THERMAL FATIGUE IN NEW LOWER HARDENING
TEMPERATURE HOT WORK STEELS

M. Pirovano

TTN SpA, Nerviano (Mi), Italy

INTRODUCTION

In the last years we assisted to the birth of die casting moulds with more

and more complex shape and with large dimensions.

Large and complex moulds create new problems to heat treatments to fit

process parameters (such as holding times or cooling rates), with quality
results (such as grain size and homogeneous hardness and toughness on all
die areas).

Steel mills are pushed to find new steel grades to cope with these un-

desired phenomena. The result of this research is the new 5% Cr grades.
The improvement, respect to old steels such as H11, is much more in the
production technology than in the chemical composition. The fine structure
and the primary carbides size allow a lower austenitization temperature and
therefore slow graingrowth.

Some of the new steels were compared with H11. The samples were

treated using following procedure to get hardness of approximatly 45 HRC.
The toughness and heat checking resistance of the samples were tested.

THE DIE CASTING STRESSES

The two main failure causes in die casting dies are heat checking and

gross cracking. The latter is surely less frequent, but more catastrophic on
its consequences. Resistance to gross cracking means high toughness; on
the other hand, it is not possible to reduce the hardness too much, because
of heat checking and thermal resistance.

565

566

6TH INTERNATIONAL TOOLING CONFERENCE

Thermal fatigue strength requires high elastic limit, high temper resis-

tance, good thermal conductivity and low thermal expansion.

The achievement of the best compromise is the main goal of the heat

treatment, but the development of more complicated shapes and big masses
makes it a hard challenge.

Different sections of die are hard to cool down at the same time, because

with a high-speed rate there is a high probability of gross cracking, due to
the high temperature difference between thin and large sections. On the
contrary, a low speed rate gets intermediate structures on massive parts.

Also the heating of the die can be a critical step of the process. It is

necessary to hold the die at the austenitization temperature long enough to
solve carbides in the core of die. However the longer the time, the larger is
the grain size; it is in fact well known that the grain size grows approximately
with the square root of the time.

THE STEEL DEVELOPMENT

As mentioned before, steel mills are trying to develop new grades to

satisfy the market requests. We choose three of these new steels, each from
a different producer. To increase the quality respect to H11, two different
directions were followed:

Thermal Fatigue in New Lower Hardening Temperature Hot Work Steels

567

chemical composition

cleanliness and microstructure

The composition of H11 steel is shown in Table 1.

Table 1.

Chemical composition of H11 steel,wt%.

C

Cr

Mo

Si

V

H11

0.38

5.3

1.3

1.0

0.4

The main difference between H11 and new steels is the silicon content.

In new steels silicon content is reduced to 0.2%. It is well known that the
silicon increases austenitizing temperature. One of the aims in developing
new steels was to reduce austenitiaing temperature, therefore the amount of
silicone was reduced.

In fact the main quality improvement comes from new production tech-

nologies such as Electro Slag Remelting (ESL), Vacuum Arc Remelting
(VAR) or Vacuum Melting Remelting (VMR). By using these technologies,
impurities such as undesired chemical compound or structural defects are
under control.

The reduction in columnar structure and segregation facilitates the solu-

tion of carbides therefore the austententizing temperature can be low. By
lowering austentising temperature the grain growth can be controlled effec-
tively.

HEAT TREATMENTS

Heat treatment was carried out according the specification of steel mills.

The suggested austenitizing temperature was 990

◦

C. As mentioned before

this temperature allows the core heating with grain growth in small parts of
the die.

The heating is performed in an inert atmosphere; it is stopped for some

time at some fixed pre- heating temperatures. Different producers suggest
different pre-heating temperatures, but they are chosen for the same goal.
One aim was to avoid create thermal stresses. When the surface get the
temperature, the core could be still very cold, and the different thermal
expansion can cause a gross cracking. Pre-heating reduces this stresses be-

568

6TH INTERNATIONAL TOOLING CONFERENCE

cause it reduces the temperature variations. The first pre-heating is normally
around 600

◦

C, which is also the temperature when the irradiative heat trans-

fer becomes more effective, causing a change in the heating rate. The last
pre-heating must be the highest possible to avoid grain growth the last step.

Some steel-mills suggest only one pre-heating, some others three, but the

basic idea is always the same. Of course holding times depend on parts
dimensions.

Cooling down is normally performed in salt bath but, depending upon

the mould dimension, sometimes it is possible to use a vacuum furnace.
The salt bath has also the advantage to harden in two steps, stopping at
an intermediate temperature of about 550–580

◦

Cand then cooling down to

room temperature.

In this steels due to so called non transformation window, it is possible

to stay long time at the intermediate temperature without forming perlite or
bainite.

There is an advantage in using salt as a cooling media. Two salt bath

can be used, one at the intermediate temperature and the other at hardening
temperature. The same result can be achieved with a vacuum furnace. In
the case of a large part it is enough to use one thermocouple near the surface
and the other near the core.

Cooling starts fast and when the inner thermocouple is reading the medium

temperature, walls start to heat to get the same temperature even on the
surface. At this point it’s possible to cool down again. In Fig 1 are reported
temperatures versus time in such a bath.

With a proper heat treatment, these steels don’t have considerable amount

of retained austenite, so a cooling below 0

◦

Cit’s not usually necessary, but

nevertheless it’s very important to get the room temperature before starting
with the first tempering process.

Tempering is normally done in three steps; the first is about 50 degrees

below the maximum tempering temperature, The tempering temperature
is chosen according required hardness. The last step is just a final stress
relieving, again few degrees below the hardening temperature.

Thermal Fatigue in New Lower Hardening Temperature Hot Work Steels

569

Figure 1.

The vacuum salt bath simulation; yellow line is the outer thermocouple, green

the inner and blue the pressure

TESTS

SAMPLE PREPARATION

All 4 steels were taken in the annealed state and tempered following the

steel- mill specifications. The required hardness was 45–46 HRC. The main
parameters are summarized in the following Table 2.

Table 2.

Heat treatment times and temperatures.

1st

pre-h

2nd

pre-h

3rd

pre-h

Austen.

1st temp

(hold. time)

2nd temp

(hold. time)

3rd temp

(hold. time)

H11

700

880

1020

550 (4)

590 (4)

570 (4)

S1

700

990

550 (4)

600 (4)

590 (3)

S2

690

830

920

1000

530 (4)

590 (4)

600 (4)

S3

650

830

990

550 (4)

590 (4)

600 (4)

All temperatures are expressed in Celsius degrees and times in hours.

All the hardening diagrams are reported in Fig. 3. After the tempering

the samples were prepared for the two tests. All samples had a martensitic
structure with a low bainite content. The grain sizes were examined with
different etchent. The grain sizes were similar because with small samples
the dramatic differences were not expected.

570

6TH INTERNATIONAL TOOLING CONFERENCE

Figure 2.

The hardened structure

TOUGHNESS TEST

The standard transverse chary - V test was used for toughness test. Al-

tough the numbers of the samples were small (three samples for each mate-
rial), the following observations were made:

The hardness was not close to what was expected except for H11. S2
and S3 had higher hardness while S1 was softer than expected.

H11 had considerably lower toughness than the others. S1 had better
performance while the difference between the toughness of S2 and S3
might be due to difference in their hardness.

Thermal Fatigue in New Lower Hardening Temperature Hot Work Steels

571

Figure 3.

The hardening diagrams

Figure 4.

Hardness

HEAT CHECKING TEST

The laboratory simulation of heat checking is not standardized because it

is a complex phenomenon influenced by many parameters such as heating

572

6TH INTERNATIONAL TOOLING CONFERENCE

Figure 5.

Charpy-V transverse toughness

and cooling rate, minimum and maximum temperatures, oxidation and so
on. Each lab has their own apparatus for heat checking test.

The schematics of our apparatus is described in Fig 6. The sample is a

Figure 6.

Scheme of the test apparatus

cylinder of 40 mm in diameter and 10 mm in height. It rotates at a speed of
4.5 rounds per minute. During the rotation the outer surface is heated by

Thermal Fatigue in New Lower Hardening Temperature Hot Work Steels

573

induction on an angle of about 90 degrees, and then it is quenched first with
flowing water (2.5 litres per minute) and then by immersion.

Temperature is read by a pyrometer just after the heating, and the inductor

power was fixed in a way to get 700-710

◦

C. The lowest temperature is not

directly measured but a pre-testing with a thermocouple showed that at 1mm
below the surface if the maximum temperature is 700

◦

C, the minimum is

80

◦

C.

Samples were run for 90 minutes, that means 400 cycles.
The outer surface is grinded, while one of the two sides is polished. After

the test, the same side is again polished to remove the scale caused by water
and oxidation. Then it was examinated under optical microscope, observing
cracks propagating from the external diameter. The following quantities
were considered:

Im, the medium crack length, in micron, that’s the sum of the single
cracks length divided by the cracks number

ρ as the cracking density, that means the total cracks numbers divided
by total crack length.

P

max

, the maximum depth, which is not always the maximum length

because sometimes the crack continues below the surface

All three quantities give indications on the heat checking, but it is not

easy to establish which one is more indicative of the tool behaviour during
its life.

Results are reported in the following table:

I

m

[mm]

ρ [mm

−1

]

P

max

[µm]

C [µm

2

mm

−1

]

H11

11,76

27,37

31,51

10.142

S1

13,36

21,42

36,57

10.465

S2

12,86

24,00

20,57

6.348

S3

9,42

22,50

21,89

4.639

The following conclusions can be made:

H11 has the highest crack density, even if differences are not so high,
and has long cracks. Statistic is high enough (more than 100 measures

574

6TH INTERNATIONAL TOOLING CONFERENCE

for each steel) to exclude that a high P

max

is due to a random phe-

nomenon. In fact some cracks cannot be measured completely because
they continue below the surface. But if we discard the longest cracks
and we consider the 4th or the 6th, P

max

does not change significantly.

S1 has the longest cracks and also the highest media, but the lowest
density. The opposite was expected, because with a high toughness
and soft material large numbers of short cracks were expected.

S2 and S3 show the best performances, especially P

max

, and this is

strange again for the same reason explained before.

None is the best in all three measured quantities, and it’s not too easy to
evaluate which would have the best performance on application. We define

C = P

max

ρI

m

as the global cracking factor [1], that, with the same number of cycles, gives
the tendency to heat checking. It is in fact shown that this quantity has a
better confidence with results in applications.

S3 and S2 showed the best results, while S1 and H11 were similar. S1

was also the softest.

Therefore it seems that the elastic limit is very important. It seems also

important to get the right hardness during tempering; we cannot of course
say that there is an optimum hardness value for all applications, because a
simple test cannot simulate the different industrial conditions, but it seems

Thermal Fatigue in New Lower Hardening Temperature Hot Work Steels

575

clear that little differences in hardness can cause dramatic differences in heat
checking resistance; in fact, from our industrial experience, we do not think
these differences can be explained only with the material itself.

It has to be taken into account that the tests were carried out with simple

parts not complex mould. It is expected that H11 perfomrs worse than others
because the other steels had homogenios final structure.

It is also expected better performance from new steels because of their

casting technology. We are also planning a test to measure K’

c

, to study

cracking propagation in heat checking. We hope to present these results
during the conference.

REFERENCES

[1] Callegari, Venturelli; Una nuova apparecchiatura e metodologia per prove di pirocric-

catura su materiali metallici; BTF Ottobre 1977 TTN SpA, Nerviano (Mi), Italy