Difference Analysis in Steel Cleanness between Two RH Treatment Modes for SPHC Grade

Transcription

1 , pp Difference Analysis in Steel Cleanness between Two RH Treatment Modes for SPHC Grade Min WANG, 1,2) * Yan-ping BAO, 1) Li-hua ZHAO, 1) Quan YANG 2) and Lu LIN 1) 1) State Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, Haidian District, Xueyuan Road 30#, Beijing, China. 2) National Engineering Research Center of Flat Rolling Equipment, University of Science and Technology Beijing, Haidian District, Xueyuan Road 30#, Beijing, China. (Received on January 20, 2015; accepted on April 2, 2015) This paper studied the steel cleanness difference of grade SPHC by comparing two RH modes. Experiments were carried out at 210 ton BOF, RH and 60 ton tundish by sampling systematically. Under mode I, free oxygen was killed during BOF tapping; and inclusions had sufficient removal during RH vacuum treatment; under mode II, RH vacuum carbon de-oxidation was adopted firstly; and then residual oxygen was killed by grain aluminum. Results showed that, (1) total generation amount of inclusions in mode II was 1/2 of that in mode I due to low residual oxygen after vacuum carbon de-oxidation, but steel cleanness in mode I was better in tundish due to long inclusions removal time; (2) under mode I, total oxygen in tundish could be controlled below 20 ppm; and cast-ability in continuous casting process reached 15 heats; (3) under mode II, nitrogen could be controlled below <15 ppm, which was 1/2 of that in mode I; (4) by adopting mode II to reduce total generation amount of inclusions and nitrogen picking up, also to prolong 20 to 30 minutes calming time before casting to guarantee inclusions removal, low nitrogen, low total oxygen and good cast-ability could realize. KEY WORDS: RH treatment modes; SPHC; vacuum carbon de-oxidation; Al 2O 3 inclusions; steel cleanness. 1. Introduction * Corresponding author: worldmind@163.com DOI: RH (Ruhrstahl Heraeus) possessed many refining functions such as decarburization, dehydrogenation, nitrogen removal, inclusions removal and chemical compositions adjustment. 1 3) It became more and more important for high quality steel production, especially for low carbon and ultra low carbon steel. 4 6) As requirement of different grades, different RH treatment modes were developed. Light treatment mode was suitable for low carbon steel, which adopted vacuum carbon de-oxidation to reduce free oxygen in liquid steel and alloy consumption. Deep decarburization mode was suitable for ultra low carbon steel such as interstitial free steel with carbon below , which adopted high vacuum atmosphere to guarantee sufficient reaction of carbon and oxygen in liquid steel; the carbon can be reduce below within 15 minutes RH vacuum treatment. About rapid decarburization controlling during RH vacuum treatment process, many relevant models were proposed. 7 11) RH OB heating mode by Al O reaction always was adopted when temperature of liquid steel was lower than target temperature; oxygen was injection into liquid steel by lance situated in top of vacuum chamber and then aluminum was added into liquid steel to reacted with soluble oxygen; the chemical reaction energy was used to heat liquid steel. Normal treatment mode was suitable for inclusions removal and dehydrogenation; under this mode, free oxygen in liquid steel was killed before RH treatment; and inclusions aggregated sufficiently and easily were removed during vacuum process due to early formation of inclusions. 12,13) In order to meet high quality and low cost requirement of different steel grades, different RH treatment modes were applied crossly. How to reach RH functions fully, exploratory research should be carried out. This paper compared the steel cleanness difference of low carbon steel (SPHC) under two different RH treatment modes (mode I: normal treatment; mode II: light treatment). By this study, more controlling characteristics and their effect on steel cleanness under different RH treatment modes were summarized. 2. Experimental Procedure and Methods Experiment was performed on low carbon Al-Killed steel with grade SPHC produced by QIAN AN STEEL works. Chemical composition of grade SPHC was listed in Table 1. The production process was described as following: 210 Table 1. Chemical composition of grade SPHC. SPHC C Si Mn P S Alt N range < target ISIJ 1652

2 Table 2. Process control difference of two modes. Mode BOF end-point BOF taping RH process I [C]<0.035% II [C]=0.04% 0.06% (1) De-oxidation by ferro-aluminium and [Al] s>0.015% (2) Ferromanganese alloying (1) No de-oxidation (2) Ferromanganese alloying (1) Remove inclusions by circulation (2) Adjust chemical composition and temperature of liquid steel (1) Vacuum carbon de-oxidation; (2) Addition carbon and Al-killed (3) Adjust chemical compositions and temperature of liquid steel (4) Remove inclusions by circulation ton BOF RH 60 ton tundish mm*210 mm CC. In order to compare the difference of two different RH treatment modes and their effect on steel cleanness, we carried out 6 heats industrial experiments for two different RH treatment modes. The process control difference of two modes was listed in Table 2. In mode I, free oxygen in liquid steel was killed during tapping with acid-soluble aluminum ([Al] s) above 0.015%; and carbon didn t be removed during RH treatment due to no free oxygen in liquid steel; so, BOF end point kept carbon content below 0.035% to meet requirement of final product. The main functions of vacuum treatment under mode I were inclusions removal and compositions adjustment. In model II, keep rimming steel during tapping and carbon content at BOF end-point was controlled between 0.04% to 0.06% to meet vacuum carbon de-oxidation during RH vacuum treatment; a part of free oxygen reacted with carbon in vacuum atmosphere; and residual oxygen was killed by aluminum addition. Steel cleanness difference between two modes was evaluated by systematic sampling. Two types of samplers were chosen. Pail sampler (diameter 120 mm, height 100 mm) was for inclusions analysis; pin sampler (diameter 8 mm, length 100 mm) was for oxygen and nitrogen analysis. The detail sampling scheme was listed in Table 3. Morphologies and compositions of inclusions were observed by scanning electron microscope (SEM, ZEISS ULTRA55) and energy dispersive spectroscopy (EDS, OXFORD INSTRUMENTS INCA X-MAX50). Total oxygen and nitrogen were detected by analyzer (NCS ON-3000). Chemical compositions were analysis by spectrometer. Activity oxygen in liquid steel was determined by oxygen probe. 3. Results Table 3. Sampling scheme of experiment. Sample Pin sampler Pail-sampler Mode I Mode II 6 samples (RH arrived, RH vacuum 5 min/ 10 min/15 min/20 min, tundish) 7 samples (RH arrived, RH decarburization end, Aluminum addition 3 min/ 6 min, RH vacuum destroyed, RH finished, tundish) 2 samples (RH finished, tundish) 2 samples (RH finished, tundish) Experimental results were shown in following tables. Labels F and Q represented the experimental heats of mode I and mode II respectively. Experimental process results data record were showed in Table 4. For oxygen content detection, values with star on top right represented free oxygen in liquid steel analyzed by oxygen probe; other values were total oxygen (T.O) detected by oxygen analyzer. Free oxygen at BOF endpoint was between 700 ppm and 850 ppm under mode I due to low carbon content in liquid steel. At BOF endpoint, carbon content ranges were controlled to and to in two modes respectively. In mode I, oxygen was killed during tapping by addition Al Fe (containing aluminum 42%); carbon hardly was removed during RH vacuum treatment and increased gradually in subsequent process; so carbon content was controlled low limited value of the grade requirement at BOF end point. In mode II, carbon and free oxygen decreased simultaneously during RH vacuum carbon de-oxidation and then grain aluminum was added into liquid steel to kill residual oxygen and adjust acid-soluble aluminum; Mn Fe alloy (with manganese 78%) and carburant were added into liquid steel to adjust compositions of liquid steel. Alloys consumption was summarized in Table 5. In heat F1, 860 kg ferro-aluminium alloy and 366 kg ferro-manganese alloy were added into liquid steel during BOF tapping to kill oxygen and adjust acid-soluble aluminum and manganese; additional 102 kg grain aluminum and 181 kg ferro-manganese were supplied during RH process due to low acid-soluble aluminum and manganese in liquid steel after RH arrived. Alloy control of heats F2 and F3 were similar with F1, but no additional grain aluminum supply during RH process. In mode II, entire alloy such as ferro-manganese, grain aluminum and carburant were added during RH process based on the requirement of final chemical compositions; no alloys was added during tapping process. Pail samples from RH finished and tundish were used to inclusions analysis. In two modes, almost all inclusions were Al 2O 3; some Al 2O 3 with MnS shell appeared in tundish samples. The morphologies of inclusions were showed in Figs. 1 to 4. In mode I, Al 2O 3 inclusions at RH finished were mainly below 5 μm (see Figs. 1(a) 1(f)); and some rod-like Al 2O 3 and cluster Al 2O 3 with size above 10 μm also were found (see Figs. 1(g) and 1(h)). At tundish sample, the morphologies of inclusions gradually became spherical; and inclusions sizes mainly were below 5 μm (see Fig. 2). In mode II, the sizes of inclusions were larger than that in mode I; and most of Al 2O 3 were polygonal and rod-like (see Fig. 3). Some duel phase inclusions were found in tundish sample (see Fig. 4(d)); the inner and outer cores both were pure Al 2O 3, which also was found at sample of RH finished in mode I (see Fig. 1(e)); the formation mechanism will be explained in following section. 4. Discussion 4.1. Generation Amount of Inclusions Comparison in Two Modes When deoxidizer (Ferro-aluminum alloy or Al grain) was ISIJ

3 Table 4. Results data record of two modes. F1 F2 F3 Q1 Q2 Q3 time [C]/ 10 6 [Al] s/ 10 6 [O]/ 10 6 [N]/ 10 6 description BOF-end * BOF end point 20: RH arrived 20:55 Vacuum start 21: : Pin-samples 21: : RH finished 21: Tundish cast on 21:49 Pail sample in tundish BOF-end * BOF end point composition 21: * RH arrived 21: : : Pin-samples 21: : RH finished 22: Tundish cast on 22:27 Pail sample in tundish BOF end * BOF end point 22: RH arrived 22:08 Vacuum start 22: : Pin-samples 22: RH finished 22: Tundish cast on 23:02 Pail-sample in tundish BOF end BOF end point composition 18: * 18:11 285* Free oxygen and temperature determination 18: * Free oxygen and temperature determination after Al-Killed 18: : Pin-samples 18: : Tundish cast on 19:15 Pail-sample in tundish BOF end BOF end point composition 18: * 18:44 Free oxygen and temperature determination 18: * 18: * Free oxygen and temperature determination after Al-Killed 18: : Pin-samples in tundish 19: : Tundish cast on 19:42 Pail-sample BOF end BOF end point composition 19: * 19:14 19:18 Free oxygen and temperature determination 19:22 224* 19: * 19: : Pin-samples 19: : Tundish cast on 20:14 Pail-sample in tundish 2015 ISIJ 1654

4 Table 5. Summary of alloy consumption at different location. Mode Location ferro-aluminium/kg ferro-manganese/kg Carburant/kg Al grain/kg F1 F2 F3 Q1 Q2 Q3 BOF tapping RH BOF tapping RH BOF tapping RH 190 BOF tapping RH BOF tapping RH BOF tapping RH Fig. 1. Typical Al 2O 3 morphologies at RH finished in Mode I. Fig. 2. Typical Al 2O 3 inclusions morphologies at tundish in Mode I. added into liquid steel, most of free oxygen in liquid steel transformed to be Al 2 O 3 inclusions. As removal of Al 2 O 3 inclusions, total oxygen decreased. So, generation amount of inclusions was closely relevant with free oxygen before final de-oxygen and re-oxidation amount, as described in Eq. (1). De-oxidation amount and re-oxidation amount is calculated by Eqs. (1) to (3). The yield rate of alloy is calculated by Eq. (4) ISIJ

5 Fig. 3. Typical Al 2O 3 morphologies at RH finished in Mode II. Fig. 4. Typical Al 2O 3 morphologies at tundish in Mode II. Mtotal = mde-oxidation + mre-oxidation m de-oxidation 102 = [ O] Free (1)... (2) mre-oxidation = 54 mal γ Al Wsteel [ O] Free Wsteel ([ Al] s-tundish [ Al] s-bof ) 48 Wsteel (3) 54 W O + W Al s tundish Al s BOF 48 [ ] ([ ] [ ] ) steel Free steel η Al = mal (4) Here, M total represented total generation amount of inclusions, ppm; m de-oxidation represented de-oxidation inclusions γ Al amount, ppm; it was calculated by free oxygen in liquid steel before final oxygen killed; m re-oxidation represented reoxidation amount of inclusions, ppm; it was calculated by alloy aluminum loss. η Al represented the effective yield rate of aluminum, %; γ Al represented the weight content of aluminum in ferro-aluminum or grain aluminum respectively; W steel represented the weight of liquid steel, kg; [Al] s and [O] Free represented acid-soluble aluminum and free oxygen in liquid steel, %; m Al represented the weight of aluminum or ferro-aluminum addition into liquid steel, kg. In Eq. (4), left part of the denominator represented aluminum consumption for de-oxidation; right part of the denominator represented acid-soluble aluminum adjustment. For calculation values in Eqs. (2) and (4), 102 is employed as the molecular weight of Al 2O 3, using 16 and 27 respectively as the atomic weight of the oxygen and the aluminum. 54 and 48 represented total atomic weight of aluminum and oxygen in Al 2O 3 respectively. As shown in Fig. 5, total generation amount of inclusions were , and respectively under mode I; de-oxidation product amount reached maximum value due to high free oxygen at BOF end 2015 ISIJ 1656

6 Fig. 5. Generation amount of inclusions in liquid steel in two modes. Fig. 7. Comparison for effective inclusions removal time. Fig. 6. Yield rate of aluminum in different modes. Fig. 8. Soluble aluminum loss in different modes. point. Re-oxidation product amount was between and In mode II, de-oxidation product amount decreased greatly due to low residual free oxygen after RH vacuum carbon de-oxidation; total generation amount of inclusions in mode II was 1/2 of that in mode I. Further analyses, yield rate of aluminum in different modes were compared in Fig. 6. In mode I, nearly 50% aluminum was used to kill oxygen; average 25% aluminum was used to acid-soluble aluminum adjustment. The loss yield rate of aluminum was between 17% and 43%. In mode II, due to low free oxygen after RH vacuum carbon deoxidation, less de-oxidation production formed during this process; and total aluminum consumption was half of that in mode I; and the loss yield rate of aluminum was between 30% and 43%, which was higher than that in mode I due to high oxidation of top slag under mode II. Figure 7 showed effective inclusions removal time between two modes. Here, effective removal time represented time difference between final de-oxidation and cast-on. In mode I, effective inclusions removal time was between 65 minutes and 90 minutes due to its early final de-oxidation, which almost twice of that in mode II. It meant that inclusion had more time to remove in mode I than that in mode II. Although generation amount of inclusions was less in mode II, the effective removal time was short; inclusions at tundish under mode II still was more and larger than that in mode I. So, both of the removal time and inclusion generation amount controlling was important for steel cleanness. Figure 8 showed the aluminum loss in different modes. During RH circulation process, liquid steel stirred sharply; acid-soluble aluminum loss was between per minute and per minute under mode I, which was 2 to 5 times of that from RH finished to tundish. Acid-soluble aluminum loss rate increased as increasing of initial acid-soluble aluminum. Acid-soluble aluminum loss rate in mode II was lower than that in mode I due to short circulation time under vacuum treatment and low initial acid-soluble aluminum after final de-oxidation. Keeping liquid steel cleanness needed low aluminum loss rate and high inclusions removal ISIJ

7 rate especially with high oxidation of top slag. The average content of (T.Fe+MnO) in top slag under mode I and mode II were 6% and 11% respectively. Although oxidation of top slag under mode I was lower than in mode II, acid-soluble aluminum loss was high due to early de-oxidation and long circulation time under vacuum treatment. So, the stirring of liquid steel should not be too sharply after final de-oxidation. Calming liquid steel was beneficial for preventing aluminum loss and re-oxidation Characteristic Comparison of Inclusions in Two Modes Total 95 inclusions were analyzed by SEM/EDS; the number percent of different types inclusions were shown in Fig. 9. Four different kinds of Al 2O 3 inclusions were found after Al-killed in experimental samples. First was rod-like Al 2O 3; second was polygonal Al 2O 3; third was spherical Al 2O 3; fourth was cluster Al 2O 3. In mode I, 5.2% cluster Al 2O 3 existed at RH finished samples, but it disappeared during tunish process; polygonal Al 2O 3 decreased and spherical Al 2O 3 increased greatly from RH end to tundish. In mode II, after RH finished, the rod-like Al 2O 3 and polygonal Al 2O 3 occupied 61.1% and 38.9% respectively; spherical and cluster Al 2O 3 were not found; but during tunish process, spherical and cluster Al 2O 3 appeared and rod-like Al 2O 3 reduced by 27.8%. Comparing the inclusions of tunish samples in two modes, 12.8% cluster Al 2O 3 still existed under mode II, which didn t existed under mode I. The de-oxidation product (Al 2O 3 inclusions) seemed to have better removal under mode I than mode II. Inclusions types and quantity at RH finished in mode I was similar with that at tundish under mode II. This point also was verified by comparison of effective removal time (see Fig. 7). Inclusions size distribution was shown in Fig. 10. From RH finished to tundish, large sizes of inclusions decreased gradually; in mode II, the descend trend was not obvious; in mode I, inclusions with size above 10 μm disappeared in tundish. So, the large size inclusions removed more sufficiently under mode I due to long effective removal time. Comparing Figs. 5, 8 and 10, we can see that re-oxidation amount in mode I was almost twice of that in mode II, which also was verified by acid-soluble aluminum loss in Fig. 8; but, the size and number of inclusions in tundish under mode I were better than that in mode II; it meant that, most re-oxidation product entered into top slag but not remaining in liquid steel under mode I; and it would not contaminate liquid steel directly. A reasonable explanation was that a large of acid-aluminum loss during RH circulation finally transformed to be Al 2O 3 on the surface of top slag and liquid steel; and re-oxidation happened on the surface of top slag and liquid steel. Figure 11 showed the cluster Al 2O 3 formation process. When aluminum alloy added into liquid steel, great supersaturation of [Al] and [O] made Al 2O 3 nucleation formed quickly; after nucleation formed stably and super-saturation decreased until it can t content nucleation condition, the diffusion growth become important. Diffusion process made inclusions size grow and agglomeration formed due to collision of different sizes inclusions. Experimental result showed that, if inner core of inclusion was spherical, outer shell and final morphologies of the inclusions approached to be a sphere in the section; and a polygonal inner core formed a polygonal inclusion. So, initial shapes of cores were extremely important for inclusions final shapes during diffusion growth process. Initial morphologies differences depended on the crystal structure of inclusions nucleation. So, how to control crystal structure of initial inclusion s nucleation related with the final morphologies of inclusions Total Oxygen and Nitrogen Control in Two Modes When free oxygen existed in liquid steel, surface active elements such as oxygen and sulfide would prevent nitrogen picking up during tapping or RH treatment. As shown in Fig. 12, nitrogen in liquid steel had great difference under two modes. Under mode II, nitrogen content was almost half of that in mode I; nitrogen content was mainly below , nitrogen picking up was less due to late final deoxidation. In mode I, there were more sources of picking up nitrogen due to early final de-oxidation. It was beneficial to keep low nitrogen by adopting mode II. As shown in Fig. 13, total oxygen decreased entirely with prolonging of the Al-killed time. Under mode I, total oxygen was controlled below within 25 minutes by RH vacuum treatment. In mode II, due to vacuum carbon deoxidation during RH process, inclusions removal time was short after Al-Killed; the range of total oxygen after RH end fluctuated between and Although initial Fig. 9. Number percent of different types inclusions. Fig. 10. Inclusions size distribution in different modes ISIJ 1658

8 Fig. 11. Cluster Al 2O 3 formation process. Fig. 12. Nitrogen comparison during RH process under two modes. Fig. 13. Total oxygen comparison during RH process under two modes. total oxygen before Al-Killed was low in mode II, the final total oxygen at RH end was higher than that in mode I. So, initial total oxygen and removal time of inclusions were both important for steel cleanness control Characteristic Comparison of Two RH Treatment Modes Based on experimental analysis of previous sections, characteristics of two RH treatment modes were summarized as Table 6. Under mode I, inclusions had sufficient removal time due to early de-oxidation; total oxygen in tundish could be controlled below 20 ppm; cast-ability in continuous casting process could reach 15 heats which was much better than that in mode II; but under mode I, nitrogen picking up and ferro-aluminum consumption were more than that in mode II. So, reasonable RH treatment mode should be chosen based on requirement of production. If low nitrogen and less alloy consumption were important, mode ISIJ

9 Table 6. Characteristic comparison of two RH modes. Modes Strong point Weak point Mode I Mode II 1. Inclusions less and sufficient removal due to early de-oxidation; 2. Low oxidation of top slag; (T.Fe+MnO)=5% 7% 3. Low total oxygen; (<20 ppm in tundish) 4. Good cast-ability (12 15 heats continuous casting) 1. Reduce over-oxidation of liquid steel at BOF endpoint 2. Less alloy consumption due to vacuum carbon de-oxidation; (210 kg grain aluminum per heat) 3. Low nitrogen content due to late final de-oxidation. (N<15 ppm in tundish) 1. Low carbon BOF endpoint result of over-oxidation and refractory life reduce; 2. Nitrogen pick up more; (N=25 30 ppm in tunidsh) 3. Aluminum alloy consumption high. (892 kg ferro-aluminum per heat) 1. High total oxygen due to short inclusion removal time; (T.O=20 25 ppm in tundish) 2. High oxidation of top slag. ((T.Fe+MnO)>10% ) 3. Weak cast-ability (6 10 heats continuous casting) II was suitable. When low total oxygen and good cast-ability were necessary, mode I was better. By adopting mode II to reduce total generation amount of inclusions and nitrogen picking up, also to prolong 20 to 30 minutes calming time before casting to guarantee inclusions removal, low nitrogen, low total oxygen and good cast-ability could realize. 5. Conclusion By comparing steel cleanness of two RH treatment modes, the difference and control conditions were summarized. Under mode I, the removal time of inclusions was sufficient and liquid steel possessed low total oxygen and good cast-ability. Total oxygen could be controlled below within 25 minutes in mode I. Aluminum alloy consumption in mode I was nearly 4 times of that in mode II. Under mode II, a part of free oxygen was removed by vacuum carbon deoxidation, alloy cost was low; nitrogen in liquid steel could be controlled below 15 ppm. By adopting mode II to reduce total generation amount of inclusions and nitrogen picking up, also to prolong 20 to 30 minutes calming time before casting to guarantee inclusions removal, low nitrogen, low total oxygen and good cast-ability could realize. Acknowledgements This work was supported by National Natural Science Foundation of China (No ), Fundamental Research Funds for the Central Universities (FRF-TP A2) and State Key Laboratory of Advanced Metallurgy Foundation in China (KF13-09). The authors wish to express their appreciation to the foundation for providing financial support that guarantees the study successfully to be carried out. REFERENCES 1) K. Kameyama, H. Nishikawa, M. Aratani, R. Asaho, N. Tamura and K. Yamaguchi: Kawasaki Steel Tech. Rep., 26 (1992), 92. 2) M. V. Ende, Y. M. Kim, M. K. Cho, J. Choi and I. H. Jung: Metall. Mater. Trans. B, 42B (2011), ) D. Q. Geng, H. Lei and J. C. He: Metall. Mater. Trans. B, 41B (2010), ) T. Kuwabara, K. Umezawa, K. Mori and H. Watanabe: Trans. Iron Steel Inst. Jpn., 28 (1988), ) Y. Z. Li, N. F. Xue and M. D. Wang: Steelmaking, 16 (2000), 42 (in Chinese). 6) F. J. Hahn, H. P. Haastert and W. Bading: Iron Steel Maker, 17 (1990), No. 3, 43. 7) B. Deo, A. Karamchetty, A. Paul, P. Singh and R. P. Chhabra: ISIJ Int., 36 (1996), ) S. Inoue, Y. Furuno, T. Usui and S. Miyahara: ISIJ Int., 32 (1992), ) K. Yamaguchi, T. Sakuraya and K. Hamagami: Kawasaki Steel Tech. Rep., 32 (1995), ) K. Yamaguchi, Y. Kishimoto, T. Sakuraya, T. Fujii, M. Aratani and H. Nishikawa: ISIJ Int., 32 (1992), ) X. G. Ai, Y. P. Bao, W. Jiang, J. H. Liu, P. H. Li and T. Q. Li: Int. J. Miner. Metall. Mater., 17 (2010), ) M. Wang, Y. P. Bao, H. Cui, W. S. Wu, H. J. Wu, B. Chen and C. X. Ji: J. Univ. Sci. Technol. Beijing, 33 (2011), ) B. S. Liu, G. S. Zhu, H. X. Li, B. H. Li, Y. Cui and A. M. Cui: Int. J. Miner. Metall. Mater., 17 (2010), ISIJ 1660