Development of Bake-Hardenable Al-Killed Steel Sheet by Box Annealing Process

UDC 669.14.122.2.153:621.785.3:539.5

Development of Bake-hardenable Al-killed

by Box-annealing Process*

Steel Sheet

By Atsuki OKAMOTO,** Masashi TAKAHASHI ** and Takao HINO***

Synopsis

The bake-hardenable cold-rolled steel sheets for the outer panels of auto-bodies exhibiting an excellent resistance to the denting have been developed in the box-annealing process, by examining the effects of chemical compositions and box-annealing conditions on the bake-hardenability and on Snoek peak height of Al-killed steels. It has been found that lowering the total carbon content to less than 0.02% or raising the annealing tem-

perature to air intercritical range bears a bake-hardenable steel sheet in the box-annealing process. Additions of silicon and phosphorus and reduc-tion of manganese are also found to increase the bake-hardenability. The bake-hardening is due to the strain aging caused by the solute carbon of approximately 10 ppm. Furthermore, it is shown that the bake-hardenable steel sheets can be successfully manufactured in steel mills and exhibits non-aging quality as well as the high r-value.

I. Introduction

Since the continuous casting machines have per-vaded to Japanese steel works, most of the cold-rolled steel sheets for outer panels of auto-bodies have been made by the box-annealed Al-killed steel sheets, be-cause they show non-aging property, high r-value and low yield strength and, as the results, exhibit superior formability to the conventional rimmed or capped steel sheets.

Recently, use of the thinner panels has been at-tempted for the weight reduction of the auto-bodies. In order to maintain the same dent-resistance with the thinner sheet, high strength steels must be applied to the panels, because their denting can be avoided by the increase in their yield strength.l,2~ The rephos-

phorized Al-killed steel with the tensile strength of 400 MPa has been one of the steels anticipated, be-cause it shows high r-value and high strength and is

produced at a low cost. However, its applications for the stamping operations resulted in the poor shape-fixability of the formed panel. It means that the steel does not fit closely to the stamping die and causes the surface deflection to the panel because of its high

yield strength. Therefore the steel with low yield strength before the stamping and high yield strength in the finished panel is required.2~ As the steel is

paint-baked for nearly 20 min at above 170°C after the stamping, the bake-hardenable steel is expected to be valid for the gauge reduction without spoiling the dent-resistance of the outer panels.

The conventional rimmed steel is known to be bake-hardenable3~ because it contains free nitrogen of nearly 20 ppm. However, it is inadequate for the outer panels because it shows yield point elongation

after room temperature aging and thereby it is in danger of stretcher strain. Though the continuously

annealed dual phase steel is also bake-hardenable,4) it exhibits poor r-value and too high strength.

As the slow heating of Al-killed steel is the most

suitable annealing process to produce a sheet with the high r-value, we have tried to develop a bake-hardenable Al-killed steel by the box-annealing pro-

cess controlling the solute carbon content.

II. Effect of Chemical Compositions and An- nealing Temperature

1. Materials and Experimental Procedures

Eighteen steels were vacuum-melted in laboratory furnace. Their chemical compositions are shown in Table 1. The basal steel (steel C 1) exhibits the chemical composition of commercial low carbon Al-killed steel sheet. The steel C l2 exhibits that of com-mercial rimmed steel sheet.

The steels were forged and then hot-rolled from 20 mm to 2.8 mm in thick by five passes. The steels were soaked for 30 min at 1 250°C before hot-rolling and cooled rapidly with spray to room temperature immediately after the hot-rolling. The finishing tem-

perature was nearly 900°C. After the cooling, the steels were reheated to 450°C, soaked for 30 min and then furnace-cooled in order to precipitate carbon as cementites. Then they were pickled and cold-rolled to 0.8 mm thick with the reduction of 71 %. The annealings were conducted in an electric furnace for 4 h at 680°, 710° or 740°C, with the heating rate of 40°C/h and the cooling rate of 80°C/h. The annealed steels were temper-rolled by an elongation of 1.5%, and then following examinations were conducted by using JIS No. 5 specimen. 1. Room-temperature Aging

Tensile testings were carried out for steels imme-diately after the temper-rolling and after the acceler-ated aging treatment at 50°C for 30 days. The treat-ment is called A.A.T. hereafter. The steel with the

yield point elongation (Y.P.E.) of less than 0.7% was judged as substantially non-aging quality. The changes in yield strength and elongation during A.A.T. were examined also. 2. Bake-hardenability

The specimens of the temper-rolled steels were stretched in tension to an elongation of 2 %, immersed

*

**

***

Based on the paper presented to the symposium of the 100th ISIJ Meeting, October 1980, at Kyushu University, Fukuoka, (publish-ed in Tetsu-to-Hagane, 66 (1980), A209, in Japanese). English manuscript received April 1, 1981. Central Research Laboratories, Sumitomo Metal Industries, Ltd., Nishinagasu-hondori, Amagasaki 660. Head Office, Sumitomo Metal Industries, Ltd., Kitahama, Higashi-ku, Osaka 541.

(802) Technical Report

Transactions ISI1, Vol. 21, 1981 (803)

in oil bath for the soaking period of 20 min at 170°C, and then stretched to the fracture. The pre-straining and the heating at 170°C simulates the stamping op-eration and the paint-baking process, respectively. The work hardenability with 2 % strain and the bake-hardenability were measured as shown in Fig. 1.

For the examination of the relation between the bake-hardenability and the room temperature aging

property, the basal steel C 1 was also annealed at con-tinuous annealing cycles which involved the rapid heating to 700°C and the air-cooling with the over-aging treatments at 400°C for various periods.

Relation between Snoek peak height and the bake-hardenability was also examined for several steels an-nealed at various heat cycles. In addition to the above-mentioned three box-annealing cycles, the cool-ing rates from 740°C to room temperature were varied in the range between 20°C/h and 200°C/h, and more-over on their cooling stages water-quenching from 300°C was conducted in order to vary the amount of solute carbon. Then their bake-hardenabilities and Snoek peak heights were measured. The Snoek peak of internal friction was measured by transversal vibra-tion method with the test piece of 10 mm wide and 120 mm long. The frequency was nearly 330 Hz. The peak temperature, TTp, was nearly 100°C. Car-bon in solution was estimated by using the following equation modified by Aoki et a1.5~ for the transversal vibration measurement of commercial sheet steels:

C (wt%) = 0.0044 x Tp x Q;

2. Results The effects of chemical compositions and annealing

temperature on the bake-hardenability are shown in Fig. 2. The Y.P.E. after A.A.T. for the steels an-nealed at 740°C is also shown in the figure. It is obvious that when the steel contains normal carbon content, the annealing at above Al temperature, that is an intercritical annealing at 740°C, is necessary to show the bake-hardenability of more than 40 MPa. When the steel is annealed at a subcritical tempera-ture, decrease of carbon content to less than 0.02%

gives larger bake-hardenability. For the three an-nealing temperatures, increase of silicon or phospho-rus contents gives larger bake-hardenability while increase of manganese or chromium content gives smaller bake-hardenability to the steels. Within them,

Table 1. Chemical compositions of specimens. (wt %)

Fig. 1. Measurement of bake-hardenability

work-hardenability by tensile testing.

(4Y.S.) and

Technical Report

( 804 ) Transactions ISIJ, Vol. 21, 1981

the effect of manganese on the bake-hardenability of the intercritically annealed steel is remarkably large. The effect of aluminum content is not large as far as the steel contains enough aluminum to combine nitro-gen as aluminum nitride. However, when aluminum is not added, the steel shows large bake-hardenability of more than 70 MPa, due to the presence of inter-stitial nitrogen. The Y.P.E, after A.A.T. is less than 0.7% for steels containing enough aluminum, while it is larger than 2 % for the steel without aluminum. The effects of sulfur, nitrogen, boron, copper and niobium contents on the bake-hardenability were in-vestigated, but no appreciable effect was observed.

Photograph 1 shows optical micrographs of the an-nealed steels. The cementite particles are coarsened slightly as the annealing temperature rises from 680° to 710°C. A remarkable change in the morphology of the cementite occurs as the annealing temperature surpasses the Al temperature. This is due to the existence of austenite phase as the absorbing sites for carbon during the soaking period.

The distinct variations in the morphology of ce-

mentite with chemical compositions were not rec-ognized except for carbon and manganese. The ef-fect of manganese content is shown in Photos. 1(d) and (e). The cementite or pearlite particles are finely dispersed in the high manganese steel. The pearlite was more often observed in high manganese steel also. These effects of manganese are possibly due to the stability of austenite phase with manganese.

The relation between the bake-hardenability and the room temperature aging property is shown in Fig. 3 including the results of steel C 1 annealed at the continuous annealing cycles. The yield point elonga-tion (Y.P.E.) after A.A.T. increases with the increase in the bake-hardenability. It implies that both prop-erties are caused by a similar reaction, that is strain aging by carbon in solution. However, steel C l2 which corresponds with a rimmed steel exhibits larger Y.P.E. than the other Al-killed steels with the same bake-hardenability, which indicates that the strain aging by carbon gives rise to a smaller Y.P.E. than that by nitrogen does. The figure also suggests that the maximum bake-hardenability with the Y.P.E. of

Fig. 2. Effect

point

of chemical compositions and annealing temperature on the

elongation after aging treatment at 50°C, for steels cooled at

bake-hardenability 80°C/h from the

and the yield

anneals.

Photo. 1. Optical micrographs of annealed steels, showing the morpholo gies of cementite particles. (Picral etch)

Technical Report

Transactions ISIJ, Vol. 21, 1981 (805)

less than 0.7% after the room temperature aging is approximately 60 MPa for both the box and con-tinuously annealed Al-killed steels.

Figure 4 shows relation between the bake-harden-ability and the decrease in total elongation by A.A.T. after temper rolling. It is clear that the decrease in total elongation by the room temperature aging is inevitable in the bake-hardenable steel sheets. The increase in yield strength by A.A.T. after temper roll-ing was examined also. However, no remarkable

correlation with the bake-hardenability was recog-

nized. The relation between the bake-hardenability and

Snoek peak height is shown in Fig. 5. It is obvious that the bake-hardenability is mainly controlled by the

solute carbon in steel. In general, the bake-harden-

ability of more than 40 MPa is attained by the 10

ppm carbon in solution. In the figure, the correla-tion between the two values differs with the chemical compositions of the steel. Addition of silicon or phos-

phorus to the basal steel gives larger bake-harden-ability with the same Snoek peak height and, on the

contrary, the increase in manganese or carbon con-tent decreases the bake-hardenability. The result im-

plies that the solute carbon calculated from Snoek

peak height is not entirely equivalent to the carbon which causes the bake-hardenability.

III. Effect of Annealing Conditions on Carbon in Solution

1. Materials and Experimental Procedure

On the basis of the former results, three Al-killed steels were melted in an oxygen converter and cast con-tinuously into slabs. Their chemical compositions are shown in Table 2. They are the low manganese and rephosphorized Al-killed steels with the different car-bon contents. The addition of silicon was limited to small amounts to avoid the surface imperfections after box-annealing. In the steel making of steels A 1 and A2, the vacuum-degassing was conducted with DH machine to lower the carbon contents.

The slabs were hot-rolled to 2.8 mm in thick with the coiling temperature of 550°C in steel works. The hot-rolled steels were pickled, cold-rolled to 0.8 mm

Fig. 3. Relation between bake-hardenability and

after aging treatment at 50°C for 3 days.

Y.P.E.

Fig. 4. Relation between bake-hardenability and decrease

in total elongation by aging treatment at 50°C, for

3 days.

Fig. 5. Relation between Snoek peak height and bake-

hardenability for various steels with different chemi-

cal compositions.

Technical Report

(806) Transactions ISIT, Vol. 21, 1981

and then annealed in laboratories. The heating rate and the soaking time were fixed to 40°C/h and 4 h, respectively. The annealing temperature ranged from 660° to 760°C and the cooling rate was also varied from 10°C/h to air-cooling. The annealed samples were processed to measure Snoek peak height of inter-nal friction by the transversal vibration method.

2. Results

Relations between the annealing conditions and Snoek peak height are shown in Fig. 6. For steel Al containing 0.01 % carbon, large amount of carbon is in solution even after the steel is cooled slowly at the rate of 10°C/h. While in steel Bl containing 0.06% carbon, both the intercritical annealing and the cool-ing at the rate of 80°C/h is necessary to hold a suf-ficient amount of carbon in solution.

As the cooling rate of tight-coil annealing is in the range of 10° and 20°C/h and that of open-coil an-nealing is nearly 80°C/h and, moreover, the tight-coil annealing at an intercritical temperature is in danger of sticking, the open-coil annealing furnace is recom-mended with steel B 1 to manufacture the bake-hard-enable steel.

Iv. Mill Manufacturing

The three hot rolled steels of which chemical com-

positions are shown in Table 2 were pickled, cold-rolled, box-annealed and then temper-rolled in steel works. The elongation of the temper-rolling ranged from 0.8 to 1.2%. The box-annealing conditions and the mechanical properties of the steels manufactured are shown in Table 3. The typical mechanical prop-erties of the conventional deep drawing quality steel

is also shown in the table for comparison. The de-veloped steels exhibit higher yield and ultimate tensile strengths, lower total elongation and larger bake-hardenability, maintaining high r- and n-values than the D.D.Q, steels.

The room temperature aging behavior of steel A2 at 30°C is shown in Fig. 7. In contrast to that of a commercial rimmed steel sheet, either the age-harden-ing behavior or the reappearance of Y.P.E. is scarcely observed in steel A2. Small increase in the yield strength and small decrease in the elongation are ob-served only at the initial stage of the aging in the steel. On the process to measure the bake-hardenability of the three steels, the pre-strained tensile specimens were heat treated for various periods and at different temperatures in an oil bath. The effects of heat treat-ment on 4Y. S., that is a bake-hardenability, and on 4U.T.S. for steel A2 are shown in Fig. 8. By the heat treatment at 170°C for 20 min, the 4Y.S. of 40 MPa and the 4U.T.S. of 20 MPa were obtained, but the hardenabilities were not influenced by the small de-crease in the temperature and the aging time.

The figure suggests that there are two processes in the age hardening. One is that below 120°C and the other is that above 170°C. The heat of the paint-baking seems to be sufficient to complete former pro-cess but not to reach the threshold of the latter pro-cess. The relations between time and temperature ex-hibiting the same 4Y.S. of 30 MPa for the three steels are shown in Fig. 9. Three straight lines of the similar slopes were obtained below 80°C. The ac-tivation energy was calculated from the slope to be 20.5 kcal/mol, which is approximately equal to that for the diffusion of carbon. The figure indicates that the increase in 4Y. S. is faster in the order of steels A2, B 1 and Al, which implies that the amount of carbon in solution is larger in that order.

V. Discussion

1. Effect of Carbon Content on the Bake-hardenability

The relation between the carbon content and the

Table 2. Chemical compositions of verter. (wt%)

steels melted in con-

Fig. 6.

Effect of anneal

height for two

phorized steels

tents.

ing conditions on Snoek peak

low manganese and rephos-

with different carbon con-

Technical Report

Transactions ISIJ, Vol. 21, 1981 (807)

box-annealing temperature required for manufactur-

ing a bake-hardenable Al-killed steel is summarized in the Fe-C diagram and shown by the shaded areas

A and B in Fig. 10. The typical morphologies of the cementite particles after the annealing are compared in Photo. 2.

In the conventional process, the steel containing nearly 0.05% carbon is usually box-annealed at a subcritical temperature corresponding to the area C. In this case, the carbon of more than 100 ppm is dissolved into the ferrite matrix at the soaking period. However, most of the carbon is precipitated as ce-mentite at the slow cooling stage because there exist many cementite particles, as shown in Photo. 2 (c), which act as the precipitation sites for carbon. When the steel is annealed in the intercritical temperature, in the area B, the cementite particles exist only on the ferrite grain boundaries and thereby the number of

precipitation sites is remarkably decreased. As the result, the steel can retain some amount of the carbon in solution after the box-annealing if the cooling rate is controlled to nearly 80°C/h. When the total car-bon content of the steel is decreased to the area A, as no cementite particle exists at the soaking period, some special grain boundaries are the only possible sites for carbon to precipitate, as speculated from Photo. 2(a), so that the large amount of carbon is kept in solution even after the slow cooling.

These effects of the morphology of cementites on the solute carbon have also been suggested in the

previous works. Butler6~ has examined the effect of cooling rate from austenite to room temperature on the carbon in solution and found that the increase in the distance between carbide particles or the grain

Table 3. Results of mill production. (JIS No. 5 specimen)

Fig . 7. Room temperature aging behaviors of a bake-hard-

enable Al-killed steel sheet and a commercial

rimmed steel sheet manufactured in steel mill.

Fig. 8. Effect of aging time and temperature after 2 % pre- straining on JY. S. and increase in ultimate tensile

strength, 4U.T.S., of steel A2, open coil annealed.

Fig. 9. Relation between time and

the same AY.S. of 30 MPa

bake-hardenable steel sheets.

temperature to show

for three box-annealed

(808) Transactions ISIJ, Vol. 21, 1981

diameter retains the larger amount of carbon in solu-tion after the slow cooling from the anneal. He has also showed6~ that the rise in the hot strip coiling tem-

perature from 566° to 677°C increases the aging index of a box-annealed Al-killed steel sheet from 1.4 to 23 MPa, due to the change in carbide dispersion. Rich-ards and Barrett7~ have investigated the effect of the hot strip coiling temperature of a rimmed steel on the carbon in solution after continuous annealing and ob-tained the results that the higher the coiling tempera-ture, the larger the carbide particles, the greater the intercarbide particle spacing and the higher the car-bon retained in solution. As for the effect of total carbon content in steel, Takashina and Harada8 have indicated that the largest amount of carbon is retained in solution after the box-annealing at 700° or 760°C of a rimmed steel containing 0.023% carbon besides

nitrogen. However, the details of the experimental

procedure have not been shown. These former works have suggested the presence of

a particular relation between the microstructure and carbon in solution, although most of their results have been obtained with rimmed steels and duplicated with the effect of nitrogen. As a rimmed steel sheet is known to retain larger amount of carbon in solution than an Al-killed steel,8'26~ more efforts to increase the solute carbon are required to manufacture a bake-hardenable Al-killed steel sheet for the practical use. The controls in the amounts of manganese, silicon and

phosphorus beside carbon are some of those efforts.

2. Effect of Substitutional Elements

The addition of silicon or phosphorus and the re-duction of the manganese content in steels are very valid to increase the bake-hardenability of box-an-nealed steels, as shown in Fig. 2.

These effects by the substitutional elements can be caused by several means, the most likely of which are : 1) changing the equilibrium solubility limit of car-

bon in ferrite, thereby changing the driving force for carbon to precipitate as cementite at about

300°C in the cooling stage; 2) changing the activation energy to form a new

lattice of cementite in the cooling stage; 3) changing the apparent diffusion rate of carbon in

ferrite; and 4) changing the morphology of cementite especially

in the intercritically annealed steel, thereby changing the nucleation sites.

There does not seem to be any reliable datum for the effect of alloying elements on the solubility limit of carbon in ferrite.9~ For example, as an effect of silicon on the solubility limit of carbon, Leak et al.10~ have reported that silicon decreases the solubility limit and, on the other hand, Borchers and Konigtl) have obtained the opposite result, though the both re-searchers have measured it by the internal friction method. Most of the research indicates that man-

ganese decreases the solubility,11-13) though there exist the opposite results14~ also. As the carbon is largely supersaturated in the cooling stage, the authors con-sider that the difference in the equilibrium solubility limit of carbon is not the predominant cause to change the amount of retained carbon in solution.

Leslie15~ has examined the quench aging process of low carbon iron containing 3 % silicon or 0.45% man-

ganese, finding that silicon raises and manganese low-ers the aging temperature at which the metastable carbide appears. He has then attributed these effects of silicon and manganese to the changes in the carbon activity by their additions. Fujita et al.16~ have also found that the addition of silicon to 0.02% and 0.2% decreases the precipitation rate of carbon in quench aging process. They interpreted the phenomenon by the temporary trapping of migrating carbon atoms by silicon atoms.

In the author's limited knowledge, it seems that silicon delays and manganese accelerates the precipita-tion of carbon from the supersaturated solid solution.

Fig. 10, Schematic presentation of the relation between

carbon content and box-annealing temperature to

manufacture bake-hardenable steels.

Photo. 2. Optical micrographs of box-annealed

showing the morphologies of cementite

correspondence with Fig. 10.

steel sheets,

particles in

These effects of substitutional elements are most prob-ably correlated with the fact that the cementite can concentrate large amount of manganese to stabilize itself,17,18> while it excludes silicon.15,18,19) In order to increase the volume fraction of cementite, the dif-fusive silicon away from the growing particles is neces-sary.15) Though there is no available report on phos-

phorus for the partition between ferrite and cementite, the same effect as silicon can be expected for it. Moreover, as phosphorus segregates to ferrite grain boundaries in the soaking period and occupies the sites favorable for carbon to precipitate, the precipitation of carbon is believed to be suppressed with the increase in the phosphorus content. It is reported by Kawa-moto et a1.20) that the addition of phosphorus increases the bake-hardenability of box-annealed Al-killed steel sheet. The morphology of cementite particles may play an important role especially for the intercritically an-nealed steels. The large effect of manganese content on the bake-hardenability of the intercritically an-nealed steels, shown in Fig. 2, is mainly attributed to the smaller intercarbide particle spacing in high man-

ganese steel (C8), as shown in Photo. 1(e).

3. Relation between the Bake-hardenability and Snoek Peak Height

A good correlation between the bake-hardenability and Snoek peak height shown in Fig. 5 indicates that

the former property is mainly controlled by the carbon

in solution. However, there is a little difference in the relationship among the steels. The higher the

manganese or carbon content is, the smaller the bake-hardenability at the same Snoek peak height is, while

the higher the silicon or phosphorus content is, the

larger it is. Since the most of the data shown in Fig. 5 are for the intercritically annealed steels, the effect

of the cementite morphology on the bake-hardena-hility must be taken into account in order to understand

the effect of manganese and carbon precisely. The reason why the larger bake-hardenability was

obtained with the same Snoek peak height for the

steels containing silicon or phosphorus is uncertain. As the addition of silicon'°" or phosphorus21) to a

large amount broadens the Snoek peak, the resolution of Snoek peak is necessary to evaluate the amount of

solute carbon. In this experiment, it was not con-ducted because the alloying amounts were small. As

the result, the solute carbon is thought to be estimated

less in the steels containing silicon or phosphorus. The effect of these elements on the stability of car-

bide must be taken into account also. As the binding energy of Fe3C, 0.42 eV,22) is a little lower than the

interaction energy between carbon and dislocation,

0.5 eV,23) the resolution of carbide in the deformed steel is thought to have occurred, according to the

stability of carbide. When the carbide contains man-

ganese, it may hardly be resolved, because the carbide is stabilized by manganese. On the contrary, when

the steel contains silicon or phosphorus, carbide can

probably be resolved to increase the bake-harden-ability.

Transactions ISIJ, Vol. 21, 1981 ( 809)

In the author's limited knowledge, these two ef-

fects of the alloying elements are thought to have in-fluenced the different correlation between the bake-hardenability and Snoek peak height. The effect of

the elements on the equilibrium solution limit of car-bon in ferrite at 170°C can be negligible, because that

for pure iron with Fe3C at 170°C is reported as ap-

proximately 0.5 ppm,22) which is far less than the amount of solute carbon in the bake-hardenable steel, that is 10 ppm.

4. Compatibility of Non-aging Property and Bake-harden- ability

As both the room temperature strain aging phe-nomenon and the bake-hardenability are controlled by the diffusion process of carbon to dislocations, the compatibility of non-aging and bake-hardenable prop-erties must be discussed.

The time-temperature relationship for the strain aging has been derived by Hundy.24) He has used the equation of Cottrell and Bilby25) to give the rate at which interstitial atoms segregate to dislocations as a function of temperature and diffusion coefficient. Hundy's relationship is roughly comprehended with the activation energy for the diffusion of carbon or nitrogen. Figure 11 shows the schematic time-temperature relationship of aging treatment to give the same yield

point elongation after the same temper rolling. The lines which are not exactly straight are plotted by using Hundy's equation for carbon23) ;

log (ti/t2) = 4 400 (1 / T'1-1 / TT2) - log (T1/ T'2)

where t1 and t2 are times and T1 and T2 are absolute temperatures. The non-aging property means prac-

tically no yield point elongation after one month at 30°C. The time-temperature relationship for it is

described by line S in the figure. The accelerated aging treatment at 50°C for 3 days used in the pres-

ent experiment is also on the line. The large bake-hardenability can be obtained when the yield point

elongation appears after the heat treatment at 170°C for 20 min. Its time-temperature relation is de-

scribed by line B. It is important that line B locates at the longer time area of line S, that is non-aging

area. A commercial galvanized steel sheet contains large amount of carbon in solution and thus shows

both the room temperature aging property and the

large bake-hardenability. While, a box-annealed Al-

killed steel sheet exhibits no bake-hardenability and non-aging property. It is recognized from the figure

that the control of solute carbon to the moderate

amount for steel as to bring the start point of yield

point elongation into the region between the lines S and B is necessary to manufacture a non-aging and bake-hardenable steel. The moderate amount of car-

bon in solution is thought to be nearly 10 ppm. The above discussion is based on the schematic

strain aging phenomenon. Actually the time-tem-

perature relationship does not hold good at high tem-

perature and the type of pre-straining must be dif

Technical Report

(810) Transactions ISIJ, Vol. 21, 1981

ferent between the cases of the room temperature aging and the paint-baking.

Figure 8 and its Arrhenius plot in Fig. 9 suggest that there are several stages of strain aging, one of which is below 120°C and controlled by the simple diffusion of carbon, and the strain aging at a higher temperature is faster than that speculated in Fig. 11. This is advantageous to produce a non-aging bake-hardenable steel.

It is well known that the reappearance of yield

point elongation is largely affected by the type of pre-straining.3,27~ The temper rolling restricts its reap-

pearance remarkably in comparison with the normal tensile deformation. Figure 12 shows the effect of pre-straining on the yield behavior of steel A2 after aging treatments at 50°C. The received steel sheet had been temper-rolled to an elongation of 0.8% in steel works. The sheet was pre-strained in tension to 2 % or temper rolled to 1.5 % in laboratory before the aging treatments. It is clear that the reappearance of yield point elongation and the increase in yield strength are faster when the steel is pre-strained in ten-sion. The stamping of steel sheet for outer-body panels of automobile is, in general, a sort of biaxial tensile deformations whose strains are around 2%. It is true especially at the exposed area. The denting is also the biaxial tensile deformation to increase the surface area. As the directions of stress applied at stamping and at denting are nearly equal, the authors believe that the bake-hardenability to value the im-

provement of dent-resistance of the panel by paint-baking is better measured by the uniaxial tensile test-ing to the same direction for the uniaxially 2 % pre-strained tensile specimen than by tensile testing for the temper-rolled specimen.

Above two considerations indicate that the non-aging and bake-hardenable steel sheet is produced in easier way than that expected from Fig. 11. Prac-

tically, its mass production in steel works by box-annealing has proved that it exhibits substantially

non-aging quality. And now the steel with the ulti-mate tensile strength of more than 350 MPa is suc-

cessfully used for outer body panels of automobiles. The steel can also be produced by controlling the

amount of nitrogen in rimmed steel. However, the strain aging by nitrogen results in the larger yield

point elongation than that by carbon as shown in Fig. 3. The more strict control in the amount of

nitrogen can be required to produce a non-aging and bake-hardenable rimmed steel.

VI. Summary and Conclusions

(1) The bake-hardenable and substantially non-aging Al-killed steel sheet with high r-value has been developed by the box-annealing process. The steel is expected to be applicable to the outer body panels of automobiles and exhibits both the sufficient form-

Fig. 11.

Relation between time and temperature for steels

to show the same yield point elongation by the

diffusion of carbon atoms. Schematic diagram

obtained by the calculation of Hundy's equation.

Fig. 12. Effect of the type of pre-straining on the

behavior of steel A2, open-coil annealed.yield

Technical Report

Transactions Is", Vol. 21, 1981 (811)

ability and the dent-resistance. (2) The bake-hardenability is mainly due to the

strain aging of carbon. The adjustment of the bake-hardenability in the range of 35 and 60 MPa offers the non-aging quality to the box-annealed or con-tinuously annealed steel. The adjustment corresponds to the control of carbon in solution to approximately 10 ppm.

(3) There are two ways to retain carbon in solu-tion after the box annealing. One is to decrease the total carbon content to less than 0.02%. Another is to raise the annealing temperature to intercritical range without changing the total carbon content and increase the cooling rate to nearly 80°C/h. In the latter way, the open coil annealing furnace must be used. The relations between the carbon content and the annealing temperature is shown in Fig. 10.

(4) The additions of silicon and phosphorus and the reduction in manganese content are also effective to increase the bake-hardenability of the box-annealed steel, while the addition of chromium decreases it.

(5) The effects of the total carbon content and the annealing temperature on the bake-hardenability are correlated with the distance between cementite particles as the precipitation sites for carbon. The effect of the substitutional elements is ascribed mainly to the change in the activation energy to form and

grow a new lattice of cementite in the cooling stage.

Acknowledgements

The authors wish to thank Dr. T. Yukitoshi and Dr. Y. Hayashi of Sumitomo Metal Industries, Ltd. for their continued support and encouragement. They are also grateful to Messrs. S. Tsunematsu and S. Nakai of Wakayama and Kashima Steel Works for their cooperation in the mill productions.

REFERENCES

1) K. Yoshida: Trans. ISIJ, 19 (1979), 257. 2) M. Kojima and Y. Hayashi: SAE Preprint 800369, Feb.,

3) 4)

5)

6) 7)

8)

9)

10) 11)

12)

13) 14)

15)

16)

17)

18)

19)

20)

21)

22) 23)

24) 25)

26)

27)

(1980). J. D. Baird: Iron Steel, 36 (1963), 186. K. Araki, S. Fukunaka and K. Uchida: Trans. ISIJ, 17

(1977), 701. K. Aoki, S. Sekino and T. Fujishima: J. Japan Inst. Met., 26 (1962), 47.

J. F. Butler: Trans. Met. Soc. RIME, 224 (1962), 89. P. N. Richards and K. V. Barrett: Trans. ASM, 58 (1965), 601. K. Takashina and S. Harada : Tetsu-to-Hagane, 54 (1968), 5191. Y. Imai, T. Masumoto and M. Sakamoto: Bull. Japan Inst. Metals, 7 (1968), 137. D. A. Leak and G. M. Leak: JISI, 189 (1958), 256. H. Borchers and W. Konig: Arch Eisenhuttenw., 34 (1963), 453. S. N. Polyakov and K. F. Starodubov: Fizika Metal., 6

(1958), 1110. G. Lagerberg: Acta Met., 7 (1959), 137. E. F. Petrova, M. I. Lapshina and L. A. Shvartsman: Metalloved i Termichesk. Obrabotka Metal, (1960), No. 4, 22. W. C. Leslie : Acta Met., 9 (1961), 1004. F. E. Fujita, Y. Ono and Y. Inokuti: J. Japan Inst. Metals, 32 (1968), 328. R. W. Gurry, J. Christakos and L. S. Darken: Trans. ASM, 53 (1961), 187. T. Sato and T. Nishizawa: J. Japan Inst. Metals, 19 (1955), 385. T. Kunitake: J. Japan Inst. Metals, 29 (1965), 653 and 30

(1966), 481. K. Kawamoto, M. Miyahara, N. Hoshino, K. Kohrida, S. Nomura and M. Sato: Trans. ISIJ, 20 (1980), B-278. W. Dickenscheid and H. J. Seemann : Rev. Met., LV

(1958), No. 9, 872. C. A. Wert: J. Metals, 188 (1950), 1242. K. Kamber, D. Keefer and C. West: Acta Met., 9 (1961), 403. B. B. Hundy: JISI, 178 (1954), 34. A. H. Cottrell and B. A. Bilby: Proc. Phys. Soc., 62A (1949), 49. C. Shiga, T. Sasaki and K. Matsumura : Kawasaki Steel Technical Report, 5 (1973), 191. H. P. Tardif and C. S. Ball: JISI, 182 (1956), 9.