bridge clearances, water-level fluctuations, and regional operational conditions heavily influence speed optimization. Consequently, inland vessel optimization requires significantly greater regional and situational specificity than ocean-going vessels.

Current ship speed optimization approaches can broadly be classified into three categories with increasing levels of sophistication.

Static optimization models: These represent the earliest and simplest class of optimization approaches. They provide a fixed pre-voyage optimal solution based on constant environmental conditions and predefined operating parameters. Such models are useful for benchmarking energy efficiency, supporting strategic planning, and developing macro-level policy frameworks. However, they are unable to account for uncertainty or deviations between assumed and actual operating conditions during a voyage.
Offline dynamic optimization models: Offline dynamic optimization is also performed prior to voyage initiation but incorporates time-varying voyage-specific parameters such as forecast weather, sea conditions, and energy prices. These models generate voyage-specific speed plans tailored to anticipated operating conditions. They are particularly suitable for medium- to long-distance routes where environmental variations are relatively predictable. Compared to static models, offline dynamic models better capture temporal variability in operating conditions. However, once the voyage begins, the optimization plan cannot adapt continuously to actual ship conditions or unforeseen changes during operation.
Online dynamic optimization models: Online dynamic optimization models operate in a closed-loop framework using continuously or periodically updated real-time data acquired from onboard sensors and environmental monitoring systems during the voyage. Because these models continuously update predictions and re-optimize the remaining voyage, they perform particularly well under uncertain and rapidly changing conditions. However, several important limitations remain. First, sensor malfunction or communication delays can significantly degrade prediction reliability. Second, the high computational requirements and rapid response demands make implementation challenging on shipboard edge-computing systems. Third, excessively frequent optimization may produce unstable operational decisions, while repeated speed adjustments can increase crew workload and accelerate wear in propulsion systems.

These speed optimization models are often integrated with fuel consumption models to estimate energy usage and associated GHG emissions. Although such integrated approaches establish the relationship between speed and energy efficiency, their overall reliability depends heavily on the accuracy of the underlying speed prediction models. Ship energy consumption models are generally categorized into three major classes.

Mechanism-based models: These analytical or empirical models are derived from mathematical relationships among vessel speed, engine power, and fuel consumption. Environmental disturbances such as currents, waves, water depth, and wind influence vessel resistance and motion characteristics, thereby affecting energy consumption. By explicitly incorporating the governing physical processes, mechanism-based models can effectively represent complex navigational environments.
Data-driven models: Unlike mechanism-based approaches that rely on predefined physical relationships, data-driven models use machine learning techniques to learn nonlinear relationships directly from historical operational data. These datasets typically include Automatic Identification System (AIS) data, bridge automation system data, and meteorological information. Such models are particularly effective at capturing highly nonlinear operational behaviors and complex environmental interactions that are often difficult to represent accurately using purely mechanistic approaches.
Hybrid mechanism and data-driven models: To ensure that data-driven optimization remains physically realistic and does not violate fundamental engineering principles, physical constraints derived from mechanism-based or “White Box” models are often incorporated into machine-learning or “Black Box” models. This integration produces hybrid or “Grey Box” models that combine the predictive capability of data-driven approaches with the physical consistency of mechanistic modeling. These hybrid frameworks provide improved predictive accuracy while maintaining physical plausibility, making them particularly suitable for high-credibility operational optimization applications.

It is also important to recognize that speed optimization requirements vary across different phases of a ship’s operational life cycle, including the design and commissioning phase, mid-service operational phase, and ageing or end-of-life phase. Each phase is associated with distinct objectives, constraints, and operational priorities. During the design and commissioning phase, the primary objective is to maximize fuel efficiency through optimal propulsion-system matching and refinement of propulsion-efficiency characteristics. Mechanism-based approaches are generally preferred at this stage because of their strong physical interpretability. As ships enter active service and operational datasets become available, data-driven and hybrid approaches become increasingly valuable. These methods allow continuous model refinement using real-time information related to vessel condition, operating environment, and energy consumption patterns. In the ageing and end-of-life phase, additional considerations become important, including equipment degradation, maintenance costs, fuel substitution strategies, and evolving regulatory requirements. Progressive degradation throughout a vessel’s service life can substantially influence the effectiveness and reliability of speed optimization strategies.

Despite significant progress in optimization methodologies, most existing models remain highly scenario-specific and exhibit limited generalizability. Many current approaches rely heavily on historical datasets associated with particular ship types, operating routes, or individual voyages, thereby restricting their transferability across different operating conditions. Consequently, there is a growing need for physics-informed machine-learning approaches and modular optimization architectures capable of adapting across varying ship classes, waterways, drafts, loading conditions, and operational environments.

Ship speed optimization also involves multiple stakeholders, including shipowners, ship managers, cargo owners, charterers, and port authorities. Operational decisions related to energy efficiency are therefore influenced not only by technical considerations, but also by commercial agreements, freight-market dynamics, and cargo delivery requirements. As a result, even when optimization strategies demonstrate clear technical benefits in reducing energy consumption, their practical implementation may be constrained or rejected if recommended operating speeds conflict with charter-party agreements, cargo handling schedules, or delivery time commitments.

Human factors further influence the adoption of speed optimization systems. Crew members may show limited acceptance of optimization recommendations that conflict with their operational experience, particularly when concerns arise regarding safety, lack of transparency in algorithmic decision-making, or perceived reductions in professional autonomy. Studies on maritime automation indicate that trust in automated systems is strongly influenced by factors such as professional identity, operational experience, and age, with highly experienced crew members often exhibiting greater skepticism toward automated recommendations. Under high-workload conditions or during frequent task switching, crews generally prioritize safety and schedule reliability over energy-efficiency objectives. Resistance to adoption may further increase when decision-support systems require additional manual data input or involve complex operational procedures.

Source:

Fan et al. (2026) Review of speed optimisation for ship energy efficiency: methods, taxonomy and challenges. Energy. https://doi.org/10.1016/j.energy.2026.141138.
Luo et al. (2024) Accuracy and applicability of ship's fuel consumption prediction models: A comprehensive comparative analysis. Energy. https://doi.org/10.1016/j.energy.2024.133187.