International Journal of Offshore and Polar EngineeringVol. 14, No. 2, June 2004 (ISSN 1053-5381)Copyright © by The International Society of Offshore and Polar Engineers

Effects of Cooling Rate and Isothermal Holding on Precipitation BehaviorDuring Solidification of Nb-Ti Bearing HSLA Steels

Hyun Jo Jun and C. G. ParkDepartment of Materials Science and Engineering, Pohang University of Science and Technology

Pohang, Korea

K. B. KangTechnical Research Laboratories, POSCO, Pohang, Korea

ABSTRACT

The effects of the cooling rate and isothermal holding on precipitation behavior during solidification have been investigatedin 0.063C-0.017Ti-0.056Nb HSLA steels. The precipitates identified in an as-cast slab were semidendritic, dendritic or rod-like Nb-rich (Nb,Ti)(C,N). The morphology and chemistry were quite different when these precipitates formed after reheatingand/or hot-rolling processes. No precipitation has been observed at the end of the solidification and the continuous coolingdown to 800°C. ICP and TEM analyses indicated that most of the Nb and Ti was mainly precipitated into carbonitrides bythe isothermal holding at a temperature range of 900° to 1000°C. In the case of a continuous casting process, the isothermalholding region corresponds to a certain flat cooling region, probably due to the latent heat of solidification. The Nb-richcarbonitrides formed during solidification are associated with the microsegregation of Ti and Nb in the interdendriticregion.

INTRODUCTION

Micro-alloying elements in steels, such as Ti, Nb and V, canfacilitate grain refinement through precipitation in austenite, andcontribute to dispersion hardening through precipitation in fer-rite during or after � → � transformation. Especially, titaniumhas frequently been added to high-strength, low-alloy (HSLA)steels to enhance the control of austenite during the weldingor reheating process. In addition, TiN precipitates of a suitablesize can suppress austenite grain coarsening in subsequent high-temperature processes such as welding or reheating, thus improv-ing the toughness of final steel products (Cuddy, 1983). Multi-micro-alloying, then, can lead to the formation of complex com-pounds, which can influence the mechanical properties of theHSLA steels.

Very large precipitates could also be found in Nb-Ti HSLAsteels produced by a continuous casting process (Chen, 1987;Zhou, 1996). It is, however, unlikely that these large precipi-tates can play any useful role in refining grain size. Instead,their harmful effects on the distribution and formation of smallerprecipitates can weaken the role of Nb and Ti as grain refinersduring the subsequent reheating and hot rolling processes (Hong,2003).

The objectives of the present study are both to identify theprecipitates formed in a continuously cast slab, and to examine theeffects of post-solidification cooling on the precipitation behaviorin Nb-Ti bearing HSLA steels.

Received August 18, 2003; revised manuscript received by the editorsFebruary 9, 2004. The original version (prior to the final revisedmanuscript) was presented at the ISOPE Symposium on High-Per-formance Materials in Offshore Industry, the 13th International Off-shore and Polar Engineering Conference (ISOPE-2003), Honolulu,Hawaii, USA, May 25–30, 2003.

KEY WORDS: Precipitate, carbonitride, solidification, cooling rate, iso-thermal holding.

EXPERIMENTALS

Materials

Table 1 gives the chemical composition of the investigated slab.POSCO (Pohang Iron & Steel Co. Ltd.) produced the slab of Nb-Ti bearing HSLA steel.

Simulation for Solidification and Cooling

Simulation of the solidification and cooling was conducted usingthe Gleeble 3500 system (Dynamic System Inc.). Rod samples120 mm in length and 10 mm in diameter were machined fromthe 1/4 and 3/4 positions of the slab thickness. The quartz tubewas used for shielding the sample during the melting and solid-ification processes. The samples welded with an R-type thermo-couple were heated from room temperature to melting temperature(1470°∼1480°C). In order to fully dissolve Nb and Ti in steels,the melt was held for 3 min and cooled down as follows, as shownin Fig. 1: (1) Cooling to 800°C at various cooling rates (10°, 100°,200°, 300°C/min) followed by water quenching in order to iden-tify the effect of the cooling rate on precipitation, and (2) coolingto various holding temperatures from 700° to 1200°C at the cool-ing rate of 100°C/min, followed by isothermal holding for 30 minin order to evaluate the precipitate-forming temperature.

Quantitative Analysis and TEM Observations of Precipitates

The quantitative amount of Nb and Ti precipitated was mea-sured using inductively-coupled plasma (ICP) spectroscopy. Atfirst, a few grams of samples were electro-chemically dis-solved, and then filtered with a polycarbonate membrane fil-

C Si Mn Nb Ti N

Slab 0.063 0.2 1.58 0.056 0.017 0.006

Table 1 Chemical composition of investigated slab (wt.%)

International Journal of Offshore and Polar Engineering, Vol. 14, No. 2, June 2004 133

Fig. 1 Thermal schedules of simulation for solidification andcooling: (a) continuous cooling, and (b) isothermal holding

ter. The filtered precipitates were melted in a buffer solution(Na2CO3 : H3BO3 = 3 : 1) and mixed with 10% HCl solution toperform the ICP analysis.

The morphology, distribution and chemical composition of theprecipitates were examined using TEM-extracted replicas. Thereplica film (carbon or aluminum) was deposited on the sur-face of the sample to extract precipitates from various samplesurfaces after mechanical polishing and electro-chemical etch-ing. The replicas released by electro-polishing were then exam-ined using FEG-TEM (JEM-2010F) equipped with ATW-EDS(Oxford Inc.)

RESULTS

Precipitates Formed in Cast Slab

The morphology of the precipitates formed in a continuouslycast slab was quite complex. Fig. 2 shows 3 different types of typ-ical precipitates: semidendritic (a), dendritic (b), and rod-like (c).The chemistry of these precipitates was investigated using nano-beam EDS analysis. All precipitates were identified as complexcarbonitrides mixed with Nb and Ti (Fig. 2d and e).

Most of the precipitates formed in the slab were revealed assemidendritic; these were quite small, ranging from 50 to 200 nm.Large dendritic precipitates varied from 0.5 to 1 �m were rarelyobserved. Rod-like precipitates with a length of up to a few �mwere usually formed along austenite grain boundaries. An impor-tant fact is that these dendritic carbonitrides are not Ti-rich, butNb-rich. The fraction of Ti in these precipitates, Ti/(Ti+Nb), wasapproximately 0.3.

The morphology and chemistry of the precipitates formed inthe cast slab were quite different from those of the precipitatesformed after reheating and/or hot rolling. It has been well reportedthat the precipitates of Nb-Ti HSLA steels are identified as largecuboidal TiN formed after the reheating process, as shown inFig. 3a (Hong, 2003). Plus, the strain-induced NbC carbides usu-ally found in hot-rolled steels are additionally nucleated heteroge-neously at the pre-exited TiN, as shown in Fig. 3b (Hong, 2002).Thus, the present results imply that the precipitates found in castslabs are quite different.

Effects of Cooling Rate on Precipitation

It is generally well known that the dendritic shape is a typi-cal morphology of solidification (liquid → solid transformation),which is dependent on the cooling rate. It is then suggested thatprecipitation starts simultaneously during the solidification fromliquid to solid, and that precipitation is also dependent on thecooling rate. Thus, simulation for solidification and cooling wereperformed down to 800°C at different cooling rates (10°, 100°,200°, and 300°C/min).

(a)

(b)

(c)

(d)

(e)

100nm

100nm

200nm

Ti/(Ti+Nb) ~ 0.3

Fig. 2 Al replicas showing morphology of precipitates in slab:(a) semidendritic, (b) dendritic, (c) rod-like and (d)–(e) EDS spec-tra of (b) and (c), respectively (Al and Cu peaks generated fromAl film and Cu grid)

At first, the carbon replica of the sample quenched at the endof solidification did not reveal any precipitate, which confirmedthe complete dissolution of Nb and Ti.

The amounts of Nb and Ti precipitated in the cast slab and thesimulated samples were measured by the ICP method (Fig. 4).In the cast slab, both Nb and Ti were fully precipitated into car-

134 Effects of Cooling Rate and Isothermal Holding on Precipitation Behavior During Solidification of Nb-Ti Bearing HSLA Steels

(a)

(b)

(c)

200nm

A : (Nb,Ti)C

B : (Ti,Nb)(N,C)

Fig. 3 TEM micrographs showing typical morphology of precipi-tates formed after (a) reheating and (b) hot-rolling processes, and(c) EDS spectra of precipitate shown in (b) (Hong, 2002, 2003)

bide or nitride (Fig. 4A). In the simulated samples, however, theamounts of both Nb and Ti were significantly lower than those inthe cast slab (Fig. 4B). The fraction of Nb and Ti contained in theprecipitates of the simulated samples was evaluated as 5% to 10%and 10% to 30%, respectively, of total additions. The amountsgradually decreased with an increasing cooling rate, because ofinsufficient diffusion time with the increase of the cooling rate.The low levels of the precipitated amounts in the simulated sam-ples suggest that the solidification and the following continuouscooling (down to 800°C) are not the major precipitation region.

In order to evaluate the morphology and chemistry of the pre-cipitates depending on cooling rates, carbon replicas of the 2 sam-ples at the cooling rate of 10° and 300°C/min were observed usingFEG-TEM with nano-beam EDS. The precipitates were rarely

slab 10C/m 100C/m 200C/m 300C/m

0.00

0.01

0.02

0.03

0.04

0.05

0.06BA

Total Ti addition

Total Nb addition

Am

ou

nt (w

t.%

)

Cooling rate

Nb

Ti

Fig. 4 Amounts of Nb and Ti precipitated in precipitates formedin cast slab, and in simulated samples measured by dissolutionand filtration

observed, compared to those of the cast slab. The morphologyof most precipitates in the simulated samples was identified assemidendritic (Fig. 5). The size of the precipitates observed inthe sample with slow cooling rate (10°C/min) varied from 50 to150 nm, which was larger than that of the precipitates (below50 nm) in the sample with fast cooling rate (300°C/min). Thechemistry of the precipitates observed in the 2 samples was iden-tified as Nb-rich (Nb,Ti)(N,C), which was similar to that of theprecipitates found in the cast slab. However, the atomic fraction ofTi, Ti/(Ti+Nb) was different between the 2 samples: about 1/3 inthe slow-cooled sample and about 1/4 in the fast-cooled sample.These results imply that the amount of Nb and Ti in the precipi-tates decreases with an increasing cooling rate, which is in goodagreement with the previous results obtained from ICP analysis.

Precipitate Forming Temperature During Solidification

In order to evaluate which temperature region is an effec-tive region for precipitation during solidification and cooling,isothermal holding experiments were performed using the Glee-ble 3500 system. The precipitates were examined in the simulatedsamples, which were isothermally held at various temperatures(700°∼1200°C) for 30 min.

Fig. 6 shows the amount of Nb and Ti contained in the pre-cipitates measured by the ICP method. Most of the Nb and Tiwas precipitated in the samples held at the temperature regionof 900°∼1000°C. In the samples held at 700° and 800°C, theamounts were approximately 50% of the total additions. No pre-cipitate was observed in the sample held at 1200°C.

TEM observation of the precipitates also revealed results similarto those measured by the ICP analysis. The precipitates of the sam-ple held at 800°C in Fig. 7 showed quite a small (below 50 nm insize) and irregular dendritic shape because of low diffusivity in lowtemperatures. The precipitates of the sample held at 1000°C had 3morphologies: semidendritic (a), dendritic (b) and rod-like (c), asshown in Fig. 8. These results are in good agreement with thoseof the precipitates formed in the cast slab (Fig. 2).

The chemistry of precipitates, as shown in Fig. 8(d) was similarto that of the cast slab. No precipitate could be observed in thesample held at 1200°C. These results indicate that the isothermalholding at 900°∼1000°C during solidification and cooling is themajor precipitation region of Nb and Ti.

International Journal of Offshore and Polar Engineering, Vol. 14, No. 2, June 2004 135

(a)

(b)

10 °C/min

300 °C/min

Ti/(Ti+Nb)~0.33

Ti/(Ti+Nb)~0.24

Fig. 5 TEM micrographs and EDS spectra of samples cooled atrate of (a) 10°C/min, and (b) 300°C/min

700 800 900 1000 1100 1200

0.00

0.01

0.02

0.03

0.04

0.05

0.06

Total Ti addition

Total Nb addition

CR=100 0C/min

Holding time: 30 min

Am

ou

nt (w

t.%

)

Holding temperature (C)

Nb

Ti

Fig. 6 Amounts of Nb and Ti in precipitates depending on hold-ing temperatures

80nm

(a) (b)

Fig. 7 TEM micrographs showing precipitates in simulated sam-ple held at 800°C for 30 min

DISCUSSIONS

ICP and TEM analyses on the precipitates indicated that mostof the Nb and Ti were precipitated into carbonitrides by isother-mal holding from 900° to 1000°C. In case of a continuouscasting process, the isothermal holding region corresponds to a

(a)

(b)

(c)

(d) Ti/(Ti+Nb)~0.3

Fig. 8 TEM micrographs showing precipitates and EDS spectrumholding at 1000°C for 30 min

136 Effects of Cooling Rate and Isothermal Holding on Precipitation Behavior During Solidification of Nb-Ti Bearing HSLA Steels

Fig. 9 Thermal history of continuous casting process (Subrama-nian, 1985)

certain very slow cooling region, probably due to the latent heatof solidification. The thermal history measured during the contin-uous casting process showed the flat cooling region to be similarto the isothermal holding region at 1050°∼1150°C (Subramanian,1985), as shown in Fig. 9. The temperature of this flat region issomewhat higher than that of the simulation results, which seemsto be caused by the latent heat of solidification.

Nano-beam EDS results reveal that the precipitates observedin both the cast slab and the simulated samples are not Ti-rich,but Nb-rich. If these precipitates are formed at the end of solidi-fication around 1400°C, they should be the Ti-rich carbonitrides,as predicted by the regular solution model (Houghton, 1982).Houghton also proposed that TiN particles are precipitated firstin the liquid and at high temperatures above 1300°C in austenite,followed by the precipitation of Nb and Ti carbonitrides as thetemperature decreases. No precipitate, however, was observed inthe high-temperature region in the present study. Although embryomay form in the high-temperature region of austenite, the presentresults reveal that most of the Ti and Nb is precipitated in thesupersaturated austenite at around 1000°C. This is in good agree-ment with the equilibrium composition at around 1000°C calcu-lated by Thermo-calc, TCFE2000 (Fig. 10).

Nb-rich dendritic carbonitrides are not observed after thereheating and hot-rolling processes, as already mentioned byHong (2002, 2003). The Nb-rich dendrite carbonitrides observed

600 800 1000 1200 1400 1600

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

Composition

range observed by EDS

Am

oun

t (w

t.%

)

Temp. (0C)

Nb

Ti

Fig. 10 Equilibrium composition of complex carbonitride calcu-latd by Thermo-Calc, TCFE2000

in the present study may be attributed to the microsegregationof Nb and Ti in the interdendritic liquid, since these were onlyformed during solidification. Assuming no diffusion in the solidand complete mixing in the liquid (Flemings, 1974), the local seg-regation can be calculated as:

CL/C0 = f�k−1L (1)

where CL and C0 are liquid and initial composition, respectively,fL is liquid fraction, and k is the equilibrium distribution coeffi-cient. The k values of Nb and Ti are 0.3 and 0.61, respectively(Elliott, 1965). The segregation ratios CL/C0 for Nb and Ti in5% liquid fraction are 8 and 3, respectively. This calculation indi-cates that the stronger segregation ratio of Nb may be responsi-ble for the formation of Nb-rich dendritic carbonitrides. Althoughthe interdendritic liquid homogenizes partially during solidifica-tion and cooling, the locally supersaturated austenite may remainbecause of the absence of precipitation during the solidificationand cooling. Zhou (1996) also reported that MnS and TiN formedin liquid could catalyze the formation of complex carbonitridesduring the solidification and cooling. However, there was no for-mation of MnS and TiN in the present study.

It is believed that the dendritic carbonitrides formed in anas-cast slab can affect the size and distribution of the precipi-tates formed during the following reheating and hot-rolling pro-cesses. From the previous results (Hong, 2003), it is predicted thatthe Nb-rich dendritic carbonitrides will dissolve, and then repre-cipitate to 2 kinds of carbonitrides—cuboidal TiN and Ti-rich(Ti,Nb)(N,C)—during the reheating process because of the ther-mal instability of Nb-rich carbonitrides. These reprecipitated car-bonitrides can play important roles in the precipitation kinetics ofNbC, as heterogeneous nucleation sites during hot rolling (Hong,2002; Jun, 2003). Detailed discussion of the relationship betweenthe precipitates of an as-cast slab and those of reheating and hot-rolling samples will be found elsewhere (Jun, 2003).

CONCLUSIONS

The simulation for solidification and cooling was performed inNb-Ti bearing HSLA steel. The morphology and microchemistryof the precipitates observed in a continuously cast slab and simu-lated samples were investigated by TEM observation with nano-beam EDS. The main results are:

• The precipitates in a continuously cast slab show 3 differentmorphologies, semidendritic, dendritic and rod-like. The chem-istry of these precipitates is Nb-rich, which is thermodynamicallystable at temperatures lower than 1000°C.

• The amounts of Nb and Ti in the precipitates are almost inde-pendent of the cooling rate during solidification. ICP and TEManalyses indicate that most Nb and Ti is mainly precipitated intocarbonitrides by the isothermal holding at a temperature range of900° to 1000°C.

• In case of a continuous casting process, the isothermal hold-ing region corresponds to a certain flat cooling region, probablydue to the latent heat of solidification.

• The Nb-rich carbonitrides formed during solidification areassociated with the microsegregation of Ti and Nb in the inter-dendritic region.

ACKNOWLEDGEMENTS

The authors would like to thank POSCO (Pohang Iron & SteelCo. Ltd.) for technical and financial support.

International Journal of Offshore and Polar Engineering, Vol. 14, No. 2, June 2004 137

REFERENCES

Chen, Z, Loretto, MH, and Cochrane, RC (1987). “Nature ofLarge Precipitates in Titanium-containing HSLA Steels,” MatlsSci and Tech, Vol 3, pp 836–844.

Cuddy, LJ, Lally, JS, and Porter, LF (1983). “Improvement ofToughness in the HAZ of High-heat-input Welds in ShipSteels,” Proc Conf HSLA Steels Tech and Appl, Am Soc forMetals, Ohio, pp 697–703.

Elliott, RP (1965). Constitution of Binary Alloys, McGraw-Hill,New York.

Flemings, MC (1974). Solidification Processing, McGraw-Hill,New York.

Hong, SG, Jun, HJ, Kang, KB, and Park, CG (2003). “Evolutionof Precipitates in the Nb-Ti-V Microalloyed HSLA Steels Dur-ing Reheating,” Scripta Materialia, Vol 48, pp 1201–1206.

Hong, SG, Kang, KB, and Park, CG (2002). “Stain-induced Pre-cipitation of NbC in Nb and Nb-Ti Microalloyed HSLA Steels,”Scripta Materialia, Vol 46, pp 163–168.

Houghton, DC, Weatherly, GC, and Embery, JD (1982). “Char-acterization of Carbonitrides in Ti bearing HSLA Steels,”Proc Conf Thermomech Process of Microall Austenite, AIME,Warrendale, Pennsylvania, pp 267–292.

Jun, HJ, Hong, SG, Kang, KB, and Park, CG (2004). “Effectsof Cooling Rate and Isothermal Holding on the PrecipitationBehavior During Continuous Casting of Nb-Ti Bearing HSLASteels,” Scripta Materialia, Vol 49, pp 1081–1086.

Subramanian, SV, Shima, S, Ocampo, G, Castillo, T, Embury,JD, and Purdy, GR (1985). “The Effect of Nitrogen Con-tent on the Evolution of Precipitates and on the Process-ing of Titanium/Niobium Bearing HSLA Steel Slab,” ProcConf HSLA Steels; Metallur and Appl, Beijing, ASM Int,pp 151–161.

Zhou, C, and Priestner, P (1996). “The Evolution of Pre-cipitates in Nb-Ti Microalloyed Steels During Solidificationand Post-Solidification Cooling,” ISIJ Int, Vol 36, No 11,pp 1397–1405.