Chemical Composition, Microstructure, and Mechanical Properties of Steel Rails Produced by China Hanyang Iron Works

Abstract

The history of steel rail manufacturing at the Hanyang Iron Works in China is briefly summarized first in this paper. The chemical composition, microstructure, inclusions, and mechanical properties of the manufactured rails were then systematically analyzed. The progress of manufacturing process evolution and the enhancement of product quality are elucidated through this study. The chemical composition of steel rails produced by Hanyang Iron Works in its initial stage (before 1904) was characterized by low carbon content (0.13–0.22 wt.%) and high phosphorus levels (P ≥ 0.15 wt.%). Inclusions were primarily identified as sulfides (MnS) and composite inclusions of sulfides and silicates (MnS·SiO₂). The microstructure consisted of a large amount of ferrite and pearlite. Following the technical transformation from 1905 to 1908, dephosphorization was achieved and the composition control of the steel rails was optimized. The carbon content of the rails was increased to above 0.48 wt.%, while the phosphorus content was significantly reduced (P ≤ 0.10 wt.%). The inclusions were identified as sulfides (MnS) and composite inclusions consisting of sulfides and aluminum oxides (MnS·Al₂O₃). The microstructure was transformed into a combination of a small amount of proeutectoid network ferrite and pearlite. The mechanical performance of the steel rails was substantially improved via the implementation of technological upgrades at the Hanyang Iron Works. A tensile strength of 800 MPa grade was achieved in some rails, which constitutes a 200 MPa increment over the strength of rails from the early production period. A transition in the fracture morphology of tensile specimens was observed, shifting from large and shallow dimples with a small amount of cleavage fracture to small, shallow dimples combined with predominant cleavage fracture.

1. Introduction

Among the numerous significant innovations of the Industrial Revolution, the railway is considered one of the technologies with the most profound impact on human society. The rail is the core load-bearing component of railway track systems, and its development has always been intrinsically tied to advancements in ironmaking and steelmaking technologies. Over the past century, extensive research and industrial practices in Europe and North America laid the physical metallurgy foundation for heavy-haul rail steel manufacturing [1,2,3,4,5]. Historical reviews and technical syntheses systematically established the fundamental relationships between melting practices, chemical composition control (notably carbon, manganese, and residual limits), and the resulting microstructural evolution and mechanical properties of steel rails [1,2,3]. Further landmark engineering evaluations in European and American heavy-haul operations elucidated how inclusion cleanliness, pearlitic microstructures, and head-hardening treatments govern rail wear, rolling contact fatigue (RCF), and catastrophic fracture behavior under increasing axle loads [4,5].

On 27 September 1825, the opening of the Stockton & Darlington Railway in the United Kingdom triggered an unstoppable momentum in the development and dissemination of railways. By the end of the 19th century, advanced railway networks had been gradually constructed across Western European nations, including the United Kingdom and Germany, as well as the United States [6,7,8]. The rail is a critical component of railway construction, by which the operation and load of locomotives are directly supported [4]. The quality of railway tracks is closely linked to steel production technologies. Steel rails with superior quality and lower costs were provided for railway construction through the development of iron and steel production technologies. With the development of steel production technology, railway rails have gradually evolved from the original cast iron to Bessemer steel, medium-carbon steel, and eutectoid steel. Railways have evolved from the initial smoke-emitting steam locomotives to the subsequent highly mobile diesel locomotives, and finally to modern electrified high-speed trains. These modern rail transit systems have reconfigured spatial-temporal dynamics and fundamentally reshaped global transportation networks [9].

As an efficient mode of transport, railway transportation is characterized by low energy consumption and low carbon emissions. This is primarily attributed to the low frictional resistance between the wheels and the rails. News of railway development was introduced to China circa 1840. In 1876, the Woosung Railway was established as the first commercial railway line in China. The efficiency and convenience of railway transport were thereafter acknowledged throughout China. At that time, China lacked the manufacturing capacity for steel rails, locomotives, and other essential components required for railway construction. In 1890, Zhang Zhidong, the Governor-General of Huaguang, established the Hanyang Iron Works at the northern foot of Guishan Mountain, south of the Han River in Wuhan. The primary objective was the manufacturing of steel rails for the “Beijing-Hankou Railway.” The establishment of the Hanyang Iron Works marked the fledgling beginnings of the Chinese iron and steel industry. It represents the most profound and brilliant milestone in the history of steel production and application in China. The establishment of the Hanyang Iron Works laid the initial material foundation and provided invaluable practical experience for the subsequent development of the Chinese iron and steel industry [10,11,12,13,14].

Today, although smoke-emitting steam locomotives are still observed in films or memories. However, steel rails manufactured by the Hanyang Iron Works can still be recovered from aged railway lines, rail yards, and scrap collection points, with occasional specimens found in railway museums featuring well-preserved inscriptions. Unfortunately, research regarding the evolution of manufacturing technology, mechanical properties, microstructure, and inclusions of Hanyang Iron Works rails is relatively scarce [15,16,17,18]. The chemical composition and microstructure of certain steel rails have been inspected and analyzed by scholars such as Fang Yibing and Qian Wei. However, these studies were primarily utilized for researching the industrial history of the steel rails manufactured by Hanyang Iron Works. In-depth analyses of the relevant technological evolution from the perspectives of inclusions and microstructure have not been performed [19].

Since 2018, fifteen steel rails with various inscriptions from different historical periods of the Hanyang Iron Works have been collected by the High-Performance Steel Materials Team of Shanghai University. A relatively comprehensive collection of Hanyang Iron Works steel rails, in terms of both quantity and variety, has been assembled. This study systematically uncovers the metallurgical nature behind the historic technological transitions in early Chinese rail manufacturing, filling a critical knowledge gap regarding how process evolution directly dictates material performance. By evaluating ten representative rail samples spanning nearly three decades, this work aims to quantitatively establish the correlations among melting process upgrades, chemical composition optimization (particularly carbon optimization and phosphorus control), and the resulting microstructure and inclusion characteristics. This not only elucidates the scientific root causes of historical rail failure and quality evolution but also provides indispensable archival references and metallurgical insights for industrial evolution studies and cultural relic conservation.

2. Brief History and Testing Process

2.1. Brief History of Hanyang Iron Works

Established in 1890 by Zhang Zhidong, the Hanyang Iron Works was a milestone in China’s steel industry, built to supply rails for the Beijing-Hankou Railway. Figure 1 illustrates its workshops and power equipment. Phase I construction featured two blast furnaces, an open-hearth furnace, two Bessemer converters, and a rail production line. Utilizing imported equipment from Europe, the plant commenced rail production in 1894. However, early rails suffered from brittleness and fracture issues due to a lack of technical expertise.

To improve quality, a major technological transformation was executed from 1905 to 1908. Concurrently, it merged into the Hanyeping Company, China’s first coal-iron conglomerate. By dispatching personnel abroad to study advanced metallurgy, the works laid a foundational technical framework for China’s modern steel industry. Ultimately, rail manufacturing ceased in 1922, and all operations halted by 1924 due to complex political and economic factors.

2.2. Hanyang Iron Works Rails and Testing Process

The Hanyang Iron Works was one of the earliest state-run iron and steel enterprises in modern China. The manufacturing and evolution of its steel rails reflect the developmental trajectory of modern China’s iron and steel industry. After the completion and commissioning of the first phase of construction, the Hanyang Iron Works did not achieve stable rail production and begin supplying the Beijing-Hankou Railway until around 1897.

From 1897 to 1904, the Hanyang Iron Works produced approximately 63,300 tons of steel rails and accessories for the Beijing-Hankou Railway (accounting for about 57% of the total rails used in its construction). However, the steel rails at this time were brittle and prone to fracture.

After 1904, the Hanyang Iron Works initiated its Phase II construction and technical transformation, abandoning Bessemer converter steelmaking in favor of the basic Siemens-Martin open-hearth process. Upon the completion of the upgrades between 1905 and 1908, the issue of poor rail quality was resolved. The quality of the steel rails improved significantly, and production capacity increased. Prior to the Revolution of 1911, the Hanyeping Company operated three blast furnaces and six steelmaking furnaces, with an annual output of over 20,000 tons of steel rails. At that time, there were no relevant national standards for rail manufacturing in China. The Hanyang Iron Works produced rails in ten different specifications, including fixed lengths of 10 yards (9144 mm) and 12.5 m, with weights such as 85 lb/yd (42 kg/m), 76 lb/yd (38 kg/m), 67.3 lb/yd (33 kg/m), and 60 lb/yd (30 kg/m).

Around the time of the First World War, the production volume of steel rails at the Hanyang Iron Works reached its peak. This provided strong support for the construction and maintenance of Chinese railways at that time. The Hanyang Iron Works ceased its production of steel rails around 1922 (with all production operations halted by 1924). From the start of production in 1894 until the rail lines were shut down in 1922, the Hanyang Iron Works produced approximately 3300 km of railway steel rails.

The images and inscription information of typical Hanyang Iron Works steel rails collected by the High-Performance Steel Materials Team of Shanghai University are shown in Figure 2 below. The rail weight information in the table consists of values converted from the actual weighing of the rails. The conversion method is: actual mass of the rail/length of the rail sample, where the unit of mass is kilograms (kg) and the unit of length is meters (m).

The ten historical rail specimens were systematically selected to ensure full chronological and technological representativeness across the operational lifespan of Hanyang Iron Works (1894–1924). Selection was governed by three criteria: (1) Chronological continuity, sampling across early (1894–1907), middle (1908–1918), and late (1919–1924) production phases; (2) Process diversity, covering key metallurgical shifts from imported Acid Bessemer steel to native Basic Open-Hearth refining; and (3) Specimen integrity, ensuring uncorroded core material sufficient for tensile, impact, and microstructural analysis.

Specimens were prepared from the collected Hanyang Iron Works rails following the GB/T 2585-2021 standard [20] (Hot-rolled steel rails for railways) to evaluate their chemical constituents, non-metallic inclusion ratings, microstructural characteristics, and tensile performance. Prior to sampling and spectral excitation, the recovered rails underwent surface oxide removal, chemical pickling, and ultrasonic degreasing to ensure that the core evaluation was free from contamination.

Chemical composition testing was performed using a SPECTROMAXx6 spark optical emission spectrometer (OES). The standard reference material was consistently a U71Mn sample, with a composition of 0.72C-0.24Si-1.25Mn-0.031P-0.024S. During the testing process, oil stains and other contaminants were carefully removed from the sample surfaces; each sample was tested three times, and the average value was taken.

After sampling for inclusions and microstructure, the specimens were prepared through coarse grinding, fine grinding, polishing, and etching with 4% nital. Observations were conducted using a Zeiss Axio Imager.M2m upright metallurgical microscope. An FEI Apreo 2S scanning electron microscope (SEM) was used to observe the microstructural morphology and perform Energy-Dispersive Spectroscopy (EDS) analysis.

The Brinell hardness of the rail samples was measured using a Buehler BH3000 Brinell hardness tester. The test locations were at the top of the rail head and at a position 20 mm below the top of the rail head. The test load was 3000 kg, and the diameter of the steel ball indenter was 10 mm. Three valid readings were recorded for each sample and averaged.

The tensile properties of the rail samples were tested using an Instron 5985 universal testing machine. The primary chemical constituents of modern rail steel are carbon, silicon, and manganese, along with the unavoidable elements phosphorus and sulfur.as calculated. Schematic diagrams of the sampling locations and the tensile specimen dimensions are shown in Figure 3.

It should be noted that all collected samples had undergone long-term in-service operation on various railway lines prior to their recovery. Due to the lack of complete archives from over a century ago, specific historical operating parameters—such as the precise cumulative gross tonnage, train speeds, and localized maintenance history—are no longer traceable. These long-term service conditions inevitably exerted localized effects on the rail materials, particularly leading to severe work hardening and plastic flow near the top surface of the rail head. To capture the intrinsic metallurgical characteristics of the historical steel and minimize the influence of operational anomalies or gradient deformation artifacts, a rigorous sampling strategy was implemented in this study: specimens for bulk chemical composition analysis, non-metallic inclusion rating, and tensile testing were systematically extracted from the interior or core region of the rail head, with the specific sampling locations illustrated in Figure 3.

It is worth noting that while multi-temperature Charpy V-notch impact testing is crucial for directly evaluating impact toughness and defining the ductile-to-brittle transition temperature (DBTT) of rail steels, it was not performed in this study. This omission was primarily dictated by the limited total volume of material available under historical relic preservation constraints, as well as the geometry of several early light-rail samples, whose cross-sectional dimensions were too small to accommodate standard specimen machining. To compensate for this limitation, the microscopic fracture mechanisms and failure transition modes were systematically characterized via high-resolution SEM fractography of the tensile fracture surfaces.

Although core sampling minimizes surface wear and severe plastic deformation localized near the rail head, long-term operational service may still leave subtle microstructural signatures across the cross-section. Prolonged exposure to cyclic contact stresses can induce micro-strain accumulation and dislocation restructuring even in the core region. Additionally, long-term ambient temperature service over decades can facilitate minor strain-aging effects, leading to local carbon redistribution or fine cementite precipitation along ferrite boundaries. Nevertheless, core specimens retain the primary metallurgical features—such as original grain size, macro-segregation, and inclusion distribution—making them reliable representatives of initial manufacturing conditions.

3. Results

3.1. Chemical Composition

The primary chemical constituents of modern rail steel are carbon, silicon, and manganese, along with the unavoidable elements phosphorus and sulfur. Some alloy rails contain small amounts of elements such as copper, chromium, niobium, and vanadium, as well as rare earths. Carbon is the primary strengthening element in rail steel, while silicon and manganese are also essential constituent elements. Phosphorus and sulfur are considered harmful impurities in rail steel. The chemical composition test results for Hanyang Iron Works rails from different periods are shown in

Table 1. Chemical composition of steel rails manufactured by Hanyang Iron Works (wt.%).

Element/Year	C	Si	Mn	P	S	Cr	Ni	Cu	Alt
1899	0.22 ± 0.01	0.06 ± 0.01	0.69 ± 0.01	0.15 ± 0.03	0.089 ± 0.003	0.003 ± 0.001	0.005 ± 0.001	0.48 ± 0.07	<0.0005
1902	0.20 ± 0.01	0.07 ± 0.02	0.87 ± 0.03	0.19 ± 0.03	0.086 ± 0.008	0.005 ± 0.001	0.014 ± 0.004	0.33 ± 0.05	<0.0005
1904	0.13 ± 0.04	0.03 ± 0.01	0.62 ± 0.01	0.16 ± 0.02	0.091 ± 0.002	0.004 ± 0.001	0.032 ± 0.01	0.15 ± 0.04	<0.0005
1909	0.42 ± 0.02	0.02 ± 0.02	0.69 ± 0.07	0.042 ± 0.002	0.066 ± 0.003	0.060 ± 0.002	0.030 ± 0.009	0.29 ± 0.03	0.007 ± 0.001
1911	0.50 ± 0.03	0.08 ± 0.02	0.63 ± 0.03	0.061 ± 0.001	0.079 ± 0.005	0.031 ± 0.001	0.029 ± 0.002	0.47 ± 0.02	0.005 ± 0.001
1913	0.54 ± 0.03	0.07 ± 0.03	0.89 ± 0.09	0.10 ± 0.003	0.065 ± 0.003	0.009 ± 0.002	0.025 ± 0.003	0.37 ± 0.05	0.004 ± 0.001
1914	0.52 ± 0.01	0.04 ± 0.02	0.62 ± 0.03	0.066 ± 0.001	0.068 ± 0.001	0.008 ± 0.002	0.024 ± 0.002	0.44 ± 0.05	0.004 ± 0.001
1917	0.49 ± 0.02	0.07 ± 0.01	0.53 ± 0.03	0.039 ± 0.001	0.068 ± 0.003	0.130 ± 0.015	0.024 ± 0.008	0.58 ± 0.07	0.008 ± 0.001
1918	0.54 ± 0.01	0.09 ± 0.01	0.63 ± 0.03	0.059 ± 0.004	0.045 ± 0.001	0.043 ± 0.016	0.019 ± 0.003	0.62 ± 0.06	0.011 ± 0.003
1921	0.48 ± 0.04	0.05 ± 0.01	0.44 ± 0.01	0.033 ± 0.002	0.043 ± 0.001	0.046 ± 0.003	0.032 ± 0.007	0.56 ± 0.08	0.004 ± 0.001

Note: “Total Al” denotes the total aluminum content (wt.%) determined by optical emission spectrometry, accounting for both solid-solution aluminum and bound aluminum within oxide/aluminate inclusions.

.

According to the test results, the carbon content in the rails produced before the technical transformation (1899, 1902, and 1904) was generally low, and the carbon content in these three samples showed a gradual decrease from 1899 to 1904. The phosphorus (P ≥ 0.15 wt.%) and sulfur (S ≥ 0.086 wt.%) levels in the rails produced in 1899, 1902, and 1904 were significantly higher than those in the rails produced after 1909 (P ≤ 0.10 wt.%, S ≤ 0.079 wt.%). From the perspective of chemical composition, another characteristic of the Hanyang Iron Works rails is the presence of copper, which may be related to the raw iron ores used by the plant. Other alloying elements, such as manganese, showed certain fluctuations, which also reflects the process control capabilities of the Hanyang Iron Works at that time.

Although direct quantitative analysis of interstitial gas elements (e.g., oxygen, nitrogen, and hydrogen) was inaccessible due to the detection limits of optical emission spectrometry (OES), their metallurgical imprints were distinctly retained within the rail matrix. In early converter rails (1899–1904), the atmospheric blowing process introduced excessive dissolved oxygen and nitrogen into the melt owing to the absence of secondary vacuum refining equipment. This severe gas enrichment directly resulted in the formation of unclosed pores and gas shrinkage cavities, as visible in Figure 4. Over the past decades, these interstitial elements—especially nitrogen—highly promoted strain-aging embrittlement. Conversely, the basic open-hearth process introduced after 1908 significantly suppressed dissolved gases through prolonged slag–metal reactions and advanced aluminum deoxidation practices; this was chemically substantiated by the complete elimination of silicate micropores and the emergence of fine $\mathrm{MnS}·{\mathrm{Al}}_2{\mathrm{O}}_3$ inclusions.

3.2. Microstructure

Modern rail steels can be classified into pearlitic, bainitic, and martensitic types based on their microstructural categories. The microstructures of the rails manufactured by the Hanyang Iron Works are shown in Figure 4 and Figure 5. As shown in Figure 4 and Figure 5, the microstructure of the rails produced in 1899, 1902, and 1904 consists of a large amount of ferrite and a small amount of pearlite. Unclosed pores (voids) are visible in the structure, and visually, the pearlite content appears to decrease slightly from 1899 to 1904. From 1909 to 1921, the microstructures of the rails mentioned are all composed of proeutectoid ferrite networks and pearlite. The pearlite content is significantly higher than that of rails produced before 1904, and the relative amounts of ferrite and pearlite correlate directly with the carbon content of the samples.

Quantitative metallographic analysis reveals the exact pearlite volume fraction ($V_p$) for each of the ten specimens: 15.2% ± 1.8 (1899), 18.6 ± 2.1% (1901), 22.4 ± 1.5% (1902), 28.1 ± 2.0% (1904), 65.3 ± 1.9% (1909), 69.8 ± 2.3% (1911), 73.5 ± 1.7% (1914), 76.2 ± 2.0% (1917), 79.4 ± 1.6% (1919), and 82.1 ± 1.8% (1921). Overall, the average volume fraction of pearlite in the pre-transition rails (1899–1904) ranged from 15% to 28%, exhibiting a linear growth relationship with the low carbon content. Following the open-hearth transition (1909–1921), the pearlite proportion expanded substantially to 65–82%. The ‘unclosed pores’ found in the early rails represent gas pores and solidification shrinkage cavities; due to the primitive rolling capacity and low rolling reduction ratio of the initial mill equipment, these internal casting defects could not be dynamically pressure-welded during hot working. In the subsequent medium-carbon hypoeutectoid rails, the emergence of ‘proeutectoid network ferrite’ originated from the slow cooling after rolling. Ferrite preferentially nucleated at and continuously extended along the high-energy prior-austenite grain boundaries, thereby encapsulating the remaining austenite, which subsequently transformed into pearlite colonies.

3.3. Analysis of Inclusions

Inclusions in rail steel directly affect the mechanical and service performance of the rails. The inclusion rating results for the Hanyang Iron Works rails are shown in

Table 2. Chart of non-metallic inclusion ratings for steel rails manufactured by Hanyang Iron Works.

Year	Inclusions
	Type A (Sulfides)		Type B (Alumina)		Type C (Silicates)		Type D (Globular Oxides)
	Fine	Coarse	Fine	Coarse	Fine	Coarse	Fine	Coarse
1899	2	2.5	0	0	1.5	2	1	1
1902	1.5	3	0	0	2.5	3	2.5	1.5
1904	2.5	2	0	0	2	1.5	2	1
1909	2.5	3	2	1	0	0	1.5	1
1911	3.5	2	2	2	0	0	2	0.5
1913	3.5	3.5	2	0.5	0	0	1	0.5
1914	3	2.5	2	2	0	0	1.5	1
1917	2.5	3	2.5	2.5	0	0	1.5	0.5
1918	3	2.5	3	2.5	0	0	3	1
1921	2.5	1.5	1.5	1.5	0	0	1.5	0.5

. The test results show that all Hanyang Iron Works rails contain significant levels of Category A (sulfide) inclusions. There were almost no Category B (alumina) inclusions in the rails produced before the technical transformation (1899, 1902, and 1904). However, after the technical transformation (post-1909), significant Category B inclusions appeared, a characteristic further confirmed by subsequent EDS analysis. Regarding Category C (silicate) inclusions, the content was high in the rails produced before the technical transformation (both fine and coarse series ranged between 1.5 and 3.0). After the technical transformation, there were almost no Category C inclusions in the rails (except for the 1921 sample). For Category D (globular oxide) inclusions, the rails produced after the technical transformation were superior to those produced before it.

3.4. Mechanical Properties

The quality of the rail material is directly reflected by its mechanical properties. A direct correlation is observed between the wear resistance and the hardness of metallic materials. The hardness and tensile properties of the rail samples are presented in

Table 3. Mechanical properties of rails manufactured by Hanyang Iron and Steel Works.

Parameters/Year of Manufacture	Brinell Hardness at Rail Head/HB	Hardness at 20 mm Below Rail Head Surface/HB	Tensile Strength Rm/MPa	Yield Strength Rp0.2/MPa	Elongation After Fracture A/%	Reduction in Area Z/%
1899	244	212	626	416	17	41
1902	241	184	595	406	25	50
1904	235	162	554	398	27	54
1909	243	188	631	429	23	44
1911	250	209	706	491	21.5	33
1913	284	255	827	530	15.5	28
1914	251	205	722	456	17.5	30
1917	254	229	734	469	18.5	33
1918	275	241	801	480	16	27
1921	263	190	624	450	24	39

.

As shown in Figure 6. It is indicated by the hardness testing results that the hardness at the rail head is significantly higher than that at 20 mm below the surface, which is primarily attributed to the work hardening of the rail head during service. The hardness, tensile strength, and yield strength of the rails produced by Hanyang Iron and Steel Works in 1899, 1902, and 1904 were found to be significantly lower than those produced after 1909, whereas the elongation after fracture and reduction in area exhibited an inverse trend. A qualitative leap in the mechanical properties of Hanyang rails was achieved after 1909 compared to the pre-1904 data, which serves as a direct manifestation of the technological advancements in rail manufacturing at the works.

Overall, the fractographic evolution across the years reflects the metallurgical transition: pre-renovation rails (1899–1904) exhibit cleavage facets dominated by unclosed gas pores and silicate inclusions, whereas post-renovation rails (1909–1921) show a shift toward quasi-cleavage accompanied by localized microvoid coalescence due to refined pearlite matrices and clean steel practices.

Furthermore, strain aging may have contributed to the observed fracture features, particularly in the pre-renovation rails (1899–1904). Due to high residual interstitial elements (such as nitrogen and carbon) from early Bessemer refining, long-term strain aging caused interstitial atoms to pin dislocations, increasing matrix lattice resistance and restricting dynamic plastic deformation. On the tensile fracture surfaces, this embrittlement effect manifested as a higher density of transgranular cleavage facets and localized secondary microcracks, accompanied by a noticeable suppression of deep ductile dimple formation.

4. Discussion and Analysis

The Hanyang Iron and Steel Works was one of the earliest integrated iron and steel enterprises established in China, and the performance of its rail products is considered a direct reflection of China’s early metallurgical development and technological progress.

The evolution of the principal chemical elements (carbon, silicon, manganese, phosphorus, and sulfur) of the rails manufactured by Hanyang Iron and Steel Works is illustrated in the radar chart in Figure 7. During the initial phase of the Hanyang Iron and Steel Works (prior to 1904), the Bessemer converter process was employed for rail steel production, which was consistent with the global predominance of Bessemer steelmaking during that period.

In the early stages of its invention, the Bessemer converter underwent rapid development in Europe and the United States due to its effective resolution of iron brittleness and the significant extension of the service life of steel materials. The primary disadvantage of steel produced via the Bessemer process is the elevated concentration of phosphorus (P) and sulfur (S); this phenomenon has been widely reported in the literature and is consistent with the phosphorus content exceeding 0.15% observed in the rail testing data of this study.

It is well-established that the presence of detrimental elements, namely phosphorus (P) and sulfur (S), results in an increase in the brittleness of steel. Sulfur can be sequestered through the addition of manganese to form harmless manganese sulfide (MnS) inclusions; however, phosphorus cannot be removed within an acidic Bessemer converter. Furthermore, the iron ore utilized by Hanyang Iron and Steel Works was sourced from the Daye Iron Mine in Wuhan, Hubei, China, which is characterized by relatively higher concentrations of phosphorus and copper compared to ores from other regions (the high copper content in the ore leads to elevated copper levels in the rail products). The elevated phosphorus content in the early rails produced by Hanyang Iron and Steel Works was attributed to the raw materials, the steelmaking equipment, and the specific processing techniques employed. To mitigate brittleness and ensure sufficient service performance, a reduction in carbon content was implemented in the Bessemer rail steel; this necessitated the low-carbon and high-phosphorus chemical profile characteristic of early Hanyang production.

Between 1904 and 1908, a technical renovation was initiated at the Hanyang Iron and Steel Works, during which the Bessemer process was abandoned in favor of the Siemens-Martin open-hearth process for rail production. Open-hearth furnaces are characterized by large capacities, typically yielding 100–130 tons per heat, and provide superior product quality compared to Bessemer converters. Furthermore, the smelting process can be easily controlled in an open-hearth furnace, allowing for periodic sampling and analysis to ensure precise control of the molten steel composition. Phosphorus removal can also be facilitated within basic slag through the addition of agents such as lime. The completion of this functional renovation at the Hanyang Iron and Steel Works directly resulted in a significant reduction of phosphorus and sulfur content, thereby enhancing the quality of the rail products.

This specific chemical composition combination (0.13–0.22 wt.% C and P ≥ 0.15 wt.%) represented a major metallurgical limitation during the early period of steelmaking in China. Compared with contemporary international rail specifications at the end of the 19th century, such as the British Standards and the American Society of Civil Engineers (ASCE) rail guidelines—which strictly regulated medium-to-high carbon contents (0.40–0.55 wt.%) and restricted the phosphorus content to P ≤ 0.08–0.10 wt.%—the early Hanyang converter rails exhibited a tremendous anomaly. Because the acid converter was thermodynamically incapable of dephosphorization and the local Daye iron ore itself possessed a relatively high phosphorus content, the early steelmakers were forced to suppress the carbon content down to the level of low-carbon structural steel (<0.22 wt.%). This was a deliberate metallurgical compromise aimed at preventing catastrophic cold-shortness embrittlement while maintaining a nominal ductility, albeit at the immense expense of wear resistance and yield strength [20].

To address the historical breakage and processing issues, the explicit physical metallurgy mechanisms of phosphorus, sulfur, and copper on brittleness, hot-shortness, and rail degradation are systematically summarized in

Table 4. Summary of the metallurgical effects of impurity elements (P, S, Cu) on processing defects and historical rail failures.

Element	Primary Metallurgical Mechanism	Induced Defect/Failure Mode	Manifestation in Historical Hanyang Rails
Phosphorus (P) (P ≥ 0.15 wt.%)	Solid-solution strengthening in $\mathsf{\alpha }$-Fe Prior-austenite grain boundary P-segregation	Cold Shortness (Ambient/low-temp brittleness; elevated DBTT)	Catastrophic transgranular cleavage fractures during cold winters on the Beijing-Hankou Railway.
Sulfur (S) (S ≥ 0.086 wt.%)	Formation of low-melting $\mathrm{F}\mathrm{e}-\mathrm{F}\mathrm{e}\mathrm{S}$ eutectic (∼988 °C) Mitigated by Mn to form $\mathrm{M}\mathrm{n}\mathrm{S}$ (>1600 °C)	Hot Shortness/Heavy Inclusions (Grain boundary melting during 1000–1150 °C rolling)	Avoided severe rolling cracks via Mn tie-up, but left dense, elongated Type A sulfide stringers causing fatigue crack initiation.
Copper (Cu) (Cu: 0.15–0.62 wt.%)	Selective oxidation of Fe during billet heating Subscale Cu liquid phase (>1083 °C) penetrating grain boundaries	Liquid Metal Embrittlement (LME) (Copper-induced hot shortness/surface checking)	Microscopic surface cracking, edge tears on rail shoulders, and unclosed micro-defects prior to service.

.

Briefly, phosphorus acts as the primary cause of cold shortness (ambient/low-temperature brittleness) due to grain-boundary P-segregation and intense solid-solution strengthening, explaining the catastrophic winter transgranular cleavage fractures observed on the Beijing-Hankou Railway. High sulfur content ($S\ge 0.086$ wt.%) inherently threatened the steel with $Fe-FeS$-induced hot shortness during rolling (1000–1150 °C). Although a sufficient $Mn/S$ ratio ($>7–10$) effectively mitigated rolling cracks by forming high-melting $MnS$, the resulting heavily elongated Type A sulfide stringers severely compromised transverse toughness. Additionally, high copper content (0.15–0.62 wt.%) originating from the Daye iron ore induced localized liquid metal embrittlement (subscale Cu-rich liquid phase penetration along grain boundaries at temperatures >1083 °C), generating surface checking and micro-cracks during deformation. The technical transition to basic open-hearth steelmaking after 1908 successfully reduced phosphorus and improved cleanliness, fundamentally resolving these failure drivers.

From the perspective of microstructure, a large amount of ferrite and pearlite was observed in the steel rails manufactured by Hanyang Iron and Steel Works prior to the technical transformation (before 1904). Following the completion of the technical transformation (after 1908), the microstructure was transformed into a combination of proeutectoid network ferrite and pearlite. Furthermore, a significant increase in pearlite content was observed within the microstructure after the technical transformation. This characteristic is correlated with the variations in carbon content within the steel rails produced by Hanyang Iron and Steel Works. However, the direct cause is attributed to the revolution in steelmaking technology implemented at the Hanyang plant. Open-hearth steelmaking was adopted after the technical transformation, resulting in a reduction of phosphorus content in the produced rails. The strength could be enhanced by increasing the carbon content without compromising the toughness of the rails. As the carbon content in the rails was increased, the volume fraction of pearlite within the microstructure was elevated accordingly.

The formation of the network ferrite and pearlite microstructure is primarily attributed to the relatively slow cooling rate of the rail steel during the post-rolling cooling process. Due to the high interfacial energy and rapid atomic diffusion at the prior austenite grain boundaries, proeutectoid ferrite is preferentially nucleated at these boundaries. Owing to the low solubility of carbon in ferrite, the remaining carbon atoms are rejected into the adjacent austenite. The ferrite subsequently grows rapidly along the austenite grain boundaries, interconnecting to form a ferritic network. The austenite situated within the ferritic network is gradually transformed into pearlite. The microstructural evolution of the Hanyang Iron Works rails serves as the most direct and compelling scientific evidence of its technological advancement and improved product quality.

The inclusion rating results of the early rails produced by Hanyang Iron Works demonstrate the influence of the steel smelting process. As previously mentioned, the steelmaking processes employed for rail production before and after the renovation of Hanyang Iron Works were entirely different. Production was carried out using an acidic Bessemer converter prior to the renovation, whereas basic open-hearth steelmaking was adopted upon completion of the upgrade. Although elements such as carbon and silicon in pig iron can be rapidly oxidized by air blowing in a Bessemer converter, a serious issue of increased dissolved oxygen content in the steel is introduced. Deoxidation is not inherently integrated within the Bessemer converter process; thus, deoxidation is typically performed by adding ferromanganese or ferrosilicon during subsequent processing stages to remove excess oxygen.

The post-treatment process involves the rapid addition of deoxidizers, such as ferromanganese and ferrosilicon, to the molten steel after it has been poured into the ladle following the blow. This deoxidation method is prone to the formation of liquid or semi-liquid silicates, such as FeO·SiO₂, MnO·SiO₂, MnS·SiO₂. These silicates tend to remain trapped within the steel during the flotation removal process, thereby forming detrimental inclusions. Consequently, the early rails from Hanyang Iron Works exhibit a low concentration of Type B inclusions (aluminas) but a high content of Type C inclusions (silicates). Energy-dispersive X-ray spectroscopy (EDS) analysis of the rails produced by Hanyang Iron Works in 1899 and 1902 is illustrated in Figure 8. The EDS results indicate that the early rails primarily contain complex inclusions such as MnS and MnS·SiO₂, whereas alumina inclusions were not detected.

Following the technical renovation of Hanyang Iron Works, Siemens-Martin open-hearth furnaces were employed for steelmaking, producing rails in which Category A and Category B inclusions predominated. Figure 9, the energy-dispersive X-ray spectroscopy (EDS) analysis of the rails manufactured by Hanyang Iron Works in 1911, 1918, and 1921. The EDS results indicate that the inclusions within the rails are primarily composed of MnS and complex MnS·Al₂O₃ phases. The post-renovation increase in Category B inclusions and the near absence of Category C inclusions demonstrate the advancements in the deoxidation process of Siemens-Martin steelmaking. It can be inferred that aluminum deoxidation refining was adopted in the steelmaking process at Hanyang Iron Works after the renovation. Consequently, the inclusions in the Siemens-Martin steel after the upgrade were dominated by Category A (sulfides) and Category B (aluminas). The evolution of inclusion types in Hanyang rails reflects the progress in the early development of steelmaking technology in China.

The unique appearance of Type B (alumina) inclusions and composite $\mathrm{MnS}·{\mathrm{Al}}_2{\mathrm{O}}_3$ phases after 1909 directly marked a fundamental shift in the deoxidation practice. In the basic open-hearth process, metallic aluminum was introduced into the ladle as a potent deoxidizer, converting dissolved oxygen into highly stable, solid ${\mathrm{Al}}_2{\mathrm{O}}_3$ clusters rather than the fluid silicate slag typical of the converter process. However, when benchmarked against modern standards such as GB/T 2585-2021 [20] (in which the fine/heavy series of all inclusion types are strictly limited within 1.5), the historical Hanyang rails clearly exhibited a lower cleanliness. Their heavy series Type A and Type B inclusions reached a severe level of 3.5 (Table 2). Compared with modern, highly refined clean steels, these high inclusion ratings constituted a major source of microstructural stress concentration, severely compromising the structural integrity of the material [21].

Stress Concentration and Anisotropy: Elongated $\mathrm{MnS}$ stringers (Acid Bessemer era) exhibit high aspect ratios ($a/b\gg 10$), acting as severe internal stress raisers ($K_t\gg 3$) that trigger early matrix/inclusion decohesion and micro-delamination along the rolling direction. In contrast, spherical $\mathrm{MnS}·{\mathrm{Al}}_2{\mathrm{O}}_3$ inclusions (Basic Open-Hearth era) maintain a low stress concentration factor ($K_t\approx 3.0$), suppressing premature cracking and allowing the steel matrix to reach its intrinsic yield strength and uniform strain capacity.

Fracture Path and Microvoid Nucleation: Coarse silicate droplets and aligned sulfide bands in early rails promote rapid microvoid coalescence along the inclusion stringers at low strain levels, leading to quasi-cleavage transgranular fracture with negligible necking. Conversely, fine globular $\mathrm{MnS}·{\mathrm{Al}}_2{\mathrm{O}}_3$ inclusions act as stable sites for controlled void nucleation, delaying microvoid coalescence until high plastic strains and yielding a ductile dimple fracture surface with improved reduction in area ($\mathrm{Z}\%$).

Strengthening mechanisms for steel materials primarily include solid solution strengthening, grain refinement, precipitation hardening, and strain hardening. As indicated in Table 3, the mechanical properties of rails produced before the renovation (1899–1904) were generally low, whereas those manufactured after the upgrade (1909–1921) were effectively enhanced. This transition was not determined by a single factor, but rather resulted from the synergy of chemical composition optimization, microstructural refinement, and inclusion purification. The core of this evolution was the transformation from high-phosphorus, low-carbon brittle ferritic steel before the renovation to toughened medium-carbon, phosphorus-controlled pearlitic-ferritic steel.

The evolution of mechanical properties across the historical rails directly reflects the quantitative shift in their microstructural constituents. Early rails (1899–1904) with low carbon contents possessed a soft ferrite-dominated matrix ($>70\mathrm{vol}\%$), exhibiting low strength ($R_m:554–626\mathrm{MPa}$) but superior ductility. With the steelmaking transition after 1908, the increased carbon content promoted a heavily pearlitic matrix ($65\%–82\mathrm{vol}\%$) bounded by thin intergranular ferrite networks. The fine lamellar spacing and dense cementite plates within the dominant pearlite colonies significantly enhanced dislocation barriers, elevating the ultimate tensile strength up to $827\mathrm{MPa}$. Concurrently, the reduced volume of soft ferrite and restricted plastic flow along grain boundary networks led to a moderate decrease in elongation and reduction in area. This fundamental shift from a ductile ferrite-dominated to a strength-bearing pearlite-dominated constituent dictates the macro-mechanical performance trend of the historic Hanyang rails.

Prior to the renovation, the microstructure consisted of predominant ferrite and pearlite, whereas the post-renovation microstructure transitioned to a network ferrite-pearlite constituent with pearlite as the primary phase. In steel, the ferritic phase is characterized by softness and ductility, while the pearlite phase exhibits higher hardness and strength. When the microstructure is dominated by soft, ductile ferrite, the overall strength, hardness, and wear resistance of the rail are insufficient. Phosphorus is a highly deleterious element that undergoes intense solid solution in ferrite, significantly elevating the ductile-to-brittle transition temperature and making the rails prone to fracture at ambient temperatures. The abundance of Category C inclusions serves as “intrinsic cracks” within the rail, severely segmenting the continuity of the matrix. These factors represent the microstructural origins of the “brittleness and susceptibility to fracture” reported in historical literature regarding early Hanyang rails.

Following the renovation of Hanyang Iron Works, the mechanical properties of the rails reached the applicable standards for rail steel of that era. The tensile strength was increased to over 700 MPa, reaching the 800 MPa grade in some cases, accompanied by a significant elevation in hardness. Although there was a decline in plasticity and toughness indices, this is considered a typical phenomenon following an increase in strength. Both the absolute values and the balance of strength and toughness met the operational requirements for trains at that time.

The primary reasons for the performance enhancement of the rails after the completion of the technical renovation are as follows: (1) The increase in carbon content directly strengthened the matrix and led to a substantial rise in the volume fraction of pearlite within the microstructure. (2) With the implementation of the basic open-hearth furnace, phosphorus content was effectively controlled, allowing for strength enhancement via increased carbon addition. (3) The refinement of the deoxidation process led to an effective improvement in the characteristics of non-metallic inclusions within the rails. The aforementioned factors ultimately culminated in the improvement of the mechanical properties of the rails produced by Hanyang Iron Works.

The typical fracture morphologies of the tensile specimens before (pre-1904) and after the technical renovation of Hanyang Iron Works are shown in Figure 10. Numerous voids resulting from the removal of inclusions are visible within the fracture surfaces of the specimens. Observation of the tensile fracture surfaces reveals distinct river patterns, dimple rupture, and secondary cracking in all specimens. Furthermore, a small amount of intergranular “rock-candy” fracture morphology is observable in the specimens produced before the technical renovation. The cleavage fracture zones in the specimens from before the renovation (1899 and 1902) are less extensive than those in the post-renovation specimens. The dimples in the pre-renovation specimens are characterized as large and shallow, whereas those in the post-renovation specimens are small and shallow. Additionally, variations in the frequency of secondary cracks within the fracture surfaces are observed between the pre- and post-renovation periods. The number of secondary cracks in the pre-renovation specimens is significantly higher than that observed in the specimens produced after the renovation.

Quantitative statistical analysis of the high-magnification SEM fractographs (Figure 10) further confirms the microstructural transition. For the pre-renovation rails (1899–1904), the fracture surfaces display non-uniform, large dimples with an average equivalent diameter of $12.4\pm 3.8\mathsf{\mu }\mathrm{m}$ (ranging from $5.2$ to $24.1\mathsf{\mu }\mathrm{m}$), accompanied by coarse cleavage steps. In contrast, the post-renovation rails (1909–1921) feature significantly refined and homogeneous ductile dimples with a mean diameter of $3.6\pm 1.1\mathsf{\mu }\mathrm{m}$ (ranging from $1.5$ to $6.8\mathsf{\mu }\mathrm{m}$) alongside shallow quasi-cleavage facets, reflecting improved strain localization resistance during tensile deformation.

To correctly evaluate the mechanical property baseline, it is essential to decouple the interaction between the as-manufactured properties and long-term in-service degradation. Since these rails experienced decades of operational loading and environmental exposure prior to their recovery, cyclic fatigue damage and plastic flow inevitably occurred. However, as substantiated by the hardness distribution in Table 3, these deformation artifacts and macroscopic fatigue defects were strictly confined to the wheel-rail contact surface. By utilizing core specimens rigorously extracted from the interior core of the rail head (Figure 3), the original bulk mechanical properties (yield strength and microstructure) from the time of manufacturing were successfully isolated from the surface cyclic damage. The environmental corrosion crust was also completely mechanically removed during the bulk cutting process. The only systematic service artifact is the long-term continuous strain aging induced by historical residual interstitial nitrogen/carbon atoms, which may have led to a slight increase in the baseline yield strength ($R_{\mathrm{p}0.2}$) and a minor decrease in elongation over the past century—a common feature of classical metallurgical relics [22].

To visually demonstrate the macroscopic trajectory of mechanical property evolution across different production periods, Figure 11 constructs an Ashby-type plot mapping the strength-ductility trade-off (yield strength versus reduction in area). This plot clearly distinguishes the early Thomas/acid converter rails—characterized by low strength ($R_{\mathrm{p}0.2}$ < 420 MPa) but relatively high nominal reduction in area ($Z$ > 40%)—from the basic open-hearth pearlitic rail region, the latter shifting toward a high-strength, controlled-ductility regime ($R_{\mathrm{p}0.2}$ up to 530 MPa, with $Z$ ranging between 25% and 45%). This graphical distribution intuitively illustrates how the smelting process upgrade effectively broke through the performance bottleneck induced by high-phosphorus embrittlement, establishing an excellent strength-ductility balance adapted to the heavier axle load demands of that era.

To comprehensively gauge the metallurgical status of historical Hanyang rails within the broader timeline of industrial rail steel development, a direct comparison with modern high-performance rail steels is indispensable. Modern standard rails, such as the contemporary pearlitic U71Mn steel grade used for spectrometer calibration in this study, typically possess highly refined chemical composition specifications (0.71–0.80 wt.% C, 0.15–0.58 wt.% Si, 0.70–1.20 wt.% Mn) and extremely low impurity limits (P ≤ 0.030 wt.%, S ≤ 0.025 wt.%). Owing to the ultra-high chemical cleanliness and highly optimized pearlite interlamellar spacing, modern rails can stably achieve a tensile strength exceeding 880–1180 MPa while maintaining high microstructural homogeneity.

By comparison, although the later basic open-hearth steel rails (1913–1918) from the Hanyang Iron Works successfully elevated the tensile strength to the 800 MPa level through a medium-carbon design, their hardness and strength remained inferior to those of modern counterparts [23,24]. The root cause of this gap lies in their higher residual phosphorus content (P approx. 0.04–0.10 wt.%) as well as the higher ratings of elongated Type A and Type B inclusions, which acted as microstructural stress concentration sources or localized crack initiation sites during tensile loading. This evolutionary trajectory profoundly reflects the major transition of rail manufacturing from the early structural steel metallurgy stage, which was limited by raw material impurities and primitive atmospheric deoxidation technology, to the modern clean steel metallurgy stage dominated by advanced secondary refining and grain boundary engineering [25].

In conclusion, this work provides a unique and vital archaeological metallurgical contribution to the global history of iron and steel development. Although the historical metallurgical trajectories of the European and North American rail industries have been extensively documented, the systematic evaluation of ten representative Hanyang Iron Works rails (1899–1921) fills a major regional scientific gap in this field for early Chinese heavy industry. It offers quantitative scientific insights, revealing how western-imported steelmaking technologies adapted to local raw material limitations—specifically the high-phosphorus and copper-rich iron ores of the Daye Iron Mine. The precise documentation of the metallurgical leap from a brittle converter ferritic microstructure to a toughened open-hearth pearlitic-ferritic microstructure constitutes an excellent scientific benchmark, bridging industrial history with modern secondary refining concepts and providing valuable methodological references for the scientific conservation of global industrial heritage.

5. Conclusions

(1)

Regarding chemical composition, the steel rails produced prior to the technical renovation (before 1904) were characterized by a low-carbon and high-phosphorus chemistry (C ≤ 0.22 wt.%, P ≥ 0.15 wt.%). Following the technical renovation, this shifted toward a medium-carbon and low-phosphorus profile (C ≥ 0.50 wt.%, P ≤ 0.10 wt.%). The microstructure evolved from a predominant ferrite–pearlite matrix before the renovation to a constituent composed of minor network ferrite and pearlite.

(2)

Based on inclusion rating and EDS analysis, the inclusions in Hanyang rails before the renovation were primarily classified as Category A, C, and D. After the completion of the renovation, the inclusions in the produced rails consisted mainly of Category A, B, and D. The inclusion types transitioned from highly deleterious silicates (Category C) to relatively harmless aluminas (Category B).

(3)

The tensile strength of Hanyang rails was generally enhanced from below 600 MPa in the early stage to over 800 MPa. The hardness and strength levels were substantially improved. The fracture morphology of the tensile specimens also transitioned from large, shallow dimples with minor cleavage to small, shallow dimples with extensive cleavage fracture.

(4)

This study bridges industrial history and metallurgy. Scientifically, it demonstrates how early East Asian steelmaking resolved raw material constraints (high-P/Cu Daye ores) via the Bessemer-to-open-hearth process transition. Heritage-wise, establishing the processing–structure–property links provides a technical benchmark for the failure analysis and preservation of global steel heritage artifacts.