Surface Quality Improvement of AA6060 Aluminum Extruded Components Through Liquid Nitrogen Mold Cooling

metals

Article Surface Quality Improvement of AA6060 Aluminum Extruded Components through Liquid Nitrogen Mold Cooling

Andrea Francesco Ciufﬁni 1,* ID , Silvia Barella 1 ID , Cosmo Di Cecca 1, Andrea Gruttadauria 1, Carlo Mapelli 1, Luca Merello 1, Giacomo Mainetti 2 ID and Massimo Bertoletti 2

1 Dipartimento di Meccanica, Politecnico di Milano, via La Masa 34, 20156 Milano, Italy, [email protected] (S.B.); [email protected] (C.D.C.); [email protected] (A.G.); [email protected] (C.M.); [email protected] (L.M.) 2 A.t.i.e. Uno Informatica S.r.l., via Macon 30, 23900 Lecco, Italy, [email protected] (G.M.); [email protected] (M.B.) * Correspondence: andreafrancesco.ciufﬁ[email protected]; Tel.: +39-340-1253162

Received: 5 May 2018; Accepted: 29 May 2018; Published: 1 June 2018

Abstract: 6xxx aluminum alloys are suitable for the realization of both structural applications and architectural decorative elements, thanks to the combination of high corrosion resistance and good surface finish. In areas where the aesthetic aspects are fundamental, further improvements in surface quality are significant. The cooling of the extrusion mold via internal liquid nitrogen fluxes is emerging as an important innovation in aluminum extrusion. Nowadays, this innovation is providing a large-scale solution to obtain high quality surface finishes in extruded aluminum semi-finished products. These results are also coupled to a significant increase in productivity. The aim of the work is to compare the surface quality of both cooled liquid nitrogen molds and classically extruded products. In this work, adhesion phenomena, occurring during the extrusion between the mold and the flowing material, have been detected as the main causes of the presence of surface defects. The analysis also highlighted a strong increase in the surface quality whenever the extrusion mold was cooled with liquid nitrogen fluxes. This improvement has further been confirmed by an analysis performed on the finished products, after painting and chromium plating. This work on the AA6060 alloy has moreover proceeded to roughness measurements and metallographic analyses, to investigate the eventual occurrence of other possible benefits stemming from this new extrusion mold cooling technology.

Keywords: aluminum; extrusion; liquid nitrogen cooling; surface quality; aesthetic

1. Introduction Extrusion is one of the most widely used forming processes for aluminum and its alloys [1,2]. In this process, the internal liquid nitrogen cooling of the extrusion mold represents a signiﬁcant innovation of recent years [3,4]. Indeed, during the extrusion process, different thermal exchanges occur, for example, heat generation due to material deformation, heat generation due to the friction between the billet and the liner and between the material and the dye, heat generation due to the internal friction between the ﬂowing material and the dead-metal zone, and heat conduction through the components of the system (billet, ram, chamber, dye, etc.). Furthermore, due to the high output speed of this forming technology, the output speed of the is semi-product also high. Thus, the contact time between the mold and the extruded material is very short and the temperature increments are concentrated within a thin surface layer. This will extensively increase the generation of surface defects [5–10].

Metals 2018, 8, 409; doi:10.3390/met8060409 www.mdpi.com/journal/metals Metals 2018,, 8,, xx 2 of 14

Metalsconcentrated2018, 8, 409 within a thin surface layer. This will extensively increase the generation of surface2 of 13 defects [5–10]. The new liquid nitrogen cooling technology, used directly in the matrix, allows the exploitation The new liquid nitrogen cooling technology, used directly in the matrix, allows the exploitation not only of the nitrogen inertizing effect, but also ofof itsits useuse asas coolant,coolant, providingproviding aa muchmuch broaderbroader not only of the nitrogen inertizing effect, but also of its use as coolant, providing a much broader range of benefits. The liquidliquid nitrogennitrogen entersenters intointo thethe channelschannels insideinside thethe matrixmatrix atat aa temperaturetemperature ofof range of benefits. The liquid nitrogen enters into the channels inside the matrix at a temperature of −196 °C and, during evaporation, it cools the mold. After the change in state, the gas creates an inert −196 ◦C and, during evaporation, it cools the mold. After the change in state, the gas creates an inert atmosphere, which inhibits the formation of oxides. The most significant advantage of this process, atmosphere, which inhibits the formation of oxides. The most significant advantage of this process, which justifies the higher installation and operating costs, is the increase in productivity due to the which justifies the higher installation and operating costs, is the increase in productivity due to the decrease in heat generation during extrusion. Indeed, the modern liquid nitrogen extrusion mold decrease in heat generation during extrusion. Indeed, the modern liquid nitrogen extrusion mold cooling systems optimize the consumption of nitrogen andand guaranteeguarantee sufficientsufficientlyly preciseprecise controlcontrol overover cooling systems optimize the consumption of nitrogen and guarantee sufficiently precise control over thethe wholewhole processprocess [3,4,11–13].[3,4,11–13]. AA schematicschematic sketchsketch of this technology is reported in Figure 1. the whole process [3,4,11–13]. A schematic sketch of this technology is reported in Figure1.

Figure 1. Schematic drawing of the liquid nitrogen extrusion mold cooling. Figure 1. SchematicSchematic drawingdrawing ofof thethe liquidliquid nitrogennitrogen extrusionextrusion moldmold cooling.cooling.

Furthermore, both the corrosion resistance and thethe aesthetic aspect of the final final products may be relevant inin somesome application fieldsfields ofof aluminumaluminum alloy extruded semi-products. Thus, surface quality may prove toto bebe crucialcrucial [[14–17].14–17]. Since thethe temperaturetemperature increments increments of of the the extrusion extrusio processn process are are concentrated concentrated in a thinin a surfacethin surface layer, theylayer,layer, drastically theythey drasticallydrastically affect theaffectaffect surface thethe surfacesurface finishing finishingfinishing [2,4,18, 19[2[2].,4,18,19].,4,18,19]. It is thus ItIt clearisis thusthus that clearclear the usethatthat of thethe liquid useuse ofof nitrogen liquidliquid coolingnitrogen of cooling the mold of andthe mold good and temperature good temperatur control ofe thecontrol overall of the process overall will process beneficially will beneficially affect these propertiesaffect these as properties well. as well. InIn thethe presentpresent work,work, the the AA6060 AA6060 alloy alloy (0.30–0.60 (0.30–0.60 wt wt % % Si, Si, 0.35–0.60 0.35–0.60 wt wt % Mg% Mg and and 0.1–0.3 0.1–0.3 wt %wt Fe) % isFe) tested. is tested. This This aluminum aluminum alloy alloy is extremely is extremely ductile ductile and therefore,and therefore, is one is of one the of most the most commonly commonly used aluminumused aluminum alloys. alloys. Further, Further, its mechanical its mechanical properties properties can be adjusted can be adjusted by subsequent by subsequent heat treatment heat throughtreatmenttreatment precipitation throughthrough precipitationprecipitation hardening. hardening.hardening. It is used inItIt manyisis usedused technical inin manymany applications, technicaltechnical applications,applications, such as automotive, suchsuch asas aerospaceautomotive, and aerospace structural and frame structural structures frame [20 structures–22]. [20–22]. This workwork aimsaims toto assess assess the the occurrence occurrence of of a relationshipa relationship between between the the surface surface quality quality and and the usethe ofuse internal of internal liquid liquid nitrogen nitrogen cooling cooling of the extrusion of the extrusion mold. Indeed, mold. for Indeed, structural for frame structural applications, frame suchapplications, as in this such case, as the in extruded this case, surface the extruded can be threated surface viacan thebe followingthreated via aesthetic the following surface processesaesthetic thatsurface might processes highlight that surface might defects. highlight In the surface present defect work,s. inIn detail, the present the extruded work, semi-productsin detail, the extruded undergo asemi-products 20 µm painting undergo process a 20 and μm a painting 1 µm thick process chromium and a 1 plating μm thick process chromium (Figure plating2). process (Figure 2).

Figure 2. Flowchart of the production process of extruded AA6060 structural frames. Figure 2. FlowchartFlowchart ofof thethe productionproduction processprocess of extruded AA6060 structural frames.

Metals 2018, 8, x 3 of 14

MetalsThus,2018, 8 ,all 409 the defects (such as pick-up, dye-lines and blisters), which are not completely hidden3 of 13 within these two overlaying layers, are highlighted in the final products by these aesthetic processes. Metals 2018, 8, x 3 of 14 The surfaceThus, all finishing the defects can (suchaccordingly as pick-up, prove dye-lines a major instance and blisters), in this which application are not field completely and even hidden small withinprogresses theseThus, could two all overlayingemergethe defects as significant.(such layers, as arepick-up, highlighted dye-lines in and the blisters), final products which are by not these completely aesthetic hidden processes. The surfacewithin these finishing two overlaying can accordingly layers, are prove highlighted a major in instance the final inproducts this application by these aesthetic field and processes. even small progresses2. ExperimentalThe surface could finishing Procedure emerge can as significant.accordingly prove a major instance in this application field and even small progresses could emerge as significant. During this work, two main sets of data were collected—one during the production trials of forty 2. Experimental Procedure semi-finished2. Experimental extruded Procedure products, and the other one through laboratory analysis, performed to assess their Duringquality. this work, two main sets of data were collected—one during the production trials of forty During this work, two main sets of data were collected—one during the production trials of forty semi-finishedThesemi-finished process extruded parameters extruded products, products, were andmonitored and the the other other by oneoneprocess-control through laboratory laboratory software, analysis, analysis, which performed manages performed to dataassess to assessabout materialtheirtheir quality. flow quality. speed, temperature and the liquid nitrogen valve opening. In detail, data about dye temperatureThe processThe was process parameterscollected parameters through were were monitored four monitored thermocouple by by process-controlprocess-controls (TC) installed software, software, within which which the manages managesmold data(Figure dataabout 3) about and twomaterial lasermaterial flowpyrometers. flow speed, speed, temperature temperature and and the the liquid liquid nitrogennitrogen valve valve opening. opening. In Indetail, detail, data data about about dye dye temperaturetemperature was was collected collected through through four four thermocouples thermocouples (TC) (TC) installed installed within within the the mold mold (Figure (Figure 3) and3) and two lasertwo laser pyrometers. pyrometers.

Figure 3. Schematic drawing of the thermocouples’ positions within the dye. FigureFigure 3. Schematic 3. Schematic drawing drawing of of the the thermocothermocouples’uples’ positions positions within within the thedye. dye.

The thermocouplesThe thermocouples used used in in thisthis this workwork work werewere were Chromel/Alumel Chromel/AlumelChromel/Alumel (K) (K) thermocouples thermocouples with with mineral mineral oxide insulation and Inconel 600 as the sheath material. The sheath had a 3.2 mm diameter and 2000 oxide insulation insulation and and Inconel Inconel 600 600 as as the the sheath sheath ma material.terial. The The sheath sheath had hada 3.2 a mm 3.2 mmdiameter diameter and 2000 and mm length. The laser pyrometer temperature method measured the emitted thermal radiation, which mm2000 length. mm length. The laser The pyrometer laser pyrometer temperature temperature method method measured measured the emitted the emitted thermal thermal radiation, radiation, which directly corresponded to the temperature and surface emissivity of the target. In detail, the pyrometer which directly corresponded to the temperature and surface emissivity of the target. In detail, directlysensors corresponded detected the to amount the temperature of infrared and radiatio surfacen emitted emissivity by the of measured the target. object. In detail, The pyrometers the pyrometer sensorsthe pyrometerwere detected installed sensors the in frontamount detected of the of exit infrared the of amountthe extrusionradiatio of infraredn dye emitted (Figure radiation by 4). the measured emitted by object. the measuredThe pyrometers object. Thewere pyrometers installed in were front installed of the exit in of front the ofextrusion the exit ofdye the (Figure extrusion 4). dye (Figure4).

Figure 4. Pyrometers’ positions.

Figure 4. Pyrometers’ positions. Figure 4. Pyrometers’ positions.

Metals 2018, 8, 409 4 of 13

On the other hand, the surface defects of the extruded semi-products were characterized in the junction zone of the billets, since this is an accumulation area for the defects [1,2]. The samples were drawn from the extruded semi-products and polished following the common standards. The chemical compositions of the samples were acquired through a scanning electron microscope (SEM) (Carl Zeiss AG, Oberkochen, Germany) Zeiss EVO 50®, using a back-scattered electron (SEM-BSE) (INCA Oxford Instrument, Oxford, UK) detector and energy dispersive X-ray microanalysis (SEM-EDS) (INCA Oxford Instrument, Oxford, UK). In addition, the same instrument was exploited to collect morphological images through a secondary electron detector (SEM-SE) (INCA Oxford Instrument, Oxford, UK). The samples were etched electrochemically using a Barker’s reagent. The electrolytic solution was prepared by mixing 40 parts of water (H2O) for each volume-part of tetrafluoboric acid (HBF4). The electrochemical anodization was performed by applying a current density of 0.2 A/cm2 (20 V dc) for 40–80 s. The metallographic analysis was executed using a cross polarized incident light. Afterwards, the index used to evaluate the average grain dimensions was calculated via the Heyn intercept method, following the ASTM E112-13 standard. Roughness measurements were performed on the surface of the body of the extruded semi-product, hence, not in the junction area. Profile roughness measurements were taken perpendicularly to the extrusion direction, using a stylus Mahr PGK MFK-250® (Mahr GmbH, Göttingen, Germany) tester with a tip radius of 2 µm and a vertical measure range of +/−250 µm. The data were collected following the standard, UNI EN ISO 4288-2000. The translational speed of the measurement was 0.5 mm/s along an exploration length of 5.6 mm. The base length was 0.8 mm and the total evaluated length was 4.0 mm. The collected data were filtered using a Gaussian filter with a wavelength cutter of 0.8 mm. Surface texture measurements were performed using an Alicona InfiniteFocus® (Alicona Imaging GmbH, Graz, Austria) microscope, which exploits focus-variation technology. The data were acquired in a similar position to those collected for the profile roughness measurements. The surface texture measurements parameters are listed as follows: the total evaluated area measured 4.0 × 4.0 mm and the base length was 0.8 mm. The collected data were filtered again using a Gaussian filter with a wavelength cutter of 0.8 mm.

3. Results The data collected during the extrusion trials are summarized in Table1. This sets out the data regarding the most meaningful specimens that were analyzed in this study. It should be underlined that the samples numbered from 1 to 15 were processed through a conventional extrusion, and then, the liquid nitrogen valve was opened and the specimens numbered up to 25 were subjected to transient conditions. Thus, only the semi-products numbered from 26 to 40 were extruded with the liquid nitrogen mold cooling system working at full capacity.

Table 1. Data featuring the extrusion trials of the specimens. Samples 5, 9, 13 were extruded conventionally. Samples 21 and 25 were processed during the transition step. Specimens 33 and 37 were extruded when the liquid nitrogen mold cooling worked at full capacity. The thermocouples (TC) were positioned within the mold as reported in Figure3.

Billet Average Billet Average Liquid Nitrogen Sample TC1 (◦C) TC2 (◦C) TC3 (◦C) TC4 (◦C) Speed (m/min) Temperature (◦C) Valve Aperture (%) 5 15.94 526 518 507 497 527 0 9 17.06 534 527 514 503 534 0 13 19.62 545 528 516 505 538 0 21 22.18 528 504 482 482 522 60.46 25 16.74 528 479 456 450 505 20.05 33 15.64 511 474 447 447 497 11.72 37 15.51 506 463 439 441 485 27.08 Metals 2018, 8, 409 5 of 13 Metals 2018, 8, x 5 of 14

TheThe datadata regardingregarding thethe characterizationcharacterization ofof defectsdefects areare reportedreported below.below. TwoTwo differentdifferent speciesspecies ofof defectsdefects were were detected detected on on the the surface surface in in the the junction junction area area of theof the extruded extruded semi-products. semi-products. Their Their SEM-BSE SEM- imagesBSE images are shown are shown in Figure in Figure5, and 5, the and acquired the acquired chemical chemical compositions compositions are quoted are quoted in Table in2 Table. 2.

Figure 5. SEM-BSE images of extrusion defects detected in the junction area of the billets: sample 5 (a), sample 13 (bFigure). A, B, 5.C SEM-BSEare the point images in which of extrusion the chemical defects analysis detected was in theperformed junction (Table area of 2). the billets: sample 5 (a), sample 13 (b). A, B, C are the point in which the chemical analysis was performed (Table2). Table 2. SEM-EDS chemical analysis of the areas highlighted in Figure 5, the element concentrations Tableare expressed 2. SEM-EDS as wt chemical %. analysis of the areas highlighted in Figure5, the element concentrations are expressed as wt %. Area % Mg % Al % Si % S % V % Cr % Mo % Fe Area5-1 L_05_A % Mg- % Al3.89 % Si2.50 %1.47 S 2.11 % V 9.67 % Cr %- Mo80.36 % Fe 5-1 L_05_A5-1 L_05_B -0.58 3.8998.87 2.500.55 1.47- 2.11- 9.67- - -- 80.36 5-113-3 L_05_B L_02_A 0.58 - 98.87 2.38 0.55 2.67 - - 1.73 - 11.26 - 3.12 - 78.76 - 13-313-3 L_02_A L_02_B -0.57 2.3897.64 2.670.49 -- 1.73- 11.26- 3.12- - 78.76 13-313-3 L_02_B L_02_C 0.57 0.75 97.6498.68 0.490.57 ------13-3 L_02_C 0.75 98.68 0.57 - - - - -

Therefore, the analyzed defects can be classified as pick-up and dye pick-up. Pick-ups (Figure 5a—pointTherefore, C) are the intermittent analyzed score defects lines can of varying be classified lengths, as which pick-up end in and a fleck dye of pick-up. aluminum Pick-ups debris. (FigureThey can5a—point be easily C) identified are intermittent by their score chemical lines composition, of varying lengths, which which is the same end in as a that fleck of of the aluminum extruded debris.aluminum They alloy can [23]. be easily The exogenous identified bydeposited their chemical layers were composition, composed which by a isdifferent the same material as that and of thethey extruded can be recognized aluminum as alloy dye [pick-up23]. The (Figure exogenous 5b—p depositedoint A, Figure layers 5a—point were composed A), which by are a different particles materialworn-out and of the they extrusion can be recognized dye and deposite as dye pick-upd on the (Figure semi-product5b—point surface A, Figure [23–27].5a—point A), which are particlesThereafter, worn-out the products, of the extrusion subjected dye to and the deposited whole production on the semi-product cycle, including surface [23painting–27]. and chromiumThereafter, plating, the were products, also analyzed. subjected The to thetypica wholel section, production recorded cycle, via includingSEM-BSE, paintingof the surface and chromiumdefects after plating, these processes, were also is analyzed. shown in The Figure typical 6. It section,is worth recorded highlighting viaSEM-BSE, that the paint of the layer surface does defectsnot follow after the these defect processes, profile; is it shown would in thus Figure cover6. It an is worthd hide highlighting the defects featured that the paint by heights layer does up to not 15 followμm. the defect profile; it would thus cover and hide the defects featured by heights up to 15 µm.

Metals 2018, 8, 409 6 of 13 Metals 2018, 8, x 6 of 14 Metals 2018, 8, x 6 of 14

FigureFigure 6. 6. SEM-BSE images of sections of the final products defects: sample 12 (aa), sample 25 (bb). Figure 6.SEM-BSE SEM-BSE images images of of sections sections of ofthe theﬁnal final productsproducts defects:defects: sample 12 ( a),), sample sample 25 25 ( (b).). Through optical microscopy, a metallographic analysis was performed on the tested samples. ThroughThrough opticaloptical microscopy,microscopy, aa metallographicmetallographic analysisanalysis waswas performedperformed onon thethe testedtested samples.samples. Figure 7 sets out two micrographs obtained after threating the samples via Barker electrochemical FigureFigure7 7sets sets out out two two micrographs micrographs obtained obtained after after threating threating the the samples samples via via Barker Barker electrochemical electrochemical etching. However, this analysis did not highlight any detail about the microstructure, which could etching.etching. However,However, this this analysis analysis did did not not highlight highligh anyt any detail detail about about the the microstructure, microstructure, which which could could be be relevant or have influence on the generation of defects. Indeed, at the grain boundaries no relevantbe relevant or have or have inﬂuence influence on the on generation the generation of defects. of defects. Indeed, atIndeed, the grain at boundariesthe grain boundaries no secondary no secondary phase or trace of segregation phenomena is present [28]. phasesecondary or trace phase of segregation or trace of segregation phenomena phenomena is present [28 is]. present [28].

Figure 7. Micrographic analysis of the pick-ups presents within the defect area the specimen 5 (a) and FigureFigure 7.7. MicrographicMicrographic analysisanalysis ofof thethe pick-upspick-ups presentspresents withinwithin thethe defectdefect areaarea thethe specimenspecimen 55 ((aa)) andand specimen 33 (b). specimenspecimen 3333 ((bb).). Moreover, the microstructures appeared homogeneous across the junction area sections (Figure Moreover, the microstructures appeared homogeneous across the junction area sections (Figure 8a). Moreover,Thus, no therelationship microstructures between appeared the microstr homogeneousucture acrossand the the junctionsurface areadefects sections was (Figureidentified.8a). 8a). Thus, no relationship between the microstructure and the surface defects was identified. Thus,Moreover, no relationship no differences between in the the morphologies microstructure of andthe thecrystal surface grains defects were wasdetected; identiﬁed. all the Moreover, junction Moreover, no differences in the morphologies of the crystal grains were detected; all the junction noareas differences of the tested in the specimens morphologies featured of coarse the crystal equiax grainsial microstructures, were detected; withou all thet junction been affected areas by of any the areas of the tested specimens featured coarse equiaxial microstructures, without been affected by any testedvariations specimens in the featuredworking coarseparameters. equiaxial microstructures, without been affected by any variations in variations in the working parameters. the workingHowever, parameters. few millimeters away from the junction areas, the microstructures of the specimens However, few millimeters away from the junction areas, the microstructures of the specimens wereHowever, very different few millimeters (Figure 8b). away The from grains the junctiondisplayed areas, a lamellar the microstructures structure, oriented of the specimensalong the were very different (Figure 8b). The grains displayed a lamellar structure, oriented along the wereextrusion very differentdirection, (Figure usually8b). featuring The grains the displayedextruded semi-products. a lamellar structure, This occurred oriented alongbecause the the extrusion process extrusion direction, usually featuring the extruded semi-products. This occurred because the process direction,arrest leaves usually the junction featuring area the for extruded a much longer semi-products. time in contact This occurredwith the mold because after the being process deformed. arrest arrest leaves the junction area for a much longer time in contact with the mold after being deformed. leavesThus, thea recrystallization junction area for phenomenon a much longer was time promot in contacted, withas testified the mold by after the beingloss of deformed. the extrusion Thus, Thus, a recrystallization phenomenon was promoted, as testified by the loss of the extrusion adeformation recrystallization pattern phenomenon of the grains was [29]. promoted, as testiﬁed by the loss of the extrusion deformation deformation pattern of the grains [29]. pattern of the grains [29].

Metals 2018, 8, 409 7 of 13 Metals 2018, 8, x 7 of 14 Metals 2018, 8, x 7 of 14

FigureFigure 8. MicrographicMicrographic analysisanalysisof of sample sample 9 in9 thein junctionthe junction area (areaa) and (a in) and a different in a different area representative area representativeofFigure the rest 8. of Micrographicthe of the semi-product rest of theanalysis semi-product (b). of sample (b). 9 in the junction area (a) and in a different area representative of the rest of the semi-product (b). Moreover, the grain size number G, an index measuring the average grain dimensions, was Moreover, the grain size number G, an index measuring the average grain dimensions, calculatedMoreover, through thethe grainHeyn sizeintercept number method G, an and index the resultsmeasuring are reported the average in Table grain 3. dimensions, was wascalculated calculated through through thethe Heyn Heyn intercept intercept method method and andthe results the results are reported are reported in Table in Table3. 3. Table 3. Grain size number (G) index. TableTable 3. 3.Grain Grain sizesize number (G) (G) index. index. Sample Extrusion Mean Temperature (°C) G (NL/mm) ◦ Sample5Sample Extrusion Extrusion Mean Mean526 Temperature ( (°C)C) G G (N14 (NL /mm)L/mm) 9 5 5534526 526 13.5 14 14 13 9 9545534 534 14.5 13.5 13.5 21 1313 528545 545 14 14.5 14.5 25 2121 528528 528 14 14 14 25 528 14 25 528 14 33 33511 511 14 14 37 3337 506511 506 14 14 14 37 506 14 Furthermore, a SEM-SE morphological analysis wasperformed, to qualitatively assess the Furthermore, a SEM-SE morphological analysis wasperformed, to qualitatively assess the differencesFurthermore, between athe SEM-SE different morphological mold cooling analysis methodologies. wasperformed, The mostto qualitatively interesting assessresults the differences between the different mold cooling methodologies. The most interesting results concerned concerneddifferences billets between extruded the with different the same mold working cooling para meters.methodologies. In Figure 9,The the junctionmost interesting areas of billets results 5-2billetsconcerned (air cooled extruded billets dye) with andextruded 25-4 the (liquid samewith the workingnitrogen same working cooled parameters. die) para aremeters. Indisplayed Figure In Figure 9in, detail. the 9, the junction Comparedjunction areas areas to the of billets25- billets 45-2 sample,5-2 (air (air cooled thecooled 5-2 dye) specimendye) andand 25-425-4 showed (liquid much nitrogen nitrogen larger cooled defe cooledcts die) and die) are a arelarge displayed displayed quantity in detail. of in the detail. Comparedmaterial Compared adhered to the to25- the onto25-44 sample, sample,the surface. the the 5-2 5-2 specimen specimen showed showed much much larger larger defe defectscts and and a large a large quantity quantity of ofthe the material material adhered adhered ontoonto the the surface. surface.

Figure 9. SEM-SE morphological analysis of the junction areas of the billets comparing one produced throughFigureFigure 9.an 9.SEM-SE air SEM-SE cooled morphological morphological die 5-2 (a) and analysis analysis one produced ofof the junctionvia the liqui areas areasd nitrogenof of the the billets billets cooled comparing comparing die 25-4 one(b one). produced produced throughthrough an an air air cooled cooled die die 5-2 5-2 ( a()a) and and one one producedproduced viavia the liqui liquidd nitrogen nitrogen cooled cooled die die 25-4 25-4 (b ().b ).

Metals 2018, 8, 409 8 of 13

The results of the roughness tests, performed on the surface of the body of the extruded semi-product—hence not in the junction areas—are shown in Table4. The mean roughness grew with the mold cooling liquid nitrogen valve aperture during the extrusion trials. On the other hand, the measured maximum roughness displayed a decreasing trend under the same operative conditions.

Table 4. Roughness tests results (Ra: Roughness Average; Rz: Average Maximum Height of the Proﬁle; Rmax: Maximum Roughness Depth).

Sample Ra (µm) Rz (µm) Rmax (µm) 13 0.40 3.02 4.31 25 0.44 4 5.12 33 0.53 3.54 3.89

The surface texture data, collected via microscopy focus-variation technology, are set out below—both the parameters describing the roughness (Table5) and the height histogram of the surfaces and its statistics (Figure 10 and Table6). The results are similar to those for the proﬁle roughness. The mean values of the surface roughness slightly increased with use of the liquid nitrogen cooling system. On the other hand, the parameters describing the maxima and the minima of the surface, underwent a decreasing trend with the use of the liquid nitrogen cooling system. In detail, the skewness values indicated that the height distribution of the liquid nitrogen-cooled sample was more “valley-tailed” than the traditionally extruded surface. For both distributions, the kurtosis parameter highlighted a leptokurtic behavior with a positive excess kurtosis. The higher value of the sample traditionally extruded testiﬁes the presence of larger tails in its height distribution, due to the occurrence of larger surface defects in the sample produced through this technology, as observed via other characterization techniques as well.

Table 5. Surface roughness parameters.

Sample Parameter Description (u) 5 37 Sa Average height of selected area nm 361 471 Sq Root-mean-square height of selected area nm 467 587 Sp Maximum peak height of selected area µm 1.99 1.41 Sv Maximum valley depth of selected area µm 3.98 2.48 Sz Maximum height of selected area µm 5.97 3.89 S10z Ten-point height of selected area µm 4.66 3.80 Ssk Skewness of selected area −0.45 −0.78 Sku Kurtosis of selected area 3.69 3.13 Sdq Root mean square gradient 0.08 0.09 Sdr Developed interfacial area ratio % 0.34 0.44 FLTt Flatness using least squares reference plane µm 5.97 3.89

Table 6. Statistics of the height histograms of specimens’ surfaces.

Sample Height Histogram Statistics (u) 5 37 Number of Elements 8,816,414 8,253,132 Classes 120 195 Class Width µm 0.05 0.02 Mean Value µm 0.0186 −0.0225 Standard Deviation µm 0.4667 0.5870 Minimum Value µm −3.9756 −2.4777 Maximum Value µm 2.0244 1.4233 Metals 2018, 8, 409 9 of 13 Metals 2018, 8, x 9 of 14

16 14 12 10 8 6 4

Relative Frequency (%) Frequency Relative 2 0 0.08 0.36 0.64 0.92 1.20 1.48 -2.44 -2.16 -1.88 -1.60 -1.32 -1.04 -0.76 -0.48 -0.20 Height (µm)

5 37

FigureFigure 10. 10. HeightHeight histogram histogram of of the the samples’ samples’ surfaces: surfaces: traditionally traditionally extruded extruded (specimen (specimen 5) 5) and and the the liquidliquid nitrogen nitrogen cooled cooled (specimen (specimen 37). 37).

4. Discussion 4. Discussion The experimental data from the current study testify that the liquid nitrogen cooling effect is not The experimental data from the current study testify that the liquid nitrogen cooling effect is instantaneous and a transition time is required before a steady working regime temperature can be not instantaneous and a transition time is required before a steady working regime temperature can reached and maintained. This is related to the mold’s thermal inertia. Furthermore, when using this be reached and maintained. This is related to the mold’s thermal inertia. Furthermore, when using technology, the working temperature becomes lower relative to conventional air cooling of the this technology, the working temperature becomes lower relative to conventional air cooling of the extrusion mold. However, the differences between temperatures measured at the different dye holes extrusion mold. However, the differences between temperatures measured at the different dye holes broaden. This phenomenon may be related to the inhomogeneity of the heat removal, linked to the broaden. This phenomenon may be related to the inhomogeneity of the heat removal, linked to the non-optimal sizing of the cooling channel. Therefore, this aspect should be taken into account during non-optimal sizing of the cooling channel. Therefore, this aspect should be taken into account during the design of the mold, to achieve easy and uniform temperature control [30]. the design of the mold, to achieve easy and uniform temperature control [30]. The experimental data regarding the defects analysis in the junction area of the semi-finished The experimental data regarding the defects analysis in the junction area of the semi-finished extruded billets, found the presence of two varieties of defects. The first featured aluminum flecks of extruded billets, found the presence of two varieties of defects. The first featured aluminum flecks debris adherent to the surface. The chemical composition confirmed that the origin of this debris was of debris adherent to the surface. The chemical composition confirmed that the origin of this debris the extruded material, since it is identical. The adhesion between the dye and the extruded material was the extruded material, since it is identical. The adhesion between the dye and the extruded generated these defects, as testified by their peculiar morphologies. These defects can thus be material generated these defects, as testified by their peculiar morphologies. These defects can thus be recognized as pick-ups [23]. recognized as pick-ups [23]. Further to this, a second category of defects was detected and it identified as dye pick-up. This Further to this, a second category of defects was detected and it identified as dye pick-up. observation was proven by the chemical composition analysis of the deposited material. Indeed, this This observation was proven by the chemical composition analysis of the deposited material. Indeed, composition is similar the composition of the AISI H13 tool steel, which is reported in Table 7. Since this composition is similar the composition of the AISI H13 tool steel, which is reported in Table7. the extrusion dye is made in this material, its provenance is clearly proved [23–27]. Since the extrusion dye is made in this material, its provenance is clearly proved [23–27]. Moreover, any metallurgical origin of the analyzed defects can be excluded. Indeed, both the Moreover, any metallurgical origin of the analyzed defects can be excluded. Indeed, both the SEM/BSE analysis and optical microscopy did not detect the presence of any other phases or SEM/BSE analysis and optical microscopy did not detect the presence of any other phases or intermetallic precipitates, any abnormal microstructural features or any abnormal chemical element intermetallic precipitates, any abnormal microstructural features or any abnormal chemical element concentrations. It follows that the origin of the detected defects is linked to wear issues and, in detail, concentrations. It follows that the origin of the detected defects is linked to wear issues and, in detail, it can be recognized as relating totally to the processing parameters and the tools [23–27]. it can be recognized as relating totally to the processing parameters and the tools [23–27].

Table 7. Extrusion dye material (AISI H13 tool steel) chemical composition, as designed by standards, Table 7. Extrusion dye material (AISI H13 tool steel) chemical composition, as designed by standards, the element concentrations are expressed as wt %. the element concentrations are expressed as wt %. Steel % C % Si % Mn % Ni % V % Cr % Mo % Fe AISISteel H13 % C 0.40 % Si1.0 % Mn 0.4 % 0.3Ni % 1.0 V % 5.0 Cr % 1.5 Mo Bal.% Fe AISI 0.40 1.0 0.4 0.3 1.0 5.0 1.5 Bal. It should beH13 highlighted that no correlation between the variation of extrusion parameters during the production trials and the species defects were identified. The generation of these kinds of

Metals 2018, 8, 409 10 of 13

It should be highlighted that no correlation between the variation of extrusion parameters during the production trials and the species defects were identified. The generation of these kinds of defects would not be totally prevented using liquid nitrogen cooling of the extrusion dye. On the other hand, the occurrence of new defect classes can be excluded by using this technology. In addition, the liquid nitrogen cooling of the extrusion dyes resulted in a beneficial effect on the amount and on the size of the extrusion defects with patent improvements. Indeed, although the analyzed extrusion defects were few and topologically isolated, the use of this technology further lowered their occurrence and dimensions [11,12]. However, considering their application, in which the aesthetic aspect plays a predominant role, even isolated and very small defects would cause the rejection of the extruded semi-product (Figure 11). Taking into account the further steps involved in the realization of the finished product, we can set the defect acceptance threshold at a height of 15 µm. Indeed, since the paint layer does not follow the defectMetals 2018 profile,, 8, x it could cover and hide these small defects. 11 of 14

Figure 11. Semi-finishedSemi-ﬁnished ( (aa)) and and finished ﬁnished (b) products:products: the the production cy cycle,cle, after extrusion, is completed by brushing, acid degreasing,degreasing, painting andand chromiumchromium plating.plating.

Considering this finishing process, the most significant roughness data were those describing the maxima of the roughness measurements of the surfaces and the distribution of the height histogram. These data (Tables4–6 and Figure 10) demonstrate the beneficial effect of the liquid nitrogen cooling of the extrusion dye, and an improvement in this key aspect can be appreciated. In addition, the surface roughness maps of the reconstructed surfaces likewise highlight this aspect (Figure 12). The surface of the liquid nitrogen cooled specimen was very smooth and featured valleys along the extrusion direction. On the other hand, on the surface of the conventionally extruded semi-product, peak-like defects were detected emerging from the surface, harming the subsequent finishing processes. Liquid nitrogen dye cooling technology improves the extrusion process through a complex mechanism, which can be described as follows: the debris, which generates the previously analyzed defects, is in both cases related to the adhesion phenomena occurring during extrusion between the dye and the flowing aluminum. However, this debris is much less present when the liquid nitrogen cools down the mold, even if the measured temperature is the same as for the air cooling.

Figure 12. Surface roughness maps. The surface of the conventionally extruded semi-product (a) is featured by peak-like defects emerging from the surface. The surface of the liquid nitrogen cooled specimen (b) features the presence of valleys along the extrusion direction.

Metals 2018, 8, x 11 of 14

Metals 2018, 8, 409 11 of 13

Hence, the improvement in the surface finishing, obtained through the liquid nitrogen cooling technology, is not only related to the lowering of the working temperature. A different kind of contribution, which could increase the surface finishing, is given by the mechanical action that the nitrogen exerts on the flowing extruded material. Indeed, while absorbing the heat cooling the dye, nitrogen undergoes a phase transformation from the liquid to the gaseous state. This phase transition is known to be accompanied by a very large increase in volume. However, sinceFigure the cooling 11. Semi-finished circuit is pressured, (a) and finished the phase (b) transformationproducts: the production generates cy acle, detaching after extrusion, pressure is on the flowingcompleted material, by brushing, which keeps acid degreasi it separatedng, painting from the and mold. chromium plating.

FigureFigure 12. 12. SurfaceSurface roughness roughness maps. maps. The The surface surface of of the the conventionally conventionally extruded extruded semi-product semi-product ( (aa)) is is featuredfeatured by by peak-like peak-like defects defects emerging emerging from from the the su surface.rface. The The surface surface of of the the liquid liquid nitrogen nitrogen cooled cooled specimenspecimen ( (bb)) features features the the presence presence of of valley valleyss along along the the extrusion extrusion direction. direction.

Moreover, liquid nitrogen cooling has another beneficial effect for surface finishing. In detail, it acts as an inert atmosphere enveloping the flowing material, avoiding its oxidation. This plays a central role in wear phenomena, in which aluminum is involved, since its oxide is very easily generated and is extremely hard. On the other hand, in comparison, aluminum is extremely soft and easily damageable with respect to its oxide [4]. Therefore, these beneficial effects on surface finishing, related to the use of liquid nitrogen mold cooling, result from a combination of the cooling effect, the inertizing effect and a mechanical detaching pressure generated by the liquid-to-gaseous transition.

5. Conclusions Production trials of aluminum alloy extruded semi-products were performed both with conventional air cooling of the extrusion mold and with newly installed liquid nitrogen cooling Metals 2018, 8, 409 12 of 13 of the mold. The semi-products analysis led to the identification of two main categories of defects found in the joint areas: pick-up and dye pick-up. The metallographic analysis pointed out the absence of any hot shortness, grain boundary segregation or coarse intermetallic compound formation phenomena occurring within aluminum semi-finished products as concurrent causes for pick-up generation. The liquid nitrogen dye-cooling did not have any influence on the extruded aluminum grain size. However, the use of liquid nitrogen cooling reduced the adhesion phenomena between the flowing aluminum and the mold, hindering the generation of the detected defects. Moreover, liquid nitrogen dye cooling provided inertizing, detaching and cooling effects. Consequently, the best results, in terms of absence of defects, were achieved at the lowest working temperatures through the use of liquid nitrogen cooling. Furthermore, low working speeds also hindered the generation of defects. Finally, liquid nitrogen mold cooling, compared to classic air cooling, granted smoother surfaces featured by a high density of valleys along the extrusion direction and the almost complete absence of peak-like defects emerging from the surface which would lead to flaws in the subsequent finishing processes.

Author Contributions: A.F.C., S.B., C.D.C., A.G., L.M. took part to the experimental campaign relative to the metallographic investigation and the defects detection and analysis. C.M. is the research group leader and supervised the work, G.M. and M.B. performed the extrusion an collect the process data using the software produced by A.t.i.e. uno. Funding: This research received no external funding. Acknowledgments: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Conflicts of Interest: The authors declare no conflicts of interest.

References

1. Sheppard, T. Extrusion of Aluminum Alloys; Kluwer Academic Publishers: Norwell, MA, USA, 1999. 2. Saha, P.K. Aluminum Extrusion Technology; ASM International: Materials Park, OH, USA, 2000. 3. Donati, L.; Segatori, A.; Reggiani, B.; Tomesani, L.; Bevilacqua Fazzini, P.A. Effect of liquid nitrogen die cooling on extrusion process conditions. Key Eng. Mater. 2012, 491, 215–222. [CrossRef] 4. Ko, D.-H.; Kang, B.-H.; Ko, D.-C.; Kim, B.-M. Improvement of mechanical properties of Al6061 extrudate by

die cooling with N2 gas during hot extrusion. J. Mater. Process. Technol. 2013, 27, 153–161. [CrossRef] 5. Bjork, T.; Bergstrom, J.; Hogmark, S. Tribological simulation of aluminum hot extrusion. Wear 1999, 224, 216–225. [CrossRef] 6. Bastani, A.F.; Aukrust, T.; Brandal, S. Optimisation of ﬂow balance and isothermal extrusion of aluminum using ﬁnite-element simulations. J. Mater. Process. Technol. 2011, 211, 650–667. [CrossRef] 7. Ma, X.; De Rooij, M.B.; Schipper, D.J. Modelling of contact and friction in aluminum extrusion. Tribol. Int. 2010, 43, 1138–1144. [CrossRef] 8. Ma, X.; De Rooij, M.B.; Schipper, D.J. Friction conditions in the bearing area of an aluminum extrusion process. Wear 2012, 279, 1–8. [CrossRef] 9. Saha, P.K. Thermodynamics and tribology in aluminum extrusion. Wear 1998, 218, 179–190. [CrossRef] 10. Schikorra, M.; Donati, L.; Tomesani, L.; Kleiner, M. The role of friction in the extrusion of AA6060 aluminum alloy, process analysis and monitoring. J. Mater. Process. Technol. 2007, 191, 288–292. [CrossRef] 11. Celani, P. Liquid nitrogen technology. In Proceedings of the 8th Aluminum Two Thousand International Congress, Milano, Italy, 14–18 May 2013. 12. Ward, T.J.; Kelly, R.M.; Jones, G.A.; Heffron, J.F. The effects of nitrogen—Liquid and gaseous—On aluminum extrusion productivity. JOM 1984, 36, 29–33. [CrossRef] 13. Twigg, R.; Oesterreich, R. Extrud-AlTM—An alternative approach to nitrogen die cooling. In Proceedings of the 8th International Extrusion Technology Seminar, Orlando, FL, USA, 18–21 May 2004. 14. Cotell, C.M.; Sprague, J.A.; Smidt, F.A., Jr. ASM Handbook Vol. 5: Surface Engineering. Surface Engineering of Aluminum and Aluminum Alloys; ASM International: Materials Park, OH, USA, 1994; pp. 784–804. Metals 2018, 8, 409 13 of 13

15. Gali, O.A.; Shafiei, M.; Hunter, J.A.; Riahi, A.R. The influence of work roll roughness on the surface/near-surface microstructure evolution of hot rolled aluminum—Magnesium alloys. J. Mater. Process. Technol. 2016, 237, 331–341. [CrossRef] 16. Rao, K.S.; Rao, K.P. Pitting Corrosion of Heat-treatable Aluminum Alloys and Welds: A Review. Trans. Indian Inst. Met. 2004, 57, 593–610. 17. Rodríguez-Diaz, R.A.; Uruchurtu-Chavarín, J.; Valdez-Villegas, S.; Valdez, S.; Juárez-Islas, J.A. Corrosion Behavior of AlMgSi Alloy in Aqueous Saline Solution. Int. J. Electrochem. Sci. 2015, 10, 1792–1808. 18. De Rooij, M.B.; Ma, X.; Den Bakker, A.J.; Werkhoven, R.J. Surface quality prediction in aluminium extrusion. Key Eng. Mater. 2012, 491, 27–34. [CrossRef] 19. Li, L.; Zhou, J.; Duszczyk, J. Prediction of temperature evolution during the extrusion of 7075 aluminium alloy at various ram speeds by means of 3D FEM simulation. J. Mater. Process. Technol. 2004, 145, 360–370. [CrossRef] 20. Miller, W.S.; Zhuang, L.; Bottema, J.; Wittebrood, A.J.; De Smet, P.; Haszler, A.; Vieregge, A. Recent development in aluminium alloys for the automotive industry. Mater. Sci. Eng. A 2000, 280, 37–49. [CrossRef] 21. Berndt, N.; Frint, P.; Wagner, M.F.-X. Influence of Extrusion Temperature on the Aging Behavior and Mechanical Properties of an AA6060 Aluminum Alloy. Metals 2018, 8, 51. [CrossRef] 22. Tekkaya, A.E.; Schikorra, M.; Becker, D.; Biermann, D.; Hammer, N.; Pantke, K. Hot profile extrusion of AA-6060 aluminum chips. J. Mater. Process. Technol. 2009, 209, 3343–3350. [CrossRef] 23. Garbacz, P.; Giesko, T.; Mazurkiewicz, A. Inspection method of aluminum extrusion process. Arch. Civ. Mech. Eng. 2015, 15, 631–638. [CrossRef] 24. Björk, T.; Westergård, R.; Hogmark, S. Wear of surface treated dies for aluminum extrusion—A case study. Wear 2001, 249, 316–323. [CrossRef] 25. Birol, Y. Analysis of wear of a gas nitrided H13 tool steel die in aluminum extrusion. Eng. Fail. Anal. 2012, 26, 203–210. [CrossRef] 26. Qamar, S.Z.; Arif, A.F.M.; Sheikh, A.K. Analysis of Product Defects in a Typical Aluminum Extrusion Facility. Mater. Manuf. Process. 2004, 19, 391–405. [CrossRef] 27. Arif, A.F.M.; Sheikh, A.K.; Qamar, S.Z.; Raza, M.K.; Al-Fuhaid, K.M. Product defects in aluminum extrusion and their impact on operational cost. In Proceedings of the 6th Saudi Engineering Conference, KFUPM, Dhahran, Saudi Arabia, 14–17 December 2002; pp. 137–154. 28. Wang, L.; Cai, J.; Zhou, J.; Duszczyk, J. Characteristics of the Friction between Aluminium and Steel at Elevated Temperatures During Ball-on-Disc Tests. Tribol. Lett. 2009, 36, 183–190. [CrossRef] 29. Kayser, T.; Klusemann, B.; Lambers, H.; Maier, H.J.; Svendsen, B. Characterization of grain microstructure development in the aluminum alloy EN AW-6060 during extrusion. Mater. Sci. Eng. A 2010, 527, 6568–6573. [CrossRef] 30. Tercelj, M.; Kugler, G.; Turk, R.; Cvahte, P.; Fajfar, P. Measurement of temperature on the bearing surface of an industrial die and assessment of the heat transfer coefficient in hot extrusion of aluminum: A case study. Int. J. Veh. Des. 2005, 39, 93–109. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).