Correlation of structure development and electric characteristics within high pressure torsion processed copper

Copper of a high purity features excellent electric conductivity, but generally very low mechanical properties. Nevertheless, optimized deformation/thermomechanical treatment can introduce favorable combinations of both. The presented study characterizes the correlation of microstructure development and electric properties within copper processed by the severe plastic deformation method of high pressure torsion, the primary advantage of which is that it enables to achieve grains with the sizes in the ultra‐fine, or even nano scales. The study investigates structure development during progressive deformation. In other words, samples processed by single and double high pressure torsion revolutions were evaluated from the viewpoints of grain sizes and grain boundaries, and the results were correlated with the experimentally measured electric conductivity. The single high pressure torsion revolution contributed to grain size decrease, while the structure after double revolution exhibited very fine grains, especially at the sample periphery featuring the highest imposed strain. Both the samples also exhibited increases in microhardness (especially after double revolution), and electric conductivity higher than 100 % IACS. The results confirmed that copper conductors featuring enhanced mechanical properties and favorable electric conductivity can be manufactured by severe plastic deformation.


| INTRODUCTION
Copper of a high purity, the superior conductivity for which is typically observed, is preferably produced by electrolytic refining, as it enables to acquire the purity higher than 99.99 % (the presence of impurities, primarily silicon, iron, arsenic, and oxygen, within the structure tends to deteriorate the electric conductivity as it aggravates the movement of free electrons) [1].Such a highly pure copper features the yield strength of only 10 MPa to 20 MPa.Therefore, additional elements are usually added to introduce strengthening and increase the mechanical properties (usually at the expense of the electric conductivity).The most advantageous ways how to improve the mechanical properties and performance of copper while maintaining favorable electric conductivity are: i) to add minor amounts of specific alloying elements, which do not seriously deteriorate the conductivity (e. g. adding small amounts of argentum or cadmium to enhance the performance and properties of copper at elevated temperatures) [2].ii) to manufacture copper-based composites with additions of strengthening elements (e. g. powder-based copper composites with chromium, alumina, or carbon and its nanotubes, copper clad composites with aluminum, magnesium and aluminum, niobium, high entropy alloys, etc. [3][4][5][6][7][8].iii) to use optimized thermomechanical (deformation) treatment (e. g. via rotary swaging, or methods of severe plastic deformation) [9][10][11][12].The mentioned methods of plastic deformation can favorably strengthen the processed materials as they impart the generation of dislocations, which subsequently form dislocation cells and walls and consequently subgrains (typically defined with low angle grain boundaries), which finally develop into individual grains (typically defined with high angle grain boundaries) [13].For strengthening (and related grain refinement) of severely deformed metallic materials, the spacing between the deformation-induced boundaries and their angles of misorientation are the key controlling parameters [14].However, the ratio of the individual contributions of dislocation strengthening by the presence of low angle grain boundaries, and Hall-Petch strengthening by the presence of high angle grain boundaries is strongly dependent on the overall imposed (shear) strain and other structural factors, such as texture, i. e. grains' orientations [15].
Probably the most widely used severe plastic deformation method is equal channel angular pressing and its modifications (e. g. with (partial) back pressure, nonequal channel angular pressing, twist channel (multi)angular pressing, etc. [16][17][18][19][20].The favorable effects of the equal channel angular pressing method on the mechanical properties and structures of commercially pure copper, as well as copper-based alloys were documented by others (e. g. [21][22][23][24]).The primary advantage of the equal channel angular pressing based severe plastic deformation methods is that they can be applied to process bulk samples (i.e. samples of relatively large volumes -as regards the severe plastic deformation methods).Other effective methods of severe and intensive plastic deformation, which can be applied to process samples of relatively larger volumes of materials, and are also industrially applicable, are e.g. equal channel angular pressing-conform, accumulative roll bonding, or rotary swaging [25][26][27][28][29][30].Probably the most effective severe plastic deformation methods, as regards the grain refinement, are the methods of friction stir processing, and high pressure torsion [31][32][33][34].However, they are primarily intended for experimental application due to the limited material volume, which can be processed at a time.
High pressure torsion -probably the first ever invented severe plastic deformation method -is based on deforming a round specimen (dimensions variable) between two anvils via shear strain under a very high pressure [35][36][37].The application of the high shear strain supports the tendency of the processed material to form localized shear bands.Nevertheless, this tendency is compensated by the applied (very) high pressure, which contributes to structure homogenization [38,39].By the effect of these phenomena, high pressure torsion imparts substantial grain fragmentation and substructure development (depending on the total imposed strain).Researchers have investigated the effects of high pressure torsion on commercially pure copper from the viewpoints of (sub)structure development, as well as structure stability after processing (e. g. [40,41]).However, as far as the authors' knowledge reaches, there is a lack of studies focusing on the investigation of the effects of high pressure torsion on the electric properties of commercially pure copper.

| AIM OF THE INVESTIGATION
Based on the above mentioned, the herein presented study focuses on investigating the effects of processing of electro-conductive commercially pure copper by single and double high pressure torsion revolutions at room temperature.After processing, the structures and substructures of the deformed samples were examined and compared to the structure of the original commercially pure copper.The acquired data on grain size, grains' orientations, and morphology were further put in correlation with the Vickers microhardness values measured along the cross-sections of the deformed samples, and experimentally acquired values of the electric conductivity (assessed via the IACS, i. e. international annealed copper standard, values).

| MATERIAL AND EXPERIMENTAL DETAILS
The original material was a commercially available electro-conductive commercially pure copper with the content of impurities of 0.015 wt.% of phosphorus, 0.002 wt.% of oxygen, and 0.002 wt.% of zinc.The original structure was relieved of its deformation history and possible presence of residual stress by applying a heat treatment at 600 °C for 30 min.The heat treated billet was then cut into pieces, i. e. samples for high pressure torsion processing, which were characterized with the diameter of 20 mm and height of 6 mm.The samples were high pressure torsion processed at room temperature under the pressure of 4 GPa and rotation speed of 1 minÀ 1 .One sample was processed by single revolution (i.e. 360°rotation, sample HPT1), whereas another one was processed by double revolution (i.e. 720°rotation, sample HPT2).
The processed samples were at first subjected to scanning electron microscopy analyses using a Tescan FERA 3 device; the electron backscatter diffraction was primarily used for the structure observations.To prepare samples for the electron backscatter diffraction analyses, the high pressure torsion processed samples were transversally cut so that the cross-sectional planes became visible, as in the schematic image of the cut sample with marked locations of the performed structure and microhardness analyses, Figure 1.The cut high pressure torsion processed samples, i. e. their cross-sectional planes, were ground manually and subsequently polished using diamond solutions (coarseness of 3 μm and 1 μm), and finally polished electrolytically.For the analyses, two types of electron backscatter diffraction scans with the areas of 300 μm ×300 μm and 100 μm ×100 μm were acquired in each examined location (scan step of 0.1 μm).The scans with the larger area were used to characterize the grain size and orientations via texture as they captured a larger number of grains to provide relevant results, whereas the scans with the smaller areas were used to characterize the microstructure and individual grains in a greater detail.A larger electron backscatter diffraction scan with the area of 500 μm ×350 μm was acquired to characterize the coarse-grained original copper.For evaluation of the results, the limits of 5°to 15°f or low angle grain boundaries, and > 15°for high angle grain boundaries, were used.
As the HPT2 sample processed by double revolution featured a heavily refined microstructure, additional detailed sub-structure analyses for the HPT2 sample were performed using a JEOL 2000 FX transmission electron microscope.The foils for the analyses were ground manually and then polished electrolytically using a Lec-troPol-5 device by Struers GmbH.
The electric properties of the processed samples were assessed using a SIGMATEST 2.070 device by FOERSTER TECOM s.r.o, Prague, Czech Republic.This high-tech eddy current measuring device is highly convenient as it is portable and enables to measure the electroconductivity of the majority of materials using its probe based on the complex impedance.It is very advantageous especially when determining the electroconductivity of samples characterized with small dimensions, such as the HPTprocessed ones.Firstly, the device had to be calibrated using default specimens with different electric properties, and then it was ready to be used for the intended measurements, i. e. measurements of the electric conductivity of the processed samples.The measured electric conductivity value (both in MS m À 1 and % IACS) was directly shown on the screen of the device.
Last but not least, HV 0.2 Vickers microhardness (i.e. load of 200 g) along two lines across the cross-sections of the processed samples, both along the periphery and axial line, was measured using a Zwick Roell device, Figure 1.The loading time for each indent was 10 s.

| Characterization of grains
Characterization of grains was performed for all the examined materials, i. e. the original copper, as well as both the HPT1 and HPT2 deformed samples.In particular, we observed the size, morphology, and orientations of the grains.Firstly, the original copper was evaluated; the results of the analyses of grains' orientations and fractions of low angle grain boundaries, high angle grain boundaries, and < 111 > 60°twin boundaries were depicted via the orientation image map, Figure 2a.The grain size distribution was evaluated by the parameter of max.feret diameter, Figure 2b.The texture within the original copper was depicted via the inverse pole figure, Figure 2c.As ensues from the results of the grains characterizations, the average grain size within the original material was 37.9 μm.Nevertheless, the size of the largest grains exceeded 200 μm.The grain boundaries were primarily of the high angle grain boundaries type (98.4 %, 65.5 % of which were the < 111 > 60°twin boundaries).The results of the texture analyses confirmed the data acquired based on the orientation image analyses, i. e. that the original copper did not feature any significant presence of preferential texture, Figure 2c.
The orientation image maps and grain size distribution charts for the peripheral and axial regions of the HPT1 sample were evaluated, Figure 3a-d.The structure observations revealed that the average grain size refined substantially after processing via a single revolution at room temperature; when compared to the original grain size value of 39.7 μm, the average grain size at the periphery of the HPT1 sample refined to 7.4 μm, whereas in the axial region of the sample it refined down to 18.0 μm.The grain size distribution was inhomogeneous within the HPT1 sample, Figure 3b-d.The structure featured large unfragmented grains with the diameters exceeding 80 μm in both the scanned regions.On the other hand, the characters, i. e. morphology of the grains, exhibited obvious differences.The structure in the peripheral region of the sample featured fragmented grains with evidently well-developed substructure; this can be observed from different shadings of colors within the individual grains defined with high angle grain boundaries, Figure 3a.Contrary to this, the grains in the axial region of the sample were also fragmented and refined, but featured more or less equiaxed shapes without visible traces of shearing.The fraction of low angle grain boundaries was lower in the axial region, too (compare 21.9 % to 36.9 % of low angle grain boundaries observed within the peripheral region).In other words, the substructure within the axial region of the sample was not so highly developed as at the periphery.Also, the grains in the axial region featured a greater fraction of < 111 > 60°t win boundaries when compared to the peripheral region (42.4 % vs. 23.6 %).
The orientation image maps and grain size distribution charts for the peripheral and axial regions of the HPT2 sample were further evaluated, Figure 4a-d,  e.This sample evidently featured massively refined grains, especially in the peripheral region, the average grain size in which was as low as 0.9 μm.In other words, the microstructure in the peripheral region of the HPT2 sample featured ultra-fine grains.Contrary to the HPT1 sample, the grains orientations within which were more or less random, the grains within the HPT2 sample exhibited the tendency to form the ideal < 110 > j j shear direction texture fiber, Figure 4c.The morphology of the microstructure within the axial region of the HPT2 sample was different compared to the periphery.Although the grains were also greatly refined, the average grain size within the axial region was 2.9 μm.The microstructure primarily consisted of newly forming recrystallized grains, but it still featured traces of larger grains with the sizes of up to ~25 μm as the extent of the imposed strain in this location was not as great as at the periphery of the sample.Nevertheless, the substructure within the larger grains was highly developed, the fraction of low angle grain boundaries in the axial region of the HPT2 sample was 61.2 %, Figure 4d.The microstructures in both the examined regions also still exhibited certain portions of < 111 > 60°twin boundaries (2.18 % in the peripheral, and 9.70 % in the axial region).

| Substructure analysis
The extent of microstructure refinement was also documented by an additional scanning transmission electron microscopy analysis performed on a foil acquired from the peripheral location of the HPT2 sample.The analyses confirmed that the microstructure within this region of the sample featured very fine and even ultra-fine grains, Figure 5a, b.The detailed depiction of the microstructure shows that the majority of the grains was recrystallized and their interiors featured relatively low densities of dislocations, Figure 5b.This finding is in accordance with the above characterized results of the electron backscatter diffraction observations, which revealed that almost 90 % of the grain boundaries within this region of the HPT2 sample were high angle grain boundaries, Figure 4a.However, among the recrystallized grains, the microstructure still featured also a fraction of fine grains with evident development of substructure and high densities of dislocations.These substructured grains were typically finer than the larger ones, i. e. those with low densities of dislocations in their interiors, Figure 5a, b.Based on this finding one can assume that the finer substructured grains were those, which were affected by the imposed shear strain more severely, and were among the first to undergo another step in the fragmentation-restoration chain [42].

| Vickers microhardness
The measurements of Vickers microhardness for the processed samples were performed along the two characteristic lines across the cross-sections of the deformed samples, Figure 6.The average microhardness value acquired for the original copper was 43.9 HV 0.2.The values of microhardness increased substantially for both the HPT1 and HPT2 samples, Figure 6.The average microhardness values for the peripheral and axial regions of the HPT1 sample were 110.8 HV 0.2 and 100.1 HV 0.2, and the respective standard deviations for these values were 10.5 and 5.3.For the HPT2 sample, the average microhardness values for its peripheral and axial regions were 136.1 HV 0.2 and 131.8 HV 0.2, and the respective standard deviations were 8.8 and 6.6.Processing via a double revolution thus not only contributed to an increase in the Vickers microhardness, but also to homogenization of the structure across the cross-section.Nevertheless, differences between the peripheral and axial regions of the HPT2 sample were still evident.This fact can primarily be attributed to the geometry of the sample.The herein presented high pressure torsion device is designed to process samples with the original thickness of 6 mm contrary to others, whose high pressure torsion equipment is designed for processing of very thin samples with thicknesses ranging between 0.5 mm and 1 mm, e. g. [34,[43][44][45][46][47][48].The disadvantage of such a relatively large sample is, indeed, that a higher extent of imposed shear strain is needed for the strain to homogeneously affect the bulk volume of the sample.On the other hand, the indisputable advantage of such a bulk material volume is that the processed material can be directly used to produce little parts, such as specific components in the microelectronics.Considering this hypothesis (to confirm the suitability of the copper material processed in the proposed way for prospective usage in the microelectronics) the electric conductivity of the high pressure torsion processed samples was further measured.

| Electric conductivity
The measurements of the electric conductivity for each processed sample were performed ten times to also enable determination of the standard deviation from the average calculated electric conductivity value.The values measured for the processed samples were also compared with the electric conductivity acquired for the original copper, Table 1.When compared to the original copper, the electric conductivity evidently slightly increased for both the processed samples.The sample processed via a double revolution exhibited a slightly higher electric conductivity than the one processed with a single revolution, although the average grain size within the sample decreased significantly and thus the absolute volume of grain boundaries, which can be considered to act as barriers for the movement of electrons, increased [49].Nevertheless, as documented by the microscopic observations, the level of substructure development decreased substantially within the majority of the volume of the HPT2 sample.Therefore, the deteriorating effect of grain boundaries on the movement of electrons was compensated by a significant reduction of the occurrence of a well-developed substructure.The HPT1 sample, on the other hand, did not exhibit such an extent of grain refinement when compared to the HPT2 sample, but the electric conductivity for the sample increased, when compared to the original copper, too.This phenomenon was most probably related to the development of micro and nano twins within the structure.The deformed structure of the HPT1 sample featured quite a high fraction of 60°twin boundaries, Figure 3a, c.These boundaries were most probably fragmented remnants of the original twin boundaries observed within the original copper, Figure 2a.Fragmenting the twin boundaries and decreasing the twin spacing positively affects the electric conductivity [50].Moreover, the presence of twin boundaries favorably decreases the accumulation of lattice defects at regular grain boundaries.The mutual effect of these phenomena thus most probably contributed to the increased electric conductivity observed for the HPT1 sample.

| CONCLUSIONS
The presented study aimed to characterize the development of microstructure and correlate the structural features with the Vickers microhardness and electric conductivity within commercially pure copper processed by the high pressure torsion method.The samples were processed by single and double high pressure torsion revolutions.Based on the acquired results, the following bullet points can be summarized: -single high pressure torsion revolution imparted significant grain size decrease; -double revolution imparted formation of ultra-finegrained structure (grain size of sub-micron level at sample periphery); -both samples exhibited increased microhardness reaching up to 136 HV 0.2 after a double revolution, increased number of revolutions increased structure homogeneity; -electric conductivity was higher than 100 % IACS for both samples, primarily due to fragmentation of twin boundaries (single revolution sample), and increase in high angle grain boundary fraction and restoration of substructure (double revolution sample).
Based on the acquired results showing that high pressure torsion processing of such a bulk sample (ø 20 mm×6 mm) is promising for simultaneous increase in mechanical properties and electric conductivity, prospective usage of such a material in the microelectronics can be assumed.

1
Schematic depiction of analyzed locations.B I L D 1 Schematische Darstellung der untersuchten Bereiche.

F I G U R E 2
Characterization of grains for original Cu: a) orientation image map depicting grains orientations and boundaries; b) area-weighted grain size distribution chart; c) texture depicted via inverse pole figures.B I L D 2 Korncharakterisierung für Original Cu: a) Orientierungsbildmikroskopie-Karte mit Darstellung der Kornorientierungen und -grenzen; b) Diagramm der flächengewichteten Korngrößenverteilung; c) Darstellung der Textur mittels inverser Polfiguren.

F
I G U R E 5 a, b.Scanning transmission electron microscopy bright field images from peripheral region of HPT2 sample.B I L D 5 a, b.Rastertransmissionselektronenmikroskopie-Hellfeldbilder vom Randbereich der HPT2-Probe.F I G U R E 6 Vickers microhardness measured along processed samples.B I L D 6 Gemessene Mikrohärte nach Vickers entlang der bearbeiteten Proben.

T A B L E 1
Experimentally acquired electroconductivity for original copper and processed samples.T A B E L L E 1Experimentell ermittelte elektrische Leitfähigkeit für Kupfer und bearbeitete Proben.