la metallurgia italiana - Gruppo Italiano Frattura

Memorie >> 

Alluminio e leghe 

CORRELATION BETWEEN 

MICROSTRUCTURE AND 

MECHANICAL PROPERTIES 

OF Al-Si CAST ALLOYS 

F. Grosselle, G. Timelli, F. Bonollo, A. Tiziani, E. Della Corte 

The influence of microstructure and process history on mechanical behaviour of cast Al-Si alloys is reported. 

In the present work, the EN-AC 46000 and 46100 aluminium alloys have been gravity cast using a stepbar 

permanent mould, with a range of thickness going from 5 to 20 mm. Metallographic and image analysis 

techniques have been used to quantitatively examine the microstructural parameters of the α-Al phase 

and eutectic Silicon. Microstructure has been also correlated with the results coming from the numerical 

simulation of the casting process. The results show that SDAS and length of eutectic silicon particles increase 

with section thickness, and consequently mechanical properties decrease. 

KEYWORDS: aluminium alloys; EN-AC 46000; EN-AC 46100; SDAS; eutectic Si; microstructure; numerical 

simulation; permanent mould casting 

INTRODUCTION 

Mechanical properties of Al-Si cast alloys depend on several microstructural 

parameters. Grain size, secondary dendrite arm 

spacing (SDAS), distribution of phases, the presence of secondary 

phases or intermetallic compounds, the morphology of silicon 

particles (size, shape and distribution) and, finally, defects play a 

key role in the determination of the elastic and plastic behaviour 

of aluminium alloys [1-3]. 

In general, castings having a finer microstructure (quantitatively 

described by low SDAS values), induced by high solidification 

rate, show better mechanical properties. Many correlations between 

mechanical behaviour (UTS, YS, elongation) and SDAS can be 

found in literature [3-4]. It is worth mentioning that, on industrial 

production, the control of solidification rate (and therefore the 

SDAS values) is quite difficult to achieve [5], as consequence of 

the geometrical complexity and of the different wall thickness in 

the real-shaped casting. For this reason, reference castings are frequently 

employed when the solidification rate has to be accurately 

controlled and different microstructures have to be achieved. Therefore, 

in these castings, the solidification conditions can be set up 

by varying the thickness and the material of the mould, as well as 

Fabio Grosselle, Giulio Timelli, Franco Bonollo, Alberto Tiziani 

Dipartimento di Tecnica e Gestione dei Sistemi Industriali 

DTG, Università di Padova, Stradella S. Nicola, 3 I-36100 Vicenza, 

Italia (Email: grosselle@gest.unipd.it; 

Tel. 00 39 0444 99 87 54; Fax No. 00 39 0444 99 88 89) 

Emilia Della Corte 

Enginsoft Spa, via Giambellino 7, I-35129 Padova, Italia 

(Email: e.dellacorte@enginsoft.it; Tel. 00 39 049 77 05 311) 

the sample size [3,6]. In this way, the factors affecting SDAS, the 

relationship between SDAS and mechanical properties of cast aluminium 

alloys can be easily better assessed and these information 

can be subsequently transferred to real-shaped casting. 

On the other hand, it is well known that SDAS is not the only factor 

affecting the mechanical behaviour of an alloy. For instance, 

the deformation behaviour of cast aluminium alloys is also affected 

by eutectic Si particles and intermetallic compounds which 

determine the initiation and the evolution of fracture [7-8]. In particular, 

for defect free castings, tensile fracture is initiated by cleavage 

of either brittle intermetallic particles or eutectic Si particles. 

The cleavage cracks are mainly perpendicular to the macroscopic 

principal strain, regardless of the particle orientation. Platelet particles 

with their length perpendicular to the tensile direction break 

because of cleavage along their length [9-10]. Therefore, it is easy 

to hypothesize that the fracture mechanism depends on the size 

and shape of Si or Fe-rich brittle phases. In detail, large and acicular 

particles are deleterious for mechanical properties reducing 

elongation to fracture and ultimate tensile strength [9]. 

In un-modified Al–Si cast alloys, the eutectic Si particles have a 

coarse, acicular and polyhedral morphology and the final mechanical 

properties of an alloy is characterized by their distribution 

in the microstructure. It was established that the size distribution 

of eutectic Si particles follows the lognormal distribution [8]. The 

probability density function of the three-parameters lognormal distribution 

can be written as: 

(1) 

where d is the diameter of Si particles, τ the threshold, σ the shape 

and μ is the scale parameter. 

la metallurgia italiana >> giugno 2009 25

Alluminio e leghe

Alloy 

EN-AC 

46000 

EN-AC 

46100 

Memorie >> 

Al 

Bal. 

Bal. 

Si 

8.8 

10.8 

s 

Tab. 1 

Chemical composition of the alloys studied in the 

present work (wt.%). 

Composizione chimica delle leghe analizzate nel presente studio (wt.%). 

The tensile specimens were 100-mm long, 20-mm wide, and 3-mm 

thick, with a gage length of 30 mm and a width of 10 mm, according 

to ASTM-B577. 

The tensile tests were done on a computer controlled tensile testing 

machine. The crosshead speed used was 2 mm/min (strain 

rate~10 -3 s -1 ). The strain was measured using a 25-mm extensometer. 

At least three specimens were tested for each zone. When the 

experimental data differed by more than 5 pct, another tensile specimen 

was tested. 

The samples cut from the cross section of the gage length were 

mechanically prepared to a 1-µm finish with diamond paste and, 

finally, polished with a commercial fine silica slurry for metallographic 

investigations. Microstructural analysis was carried out 

using an optical microscope and quantitatively analyzed using 

an image analyzer. To quantify the microstructural features, the 

image analysis was focused on the secondary dendrite arm spacing 

(SDAS), and on the size and aspect ratio of the eutectic silicon 

particles. Size is defined as the equivalent circle diameter (d); the 

aspect ratio (α) is the ratio of the maximum to the minimum Ferets. 

To obtain a statistical average of the distribution, a series of at least 

10 photographs of each specimen were taken; each measurement 

included more than 1000 particles. The secondary phases, such 

as the Mg 2 Si and CuAl 2 particles, and the iron-rich intermetallics 

were excluded from the measurements and further analysis. Average 

SDAS values were obtained using the linear intercept method, 

which involves measuring the distances between secondary 

dendrite arms along a line normal to the dendrite arms. 

Casting simulation 

The MAGMASOFT® v4.6 (2007) commercial software, with its 

module for gravity die casting, was used for numerically simulating 

the filling and solidification behaviour of analysed castings. 

The characteristics of the software used in this study are as follows: 

- ease of physical interpretation of various steps of algorithms; 

- conservation of physical properties; 

- reduction of solving time. 

Basic governing equations of the software are continuity equation, 

Navier–Stoke’s equation, energy equation and volume of fluid 

(VoF) method for the free surface movement during the die filling. 

The numerical code employs the finite volume approach to convert 

differential equations into algebraic ones and solve them on 

a rectangular grid. The CAD model of the step casting was drawn 

and imported in the simulation software where a controlled volume 

mesh of 132000 cells for the die cavity was automatically 

generated by the software. The initial conditions for numerical 

simulation were defined to reproduce the casting parameters. The 

pouring temperature was set at 720°C, while, for the die, the temperature 

for the first cycle was assumed to be at a uniform temperature 

of 250°C. In the subsequent cycles, the initial temperature 

in the die is taken to be the predicted temperature distribution at 

Cu 

3.0 

1.8 

Fe 

0.8 

0.8 

Mg 

0.21 

0.13 


Mn 

0.26 

0.18 

Ni 

0.087 

0.089 

la metallurgia italiana >> giugno 2009 27 

Ti 

0.039 

0.046 

Zn 

0.90 

1.29 

Others 

0.05 

0.05 

the end of the previous cycle. A number of 10–15 cycles were taken 

after the start up to reach a quasi-steady-state temperature in 

the die. The thermal conductivity of the die varied in the range of 

33.4–31.5 W/mK, in the working temperature range of 450–520°C. 

The other physical constants and properties of the die and the alloys, 

and their evolution with temperature, were chosen among 

those present in the software database, as well as the heat transfer 

coefficients (HTC), taking into account affecting parameters, like 

the type and thickness of coating, and the pouring temperature. To 

define the whole set of boundary conditions in the model, the process 

parameters (e.g. regarding the filling and cooling cycle) and 

the cycle time, acquired from the casting process, were imported 

in the software, increasing the reliability of numerical simulation. 

Virtual thermocouples were inserted in the different zones of the 

die in order to control the temperature profiles and to compare 

these values with the real ones. Solidification time was assessed 

via numerical simulation code in order to predict the final microstructure 

of the casting. The mechanical properties of the aluminium 

cast alloy were predicted by using the newly developed 

add-on module to the simulation software [12]. 

RESULTS AND DISCUSSION 

In Fig. 4, typical microstructures of the step castings are reported 

with reference to the different step, which varies in a range 

of thickness between 5 to 20 mm. While Fig. 4a shows the as-cast 

microstructure of EN-AC 46000 alloy, in Fig. 4b the microstructure 

of EN-AC 46100 alloy is presented. The microstructure of the step 

castings analysed consists of a primary phase, α-Al solid solution, 

which precipitates from the liquid as the primary phase in the form 

of dendrites. The as-cast Al–9Si–3Cu alloy (EN-AC 46000) shows 

primary α-Al grains in the matrix of the eutectic structure (Fig. 4a). 

The eutectic structure is a mixture of the α-Al and eutectic silicon 

phase. The eutectic silicon can be seen in the interdendritic regions. 

The Al–11Si–2Cu alloy (EN-AC 46100) shows the mixed structure 

of the α-Al grains and eutectic (Fig. 4b). Moreover, the addition of 

silicon increased the fraction of the eutectic in the interdendritic 

region. Intermetallics compounds, such as Fe- and Cu-rich intermetallics, 

were also observed. A low level of microdefects, in the 

form of microshrinkage, was found in the specimens analysed. 

The scale of microstructure in different zones of the castings was 

characterized by means of SDAS measurements and then correlated 

with mechanical properties. These data are further described. 

Fig. 4 presents calculated solidification times, from numerical simulation, 

with the corresponding microstructure within step castings 

of EN-AC 46000 and 46100 alloys. A general coarsening of 

microstructure occurs in thicker regions, quantified by SDAS values, 

as the result of the increased solidification time in both alloys. 

For every section thickness, the SDAS values were higher in the 

samples extracted from the inner section than the specimens from 

the external one. 

Similar values of SDAS and solidification time for the two alloys 

were obtained as consequence of similar thermal properties. 

Solidification times were also estimated by means of SDAS measurements 

using equation [13]:

Alluminio e leghe

Alloy 

EN-AC 46000 

EN-AC 46100 

Memorie >> 

Section 

A.i 

B.i 

C.e 

C.i 

D.e 

D.i 

A.i 

B.i 

C.e 

C.i 

D.e 

D.i 

SDAS (μm) 

16.4 (0.8) 

23.9 (1.8) 

23.5 (1.0) 

30.8 (0.9) 

25.0 (1.3) 

32.5 (1.9) 

18.6 (2.4) 

22.4 (1.6) 

26.8 (1.8) 

32.8 (1.3) 

26.9 (1.9) 

35.3 (2.1) 

Calculated solidification time (s) 

14 

40 

38 

82 

46 

95 

20 

33 

55 

98 

56 

120 

Similar behaviour of the eutectic Si size was observed in the EN- 

AC 46100 alloy. 

The impact of the solidification time on the aspect ratio of eutectic 

Si particles is negligible, as shown in Fig. 5b. Every step shows 

similar distributions of the aspect ratio of the eutectic Si particles. 

The irregular growth mechanism of un-modified eutectic Si particles 

confirms to be independent from the solidification rate, at 

least for the range of solidification rate investigated [14]. 

Generally, the distributions of the aspect ratio of the eutectic Si 

particles in EN-AC 46100 alloy show similar behaviour. 

The correlation between solidification time and Si particles parameters 

is reported in Fig. 6. For both alloys, the average diameter 

increases significantly by increasing the solidification time from 20 

to 40 seconds, while for longer times the values are steady in the 

range of 6.5 to 7 µm (Fig. 6a). On the other side, the aspect ratio 

seems to be independent from the solidification time (Fig. 6b), as 

previously demonstrated by means of the distribution plots in Fig. 

5b. 

Contrary, a relationship can be found between the aspect ratio and 

the Si content. If the Si amount is increased from 9 wt.% in the 

EN-AC 46000 alloy to 11 wt.% and EN-AC 46100 alloy, the aspect 

ratio of eutectic Si particles increases, indicating a more intense 

growing along the main axis direction of the Si particles. 

In Tab. 3, the results of the mechanical investigation are shown. 

The values of the standard deviation confirm the presence of a 

low amount of microdefects, which affect the mechanical properties. 

However, no macrodefects were observed through the X-ray 

investigation. 

If A.i and D.i sections are considered and compared, a reduction 

of 23 and 15% in UTS and 71 and 54% in elongation to fracture is 

observed for the EN-AC 46000 and 46100 alloys respectively, as a 

consequence of the different microstructure scale. In the EN-AC 

46000 alloy, the UTS varies from 167 to 218 MPa and the elongation 

to fracture from 0.4 to 1.4%, while the change is in the range 

of 160 to 188 MPa for UTS and from 0.6 to 1.3% for elongation to 

fracture, in the EN-AC 46100 alloy. On the other hand, the solidi- 


Equivalent Diameter, d (μm) 

5.3 (1.8) 

6.4 (3.0) 

6.4 (2.7) 

6.8 (3.1) 

7.4 (3.8) 

6.8 (3.3) 

5.5 (1.8) 

6.2 (2.6) 

6.7 (2.9) 

7.6 (3.8) 

6.7 (2.9) 

6.8 (3.0) 

Aspect Ratio, α 

3.0 (1.5) 

2.9 (1.6) 

2.9 (1.3) 

2.7 (1.4) 

3.0 (1.5) 

3.1 (1.4) 

4.0 (2.0) 

3.7 (2.0) 

3.6 (1.8) 

3.1 (1.7) 

4.0 (2.0) 

4.1 (2.3) 

s 

Tab. 2 

Average values of SDAS, equivalent diameter and aspect ratio of eutectic silicon particles obtained from different sections 

of the step castings (standard deviation in parentheses); solidification times, calculated with a numerical simulation approach, are 

also reported. Data refer to EN-AC 46000 and 46100 alloys. 

Valori medi di SDAS, del diametro equivalente e del rapporto d’aspetto per le particelle di silicio eutettico, ottenuti dalle diverse sezioni 

del getto a gradini (i valori di deviazione standard in parentesi); sono inoltri riportati i valori del tempo di solidificazione per le varie zone, 

calcolati mediante la simulazione di processo. I dati si riferiscono alle leghe EN-AC 46000 e 46100. 

s 

Fig. 6 

Variation in (a) equivalent diameter and (b) 

aspect ratio of eutectic silicon particles as function of 

solidification time. 

Variazione dei valori di (a) diametro equivalente e (b) del 

rapporto d’aspetto delle particelle di silicio eutettico in 

funzione del tempo di solidificazione. 


a 

b

Alluminio e leghe

CONCLUSIONS 

Memorie >> 

The effects of microstructural parameters such as SDAS, size and 

morphology of eutectic silicon particles on mechanical properties 

of EN-AC 46000 and 46100 alloys have been investigated. In addition, 

the validation of the results provided by a numerical simulation 

approach has been performed. Based on the results obtained in 

the present study, the following conclusions can be drawn. 

- The equivalent diameter and the aspect ratio of eutectic Si particles 

follow three-parameter lognormal distributions. 

- While the distribution of the equivalent diameter depends on solidification 

time, the distribution of the aspect ratio is less sensible, 

indicating the irregular growing mechanism of un-modified eutectic 

silicon. 

- The average size of eutectic silicon particles is similar in both EN- 

AC 46000 and 46100 alloys, while the aspect ratio of EN-AC 46100 

is higher, probably due to higher Si content. 

- The mechanical properties, i.e. the UTS and elongation to fracture, 

depend on SDAS and on the aspect ratio of the eutectic Si particles, 

which seems an alloy-related parameters. Increasing the SDAS and 

the aspect ratio values, the UTS and the elongation to fracture decrease. 

- The difference in the mechanical properties of the two alloys is the 

consequence of different chemical composition. Higher Cu and Mg 

contents in the EN-AC 46000 alloy allows to increase the YS, while 

a lower Si amount permits to enhance the ductility, reaching higher 

a 

b 

s 

Fig. 8 

Average (a) UTS and (b) elongation to fracture as a 

function of the combined parameter SDAS x Aspect ratio; 

coefficient of determination, R2, are given. Data refer to EN- 

AC 46000 and 46100 alloys. 

Valori medi di (a) UTS e (b) allungamento a rottura in funzione del 

prodotto tra SDAS e rapporto d’aspetto. E’ inoltre riportato in 

coefficiente di determinazione, R2. I dati si riferiscono alle leghe 

EN-AC 46000 e EN-AC 46100. 


s 

Fig. 9 

Comparison between experimental and simulated 

mechanical properties in EN-AC 46000 step casting. The 

images refer to (a) YS, (b) UTS and (c) elongation to fracture. 

Confronto tra i valori delle proprietà meccaniche ottenuti 

sperimentalmente e mediante simulazione di processo per il 

getto colato con lega EN-AC 46000. Le immagini si riferiscono a 

(a) YS, (b) UTS e (c) allungamento a rottura. 

UTS and elongation to fracture values than the EN-AC 46100 alloy. 

- Since numerical simulation results reproduce the experimental 

data with a good accuracy, it can be stated that numerical simulation 

is a useful tool for the reduction of time and costs in the design 

stage. 

- The present investigation has been carried out on un-modified 

gravity cast alloys; in the case of higher cooling rate, modification 

or heat treatment, attention should be also paid to the size of eutectic 

Si particle, as a parameter affecting the mechanical behaviour. 

ACKNOWLEDGMENTS 

The European Project NADIA- New Automotive components 

Designed for and manufactured by Intelligent processing of light 


a 

b 

c

Alluminio e leghe

Memorie >> 


ANALISI DELL’EVOLUZIONE 

MICROSTRUTTURALE DURANTE IL 

PROCESSO DI ESTRUSIONE DELLA LEGA 

AA6060 MEDIATE SIMULAZIONI FEM 

M. El Mehtedi, L. Donati, S. Spigarelli, L. Tomesani 

La previsione della microstruttura finale dopo l’estrusione delle leghe di alluminio è un argomento che ha 

suscitato un grande interesse negli ultimi anni, visto che le proprietà meccaniche e la qualità degli estrusi 

sono fortemente dipendenti dall’evoluzione e del tipo di microstruttura. Questo lavoro pone come obiettivo 

lo studio dell’evoluzione microstrutturale della lega di alluminio AA6060 durante l’estrusione, mediate 

simulazioni FEM utilizzando il Codice Deform 3D. Allo scopo di determinare i coefficienti dei modelli 

di ricristallizzazione da inserire nel codice FEM, sono state prodotte delle prove sperimentali mediante 

l’estrusione inversa di coppe a diverse temperature e velocità di deformazione. Dall’analisi metallografica 

dei campioni estrusi è stato possibile determinare i coefficienti del modello dinamico di ricristallizzazione in 

dotazione al codice FEM Deform 3D. Le coppe sono state successivamente trattate termicamente in forno per 

far avvenire la ricristallizzazione statica, e sono stati determinati i coefficienti del modello di ricristallizzazione 

statica. Una volta convalidati i modelli, si è passato alla simulazione del processo reale di estrusione di una 

billetta cilindrica. L’evoluzione della microstruttura presenta delle zone con dei grani allungati ed altre con dei 

grani ricristallizzati con fenomeni di accrescimento. I risultati delle simulazioni sono stati confrontati con le 

microstrutture delle billette estruse, mostrando una buona corrispondenza. 

PAROLE CHIAVE: alluminio e leghe, estrusione, deformazioni plastiche, simulazione numerica, processi 

INTRODUZIONE 

I modelli di previsione della microstruttura hanno suscitato un 

grande interesse da parte delle industrie negli ultimi anni, specialmente 

per quanto riguarda le leghe leggere ove le proprietà 

meccaniche sono fortemente dipendenti dalla microstruttura 

finale [1,2]. Inoltre è ben noto come la dimensione del grano e 

la precipitazione di fasi secondarie influenzano diversi aspetti 

del prodotto finale, quali, ad esempio, l’effetto estetico, la resistenza 

a trazione, la formabilità, la resistenza a fatica e a corrosione. 

La struttura a grana fine è particolarmente richiesta 

specialmente quando il prodotto viene sottoposto ad elevati 

carichi di fatica oppure viene messo in opera in atmosfera corrosiva 

[3,4]. Le proprietà meccaniche dei profilati in alluminio 

sono fortemente legate all’evoluzione della microstruttura du- 

M. El Mehtedi, S. Spigarelli 

Dipartimento di Meccanica, Università Politecnica delle Marche, 

60131 Ancona, Italia - e-mail: elmehtedi@univpm.it 

L. Donati, L. Tomesani 

DIEM, Università degli studi di Bologna, 40136 Bologna, Italia 

rante tutto il ciclo produttivo, dalla billetta prodotta per fusione 

fino al ciclo di invecchiamento per le leghe da trattamento 

termico [5]; precipitati intermetallici grossolani, zone libere 

da precipitati (Precipitate Free Zones), la distribuzione e le dimensioni 

dei grani, l’ingrossamento del grano rappresentano 

alcuni problemi legati all’evoluzione microstrutturale delle 

leghe di alluminio che possono indurre ad un prodotto finito 

povero dal punto di vista meccanico. L’ottenimento della microstruttura 

ottimale si è spesso basato sull’esperienza tramandata 

e solo di recente l’interesse verso i processi di simulazione 

è significativamente aumentato. 

La ricristallizzazione nelle leghe di alluminio è stata studiata 

in maniera molto approfondita negli ultimi decenni, soprattutto 

per gli aspetti legati al comportamento di queste leghe 

durante la deformazione a caldo [7,8], ma non esiste nessun lavoro 

fornisce delle equazioni affidabili oppure dei coefficienti 

da utilizzare nei sistemi di simulazione con gli elementi finiti. 

Durante tutto il processo termo-meccanico avvengono diversi 

meccanismi metallurgici come la ricristallizzazione statica 

(SRX), ricristallizzazione dinamica (DRX), la ricristallizzazione 

geometrica dinamica (GDRX), la crescita dei grani ed infine la 

precipitazione di fasi secondarie [9,10]. Durante il processo di 

lavorazione delle leghe di alluminio ed in particolare le leghe 


Alluminio e leghe

Memorie >> 

t – tempo (sec); d 0 – diametro iniziale del grano (μm); d – diametro 

del grano (μm); d rex – diametro del grano ricristallizzato 

(μm); d ave – diametro medio del grano (μm); X –frazione in 

volume ricristallizzata (X drx dinamica, per X=0 non è avvenuta 

ricristallizzazione; per X=1, la struttura è al 100% ricristallizzata); 

ε - deformazione reale; ε c – deformazione critica (al di 

sopra della quale inizia la ricristallizzazione dinamica); ε p – deformazione 

di picco (corrispondente al valore massimo di tensione); 

ε 0.5 – deformazione al 50% ricristallizzato; έ - velocità di 

deformazione (sec-1); t 0.5 – tempo al 50% ricristallizzato (sec); 

a 1–10 – coefficienti del materiale ottenuti sperimentalmente; h 1–8 

, n 1–8 e m 1–9 sono gli esponenti del grano, della deformazione e 

della velocità di deformazione rispettivamente; Q 1–8 – energie 

di attivazione; β d - kd sono dei coefficienti del materiale ottenuti 

sperimentalmente. 

PROCEDURE SPERIMENTALI E DISCUSSIONE 

Prima fase-prove di laboratorio 

Nella prima fase, è stata messa a punto una procedura sperimentale 

allo scopo di valutare la distribuzione e la grandezza 

dei grani dopo deformazione della lega AA6060 ricevuta allo 

stato di omogeneizzazione con la seguente composizione chimica 

(% peso): 

Grazie alla strumentazione mostrata in Fig. 1, sono state prodotte 

delle coppe a spessore variabile mediante estrusione inversa 

alle temperature 250, 350, 450 e 550° C con diverse velocità 

di traversa 0.1 e 5 mm/sec. 

Le condizioni di prova sono state scelte per ottenere una deformazione 

locale compresa tra 0-3.88 con una velocità di deformazione 

0-5 s-1; la temperatura dei campioni era tipica dei 

processi di estrusione al livello industriale. Lo stampo ed il 

campione sono stati scaldati in un forno a resistenza e la temperatura 

del campione veniva controllata mediante termocoppia 

a contatto. I campioni sono stati spenti in acqua immediatamente 

dopo deformazione per conservare la microstruttura 

inalterata. I campioni deformati sono stati tagliati, inglobati e 

lucidati; per l’osservazione al microscopio ottico con luce polarizzata, 

i campioni sono stati attaccati elettroliticamente con 

l’attacco Barker, infine è stata effettuata un’analisi accurata 

della distribuzione e della grandezza dei grani. In parallelo, 

è stato simulato tutto il processo sperimentale di deformazione 

mediante simulatore di elementi finiti FEM Deform 3D allo 

scopo di valutare la distribuzione della deformazione e la velocità 

di deformazione in tutti i campioni. Grazie al confron- 

s 

Fig. 1 

Strumentazione dell’estrusione inversa. 

Inverse extrusion equipment. 


Lega 

AA6060 

Temperatura 

250° C 

350° C 

450° C 

550° C 

n8 -0.364 

-0.985 

-0.722 

-0.420 

m8 -0.213 

-0.105 

-0.084 

0.046 

s 

Tab. 1 

Coefficienti ottenuti dell’equazione (5). 

Regressed coefficienst of equation (5). 

a8 1.93E+15 

7.22E+12 

1.34E+11 

8.26E+09 

to delle misurazioni sperimentali dei grani con le condizioni 

locali di deformazione sono stati ricavati i diversi coefficienti 

del modello di ricristallizzazione dinamica: a 2 =0.05, ε 0.5 =0.15, 

a 5 =0.15, β d =1 kd=1, a 10 =1, n 8 , m 8 , a 8 (Tab. 1), h 5 =n 5 =m 5 =Q 5 =0. 

Una più ampia e dettagliata descrizione della procedura di calcolo 

dei coefficienti è riportata da Donati et.al. [13]. 

La microstruttura del materiale allo stato di fornitura è costituita 

da grani ben definiti ed equiassici con un diametro medio 

di circa 135 μm decorati da grosse particelle lungo il bordo 

di grano; si nota inoltre la presenza di composti intermetallici 

all’interno dei grani. La Fig. 2a mostra la distribuzione della 

microstruttura nei campioni deformati, si nota la presenza di 

grani fortemente allungati in prossimità del punzone dove 

il valore stimato della deformazione è 2.5-3.88, mentre nelle 

zone vicine allo stampo si ha una struttura meno deformata 

e più equiassica con valori della deformazione minori a 0.8; 

si evidenzia una riduzione nello spessore dei grani passando 

dalla zona vicina al punzone da 20-50 μm fino alla zona a contatto 

dello stampo con valori di 80-110 μm. Questa diminuzione 

è dovuta al tipo di deformazione che ha subito il materiale 

(lungo la direzione assiale), mentre solo in alcune piccole zone 

si vede l’effetto della ricristallizzazione geometrica dinamica 

(GDRX) che ha generato nuovi grani senza nucleazione; infatti, 

i grani sono fortemente deformati lungo la direzione di 

estrusione e quando lo spessore del grano raggiunge quello 

del sottograno si formano questi nuovi grani equiassici. 

Confrontando i risultati sperimentali delle dimensioni dei 

grani e la loro distribuzione con i dati forniti da modello di 

previsione FEM (fig.2b) lungo tutta la sezione dei campioni, 

si ha un errore medio di circa il 12% con una massima deviazione 

del 57%, conseguendo un interessante accordo con i dati 

sperimentali, come riportato in [13]. La maggior deviazione si 

trova alle basse temperature di deformazione e nella zona alta 

delle coppe dove un non perfetto centraggio del punzone (nelle 

prove sperimentali) potrebbe indurre a delle deformazioni 

non omogenee producendo degli spessori diversi. 

Seconda fase - Trattamento termico 

In questa seconda fase, sono stati trattati termicamente 8 campioni 

in forno a 550°C per 30 minuti; tali campioni, raffreddati 

in aria, successivamente sono stati riportati a 180°C per 10 

ore. Seguendo la stessa procedura della prima fase, i campioni 

sono stati preparati per l’analisi metallografica per la misura 

del diametro medio del grano. Fig. 3 mostra la presenza di 

grani equiassici in tutti i campioni trattati, si nota la completa 

ricristallizzazione statica ed un aumento del diametro medio 

del grano; in particolare si nota una crescita dei grani nei campioni 

deformati alle basse velocità ed alle alte temperature, 

e a temperature superiori a 450°C siamo in presenza di una 

crescita abnorme dei grani nella zona interna del bicchierino 

(PCG) (Fig. 3-c). A 350°C il diametro medio del grano va da 

155 μm vicino al punzone fino a 220 μm sul fondo del bicchiere 


Alluminio e leghe

Memorie >> 

s 

Fig. 4 

Mostra alcune billette estruse a 0.3 mm/s. 

the billets extruded at 0.3 mm/s. 

della billetta in uscita iniziava immediatamente a raffreddarsi 

mentre la parte rimasta all’interno del container rimaneva 

ancora calda fino alla totale estrazione. Vista l’impossibilità di 

raffreddare immediatamente la billetta dopo l’estrusione si è 

tenuto conto durante le successive analisi sia della ricristallizzazione 

statica (dovuta alla prolungata permanenza a circa 

450°C) sia dell’accrescimento del grano. Il carico massimo 

raggiunto durante la fase di estrusione era di 3MN. Il flusso 

del materiale e la struttura cristallina sono stati analizzati in 

tutte le zone della billetta, poiché l’estrusione è stata fermata al 

raggiungimento del 50% dell’altezza della billetta come corsa 

del punzone (Fig. 4). 

Le billette sono state tagliate lungo un piano di simmetria e 

preparate per le macro e micro analisi; come si vede in Fig. 5, 

la macro struttura dimostra una morfologia diversa dei grani 

a seconda la posizione in cui si trovano. Sono state individuate 

4 zone: 

zona I, detta zona morta; in questa zona il materiale non subisce 

scorrimento né deformazioni significative all’avanzare del 

pistone (esclusa la fase iniziale del processo). 

La microstruttura nella zona (A1) (Fig. 

5), è dominata da grani equiassici con un 

diametro medio 110 μm, non si evidenzia 

presenza di ricristallizzazione né fenomeni 

di accrescimento vista la quasi assenza 

della deformazione. 

Zona II, in cui il materiale subisce un forte 

scorrimento; i grani sono sottoposti a 

una deformazione di taglio a causa delle 

condizioni al contorno. I grani nella zona 

(A2) risultano molto deformati ed allungati 

lungo le linee di flusso del materiale 

ed orientati verso il foro di uscita della 

matrice, hanno lunghezza media 320 μm 

e spessore medio di 54 μm, con un valore 

del diametro medio equivalente di circa 

120 μm. Non risultano fenomeni significativi 

di ricristallizzazione e di crescita dei 

grani deformati. 

Zona III; al contrario della zona II il ma- 

s 

Fig. 6 

analisi EBSD della zona (A5). 

EBSD analysis of (A5) location. 


s 

Fig. 5 

Macro e micro analisi lungo una sezione della 

billetta estrusa. 

Macro and micro analysis of the extruded rest. 

teriale scorre direttamente verso il foro di uscita della matrice, 

e i grani vengono semplicemente traslati fino al foro di uscita; 

il materiale in questa zona è soggetto ad elevata pressione 

idrostatica ed elevata temperatura, mentre la deformazione e 

la velocità di deformazione aumentano man mano che il materiale 

si avvicina al foro di uscita. All’uscita (A5), il materiale 

presenta una struttura composta da grani ricristallizzati con 

un diametro medio di 50 μm, ma persiste la presenza di grani 

fortemente allungati di spessore medio pari a circa 100 μm, 

come si vede anche in modo chiaro dall’analisi EBSD fatta nella 

stessa zona (A5), mostrata in Fig. 6. 

Infine, nella zona IV (zona di uscita dallo stampo), a causa 

dell’influenza dei valori elevatissimi di tutti i parametri del 

processo visti finora, la struttura si presenta abbastanza complessa; 

al centro valgono le condizioni descritte per la zona 

(A5), mentre vicino alla superficie, siamo in presenza di una 

corona composta da grani completamente ricristallizzati con 

una crescita abnorme (A4) con dei grani che raggiungono i 

1300 μm di diametro medio. Questo fenomeno è causato dalle 

condizioni estreme di deformazione, velocità di deformazione, 

elevato attrito (che causa un aumento della temperatura 

nella zona periferica) e di raffreddamento. Lo stesso fenomeno 

è stato osservato nelle prove fatte sui bicchierini nella fase pre- 


Alluminio e leghe

Memorie >> 

simulare un processo reale di estrusione. Dal confronto tra i risultati 

sperimentali e quelli delle simulazioni, risulta che i dati 

delle simulazioni hanno una buona concordanza con quelli 

sperimentali, eccetto che nelle zone periferiche dove appaiono 

fenomeni di accrescimento del grano, che dimostrano una 

carenza del codice. 

RIFERIMENTI BIBLIOGRAFICI 

1] T. Sheppard, 2006. Prediction of structure during shaped 

extrusion and subsequent static recrystallisation during the 

solution soaking operation, Journal of Materials Processing 

Technology vol.177, pp. 26–35. 

2] A.R. Bandar, S. R. Claves, J. Lu, K. Matous, W. Z. Misiolek, 

A. M. Maniatty 2004 “Microstructural Evaluation of 6xxx Aluminum 

Alloys for Computer-Simulated Texture Prediction”, 

Aluminum Extrusion Technology Seminar ET 2004, Orlando, 

vol 1, pp169-176. 

3] B. Dixon, “Extrusion of 2xxx and 7xxx alloys”, Aluminum 

Extrusion Technology Seminar, Chicago,2000, vol1, pp281- 

2940. 

4] T. Sheppard, “Development of structure recrystallization kinetics 

and prediction of recrystallised layer thickness in some 

Al-alloys”, Aluminum Extrusion Technology Seminar, Chicago,1996, 

vol1, pp163-170. 

5] J.M.C. Mol, J. van de Langkruis, J.H.W. de Wit and S. van 

der Zwaag “An integrated study on the effect of pre- and 

post-extrusion heat treatments and surface treatment on the 

filiform corrosion properties of an aluminium extrusion alloy” 

Corrosion Science, Volume 47, Issue 11, November 2005, Pages 

2711-2730 

6] F. J. Humphreys, M. Hatherly “Recrystallization and Related 

Annealing Phenomena” Pergamon Press Inc, Oxford, 1995 

ISBN 978-0080418841 

ANALYSIS OF THE MICROSTRUCTURAL EVOLUTION 

DURING HOT EXTRUSION OF AA6060 BY MEANS 

OF FEM SIMULATION 

Keywords: FEM, recrystallization, extrusion, aluminium 

alloy 

In this work an experimental methodology to evaluate the prediction of 

recrystallized structures in aluminum extrusion was presented and validated. 

In the first part of the work an experimental procedure to investigate 

the evolution of recrystallization in aluminum alloys is presented 

and discussed. Several cups, obtained by means of inverse extrusion, were 

produced at different temperatures and process speeds. The specimens 

were analyzed in order to examine the grain size distribution. The coefficients 

for dynamic recrystallization models were obtained by regression 

ABSTRACT 


7] Gourdet, S. Montheillet, F. “Experimental study of the recrystallization 

mechanism during hot deformation of aluminium” 

Materials Science and Engineering A: Structural Materials: 

Properties, Microstructure and Processing, v 283, n 1-2, 

May, 2000, p 274-288 

8] J. G. Byrne, Recovery, Recrystallization, and Grain Growth, 

(New York: MacMillman, 1965), 93-109. 

9] R. D. Doherty, D. A. Hughes, F. J. Humphreys, J. J. Jonas, D. 

Juul Jensen, M. E. Kassner, W. E. King, T. R. McNelley, H. J. Mc- 

Queen and A. D. Rollett “Current issues in recrystallization: 

a review” Materials Science and Engineering A, Volume 238, 

Issue 2, 15 November 1997, Pages 219-274 

10] T. Pettersen, B. Holmedal, E. Nes, “Microstructure development 

during hot deformation of aluminum to large strains” 

Metallurgical and Materials Transactions A: Physical Metallurgy 

and Materials Science, v 34, n 12, December, 2003, p 2737- 

2744 

11] J. Fluhrer, “DEFORMTM 3D User’s Manual Version 6.0” 

Scientific Forming Technologies Corporation, 2006 

12] G. Shen, S.L Semiatin, and R. Shivpuri, “Modeling Microstructure 

Development during the Forging of Waspaloy”, 

Metallurgical and Materials Transactions A, 26A (1995), 1795- 

1803. 

13] L. Donati, J. Dzwonczyk, J. Zhou, L. Tomesani “Microstructure 

prediction of hot-deformed aluminum alloys” accepted 

for publication on Key Engineering Materials (Proceedings of 

Extrusion Workshop and Benchmark 2007), Trans Tech (2007); 

14] T. Sheppard, Metallurgical Aspects of Direct and Indirect 

Extrusion, Proc. of the 4th Aluminum Extrusion Technology 

Seminar, (1984), 107-124 

15] M. Schikorra, L. Donati, L. Tomesani, M. Kleiner “The role 

of friction in the extrusion of AA6060 aluminum alloy, process 

analysis and monitoring”, Journal of Materials Processing 

Technology, 191 (2007) pp. 288–292. 

analysis after thermo-mechanical FEM simulations of the experiments 

realized with the code Deform 3D. A complete set of coefficients was regressed 

for the available microstructure evolution models inside the code 

environment. The specimens were then heated in a furnace and cooled 

in order to reproduce static recrystallization of the material. The grain 

distribution was examined and the coefficients for the equation for static 

recrystallization prediction were regressed, too. In the second part of the 

work the extrusion of a round-shaped profile is described and the grain 

size distribution on the profile and on the billet rest is analyzed. The obtained 

models were applied to the real extrusion of a round profile and a 

comparison between experimental measurements and simulation results 

was performed. The simulated results were in very good agreement with 

experimental data, except in zones where peripheral coarse grain and 

grain growth appeared. Here, a further investigation effort and specific 

modeling equations are required. 


Memorie >> 

Acciaio 

PRECIPITATION STRENGTHENING 

PRODUCED BY THE FORMATION 

IN FERRITE OF Nb CARBIDES 

M. A. Altuna, A. Iza-Mendia, I. Gutierrez 

A Nb microalloyed steel has been thermomechanically processed at laboratory through the use of plane strain 

compression sequences followed by simulated coiling. Tensile samples have been machined from the obtained 

specimens in order to investigate the effect of different variables: recrystallisation or accumulated strain before 

transformation, holding in austenite and coiling temperature on the final mechanical behaviour. Transmission 

electron microscopy observation of the precipitates has been carried out after coiling at different temperatures. 

It has been shown that when Nb remains in solution in austenite after hot deformation, it can precipitate in 

ferrite, leading to an important strengthening effect which is directly related to the concentration of Nb in 

solution before transformation and coiling temperature. 

KEYWORDS: thermomechanical processing, niobium, coiling, precipitation strengthening, ferrite, microstructure, 

tensile 

INTRODUCTION 

Niobium delays austenite recrystallization during hot rolling 

interpass times due to solute drag when being in solution and 

also as a consequence of carbonitride strain induced precipitation 

[1,2,3]. These two phenomena result in a final ferrite 

grain refinement [4,5,6,7]. Precipitates formed in austenite 

are subjected to a relatively fast coarsening and lose part of 

their potential efficiency as ferrite strengtheners. Precipitation 

of Nb in ferrite is expected to be significantly finer and 

contribute to increase the tensile properties. Although extensive 

investigations have been carried out to characterize Nb 

(C,N) precipitation in austenite, there are fewer studies concerning 

precipitation of NbC in ferrite. Additionally, there are 

some controversial results, depending on the source. Some 

authors claim that when some Nb is left in solution after hot 

working, precipitation can take place during the coiling [8,9], 

while others consider that homogeneous precipitation of 

NbC is suppressed below about 700ºC [10,11] and that there 

is no precipitation of Nb during coiling [12,13]. 

The purpose of this investigation was to study the eventual 

precipitation of Nb in ferrite during coiling and in the case of 

precipitation to estimate its contribution to the strength. 

EXPERIMENTAL 

An industrial Nb-steel [14] with composition: 0.06C - 0.35Si 

M. A. Altuna, A. Iza-Mendia, I. Gutierrez 

CEIT and TECNUN (Univ. Navarra), Donostia-San Sebastián, Spain 

Paper presented at the 3rd International Conference 

Thermomechanical Processing of Steels, organised by AIM; Padova, 

10-12 September 2008 

- 1Mn - 0.05Al - 0.0056N - 0.056Nb - 0.002Ti (wt-%) has been 

the base for the present work. Multipass thermomechanical 

plane strain compression tests were performed after reheating 

the specimens at 1250ºC for 15 minutes in order to assure 

the total dissolution of Nb, followed by fast cooling to 

the deformation temperature. 

Two different deformation sequences were applied in order 

to condition the austenite. The thermal cycle of the former 

sequence is shown schematically in Fig. 1. One deformation 

pass was applied at 1100ºC at a strain rate of 1 s -1 and 

a strain ε=0.3, followed by holding time at the same temperature 

during 20s in order to ensure the recrystallization 

s 

Fig. 1 

Example of the thermal cycle applied for onepass 

deformation sequence plus coiling. 

Esempio di ciclo termico, per sequenza di deformazione a 

passaggio singolo seguito da avvolgimento. 


Acciaio

Memorie >> 

s 

Fig. 4 

Ferrite-pearlite microstructures obtained after 

two-pass deformation sequence and coiling at different 

temperatures. The microstructure of the reference 

test (1h at 870ºC) is also shown for comparison. 

Microstrutture di ferrite – perlite ottenute dopo una 

sequenza di deformazione a due passaggi e avvolgimento 

a diverse temperature. A fini comparativi viene mostrata 

anche la microstruttura delle prove di riferimento (1h a 

870ºC). 

650ºC, leading to some polygonal ferrite being formed at prior 

austenite grain boundaries. 

Subsequent fast cooling to 500ºC and simulated coiling at this 

temperature leads to some acicular ferrite and finally pearlite 

produces from the carbon enriched austenite. 

The microstructures obtained at different coiling temperatures 

after a two-pass deformation sequence, are shown in Fig. 4. 

Ferrite-pearlite microstructures are obtained at coiling temperatures 

between 750 and 600ºC. As can clearly be seen, the application 

of the second deformation pass produces a significant 

refinement of the final microstructure which can be attributed 

to the accumulated strain in austenite before transformation. 

The volume fraction of ferrite and mean linear intercept are 

shown in Tab. 1 for the different tests. Pearlite is in the range 

3-7% whereas ferrite grain size varies with the applied conditions. 

Decreasing the coiling temperatures produces some 

grain refinement for the one-pass deformation sequence, passing 

from 33 μm when coiling is performed at 750ºC to 23 μm 

at 600ºC. The effect of the coiling temperature on the ferrite 

grain size when applying a two pass deformation sequence is 

negligible leading to a value around 14 μm. 

Holding the specimen during 1h at 870ºC has a small effect on 

the ferrite grain size after one-pass deformation sequence, but 

gives higher grain sizes after two-pass. 

Tensile data have been plotted in Fig. 5 as a function of coiling 

temperature for the two deformation sequences. The reference 

tests have been identified by R. It is evident that the 

decrease of the coiling temperature has a strengthening effect, 

excepting for the case in which a 1h holding at 870ºC was applied 

before cooling to the coiling temperature. The difference 

in yield stress between the two materials coiled at 650ºC with 

Acciaio 

s 

Fig. 5 

Mechanical behaviour, as a function of the 

coiling temperature. R= reference test (1h at 870ºC): 

a) 1 deformation pass and b) two deformation pass 

sequences. 

Comportamento meccanico, in funzione della 

temperatura di avvolgimento. R = prova di riferimento 

(1h a 870ºC): a) un passaggio di deformazione e b) due 

passaggi di deformazione. 

s 

Fig. 6 

Model predictions [15] for the precipitation of 

Nb in austenite at 870ºC considering precipitation on a 

recrystallized austenite or strain induced precipitation. 

Previsioni da modello [15] per la precipitazione di Nb 

in austenite a 870°C considerando la precipitazione 

su austenite ricristallizzata o precipitazione indotta da 

deformazione. 


Acciaio

Memorie >> 

s 

Fig. 10 

TEM image of a thin foil showing fine 

precipitates homogeneously distributed in ferrite. 

1-pass sequence and coiling at 600ºC. 

Immagine TEM di una lamina sottile che mostra 

precipitati fini distribuiti omogeneamente nella ferrite. 

Sequenza di deformazione a passaggio singolo, 

permanenza 1h a 870ºC e raffreddamento a 650ºC. 

temperature goes in agreement with the observed increase of 

the density of fine precipitates in ferrite. The size of the precipitates 

has been determined from TEM images obtained on 

thin foils. The particle size distributions for different processing 

conditions are shown in Fig. 11. Only particles with sizes 

lower than 20 nm have been taken into consideration for this 

quantification. 

The finest precipitation has been observed in the sample 

coiled at 750ºC, leading to a mean value of 6 nm, with about 

40% of the precipitates with sizes in the range 4-6 nm. The 

coarsest particles are observed in the specimen held during 

1h at 870ºC and coiled at 650ºC. In this case, the mean particle 

size is around 10nm. In the specimen coiled at 600ºC, 

around the 55% of the precipitates have sizes between 6 and 

8 nm, but an important fraction of particles are over this 

range. The result is a mean precipitate size of about 9 nm. It 

has to be mentioned that was not always possible to obtain 

a dark field image of the precipitates because quite often the 

contribution of the precipitates to the diffraction pattern was 

quite weak. As a consequence, the obtained particle sizes are 

probably slightly overestimated. 

Nb microalloying is generally used to condition the austenite 

by thermomechanical processing and obtain fine ferrite 

grain sizes leading to improved mechanical properties. In 

addition to this, it is generally accepted that Nb in solution 

in austenite leads to some strengthening of the final microstructure 

that cannot be attributed to the ferrite grain size 

refinement. However, there is some controversy when trying 

to explain the metallurgical mechanism which is responsible 

of this strengthening. Some authors attribute it to a high dislocation 

density at the interior of the non-polygonal ferrite 

grains usually observed in Nb containing steels, while others 

attribute it to precipitation. 

In the present case, the ferrite microstructure is relatively 

coarse because the deformation sequences were designed 

aiming to investigate the eventual precipitation of Nb in ferrite 

over reaching austenite refinement. Some ferrite grains 

Acciaio 

s 

Fig. 11 

Particle size distributions obtained by TEM 

after one-pass deformation sequence a) 1h at 

870ºC+coiling at 650ºC, b) coiling at 750ºC and c) 

coiling at 600ºC. Only particles with sizes lower than 

20 nm have been considered. 

Distribuzioni della dimensione delle particelle ottenute 

mediante TEM dopo sequenza di deformazione a 

passaggio singolo a) 1h a 870ºC + raffreddamento a 

650ºC, b) avvolgimento a 750°C e c) avvolgimento a 

600°C. Sono state considerate solo le particelle con 

dimensioni inferiori a 20 nm. 

present irregular shapes while others are polygonal. This is 

true for all the processing conditions, including the one in 

which the holding in austenite before coiling has produced 

an important precipitation before transformation. This indicates 

that even a low fraction of Nb in solution is able to 

produce non-polygonal grains. 

TEM observations do not indicate the presence of a high volume 

fraction of dislocations, but clearly demonstrate precipi- 


a 

b 

c

Acciaio

Memorie >> 

AKCNOWLEDGEMENTS 

This work has been carried out with a financial grant from 

the Research Fund for Coal and Steel of the European Community 

and from the Programa de Acciones Complementarias 

CICYT MAT2004-0048-E, Ministerio de Educación y 

Ciencia, Spain. The authors would like to acknowledge L. 

Mujica, S. Martin for performing Thermomechanical testing 

and to C. Iparraguirre for her contribution to TEM work. 

REFERENCES 

1] B. DUTTA, C. M. SELLARS, Materials Science and Technology, 

3, (1987), p. 197. 

2] W. J. LIU, J. J. JONAS, Metallurgical Transactions A, 20A, 

(1989), p. 689. 

3] R. ABAD, A. I. FERNANDEZ, B. LOPEZ, J. M. RODRIGU- 

EZ-IBABE, ISIJ International, 41, (2001), p. 1373. 

4] E. J. PALMIERE, C. I. GARCIA, A. J. DEARDO, Metallurgical 

and Materials Transactions A, 27A, (1996), p. 951. 

5] O. KWON, A. J. DEARDO, Acta Metallurgica et Materialia, 

39, (1991), p. 529. 

6] Q. B. YU, Z. D. WANG, X. H. LIU, G. D. WANG, Materials 

Science and Engineering A, 379A, (2004), p. 384. 

7] S. YAMAMOTO, C. OUCHI, T. OSUKA, “Thermomechanical 

processing of microalloyed austenite”, Ed. A.J. DeArdo, 

G.A. Ratz, P.J. Wray, The Metallurgical Society of AIME, 

Warrendale, PA, (1982), p. 613. 

8] R. D. K. MISRA, H. NATHANI, J. E. HARTMANN, F. SICILI- 

ANO, Materials Science and Engineering A, 394A, (2005), p. 339. 

MIGLIORAMENTO DELLE CARATTERISTICHE 

MECCANICHE MEDIANTE PRECIPITAZIONE 

PRODOTTO DALLA FORMAZIONE DI CARBURI DI Nb 

NELLA FERRITE 

Parole chiave: acciaio, processi, precipitazione 

Un acciaio microlegato al Nb è stato trasformato termomeccanicamente 

in laboratorio attraverso sequenze di compressione planare seguite 

da avvolgimento simulato. Da questo materiale sono stati ricavati 

provini di trazione mediante lavorazione dei pezzi ottenuti, al fine di 

ABSTRACT 

Acciaio 

9] S. SHANMUGAM, N. K. RAMISETTI, R. D. K. MISRA, 

T. MANNERING, D. PANDA, S. JANSTO, Materials Science 

and Engineering A, 460A, (2007), p. 335. 

10] A. J. DEARDO, International Materials Reviews, 48, 

(2003), p. 371. 

11] T. SAKUMA, R. W. K. HONEYCOMBE, Metal Science, 

18, (1984), p. 449. 

12] V. THILLOU, M. HUA, C. I. GARCIA, C. PERDRIX, A. J. 

DEARDO, Materials Science Forum, 284-286, (1998), p. 311. 

13] H. J. KESTENBACH, S. S. CAMPOS, J. GALLEGO, E. 

V. MORALES, Metallurgical and Materials Transactions A, 

34A, (2003), p. 1013. 

14] I. GUTIERREZ, M. A. ALTUNA, G. PAUL, S.V. PARK- 

ER, J. H. BIANCHI, P. VESCOVO, C. MESPLONT, M. WO- 

JCICKI, R. KAWALLA ‘ Mechanical Property Models for 

high strength complex microstructures (MEPMO), RFCS 

project, contract number: RFS-CR-03009. Draft final report, 

march (2007). 

15] B. López, MOFIPRE model, CEIT internal Report, 

(2007). 

16] F.B. PICKERING, Materials Science and Engineering, Ed. 

R.W. Cahn, P. Haasen, E.J. Kramer, Vol. 7, Constitution and 

Properties of Steels, Ed. F.B. Pickering, VCH, (1993), p. 47. 

17] E. OROWAN, “Internal stresses in metals and alloys”, 

Ed. The Institute of Metals, (1948), London, p. 451. 

18] M. F. ASHBY, Acta Metallurgica, 14, (1966), p. 679. 

19] T. GLADMAN, “The Physical Metallurgy of Microalloyed 

Steels”, Ed. The Institute of Materials, (1997), London. 

20] P. BUESSLER, P. MAUGIS, O. BOUAZIZ, J.-H. SCHMITT, 

Iron and Steelmaker, 30, (2003), p. 33. 

studiare l’effetto delle diverse variabili sul comportamento meccanico 

finale: 

ricristallazione o tensioni accumulate prima della trasformazione, 

permanenza in campo austenitico e a temperatura di avvolgimento. 

Dopo l’avvolgimento a differenti temperature è stata effettuata l’osservazione 

dei precipitati al microscopio elettronico a trasmissione. 

È stato dimostrato che, quando il Nb rimane in soluzione nell’ austenite 

dopo deformazione a caldo, può precipitare nella ferrite, portando ad 

un importante effetto di miglioramento delle caratteristiche meccaniche 

che è direttamente collegato alla concentrazione di Nb in soluzione 

prima della trasformazione e della temperatura di avvolgimento. 


Memorie >> 

Acciaio 

EFFECT OF PRE-STRAINING AND 

BAKE HARDENING 

ON THE MICROSTRUCTURE 

AND MECHANICAL PROPERTIES 

OF CMnSi TRIP STEELS 

L.C. Zhang, I.B. Timokhina, A. La Fontaine, S.P. Ringer, P.D. Hodgson, E.V. Pereloma 

The effects of pre-straining and bake hardening on the mechanical behaviour and microstructural changes were 

studied in two CMnSi TRansformation-Induced Plasticity (TRIP) steels with different microstructures after 

intercritical annealing. The TRIP steels before and after pre-straining and bake hardening were characterised 

by X-ray diffraction, optical microscopy, transmission electron microscopy, three dimensional atom probe 

and tensile tests. Both steels exhibited discontinuous yielding behaviour and a significant strength 

increase with some reduction in ductility after pre-straining and bake hardening treatment. The 

following main microstructural changes are responsible for the observed mechanical behaviours: a decrease 

in the volume fraction of retained austenite, an increase in the dislocation density and the formation of 

cell substructure in the polygonal ferrite, higher localized dislocation density in the polygonal ferrite 

regions adjacent to martensite or retained austenite, and the precipitation of fine iron carbides in bainite and 

martensite. The mechanism for the observed yield point phenomenon in both steels after treatment was 

analysed. 

KEYWORDS: transformation-induced plasticity steel, retained austenite, bake hardening, mechanical behaviour, 

microstructure, three-dimensional atom probe 

INTRODUCTION 

In order to reduce weight and save energy, there has been 

an increasing interest in application of high strength 

steels for car structural components. As one of the 

advanced high strength steels, transformation-induced 

plasticity (TRIP) steels have attracted the attention of both 

the steelmaking and automotive industries over the last 

L.C. Zhang, E.V. Pereloma 

School of Mechanical, Materials and Mechatronics Engineering, 

Faculty of Engineering, University of Wollongong, Wollongong, 

NSW 2522, Australia 

I.B. Timokhina, P.D. Hodgson 

Centre for Material and Fibre Innovation, Faculty of Science and 

Technology, Deakin University, Geelong, Victoria 3217, Australia 

A. La Fontaine, S.P. Ringer 

Australian Key Centre for Microscopy & Microanalysis, The 

University of Sydney, NSW 2006, Australia 

Paper presented at the 3rd International Conference 

Thermomechanical Processing of Steels, organised by AIM, Padova, 

10-12 september 2008 

two decades. TRIP steels consisting of a complex 

multiphase microstructure (polygonal ferrite, carbide-free 

bainite, retained austenite and martensite) offer an excellent 

combination of strength (700–1000 MPa) and ductility 

(30–40% total elongation) [1]. The interaction of the multiple 

phases present in the microstructure during deformation 

and the strain-induced transformation of the metastable 

retained austenite to martensite [2,3] are thought to 

be responsible for these mechanical properties. At present, 

several studies have been undertaken to investigate the 

effect of the chemical composition and processing parameters 

of TRIP steel sheets [2,4,5], because these are the two 

main factors that can affect the volume fraction and stability 

of the retained austenite in TRIP steels. 

On the contrary, little work has been conducted on the 

variations of the mechanical properties and microstructures 

during the paint baking of deformed panel parts 

of TRIP steel sheets. When the steels are used for outer 

body parts and subjected to the paint baking cycle, additional 

strengthening of approximately 100–200 MPa 

arises from bake hardening and results in good shape 

fix-ability and improved dent and crash resistance [6]. 

Recently, research on bake hardening of intercritically an- 


Acciaio

RESULTS 

Memorie >> 

A. Mechanical properties 

The mechanical properties of the steels in both as-received 

and PS/BH state are given in Fig. 1 and Tab. 2. Both steels 

exhibited a good combination of strength and ductility 

in the as-received condition with higher values of 

strength and ductility for TRIP01 steel compared with the 

TRIP02 (Tab. 2). In this state both steels showed a continuous 

yielding behaviour (Fig. 1). A continuous exponential 

decrease was distinct in the strain-hardening rate curves 

for both steels (Figs. 2 (a) and (b)). After the PS/BH treatment, 

both steels demonstrated a yield-point phenomenon 

on the stress-strain curves. After yield point elongation, the 

Acciaio 

Steel Conditions UTS (MPa) YS (MPa) Total El (%) Uniform El (%) BH response (MPa) 

TRIP01 as-received 880 ± 15 608 ± 20 40 ± 2 

30 ± 2 

N/A 

4 % P S / B H 891 ± 15 803 ± 20 29 ± 2 

21 ± 2 

82 ± 2 

TRIP02 as-received 795 ± 10 520 ± 10 31 ± 2 

27 ± 3 

N/A 

5 % P S / B H 810 ± 8 735 ± 5 27 ± 2 

21 ± 3 

60 ± 3 

UTS: ultimate tensile strength, YS: yield strength, Total El: total elongation, Uniform El: uniform elongation, BH response: 

bake hardening response. 

s 

Tab. 2 

Mechanical properties of the investigated TRIP steels after different processing. 

Proprietà meccaniche degli acciai TRIP investigati dopo i diversi processi. 

s 

Fig. 2 

Variations of the strain hardening rate with the 

true strain of the (a) TRIP01 and (b) TRIP02 steels. 

Variazioni della velocità di incrudimento con la sollecitazione 

negli acciai (a) TRIP01 e (b) TRIP02. 

s 

Fig. 3 

Optical micrographs of the (a) TRIP01 and (b) 

TRIP02 steels after PS/BH treatment. 

Micrografie ottiche degli acciai (a) TRIP01 e (b) TRIP02 

dopo trattamento di PS/BH. 

flow stress started to increase until necking occurred. 

The strain-hardening rate curve of the TRIP02 steel 

after PS/BH treatment still showed a continuous exponential 

decrease (Fig. 2(b)). On the contrary, the 

strain-hardening rate of the TRIP01 steel after PS/BH 

treatment decreased sharply to negative values and 

then started to increase at around 6.5% true strain (Fig. 

2(a)). As seen from Tab. 2, the PS/BH treatment led to a significantly 

higher yield strength (~200 MPa) and a slightly 

higher ultimate tensile strength (~15 MPa) in both steels. 

However, the total elongation decreased by ~11% and ~4% 

for the TRIP01 and the TRIP02 steels, respectively, and the 

uniform elongations reduced by ~9% for TRIP01 and ~6% 

for the TRIP02. Both TRIP steels displayed a noticeable 

bake-hardening response (~82 and ~60 MPa for the 

TRIP01 and the TRIP02, respectively). 

B. Microstructural features 

Optical micrographs revealed that both TRIP steels showed 

a complex multiphase microstructure (Fig. 3). In the as-received 

condition, the microstructure of the TRIP01 steel consisted 

of ~ 30 ± 3% polygonal ferrite (PF), ~56 ± 3% bainite, 

~10 ± 3% retained austenite (RA) with an average carbon 

content of 1.21 ± 0.04 wt.% and the remaining of martensite. 

The polygonal ferrite grain size was 3 ± 1.5 μm. The TRIP02 

steel contained ~70 ± 3% polygonal ferrite and ~20 ± 3% retained 

austenite with an average carbon content of 1.20 ± 

0.05 wt.%. The remaining small volume fraction contained 

martensite and bainite. The average size of polygonal ferrite 

grains was 4 ± 1.5 μm. The volume fraction of retained 

austenite decreased from ~10% in as-received condition 

to ~8% in PS/BH state for the TRIP01 and from ~20% 

in the as-received condition to ~12% in the PS/BH state 

for the TRIP02 steel. In addition, the PS/BH treatment led to 

an increase in carbon content of retained austenite to 1.28 ± 

0.04 wt.% (TRIP01) and 1.30 ± 0.03 wt.% (TRIP02). It is noted 

that bainitic ferrite grains in the TRIP02 were predominantly 

long and parallel laths while those in the TRIP01 steel had 

random orientation. 

Fig. 4 illustrated a more detailed microstructure of the 

steels after different processing. The microstructure in the 

as-received state was composed of polygonal ferrite, carbide-free 

bainite, retained austenite and martensite (Figs. 

4 (a) and (b)). After the PS/BH treatment, a lot of dislocations 

were observed in the polygonal ferrite, as seen 

from Figs. 4 (c) and (d). The most affected regions of 

PF grains were in the vicinity of martensite or retained 

austenite crystals, as denoted by the black arrows. At the 

same time, the formation of dislocation cells in the polygo- 


Acciaio

Steel 

TRIP01 

TRIP02 

Memorie >> 

Phase 

Polygonal ferrite 

Bainitic ferrite after PS/BH 

Martensite 

Polygonal ferrite 

Retained austenite 

Bainite Bainitic ferrite 

Retained austenite 

of both steels was detected: (i) carbon-depleted region with 

an average carbon content of between 0.003 at.% and 0.02 

at.% (Figs. 6 and 7), (ii) a carbon-enriched region with average 

carbon content of between 3 at.% and 5.5 at.% (Fig. 7), 

and (iii) a volume or region, which can be described as a 

mixture of carbon-enriched (5.10 ± 0.20 at.%) and carbondepleted 

(0.20 ± 0.03 at.%) regions (Fig. 7). A summary 

of their major elemental concentrations is given in Tab. 

3. Therefore, from this the first region was polygonal 

Concentration 

C 

< 0.003 

0.11±0.01 

3.22±0.01 

0.02±0.01 

4.60±0.02 

0.20±0.03 

5.10±0.20 

Acciaio 

s 

Tab. 3 

Concentrations of the major alloying elements (at.%) in the phases of as-received steels, except where indicated 

otherwise, determined using APT and calculated based on the number of atoms. 

Concentrazioni dei maggiori elementi alliganti (at.%) nelle fasi degli acciai come ricevuti, ad eccezione di dove diversamente indicato, 

determinati utilizzando APT e calcolati sulla base del numero di atomi. 

s 

Fig. 7 

(a) Carbon atom map of as-received 

TRIP02 sample showing bainitic ferrite and 

retained austenite and (b) C, Mn and Si compositional 

profiles across along green (nearly horizontal) line of 

selected box shown in (a). 

Mappa dell’atomo di carbonio del campione di acciaio 

TRIP02 come ricevuto che mostra ferrite bainitica e 

austenite residua e (b) profili di composizione di C, Mn 

e Si lungo la linea verde (quasi orizzontale) del riquadro 

selezionato mostrato in (a). 

ferrite, the second retained austenite or martensite, and 

the third region was bainite, where the carbon-rich region 

was retained austenite/martensite and the carbondepleted 

region was bainitic ferrite. In both steels the 

Si content of polygonal ferrite was higher than the 

nominal Si content while the Mn content was lower (Tab. 

3). In addition, small iron carbide particles appeared in 

the as-received TRIP01 steel (Fig. 6). As shown in the APT 

map of the as-received TRIP02 steel (Fig. 7), the retained 

austenite crystals may appear in bainite, which had 

a plate-like shape with a thickness of 16 ± 2 nm (Fig. 7) 

and contained a high carbon content (Tab. 3). The bainitic 

ferrite was characterised by a significantly higher carbon 

content (~0.2 at.%), lower Si content (~2.8 at.%) and higher 

Mn level (~1.5 at.%) than polygonal ferrite (Tab. 3). The bainitic 

ferrite in the TRIP01 steel after PS/BH treatment had 

higher than expected Si content (Tab. 3). This might be associated 

with the decomposition of bainitic ferrite during 

bake hardening resulting in the formation of iron carbides 

and rejection of Si back into the BF matrix (Fig. 8). 

After the PS/BH treatment, the formation of Cottrell atmospheres 

at dislocations in the polygonal ferrite and bainitic 

ferrite was evident from the atom probe maps (Figs. 8). The 

rod-like shape of carbon segregation to dislocations is clear 

from two atom maps taken in perpendicular directions 

(Figs. 8 (b) and (c)). It is also clear from Fig. 8 (a) that C segregation 

was non-uniform along the dislocations forming 

a complex tangle, with the formation of iron carbides at 

dislocation intersections. The carbon concentration in the 

core of the atmosphere reached up to ~7 at.%. At the same 

time, concentration profiles across the coarsest carbides 

revealed a carbon content at the centre of ~21 at.% which 

is close to the level in Fe 3 C. Carbon segregations to the 

particular plane in the coarse retained austenite crystals 

were observed in the TRIP02 after PS/BH (Fig. 9 (a)). 

The angle between the saturated planes was ±35°. The 

compositional profiles across these planes showed the 

carbon enrichment to 7±0.08 at.%, while Si and Mn concentrations 

were similar to the matrix composition (Fig. 9 (b)). 

DISCUSSION 

Si 

4.02±0.01 

3.30±0.1 

3.61±0.01 

3.90±0.03 

3.90±0.05 

2.80±0.02 

4.40±0.05 

Both steels showed continuous yielding in the as-received 

condition, while they exhibited a yield-point phenomenon 


Mn 

1.26±0.01 

1.44±0.04 

2.28±0.01 

0.90±0.03 

1.02±0.03 

1.50±0.02 

1.20±0.03

Acciaio

Memorie >> 

phenomenon in TRIP02 is associated with the increase of 

the number of mobile dislocations due to the generation of 

the new ones. This difference in steels behaviour is due to 

the main microstructural difference of both steels. As seen 

from Fig. 3, there was about 70% polygonal ferrite in the 

TRIP02, in which numerous new dislocations easily formed 

during pre-straining, as well as during tensile testing 

of pre-strained and bake-hardened samples. There was 

much less polygonal ferrite (~30%) and much more bainite 

(~56%) in TRIP01, which was not as susceptible to prestraining 

as PF and in which the formation of Cottrell 

atmospheres took place [4]. 

CONCLUSIONS 

A pre-straining and bake hardening treatment led to a significant 

strength increase in CMnSi TRIP steels with some 

reduction in ductility. The observed effects of pre-straining 

and bake hardening on the mechanical behaviour of the 

steels were associated with the main changes in 

their microstructure: microstructural changes in polygonal 

ferrite, such as an increase in the dislocation density and 

the formation of cell substructure, a localized increase of 

the dislocation density in the polygonal ferrite regions adjacent 

to martensite and the formation of fine precipitation 

in bainite and martensite. It was concluded that the 

observed yield point phenomenon was due to dislocation 

unlocking in the TRIP01 steel, which contained bainite as a 

dominant phase. Contrarily, formation of new mobile dislocations 

was the reason for the yield drop in TRIP02 steel, 

in which polygonal ferrite was a major phase. 

ACKNOWLEDGEMENT 

The authors would like to acknowledge the support 

of the Ford Motor Company and Australian Research 

EFFETTO DI PRE-TENSIONAMENTO E BAKE- 

HARDENING SULLA microstruttura E SULLE 

PROPRIETA’ MECCANICHE E DI UN ACCIAIO TRIP 

CMnSi 

Parole chiave: acciaio, deformazioni plastiche, proprietà 

Nel lavoro sono stati studiati gli effetti di pre-tensionamento e bake 

hardening sul comportamento meccanico e sui cambiamenti microstrutturali 

in due acciai CMnSi TRIP (TRansformation-Induced Plasticity) 

con diverse microstrutture dopo ricottura intercritica. L’acciaio TRIP, 

prima e dopo i processi di pre-tensionamento e bake-hardening, sono 

stati caratterizzati mediante diffrazione a raggi X, microscopia ottica, 

microscopia elettronica a trasmissione, sonda atomica tridimensionale e 

Acciaio 

Council (ARC) Linkage scheme. We also acknowledge the 

technical assistance from the AMMNF. 

REFERENCES 

ABSTRACT 

1] Y. SAKUMA, O. MATSUMURA and H. TAKECHI, Metall. 

Mater. Trans. A 22 (1991), p.489. 

2] P.J. JACQUES, J. LADRIÈRE and F. DELANNY, Metall. 

Mater. Trans. A, 32 (2001), p.2759. 

3] V.F. ZACKAY, E.R. PARKER, D. FAHR and R. BUSH, 

Trans. ASM, 60 (1967), p.252. 

4] B.C. DE COOMAN, Curr. Opin. Solid State Mater. Sci. 8 

(2004), p.285. 

5] D.V. EDMONDS, K. HE, F.C. RIZZO, B.C. DE COOMAN, 

D.K. MATLOCK, and J.G. SPEER, Mater. Sci. Eng. A 438- 

440 (2006), p.25. 

6] L.J. BAKER, S.R. DANIEL and J.D. PARKER, Mater. Sci. 

Tech. 18 (2002), p. 355. 

7] I.B. TIMOKHINA, P.D. HODGSON and E.V. PERELO- 

MA, Metall. Mater. Trans. A 38 (2007), p.2442. 

8] A.K. DE, S. VANDEPUTTE and B.C. DE COOMAN, 

Scripta Mater. 44 (2001), p.695. 

9] B.D. CULLITY, Elements of X-ray diffraction, Addison- 

Wesley, London (1978) p.411. 

10] D.J. DYSON and B. HOLMES, Iron Steel Inst. 208 (1970), 

p.469. 

11] P.B. HIRSCH, R.B. NICHOLSON, A. HOWIE, D.W. PA- 

SHLEY and M.J. WHELAN, Electron microscopy of thin 

crystals, Butterworths, London (1965), p. 51. 

12] M.K. MILLER, Atom Probe Tomography, in: Handbook 

of Microscopy for Nanotechnology, eds. N. YAO and Z.L. 

WANG, Kluwer Academic Press, New York (2005), p.236. 

13] E.V. PERELOMA, I.B. TIMOKHINA, M.K. Miller and 

P.D. HODGSON, Acta Mater. 55 (2007), p.2587. 

14] D. KALISH and M. COHEN, Mater. Sci. Eng. 6 (1970), 

p. 156. 

prove a trazione. Entrambi gli acciai hanno mostrato comportamento discontinuo 

allo snervamento e un significativo aumento della resistenza 

neccanica con una riduzione della duttilità dopo il trattamento di pretensionamento 

e bake-hardening. I seguenti principali cambiamenti microstrutturali 

sono responsabili dei comportamenti meccanici osservati 

una diminuzione della frazione in volume di austenite residua, un aumento 

della densità delle dislocazioni e la formazione di una sottostruttura 

a celle nella ferrite poligonale, una maggiore densità di dislocazioni 

localizzata nelle regioni a ferrite poligonale adiacenti alla martensite o 

all’austenite residua, una precipitazione di carburi di ferro nella bainite 

e nella martensite. 

Infine è stato analizzato il meccanismo relativo alla fenomenologia connessa 

al limite di snervamento osservato in entrambi gli acciai dopo 

trattamento. 


Memorie >> 

Metallurgia delle polveri 

MATHEMATICAL MODELING 

OF HEAT TREATING 

POWDER METALLURGY STEEL 

COMPONENTS 

V. S. Warke, M. M. Makhlouf 

A mathematical model to predict the response of powder metallurgy steels to heat treatment is presented 

and discussed. The model is based on modification of commercially available software that was originally 

developed for wrought alloys so that it can account for the effect of porosity. An extensive database had to be 

developed specifically for PM steels and includes porosity- and temperature-dependent phase transformation 

kinetics, and porosity- and temperature-dependent phase-specific mechanical, physical, and thermal properties. 

This extensive database has been developed for FL-4065 PM steel and has been used in the model to predict 

dimensional change, distortion, and type and quantity of metallurgical phases that develop in a typical PM 

component upon heat treatment. The model predictions are compared to measured values and are found to be in 

excellent agreement with them. 

KEYWORDS: powder metallurgy, modelling, steel, heat treatment 

INTRODUCTION 

Components that are manufactured by the powder metallurgy 

process (PM) experience considerable changes during heat 

treatment. These include changes in their mechanical properties, 

dimensions, magnitude and sense of residual stresses, and 

metallurgical phase composition. Since most of the quality assurances 

criteria that these components have to meet include 

prescribed minimum mechanical properties and compliance 

with dimensional tolerances, it is necessary for producers of 

PM components to be able to accurately predict these changes 

in order to take appropriate measures to insure the production 

of parts that meet the required specifications. Several software 

packages that are capable of predicting the heat treatment response 

of wrought steels are available commercially [1-3], but 

none of them can predict the response of PM components. In 

this work, we developed a finite element-based model to predict 

the response of PM steels to heat treatment. The model is 

based on a modification of the commercially available software 

DANTE [3]. DANTE is comprised of a set of user-defined subroutines 

and can be linked to the finite element solver ABAQUS. 

Virendra S. Warke, Makhlouf M. Makhlouf 

Department of Mechanical Engineering - Worcester Polytechnic 

Institute - Worcester, MA 01609 

Paper presented at the International Conference 

“Innovation in heat treatment for industrial competitiveness”, 

organised by AIM, Verona, 7-9 May 

The DANTE subroutines contain a mechanics module, a phase 

transformation module, and a diffusion module that are coupled 

to a stress/displacement solver, a thermal solver, and a 

mass diffusion solver, respectively. A block diagram showing 

the combined DANTE/ABAQUS model is shown in Fig. 1. The 

model requires an extensive database, which includes temperature- 

and porosity-dependent phase transformation kinetics, 

and temperature- and porosity-dependent phase-specific mechanical, 

physical, and thermal properties of the steel. We de- 

s 

Fig. 1 

Solution procedure for the DANTE/ABAQUS 

model. 

Procedura di soluzione per il modello DANTE/ABAQUS. 


Metallurgia delle polveri

Memorie >> 

s 

Fig. 4 

Measured strain vs. time for the austenite to 

bainite transformation at different isothermal holding 

temperatures. 

Misura della deformazione in funzione del tempo per la 

trasformazione da austenite a bainite durante trattamenti 

isotermi a temperature diverse. 

s 

Fig. 6 

TTTP diagram for 90% and 100% density FL- 

4605 PM steel plotted for the austenite to bainite and 

the austenite to martensite transformations. 

Diagramma TTTP per densità 90% e 100% dell’acciaio 

FL- 4605 PM tracciato per le trasformazioni da austenite 

a bainite e da austenite a martensite. 

hibit transformation plasticity, i.e., plastic flow caused by change 

in the relative proportions of the various metallurgical phases in 

the microstructure brought about by the phase transformation. 

We used Low Stress Dilatometry in order to characterize this 

transformation-induced plasticity in FL-4605 PM steel. The procedure 

entails applying an external compressive static load to 

a standard specimen in a Gleeble machine just before the start 

of the transformation. We chose the magnitude of the applied 

load such that the magnitude of the resulting stress is less than 

the flow stress of austenite at the temperature of application of 

the load. We performed transformation induced plasticity measurements 

on samples of three densities (90%, 95%, and 100% of 

the theoretical density of the material) for each of the austenite 

to martensite and the austenite to bainite transformations. 


s 

Fig. 5 

Measured strain vs. temperature at different 

cooling rates during continuous cooling transformation 

measurements. 

Misura della deformazione in funzione della temperatura 

per velocità di raffreddamento differenti durante le 

misure di trasformazione in raffreddamento continuo. 

s 

Fig. 7 

True stress vs. true strain curve for austenite 

measured at three different temperatures for samples 

with 90% density at 1 s-1 strain rate. 

Curva sollecitazione vs. deformazione per austenite, 

misurata a tre temperature diverse per campioni con 

densità 90% e velocità di deformazione 1 s-1 . 

Fig. 8 (a) and Fig. 8(b) show the measured dilatation data for 

the 100% density sample at three the different applied stresses 

during the austenite to bainite and the austenite to martensite 

transformations, respectively. We performed similar measurements 

on samples with 90 % and 95% of theoretical density and 

used a fitting routine to fit this data to mathematical equations 

that were then used to create a transformation induced plasticity 

database. 

MODEL CONSTRUCTION AND MATHEMATICAL 

SIMULATIONS 

The capabilities of the model are demonstrated using the test 

part shown in Fig. 9. The dimensions of the part are summarized 



Memorie >> 

a 

b 

s 

Fig. 10 

Coordinates of the circular hole before and 

after heat treatment. (a) Measured using a CMM. (b) 

Predicted by the model. 

Coordinate del foro circolare prima e dopo trattamento 

termico. (a) Misurato mediante CMM. (b) Previsione del 

modello. 

FL-4605 PM steel component of configuration and dimensions 

similar to those used in the model. 

Sample production – We admixed AUTOMET 4601 steel powder 

with Asbury 1651 graphite powder to yield 0.5 wt% carbon 

in the final product. We added zinc sterate to the powder as a lubricant, 

then placed the admixed powder in a die and pressed it 

using 216 tons of pressure. We sintered the green parts at 1100°C 

for 30 minutes in a continuous sintering furnace under a controlled 

atmosphere and then we air-cooled them to room temperature. 

The average measured density of the parts was 95% 

of theoretical density with negligible variation within each part 

and from part to part. 

Heat treatment – The heat treatment cycle for the sintered parts 

consisted of furnace heating to 850°C, holding at this temperature 

for 20 minutes, and then quenching in oil with the parts 

placed in an upright position with their thinner section point- 


s 

Fig. 11 

Change in radius at different locations around 

the circular hole as predicted by the model 

and as measured by a CMM. 

Variazione del raggio in diverse posizioni intorno ai fori 

circolari secondo la previsione del modello e misurato 

mediante CMM. 

s 

Fig. 12 

Section lines showing where cuts were made 

for measuring retained austenite. 

Linee di sezione che mostrano dove sono stati effettuati i 

tagli per misurare l’austenite residua. 

ing down. Twenty parts were heat treated in an internal quench 

batch furnace under an endothermic atmosphere with 0.5 wt. 

% carbon. 

In order to characterize the amount of distortion in the parts 

caused by the heat treatment we designed a fixture to hold the 

parts at the same location in a coordinate measurement machine 

(CMM). We measured the circular hole before and after 

heat treating the sample at locations around the periphery in 5° 

increments. We repeated the measurements at four depths along 

the thickness of the part and in each case we converted the xy 

measurement into a radius, r, and an angle θ at each measured 

point. We then normalized the radius by dividing it by the average 

radius before heat treatment. Fig. 10 and Fig. 11 compare the 

measured and model-predicted changes in dimensions of the 

central hole due to heat treatment. 



Memorie >> 

Pressocolata 

DIE-CASTING: S.D.C. STEEL, A 

CONTINUOUS METALLURGIC INNOVATION 

TO MEET WITH THE PROBLEMS 

A. Grellier, F. Piana, G. Gay 

The S.D.C. steel grade has been especially designed for large-size toolings used in die-casting applications with 

the objective of increasing the fatigue resistance of the material and the tool life. A fundamental and integral 

approach has been undertaken to understand and measure the temperature and stress conditions at the surface 

during service, get a full multi-scale description of the steel microstructure and define the relationship between 

microstructure and its evolution and thermal fatigue resistance. The new grade with the associated optimised 

heat treatment offers superior mechanical properties and shows improved performance in its die-casting 

industrial applications. 

KEYWORDS: Die-casting, steel grade, microstructure, multi-scale observation, heat-treatment, hardenability, forging 

process modelling. 

INTRODUCTION 

For light alloys casting and more specifically pressure 

die-casting, the 5% chromium steel family grades (H11, 

X37CrMoV5, 1-2343) are widely used. Their performance 

measured by the total number of manufactured parts and 

the amount and cost of repair operations has been through 

the years increased by better casting process control and 

progress in the metallurgy and the heat treatment of the 

moulds. Higher hardness has a positive effect on tool life, 

but remains limited to guarantee a minimum toughness. A 

breakthrough was necessary in the metallurgical conception 

of the grade: beyond the measurement of conventional 

mechanical properties, our research program was focused 

on the global understanding of all the metallurgical and 

thermo-mechanical phenomenas involved. Hereafter are 

described some aspects of our scientific analysis, and the 

basic properties of the new improved S.D.C. steel. 

THERMOMECHANICAL LOCAL CONDITIONS OF 

THE WORKING SURFACE DURING SERVICE 

During parts production in the pressure die-casting process, 

the tool surface is submitted to thermal shocks at two times 

André Grellier 

Aubert&Duval - R&D Department – BP1 

F-63770 - Les Ancizes, France 

Fulvio Piana 

Aubert&Duval Italia-Viale Leonardo da Vinci 97- 

20090 Trezzano sul Naviglio (MI) 

Gérald Gay 

Aubert&Duval - Application Engineering Dpt. 

22 rue Henri Vuillemin B.P.63 – 92233- Gennevilliers 

within every cycle: 

- a hot shock when liquid metal is injected inside the 

mould 

- a cold shock when lubricant is sprayed on the surface. 

Temperature gradients induce high shear stresses at the surface 

of the material: 

- compression stresses during the hot shock which induce 

plastic deformation in compression mode, 

- tension stresses during cold shock with associated plastic 

deformation in tension mode and further crack initiation; at 

this stage, cracking may occur. 

In industrial conditions, the heat flux through the tool surface 

has been measured (Fig. 1) and the surface temperature 

s 

Fig. 1 

Typical thermal recording of one cycle during 

production of pressure-die cast parts. 

Registrazione tipica della temperatura dello stampo in 

esercizio. 


Pressocolata

Memorie >> 

sion Electron Microscopy in high resolution mode and with 

observations on replicas, X-Ray diffraction on precipitates 

extracted by chemical dissolution, and Small Angle Neutrons 

Scattering (SANS) has shown that precipitates of the 

5%Cr steels may be classified in two families of sizes: about 

3 nm and 25/50 nm. Mechanical properties appear to be 

directly dependant from the density of the first family carbides 

of the (V,Mo)C type ; their density has been assessed 

to be about 1023/m 3 to 1024/m 3 . 

The new S.D.C. grade is derived from the 5% Cr standard 

steels with a nickel addition and a molybdenum and vanadium 

content adjustment in order to approach the perfect 

microstructure described before : 

- the combination of the effect of all elements, overall nickel 

and molybdenum, provides a very high hardenability: 

bainite features with wide ferrite lathes and elongated harmful 

carbides are avoided even on large parts corresponding 

to lower quench cooling rates, 

- martensite needles are small and dislocation density is 

high in the matrix, 

- the density of the nanometric (small size family) carbides 

has been carefully adjusted by the vanadium content -0.65%- 

for an austenising temperature of 1030°C. 

FINAL GENERAL PROPERTIES OF THE S.D.C. 

STEEL 

S.D.C. quenched from 1030°C shows a very high hardenability 

(Fig. 3). As the common annealing treatment of conventional 

5%Cr steels is not applicable to S.D.C., a specific cycle 

has been defined, which confers a hardness of less than 220 

HB suitable for rough machining. The microstructure (Fig. 

4) cannot be easily rated with conventional micrographic 

standard charts. Small-sized carbides are dispersed with a 

needle heredity distribution. 

The curve defining hardness as a function of tempering 

temperature (Fig. 5) is not very different from the standard 

s 

Fig. 4 

Microstructure in the annealed condition - 

original magnification: x500. 

Microstruttura in condizioni di tempra. 

Ingrandimento: x500. 


H11 reference. Charpy V impact test energy has been measured 

on a significative number of large size products in the 

conditions of industrial heat treatment and in conformity 

with NADCA Standard (Fig. 6). The results show an unusual 

homogeneity between specimens taken from the same 

place in the block, between different positions in the block 

(core, near-surface…) and between different products. This 

steel is from far less sensitive that nickel-free grades to the 

s 

Fig. 5 

S.D.C. Hardness after tempering - Previous 

1030°C austenisation. 

Durezza dopo il trattamento di tempra. Previa 

austenitizzazione. 

s 

Fig. 6 

Charpy V impact test results on blocks in 

industrial heat treating conditions. 

Risultati dei test Charpy V, sui blocchi in condizioni di 

trattamento termico industriale. 



Memorie >> 

s 

Fig. 10 

Microstructure in the fully heat treated 

condition of a 1000x380 section - a) standard etch 

original magnification: x500; b) specific etching for 

grain size visualisation original magnification: x 100. 

Microstruttura nelle condizioni di trattamento termico 

della sezione 1000 x 380 - a) incisione standard 

ingrandiamento: x 500; 10-b: incisione specifica 

per la visualizzazione della dimensione del grano 

ingrandiamento: x 100. 

BEST PRACTICES FOR HEAT TREATING 

a 

b 

S.D.C. steel must be heat treated with similar equipment 

and procedures as common 5%Cr steels with the additional 

following instructions: 

- the austenisation temperature must be 1025/1030°C after 

a pre-heating in he 750/800°C range, 

- for best service performance, the target for hardness must 

be by 1 to 1.5 HRC more than the hardness of the reference 

5% Cr steel for the same application. 

- the first tempering must be at 550°C, 

- the final expected hardness is obtained by temperature 

adjustment of the following tempering cycles, 


- the full heat treatment may include two or three tempering 

steps; a third temper is recommended for very big 

blocks and if the temperature between quenching and 

first temper and between first and second temper has not 

reached a value below 100°C. 

- SDC is less sensitive than conventional 5%Cr steels to 

variations of gas pressure and gas flow during quenching. 

FIRST RESULTS IN PRODUCTION CONTEXT 

Moulds from different sizes have been produced and put 

in production inside several plants, with an initial hardness 

in the 44 to 49 HRC range. To-day, several of them 

have produced more than 100 000 units and none of them 

is considered as ruined, so a direct comparison for the total 

production of the mould is still not available. 

Nevertheless, results are considered as better than for reference 

steels of the H11 or low-Si H11 grades: 

- production before first crack detection: better from 15 to 

40%, 

- percentage of damaged surface for a given production 

amount: better from 15 to 35%, 

- maintenance operations: decrease from 20 to 80%, 

To-day, no drawback has been identified about repairing 

by welding, distortion, or catastrophic cracking during 

heat-treatment and service. Distortion during heat treatment 

appear to be less important than for classic reference 

grades. 

Moreover, efforts have still to be done to adjust precisely 

the parameters of heat treating in partnership with every 

heat treating plant. 

CONCLUSIONS 

The S.D.C. steel is a new material for tools and more specifically 

die-casting moulds, with superior properties obtained 

as the result of the combination of several innovative 

actions: 

- a multi-scale approach for investigation in the microstructure, 

- a definition of an ideal microstructure by the understanding 

of relations between microstructure and its evolution 

under thermo-mechanical loading during service, 

- a new composition with a nickel addition and a precise 

adjustment of the balance between hardening elements 

and nickel, with a low level of trace elements 

- a specific production route with an optimised forging 

and annealing process, 

- a heat treatment with an accurate definition of parameters 

for austenising and tempering. 

To-day, the S.D.C. toolings which have been put into production 

show an improved performance compared to reference 

materials. These results have to been confirmed; 

discussions and technical collaboration between all partners 

remain absolutely necessary to go ahead on the way 

of progress. 

REFERENCES 

1] G.DOUR, M. DARGUSH, C.DAVIDSON, A.NEF: Journal 

of Material Processing Technology 169 (2005),223-233 

2] DELAGNES D., REZAI-ARIA F., LEVAILLANT C., 

GRELLIER A.,: Proc. of 5th International Conference on

la metallurgia italiana - Gruppo Italiano Frattura

Create successful ePaper yourself

Delete template?

Save as template ?