2023 Volume 64 Issue 9 Pages 2296-2301
A primary melted mark (PMM) on copper wire has been expected to identify the fire origin. However, the solidification microstructure under various cooling conditions has not been extensively studied. In this research, the effects of cooling conditions on PMM microstructure before and after heat treatment were investigated. The results represent dendrite growth direction to the surface of the melted mark with copper dendrites surrounded by (Cu+Cu2O) eutectic structure for air and water-cooling specimens. Moreover, the melted mark cooled in the air has more than twice of secondary dendrite arm spacing (SDAS) than the water-cooling sample in the case of pre-heat treatment. After heat treatment and cooled down in the furnace, air, and water: the dendritic structure disappears and Cu2O precipitates on the copper matrix. The smaller crystallite size and oxide layer cracking could be found in a higher cooling rate case. Therefore, by connecting thermal history with solidification structure and surface oxidation morphology, the change in microstructures and physical parameters of PMM could be expected to be helpful for fire investigation.
Copper wire with oxygen between 0.01 and 0.05% generally contained has been extensively used for electrical applications because of its excellent electrical conductivity.1,2) However, the solidification structure of melted marks on copper wire generated by arcing of electrical current in a fire accident is still insufficient to describe the fire behavior in order to determine the origin of the fire accident. The critical evidence used to identify the cause of the fire for the fire investigator is the “melted mark” on the copper wire that is not burned and is left behind at the fire scene.3) Generally, various melted marks are found in the fire site; primary melted mark (PMM), secondary melted mark (SMM), and fire melted mark (FMM). PMM is an electrical melted mark caused the fire, generated before the fire. SMM is an electrical melted mark created after a fire by the burning of PVC insulation.4–6) FMM is a thermal damage caused by fire melting. Lee et al.4,7) reported that PMM is generated in an atmosphere at or near room temperature and rapidly cooled, whereas SMM is generated in the high temperature of the fire atmosphere. Therefore, the cooling rate of PMM should be much higher than SMM. As a result, they proposed that it is possible to distinguish between PMM and SMM by measuring the oxygen concentration and dendrite arm spacing of melted marks on copper wire. The solidification structure of melted marks on copper simulated by short-circuiting in the atmosphere was continuously discussed by Liu et al.8) They presented that the Cu-dendrites and (Cu+Cu2O) eutectic structure under the cuprous oxide (Cu2O) surface layer is the fingerprint of PMM.
However, the solidification microstructure and post-heat treatment characteristics of PMM under different cooling conditions have not been extensively studied. Therefore, the objective of this study is to explore the characteristic features of PMM due to cooling type effects both pre- and post-heat treatment by observing the microstructure, dendrite growth direction, and investigating the cooling condition effects on oxygen concentration for the melted mark. Each microstructure was evaluated from a metallurgical point of view for the judgment of fire cases.
Commercial pure copper wires with a diameter of 2.6 mm and a length of 25 mm were used in this study. For arc simulation, an arc welding machine has been used to melt the copper wire without PVC insulation. The distance between two electrodes was fixed at one centimeter, and a voltage of 100 Volt was applied to all samples. The copper wire was rapidly melted and suddenly solidified in air (A) and water (W) (Fig. 1(a)). These two PMMs were examined for the pre-heat treatment condition. To simulate the fire environment and investigate post-treatment characteristics, the thermocouple was set under the copper wire at 2 mm from the end of the copper wire. PMMs solidified in the air were annealed in the furnace at 1000°C for 10 min, and finally cooled down to room temperature through various cooling types (Furnace, Air, and Water), as shown in Fig. 1(b).
Schematic diagram of various cooling types for melted mark on copper wire: (a) solidification process, and (b) heat treatment process.
The microstructure in the longitudinal section was then observed by first grinding with SiC paper from #500 to #4000 and secondly mirror-polishing by 0.1 um of diamond paste. Before microstructural observation with an optical microscope, the aqueous ferric chloride (120 ml distilled water, 30 ml hydrochloric acid, and 10 g ferric chloride) was used to etch the surface of the melted marks for a few minutes.9–11) The phase identification of the oxide layer for heat treatment specimens was done by X-ray diffraction (Rigaku, RINT-2100), under 40 kV/4 mA with CuKα radiation of wavelength λ = 0.1540 nm. Data were taken in the range 2θ of 20 to 80 degrees with increments of 0.02 degrees. The surface morphology and chemical composition of the heat treatment specimens were observed by SEM (Hitachi, SU3500) attached EDS.
This section describes the influence of cooling rate on dendrite growth in the melted marks of the copper wire before heat treatment. The micrographs in the longitudinal section of the melted marks on the copper wire solidified in air and water are illustrated in Figs. 2(a)–(d) and Figs. 2(e)–(h), respectively. Both specimens represent the microstructure of the non-melting region that is different from the melting region with a distinct boundary at the left side. The grain boundaries and annealing twins are observed at the non-melting zone, while the melting zone is a dendritic structure. Moreover, gas porosities are observed only in the melting zone; it was thought that gas porosities were formed during the solidification process in which oxygen and hydrogen have been dissolved into liquid copper by air or water.
The longitudinal section of melted mark solidified: (a)–(d) in air, and (e)–(h) by water where (b), (f) at close to non-melting zone, (c), (g) in the middle, and (d), (h) at close to surface of melted mark.
The dendrites begin to grow at an interface between the non-melting and melting regions, as shown in Figs. 2(b) and 2(f). It means that during solidification when the liquid copper temperature is lower than the melting point of 1083°C,12,13) the copper grains are nucleated at the solid-liquid interface of the copper wire like an epitaxial growth. Then these grow towards the interface into the surface of the melted mark to form columnar crystals, as shown in Figs. 2(b)–(d) for air-cooling and Figs. 2(f)–(h) for water cooling. Clear dendrite growth is observed throughout the specimens, but the size of dendrite is relatively smaller at the surface of melted mark of both cooling conditions. Although the dendrite growth direction for the water-cooling specimen should be growing from the surface to inside the melted mark, the liquid copper rapidly solidifies, and the dendrites quickly grow to the surface. It is because the heat transfer occurs along the long solid copper wire and the melted mark has a minimal volume. Therefore, the dendrite growth direction from the non-melting region to the surface of the melted mark in the direction opposite the heat flow becomes the predominant growth direction for both cooling types. In addition, the growth direction of dendrites for the air-cooling specimen is consistent with short-circuiting in the 110 Volt experiment of copper wire at an ambient atmosphere described by Liu et al.8,14)
3.2 Solidification microstructure characterizationFigure 3 shows the typical microstructures of the melted mark in air and water cooling. The microstructure in most regions of both specimens comprised primary copper dendrites in a bright region surrounded by a copper-oxygen eutectic structure in a dark region. It can be described according to the Cu–O phase diagram.2) When the liquid copper temperature is lower than 1083°C, the copper atoms firstly crystallize on the non-melting zones. They solidify into primary copper dendrites and grow rapidly from the interface to the surface of the melted mark to form columnar dendrites along the preferred crystallographic direction. During dendrite growth, the oxygen atoms are piled-up from the dendrite tips into the liquid and progressively create richer solutions to form eutectic composition solutions. When the temperature is below the eutectic temperature (1065°C), the liquid of eutectic composition solidifies into a (Cu+Cu2O) eutectic structure. Therefore, the final microstructure of the melted mark on copper wire under both cooling types has two constituents of primary copper dendrites and (Cu+Cu2O) eutectic structure. Although the constituents in the microstructure of air- and water-cooling specimens are similar, there is a noticeable difference in dendrite arm spacing. The secondary dendrite arm spacing (SDAS) near the surface of the air- and water-cooling specimens are 12.0 µm and 4.7 µm, respectively. The water-cooling specimen has the SDAS less than the air-cooling specimen because of the faster cooling rate in the water. When the cooling rate is high, the size of diffusion field around the dendrite in the liquid phase reduces, resulting in smaller SDAS in the water-cooling specimen. Therefore, the difference in SDAS between high and low cooling rates can describe the fire behaviors after short-circuiting.
Microstructures of the melted marks solidified: (a) in air, (b) by water.
From the solidification process of the Cu–Cu2O binary alloy, when the oxygen concentration in the liquid phase is 0.39 mass% or less, copper is solidified first, and solidification continues while increasing the oxygen concentration in the liquid phase. When the oxygen concentration in the remaining liquid phase portion reaches 0.39 mass%, the solidification is complete, and a eutectic structure is formed. As a result, the oxygen concentration in the Cu primary dendrite part is 0.08 mass%, while it is 0.39 mass% in the eutectic structure.7) Accordingly, the oxygen concentration can be calculated from the area ratio method by the following eq. (1)
\begin{equation} \text{O}\ (\text{mass%}) = 0.39 \times (\text{A} - \text{A}_{\text{D}})/\text{A} \end{equation} | (1) |
The relationship between oxygen concentration and distance to surface for air and water-cooling melted marks.
An optical microscope was used to examine the microstructures of heat treatment as shown in Fig. 1(b) specimens. Figure 5 depicts the microstructures of the melted mark before heat treatment (a) and after heat treatment with furnace cooling (b), (c), air cooling (d), and water cooling (e). A significant change in microstructure is clearly observed after annealing. Namely, the microstructure before heat treatment comprises the primary copper dendrites in a bright region and the (Cu+Cu2O) eutectic phase (Fig. 5(a)) as described in section 3.2. After heat treatment, the dendrites disappear entirely (Figs. 5(b)–(e)). This is because the copper dendrites in the same crystal orientation have to reduce their interfacial energy by dissolving and merging together to finally become a larger grain as a well-known ripening phenomenon.16) At the same time, the eutectic structure is changed into Cu2O particles precipitated on the Cu matrix all over the surface. In Fig. 6, the distribution of individual elements at center of the melting region of the annealed specimen of Fig. 5(b) was determined by EDS analysis. The obtained images show that copper is uniformly distributed on the surface while the black point is enriched by oxygen. The oxygen-rich point was identified by EDS as precipitated Cu2O. According to the Cu–O phase diagram,2) when the copper-oxygen is reheated below the eutectic temperature at Cu+Cu2O phase region, the diffusion rate of Cu2O is extremely fast resulting in the precipitation of Cu2O over the entire copper matrix.17) Furthermore, we reported the presence of Cu2O precipitates in the microstructure of real-world fire samples after exposure to fire.18)
The microstructures of melted mark: (a) before heat treatment, (b)–(e) after heat treatment then cooled in: (c) furnace, (d) air, and (e) water.
EDS mapping of composition of annealed specimen.
The microstructure of annealed specimens cooled under various cooing rates is not significantly different. The most constituent microstructure consists of Cu2O particles precipitated on a copper matrix as noticed in the small black points (Figs. 5(c)–(e)). The large black points are gas porosities generated during the melting process before heat treatment. The different colors in the bright and dark regions in the air- and water-cooled specimens display the difference in crystal orientation for each grain. The dendritic structure cannot be observed for annealed specimens.
3.5 Surface morphology after heat treatmentHeat treatment of melted marks on the copper wire was conducted at 1000°C for 10 min and cooled in the furnace, air, and water to simulate the fire environments. The surface morphologies for each cooling condition were investigated by SEM, as shown in Fig. 7. The oxide layer on the outer surface of the melted mark of the water-cooling specimen in Fig. 7(c) was almost completely cracked. In contrast, furnace and air-cooling specimens partially delaminated, as illustrated in Figs. 7(a) and 7(b). It can be described in terms of the difference in thermal expansion. As the oxide layer forms during heat treatment at high temperatures, strain-free oxide growth is assumed because of high diffusivity and high atomic mobility. When the heat treatment is complete, the strain is introduced into the copper and copper oxide layer as the specimen cools down. Suppose the cooling rate of the water-cooling specimen is too high. In that case, the ability of the crystallites to rearrange themselves cannot handle this strain, increasing the possibility of cracking of the oxide layer leaving the outer surface of the melted mark.19)
Surface morphology taken by SEM on melted marks cooled: (a) in furnace, (b) in air, (c) by water. Figure (d) shows part of a crack in the oxide layer of Fig. (c).
The XRD patterns of the copper oxide layer prepared under heat treatment at 1000°C for 10 mins and cooled down in the furnace, air and, water are demonstrated in Fig. 8. The peak positions at 2θ values of 29.42, 36.29, 42.16, 61.27, 73.40, and 77.27 degrees correspond to the (110), (111), (200), (220), (311), and (220) planes of Cu2O. These peaks are in good agreement with Cu2O powder from JCPDS file no. 00-002-1067 confirming the presence of Cu2O for all cooling types. The oxidation can be explained as copper atoms reacting with oxygen atoms to form Cu2O, referred by the following reaction:
\begin{equation} \text{2Cu} + \text{1/2O$_{2}$} \to \text{Cu$_{2}$O}{}^{\text{20)}} \end{equation} | (2) |
X-ray diffraction patterns of copper oxide on melted mark of copper wire after heat treatment for different cooling conditions.
At high heat treatment temperatures, the more reaction between copper and oxygen is produced, and the Cu2O grows predominant with increasing Cu2O layer thickness. Moreover, the broadened peaks in the XRD patterns have been used to estimate the crystallite size of Cu2O by using the Debye-Scherrer formula21–23)
\begin{equation} D = 0.9\lambda/\beta \cos \theta \end{equation} | (3) |
The plot of cooling rate versus Cu2O crystallite size.
Although the different cooling rates may not have much effect on the structure after heat treatment as discussed in the previous section, it also affects oxide cracking and the crystallite size of the oxide. Furthermore, the disappearance of dendrites is a crucial characteristic of the annealed specimens, indicating that the melted mark on copper wire has been in a fire environment for a certain time and temperature. Quantitatively, the SDAS, oxygen concentration, and crystallite size were obtained for each condition. The melted mark cooled in the air has a larger SDAS than the water-cooling sample in the case of pre-heat treatment. In terms of oxygen concentration, it is difficult to distinguish between air and water-cooling samples because it seems the difference in oxygen concentrations between both cooling types remains sufficiently unclear. This work has not investigated quantitative relationships between SDAS and oxygen concentration. However, Lee et al. discovered that oxygen concentration has an effect on SDAS, with high oxygen content decreasing SDAS. Post-heat treatment, the furnace-cooling sample has the largest crystallite size of 50 nm. That is, the presence of large crystallites indicates that the melted mark has been exposed to the fire and has cooled in the ambient fire.
The characteristic features of PMM due to cooling type effects both pre and post-heat treatment were investigated. The following results are obtained:
This work was supported by the Royal Thai Government scholarship.