**Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates**

Kuninori Kitahara1 and Akito Hara2 *1Shimane University 2Tohoku Gakuin University Japan* 

## **1. Introduction**

506 Crystallization – Science and Technology

Matsumura M. and Oh C. H. (1999). Advanced Excimer-Laser Annealing Process for Quasi

Palani I. A., Vasa N. J., and Singaperumal M. (2008). Crystallization and Ablation in

Razak S.N and Bidin N. (2010). Effect of Multipulses Exposure On Excimer Laser Annealed

Sato H. and Nishio S. (2001). Polymer laser photochemistry, ablation, reconstruction, and

Sauerbrey, R. and Petit, G. H. (1989). Theory of Etching Organics Materials by Ultraviolet

Shah, A.; Meier, J.; Vallat-Sauvain, E.; Wyrsch N.; Kroll U.; Droz C.; Graf U. (2003). Material

Shah, Z. A. (2009). Surface Treatment of Polymethylmethacrylate by Excimer Laser Ablation.

Speidell, J. L.; Pulaski D.P.; Patel R.S. (1997). Mask for laser ablation technology: New requirement and challenges, IBM Journal and research development. 8618-8646. Srinivasan, R. and Mayne-Banton, V. (1982). Self Developing Photoetching of

Streetman, B. G. & Banerjee, S. (2000). Solid State Electronic Devices (5th ed.), New Jersey:

Takaoka G. H.; Fujita K.; Ishikawa J., and Takagi T. (1989). Grain Boundary Diffusion

Viatella J. and Singh R. K. (1997). Transient recrystallization of amorphous silicon films.

Watanabe, H.; Miki H.; Sugai, S.; Kawasaki K.; Kioka, T. (1994). Crystallization process of

Yuan, Z.; Lou, Q.; Zhou, J.; Dong, J.; Wei, Y.; Wang, Z.; Zhao, H.g; Wu, G. (2009). Numerical

and solar cell research in microcrystalline silicon", Solar Energy Materials and Solar

Poly(ethyleneterephtalate) Film by Far Ultraviolet Excimer Laser Radiaion.Appl.

Control of Cu Films Prepared By The ICB Technique. Nuclear Instruments and

polycrstalline silicon by KrF2 Excimer laser annealing. Jpn. J. Appl. Phys. 33(8):

and experimental analysis on green laser crystallization of amorphous silicon thin films. Optics & Laser Technology 41 (4): 380. doi: 10.1016 /j.optlastec.2008.09.003 Zhang X and Atrens A. (1992), Rapid solidification characteristics in melt spinning.

Materials Science and Engineering: A. 159(2): 243. doi:10.1016/0921-5093(92)90295.c

Schneider, T. L. (1998). Refractive Surgery and the Excimer Laser. Jacksonville Medicine Serafetinides, A. A.; Makropoulou M. I.; Skordoulis, C. D; and Kar, A. K. (2001). Ultrashort

Pulsed Laser Ablation of Polymers. Appl. Surf. Sci. 180: 42

Cells. 78 (1–4): 469. doi: 10.1016/S0927-0248(02)00448-8

MSc Thesis. Universiti Teknologi Malaysia 2009

Annealing of Amorphous-Si Thin Film On Glass and Crystalline-Si Substrates Irradiated By Third Harmonics of Nd3+:YAG Laser, Materials Science in

Polycrystalline Silicon Thin Film. Proc. of 2010 International Conference on Enabling Science and Nanotechnology. Escienano2010, 1-3 December, 2010, KLCC,

Single-Crystal Silicon Thin-Film Devices. Thin Solid Films. 337: 123

Semiconductor Processing. 11: 107

polymerization, J. Photochem. Photobiol. 2:139

Laser Pulses. Applied Physics Letter, 55(5): 421

Malaysia

Phys. Lett. 41(6): 576

4491

Prentice Hall. ISBN 0130255386

Methods in Physics Research B37/38: 882

Materials Science and Engineering B. 47(1): 78

Silicon (Si) thin films on glass substrates have been extensively developed as a semiconductor material for electronic devices. This material is especially useful for largearea panel devices such as thin-film transistors (TFTs) on active-matrix flat panel displays. The most widely used Si films are hydrogenated amorphous Si (a-Si:H), which can be deposited at temperatures lower than the strain point of the substrate. However, improved electronic properties are required to achieve higher device performance. Using polycrystalline Si (poly-Si) films instead of a-Si:H films enhances carrier mobility by two or three orders of magnitude; thus, driver circuits can be incorporated into display panels, as shown in Fig. 1. The application of poly-Si will be extended to mobile displays with large pixel density, microprocessor–display combined panels, and thin-film solar cells.

Fig. 1. Low-temperature poly-Si liquid crystal display (2 in. diagonal). Driver circuits are integrated at the periphery of the panel.

The poly-Si used for TFT must be ≤ 50 nm thick to ensure the desired device performance. Furthermore, the crystalline fraction should be almost 100%. Such thin films cannot be deposited directly on glass; they must be formed by recrystallization of a-Si precursor films. For this purpose, manufactures have employed solid-phase crystallization (SPC) and

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 509

also performed by zone heating; its advantages include low *T*s (500–700 °C) and consistent

(a) (b)

Fig. 3. Structures of Si layers on non-crystalline layers: (a) Si on insulator, where Si substrate is coated by SiO2 film with windows for seed areas: (b) excimer-laser crystallized Si, where

Control of nucleation during the initial stage of lateral growth and crystalline orientation is essential for the growth of single or quasi-single crystals. To achieve this control, the noncrystalline underlayer is partially opened to reveal the crystalline Si substrate, which acts as the seed area, as shown in Fig. 3(a). Alternatively, the fundamental layers are trenched in order to form a grating-like shape to guide the crystalline orientation; this is known as artificial epitaxy or graphoepitaxy. However, the structures thus created require additional photolithographic processing. In contrast, the laser crystallization process enables control of the crystalline orientation without seeding or photolithography; however, the obtained crystalline quality is not always adequate for the submicron-scale processes required for LSI. For applications in electronic displays, the substrate should be glass or plastic rather than c-Si. On these substrates, Si thin films become essentially polycrystalline because of the seedless growth, as shown in Fig. 3(b). Poly-Si on glass has been applied in devices called giant microelectronics or large-area electronics (Kuriyama et al., 1992; Sameshima, 2009), which have areas several orders of magnitude larger than those of SOI devices. The use of glass or plastic substrates strongly restricts the upper limit of *T*s. Technologies are classified as high-temperature poly-Si or low-temperature poly-Si depending on *T*s (Blake, 1997). The border between them is determined by the strain point of a low alkali glass substrate at ~590

SPC requires a *T*s that is higher than 600 °C. Thus, fused quartz glass is used as the substrate because its strain point is as high as 990 °C. SPC poly-Si is used mainly in 1-in. diagonal or

Laser crystallization technology decreases *T*s to below the strain point of low alkali glass substrates and in many cases to room temperature (RT) owing to selective heating of the Si layer. Progress in laser annealing has been reviewed elsewhere (Sameshima, 2009). Poly-Si films for medium-sized LCD panels suitable for mobile information terminal applications

There are two opposing requirements for laser crystallization: ensuring uniformity in order to integrate TFTs, and eliminating grain boundaries in order to improve the performance of individual TFTs. The former and latter requirements are met by the growth of grains that are notably smaller and larger than the TFT channels, as shown in Fig. 4(b) and 4(c), respectively. Those opposing needs are expected to be reconciled by the growth of flow-

smaller liquid crystal displays (LCDs) for multimedia projectors.

SiO2

Poly-Si

Glass substrate

Melt a-Si

Melt a-Si

surface smoothness.

SiO2

glass substrate is fully coated by SiO2 film.

°C (Corning's technical glass catalog).

shaped grains, as shown in Fig. 4(d).

have been formed.

Si substrate

excimer-laser crystallization (ELC) techniques, as shown in Fig. 2 (Hayashi et al., 1984; Sameshima et al., 1986).

The grain boundaries in poly-Si thin films typically feature random configurations. However, random grain boundaries were reported to severely degrade device performance (Blake et al., 1997). Therefore, grain enlargement is desirable. To achieve this, lateral growth during crystallization is necessary. The lateral growth descrived in this manuscript corresponds to oriented overgrowth of the crystalline film on a non-crystalline layer (i.e., the glass substrate), which is regarded as a type of epitaxy (Givargizov, 1991). Epitaxy proceeds so as to minimize the free energy and is expected to effectively reduce the defect density.

In this work, we investigate the relationship between lateral growth and defects using various characterization techniques. First, we give an overview the crystallization process on non-crystalline substrates and the characterization of defects in poly-Si thin films. Next, we describe previous studies and our work on SPC, ELC, and flow-shaped growth of poly-Si on glass from the perspective of lateral growth. Finally, the growth of a quasisingle-crystal SiGe thin film on a glass substrate, that is super-lateral growth, is demonstrated.

Fig. 2. Techniques for recrystallization of a-Si films on glass substrates using (a) solid-phase crystallization and (b) excimer-laser crystallization.

## **2. Overview of lateral crystallization on non-crystalline layer**

The basic technology for poly-Si formation on a glass substrate is Si on insulator (SOI), which was developed for applications such as low power consumption large-scale integrated circuits (LSIs), three-dimensional LSIs, radiation-hardened electronics and solar cells. SOI technologies have been reviewed in detail elsewhere (Givargizov, 1991; Colinge, 2004). The substrate used for SOI is typically single-crystalline Si (c-Si) coated with a dielectric film such as SiO2 or Si3N4. Although SOI technology includes a wide range of methods, we focus on the recrystallization of a-Si thin films. The principle SOI techniques used for this purpose are zone-melting recrystallization and lateral SPC.

Zone-melting recrystallization is performed by applying a heat source to an a-Si film, which induces melting followed by solidification of the film. Common heat sources are line-shaped electric resistors, electron beams, and continuous-wave (CW) Ar ion lasers. Lateral SPC is

excimer-laser crystallization (ELC) techniques, as shown in Fig. 2 (Hayashi et al., 1984;

The grain boundaries in poly-Si thin films typically feature random configurations. However, random grain boundaries were reported to severely degrade device performance (Blake et al., 1997). Therefore, grain enlargement is desirable. To achieve this, lateral growth during crystallization is necessary. The lateral growth descrived in this manuscript corresponds to oriented overgrowth of the crystalline film on a non-crystalline layer (i.e., the glass substrate), which is regarded as a type of epitaxy (Givargizov, 1991). Epitaxy proceeds so as to minimize the free energy and is expected to effectively reduce

In this work, we investigate the relationship between lateral growth and defects using various characterization techniques. First, we give an overview the crystallization process on non-crystalline substrates and the characterization of defects in poly-Si thin films. Next, we describe previous studies and our work on SPC, ELC, and flow-shaped growth of poly-Si on glass from the perspective of lateral growth. Finally, the growth of a quasisingle-crystal SiGe thin film on a glass substrate, that is super-lateral growth, is

Electric resistance furnace Excimer laser line beam

(a) (b)

**2. Overview of lateral crystallization on non-crystalline layer** 

used for this purpose are zone-melting recrystallization and lateral SPC.

crystallization and (b) excimer-laser crystallization.

Fig. 2. Techniques for recrystallization of a-Si films on glass substrates using (a) solid-phase

The basic technology for poly-Si formation on a glass substrate is Si on insulator (SOI), which was developed for applications such as low power consumption large-scale integrated circuits (LSIs), three-dimensional LSIs, radiation-hardened electronics and solar cells. SOI technologies have been reviewed in detail elsewhere (Givargizov, 1991; Colinge, 2004). The substrate used for SOI is typically single-crystalline Si (c-Si) coated with a dielectric film such as SiO2 or Si3N4. Although SOI technology includes a wide range of methods, we focus on the recrystallization of a-Si thin films. The principle SOI techniques

Zone-melting recrystallization is performed by applying a heat source to an a-Si film, which induces melting followed by solidification of the film. Common heat sources are line-shaped electric resistors, electron beams, and continuous-wave (CW) Ar ion lasers. Lateral SPC is

a-Si poly-Si

Sameshima et al., 1986).

the defect density.

demonstrated.

also performed by zone heating; its advantages include low *T*s (500–700 °C) and consistent surface smoothness.

Fig. 3. Structures of Si layers on non-crystalline layers: (a) Si on insulator, where Si substrate is coated by SiO2 film with windows for seed areas: (b) excimer-laser crystallized Si, where glass substrate is fully coated by SiO2 film.

Control of nucleation during the initial stage of lateral growth and crystalline orientation is essential for the growth of single or quasi-single crystals. To achieve this control, the noncrystalline underlayer is partially opened to reveal the crystalline Si substrate, which acts as the seed area, as shown in Fig. 3(a). Alternatively, the fundamental layers are trenched in order to form a grating-like shape to guide the crystalline orientation; this is known as artificial epitaxy or graphoepitaxy. However, the structures thus created require additional photolithographic processing. In contrast, the laser crystallization process enables control of the crystalline orientation without seeding or photolithography; however, the obtained crystalline quality is not always adequate for the submicron-scale processes required for LSI.

For applications in electronic displays, the substrate should be glass or plastic rather than c-Si. On these substrates, Si thin films become essentially polycrystalline because of the seedless growth, as shown in Fig. 3(b). Poly-Si on glass has been applied in devices called giant microelectronics or large-area electronics (Kuriyama et al., 1992; Sameshima, 2009), which have areas several orders of magnitude larger than those of SOI devices. The use of glass or plastic substrates strongly restricts the upper limit of *T*s. Technologies are classified as high-temperature poly-Si or low-temperature poly-Si depending on *T*s (Blake, 1997). The border between them is determined by the strain point of a low alkali glass substrate at ~590 °C (Corning's technical glass catalog).

SPC requires a *T*s that is higher than 600 °C. Thus, fused quartz glass is used as the substrate because its strain point is as high as 990 °C. SPC poly-Si is used mainly in 1-in. diagonal or smaller liquid crystal displays (LCDs) for multimedia projectors.

Laser crystallization technology decreases *T*s to below the strain point of low alkali glass substrates and in many cases to room temperature (RT) owing to selective heating of the Si layer. Progress in laser annealing has been reviewed elsewhere (Sameshima, 2009). Poly-Si films for medium-sized LCD panels suitable for mobile information terminal applications have been formed.

There are two opposing requirements for laser crystallization: ensuring uniformity in order to integrate TFTs, and eliminating grain boundaries in order to improve the performance of individual TFTs. The former and latter requirements are met by the growth of grains that are notably smaller and larger than the TFT channels, as shown in Fig. 4(b) and 4(c), respectively. Those opposing needs are expected to be reconciled by the growth of flowshaped grains, as shown in Fig. 4(d).

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 511

poly-Si (described below) were calibrated according to the spectral resolution of the setup. The peak frequency of the spectra was exactly determined by fitting to a Lorentzian curve. Typical first-order Raman scattering for c-Si, Si on sapphire (SOS), and poly-Si on glass are shown in Fig. 5. In the crystalline phase, the Raman spectrum exhibits an isolated band corresponding to the degeneration of a single transverse-optical (TO) phonon mode and two longitudinal-optical (LO) phonon modes; the c-Si band exhibits a linewidth as narrow as 2.5–3.5 cm-1 at RT (Temple & Hathaway, 1973; Menendez & Cardona, 1984) owing to the *q* = 0 (*q*: wave vector) selection rule. In the amorphous phase, four continuous phonon bands are observed; they are broad because the *q* = 0 selection rule is loosened as a result of a decrease

The features of the optical phonon mode (OPM) of the crystalline phase are represented by

are smaller than a few tens of nanometers, the spatial correlation of the phonon decreases,

et al., 1981). Compressive and tensile stress in the films cause the peak to shift to higher and

(Englet, 1980). Therefore, the stress can be quantitatively estimated only when the impact of the other factors is negligible. Note, however, that the intensity of the stress depends on the crystallization technique and conditions used. Therefore, an analysis of individual crystallization techniques should be developed by comparing the Raman spectroscopy

), FWHM, intensity, and in some cases asymmetry. When the grains

) and increases the FWHM (Richter

is proportional to the magnitude of the stress

∙ Electron spin resonance ∙ Micro-Raman spectroscopy ∙ Spectral ellipsometry ∙ Secco etching and SEM

∙ X-ray diffraction

∙ Atomic force microscopy

∙ Electron backscattering diffraction ∙ Transmission electron microscopy

∙ Test device of thin film transistor

in the phonon correlation length.

lower frequencies, respectively. Moreover,

results with those obtained by SEM and TEM.

which causes the peaks to shift to a lower frequency (

Defect density ∙ Stress in film ∙ Crystal fraction ∙

Size and shape of grains ∙ Crystal orientation ∙

Surface morphology ∙ Electric characteristics ∙

Table 1. Principal terms for characterization of poly-Si thin films on glass substrates.

Defects in poly-Si are frequently observed in SEM images. Grain boundaries and some defect clusters are clearly revealed by Secco etching (Secco d'Aragona, 1972). These defects are evidently electrochemically active, which is confirmed by the fact that they disappear upon hydrogenation before etching. On the other hand, inactive defects such as twin boundaries are observed only as shallow contrasts and are independent of hydrogenation

To summarize, various electron microscopies provide direct knowledge of defects and the geometric configuration of grains in poly-Si/glass systems. However, defects cannot always be detected by those methods. Raman microscopy is useful for easy macroscopic (submicron-scale) defect analysis and quantitative analysis of stress in films. However, semiempirical analysis aided by other techniques is required to simultaneously characterize the

Geometry of grain boundaries ∙

defect density and stress in films by Raman spectroscopy.

the peak frequency (

(Kitahara et al., 2009a).

Fig. 4. Structures of thin-film transistors (TFTs) on glass substrates: (a) test device for characterizing electrical performance; geometry of grains and TFT channel for (b) grains notably smaller than the channel region, (c) a large grain containing the channel region, and (d) flow-shaped grains running parallel to the drain current.

### **3. Characterization of poly-Si films on glass**

The principal terms for the characterization of poly-Si thin films on glass substrates are summarized in Table 1. The dangling bonds in a-Si and poly-Si films were qualitatively analyzed via the spin density of unpaired electrons using electron-spin resonance. The spin density was found to be reduced by hydrogenation (Nickel et al., 1997; Spinella et al., 1998), which causes termination of the dangling bonds with H atoms. The geometry and crystal orientation of the grains were investigated using images and diffraction patterns obtained by scanning electron microscopy (SEM), transmission electron microscopy (TEM), and electron backscattering diffraction (EBSD). When growth accompanies melting, the surface morphology can vary with the displacement of the melt. The surface morphology is observed using atomic force microscopy (AFM) or cross-sectional TEM. The electronic characteristics were estimated by fabricating test n-channel and p-channel TFT devices, as shown in Fig. 4(a). The devices exhibit characteristics such as the field-effect electron and hole mobilities (FEn and FEp, respectively), threshold voltage, subthreshold voltage, and off-current.

The characterization techniques described above require the specimen be processed. Moreover, they are not simple enough to be used to characterize large numbers of films. In contrast, optical techniques are suitable for simple, non-destructive characterization. We analyzed the defects and stress in a poly-Si/glass system by Raman scattering spectroscopy (Kitahara et al., 2002, 2003, 2011a). Micro-Raman spectroscopy was conducted by using a Renishaw System 1000 in the backscattering geometry under excitation with a 514.5 nm Ar ion laser. Details of the setup and band component analysis were described elsewhere (Frost & Shurvell, 1997). The full width at half maximum (FWHM) values of spectra obtained for

(a) (b)

(c) (d)

The principal terms for the characterization of poly-Si thin films on glass substrates are summarized in Table 1. The dangling bonds in a-Si and poly-Si films were qualitatively analyzed via the spin density of unpaired electrons using electron-spin resonance. The spin density was found to be reduced by hydrogenation (Nickel et al., 1997; Spinella et al., 1998), which causes termination of the dangling bonds with H atoms. The geometry and crystal orientation of the grains were investigated using images and diffraction patterns obtained by scanning electron microscopy (SEM), transmission electron microscopy (TEM), and electron backscattering diffraction (EBSD). When growth accompanies melting, the surface morphology can vary with the displacement of the melt. The surface morphology is observed using atomic force microscopy (AFM) or cross-sectional TEM. The electronic characteristics were estimated by fabricating test n-channel and p-channel TFT devices, as shown in Fig. 4(a). The devices exhibit characteristics such as the field-effect electron and

The characterization techniques described above require the specimen be processed. Moreover, they are not simple enough to be used to characterize large numbers of films. In contrast, optical techniques are suitable for simple, non-destructive characterization. We analyzed the defects and stress in a poly-Si/glass system by Raman scattering spectroscopy (Kitahara et al., 2002, 2003, 2011a). Micro-Raman spectroscopy was conducted by using a Renishaw System 1000 in the backscattering geometry under excitation with a 514.5 nm Ar ion laser. Details of the setup and band component analysis were described elsewhere (Frost & Shurvell, 1997). The full width at half maximum (FWHM) values of spectra obtained for

Fig. 4. Structures of thin-film transistors (TFTs) on glass substrates: (a) test device for characterizing electrical performance; geometry of grains and TFT channel for (b) grains notably smaller than the channel region, (c) a large grain containing the channel region, and

G

FEp, respectively), threshold voltage, subthreshold voltage, and

S D

G

G

S D

G

(d) flow-shaped grains running parallel to the drain current.

**3. Characterization of poly-Si films on glass** 

hole mobilities (

off-current.

FEn and

S D

Source Drain (S) (D)

Channel

Gate (G)

poly-Si (described below) were calibrated according to the spectral resolution of the setup. The peak frequency of the spectra was exactly determined by fitting to a Lorentzian curve.

Typical first-order Raman scattering for c-Si, Si on sapphire (SOS), and poly-Si on glass are shown in Fig. 5. In the crystalline phase, the Raman spectrum exhibits an isolated band corresponding to the degeneration of a single transverse-optical (TO) phonon mode and two longitudinal-optical (LO) phonon modes; the c-Si band exhibits a linewidth as narrow as 2.5–3.5 cm-1 at RT (Temple & Hathaway, 1973; Menendez & Cardona, 1984) owing to the *q* = 0 (*q*: wave vector) selection rule. In the amorphous phase, four continuous phonon bands are observed; they are broad because the *q* = 0 selection rule is loosened as a result of a decrease in the phonon correlation length.

The features of the optical phonon mode (OPM) of the crystalline phase are represented by the peak frequency (), FWHM, intensity, and in some cases asymmetry. When the grains are smaller than a few tens of nanometers, the spatial correlation of the phonon decreases, which causes the peaks to shift to a lower frequency () and increases the FWHM (Richter et al., 1981). Compressive and tensile stress in the films cause the peak to shift to higher and lower frequencies, respectively. Moreover, is proportional to the magnitude of the stress (Englet, 1980). Therefore, the stress can be quantitatively estimated only when the impact of the other factors is negligible. Note, however, that the intensity of the stress depends on the crystallization technique and conditions used. Therefore, an analysis of individual crystallization techniques should be developed by comparing the Raman spectroscopy results with those obtained by SEM and TEM.

Table 1. Principal terms for characterization of poly-Si thin films on glass substrates.

Defects in poly-Si are frequently observed in SEM images. Grain boundaries and some defect clusters are clearly revealed by Secco etching (Secco d'Aragona, 1972). These defects are evidently electrochemically active, which is confirmed by the fact that they disappear upon hydrogenation before etching. On the other hand, inactive defects such as twin boundaries are observed only as shallow contrasts and are independent of hydrogenation (Kitahara et al., 2009a).

To summarize, various electron microscopies provide direct knowledge of defects and the geometric configuration of grains in poly-Si/glass systems. However, defects cannot always be detected by those methods. Raman microscopy is useful for easy macroscopic (submicron-scale) defect analysis and quantitative analysis of stress in films. However, semiempirical analysis aided by other techniques is required to simultaneously characterize the defect density and stress in films by Raman spectroscopy.

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 513

Fig. 6. Scanning electron microscopy images for solid-phase crystallized poly-Si heated at 900 °C for 1.5 h. Images were taken (a) before and after Secco etching for (b) 20 s and (c) 60 s.

Fig. 7. Transmission electron microscopy image and diffraction pattern for solid-phase

We investigated the microscopic geometry of the grains and defects by SEM and TEM. We then studied the phase variation at the initial stage macroscopically using Raman spectroscopy and spectroscopic ellipsometry. SPC was performed in a N2 flow at atmospheric pressure. Figure 6 shows SEM images before and after Secco etching after SPC at *T*c = 900 °C for 1.5 h. The surface of the as-crystallized film is relatively smooth. Grain boundaries were not clearly revealed by the short etching time (20 s) owing to the fine structures at the grain boundaries and the overlapping of grains (Spinella et al., 1998). The longer etching time (60 s) revealed feather-like grains lying parallel to the substrate. In addition, major twin boundaries appeared in the individual grains as shallow lines. The plan-view TEM image shown in Fig. 7 indicates that the twin boundaries extend along the major axis. Furthermore, a large density of micro-twins is distributed in the grains. The subspots and streaks evident in the diffraction pattern also indicate the presence of twins

Figure 8 shows the OPMs in the Raman spectra during the initial stage of SPC at *T*c = 580 °C. Only the amorphous phase is detected at 1 h. The crystal component begins to appear in the spectrum after a latent time of ~3 h. Figure 9 shows the crystal fraction in the 580 °C SPC film as a function of the heating time. In Raman spectroscopy, the volume fraction of the crystal was estimated by the area–intensity ratio of the amorphous (*I*a) and crystal (*I*c) components, i.e., *I*c/(*I*c + *I*a), which yields a relative value. In ellipsometry, the volume fraction of the crystal was analyzed using the Tauc–Lorentz model (Jellison, 1998), which yields an absolute value. Although no crystal component was found by Raman spectroscopy until 3 h had passed, the volume fraction estimated by ellipsometry began to increase at 1 h. The spectra of the imaginary dielectric function observed by ellipsometry show that the E1 and E2 bands were somewhat visible even in the latent time. This suggests that atomic

(b)

(a) (c)

(b) (a)

1 m

crystallized film heated at 900 °C for 3 h.

and {111} microfacets in the grains.

1 m

1 m

Fig. 5. Typical optical-phonon mode observed in the Raman scattering spectra for Si on sapphire (SOS), single crystalline Si (c-Si) bulk containing no stress, and poly-Si thin film on glass. Intensity is normalized to the individual peak values. Peaks are shifted to higher and lower frequencies by compressive and tensile stresses, respectively.

## **4. Solid-phase crystallization**

The heat source for SPC typically consists of an electric resistance furnace and an infrared lamp annealing furnace. The force that drives crystallization of a-Si in the metastable state is the lowering of the Gibbs free energy through the crystalline phase change. SPC poly-Si technologies are reviewed in detail elsewhere (Hatalis & Greve, 1988; Spinella et al., 1998).

The typical grain size of solid-phase crystallized poly-Si for TFTs is a few hundred nanometers, which is at least several times larger than the film thickness. This suggests that SPC on non-crystalline substrates is dominated by lateral growth rather than columnar growth which is frequently seen in chemical vapor deposition.

SPC is initiated by homogeneous nucleation after incubation. In pure Si, the activation energy of nucleation is larger than that of growth by 0.25 eV (Spinella et al., 1998). Therefore, growth proceeds immediately after nucleation. The lower limit of the nucleation temperature depends on the quality of the precursor film and is ~560 °C for highly pure a-Si. The grain size decreases with increasing crystallization temperature (*T*c), which is associated with increasing nucleation frequency. Lateral grain growth is first interrupted by collisions among grains. Then, the grain size increases due to secondary grain growth; the secondary growth rate increases with the grain boundary energy, surface energy anisotropy, grain boundary mobility, and *T*c (Thompson, 1985). Furthermore, twin formation at the early stage of crystallization preferentially accelerates the lateral growth rate in the <112> direction, producing grains with an ellipsoidal outline (Nakamura et al., 1989). Growth along the minor axis of the ellipsoid, i.e., in the <111> direction, proceeds with simultaneous formation of many micro-twins (Drosed & Washburn, 1980). Although solid-phase crystallized poly-Si films have a considerable density of defects, hydrogenation of TFTs effectively inactivates those defects, producing a FEn of 24 cm2/Vs (Little et al., 2001).

Poly-Si/glass C-Si SOS

500 510 520 530 540 Frequency (cm -1)

The heat source for SPC typically consists of an electric resistance furnace and an infrared lamp annealing furnace. The force that drives crystallization of a-Si in the metastable state is the lowering of the Gibbs free energy through the crystalline phase change. SPC poly-Si technologies are reviewed in detail elsewhere (Hatalis & Greve, 1988; Spinella et al., 1998). The typical grain size of solid-phase crystallized poly-Si for TFTs is a few hundred nanometers, which is at least several times larger than the film thickness. This suggests that SPC on non-crystalline substrates is dominated by lateral growth rather than columnar

SPC is initiated by homogeneous nucleation after incubation. In pure Si, the activation energy of nucleation is larger than that of growth by 0.25 eV (Spinella et al., 1998). Therefore, growth proceeds immediately after nucleation. The lower limit of the nucleation temperature depends on the quality of the precursor film and is ~560 °C for highly pure a-Si. The grain size decreases with increasing crystallization temperature (*T*c), which is associated with increasing nucleation frequency. Lateral grain growth is first interrupted by collisions among grains. Then, the grain size increases due to secondary grain growth; the secondary growth rate increases with the grain boundary energy, surface energy anisotropy, grain boundary mobility, and *T*c (Thompson, 1985). Furthermore, twin formation at the early stage of crystallization preferentially accelerates the lateral growth rate in the <112> direction, producing grains with an ellipsoidal outline (Nakamura et al., 1989). Growth along the minor axis of the ellipsoid, i.e., in the <111> direction, proceeds with simultaneous formation of many micro-twins (Drosed & Washburn, 1980). Although solid-phase crystallized poly-Si films have a considerable density of defects, hydrogenation of TFTs

FEn of 24 cm2/Vs (Little et al., 2001).

Fig. 5. Typical optical-phonon mode observed in the Raman scattering spectra for Si on sapphire (SOS), single crystalline Si (c-Si) bulk containing no stress, and poly-Si thin film on glass. Intensity is normalized to the individual peak values. Peaks are shifted to higher and

N orm

**4. Solid-phase crystallization** 

alized intensity (arb.

Normalized intensity (arb. units)

FWHM

lower frequencies by compressive and tensile stresses, respectively.

growth which is frequently seen in chemical vapor deposition.

effectively inactivates those defects, producing a

Fig. 6. Scanning electron microscopy images for solid-phase crystallized poly-Si heated at 900 °C for 1.5 h. Images were taken (a) before and after Secco etching for (b) 20 s and (c) 60 s.

Fig. 7. Transmission electron microscopy image and diffraction pattern for solid-phase crystallized film heated at 900 °C for 3 h.

We investigated the microscopic geometry of the grains and defects by SEM and TEM. We then studied the phase variation at the initial stage macroscopically using Raman spectroscopy and spectroscopic ellipsometry. SPC was performed in a N2 flow at atmospheric pressure. Figure 6 shows SEM images before and after Secco etching after SPC at *T*c = 900 °C for 1.5 h. The surface of the as-crystallized film is relatively smooth. Grain boundaries were not clearly revealed by the short etching time (20 s) owing to the fine structures at the grain boundaries and the overlapping of grains (Spinella et al., 1998). The longer etching time (60 s) revealed feather-like grains lying parallel to the substrate. In addition, major twin boundaries appeared in the individual grains as shallow lines. The plan-view TEM image shown in Fig. 7 indicates that the twin boundaries extend along the major axis. Furthermore, a large density of micro-twins is distributed in the grains. The subspots and streaks evident in the diffraction pattern also indicate the presence of twins and {111} microfacets in the grains.

Figure 8 shows the OPMs in the Raman spectra during the initial stage of SPC at *T*c = 580 °C. Only the amorphous phase is detected at 1 h. The crystal component begins to appear in the spectrum after a latent time of ~3 h. Figure 9 shows the crystal fraction in the 580 °C SPC film as a function of the heating time. In Raman spectroscopy, the volume fraction of the crystal was estimated by the area–intensity ratio of the amorphous (*I*a) and crystal (*I*c) components, i.e., *I*c/(*I*c + *I*a), which yields a relative value. In ellipsometry, the volume fraction of the crystal was analyzed using the Tauc–Lorentz model (Jellison, 1998), which yields an absolute value. Although no crystal component was found by Raman spectroscopy until 3 h had passed, the volume fraction estimated by ellipsometry began to increase at 1 h. The spectra of the imaginary dielectric function observed by ellipsometry show that the E1 and E2 bands were somewhat visible even in the latent time. This suggests that atomic

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 515

Figure 10 shows the relationship between the FWHM of the OPM and *T*c in the range of 625– 900 °C with annealing times of 3–8 h; the times were varied in order to achieve sufficient crystallization. The FWHM evidently decreases with increasing *T*c. Thus, the defect density decreases with increasing *T*c, which is a universal tendency for crystal growth under essentially stable thermal equilibrium. The plots for metal-induced lateral crystallization

> ◇ MILC (168 h) ○ MILC (8 h)

700 600 500 (oC)

0.8 1.0 1.2 1.4

Fig. 10. Relationship between the full width at half maximum (FWHM) of the optical phonon mode and the crystallization temperature (*T*c). Plots for poly-Si films fabricated by metal-induced crystallization (MILC) and solid-phase crystallization (SPC) are shown. Note

To summarize, Si atoms cause reconstruction toward ordering in the amorphous phase in the initial stage of annealing. SPC begins at homogeneous nucleations after the incubation time. Growth begins immediately after nucleation. The grain size increases with acceleration of lateral growth by the major twin boundaries formed at the early stage and the coalescence of grains by secondary grain growth. Growth proceeds with simultaneous formation of many micro-twins, resulting in many defects. Increasing *T*c decreases the defect density in

The nucleation temperature of a-Si is decreased by the addition of a catalytic metal; the process is referred to as MILC and catalyst-assisted SPC (Kawatsu et al., 1990; Cammarata & Thompson, 1990; Takayama et al., 2000). Although metals such as Al, Mo, Ni, Pd, and Ti have been applied to cause MILC, Ni is the most frequently used catalyst. Ni can be supplied to the seed area of an a-Si film by evaporation, sputtering, or application of an acetate solution. Annealing of the film causes lateral diffusion of Ni atoms into the amorphous area before the silicide reaction occurs. NiSi2 begins to segregate at *T*c > 420 °C. Although NiSi2 has the fluorite crystal structure, which is not the same as that of Si, the

1000/T (K-1)

(MILC) will be described in the next section.

FWHM (cm-1)

grains as the grain size is reduced.

**5. Metal-induced lateral crystallization** 

5.0

that the dependence on *T*c of MILC is opposite to that of SPC.

5.5

6.0

6.5

7.0

7.5

8.0

900

● SPC

reconstruction toward ordering before the appearance of the crystal was detected by ellipsometry.

Fig. 8. Optical phonon modes in the Raman spectra for the initial stage of solid-phase crystallized film heated at 580 °C for (a) 1 h and (b) 3 h. Thick lines and thin lines are experimental values and fitting curves, respectively.

Fig. 9. Relationship between crystal fraction and heating time for solid-phase crystallized film heated at 580 °C. Crystal fraction was estimated using spectral ellipsometry and Raman spectroscopy.

reconstruction toward ordering before the appearance of the crystal was detected by

LO TO

350 400 450 500 550 Frequency (cm-1)

> 0123456 Heating time (h)

Raman (Relative value)

Fig. 8. Optical phonon modes in the Raman spectra for the initial stage of solid-phase crystallized film heated at 580 °C for (a) 1 h and (b) 3 h. Thick lines and thin lines are

Ellipsometry

Fig. 9. Relationship between crystal fraction and heating time for solid-phase crystallized film heated at 580 °C. Crystal fraction was estimated using spectral ellipsometry and Raman

ellipsometry.

Intensity (arb. units)

experimental values and fitting curves, respectively.

0

0

20

0.2

40

0.4

Volume fraction of crystal

spectroscopy.

60

0.6

80

0.8

(a)

(b)

Figure 10 shows the relationship between the FWHM of the OPM and *T*c in the range of 625– 900 °C with annealing times of 3–8 h; the times were varied in order to achieve sufficient crystallization. The FWHM evidently decreases with increasing *T*c. Thus, the defect density decreases with increasing *T*c, which is a universal tendency for crystal growth under essentially stable thermal equilibrium. The plots for metal-induced lateral crystallization (MILC) will be described in the next section.

Fig. 10. Relationship between the full width at half maximum (FWHM) of the optical phonon mode and the crystallization temperature (*T*c). Plots for poly-Si films fabricated by metal-induced crystallization (MILC) and solid-phase crystallization (SPC) are shown. Note that the dependence on *T*c of MILC is opposite to that of SPC.

To summarize, Si atoms cause reconstruction toward ordering in the amorphous phase in the initial stage of annealing. SPC begins at homogeneous nucleations after the incubation time. Growth begins immediately after nucleation. The grain size increases with acceleration of lateral growth by the major twin boundaries formed at the early stage and the coalescence of grains by secondary grain growth. Growth proceeds with simultaneous formation of many micro-twins, resulting in many defects. Increasing *T*c decreases the defect density in grains as the grain size is reduced.

#### **5. Metal-induced lateral crystallization**

The nucleation temperature of a-Si is decreased by the addition of a catalytic metal; the process is referred to as MILC and catalyst-assisted SPC (Kawatsu et al., 1990; Cammarata & Thompson, 1990; Takayama et al., 2000). Although metals such as Al, Mo, Ni, Pd, and Ti have been applied to cause MILC, Ni is the most frequently used catalyst. Ni can be supplied to the seed area of an a-Si film by evaporation, sputtering, or application of an acetate solution. Annealing of the film causes lateral diffusion of Ni atoms into the amorphous area before the silicide reaction occurs. NiSi2 begins to segregate at *T*c > 420 °C. Although NiSi2 has the fluorite crystal structure, which is not the same as that of Si, the

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 517

T

c=500 oC

Intensity (arb.

region, and (c) uncrystallized region.

density despite the increase in *T*c.

**6. Excimer-laser crystallization** 

(b)

(a)

(c)

during MILC requires both a Ni supply and the lateral growth process.

growth length is smaller than the calculated Ni diffusion length in a-Si.

300 400 500 600 Frequency (cm -1)

Typical Raman spectra for a Ni-evaporated source region, lateral growth region, and uncrystallized region are shown in Figs. 12(a)–(c). Here, MILC was conducted at 500 °C for 8 h. The spectrum for the Ni-evaporated region indicates the presence of a large amorphous fraction. In this area, the Ni density is likely too high to effectively induce lateral crystallization. In contrast, in the majority of the lateral growth region, the OPM is accompanied only by a weak amorphous-like mode. The uncrystallized region naturally remains amorphous. These results suggest that the amorphous-to-polycrystalline transition

The relationship between *T*c and the FWHM of the OPM for the lateral growth region is shown in Fig. 10. The FWHM decreases with decreasing *T*c and reaches 5.7 cm-1 at 450 °C, which is nearly equal to the FWHM of SPC at 900 °C, 5.6 cm-1. A factor other than thermal energy is responsible for the *T*c dependence of MILC because the dependence on *T*c is opposite to that of SPC. In MILC, the Ni atoms exhibit a lateral gradation of density in a-Si due to diffusion from the source region. In particular, at *T*c < 500 °C, the observed lateral

We concluded that the lateral growth at relatively low *T*c is directed by the gradation in Ni density. Furthermore, the growth at *T*c as low as 450 °C is dominated by needle-like lateral epitaxy and results in a low defect density. Increasing *T*c restricts the length of straight growth and also enhances random growth due to SPC; these two effects increase the defect

Crystallization is performed primarily by KrF or XeCl excimer lasers, which supply intense pulsed light with durations of ~30 ns and wavelengths of 249 and 308 nm, respectively. The

Fig. 12. Typical Raman spectra for (a) Ni-evaporated source region, (b) lateral growth

mismatch of their lattice constants can be as small as 0.4%. Thus, NiSi2 operates as an excellent nucleus for Si growth. NiSi2 seeds then migrate laterally, leaving needle-like Si crystals in the a-Si. The activation energies reported for the growth rate of Ni-MILC, which range from 1.3– 2.3 eV (Makihara et al., 2003; Kitahara et al., 2009b), are smaller than those for SPC (3.4 eV) and solid-phase epitaxy (2.7 eV) (Roth et al., 1990; Spinella et al., 1998). Lower values of *T*c lead to straighter crystallization. Annealing at 450 °C reportedly results in crystals 10 μm in length and 160 nm wide (Makihara et al., 2003). Needle-like crystals coalesce and form larger grain when the remaining a-Si in the interstitial area is crystallized via SPC.

The application of MILC to devices is not easy because of the long processing time required and the existence of silicide in the crystallized region. However, TFTs have been fabricated by combining MILC with ELC. In this method, the silicide was gettered to ion-implanted regions corresponding to the source and drain electrodes and achieved a *μ*FEn value of 320 cm2/Vs (Mizuki et al., 2004).

The density and geometry of the residual defects in MILC films are expected to differ from those in SPC films because MILC is more dominated by lateral growth than SPC is. We investigated the difference in defects between SPC and MILC (Kitahara et al., 2009b). MILC was performed by evaporating Ni onto an a-Si/SiO2/fused quartz glass substrate through a metal mask with rectangular windows. The thicknesses of Ni and a-Si were 13 and 110 nm, respectively. The plan-view TEM image after MILC at 450 °C for 168 h is shown in Fig. 11(a). The image exhibits needle-like crystals with a {110} plane extending toward <111>. Note that the grains exhibited little contrast in the strain corresponding to defects. The coexistence of sharp diffraction spots and a halo in Fig. 11(a') indicate that the grain is a single crystal with a surrounding amorphous component. The TEM image after MILC at 500 °C for 168 h [Fig. 11(b)] shows accumulated needle-shaped grains with strain contrast due to defects. Small subspots in the diffraction pattern shown in Fig. 11(b') also indicate the presence of defects in the grains. However, the streaks corresponding to microfacets that were seen for the SPC films are not observed here. These results suggest that the number of defects in the grains decreases with decreasing *T*c, which is opposite to the trend for SPC.

Fig. 11. Transmission electron microscopy images for (a) metal-induced crystallization (MILC) at 450 °C for 168 h and (b) MILC at 500 °C for 168 h. Individual diffraction patterns are also shown in (a') and (b').

Fig. 12. Typical Raman spectra for (a) Ni-evaporated source region, (b) lateral growth region, and (c) uncrystallized region.

Typical Raman spectra for a Ni-evaporated source region, lateral growth region, and uncrystallized region are shown in Figs. 12(a)–(c). Here, MILC was conducted at 500 °C for 8 h. The spectrum for the Ni-evaporated region indicates the presence of a large amorphous fraction. In this area, the Ni density is likely too high to effectively induce lateral crystallization. In contrast, in the majority of the lateral growth region, the OPM is accompanied only by a weak amorphous-like mode. The uncrystallized region naturally remains amorphous. These results suggest that the amorphous-to-polycrystalline transition during MILC requires both a Ni supply and the lateral growth process.

The relationship between *T*c and the FWHM of the OPM for the lateral growth region is shown in Fig. 10. The FWHM decreases with decreasing *T*c and reaches 5.7 cm-1 at 450 °C, which is nearly equal to the FWHM of SPC at 900 °C, 5.6 cm-1. A factor other than thermal energy is responsible for the *T*c dependence of MILC because the dependence on *T*c is opposite to that of SPC. In MILC, the Ni atoms exhibit a lateral gradation of density in a-Si due to diffusion from the source region. In particular, at *T*c < 500 °C, the observed lateral growth length is smaller than the calculated Ni diffusion length in a-Si.

We concluded that the lateral growth at relatively low *T*c is directed by the gradation in Ni density. Furthermore, the growth at *T*c as low as 450 °C is dominated by needle-like lateral epitaxy and results in a low defect density. Increasing *T*c restricts the length of straight growth and also enhances random growth due to SPC; these two effects increase the defect density despite the increase in *T*c.

#### **6. Excimer-laser crystallization**

516 Crystallization – Science and Technology

mismatch of their lattice constants can be as small as 0.4%. Thus, NiSi2 operates as an excellent nucleus for Si growth. NiSi2 seeds then migrate laterally, leaving needle-like Si crystals in the a-Si. The activation energies reported for the growth rate of Ni-MILC, which range from 1.3– 2.3 eV (Makihara et al., 2003; Kitahara et al., 2009b), are smaller than those for SPC (3.4 eV) and solid-phase epitaxy (2.7 eV) (Roth et al., 1990; Spinella et al., 1998). Lower values of *T*c lead to straighter crystallization. Annealing at 450 °C reportedly results in crystals 10 μm in length and 160 nm wide (Makihara et al., 2003). Needle-like crystals coalesce and form larger grain

The application of MILC to devices is not easy because of the long processing time required and the existence of silicide in the crystallized region. However, TFTs have been fabricated by combining MILC with ELC. In this method, the silicide was gettered to ion-implanted regions corresponding to the source and drain electrodes and achieved a *μ*FEn value of 320

The density and geometry of the residual defects in MILC films are expected to differ from those in SPC films because MILC is more dominated by lateral growth than SPC is. We investigated the difference in defects between SPC and MILC (Kitahara et al., 2009b). MILC was performed by evaporating Ni onto an a-Si/SiO2/fused quartz glass substrate through a metal mask with rectangular windows. The thicknesses of Ni and a-Si were 13 and 110 nm, respectively. The plan-view TEM image after MILC at 450 °C for 168 h is shown in Fig. 11(a). The image exhibits needle-like crystals with a {110} plane extending toward <111>. Note that the grains exhibited little contrast in the strain corresponding to defects. The coexistence of sharp diffraction spots and a halo in Fig. 11(a') indicate that the grain is a single crystal with a surrounding amorphous component. The TEM image after MILC at 500 °C for 168 h [Fig. 11(b)] shows accumulated needle-shaped grains with strain contrast due to defects. Small subspots in the diffraction pattern shown in Fig. 11(b') also indicate the presence of defects in the grains. However, the streaks corresponding to microfacets that were seen for the SPC films are not observed here. These results suggest that the number of defects in the

when the remaining a-Si in the interstitial area is crystallized via SPC.

grains decreases with decreasing *T*c, which is opposite to the trend for SPC.

(a) (b)

(a') (b')

Fig. 11. Transmission electron microscopy images for (a) metal-induced crystallization (MILC) at 450 °C for 168 h and (b) MILC at 500 °C for 168 h. Individual diffraction patterns

cm2/Vs (Mizuki et al., 2004).

are also shown in (a') and (b').

Crystallization is performed primarily by KrF or XeCl excimer lasers, which supply intense pulsed light with durations of ~30 ns and wavelengths of 249 and 308 nm, respectively. The

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 519

Fig. 13. Scanning electron microscopy images for (a, b) non-hydrogenated and (c) hydrogenated excimer-laser annealed poly-Si films. Secco etching was performed for (a) 15 s and (b, c) 40 s.

Hydrogenation interfered with the appearance of grain boundaries after etching.

100 nm

Fig. 14. Typical transmission electron microscopy image and diffraction pattern for a grain of excimer-laser crystallized poly-Si; diffraction pattern is that in the circled region..

(b)

001 101

Fig. 15. (a) Crystalline orientation map in terms of the normal direction, (b) grain boundary map, and (c) inverse pole diagrams of the excimer-laser crystallized poly-Si, as observed by electron backscattering diffraction. Surface exhibits weak orientation to the {001} plane.

(c)

1 m 1 m

1 m

(c)

111

Frequency

2-5 o 5-180 o 3 CSL

(a)

(b)

001

Twin boundaries appear for some grains.

(a)

111

1 m 1 m

101

luminous flux is typically shaped as a linear beam. Overlapping irradiation by the scanning pulsed light enables uniform crystallization over a wide area. The laser energy density is adjusted to be slightly lower than that at which complete melting of the Si film occurs. In practical use, *T*s is maintained at around RT during crystallization; this forms grains a few hundred nanometers in size with high reproducibility.

Because of the short pulse duration, *T*s does not almost increase. However, rapid cooling after laser irradiation is an obstacle to lateral growth. Some techniques, including doublebeam irradiation, have been proposed to decrease the cooling rate. Applying multiple laser irradiations at *T*s < 400 °C reportedly increases the grain size to 4.5 μm and strongly aligns the surface orientation to the <111> direction (Kurimiya et al., 1993).

One of the features of ELC is a rough surface morphology containing a large density of hillocks. The hillocks are generated by a positive feedback effect between the beam-induced periodic surface roughness pattern and the interference in subsequent pulses (McCulloch & Brotherton, 1995). Another feature of ELC is that *μ*FEn is considerably higher than that of SPC. The *μ*FEn value of an excimer-laser crystallized poly-Si TFT increases with increasing grain size and reaches 320 cm2/Vs at an average grain size of 700 nm, at which the dominant factor determining *μ*FEn varies from grain boundary scattering to lattice scattering (Hara et al., 2002a). Furthermore, a *μ*FEn value of 914 cm2/Vs was reportedly obtained by positioncontrolled large grain growth (Mitani et al., 2008).

The large *μ*FEn value of excimer-laser crystallized poly-Si has been attributed to a low defect density in grains owing to a regrowth procedure through the liquid phase. In practice, TEM images typically exhibit few defects in the grains. The average dislocation density was reportedly 8 × 106 cm-2 for large (2–4 m) grains and 108–109 cm-2 even for defective grains (Christiansen et al., 2001). However, the cooling velocity of ELC can be as high as ~1010 K/s (Sameshima & Usui, 1993). Therefore, recrystallization is expected to proceed under significant deviations from thermal equilibrium; consequently, large numbers of defects are frozen in the grains. In defects in bulk c-Si, impurity complexes generated at high temperature were shown to be frozen by quenching at a rate of ~103 K/s (Takahashi et al., 1995). The cooling rate of ELC is far larger than that in the quenching experiment. Hence, in ELC, defects undetectable by TEM are presumed to reside in the grains.

We studied defects in excimer-laser crystallized poly-Si (Kitahara et al., 2007, 2009a). We irradiated a-Si films (50 nm thick) on SiO2-coated fused quartz substrates with a XeCl excimer laser with 95% overlap. Figure 13 shows the surface of the excimer-laser crystallized poly-Si observed by SEM after Secco etching. Many hillocks appear as bright spots, as shown in Fig. 13(a); they are located at the junctions of grain boundaries. The hillock interval under the optimum conditions for ELC is nearly equal to the wavelength of the laser. AFM images indicated that ridges also lay along grain boundaries.

The grain boundaries were more evident after a longer etching time, as shown in Fig. 13(b). In contrast, grain boundaries did not appear in the hydrogenated film even after a long etching time, as shown in Fig. 13(c). This result implies that the grain boundaries were electrochemically inactivated by hydrogenation. On the other hand, structures other than grain boundaries were detected as shallow lines independently of hydrogenation, which could be due to the lack of electrochemically active dangling bonds.

luminous flux is typically shaped as a linear beam. Overlapping irradiation by the scanning pulsed light enables uniform crystallization over a wide area. The laser energy density is adjusted to be slightly lower than that at which complete melting of the Si film occurs. In practical use, *T*s is maintained at around RT during crystallization; this forms grains a few

Because of the short pulse duration, *T*s does not almost increase. However, rapid cooling after laser irradiation is an obstacle to lateral growth. Some techniques, including doublebeam irradiation, have been proposed to decrease the cooling rate. Applying multiple laser irradiations at *T*s < 400 °C reportedly increases the grain size to 4.5 μm and strongly aligns

One of the features of ELC is a rough surface morphology containing a large density of hillocks. The hillocks are generated by a positive feedback effect between the beam-induced periodic surface roughness pattern and the interference in subsequent pulses (McCulloch & Brotherton, 1995). Another feature of ELC is that *μ*FEn is considerably higher than that of SPC. The *μ*FEn value of an excimer-laser crystallized poly-Si TFT increases with increasing grain size and reaches 320 cm2/Vs at an average grain size of 700 nm, at which the dominant factor determining *μ*FEn varies from grain boundary scattering to lattice scattering (Hara et al., 2002a). Furthermore, a *μ*FEn value of 914 cm2/Vs was reportedly obtained by position-

The large *μ*FEn value of excimer-laser crystallized poly-Si has been attributed to a low defect density in grains owing to a regrowth procedure through the liquid phase. In practice, TEM images typically exhibit few defects in the grains. The average dislocation density was reportedly 8 × 106 cm-2 for large (2–4 m) grains and 108–109 cm-2 even for defective grains (Christiansen et al., 2001). However, the cooling velocity of ELC can be as high as ~1010 K/s (Sameshima & Usui, 1993). Therefore, recrystallization is expected to proceed under significant deviations from thermal equilibrium; consequently, large numbers of defects are frozen in the grains. In defects in bulk c-Si, impurity complexes generated at high temperature were shown to be frozen by quenching at a rate of ~103 K/s (Takahashi et al., 1995). The cooling rate of ELC is far larger than that in the quenching experiment. Hence, in

We studied defects in excimer-laser crystallized poly-Si (Kitahara et al., 2007, 2009a). We irradiated a-Si films (50 nm thick) on SiO2-coated fused quartz substrates with a XeCl excimer laser with 95% overlap. Figure 13 shows the surface of the excimer-laser crystallized poly-Si observed by SEM after Secco etching. Many hillocks appear as bright spots, as shown in Fig. 13(a); they are located at the junctions of grain boundaries. The hillock interval under the optimum conditions for ELC is nearly equal to the wavelength of the

The grain boundaries were more evident after a longer etching time, as shown in Fig. 13(b). In contrast, grain boundaries did not appear in the hydrogenated film even after a long etching time, as shown in Fig. 13(c). This result implies that the grain boundaries were electrochemically inactivated by hydrogenation. On the other hand, structures other than grain boundaries were detected as shallow lines independently of hydrogenation, which

hundred nanometers in size with high reproducibility.

controlled large grain growth (Mitani et al., 2008).

the surface orientation to the <111> direction (Kurimiya et al., 1993).

ELC, defects undetectable by TEM are presumed to reside in the grains.

laser. AFM images indicated that ridges also lay along grain boundaries.

could be due to the lack of electrochemically active dangling bonds.

Fig. 13. Scanning electron microscopy images for (a, b) non-hydrogenated and (c) hydrogenated excimer-laser annealed poly-Si films. Secco etching was performed for (a) 15 s and (b, c) 40 s. Hydrogenation interfered with the appearance of grain boundaries after etching.

Fig. 14. Typical transmission electron microscopy image and diffraction pattern for a grain of excimer-laser crystallized poly-Si; diffraction pattern is that in the circled region..

Fig. 15. (a) Crystalline orientation map in terms of the normal direction, (b) grain boundary map, and (c) inverse pole diagrams of the excimer-laser crystallized poly-Si, as observed by electron backscattering diffraction. Surface exhibits weak orientation to the {001} plane. Twin boundaries appear for some grains.

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 521

CLC-(c)

ELC

ELC 200-400 mJ/cm2

0 2 4 6 8 10 12 FW H M (cm -1)

) for excimer-laser crystallized poly-Si films irradiated at various energy densities.

by 3 cm-1. Plots for continuous-wave laser lateral crystallization (CLC) will be

The Raman spectra for excimer-laser crystallized poly-Si exhibit remarkable features. For example, the intensity of Raman scattering is more than ten times that for SPC films with the same thickness and even that of bulk c-Si. This enhanced Raman scattering has been reported for a roughened semiconductor surface, Si nanocones, and Si nanotubes (Sridharan et al., 2003; Jayavel et al., 2006; Cao et al., 2006). Figure 18 shows a two-dimensional map of the OPM intensity for excimer-laser crystallized poly-Si. The observed area consists of a few micrometer-size grains enlarged by super-lateral growth (SLG) and nanocrystals adjacent to the SLG region (Im & Kim, 1993). The OPM intensity is clearly enhanced at grain boundaries. Thus, the large intensity for the excimer-laser crystallized films is attributed to the enhancement of Raman scattering by hillocks and ridges around the grain boundaries. Therefore, the Raman spectra of excimer-laser crystallized poly-Si reflect mainly the situation around grain boundaries. Accordingly, polishing the sample to remove the hillocks and ridges decreases the intensity to nearly one-tenth. Then, the Raman spectra begin to reflect the region inside the grain. The FWHM changes from 5.0 to 3.8 cm-1 after polishing. This supports the conclusion that Raman spectra of unpolished specimens reflect mainly the situation around grain boundaries, where the high defect density increases the FWHM.

The other feature of ELC is the hydrogenation effect. Excimer-laser crystallized poly-Si films with and without hydrogenation were examined in terms of the OPM. Figure 19 shows the

The non-hydrogenated film exhibits relaxation of the tensile stress after a short etching period; the relaxation is due to penetration of the etching solution through clusters of defects in the grains to the poly-Si/SiO2 interface. Defects inside the grain tend to accumulate rather than remain as point defects. In contrast, the hydrogenated excimer-laser

with Secco etching time; the results of SPC are also shown for comparison.

Fig. 17. Relationship between full width at half maximum (FWHM) and lower frequency

Plots for the film crystallized under the optimum conditions (excimer-laser crystallization, ELC) and that polished to remove hillocks and ridges on the surface (ELC-p) are also shown. Solid line is calculated from the space correlation model; broken lines are values shifted to

3.0 cm-1

Spatial correlation model Crystal size (nm)

3 4 10

2

ELC-p

CLC-(a) CLC-(b)

P

shift (

larger

variation in

described in the next section.

eak shift (c

Lower frequency shift (cm-1)

m

A typical plan-view TEM image and diffraction pattern of a grain of excimer-laser crystallized poly-Si are shown in Fig. 14. The major twin boundary lies across the grain. However, unlike the results for SPC, no micro-twin was found in the grain. Dark contrast due to dislocation, defect clusters, and stacking faults tends to appear at the periphery of the grain.

The EBSD pattern of the surface normal direction (ND) is shown in Fig. 15. Although the surface orientation is scattered over a wide range, the {001} orientation has the highest frequency. A low index plane tends to exist owing to the small growth rate. The grain boundary map indicates that most grain boundaries are random and a quarter of the grains contain twin boundaries corresponding to the 3 coincident-site lattice (CSL). The twin boundaries in other grains might be eliminated by grain coalescence due to secondary grain growth.

Raman spectroscopy was applied in order to obtain macroscopic and microscopic characterization of excimer-laser crystallized poly-Si. Figure 16 shows the variation in the FWHM and of the OPM with the energy density of the laser used for crystallization. The space correlation model (Richter et al., 1981) is useful for analysis of Raman spectra in order to deduce the decrease in the regularity of the crystalline structure, i.e., the presence of nanocrystals and large density defects. The relationship between the FWHM and of the OPM is plotted in Fig. 17, which also shows a line calculated using the space correlation model. The experimental plots exhibit the same tendency as the calculated values but are shifted to a lower frequency by ~3 cm-1. The magnitude of that deviation is close to the frequency shift caused by tensile stress in the film, which is induced by shrinkage of the film during solidification. The large values of both the FWHM and at low laser energies are due to the formation of nanosized crystals. At higher laser energies, the space correlation is determined by the defect density because the grains are sufficiently large. This suggests that the defect density is greater than 1012 cm-2, which is estimated on the basis of the space correlation length deduced from the FWHM.

Fig. 16. Full width at half maximum (FWHM) and peak shift ( of the optical-phonon mode as a function of the laser energy density used for crystallization.

A typical plan-view TEM image and diffraction pattern of a grain of excimer-laser crystallized poly-Si are shown in Fig. 14. The major twin boundary lies across the grain. However, unlike the results for SPC, no micro-twin was found in the grain. Dark contrast due to dislocation,

The EBSD pattern of the surface normal direction (ND) is shown in Fig. 15. Although the surface orientation is scattered over a wide range, the {001} orientation has the highest frequency. A low index plane tends to exist owing to the small growth rate. The grain boundary map indicates that most grain boundaries are random and a quarter of the grains contain twin boundaries corresponding to the 3 coincident-site lattice (CSL). The twin boundaries in other

Raman spectroscopy was applied in order to obtain macroscopic and microscopic characterization of excimer-laser crystallized poly-Si. Figure 16 shows the variation in the

space correlation model (Richter et al., 1981) is useful for analysis of Raman spectra in order to deduce the decrease in the regularity of the crystalline structure, i.e., the presence of

OPM is plotted in Fig. 17, which also shows a line calculated using the space correlation model. The experimental plots exhibit the same tendency as the calculated values but are shifted to a lower frequency by ~3 cm-1. The magnitude of that deviation is close to the frequency shift caused by tensile stress in the film, which is induced by shrinkage of the film

due to the formation of nanosized crystals. At higher laser energies, the space correlation is determined by the defect density because the grains are sufficiently large. This suggests that the defect density is greater than 1012 cm-2, which is estimated on the basis of the space

FWHM

200 300 400 500

Laser energy density (mJ/cm2)

FWHM of c-Si

of the optical-phonon

nanocrystals and large density defects. The relationship between the FWHM and

during solidification. The large values of both the FWHM and

3

Fig. 16. Full width at half maximum (FWHM) and peak shift (

mode as a function of the laser energy density used for crystallization.

4

Shift

5

6

FWHM and

(cm-1)

7

8

9

10

correlation length deduced from the FWHM.

of the OPM with the energy density of the laser used for crystallization. The

ELC

of the

at low laser energies are

defect clusters, and stacking faults tends to appear at the periphery of the grain.

grains might be eliminated by grain coalescence due to secondary grain growth.

FWHM and

Fig. 17. Relationship between full width at half maximum (FWHM) and lower frequency shift () for excimer-laser crystallized poly-Si films irradiated at various energy densities. Plots for the film crystallized under the optimum conditions (excimer-laser crystallization, ELC) and that polished to remove hillocks and ridges on the surface (ELC-p) are also shown. Solid line is calculated from the space correlation model; broken lines are values shifted to larger by 3 cm-1. Plots for continuous-wave laser lateral crystallization (CLC) will be described in the next section.

The Raman spectra for excimer-laser crystallized poly-Si exhibit remarkable features. For example, the intensity of Raman scattering is more than ten times that for SPC films with the same thickness and even that of bulk c-Si. This enhanced Raman scattering has been reported for a roughened semiconductor surface, Si nanocones, and Si nanotubes (Sridharan et al., 2003; Jayavel et al., 2006; Cao et al., 2006). Figure 18 shows a two-dimensional map of the OPM intensity for excimer-laser crystallized poly-Si. The observed area consists of a few micrometer-size grains enlarged by super-lateral growth (SLG) and nanocrystals adjacent to the SLG region (Im & Kim, 1993). The OPM intensity is clearly enhanced at grain boundaries. Thus, the large intensity for the excimer-laser crystallized films is attributed to the enhancement of Raman scattering by hillocks and ridges around the grain boundaries. Therefore, the Raman spectra of excimer-laser crystallized poly-Si reflect mainly the situation around grain boundaries. Accordingly, polishing the sample to remove the hillocks and ridges decreases the intensity to nearly one-tenth. Then, the Raman spectra begin to reflect the region inside the grain. The FWHM changes from 5.0 to 3.8 cm-1 after polishing. This supports the conclusion that Raman spectra of unpolished specimens reflect mainly the situation around grain boundaries, where the high defect density increases the FWHM.

The other feature of ELC is the hydrogenation effect. Excimer-laser crystallized poly-Si films with and without hydrogenation were examined in terms of the OPM. Figure 19 shows the variation in with Secco etching time; the results of SPC are also shown for comparison. The non-hydrogenated film exhibits relaxation of the tensile stress after a short etching period; the relaxation is due to penetration of the etching solution through clusters of defects in the grains to the poly-Si/SiO2 interface. Defects inside the grain tend to accumulate rather than remain as point defects. In contrast, the hydrogenated excimer-laser

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 523

relatively low-energy grains. At a practical grain size of a few hundred nanometers, the lateral growth of individual grains at random orientation progresses, forming high-energy grain boundaries. A large number of the dangling bonds at grain boundaries and the accumulated

Free carriers in the channels of poly-Si TFTs are frequently scattered by grain boundaries. The presence of flow-shaped crystals aligned parallel to the channel current is expected to improve the electrical performance. The advantage of this configuration is that it does not

Flow-shaped growth of Si films has been achieved via techniques such as sequential-lateral solidification (SLS), selectively enlarging laser crystallization (SELAX), and CW laser lateral crystallization (CLC) (Im et al., 1997; Crowder et al., 2000; Hatano et al., 2002; Hara et al., 2002b; Fujii et al., 2007). SLS and SELAX employ pulsed excimer lasers and pulse-modulated diode-pumped solid-state (DPSS) CW lasers (a Nd:YVO4 laser with a wavelength of 532 nm), respectively, as heat sources. In these methods, the region crystallized by the previous laser pulse acts as the seed for the next laser irradiation. Grain boundaries and subgrain boundaries run approximately parallel to the motion of the solidification interface. The general pattern consists of a branching river-like or wishbone structure (Crowder et al., 2000). The CLC procedure is based on the laser annealing method used in SOI technology. A compact CW DPSS laser or blue-ray semiconductor laser diode (Noguchi et al., 2010) is used as the heat source. Flow-shaped lateral growth is achieved by adjusting the laser scanning velocity and output power. TFTs fabricated by SELAX and CLC exhibit *μ*Fn values of 440 and 566 cm2/Vs, respectively; these values are evidently larger than those of TFTs fabricated

We investigated the grain geometry and defects in grains produced by CLC (Kitahara et al., 2009a, 2011b). CLC was performed on 150 nm thick a-Si. Figure 20 shows SEM images of CW laser lateral crystallized poly-Si after Secco etching. The growth geometry varies from the center to the periphery of the laser beam and includes flow-shaped crystals, granular crystals, and leaf-shaped crystals extending to the periphery. The grain boundaries are distinctly visible as sharp etched lines. In the flow-shaped region, most of the etched lines were generated in the crystal, and the number of lines was diminished by coalescence with other lines. Lateral growth processes repeated the generation and reduction of boundaries, as in SLS. Note that the outlines of grains in the leaf-shaped region resemble those of the solid-phase crystallized poly-Si except for the length, which extends toward the outside of the beam. It is likely that the temperature gradient enhances lateral growth in SPC. Figure 20(a')–(c') show the surface of a hydrogenated specimen after Secco etching. Most of the grain boundaries in all regions became undetectable, which implies that they were

Figure 21 shows the crystalline orientation maps and inverse pole diagrams of the Si film in terms of the ND and reference direction (RD) determined by EBSD; the RD is parallel to the laser scanning direction. The ND is distributed and tends to exhibit a high index, whereas the RD tends to be oriented toward <101> or <100>. This suggests that the growth direction is stabilized by the liquid–solid interface rather than the surface or film–substrate interface. Many of the grains began to exhibit ND orientation to <100> with increasing film thickness

defects in the grains are electrochemically and probably electronically active.

require photolithography or complicated optical laser setup.

**7. Flow-shaped growth** 

by conventional SPC and ELC.

electrochemically active.

crystallized film exhibits almost no stress relaxation. Therefore, a large number of electrochemically active defects are assumed to be present in the grains even though they were not detected by TEM. In the solid-phase crystallized films and excimer-laser crystallized films annealed at 1000 °C, the reduction in stress with etching was not as abrupt, which implies that the thermal treatment reduced the number of dangling bonds in the grains.

Fig. 18. Two-dimensional map of optical-phonon mode intensity for excimer-laser crystallized poly-Si consisting of a few micrometer-sized grains enlarged by super-lateral growth. View area is 4.2 × 4.2 m2. The intensity is apparently large at grain boundaries.

Fig. 19. Variation in peak shift with Secco etching time observed for (a) excimer-laser crystallization (ELC) and (b) solid-phase crystallization (SPC). Only the as-crystallized excimer-laser crystallized film exhibits a rapid decrease in the peak shift, which implies stress release in the film.

On the basis of the results shown above, ELC is considered to progress as follows. Excimerlaser irradiation under optimum conditions melts the film, leaving an adequate seed density. Because solidification occurs extremely quickly, a large number of defect clusters are frozen in the grains. Multiple irradiations affect the merging of relatively high-energy grains with relatively low-energy grains. At a practical grain size of a few hundred nanometers, the lateral growth of individual grains at random orientation progresses, forming high-energy grain boundaries. A large number of the dangling bonds at grain boundaries and the accumulated defects in the grains are electrochemically and probably electronically active.

## **7. Flow-shaped growth**

522 Crystallization – Science and Technology

crystallized film exhibits almost no stress relaxation. Therefore, a large number of electrochemically active defects are assumed to be present in the grains even though they were not detected by TEM. In the solid-phase crystallized films and excimer-laser crystallized films annealed at 1000 °C, the reduction in stress with etching was not as abrupt, which implies that

(arb. units) Grain boundaries

Fig. 18. Two-dimensional map of optical-phonon mode intensity for excimer-laser crystallized poly-Si consisting of a few micrometer-sized grains enlarged by super-lateral growth. View area is 4.2 × 4.2 m2. The intensity is apparently large at grain boundaries.

Hydrogenated

Ascrystallized

0 50 100 150 Etching time (s)

Fig. 19. Variation in peak shift with Secco etching time observed for (a) excimer-laser crystallization (ELC) and (b) solid-phase crystallization (SPC). Only the as-crystallized excimer-laser crystallized film exhibits a rapid decrease in the peak shift, which implies

On the basis of the results shown above, ELC is considered to progress as follows. Excimerlaser irradiation under optimum conditions melts the film, leaving an adequate seed density. Because solidification occurs extremely quickly, a large number of defect clusters are frozen in the grains. Multiple irradiations affect the merging of relatively high-energy grains with

(b) SPC

(a) ELC

Hydrogenated

1000oC annealed Ascrystallized

400-450 350-400 300-350 250-300 200-250 150-200 100-150 50-100

Intensity

the thermal treatment reduced the number of dangling bonds in the grains.

Nanocrystal region

> 5

Peak shift (cm-1)

stress release in the film.

Free carriers in the channels of poly-Si TFTs are frequently scattered by grain boundaries. The presence of flow-shaped crystals aligned parallel to the channel current is expected to improve the electrical performance. The advantage of this configuration is that it does not require photolithography or complicated optical laser setup.

Flow-shaped growth of Si films has been achieved via techniques such as sequential-lateral solidification (SLS), selectively enlarging laser crystallization (SELAX), and CW laser lateral crystallization (CLC) (Im et al., 1997; Crowder et al., 2000; Hatano et al., 2002; Hara et al., 2002b; Fujii et al., 2007). SLS and SELAX employ pulsed excimer lasers and pulse-modulated diode-pumped solid-state (DPSS) CW lasers (a Nd:YVO4 laser with a wavelength of 532 nm), respectively, as heat sources. In these methods, the region crystallized by the previous laser pulse acts as the seed for the next laser irradiation. Grain boundaries and subgrain boundaries run approximately parallel to the motion of the solidification interface. The general pattern consists of a branching river-like or wishbone structure (Crowder et al., 2000). The CLC procedure is based on the laser annealing method used in SOI technology. A compact CW DPSS laser or blue-ray semiconductor laser diode (Noguchi et al., 2010) is used as the heat source. Flow-shaped lateral growth is achieved by adjusting the laser scanning velocity and output power. TFTs fabricated by SELAX and CLC exhibit *μ*Fn values of 440 and 566 cm2/Vs, respectively; these values are evidently larger than those of TFTs fabricated by conventional SPC and ELC.

We investigated the grain geometry and defects in grains produced by CLC (Kitahara et al., 2009a, 2011b). CLC was performed on 150 nm thick a-Si. Figure 20 shows SEM images of CW laser lateral crystallized poly-Si after Secco etching. The growth geometry varies from the center to the periphery of the laser beam and includes flow-shaped crystals, granular crystals, and leaf-shaped crystals extending to the periphery. The grain boundaries are distinctly visible as sharp etched lines. In the flow-shaped region, most of the etched lines were generated in the crystal, and the number of lines was diminished by coalescence with other lines. Lateral growth processes repeated the generation and reduction of boundaries, as in SLS. Note that the outlines of grains in the leaf-shaped region resemble those of the solid-phase crystallized poly-Si except for the length, which extends toward the outside of the beam. It is likely that the temperature gradient enhances lateral growth in SPC. Figure 20(a')–(c') show the surface of a hydrogenated specimen after Secco etching. Most of the grain boundaries in all regions became undetectable, which implies that they were electrochemically active.

Figure 21 shows the crystalline orientation maps and inverse pole diagrams of the Si film in terms of the ND and reference direction (RD) determined by EBSD; the RD is parallel to the laser scanning direction. The ND is distributed and tends to exhibit a high index, whereas the RD tends to be oriented toward <101> or <100>. This suggests that the growth direction is stabilized by the liquid–solid interface rather than the surface or film–substrate interface. Many of the grains began to exhibit ND orientation to <100> with increasing film thickness

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 525

(Hara et al., 2002b). The grain boundary map in Fig. 22(a) indicates that the 3 CSL corresponding to the first-order twin boundary lies between the high-angle grain boundaries and extends in the laser scanning direction. Figure 22(b) shows the number fractions of the boundaries in terms of their misorientation angles. The 3 CSL with a tilt angle of 60º occupies a fraction of ~0.4. First-order twin boundaries reportedly are not electrically active, whereas second-order twin boundaries act as strong recombination centers (Cunningham et al., 1982). Therefore, the present 3 CSL is regarded as electrically inactive. The remaining boundaries, which occupy a fraction of ~0.6, are distributed over a

laser lateral crystallized film: (a) flow-shaped growth, (b) granular growth, and (c) SPC-like

smaller than those obtained by ELC. The difference between CLC and ELC can be attributed to the cooling rate. That of CLC is three to four orders of magnitude smaller than that of ELC; the cooling rate of ELC is determined by the scan speed and energy gradation at the

As described above, flow-shaped crystals were successfully grown by CLC. The crystalline boundaries in the resulting films, except for the 3 CSL, are electrochemically active, like those of excimer-laser crystallized films. The defect cluster density in the grains of the CW laser lateral crystallized films is smaller than that of the excimer-laser crystallized films; this is explained by the fact that the cooling rate of CLC is several orders of magnitude smaller

0.0

Fig. 22. (a) grain boundary maps and (b) number fractions of the boundaries in terms of the

Compounding Si with germanium (Ge) not only extends the application of band engineering but also reduces the melting point, which is advantageous for growth on glass

misorientation angle for the poly-Si films shown in Fig. 21. 3 coincident site lattice

**8. Growth of quasi-single crystalline SiGe thin films** 

0.2

0.4

Number fraction

0.6

0 20 40 60

Misorientation angle (o)

for three regions in the CW

values, which are somewhat

wide range of angles.

edge of the laser beam.

than that of ELC.

5 m

2-5 o 5-180 o 3 CSL

dominates the boundaries.

Figure 17 shows the relationship between the FWHM and

growth. Regions (a) and (b) exhibit very similar FWHM and

(a) (b)

Fig. 20. Scanning electron microscopy images of continuous-wave laser lateral crystallized poly-Si. Secco etching was performed for 25 s. Observations were made in areas from the center to the periphery of the laser beam, as follows: (a) flow-shaped along the laser scanning direction, (b) granular, and (c) leaf-shaped extending to the periphery. (a')–(c') Images of the etched film after hydrogenation.

Fig. 21. Crystalline orientation maps for (a) normal direction and (b) reference direction and (a', b') individual inverse pole diagrams. Poly-Si film was formed by continuous-wave laser lateral crystallization.

(a')

(b')

(c')

3 m

1 m

1 m

001

Frequency

111

101

1 m

(b)

5 m

001 101

Fig. 21. Crystalline orientation maps for (a) normal direction and (b) reference direction and (a', b') individual inverse pole diagrams. Poly-Si film was formed by continuous-wave laser

111

Fig. 20. Scanning electron microscopy images of continuous-wave laser lateral crystallized poly-Si. Secco etching was performed for 25 s. Observations were made in areas from the center to the periphery of the laser beam, as follows: (a) flow-shaped along the laser scanning direction, (b) granular, and (c) leaf-shaped extending to the periphery. (a')–(c')

1 m

3 m

Ascrystallized Hydrogenated

(c)

Images of the etched film after hydrogenation.

(a)

5 m

lateral crystallization.

001 101

111

(a') (b')

(a)

(b)

(Hara et al., 2002b). The grain boundary map in Fig. 22(a) indicates that the 3 CSL corresponding to the first-order twin boundary lies between the high-angle grain boundaries and extends in the laser scanning direction. Figure 22(b) shows the number fractions of the boundaries in terms of their misorientation angles. The 3 CSL with a tilt angle of 60º occupies a fraction of ~0.4. First-order twin boundaries reportedly are not electrically active, whereas second-order twin boundaries act as strong recombination centers (Cunningham et al., 1982). Therefore, the present 3 CSL is regarded as electrically inactive. The remaining boundaries, which occupy a fraction of ~0.6, are distributed over a wide range of angles.

Figure 17 shows the relationship between the FWHM and for three regions in the CW laser lateral crystallized film: (a) flow-shaped growth, (b) granular growth, and (c) SPC-like growth. Regions (a) and (b) exhibit very similar FWHM and values, which are somewhat smaller than those obtained by ELC. The difference between CLC and ELC can be attributed to the cooling rate. That of CLC is three to four orders of magnitude smaller than that of ELC; the cooling rate of ELC is determined by the scan speed and energy gradation at the edge of the laser beam.

As described above, flow-shaped crystals were successfully grown by CLC. The crystalline boundaries in the resulting films, except for the 3 CSL, are electrochemically active, like those of excimer-laser crystallized films. The defect cluster density in the grains of the CW laser lateral crystallized films is smaller than that of the excimer-laser crystallized films; this is explained by the fact that the cooling rate of CLC is several orders of magnitude smaller than that of ELC.

Fig. 22. (a) grain boundary maps and (b) number fractions of the boundaries in terms of the misorientation angle for the poly-Si films shown in Fig. 21. 3 coincident site lattice dominates the boundaries.

## **8. Growth of quasi-single crystalline SiGe thin films**

Compounding Si with germanium (Ge) not only extends the application of band engineering but also reduces the melting point, which is advantageous for growth on glass

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 527

in terms of their misorientation angles. The map shows that there were very few high-angle grain boundaries. Twin islands were aligned parallel to the laser scanning direction and were outlined by the 3 CSL. Although low-angle grain boundaries with a misorientation of <5º appeared, their population was relatively small. Thus, these findings indicate that a

quasi-single crystal was formed in the film.

corresponds to regions of Ge accumulation.

both low- and high-angle grain boundaries.

and thus adjusting the lattice constant.

were formed on glass substrates.

3 m

submicrometers in front of the growth region (Kitahara et al., 2011b).

Fig. 24. Reflection electron image of the surface of the Si0.7Ge0.3 film. Bright areas

Figure 27 shows a TEM image and an energy-dispersive X-ray (EDX) profile of Si and Ge. Multiple areas of dark contrast run nearly parallel to the laser-scanning direction in the TEM image. The EDX profile indicates that Ge is strongly segregated in the dark regions of the TEM image. This finding is consistent with the Ge segregation estimated from the reflection electron image in Fig. 24. Thus, Ge segregation was found to occur at many 3 CSLs and at

The 9 CSL for pure Si reportedly can construct a dangling-bond–free stable structure through the formation of 5- and 7-membered rings (Kohyama & Yamamoto, 1994). In SiGe alloys, the energy at the boundaries can be further reduced by locally varying the Ge content

The alloy's characteristics should be considered when discussing the proper growth of SiGe. In pure materials, the solid–liquid interface under a positive temperature gradient is essentially flat during solidification. In contrast, the solid phase in alloys extends into the melt owing to the instability of the solid–liquid interface, resulting in oriented cellular growth; this corresponds to the constitutional undercooling model (Chalmers, 1964). Si and Ge form a complete solid solution. From the phase diagram and the calculation of the equilibrium partition coefficient, the undercooled area was estimated to be on the order of

As described above, the growth mode of CLC was changed from flow-shaped growth to super-lateral growth by compounding Ge with Si. This result was attributed to the constitutional undercooling that is peculiar to alloys. At crystalline boundaries running along the laser scanning direction, the energy was lowered by Ge segregation, resulting in the suppression of high-angle grain boundaries. As a result, quasi-single-crystalline films

substrates. Rapid melt growth on sapphire substrates and ELC on glass substrates have been reported as techniques for the growth of SiGe thin films (Sameshima et al., 2005; Weizman et al., 2005; Koh et al., 2010; Tanaka et al., 2010) Strong lateral segregation reportedly caused the local Ge content to differ by as much as 40% from the average value (Weizman et al., 2005). Therefore, it is necessary to examine the segregation of Ge and the appropriate growth mode for alloys in order to form crystalline SiGe on glass substrates.

We investigated the crystal configuration, Ge segregation, and growth modes specific to the alloy (Kitahara et al., 2011b). SiGe thin films were crystallized on glass substrates using CLC with a DPSS CW laser. The precursor film was amorphous Si0.7Ge0.3 with a thickness of 100 nm.

A typical Raman spectrum of the Si0.7Ge0.3 film is shown in Fig. 23. The crystallization of the SiGe films was confirmed by the appearance of sharp optical phonon modes at ~300 cm-1 for the Ge–Ge bond, ~400 cm-1 for the Si–Ge bond, and ~500 cm-1 for the Si–Si bond.

Figure 24 shows a reflection electron image. In the region corresponding to the central position of the laser beam, multiple areas of bright contrast that extend for more than 100 μm run parallel to the laser scanning direction at intervals of ~1 μm. The coalescence and generation of grain boundaries observed in pure Si films were not evident here, which indicates a type of super-lateral growth. Because the reflection electron intensity depends on the atomic number, the contrast suggests strong segregation of Ge along the bright lines.

Fig. 23. Typical Raman spectrum for the optical phonon modes of Ge–Ge, Si–Ge, and Si–Si bonds in a crystallized Si0.7Ge0.3 film. Peak position for unstrained Si is indicated by 0. The lower shift of the Si–Si band from 0 depends primarily on the Ge content.

Figure 25 shows crystalline orientation maps and inverse pole diagrams of the film determined by EBSD. "ND" indicates that the crystals are strongly aligned with the <111> plane. The RD, which corresponds to the laser scanning direction, tends to align with the <101> plane. Figure 26 shows the grain boundary map and number fractions of boundaries

substrates. Rapid melt growth on sapphire substrates and ELC on glass substrates have been reported as techniques for the growth of SiGe thin films (Sameshima et al., 2005; Weizman et al., 2005; Koh et al., 2010; Tanaka et al., 2010) Strong lateral segregation reportedly caused the local Ge content to differ by as much as 40% from the average value (Weizman et al., 2005). Therefore, it is necessary to examine the segregation of Ge and the appropriate

We investigated the crystal configuration, Ge segregation, and growth modes specific to the alloy (Kitahara et al., 2011b). SiGe thin films were crystallized on glass substrates using CLC with a DPSS CW laser. The precursor film was amorphous Si0.7Ge0.3 with a

A typical Raman spectrum of the Si0.7Ge0.3 film is shown in Fig. 23. The crystallization of the SiGe films was confirmed by the appearance of sharp optical phonon modes at ~300 cm-1 for

Figure 24 shows a reflection electron image. In the region corresponding to the central position of the laser beam, multiple areas of bright contrast that extend for more than 100 μm run parallel to the laser scanning direction at intervals of ~1 μm. The coalescence and generation of grain boundaries observed in pure Si films were not evident here, which indicates a type of super-lateral growth. Because the reflection electron intensity depends on the atomic number, the contrast suggests strong segregation of Ge along the bright lines.

> 250 300 350 400 450 500 550 300 400 500 Frequency (cm-1)

> > 0 depends primarily on the Ge content.

Fig. 23. Typical Raman spectrum for the optical phonon modes of Ge–Ge, Si–Ge, and Si–Si bonds in a crystallized Si0.7Ge0.3 film. Peak position for unstrained Si is indicated by

Figure 25 shows crystalline orientation maps and inverse pole diagrams of the film determined by EBSD. "ND" indicates that the crystals are strongly aligned with the <111> plane. The RD, which corresponds to the laser scanning direction, tends to align with the <101> plane. Figure 26 shows the grain boundary map and number fractions of boundaries

Si-Ge

Si-Si

0

> 0. The

growth mode for alloys in order to form crystalline SiGe on glass substrates.

the Ge–Ge bond, ~400 cm-1 for the Si–Ge bond, and ~500 cm-1 for the Si–Si bond.

Intensity (arb. units)

lower shift of the Si–Si band from

Ge-Ge

Si0.7Ge0.3

thickness of 100 nm.

in terms of their misorientation angles. The map shows that there were very few high-angle grain boundaries. Twin islands were aligned parallel to the laser scanning direction and were outlined by the 3 CSL. Although low-angle grain boundaries with a misorientation of <5º appeared, their population was relatively small. Thus, these findings indicate that a quasi-single crystal was formed in the film.

Fig. 24. Reflection electron image of the surface of the Si0.7Ge0.3 film. Bright areas corresponds to regions of Ge accumulation.

Figure 27 shows a TEM image and an energy-dispersive X-ray (EDX) profile of Si and Ge. Multiple areas of dark contrast run nearly parallel to the laser-scanning direction in the TEM image. The EDX profile indicates that Ge is strongly segregated in the dark regions of the TEM image. This finding is consistent with the Ge segregation estimated from the reflection electron image in Fig. 24. Thus, Ge segregation was found to occur at many 3 CSLs and at both low- and high-angle grain boundaries.

The 9 CSL for pure Si reportedly can construct a dangling-bond–free stable structure through the formation of 5- and 7-membered rings (Kohyama & Yamamoto, 1994). In SiGe alloys, the energy at the boundaries can be further reduced by locally varying the Ge content and thus adjusting the lattice constant.

The alloy's characteristics should be considered when discussing the proper growth of SiGe. In pure materials, the solid–liquid interface under a positive temperature gradient is essentially flat during solidification. In contrast, the solid phase in alloys extends into the melt owing to the instability of the solid–liquid interface, resulting in oriented cellular growth; this corresponds to the constitutional undercooling model (Chalmers, 1964). Si and Ge form a complete solid solution. From the phase diagram and the calculation of the equilibrium partition coefficient, the undercooled area was estimated to be on the order of submicrometers in front of the growth region (Kitahara et al., 2011b).

As described above, the growth mode of CLC was changed from flow-shaped growth to super-lateral growth by compounding Ge with Si. This result was attributed to the constitutional undercooling that is peculiar to alloys. At crystalline boundaries running along the laser scanning direction, the energy was lowered by Ge segregation, resulting in the suppression of high-angle grain boundaries. As a result, quasi-single-crystalline films were formed on glass substrates.

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 529

Lateral crystallization is indispensable for the growth of poly-Si films on glass substrates for application in large-area devices such as electronic displays and solar cells. In this paper, an outline of the various crystallization techniques was introduced, and our study of lateral

When SPC was used, lateral growth proceeded along the major twin boundary. However, a large density of micro-twins appeared in the grains. Oriented lateral growth is expected to decrease the number of defects. In practice, MILC exhibited a decreased defect density with lowered growth temperature, which is attributed to enhanced directivity of needle-like

In ELC, growth is accompanied by melting and recrystallization of the film. TEM revealed only low-density defects in grains. Those defects consisted primarily of dislocations. However, Raman spectroscopy and chemical etching indicated the presence of a considerable density of defects in the grains and at the grain boundaries. The existence of defects in the grains was attributed to the high cooling velocity during recrystallization.

CLC is advantageous for its low cooling velocity and ability to produce oriented lateral growth. Relatively small defect densities were observed in pure Si films. Lateral growth proceeds by generation and disappearance of sub-grain boundaries along the laser scanning direction. Lateral growth was significantly enhanced by compounding Ge with Si, which results in a type of super-lateral growth. Such growth was attributed to the constitutional undercooling effect that is characteristics of alloys. The undercooling effect and the segregation of Ge decrease the energy at the boundaries running parallel to the laser

Not only Si and Si-related films but also metal oxides and organic films are beginning to be extensively developed as materials for electronic devices on glass or plastic films. Si will share this role in the development of high-performance devices. One of the remarkable achievements of Si technology is the growth of single-crystalline films on large-area substrates, although such films suffer from cracking, which must be avoided in heterostructures. Both the advance of laser technologies and the control of the physical properties of materials, such as alloying and stacking of different layers, are more promising

The authors would like to thank Prof. Tsuda H. in Osaka Prefecture Univ. for TEM observation of SPC and MILC poly-Si. This study was funded in part by the Japan Society for the Promotion of Science [Grants-in-Aid for Scientific Research, (B) 19360165 and (C)

Blake, J. G.; King, M. C.; Stevens III, J. D. & Young R. (1997). Low-Temperature Polysilicon Reshapes FPD Procudtuin, *Solid State Technology*, Vol, 40, No. 151, pp. 151-161

scanning direction, resulting in the growth of quasi-single-crystalline films on glass.

for the growth of single-crystal films on non-crystalline substrates.

**10. Acknowledgements** 

21560329].

**11. References** 

**9. Conclusions** 

growth.

growth and defects was described.

Fig. 25. Crystal orientation maps and inverse pole diagrams for (a, a') normal direction and (b, b') reference direction of Si0.7Ge0.3 film. Surface is strongly oriented to <111>.

Fig. 26. Grain boundary map and number fractions of the boundaries in terms of the misorientation angle for Si0.7Ge0.3 film. Boundaries consist mainly of 3 coincident site lattice.

Fig. 27. Transmission electron microscopy image and electron dispersion X-ray profile of Si and Ge atoms in crystallized Si0.7Ge0.3 film. Profiles were taken between A and B in the image.

## **9. Conclusions**

528 Crystallization – Science and Technology

(b)

(a)

5 m

001 101

2-5 o 5-180 o 3 CSL

A B

3 m

111

(a') (b') 111

(b, b') reference direction of Si0.7Ge0.3 film. Surface is strongly oriented to <111>.

5 m

Fig. 25. Crystal orientation maps and inverse pole diagrams for (a, a') normal direction and

0.0

0.2

0.4

Number fraction

Fig. 26. Grain boundary map and number fractions of the boundaries in terms of the

misorientation angle for Si0.7Ge0.3 film. Boundaries consist mainly of 3 coincident site lattice.

Intensity (arb.u

Intensity (arb. units)

Fig. 27. Transmission electron microscopy image and electron dispersion X-ray profile of Si and Ge atoms in crystallized Si0.7Ge0.3 film. Profiles were taken between A and B in the image.

1 m

0.6

001 101

001

0 20 40 60

Misorientation angle (o)

Si

0.0 0.5 1.0 1.5 2.0 P osition (m )

Ge <sup>A</sup> <sup>B</sup>

Frequency

111

101

Lateral crystallization is indispensable for the growth of poly-Si films on glass substrates for application in large-area devices such as electronic displays and solar cells. In this paper, an outline of the various crystallization techniques was introduced, and our study of lateral growth and defects was described.

When SPC was used, lateral growth proceeded along the major twin boundary. However, a large density of micro-twins appeared in the grains. Oriented lateral growth is expected to decrease the number of defects. In practice, MILC exhibited a decreased defect density with lowered growth temperature, which is attributed to enhanced directivity of needle-like growth.

In ELC, growth is accompanied by melting and recrystallization of the film. TEM revealed only low-density defects in grains. Those defects consisted primarily of dislocations. However, Raman spectroscopy and chemical etching indicated the presence of a considerable density of defects in the grains and at the grain boundaries. The existence of defects in the grains was attributed to the high cooling velocity during recrystallization.

CLC is advantageous for its low cooling velocity and ability to produce oriented lateral growth. Relatively small defect densities were observed in pure Si films. Lateral growth proceeds by generation and disappearance of sub-grain boundaries along the laser scanning direction. Lateral growth was significantly enhanced by compounding Ge with Si, which results in a type of super-lateral growth. Such growth was attributed to the constitutional undercooling effect that is characteristics of alloys. The undercooling effect and the segregation of Ge decrease the energy at the boundaries running parallel to the laser scanning direction, resulting in the growth of quasi-single-crystalline films on glass.

Not only Si and Si-related films but also metal oxides and organic films are beginning to be extensively developed as materials for electronic devices on glass or plastic films. Si will share this role in the development of high-performance devices. One of the remarkable achievements of Si technology is the growth of single-crystalline films on large-area substrates, although such films suffer from cracking, which must be avoided in heterostructures. Both the advance of laser technologies and the control of the physical properties of materials, such as alloying and stacking of different layers, are more promising for the growth of single-crystal films on non-crystalline substrates.

## **10. Acknowledgements**

The authors would like to thank Prof. Tsuda H. in Osaka Prefecture Univ. for TEM observation of SPC and MILC poly-Si. This study was funded in part by the Japan Society for the Promotion of Science [Grants-in-Aid for Scientific Research, (B) 19360165 and (C) 21560329].

## **11. References**

Blake, J. G.; King, M. C.; Stevens III, J. D. & Young R. (1997). Low-Temperature Polysilicon Reshapes FPD Procudtuin, *Solid State Technology*, Vol, 40, No. 151, pp. 151-161

Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 531

Hatano, M.; Shiba, T. & Ohkura, M. (2002). Selectively Enlarging Laser Crystallization

Im, J. S.; Sposili, R. S. & Crowder, M. A. (1997). Single-Crystal Si Films for Thin-Film

Jayavel, P. ; Nakamura, S.; Kesavamoorsly, R.; Srivastava, G. P.; Tomoda, W.; Koyama, T. &

Jellison, Jr. G. E. (1998). Spectroscopic Ellipsometry Data Analysis: Measure Versus

Kawazu, T.; Kudo, H.; Onari, S. & Arai, T. (1990). Low-Temperature Crystallization of

*Journal of Applied Physics*, Vol. 29, No. 12, (December 1990), pp. 2698-2704 Kitahara, K.; Yamazaki, R.; Kurosawa, T.; Nakajima, K. & Moritani, A. (2002). Analysis of

Kitahara, K.; Ohnishi, K.; Katoh, Y.; Yamazaki, R. & Kurosawa, T. (2003). Analysis of Defects

*Japanese Journal of Applied Physics*, Vol. 46, No. 19, (May 2007), pp. L448-L450 Kitahara, K.; Ohashi, Y.; Yamamoto, K. & Sasaki, N. (2009a). Electrochemical and Raman-

Kitahara, K.; Kambara, J.; Kobata, M. & Tsuda, H. (2009b). Suppression of Defects during

Kitahara, K.; Ishii, T.; Suzuki, J.; Bessyo, T.; Watanabe, N. (2011a). Characterization of

Kitahara, K.; Hirose, K.; Suzuki, J.; Kondo, K. & Hara, A. (2011b) Growth of Quasi-Single-

pp. 632139 (14 pages), doi:10.1155/2011/632139

*Digest of Technical Papers*, Boston, May 2002, Vol. 33, Issue 1, pp. 158-161 Hayashi, H.; Noguchi, T. & Oshima, T. (1984). Polysilicon Super-Thin Film Transistor (SFT), *Japanese Journal of Applied Physics*, Vol. 23, No. 11, (November 1984), L819-L820 Im, J. S. & Kim, H. J. (1993). Phase Transformation Mechanisms Involved in Excimer Laser

(October 1993), pp. 1969-1971, ISSN 0003-6951

3436, ISSN 0003-6951

ISSN 0040-6090

pp. 5055-5059

pp. 021205 (6 pages)

pages)

(February 2006), pp. R19-R21

Technology for High and Uniform Performance Poly-Si TFTs, *SID Symposium* 

Crystallization of Amorphous Silicon Films, *Applied Physics Letters*, Vol. 63, No. 14,

Transistor Devices, *Applied Physics Letters*, Vol. 70, No. 25, (June 1997), pp. 3434-

Hayakawa, Y. (2006). Surface Morphology Effects on the Optical Phonon Modes in InAsxSb1-x Epilayers on GaAs(001), *Physica Status Solidi* (b), Vol. 243, No. 4,

Calculated Quantities, *Thin Solid Films*, Vol. 313-314, (February 1998), pp. 33-39,

Hydrogenated Amorphous Silicon Induced by Nickel Silicide Formation, *Japanese* 

Stress in Laser-Crystallized Polysilicon Thin Films by Raman Scattering Spectroscopy, *Japanese Journal of Applied Physics*, Vo. 41, Part 1, No. 8, (August 2002),

in Polycrystalline Silicon Thin Films Using Raman Scattering Spectroscopy, *Japanese Journal of Applied Physics*, Vol. 11, Part 1, No. 11, (November 2003), pp. 6742-6747 Kitahara, K.; Ohashi, Y. & Matsumoto, S. (2007). Hydrogenation Effects on Chemical

Activity of Defects in Excimer-Laser-Annealed Polycrystalline Silicon Thin Films,

Scattering Characterizations of Defects in Polycrystalline Silicon Thin Films Formed by Excimer-Laser Annealing, Solid-Phase Crystallization, and Continuous-Wave Lateral Crystallization, *Japanese Journal of Applied Physics*, Vol. 48, (February 2009),

Metal-Induced Lateral Crystallization of Polycrystalline-Silicon Thin Films by Directed Lateral Growth, *Japanese Journal of Applied Physics*, Vol. 48, pp. 091203 (5

Defects and Stress in Polycrystalline Silicon Thin Films on Glass Substrates by Raman Microscopy, *International Journal of Spectroscopy*, Vol. 2011, (August 2011),

Crystal Silicon-Germanium Thin Films on Glass Substrates by Continuous Wave


Cammarata, R. C.; Thompson, C. V.; Hayzelden, C & Tu, K. N. (1990). Silicide Precipitation

Christiansen, S.; Lengsfeld, P.; Krinke, J.; Nerding, M.; Nickel, N. H. & Strunk, H. P. (2001).

Films, *Journal of Applied Physics*, Vol. 89, No. 10 (May 2001), pp. 5348-5354 Crowder, M. A.; Limanov, A. B. & Im J. S. (2000). Sub-Grain Boundary Spacing in

Colinge, J.-P. (2004). *Silicon-on-Insulator Technology: Materials to VLSI*, Springer, New York,

Cunningham, B.; Strunk, H. & Ast, D. G. (1982). First and Second Order Twin Boundaries in

Drosed, R. & Washburn, J. (1980). A New Technique for Observing the Amorphous to

*Applied Physics*, Vol. 51, No. 8 (August 1980), pp. 4108-4110, ISSN 0021-8979 Englert, Th.; Abstrerrer, G. A.; & Pontcharra, J. (1980). Determination of Existing Stress in

Frost R. L. & Shurvell H. F. (1997). Raman Microprobe Spectroscopy of Halloysite, *Clays and* 

Fujii S., Kuroki S., Kotani K. & Ito T. (2007). Enlargement of Crystal Grains in Thin Silicon

Givargizov, E. I. (1991). *Oriented Crystallization on Amorphous Substrate*, Premium Press, New

Hara, A.; Takeuchi, F. & Sasaki, N. (2002a). Mobility Enhancement Limit of Excimer-Laser-

Hara, A.; Takeuchi, F.; Takei, M.; Suga, K.; Yoshino, K.; Chida, M.; Sano, Y. & Sasaki, N.

Hatalis, M. K. & Greve, D. W. (1988). Large Grain Policrystalline Silicon by Low-

Vol. 91, No. 2, (January 2002) pp. 708-714, ISSN 0021-8979

*Journal of Materials Research*, Vo. 5, No. 10, (October 1990), pp. 2133-2138 Cao, L.; Nabet, B. & Spanier, J. E. (2006), Enhanced Raman Scattering from Individual

Chalmers, B. (1964). *Principles of Solidification*, John Wiley and Sons Inc., New York

2006), pp. 157402-1-4.

ISBN: 1-4020-7773-4

(February 1982), pp. 237-239

*Electronics*, Vol. 23, pp. 31-33

York, ISBN: 0-306-43122-X

L311-L313

ISSN 0021-8979

*Clay Minerals*, Vol. 45, No. 1, pp. 58-72

Vol. 46, No. 4B (April 2007), pp. 2501-2504

Q9.6.6

and Silicon Crystallization in Nickel Implanted Amorphous Silicon Thin Films,

Semiconductor Nanocones and Nanowires, *Physical Review Letters*, Vol. 96 (April

Nature of Grain Boundaries in Laser Crystallized Polycrystalline Silicon Thin

Directionally Crystallized Si Films Obtained via Sequential Lateral Solidification, *Materials Research Society Symposium Proceedings*, Boston, 2000, Vol. 621, pp. Q9.6.1-

Edge Defined Film Growth Silicon Ribbon, *Applied Physics Letters*, Vol. 40, No. 3

Crystalline Transformation in Thin Surface Layers on Silicon Wafers, *Journal of* 

Silicon Films on Sapphire Substrate Using Raman Spectroscopy, *Solid-State* 

Films by Continuous-Wave Laser Irradiation, *Japanese Journal of Applied Physics*,

Crystallized Polycrystalline Silicon Thin Film Transistors, *Journal of Applied Physics*,

(2002b). High Performance Polycrystalline Silicon Thin Film Transistors on Non-Alkali Glass Produced Using Continuous Wave Laser Lateral Crystallization, *Japanese Journal of Applied Physics*, Vol. 41, Part 2, No. 3B, (March 2002), pp.

Temperature Annealing of Low-Pressure Chemical Vapor Deposited Amorphous Silicon Films, *Journal of Applied Physics*, Vol. 63, No. 7, (April 1988), pp. 2260-2266,


Oriented Lateral Growth and Defects in Polycrystalline-Silicon Thin Films on Glass Substrates 533

Noguchi, T.; Chen, Y.; Miyahira, T.; Mugiraneza, J. D.; Ogino, Y.; Iida, Y.; Sahota, E. & Terao,

Richter, H.; Wang, Z. P. & Ley, L. (1981).The One Phonon Raman Spectrum in

Roth, J. A.; Olsen, G. L.; Jacobson, D. C. & Poate, J. M. (1990). Kinetics of Solid Phase Epitaxy

Samesima, T.; Usui, S. & Sekiya, M. (1986). XeCl Excimer Laser Annealing Used in the

Sameshima, S. & Usui, S. (1993). Pulsed Laser-Induced Melting Followed by Quenching of

Sameshima, T.; Watakabe, H.; Kanno, H.; Sadoh, T. & Miyao, M. (2005). Pulsed Laser

Sameshima, T. (2009). Laser Crystallization for Large-Area Electronics, *Applied Physics* A,

Secco d' Aragona, F. (1972). Dislocation Etch for (100) Planes in Silicon, *Journal of* 

Sridharan, M.; Narayandass, Sa. K.; Mangalaraj, D. & Lee, H. C. (2003). Effect of Boron Ion

Spinella, C.; Lombard, S. & Prioro, F. (1998). Crystal Grain Nucleation in Amorphous

Takahashi, H.; Suezawa, M. & Sumino, K. (1995). Generation and Dissociation of Complexes

Takayama, T.; Ohtani, H.; Miyanaga, A.; Mitsuki, T.; Ohnuma, H.; Nakajima, S. & Yamazaki,

Tanaka, T.; Toko, K.; Sadoh, T. & Miyao, M. (2010). High Quality Single-Crystalline Ge-Rich

Thompson, C. V. (1985). Secondary Grain Growth in Thin Films of Semiconductors:

*Applied Physics Express*, Vol. 3, (February 2010), pp. 031301 (3 pages) Temple, P. A. & Hathaway, C. E. (1973). Multiphonon Raman Spectrum of Silicon, *Physical* 

(September 1995), pp. 3078-3082, ISSN 0021-8979

*Review* B, Vol, 7, No. 8, (April 1973), pp. 3685-3697

pp. 03CA10 (3 pages)

6598, ISSN 0021-8979

ISSN 0040-6090

1972), pp. 948-951

201, pp. 465-474

ISSN 0021-8979

772, ISSN 0021-8979

25-28

*Letters*, Vol. 57, issue 13, pp. 1340-1342

1986), pp. 276-278, ISSN 0741-3106

Vol. 96, (April 2009), pp. 137-144

0038-1098

M. (2010). Advanced Micro-Polycrystalline Silicon Films Formed by Blue-Multi-Laser-Diode Annealing, *Japanese Journal of Applied Physics*, Vol. 49, (March 2010),

Microcrystalline Silicon, *Solid State Communications*, Vol. 39, pp. 625-629, ISSN

in Thick Amorphous Si Layers Formed by MeV Ion Implantation, *Applied Physics* 

Fabrication of Poly-Si TFT's, *IEEE Electron Device Letters*, Vol. EDL-7, No. 5, (May

Silicon Films, *Journal of Applied Physics*, Vol. 74, No. 11, (December 1993, pp. 6592-

Crystallization of Silicon-Germanium Films, *Thin Solid Films*, Vol. 487, pp. 67-71,

*Electrochemical Societies: Solid State Science and Technology*, Vol. 119, No. 7, (July

Implantation on the Structural and Optical Properties of Polycrystalline Cd0.96Zn0.04Te Thin Films, *Nuclear Instruments and Methods in Physics Research* B, Vol.

Silicon, *Journal of Applied Physics*, Vol. 84, No. 10, (November 1998), pp. 5383-5413,

of Iron and Phosphorus Atoms in Silicon, *Journal of Applied Physics*, Vol. 78, No. 5,

S. (2000). Continuous Grain Silicon Technology and Its Application for Active Matrix Display, *Digest of Technical Papers: AM-LCD 2000*, Tokyo, July 2000, pp.

SiGe on Insulator Structures by Si-Doping Controlled Rapid Melting Growth,

Theoretical Aspects, *Journal of Applied Physics*, Vol. 58, No. 2, (July 1985), pp. 763-

Laser Lateral Crystallization , *Japanese Journal of Applied Physics*, Vol. 50, (October 2011), pp, 115501 (6 pages)


Koh, H.-Y. S.; Chen, S.-L.; Griffin, P. B. & Plummer, J. D. (2010). High Quality Single-Crystal

Kohyama, M. & Yamamoto, R. (1994). Theoretical Study of Grain Boundaries in Si: Effects of

Kuriyama, H.; Nohda, T.; Ishida, S.; Kuwahara, T.; Noguchi, S.; Kiyama, S.; Tsuda, S. &

Little, T. W.; Takahashi, K.; Koide, H.; Nakazawa, T.; Yudasaka, I. & Ohshima, H. (2001).

Makihara, K.; Nozaki, H.; Asano, T. & Miyasaka, M. (2003). Metal-Induced Lateral

McCulloch, D. J. & Brotherton, S. D. (1995). Surface Roughness Effects in Laser Crystallized

Menendez, J. & Cardona, M. (1984). Temperature Dependence of the First-Order Raman

Mitani, M.; Endo, T.; Taniguchi, Y.; Katou, T.; Shimoto, S.; Ohno, T.; Tsuboi, S.; Okada, T.;

Mizuki, T.; Matsuda, J. S.; Nakamura, Y.; Takagi, J. & Yoshida, T. (2004). Large Domains of

Nakamura,A.; Emoto, F.; Fujii, E.; Yamamoto, A.; Uemoto, Y.; Senda, K. & Kano, G. (1989).

Nickel, N. H.; Anderson, G. B. & Johnson R. I. (1997). Grain-Boundary Defects in Laser-

*Physics*, Vol. 32, Part 1, No. 12B, (December 1993), pp. 6190-6195

*State Letters*, Vol. 13, Issue 8, (May 2010), pp. H281-H283

2011), pp, 115501 (6 pages)

(December 1992), pp. 4550-4554

*Phenomena*, Vol. 93, pp. 207-212

Vol. 29, No. 4, (February 1984), pp. 2051-2059

2060-2062, ISSN 0003-6951

2008), pp. 8707-8713

4248-4251, ISSN 0021-8979

1997), pp. 12062-12068, ISSN 0163-1829

0018-9383

3724-3728

Laser Lateral Crystallization , *Japanese Journal of Applied Physics*, Vol. 50, (October

Laterally Graded SiGe on Insulator by Rapid Melt Growth, *Electrochemical and Solid* 

Structural Disorder on the Local Electronic Structure and the Origin of Band Tails, *Physical Review* B, Vol. 50, No. 12, (September 1994), pp. 8502-8522, ISSN 0163-1829 Kuriyama, H.; Kuwahara, T.; Ishida, S.; Nohda, I.; Sano, K. & Iwata, H. (1992). Improving

the Uniformity of Poly-Si Films Using a New Excimer Laser Annealing Method for Giant- Microelectronics, *Japanese Journal of Applied Physics*, Vol. 31, Part 1, No. 12B,

Nakano, S. (1993). Lateral Grain Growth of Poly-Si Films with a Specific Orientation by an Excimer Laser Annealing Method, *Japanese Journal of Applied* 

Low Temperature Poly-Si TFTs Using Solid Phase Crystallization of Very Thin Films and an Electron Cyclotron Resonance Chemical Vapor Deposition Gate Insulator, *Japanese Journal of Applied Physics*, Vol. 30. No. 12B, (December 1991), pp.

Crystallization of Amorphous Silicon under Reduced Nickel Supply, *Solid State* 

Polycrystalline Silicon, *Applied Physics Letters*, Vol. 66, No. 16, (April 1995), pp.

Scattering by Phonons in Si, Ge, and a-Sn: Anharmonic Effects, *Physical Review* B,

Azuma, K.; Kawachi, G. & Matsumura, M. (2008). Ultrahigh-Performance Polycrystalline Silicon Thin-Film Transistors on Excimer-Laser-Processed Pseodo-Single Crystal Films, *Japanese Journal of Applied Physics*, Vol. 47, No. 12, (December

Continuous Grain Silicon on Glass Substrate for High-Performance TFTs, *IEEE Transactions on Electron Devices*, Vol. 51, No. 2, (February 2004), pp. 204-211, ISSN

Analysis of Solid Phase Crystallization in Amorphized Polycrystalline Si Films on Quartz Substrates, *Journal of Applied Physics*, Vo. 66, No. 9, (November 1989), pp.

Crystallized Polycrystalline Silicon, *Physical Review* B, Vol. 56, No. 19, (November


**20** 

R. Ramasamy

*India* 

 *Ocean Engineering, IITM, Chennai,* 

**Crystallization, Fractionation and Solidification** 

A mineral is naturally occurring substance with a characteristic crystal structure and chemical composition. It occurs in rocks. Frequently, it has homogeneous crystalline structure in which one or more types of atoms or molecules may be partially substituted for the original atoms and molecules without changing the structure. Thus, most rock-forming minerals are formed in solid solutions. Aggregates of minerals produce characteristic rocktextures which shed light on their trends of magmatic evolution. They crystallize from molten rock, called magma. Magmas are a natural material of hot silicate / carbonate /oxide / phosphate/ sulphide and sulphur melts from which igneous rocks form. They develop by the partial melting of deep-seated rocks at depth. Primary melts are composed of suspended crystals and dissolved gases. The crystallization of a primary melt and the isolation of the rest liquid freed from suspended solids are called the parent magma, which then produces a series of residual magmas of a secondary nature with varying compositions through differentiation and fractionation. The molten material is necessarily forced to the surface, either by hydrostatic head in the mantle where the encasing rock is denser than the melt or else via gas by propulsion. Pressure enhances the solubility of H2O and other volatiles, which lower the equilibrium temperature of solidification. These volatile substances may largely escape during the course of the ascension of the magma as well as during the course of crystallization. Rising toward the Earth's surface, the magma enters zones of lower temperatures and pressures. Decreasing temperatures tend to bring about crystallization, which produces solid crystals suspended in the liquid. Other solid fragments are incorporated from the walls and roof of the conduit through which the magma is rising and the magma changes its composition by assimilation. A closed magmatic chamber is formed by the filling up of magma in cavities at intermittent stopping places. Over the course of the magmatic crystallization of the parent magma in a closed magmatic chamber, volatiles may be concentrated and interacted with minerals during slow cooling. Magma is generally composed of eight major oxides SiO2, Al2O3, Fe2O3, FeO, MgO, CaO, Na2O and K2O, as well as from lesser proportions of TiO2, P2O5, H2O, CO2, S and other volatiles. In addition to these there are other substances constituting various gases and trace elements. As crystallization progresses, the volatiles and the more soluble silicate components are concentrated in the

**1. Introduction** 

**of Co-Magmatic Alkaline Series Sequentially** 

**Emplaced in the Carbonatite Complex of** 

**Tiruppattur, Tamil Nadu, India** 

Weizman, M.; Nickel, N.H.; Sieber, I.; Bohne, W.; Rohrich, J.; Strub, E. & Yan, B. (2005). Phase Segregation in Laser Crystallized Polycrystalline SiGe Thin Films, *Thin Solid Films*, Vol. 487, (March 2005), pp. 72-76, ISSN 0040-6090

## **Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India**

R. Ramasamy  *Ocean Engineering, IITM, Chennai, India* 

## **1. Introduction**

534 Crystallization – Science and Technology

Weizman, M.; Nickel, N.H.; Sieber, I.; Bohne, W.; Rohrich, J.; Strub, E. & Yan, B. (2005).

*Films*, Vol. 487, (March 2005), pp. 72-76, ISSN 0040-6090

Phase Segregation in Laser Crystallized Polycrystalline SiGe Thin Films, *Thin Solid* 

A mineral is naturally occurring substance with a characteristic crystal structure and chemical composition. It occurs in rocks. Frequently, it has homogeneous crystalline structure in which one or more types of atoms or molecules may be partially substituted for the original atoms and molecules without changing the structure. Thus, most rock-forming minerals are formed in solid solutions. Aggregates of minerals produce characteristic rocktextures which shed light on their trends of magmatic evolution. They crystallize from molten rock, called magma. Magmas are a natural material of hot silicate / carbonate /oxide / phosphate/ sulphide and sulphur melts from which igneous rocks form. They develop by the partial melting of deep-seated rocks at depth. Primary melts are composed of suspended crystals and dissolved gases. The crystallization of a primary melt and the isolation of the rest liquid freed from suspended solids are called the parent magma, which then produces a series of residual magmas of a secondary nature with varying compositions through differentiation and fractionation. The molten material is necessarily forced to the surface, either by hydrostatic head in the mantle where the encasing rock is denser than the melt or else via gas by propulsion. Pressure enhances the solubility of H2O and other volatiles, which lower the equilibrium temperature of solidification. These volatile substances may largely escape during the course of the ascension of the magma as well as during the course of crystallization. Rising toward the Earth's surface, the magma enters zones of lower temperatures and pressures. Decreasing temperatures tend to bring about crystallization, which produces solid crystals suspended in the liquid. Other solid fragments are incorporated from the walls and roof of the conduit through which the magma is rising and the magma changes its composition by assimilation. A closed magmatic chamber is formed by the filling up of magma in cavities at intermittent stopping places. Over the course of the magmatic crystallization of the parent magma in a closed magmatic chamber, volatiles may be concentrated and interacted with minerals during slow cooling. Magma is generally composed of eight major oxides SiO2, Al2O3, Fe2O3, FeO, MgO, CaO, Na2O and K2O, as well as from lesser proportions of TiO2, P2O5, H2O, CO2, S and other volatiles. In addition to these there are other substances constituting various gases and trace elements. As crystallization progresses, the volatiles and the more soluble silicate components are concentrated in the

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

cooling magma and the formation of minerals that make up igneous rocks.

suddenly when the melt reaches the appropriate temperature.

1. The compositional paths followed by both the melts and the growing crystal depend on

2. The composition of both the melt and the crystal at any point in the crystallization process is dependent on the extent to which the crystallization process has been

3. For minerals that exhibit a solid solution, crystallization proceeds in a continuous manner with the composition of the mineral changing along with that of the magmatic

4. For minerals that do not exhibit a solid solution, melting or crystallization will proceed in discrete discontinuous steps. As a melt cools, such discrete minerals will appear

5. Minerals that crystallize at high temperatures are rich in Mg and Fe and relatively poor

Some generalizations:

completed.

liquid.

in Si.

the initial bulk composition of the melt.

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 537

changes in structure of residual melts in spatial and temporal conditions. Slow cooling permits extensive large crystal growth under plutonic magmatic crystallization. As melting and crystallization are reversible processes, different types of phase diagrams are needed to understand how melts crystallize. Magmatic liquids have a sub-solidus structure which is created largely by varying the degrees of bonding among SiO4 tetrahedra (polymerization). The difference between a silicate liquid and a silicate mineral is that the mineral has a definite long-range order structure that is the same throughout while a silicate melt shows different types of the short-range order of polymerization throughout the melt. Additionally, the degree of the polymerization of the melt controls the viscosity of the melt. Mafic magmas that have relatively low Si contents have depolymerized melts that produce depolymerized crystals - like olivine or pyroxene - and these melts tend to have a high temperature and a low viscosity. In contrast, felsic magmas have abundant Si and Al, whereby the melts are highly polymerized owing to their high viscosity such that they crystallize as sheet and framework silicates. Because of these consequences, Mg-Fe-rich basaltic magmas erupt easily and the lavas flow several kilometres from their vents, while the Si-rich rhyolite lava erupts explosively without escaping out of its inherent gas bubbles. Volatiles are elements that dissolve in magmas but transform to gas when magma crystallizes or because of a sudden decrease in pressure. The eolution of magmatic rocks may be traced out based on the concept of "Bowen's Reaction Series". The order of crystallization of common plagioclase feldspars from cooling magma is evolved from Carich to more Na-rich plagioclase feldspars as a continuous series during decreasing temperature. The high temperature olivine would react with residual magma and change to the next mineral pyroxene in the series. Pyroxene continues to cooling, it would convert to amphibole and then biotite by adjusting the mineral crystalline lattice to achieve stability of different temperatures as discrete minerals. At lower temperatures both continuous and discrete series merge together and orthoclase, muscovite and quartz tend to crystallize. The difference in crystallization temperatures for different kinds of minerals plays a major role in the differentiation of rock composition as magma cools. Thus, various types of magmatic rocks from the early formed ultamafic peridotites to granitic rocks at the late magmatic stages are evolved (Bowen, 1956). Bowen's reaction series show that one homogeneous body of magma can form more than one kind of igneous rock. It reveals the relationship between

remaining liquid. They increase the fluidity of the magma. Moreover, they do not enter appreciably into the earlier minerals forming from the magma, but are instead finally concentrated in an aqueous solution as hydrothermal fluid. Minerals initially crystallized from the magma are denser than the magma and buoyancy forces lead to gravity settling the crystals. In some rare and denser iron-rich or carbonatite magmas, feldspar crystals may float. The viscosity of silicate magma generally increases with a decrease in the temperature - with decreases in dissolved H2O - with an increase in crystal fraction and an increase in the degree of SiO2 content.

## **2. Polymerization**

The polymerization of silicate melt takes place by the reaction of a silicate melt to form three dimensional networks of SiO4 chains. This structure of a silicate melt provides information on its structure and physical, chemical and thermal properties. From this information, it is possible to understand the conditions of crystallization and the evolution of the silicate minerals from the melt /magma. In magmatic processes, phase equilibria in melt-crystalvapour systems, diffusion in melts and the thermodynamic, electrical and rheological properties of magma systems are of particular importance. Silicate minerals essentially contain silicon and oxygen in the form of a network of tetrahedra in which a Si ion is surrounded by 4 oxygen ions. The tetrahedra occur as isolated (SiO4)4- or else joined together in various ways, such as (Si2O7)6- and rings (Si6O18)12-. Ortho-silicates have isolated (SiO4)4- tetrahedra that are connected only by interstitial cations. Sorosilicates have isolated double tetrahedral groups with (Si2O7)8-. Cyclosilicates have linked tetrahedra with (SixO3x)2x-. Inosilicates have interlocking chains of silicate tetrahedra with either SiO3 for single chains or Si4O11 for double chains. Sheet-silicates form parallel sheets of silicate tetrahedra with Si2O5. Tectosilicates have a three-dimensional framework of silicate tetrahedra with SiO2. As a silicate melt cools, minerals crystallize that are in equilibrium with the melt. Ortho-silicates like olivine generally crystallize from a silicate melt at a very high temperature in a disordered state; on the other hand, quartz and alkali feldspar crystallize from the melt at a relatively low temperature state. At the ends of the scale, the structural states of silicate melts vary widely. Similarly, within the crystallizing minerals, the olivine crystallizing at the early stage of magmatic evolution and the quartz crystallizing at very high temperature states at the late magmatic stage are deficient in Si ions in their tetrahedral co-ordinations. Thus, the structural states of co-existing minerals and the silicate melt from which they crystallize vary widely. In studying such variations, it is possible to see the trend of magmatic evolution. Magmatic evolution may reveal the cooling history of minerals and their accompanying rocks. Volcanic rock which has been suddenly quenched also reveals genetic relationships between phenocrysts and their groundmass matrix, which represents a quenched silicatemelt. There exists a genetic relationship between the structure of a silicate melt and the structure of the silicate mineral crystallizing stemming from it. For example, under an increase of alkalis and alkaline earth elements in the silicate melt, tetrahedrally co-ordinated Al3+ displays a strong preference for three-dimensional network units in the form of feldspars in the silicate melts with an increase of the Al/(Al+Si) of the melt and a decrease of SiO44- units in the melt (Mysen et al., 1985). The structure of a silicate melt depends upon the viscosity of the melt and the subsolidus structure of the crystallizing silicate mineral at a high temperature state. Under these favourable conditions, the crystalmelt equilibrium of the partition of one element between them sets. Magmatic evolution can be traced by studying the changes of the structural state of crystallizing minerals and the changes in structure of residual melts in spatial and temporal conditions. Slow cooling permits extensive large crystal growth under plutonic magmatic crystallization. As melting and crystallization are reversible processes, different types of phase diagrams are needed to understand how melts crystallize. Magmatic liquids have a sub-solidus structure which is created largely by varying the degrees of bonding among SiO4 tetrahedra (polymerization). The difference between a silicate liquid and a silicate mineral is that the mineral has a definite long-range order structure that is the same throughout while a silicate melt shows different types of the short-range order of polymerization throughout the melt. Additionally, the degree of the polymerization of the melt controls the viscosity of the melt. Mafic magmas that have relatively low Si contents have depolymerized melts that produce depolymerized crystals - like olivine or pyroxene - and these melts tend to have a high temperature and a low viscosity. In contrast, felsic magmas have abundant Si and Al, whereby the melts are highly polymerized owing to their high viscosity such that they crystallize as sheet and framework silicates. Because of these consequences, Mg-Fe-rich basaltic magmas erupt easily and the lavas flow several kilometres from their vents, while the Si-rich rhyolite lava erupts explosively without escaping out of its inherent gas bubbles. Volatiles are elements that dissolve in magmas but transform to gas when magma crystallizes or because of a sudden decrease in pressure. The eolution of magmatic rocks may be traced out based on the concept of "Bowen's Reaction Series". The order of crystallization of common plagioclase feldspars from cooling magma is evolved from Carich to more Na-rich plagioclase feldspars as a continuous series during decreasing temperature. The high temperature olivine would react with residual magma and change to the next mineral pyroxene in the series. Pyroxene continues to cooling, it would convert to amphibole and then biotite by adjusting the mineral crystalline lattice to achieve stability of different temperatures as discrete minerals. At lower temperatures both continuous and discrete series merge together and orthoclase, muscovite and quartz tend to crystallize. The difference in crystallization temperatures for different kinds of minerals plays a major role in the differentiation of rock composition as magma cools. Thus, various types of magmatic rocks from the early formed ultamafic peridotites to granitic rocks at the late magmatic stages are evolved (Bowen, 1956). Bowen's reaction series show that one homogeneous body of magma can form more than one kind of igneous rock. It reveals the relationship between cooling magma and the formation of minerals that make up igneous rocks.

Some generalizations:

536 Crystallization – Science and Technology

remaining liquid. They increase the fluidity of the magma. Moreover, they do not enter appreciably into the earlier minerals forming from the magma, but are instead finally concentrated in an aqueous solution as hydrothermal fluid. Minerals initially crystallized from the magma are denser than the magma and buoyancy forces lead to gravity settling the crystals. In some rare and denser iron-rich or carbonatite magmas, feldspar crystals may float. The viscosity of silicate magma generally increases with a decrease in the temperature - with decreases in dissolved H2O - with an increase in crystal fraction and an increase in the

The polymerization of silicate melt takes place by the reaction of a silicate melt to form three dimensional networks of SiO4 chains. This structure of a silicate melt provides information on its structure and physical, chemical and thermal properties. From this information, it is possible to understand the conditions of crystallization and the evolution of the silicate minerals from the melt /magma. In magmatic processes, phase equilibria in melt-crystalvapour systems, diffusion in melts and the thermodynamic, electrical and rheological properties of magma systems are of particular importance. Silicate minerals essentially contain silicon and oxygen in the form of a network of tetrahedra in which a Si ion is surrounded by 4 oxygen ions. The tetrahedra occur as isolated (SiO4)4- or else joined together in various ways, such as (Si2O7)6- and rings (Si6O18)12-. Ortho-silicates have isolated (SiO4)4- tetrahedra that are connected only by interstitial cations. Sorosilicates have isolated double tetrahedral groups with (Si2O7)8-. Cyclosilicates have linked tetrahedra with (SixO3x)2x-. Inosilicates have interlocking chains of silicate tetrahedra with either SiO3 for single chains or Si4O11 for double chains. Sheet-silicates form parallel sheets of silicate tetrahedra with Si2O5. Tectosilicates have a three-dimensional framework of silicate tetrahedra with SiO2. As a silicate melt cools, minerals crystallize that are in equilibrium with the melt. Ortho-silicates like olivine generally crystallize from a silicate melt at a very high temperature in a disordered state; on the other hand, quartz and alkali feldspar crystallize from the melt at a relatively low temperature state. At the ends of the scale, the structural states of silicate melts vary widely. Similarly, within the crystallizing minerals, the olivine crystallizing at the early stage of magmatic evolution and the quartz crystallizing at very high temperature states at the late magmatic stage are deficient in Si ions in their tetrahedral co-ordinations. Thus, the structural states of co-existing minerals and the silicate melt from which they crystallize vary widely. In studying such variations, it is possible to see the trend of magmatic evolution. Magmatic evolution may reveal the cooling history of minerals and their accompanying rocks. Volcanic rock which has been suddenly quenched also reveals genetic relationships between phenocrysts and their groundmass matrix, which represents a quenched silicatemelt. There exists a genetic relationship between the structure of a silicate melt and the structure of the silicate mineral crystallizing stemming from it. For example, under an increase of alkalis and alkaline earth elements in the silicate melt, tetrahedrally co-ordinated Al3+ displays a strong preference for three-dimensional network units in the form of feldspars in the silicate melts with an increase of the Al/(Al+Si) of the melt and a decrease of SiO44- units in the melt (Mysen et al., 1985). The structure of a silicate melt depends upon the viscosity of the melt and the subsolidus structure of the crystallizing silicate mineral at a high temperature state. Under these favourable conditions, the crystalmelt equilibrium of the partition of one element between them sets. Magmatic evolution can be traced by studying the changes of the structural state of crystallizing minerals and the

degree of SiO2 content.

**2. Polymerization** 


Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

magmatic evolution of co-magmatic series of rocks.

**3. Magmatic melt structure** 

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 539

alumina poor or silica under-saturated rocks (Barth, 1962) or silica under-saturated alkalialuminium silicates (Schairer and Yoder, 1960). If the crystal's settling and fractionation are irregular, the composition of the residual magma continuously changes over the course of time. The early crystallized and fractionated silicate or non-silicate mineral from the magma controls the chemical composition of the residual melt and its subsequent trend of crystallization. The composition of the parent magma and its volatile constituents decide which of the minerals start the initial crystallization. Only by studying the field relationship, petrographic features and compositions of the minerals in the rocks and their mode of occurrence and geological setting, along with experimental results, is it possible to trace the

The kinetic properties of magma depend upon the atomic mobility within variably polymerized silicate melts. Most polymerized silicic crystal fractions affect the properties of melts by resisting shear deformations. They provide bases for polymerized links to attach in melts' rich fluids and by providing porous networks for the melts and the chemical species within them to flow or diffuse from one location to another. Crystals with large interfacial angles (>60 degrees) represent a near collapse of the grain-supported structure, which can act to either press out or lock melt fractions from one another within a crystal – a melt mush at lower temperatures. Therefore, it may also be impossible to initiate the types of grain boundary movements continuously needed for both minerals in new solutions. This may have an impact on grain compositions, resulting in the zoning of the crystals that form silicic melts becoming more viscous. The magmatic melt structure is viewed as linkages of Si 4+ and Al 3+ ions in tetrahedral coordination with O2-ions. They are polymerized in the more Si-Al rich melts. Dissolved water breaks the Si-O chains and reduces the degree of polymerization and makes it less viscous. The presence of CO2 rich fluid inclusions in calcites, feldspars, pyroxenes and olivines (Ramasamy and Shapenko, 1980) could be considered to be vapour trapped phases during the growth of these minerals. In addition to H2O, the vapour phases of CO2, SO3, and P2O5 play a critical role in the viscosity of the melt. The formation of crystal nucleus and crystal growth or the accretion of atoms onto the nucleus may take place in two or more stages, depending upon the kinetic conditions of the magma. Microlites are relatively associated with higher energies of surface tension and surface energies than their associated phenocrysts. The rates of growth increase with the increasing of undercooling up to a maximum value and then diminish. High nuclei populations yield fine-grained rocks. The close associations of both fine-grained aplitic syenites and coarse-grained pegmatitic syenites with subvolcanic miarolitic crystallizations indicate wide variations in the textural pattern of the alkaline rocks in the area under study. The drastic cooling of highly viscous melts produces amorphous glass. Vitrophyric textures are produced by rapid cooling while emplacement occurs in some dolerites in this area. Complex minerals with solid solution relationships often grow when their growth rates are faster than their diffusion rates. The diffusion of chemical constituents is due to variations in the thermal energy in the magmatic chamber. Zoned plagioclases, clinopyroxenes, zoned alkaline rocks and zoned carbonatites are evolved in this manner. An increase of volatile proportions in late magmatic melts produces boiling, exsolution and / or vesiculation in the residual rocks. The sizes of comagmatic bodies and their volumes during fractionation in sequences and subsequent emplacements also play vital roles in evolution of melt structures at various levels of emplacements as well as in


The most common magma which is rich in silica is highly polymerized and has a low activity of oxygen ions and a basic magma rich in MgO and CaO has a high activity of oxygen ions. During the course of crystallization, the most abundant, least soluble mineral grows first followed by the less abundant, more soluble substances from the magma. During the course of crystallization, silicate magma displays progressive enrichments of both soda relative to lime and ferrous iron relative to magnesia in crystals and residual melts. Slowly ascending very hot magmas may cool before reaching the surface.

The crystallization of minerals from magma is a complex process because of changes in the temperature, pressure and composition (TPX). The TPX can have dramatic effects on the order of crystallization and the order-disorder structural state of various minerals in a sequence. The addition or loss of water, CO2, H2 and O2 changes the course of magmatic differentiation. The initial crystallization and separation of early-formed minerals from the residual liquid plays a critical role in the trend of magmatic evolution. The general order of crystallization from the magma is as follows: a) low silica containing basic minerals, such as olivine and calcic plagioclase, together with non-silicate minerals, b) medium-silica-bearing minerals, such as clinopyroxene or labradorite, c) high silica-bearing minerals, such as orthoclase and quartz. Depending on the initial chemical composition of the common basaltic magma, either plagioclase or pyroxene may start to crystallize first, changing the composition of the residual magma until simultaneous crystallization occurs. Slight differences in the composition of the parent basaltic magma may result in the production of an end-magma which is either over-saturated or under-saturated in silica (Barth, 1962). The early crystallization and fractionation of olivine / plagioclase / diopside, the respective residual liquid, is progressively enriched in silica over-saturated rocks (Bowen, 1956) or else alumina poor or silica under-saturated rocks (Barth, 1962) or silica under-saturated alkalialuminium silicates (Schairer and Yoder, 1960). If the crystal's settling and fractionation are irregular, the composition of the residual magma continuously changes over the course of time. The early crystallized and fractionated silicate or non-silicate mineral from the magma controls the chemical composition of the residual melt and its subsequent trend of crystallization. The composition of the parent magma and its volatile constituents decide which of the minerals start the initial crystallization. Only by studying the field relationship, petrographic features and compositions of the minerals in the rocks and their mode of occurrence and geological setting, along with experimental results, is it possible to trace the magmatic evolution of co-magmatic series of rocks.

## **3. Magmatic melt structure**

538 Crystallization – Science and Technology

7. Intermediate minerals that crystallize at intermediate temperatures are intermediate in

8. Another important observation in that those minerals which crystallize at high temperatures show less polymerization of the Si-tetrahedra than those which crystallize

9. The discrete mineral sequences starts with orthosilicates (olivine), single chain (pyroxene), double chain (amphibole) and sheet silicates (biotite), and exhibit a high degree of silica polymerization with incorporations of H2O, F and other volatiles. The crystallization of plagioclase in a continuous series also indicates an increase of silica

10. Highly polymerized silica-rich tectosilicates are more stable than silica-poor tectosilicates.

12. The initial crystallization of pyroxene under dry conditions, the residual liquid, is

13. The initial crystallization of alkali pyroxene, alkali amphibole, biotite or magnetite, from a silica undersaturated volatile-rich basic alkali magma, residual magma slowly gets enriched with silica and finally at the late magmatic stages free quartz crystallizes in

14. Rock textures are controlled by conditions of crystallization and cooling. Studying rock textures and the mode of the occurrence of rock types in the field, it is possible to

The most common magma which is rich in silica is highly polymerized and has a low activity of oxygen ions and a basic magma rich in MgO and CaO has a high activity of oxygen ions. During the course of crystallization, the most abundant, least soluble mineral grows first followed by the less abundant, more soluble substances from the magma. During the course of crystallization, silicate magma displays progressive enrichments of both soda relative to lime and ferrous iron relative to magnesia in crystals and residual melts. Slowly

The crystallization of minerals from magma is a complex process because of changes in the temperature, pressure and composition (TPX). The TPX can have dramatic effects on the order of crystallization and the order-disorder structural state of various minerals in a sequence. The addition or loss of water, CO2, H2 and O2 changes the course of magmatic differentiation. The initial crystallization and separation of early-formed minerals from the residual liquid plays a critical role in the trend of magmatic evolution. The general order of crystallization from the magma is as follows: a) low silica containing basic minerals, such as olivine and calcic plagioclase, together with non-silicate minerals, b) medium-silica-bearing minerals, such as clinopyroxene or labradorite, c) high silica-bearing minerals, such as orthoclase and quartz. Depending on the initial chemical composition of the common basaltic magma, either plagioclase or pyroxene may start to crystallize first, changing the composition of the residual magma until simultaneous crystallization occurs. Slight differences in the composition of the parent basaltic magma may result in the production of an end-magma which is either over-saturated or under-saturated in silica (Barth, 1962). The early crystallization and fractionation of olivine / plagioclase / diopside, the respective residual liquid, is progressively enriched in silica over-saturated rocks (Bowen, 1956) or else

6. Minerals that crystallize at the lowest temperatures are rich in Si and Al.

constituents, even though they crystallize in framework silicates.

ascending very hot magmas may cool before reaching the surface.

The degree of polymerization increases from orthosilicates to tectosilicates. 11. The initial crystallization of olivine, the residual magma, is enriched in silica.

relation to Si.

deficient silica.

significant amount as an end product.

correlate rocks and minerals.

at lower temperatures.

The kinetic properties of magma depend upon the atomic mobility within variably polymerized silicate melts. Most polymerized silicic crystal fractions affect the properties of melts by resisting shear deformations. They provide bases for polymerized links to attach in melts' rich fluids and by providing porous networks for the melts and the chemical species within them to flow or diffuse from one location to another. Crystals with large interfacial angles (>60 degrees) represent a near collapse of the grain-supported structure, which can act to either press out or lock melt fractions from one another within a crystal – a melt mush at lower temperatures. Therefore, it may also be impossible to initiate the types of grain boundary movements continuously needed for both minerals in new solutions. This may have an impact on grain compositions, resulting in the zoning of the crystals that form silicic melts becoming more viscous. The magmatic melt structure is viewed as linkages of Si 4+ and Al 3+ ions in tetrahedral coordination with O2-ions. They are polymerized in the more Si-Al rich melts. Dissolved water breaks the Si-O chains and reduces the degree of polymerization and makes it less viscous. The presence of CO2 rich fluid inclusions in calcites, feldspars, pyroxenes and olivines (Ramasamy and Shapenko, 1980) could be considered to be vapour trapped phases during the growth of these minerals. In addition to H2O, the vapour phases of CO2, SO3, and P2O5 play a critical role in the viscosity of the melt.

The formation of crystal nucleus and crystal growth or the accretion of atoms onto the nucleus may take place in two or more stages, depending upon the kinetic conditions of the magma. Microlites are relatively associated with higher energies of surface tension and surface energies than their associated phenocrysts. The rates of growth increase with the increasing of undercooling up to a maximum value and then diminish. High nuclei populations yield fine-grained rocks. The close associations of both fine-grained aplitic syenites and coarse-grained pegmatitic syenites with subvolcanic miarolitic crystallizations indicate wide variations in the textural pattern of the alkaline rocks in the area under study. The drastic cooling of highly viscous melts produces amorphous glass. Vitrophyric textures are produced by rapid cooling while emplacement occurs in some dolerites in this area. Complex minerals with solid solution relationships often grow when their growth rates are faster than their diffusion rates. The diffusion of chemical constituents is due to variations in the thermal energy in the magmatic chamber. Zoned plagioclases, clinopyroxenes, zoned alkaline rocks and zoned carbonatites are evolved in this manner. An increase of volatile proportions in late magmatic melts produces boiling, exsolution and / or vesiculation in the residual rocks. The sizes of comagmatic bodies and their volumes during fractionation in sequences and subsequent emplacements also play vital roles in evolution of melt structures at various levels of emplacements as well as in

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

Fig. 1. b. Structural sketch map

**5. Methodology** 

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 541

Fig. 1. c. Parent magma ascending from a deep-seated source filled in a magmatic chamber

Several fieldtrips and observations were made during the course of geological mapping. Thin sections were prepared for various rocks collected in the field and studied under a universal stage (Naidu, 1958) attached polarizing microscope for the identification and

determination of the volume proportions of the minerals.

residual magmas. In studying the order / disorder relationships and partitioning elements, it is possible to trace the cooling history of the magma.

## **4. Carbonatite complex**

The carbonatite complex of Tiruppattur (12o 00'00"-12o30'00"N and 78o25'00"-78o35'00"E) Tamil Nadu India is an ideal area to trace magmatic evolution and its magmatic zonal variations (Fig. 1). It is unique in its geological settings, in the fractional crystallization of various minerals crystallized during the course of a prolonged period of magmatic differentiation, and in the emplacement of co-magmatic zoned alkaline rocks and the immiscible separation of carbonatites and alkali syenites from common parent magma. The geological field settings present in this area indicate the occurrences of co-magmatic sequences of both silica undersaturated series of rocks and silica oversaturated series of rocks together. Both these series of rocks have co-magmatic relationships and continuous compositional variations among mineral assemblages. Progressive enrichment of alkali, silica and volatile constituents plays a critical role during the course of magmatic evolution from a highly silica undersaturated shonkinite magma to silica oversaturated alkali syenite or granite magma under favourable tectonic environment..

Fig. 1. a. Geological map of the carbonatite complex of Tiruppattur, India

## **5. Methodology**

540 Crystallization – Science and Technology

residual magmas. In studying the order / disorder relationships and partitioning elements, it

The carbonatite complex of Tiruppattur (12o 00'00"-12o30'00"N and 78o25'00"-78o35'00"E) Tamil Nadu India is an ideal area to trace magmatic evolution and its magmatic zonal variations (Fig. 1). It is unique in its geological settings, in the fractional crystallization of various minerals crystallized during the course of a prolonged period of magmatic differentiation, and in the emplacement of co-magmatic zoned alkaline rocks and the immiscible separation of carbonatites and alkali syenites from common parent magma. The geological field settings present in this area indicate the occurrences of co-magmatic sequences of both silica undersaturated series of rocks and silica oversaturated series of rocks together. Both these series of rocks have co-magmatic relationships and continuous compositional variations among mineral assemblages. Progressive enrichment of alkali, silica and volatile constituents plays a critical role during the course of magmatic evolution from a highly silica undersaturated shonkinite magma to silica oversaturated alkali syenite

> 1. Carbonatite, 2. Shonkinite, 3. Vogesite, 4. Oligoclasite, 5. Hornblende syenite, 6. Hornblende porphyritic

syenite, 7. Augite syenite,

16. Skarn, 17. Granite 18. Granodiorite,

28. Norite, 29. Granite gneiss

9. Foliated syenite,

8. Porphyritic augite syenite,

10. Foliated porphyritic syenite 11. Hornblede biotite syenite, 12. Riebeckite syenite, 13. Garnetiferous syenite, 14. Microcline garnet syenite, 15. Wollastonite syenite,

19. Biotite oligoclase pyroxenite

20. Biotite pyroxenite, 21. Ultrabasics 22. Ultramafics, 23. Dolerite, 24. Pegmatite-aplite 25. Albitite 26. Charnockite, 27. Pyroxene granulite,

is possible to trace the cooling history of the magma.

or granite magma under favourable tectonic environment..

**Sevvattur** 

Fig. 1. a. Geological map of the carbonatite complex of Tiruppattur, India

**4. Carbonatite complex** 

Several fieldtrips and observations were made during the course of geological mapping. Thin sections were prepared for various rocks collected in the field and studied under a universal stage (Naidu, 1958) attached polarizing microscope for the identification and determination of the volume proportions of the minerals.

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

iv. The two basins are enclosed with ultrabasic rocks

Science Laboratory, Indian Institute of Technology Madras, Chennai.

Sevvattur Basin

**6. Structure** 

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 543

i. Differentiation and fractionation of the parent magma for a quiescent prolonged period ii. Emplacements of a series of soda-rich comagmatic syenites and carbonatites in the

iii. Shifting of the magmatic emplacements of potash rich co-magmatic alkali series and carbonatites into the continuously uplifted and rotated Jogipatti Basin towards the SW

By measuring the relationships between the crystallographic and optic axes of the chemical compositions of the minerals and their order-disorder, the structural features are estimated. The textural and structural features of the mineral assemblages, the order of crystallization of the crystallization of the minerals and their magmatic evolutions are traced. The composition of feldspar grains, Fe-Mg ratios in olivine, pyroxene, amphiboles and biotite were calculated by measuring the refractive indices and by finding the optical properties and using the appropriate nomograms given in the texts (Winchell, 1945; Deer et al., 1965). X-ray powder diffraction analyses were made to study the order-disorder relationship by measuring the triclinicity of the feldspars and the M2, M1 and T sites in the clinopyroxenes and the M4 sites in the amphiboles. Staining techniques were adopted to distinguish calcite and ferro carbonates from dolomite, both in the laboratory as well as in the field (Dickinson, 1965). The chemical compositions of rocks were given by the conventional gravimetric analyses (author in the Geology Dept. Presidency College Chennai and Petrochemical Laboratory, Geology Faculty, Moscow State University) XRF and EDAX under a high resolution scanning electron microscope (IITM, Chennai) . The trace elements contents were determined by Atomic Absorption Spectrometry, Quartz –Grating Spectrograph and HR EDAX. These instrumental facilities were availed by the Department of Geology, Presidency College, Chennai, the Petrography Department, Faculty of Geology, Moscow State University, Russia, the Petrology laboratory, Department of Geology, Government of Tamil Nadu, Chennai and the Material

The carbonatite complex is located in Tamil Nadu, India in a NE-SW trending rift valley formed between the Javadi Hills and the Elagiri Hills (Fig. 1 b), which are located within 5 km of one another. According to Grady (1971), the rift valley, bounded by fault planes, belongs to the category of the deep main faults of Peninsular India, extending more than 200 km in distance. Before rifting, the region was folded, faulted and uplifted by major tectonic deformations, trending with fold-axial plunges in the NW, N, NE, ENE, E and ESE directions. Owing to the number of re-worked fault systems occurring, since the Archaeans, the area has been dissected into several segments of block faults with rotational movements. Hence, it is very hard to analyse the age relationships of different fold styles. However, the regional pattern of fault systems present is seen around the carbonatite complex of Tiruppattur. The culmination of a NE-SW fold system caused the coeval and collinear formation of the Sevvattur and Jogipatti basins by the refolding of ultramafic, basic granulite, charnockite and granite gneiss (Fig. 1b & 1c). The regional NE-SW and N-S fault systems caused formation of a graben structure (Grady, 1971). Due to block faulting, the Jogipatti block was moved towards the west and then towards the north. The melting of the mantle rocks occurred at depth due to E-W crustal shear, which later provides an access route for the rising magma from a mantle source with intermittent stopping in a cavity, forming a closed magmatic chamber at depth. After a prolonged period of magmatic crystallization, differentiation and fractionation with the immiscible separation of

Fig. 1. d. The partial melting of rocks of different compositions concentrates in some large cavities of magmatic chambers. Hot magma rising from the feeder chamber mixes with the already existing magma at a relatively lower temperature and turbulently circulates inside the magmatic chamber. Some of early-formed crystals stick to the walls. The pre-existing magma may dissolve with the incoming magma and mixes with it. The undissolved immiscible magma of different compositions may settle at certain portions. Fig. 1e Magma ascending through fractures breaks the wall rock into pieces of rock fragments which are assimilated by the magma while the magma changes its composition.


By measuring the relationships between the crystallographic and optic axes of the chemical compositions of the minerals and their order-disorder, the structural features are estimated. The textural and structural features of the mineral assemblages, the order of crystallization of the crystallization of the minerals and their magmatic evolutions are traced. The composition of feldspar grains, Fe-Mg ratios in olivine, pyroxene, amphiboles and biotite were calculated by measuring the refractive indices and by finding the optical properties and using the appropriate nomograms given in the texts (Winchell, 1945; Deer et al., 1965). X-ray powder diffraction analyses were made to study the order-disorder relationship by measuring the triclinicity of the feldspars and the M2, M1 and T sites in the clinopyroxenes and the M4 sites in the amphiboles. Staining techniques were adopted to distinguish calcite and ferro carbonates from dolomite, both in the laboratory as well as in the field (Dickinson, 1965). The chemical compositions of rocks were given by the conventional gravimetric analyses (author in the Geology Dept. Presidency College Chennai and Petrochemical Laboratory, Geology Faculty, Moscow State University) XRF and EDAX under a high resolution scanning electron microscope (IITM, Chennai) . The trace elements contents were determined by Atomic Absorption Spectrometry, Quartz –Grating Spectrograph and HR EDAX. These instrumental facilities were availed by the Department of Geology, Presidency College, Chennai, the Petrography Department, Faculty of Geology, Moscow State University, Russia, the Petrology laboratory, Department of Geology, Government of Tamil Nadu, Chennai and the Material Science Laboratory, Indian Institute of Technology Madras, Chennai.

## **6. Structure**

542 Crystallization – Science and Technology

Fig. 1. d. The partial melting of rocks of different compositions concentrates in some large cavities of magmatic chambers. Hot magma rising from the feeder chamber mixes with the already existing magma at a relatively lower temperature and turbulently circulates inside the magmatic chamber. Some of early-formed crystals stick to the walls. The pre-existing magma may dissolve with the incoming magma and mixes with it. The undissolved immiscible magma of different compositions may settle at certain portions. Fig. 1e Magma ascending through fractures breaks the wall rock into pieces of rock fragments which are

assimilated by the magma while the magma changes its composition.

The carbonatite complex is located in Tamil Nadu, India in a NE-SW trending rift valley formed between the Javadi Hills and the Elagiri Hills (Fig. 1 b), which are located within 5 km of one another. According to Grady (1971), the rift valley, bounded by fault planes, belongs to the category of the deep main faults of Peninsular India, extending more than 200 km in distance. Before rifting, the region was folded, faulted and uplifted by major tectonic deformations, trending with fold-axial plunges in the NW, N, NE, ENE, E and ESE directions. Owing to the number of re-worked fault systems occurring, since the Archaeans, the area has been dissected into several segments of block faults with rotational movements. Hence, it is very hard to analyse the age relationships of different fold styles. However, the regional pattern of fault systems present is seen around the carbonatite complex of Tiruppattur. The culmination of a NE-SW fold system caused the coeval and collinear formation of the Sevvattur and Jogipatti basins by the refolding of ultramafic, basic granulite, charnockite and granite gneiss (Fig. 1b & 1c). The regional NE-SW and N-S fault systems caused formation of a graben structure (Grady, 1971). Due to block faulting, the Jogipatti block was moved towards the west and then towards the north. The melting of the mantle rocks occurred at depth due to E-W crustal shear, which later provides an access route for the rising magma from a mantle source with intermittent stopping in a cavity, forming a closed magmatic chamber at depth. After a prolonged period of magmatic crystallization, differentiation and fractionation with the immiscible separation of

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

this rock produces expansion cracks in the surrounding albite. .

Syenite pegmatites and aplites are common in the carbonatite complex. They also occur as sheets along the fracture planes present in the syenites. Numerous anastomosing veins of pegmatites and aplites are seen in the syenites. Small bodies of pegmatites and aplites carry coarse-grained crystals along their borders. The mafic minerals grow towards it as shooting grains from the wall rocks and the core is only composed of felsic minerals. Massive bodies of pegmatites and aplites occupy over 50 km2 in the Jogipatti basin. The northern portion of the syenite complex of this basin is composed of riebeckite-anorthoclase syenite while its southern portion is composed of orthoclase-bearing garnetiferous pegmatitic syenite. Garnetiferous syenite exhibits a hypidiomorphic granular texture. Euhedral to subhedral grains of garnets are embedded in the potash feldspar. The grossularite-andradite garnet gradually transforms into melanite towards the southwest of this region. Melanite-

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 545

acmite-syenite are composed by a significant amount of magnetite and quartz and large plates of acmite are surrounded by the released products of magnetites. Abrupt variations from a porphyritic syenite to a non-porphyritic syenite are also seen. The impregnation of cumulates of feldspar phenocrysts into a syenite body is also seen. Many melanocratic streaks, thin slivers and patches of pyroxenites, are strewn as disseminated or as scattered material in the syenite body as mafic cumulates. The dissemination of mafic minerals is often crowded around mafic xenoliths. It seems to be the case that melanocratic rocks rich in mafic minerals are developed in this manner. Foliated fine-grained syenite occurs as a linear band in the porphyritic syenites in the Sevvattur basin. It exhibits crude foliation by the presence of abundant mica. In thin sections, it displays a xenomorphic granular texture. It is composed of oligoclase, anorthoclase, microcline, biotite, margarite and calcite. The grain size of the feldspar ranges from 0.1 mm to 0.3 mm. Discrete grains of microcline are formed as an interstitial to biotite. The replacement of oligoclase by potash feldspar is found along the cleavage planes. The development of patchy perthite is obvious: it imperceptibly grades into a pink coloured porphyritic syenite with a trachitoid texture. It occurs as a pale pink coloured rock with large plates of potash feldspar and oligoclase set in a fine-grained feldspathic matrix. Most of the feldspar platelets are square shaped with up to 5 cm sides. The breadth of the plates varies between 2 mm to 4 mm. In thin sections, feldspar displays an inequigranular poikilitic texture. Hornblende needles are present as inclusions within the plates of the feldspars and the needles orient in parallel towards the foliation planes of the rock. It is essentially composed of microcline, oligoclase and orthoclase, with accessories of chlorite, biotite, hornblende and epidote. The microcline and oligoclase are formed in two generations. Often, the feldspars are unoriented. They carry inclusions of fine needles of mafic minerals (with a helicitic texture) and show characteristic foliation. The core portion of the Sevvattur syenite body is occupied by oligoclasiteand albitite. The amphibole in the biotiteoligoclasite is altered into epidote granules and the amphibole is impregnated with felsic minerals producing a sieve texture. In some oligoclasites oligoclase and biotite, flakes exhibit a sheath texture with radiating platelets orienting towards a common centre. Along the peripheral portion of the carbonatite, albitite is exposed. In the albitite, the feldspar plates have inclusions of magnetite. The modal composition of potash feldspars varies widely and their content increases towards the core of the basin. Again, the volume percentage of the potash feldspars increases towards the south. Albitite, oligoclasite and biotite-bearing oligoclasites occur adjacent to the carbonatite body. A fine-grained monomineralic rock composed of more than 90% of albite in volume with accessories of magnetite and others exhibits the hypidiomorphic granular texture. Post magmatic growth of lueshite (sodium pyrochlore) in

carbonatite magmas, the sequences of magmatic rocks were emplaced, first, in the Sevvattur basin and the block was then down warped. A forceful southern movement of the Sevvattur block abutting over the Jogipatti block caused an uplift and a rotational effect over which the rotational axis remained normal to the fault plane trending E-W and resting on the NE-SW fault plane. By the geometrical analysis of this fault system (Ramasamy, 1982), it was estimated that the Jogipatti basin was uplifted by more than 1000 m and, apparently, rotated by about 16o towards the east, and that it caused the emplacements of continuous sequences of late co-magmatic rocks. The magmatic activities were shifted to the latter basin from the original feeding source at depth (Fig. 1 b and 1 c). The generalized features of the magmatic activities are described in Fig. 1 d and 1 e.

### **7. Field relationship, mineralogical and textural variations in the alkaline complex**

The carbonatite complex of Tiruppattur (Borodin et al., 1971; Saravanan and Ramasamy, 1971; Ramasamy, 1982) belongs to the Proterozoic Period, being around 800 Ma old (Schieicher et al., 1998). It is emplaced in two adjacent structural basins amidst Pre-Cambrian granite gneisses and charnockites (Fig. 1). Both of the basins are bounded by ultramafic rocks. The outer shell of the northern Sevvattur body is composed of a finegrained speckled hornblende syenite with chilled margins and sharp contact with ultramafics and granite gneisses. Towards the south, the specked hornblende syenite imperceptibly grades into mottled augite syenite. The syenite complex exceeds 30 km2 in extent. These syenites imperceptibly grade inwards towards the centre into their porphyritic syenite variants. The mottled porphyritic syenite is composed of large plates of feldspars containing smaller grains of feldspars in a matrix with accessories of aegirine augite, ferrohastingsite, hornblende and magnetite. Large plates of plagioclase exhibit normal zoning from andesine to sodic oligoclase. The anorthite contents of the plagioclases decrease from the portions from the north to the south and from the peripheral portions towards the centre. The normal zoning and twinning of the plagioclase lamellae can be clearly seen in large platelets, even in hand-specimens. These rocks are essentially composed of alkali feldspars measuring up to 10 cm x 7 cm x 1 cm. The accessory mafic minerals are diopside, hornblende, augite and ferro-hastingsite. Feldspathic phenocrysts are present in the primary foliation of the rock. However, those mafic minerals occurring in the south are slightly coarse-grained, subhedral to euhedral in form, and are characteristically enriched in alkali constituents and potash feldspars. In places, cumulate feldspar phenocrysts with primary flow orientations are seen in some syenite bodies. The plagioclase and potash feldspars are composed of inclusions of plagioclases with complex twinning lamellae belonging to two or three generations and oriented in different directions. Microcline micro-perthite is commonly present in mottled porphyritic syenite. Albite lamellae, as blebs, orient along crystallographic directions. The plagioclase laths are increasingly plate-like and inform towards the centre of the basin. Similarly, the sizes of the mafic minerals increase and their compositions change with increases in alkali and ferric iron contents. Some places among the syenites - owing to the concentration of mafic minerals in felsic rocks or else the mush of feldspar phenocrysts in hypo-melanocratic syenites - form cumulate textures. The flow textures of feldspar phenocrysts and amphiboles are common in most of the syenites and they orient in the same directions. The rocks of the inner portions are successively younger than the rocks of the peripheral portions. The core portions are occupied by highly differentiated magmatic layers derived from deeper levels. Small veins of fine-grained

carbonatite magmas, the sequences of magmatic rocks were emplaced, first, in the Sevvattur basin and the block was then down warped. A forceful southern movement of the Sevvattur block abutting over the Jogipatti block caused an uplift and a rotational effect over which the rotational axis remained normal to the fault plane trending E-W and resting on the NE-SW fault plane. By the geometrical analysis of this fault system (Ramasamy, 1982), it was estimated that the Jogipatti basin was uplifted by more than 1000 m and, apparently, rotated by about 16o towards the east, and that it caused the emplacements of continuous sequences of late co-magmatic rocks. The magmatic activities were shifted to the latter basin from the original feeding source at depth (Fig. 1 b and 1 c). The generalized features of the magmatic activities are described in Fig. 1 d and 1 e.

**7. Field relationship, mineralogical and textural variations in the alkaline** 

The carbonatite complex of Tiruppattur (Borodin et al., 1971; Saravanan and Ramasamy, 1971; Ramasamy, 1982) belongs to the Proterozoic Period, being around 800 Ma old (Schieicher et al., 1998). It is emplaced in two adjacent structural basins amidst Pre-Cambrian granite gneisses and charnockites (Fig. 1). Both of the basins are bounded by ultramafic rocks. The outer shell of the northern Sevvattur body is composed of a finegrained speckled hornblende syenite with chilled margins and sharp contact with ultramafics and granite gneisses. Towards the south, the specked hornblende syenite imperceptibly grades into mottled augite syenite. The syenite complex exceeds 30 km2 in extent. These syenites imperceptibly grade inwards towards the centre into their porphyritic syenite variants. The mottled porphyritic syenite is composed of large plates of feldspars containing smaller grains of feldspars in a matrix with accessories of aegirine augite, ferrohastingsite, hornblende and magnetite. Large plates of plagioclase exhibit normal zoning from andesine to sodic oligoclase. The anorthite contents of the plagioclases decrease from the portions from the north to the south and from the peripheral portions towards the centre. The normal zoning and twinning of the plagioclase lamellae can be clearly seen in large platelets, even in hand-specimens. These rocks are essentially composed of alkali feldspars measuring up to 10 cm x 7 cm x 1 cm. The accessory mafic minerals are diopside, hornblende, augite and ferro-hastingsite. Feldspathic phenocrysts are present in the primary foliation of the rock. However, those mafic minerals occurring in the south are slightly coarse-grained, subhedral to euhedral in form, and are characteristically enriched in alkali constituents and potash feldspars. In places, cumulate feldspar phenocrysts with primary flow orientations are seen in some syenite bodies. The plagioclase and potash feldspars are composed of inclusions of plagioclases with complex twinning lamellae belonging to two or three generations and oriented in different directions. Microcline micro-perthite is commonly present in mottled porphyritic syenite. Albite lamellae, as blebs, orient along crystallographic directions. The plagioclase laths are increasingly plate-like and inform towards the centre of the basin. Similarly, the sizes of the mafic minerals increase and their compositions change with increases in alkali and ferric iron contents. Some places among the syenites - owing to the concentration of mafic minerals in felsic rocks or else the mush of feldspar phenocrysts in hypo-melanocratic syenites - form cumulate textures. The flow textures of feldspar phenocrysts and amphiboles are common in most of the syenites and they orient in the same directions. The rocks of the inner portions are successively younger than the rocks of the peripheral portions. The core portions are occupied by highly differentiated magmatic layers derived from deeper levels. Small veins of fine-grained

**complex** 

acmite-syenite are composed by a significant amount of magnetite and quartz and large plates of acmite are surrounded by the released products of magnetites. Abrupt variations from a porphyritic syenite to a non-porphyritic syenite are also seen. The impregnation of cumulates of feldspar phenocrysts into a syenite body is also seen. Many melanocratic streaks, thin slivers and patches of pyroxenites, are strewn as disseminated or as scattered material in the syenite body as mafic cumulates. The dissemination of mafic minerals is often crowded around mafic xenoliths. It seems to be the case that melanocratic rocks rich in mafic minerals are developed in this manner. Foliated fine-grained syenite occurs as a linear band in the porphyritic syenites in the Sevvattur basin. It exhibits crude foliation by the presence of abundant mica. In thin sections, it displays a xenomorphic granular texture. It is composed of oligoclase, anorthoclase, microcline, biotite, margarite and calcite. The grain size of the feldspar ranges from 0.1 mm to 0.3 mm. Discrete grains of microcline are formed as an interstitial to biotite. The replacement of oligoclase by potash feldspar is found along the cleavage planes. The development of patchy perthite is obvious: it imperceptibly grades into a pink coloured porphyritic syenite with a trachitoid texture. It occurs as a pale pink coloured rock with large plates of potash feldspar and oligoclase set in a fine-grained feldspathic matrix. Most of the feldspar platelets are square shaped with up to 5 cm sides. The breadth of the plates varies between 2 mm to 4 mm. In thin sections, feldspar displays an inequigranular poikilitic texture. Hornblende needles are present as inclusions within the plates of the feldspars and the needles orient in parallel towards the foliation planes of the rock. It is essentially composed of microcline, oligoclase and orthoclase, with accessories of chlorite, biotite, hornblende and epidote. The microcline and oligoclase are formed in two generations. Often, the feldspars are unoriented. They carry inclusions of fine needles of mafic minerals (with a helicitic texture) and show characteristic foliation. The core portion of the Sevvattur syenite body is occupied by oligoclasiteand albitite. The amphibole in the biotiteoligoclasite is altered into epidote granules and the amphibole is impregnated with felsic minerals producing a sieve texture. In some oligoclasites oligoclase and biotite, flakes exhibit a sheath texture with radiating platelets orienting towards a common centre. Along the peripheral portion of the carbonatite, albitite is exposed. In the albitite, the feldspar plates have inclusions of magnetite. The modal composition of potash feldspars varies widely and their content increases towards the core of the basin. Again, the volume percentage of the potash feldspars increases towards the south. Albitite, oligoclasite and biotite-bearing oligoclasites occur adjacent to the carbonatite body. A fine-grained monomineralic rock composed of more than 90% of albite in volume with accessories of magnetite and others exhibits the hypidiomorphic granular texture. Post magmatic growth of lueshite (sodium pyrochlore) in this rock produces expansion cracks in the surrounding albite. .

Syenite pegmatites and aplites are common in the carbonatite complex. They also occur as sheets along the fracture planes present in the syenites. Numerous anastomosing veins of pegmatites and aplites are seen in the syenites. Small bodies of pegmatites and aplites carry coarse-grained crystals along their borders. The mafic minerals grow towards it as shooting grains from the wall rocks and the core is only composed of felsic minerals. Massive bodies of pegmatites and aplites occupy over 50 km2 in the Jogipatti basin. The northern portion of the syenite complex of this basin is composed of riebeckite-anorthoclase syenite while its southern portion is composed of orthoclase-bearing garnetiferous pegmatitic syenite. Garnetiferous syenite exhibits a hypidiomorphic granular texture. Euhedral to subhedral grains of garnets are embedded in the potash feldspar. The grossularite-andradite garnet gradually transforms into melanite towards the southwest of this region. Melanite-

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

with the adjoining plutons.

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 547

acmite-bearing syenite grades with an enrichment of cumulate needles of katophorite into lamprophyric vogesite, displaying a pandiomorphic texture. The shonkinite occurring in Jogipatti basin is a coarse-grained inequigranular rock composed of equal proportions of sanidine / anorthoclase and augite, with accessories of olivine, hastingsite, phlogopite, apatite and magnetite and exhibiting a lamprophyric pandiomorphic texture. In the field, it occurs as xenoliths or nodules in ultramafic rocks. The ultramafic rock which is the host rock of the shonkinite appears to be possessed of kimberlitic affinity (Ramasamy et al., 2010). Mineralogical gradation exists between the shonkinite and garnetiferous syenite. An abrupt mineralogical and compositional gradation exists between the vogesite and shonkinite. Similarly, there exists such gradation between vogesite and hornblende-biotite oligoclasites. Again, the shonkinite exhibits such a type of gradation between the garnetiferous-orthoclase syenite and the mottled augite syenite. On the whole, one syenitic member has genetic relationship with any other member in this zoned complex, either in mineralogical or compositional gradations. In a detailed petrographic and mineralogical study, it is shown that the shonkinite is considered as a parent magma for this carbonatite alkali-syenite complex (Saravanan and Ramasamy, 1995; Ramasamy et al., 2010). Dykes, ring dykes, sills, veins, cavity fillings, cone sheets and plugs of syenites and carbonatites show sharp contacts with chilled margins with their host rocks. Curvilinear tensional cracks are seen around the exposures of the syenites and carbonatites. Late magmatic aplites and pegmatites filled these cracks and show co-magmatic relationships

All of the carbonatites occurring in the three places in Sevvattur, Jogipatti and Onnakarai grade with zonal variation from ferro-carbonatite, beforsite and sovite. Calcite, dolomite and ankerite constitute the essential minerals in these rocks and biotite, phlogopite, magnetite, riebeckite, aegirine augite, acmite, wollastonite, garnet and the rare-earth minerals of pyrochlore, zircon and niobian rutile constitute the notable accessory minerals in these rocks. Two generations of magnetites, both in the form of opaque dusts and anhedral, euhedral and twinned crystals, are present in some of the carbonatites and in the shonkinite. The Fe- and Mg-rich carbonatites appear to be emplaced from relatively deeper horizons. By their characteristic minerals, the zoned cone-sheets of the carbonatites in Sevvattur exhibit various sheets and layers as hybrid sovite, biotite sovite, phlogopite sovite, aegirine sovite, pyrochlore-bearing para-ankeritic sovite, pyrochlore-rich para-ankeritic beforsite, magnetite beforsite and apatite soviet. The elongated rings and veins of the riebeckite-carbonatites are seen in the riebeckite syenites of the Jogipatti basin. The hybrid carbonatites of the Jogipatti basin are enriched with richterite, magnesio-arfvedsonite, acmite, wollastonite and scapolite. The carbonatites of Onnakarai appear to be younger than the Jogipatti carbonatite, which is intermediate and younger itself than the Sevvattur carbonatite (Ramasamy, 1973). Again, the zoned alkali syenites of Sevvattur are older than those in the Jogipatti basin. A carbonatite body crops out in arcuate cone-sheets with a sharp contact along the meeting between the ultramafic rocks and the syenites in the northern periphery of the Sevvattur body. Mineralogical variations occur as zones which are almost conformable with one another, starting from the convex side of the arc and proceeding to the concave side in the following order: a) micaceous zone, b) ankerite zone, c) beforsite zone and d) sovite zone. About 400 m west of the carbonatite exposure is evidence of the forceful injection of carbonatite, as manifested in a syenitic outcrop which has been thoroughly brecciated with feldspar grains which were twisted and ragged with the emplacement of fine-grained carbonatitic cementing media. Along the contacts of the carbonatite, the host rocks have

microcline syenite, wollastonite syenite, wollastonite-scapolite syenite and wollastonitemelanite carbonatite are some of the aplitic veins found in this complex. Fine-grained microcline melanite syenite displays a xenomorphic granular texture. Micro-grains of microcline are surrounded by accessories of anhedral albite and melanite. These microclines and anorthoclases are highly disordered feldspars, having significant trace amounts of ferric iron. The biotite-syenite has a xenomorphic granular texture with euhedral to subhedral biotite. The biotite syenite porphyry is composed of large phenocrysts of anorthoclase, displaying zonal variation from the core to the periphery and set in a micro-granular feldspathic matrix. Such syenite porphyries are also composed of potash feldspar phenocrysts with a cryptoperthitic texture at the core and with homogeneous peripheral portions. The feldspars belong to two or more generations and they occur as large plates of phenocrysts as well as very small anhedral grains in the matrix. Younger aplites trend in N85o crosscuts while older aplites trend in the N45o direction. Small pockets of zircon- and magnetite-bearing syenites are seen in this basin. The miarolitic texture is a commonly seen in riebeckite pegmatites and aplites. This texture is produced by a crystal growth mechanism that was initially governed under the influence of surface-controlled kinetics. Larger crystals tend to grow by absorbing the required amount of water from the pore fluids present in the surrounding fine-grained matrix. Therefore, the fine-grained grains in the matrix are unable to grow larger due to a lack of water. Moreover, owing to their rapid cooling, larger phenocrysts formed at high temperatures exhibit lognormal rates of growth and in sizes relative to the small grains present in the matrix, forming syenite porphyries or porphyritic rocks under sub-volcanic conditions. The sudden increase of the viscosity of the residual magma also plays a critical role in the crystallization process of felsic and mafic minerals of two or more generations. Large crystals enclosing inclusions of euhedral feldspars oriented in two or three different directions are seen in mottled augite porphyritic syenite and this feature indicates slow cooling of the syenite magma followed by rapid ascending and consolidation of the magma. On the other hand, the anorthoclase-bearing miarolitic syenitic aplites and pegmatites carry abundant miarolitic cavities in which radiating needles of riebeckite are grown towards the centre of the cavities. Cross-cutting anorthoclasebearing aplites are seen in mottled porphyritic syenites. Two generations of magnetite, apatite and zircon crystallize, both as early-formed minerals as well as very late-formed minerals. The late-formed minerals tend to have comparatively high ratios of surface area to volume. The large feldspar phenocryst platelets increasingly thin in speckled and mottled porphyritic syenites. Moreover, feldspars which crystallize very late in the mottled porphyritic syenite appear to be formed as early crystallized mineral in some of the agpaitic syenite pegmatites and aplites in the Jogipatti basin. The concentration of garnet, aegirine-augite or magnetite along a particular zone produces a banded structure. Radiating bundles and prisms of dark green riebeckite are seen in the aplitic riebeckite syenite and in the carbonatites of Jogipatti. All these varieties of syenites are more or less interrelated to each other in their mineralogy and chemical composition. The syenites of the Sevvattur basin exhibit magmatic continuity with the syenites of the Jogipatti basin. They exhibit a co-magmatic relationship and gradational, zonal, spatial and temporal variations and continuities.

An arc-like vogesite outcrop with a lamprophyric pandiomorphic texture with black needles of katophorite set in a feldspathic matrix of two generations occurs in the Sevvattur basin with close proximity to carbonatite exposures. The vogesite imperceptibly grades to ferrohastingsite syenite, speckled hornblende syenite and hornblende-biotite oligoclasite. An

microcline syenite, wollastonite syenite, wollastonite-scapolite syenite and wollastonitemelanite carbonatite are some of the aplitic veins found in this complex. Fine-grained microcline melanite syenite displays a xenomorphic granular texture. Micro-grains of microcline are surrounded by accessories of anhedral albite and melanite. These microclines and anorthoclases are highly disordered feldspars, having significant trace amounts of ferric iron. The biotite-syenite has a xenomorphic granular texture with euhedral to subhedral biotite. The biotite syenite porphyry is composed of large phenocrysts of anorthoclase, displaying zonal variation from the core to the periphery and set in a micro-granular feldspathic matrix. Such syenite porphyries are also composed of potash feldspar phenocrysts with a cryptoperthitic texture at the core and with homogeneous peripheral portions. The feldspars belong to two or more generations and they occur as large plates of phenocrysts as well as very small anhedral grains in the matrix. Younger aplites trend in N85o crosscuts while older aplites trend in the N45o direction. Small pockets of zircon- and magnetite-bearing syenites are seen in this basin. The miarolitic texture is a commonly seen in riebeckite pegmatites and aplites. This texture is produced by a crystal growth mechanism that was initially governed under the influence of surface-controlled kinetics. Larger crystals tend to grow by absorbing the required amount of water from the pore fluids present in the surrounding fine-grained matrix. Therefore, the fine-grained grains in the matrix are unable to grow larger due to a lack of water. Moreover, owing to their rapid cooling, larger phenocrysts formed at high temperatures exhibit lognormal rates of growth and in sizes relative to the small grains present in the matrix, forming syenite porphyries or porphyritic rocks under sub-volcanic conditions. The sudden increase of the viscosity of the residual magma also plays a critical role in the crystallization process of felsic and mafic minerals of two or more generations. Large crystals enclosing inclusions of euhedral feldspars oriented in two or three different directions are seen in mottled augite porphyritic syenite and this feature indicates slow cooling of the syenite magma followed by rapid ascending and consolidation of the magma. On the other hand, the anorthoclase-bearing miarolitic syenitic aplites and pegmatites carry abundant miarolitic cavities in which radiating needles of riebeckite are grown towards the centre of the cavities. Cross-cutting anorthoclasebearing aplites are seen in mottled porphyritic syenites. Two generations of magnetite, apatite and zircon crystallize, both as early-formed minerals as well as very late-formed minerals. The late-formed minerals tend to have comparatively high ratios of surface area to volume. The large feldspar phenocryst platelets increasingly thin in speckled and mottled porphyritic syenites. Moreover, feldspars which crystallize very late in the mottled porphyritic syenite appear to be formed as early crystallized mineral in some of the agpaitic syenite pegmatites and aplites in the Jogipatti basin. The concentration of garnet, aegirine-augite or magnetite along a particular zone produces a banded structure. Radiating bundles and prisms of dark green riebeckite are seen in the aplitic riebeckite syenite and in the carbonatites of Jogipatti. All these varieties of syenites are more or less interrelated to each other in their mineralogy and chemical composition. The syenites of the Sevvattur basin exhibit magmatic continuity with the syenites of the Jogipatti basin. They exhibit a co-magmatic relationship and gradational,

zonal, spatial and temporal variations and continuities.

An arc-like vogesite outcrop with a lamprophyric pandiomorphic texture with black needles of katophorite set in a feldspathic matrix of two generations occurs in the Sevvattur basin with close proximity to carbonatite exposures. The vogesite imperceptibly grades to ferrohastingsite syenite, speckled hornblende syenite and hornblende-biotite oligoclasite. An acmite-bearing syenite grades with an enrichment of cumulate needles of katophorite into lamprophyric vogesite, displaying a pandiomorphic texture. The shonkinite occurring in Jogipatti basin is a coarse-grained inequigranular rock composed of equal proportions of sanidine / anorthoclase and augite, with accessories of olivine, hastingsite, phlogopite, apatite and magnetite and exhibiting a lamprophyric pandiomorphic texture. In the field, it occurs as xenoliths or nodules in ultramafic rocks. The ultramafic rock which is the host rock of the shonkinite appears to be possessed of kimberlitic affinity (Ramasamy et al., 2010). Mineralogical gradation exists between the shonkinite and garnetiferous syenite. An abrupt mineralogical and compositional gradation exists between the vogesite and shonkinite. Similarly, there exists such gradation between vogesite and hornblende-biotite oligoclasites. Again, the shonkinite exhibits such a type of gradation between the garnetiferous-orthoclase syenite and the mottled augite syenite. On the whole, one syenitic member has genetic relationship with any other member in this zoned complex, either in mineralogical or compositional gradations. In a detailed petrographic and mineralogical study, it is shown that the shonkinite is considered as a parent magma for this carbonatite alkali-syenite complex (Saravanan and Ramasamy, 1995; Ramasamy et al., 2010). Dykes, ring dykes, sills, veins, cavity fillings, cone sheets and plugs of syenites and carbonatites show sharp contacts with chilled margins with their host rocks. Curvilinear tensional cracks are seen around the exposures of the syenites and carbonatites. Late magmatic aplites and pegmatites filled these cracks and show co-magmatic relationships with the adjoining plutons.

All of the carbonatites occurring in the three places in Sevvattur, Jogipatti and Onnakarai grade with zonal variation from ferro-carbonatite, beforsite and sovite. Calcite, dolomite and ankerite constitute the essential minerals in these rocks and biotite, phlogopite, magnetite, riebeckite, aegirine augite, acmite, wollastonite, garnet and the rare-earth minerals of pyrochlore, zircon and niobian rutile constitute the notable accessory minerals in these rocks. Two generations of magnetites, both in the form of opaque dusts and anhedral, euhedral and twinned crystals, are present in some of the carbonatites and in the shonkinite. The Fe- and Mg-rich carbonatites appear to be emplaced from relatively deeper horizons. By their characteristic minerals, the zoned cone-sheets of the carbonatites in Sevvattur exhibit various sheets and layers as hybrid sovite, biotite sovite, phlogopite sovite, aegirine sovite, pyrochlore-bearing para-ankeritic sovite, pyrochlore-rich para-ankeritic beforsite, magnetite beforsite and apatite soviet. The elongated rings and veins of the riebeckite-carbonatites are seen in the riebeckite syenites of the Jogipatti basin. The hybrid carbonatites of the Jogipatti basin are enriched with richterite, magnesio-arfvedsonite, acmite, wollastonite and scapolite. The carbonatites of Onnakarai appear to be younger than the Jogipatti carbonatite, which is intermediate and younger itself than the Sevvattur carbonatite (Ramasamy, 1973). Again, the zoned alkali syenites of Sevvattur are older than those in the Jogipatti basin. A carbonatite body crops out in arcuate cone-sheets with a sharp contact along the meeting between the ultramafic rocks and the syenites in the northern periphery of the Sevvattur body. Mineralogical variations occur as zones which are almost conformable with one another, starting from the convex side of the arc and proceeding to the concave side in the following order: a) micaceous zone, b) ankerite zone, c) beforsite zone and d) sovite zone. About 400 m west of the carbonatite exposure is evidence of the forceful injection of carbonatite, as manifested in a syenitic outcrop which has been thoroughly brecciated with feldspar grains which were twisted and ragged with the emplacement of fine-grained carbonatitic cementing media. Along the contacts of the carbonatite, the host rocks have

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

genetic relationship with that of the alkali syenites.

The mineral assemblages in the alkaline rocks occurring in the two adjacent basins exhibit continuous chemical variations from one end to the other in various members of co-

**8. Zoned carbonatite-alkali complex** 

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 549

developed at the point of contact with the alkali syenites-carbonatite-ultramafics. Feldspathic selvages in ultramafics often display the development of biotite. At the contact between the garnetiferous syenites and the ultramafics, a skarn rock has been developed at around 2000 m length and with a width ranging from 10 to 200 m. It is comprised of calcite, wollastonite, grossularite and epidote as major minerals. In these rocks, no accessory minerals characteristic of carbonatites, such as apatite, magnetite, zircon, alkali pyroxenes, amphiboles and REE minerals, are seen. The individual minerals in the skarn rock do not have any mineral lineation steeply plunging towards the centre of the basin. However, evidence such as the partial transformation of diopside into calcite, garnet or wollastonite is present for the carbonate metasomatism of one mineral into another. The rock has been extensively folded, sheared and deformed, carrying ultrabasic nodules of varying dimensions from less than 1 cm to over 40 cm. These nodules are partially carbonatized and they exist discontinuously along the folded axial planes. The intricate folding pattern with layers and veins of calcite is even seen in hand specimens. Granulitic and gneissic layers occur within it as contorted and elongated lenses. These knotty inclusions stand out boldly on the weathered surfaces. The skarn rock exhibits a characteristic ribbed structure due to weathering. The ribs are parallel to the foliation direction and are developed along the trend of the elongation of the direction of the relics of mobilized ultramafic nodules towards the N 45o direction. Specks of native copper, bornite, covellite, chalcopyrite, pyrite, galena pyrrhotite are disseminated into the nodules. The modes of the occurrence of these nodules indicate that part of the ultramafic rock occurring along the contact of the garnetiferous syenite has been transformed into a skarn rock during the emplacement of the syenite body along its contacts. Moreover, along these contacts, riebeckite-sovites, acmite and riebeckitebearing ferro-carbonatites, veins of ferro-carbonatite breccias are seen. The ferro-carbonatite breccia carrying angular fragments of ferro-carbonatite and riebeckite syenite, quartz carrying inclusions of riebeckite needles and euhedral magnetite crystals sets in very finegrained carbonate matrix. Besides breccias, there are exposures of monazite bearing riebeckite syenite, benstonite carbonatite, barite veins, veins of galena and sulphide hosted aplites and pegmatites are emplaced just inside and on the western outside of the skarn rock exposures. The ultramafic body exposed to the west of the skarn rock is seen with the extensive development of biotite and with specks of sulphide minerals aligned characteristically along the foliation or schistose planes of the rock. It is subjected to vermiculitization with the emplacement of carbonatitic veins. In places, veins of ilmenorutile, barite, apatite-ilmenorutile rock, ankeritic carbonatite and riebeckite are seen in it. The biotite is extensively developed in the ultramafic rock along its contact with garnetiferous syenite. Small pockets of massive biotite-hornblende granites with a sharp point of contact with the country rocks of granite-gneisses occur in the north and south of the carbonatite complex. An arcuate fault-bounded aegirine granite exposure imperceptibly grades towards the east as the hornblende granodiorite in the east is seen in between the two syenite basins. The other granites occurring in this complex also have a gradational variation in their mineralogy and in their chemical compositions, and they also have a

been carbonatized by the intrusion of a thin vein in which lots of carbonates and replacements of silicate minerals into calcites are observed. Iron-enriched phlogopite is found at the contact of carbonatite with the ultramafics. Schiller-type inclusions are common and abundant along the cleavages and partings of the augite crystals. Hornblende wraps round the augite and is pleochroic, from green to yellow. Small prisms of biotite are found along the margins of the pyroxene and hornblende. An economical deposit of vermiculite has been formed at the contact between the ultramafics and the carbonatites. The carbonatite exhibits a well-developed concentric flow structure with linear crystals of apatite, alkaliamphibole and magnetite or else a schlieren of hornblende and biotite and streaky patches of ferromagnesian minerals. The carbonatite imperceptibly grades into sovite, beforsite and ankerite, depending upon an increase of dolomite or ferro-carbonates. In the ferrocarbonatites, which are relatively emplaced from deep-seated sources, pyrochlore, ilmenite and ilmenorutile concentrated in the Sevvattur, Jogipatti and Onnakarai villages, respectively. The broad spectrum of the evolution of the carbonatite complex is manifested in: a) introduction of the calcite into the host rocks, b) the formation of calcium-bearing minerals like grossularite-andradite, melanite, wollastonite, richterite paragonite, margarite, sphene and epidote, c) the alkalinization of clinopyroxene and amphiboles, d) phlogopitization and e) feldspathization. The carbonatite exhibits a xenomorphic granular texture by the presence of anhedral calcite or dolomite grains. The oligoclase and potash feldspars in the plates of the calcites of carbonatites are often corroded with an embayed outline. The micro-fractures in the host rocks are filled with calcites and some mafic minerals are transformed into margarite. The potash feldspar is found as inclusion along the contact of carbonatites and syenites and the host rock is enriched with turbid granular potash feldspar. Lumps of granular clinopyroxenes are developed in the host rock of ultramafics along the point of contact with the carbonatites. The grain size of the granular clinopyroxene has been transformed into well-developed clinopyroxene with larger dimensions. Some of them grow to over 10 cm in their dimensions as single crystals. Apatite is the more common accessory mineral in this contact zone. Clinopyroxene and amphiboles are converted into biotite. Wollastonite, garnet (melanite) or scapolite are formed as small prisms interstitial to acmite and appear to be formed by magmatic crystallization (Eckermann, 1966).

Both the basins are surrounded by ultramafic rocks. They include dunite, peridotites, kimberlites, pyroxenite, biotite pyroxenite, massive amphibolites and biotite. It is not possible to classify these rocks in the field owing to their scarce exposure on the surface and so they are together classified as ultramafic rocks. Magnesite and kankar are found on the weathered surfaces of the ultramafics. Dunite and peridotites are altered to talc-steatite and calcite-bearing rocks. Massive dunite exhibits a saccharoidal texture. However, the grain sizes of these rocks vary widely. Porphyritic pyroxenite with kankar veins is also seen in some well sections. Ultramafics rich in biotite and amphibole are very common. Veins and selvages of calcite-bearing riebeckite showing a schistose structure fill the cracks developed within the ultramafics. The ultramafics are essentially comprised of pyroxenites and peridotites. Thin veins of oligoclasites, albitites and carbonatites intrude into the ultramafics, which are transformed into biotite-pyroxenite and hornblende-oligoclase-calcite-bearing pyroxenite. Many varieties of hybrid carbonatites are developed at the point of contact of the ultramafic rocks with the carbonatites. Alkali pyroxenes like aegirine-augite and augite have developed in the ultramafic rocks and in the syenites. Alkali amphiboles like magnesio-riebeckite, magnesio-arfvedsonite and richterite have developed extensively in the aplitic syenites. Grossularite, melanite, scapolite, calcite and wollastonite were

been carbonatized by the intrusion of a thin vein in which lots of carbonates and replacements of silicate minerals into calcites are observed. Iron-enriched phlogopite is found at the contact of carbonatite with the ultramafics. Schiller-type inclusions are common and abundant along the cleavages and partings of the augite crystals. Hornblende wraps round the augite and is pleochroic, from green to yellow. Small prisms of biotite are found along the margins of the pyroxene and hornblende. An economical deposit of vermiculite has been formed at the contact between the ultramafics and the carbonatites. The carbonatite exhibits a well-developed concentric flow structure with linear crystals of apatite, alkaliamphibole and magnetite or else a schlieren of hornblende and biotite and streaky patches of ferromagnesian minerals. The carbonatite imperceptibly grades into sovite, beforsite and ankerite, depending upon an increase of dolomite or ferro-carbonates. In the ferrocarbonatites, which are relatively emplaced from deep-seated sources, pyrochlore, ilmenite and ilmenorutile concentrated in the Sevvattur, Jogipatti and Onnakarai villages, respectively. The broad spectrum of the evolution of the carbonatite complex is manifested in: a) introduction of the calcite into the host rocks, b) the formation of calcium-bearing minerals like grossularite-andradite, melanite, wollastonite, richterite paragonite, margarite, sphene and epidote, c) the alkalinization of clinopyroxene and amphiboles, d) phlogopitization and e) feldspathization. The carbonatite exhibits a xenomorphic granular texture by the presence of anhedral calcite or dolomite grains. The oligoclase and potash feldspars in the plates of the calcites of carbonatites are often corroded with an embayed outline. The micro-fractures in the host rocks are filled with calcites and some mafic minerals are transformed into margarite. The potash feldspar is found as inclusion along the contact of carbonatites and syenites and the host rock is enriched with turbid granular potash feldspar. Lumps of granular clinopyroxenes are developed in the host rock of ultramafics along the point of contact with the carbonatites. The grain size of the granular clinopyroxene has been transformed into well-developed clinopyroxene with larger dimensions. Some of them grow to over 10 cm in their dimensions as single crystals. Apatite is the more common accessory mineral in this contact zone. Clinopyroxene and amphiboles are converted into biotite. Wollastonite, garnet (melanite) or scapolite are formed as small prisms interstitial to acmite and appear to be formed by

Both the basins are surrounded by ultramafic rocks. They include dunite, peridotites, kimberlites, pyroxenite, biotite pyroxenite, massive amphibolites and biotite. It is not possible to classify these rocks in the field owing to their scarce exposure on the surface and so they are together classified as ultramafic rocks. Magnesite and kankar are found on the weathered surfaces of the ultramafics. Dunite and peridotites are altered to talc-steatite and calcite-bearing rocks. Massive dunite exhibits a saccharoidal texture. However, the grain sizes of these rocks vary widely. Porphyritic pyroxenite with kankar veins is also seen in some well sections. Ultramafics rich in biotite and amphibole are very common. Veins and selvages of calcite-bearing riebeckite showing a schistose structure fill the cracks developed within the ultramafics. The ultramafics are essentially comprised of pyroxenites and peridotites. Thin veins of oligoclasites, albitites and carbonatites intrude into the ultramafics, which are transformed into biotite-pyroxenite and hornblende-oligoclase-calcite-bearing pyroxenite. Many varieties of hybrid carbonatites are developed at the point of contact of the ultramafic rocks with the carbonatites. Alkali pyroxenes like aegirine-augite and augite have developed in the ultramafic rocks and in the syenites. Alkali amphiboles like magnesio-riebeckite, magnesio-arfvedsonite and richterite have developed extensively in the aplitic syenites. Grossularite, melanite, scapolite, calcite and wollastonite were

magmatic crystallization (Eckermann, 1966).

developed at the point of contact with the alkali syenites-carbonatite-ultramafics. Feldspathic selvages in ultramafics often display the development of biotite. At the contact between the garnetiferous syenites and the ultramafics, a skarn rock has been developed at around 2000 m length and with a width ranging from 10 to 200 m. It is comprised of calcite, wollastonite, grossularite and epidote as major minerals. In these rocks, no accessory minerals characteristic of carbonatites, such as apatite, magnetite, zircon, alkali pyroxenes, amphiboles and REE minerals, are seen. The individual minerals in the skarn rock do not have any mineral lineation steeply plunging towards the centre of the basin. However, evidence such as the partial transformation of diopside into calcite, garnet or wollastonite is present for the carbonate metasomatism of one mineral into another. The rock has been extensively folded, sheared and deformed, carrying ultrabasic nodules of varying dimensions from less than 1 cm to over 40 cm. These nodules are partially carbonatized and they exist discontinuously along the folded axial planes. The intricate folding pattern with layers and veins of calcite is even seen in hand specimens. Granulitic and gneissic layers occur within it as contorted and elongated lenses. These knotty inclusions stand out boldly on the weathered surfaces. The skarn rock exhibits a characteristic ribbed structure due to weathering. The ribs are parallel to the foliation direction and are developed along the trend of the elongation of the direction of the relics of mobilized ultramafic nodules towards the N 45o direction. Specks of native copper, bornite, covellite, chalcopyrite, pyrite, galena pyrrhotite are disseminated into the nodules. The modes of the occurrence of these nodules indicate that part of the ultramafic rock occurring along the contact of the garnetiferous syenite has been transformed into a skarn rock during the emplacement of the syenite body along its contacts. Moreover, along these contacts, riebeckite-sovites, acmite and riebeckitebearing ferro-carbonatites, veins of ferro-carbonatite breccias are seen. The ferro-carbonatite breccia carrying angular fragments of ferro-carbonatite and riebeckite syenite, quartz carrying inclusions of riebeckite needles and euhedral magnetite crystals sets in very finegrained carbonate matrix. Besides breccias, there are exposures of monazite bearing riebeckite syenite, benstonite carbonatite, barite veins, veins of galena and sulphide hosted aplites and pegmatites are emplaced just inside and on the western outside of the skarn rock exposures. The ultramafic body exposed to the west of the skarn rock is seen with the extensive development of biotite and with specks of sulphide minerals aligned characteristically along the foliation or schistose planes of the rock. It is subjected to vermiculitization with the emplacement of carbonatitic veins. In places, veins of ilmenorutile, barite, apatite-ilmenorutile rock, ankeritic carbonatite and riebeckite are seen in it. The biotite is extensively developed in the ultramafic rock along its contact with garnetiferous syenite. Small pockets of massive biotite-hornblende granites with a sharp point of contact with the country rocks of granite-gneisses occur in the north and south of the carbonatite complex. An arcuate fault-bounded aegirine granite exposure imperceptibly grades towards the east as the hornblende granodiorite in the east is seen in between the two syenite basins. The other granites occurring in this complex also have a gradational variation in their mineralogy and in their chemical compositions, and they also have a genetic relationship with that of the alkali syenites.

## **8. Zoned carbonatite-alkali complex**

The mineral assemblages in the alkaline rocks occurring in the two adjacent basins exhibit continuous chemical variations from one end to the other in various members of co-

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

increase as K increases in this complex.

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 551

clinopyroxenes or mafic minerals during the course of magmatic differentiation. According to the volume proportion and mineral structure of the mafic minerals present in the various alkaline members, they are subjected to transformations of clinopyroxenes into amphibole, biotite or magnetite. There exists continuity in the compositional variations of the mafic minerals from one syenite member to another (Ramasamy, 1986a). The clinopyroxenes from the youngest alkali syenites fall into the fields of ferrosilite and the calcium-rich hedenbergite end. The clinopyroxene from the melanite-orthoclase syenite contains a high content of CaO owing to the partial transformation of the clinopyroxene into grossulariteandradite garnet (Saravanan and Ramasamy, 1995). During late magmatic stages, due to increase of agpaitic coefficient {(Na+K)/Al} with progressive depletion of Al, orthoclase transforms into highly disordered microcline incorporating ferric iron in its lattice. Release of Tiiv and Feiii in the appropriate sites of orthoclase and grossularite-andradite garnet as per the exchange of TiivFeiii ↔ TilvAliii from subsolidus garnetiferous orthoclasite transforms to melanite-microcline syentite (Ramasamy, 1986). Owing to increasing crystallization and the transformation of mafic minerals into alkaline mafic minerals, it is also dependent on the mineral proportion of mafic minerals present in the individual alkaline rocks. Replacing Ca ions from the amphiboles of Ca76Na15K9 by the substitution of (Na+K) in the M4 sites of the amphiboles to Ca27Na65K8 in the younger generation of syenites caused the crystallization of the riebeckite. Furthermore, the ratio of 100 Mg/(Mg+Fe+Mn) of the calciferous amphiboles decreases from 85 to 30% because of the changing composition from pargasite to ferrohastingsite in the younger syenites. The katophorite and richterite are restricted to vogesite and hybrid carbonatites respectively. The eckermannite, magnesio-arfvedsonite, arfvedsonite, magnesioriebeckite and riebeckite are widely distributed in potash rich syenites and carbonatites. Both the clinopyroxenes and amphiboles are deficient in Si in their tetrahedral sites and are compensated by Al and Ti ions. However, under increasing PH2O in the late magmatic syenites, Si is saturated and Na replaces Ca with increasing silica activity. The "m" value of biotite Mg/(Mg+Fe) decreases from 0.91 to 0.29 in the late magmatic syenites. The Cs (4-10 ppm), Rb (30-345 ppm) and Li (1.5-7.5 ppm) contents

The emplacement of younger alkali syenite bodies at the contacts of older syenite units plays a critical role in the transformation and metasomatism of mafic minerals into alkali-rich minerals. Again, feldsic minerals are also subjected to metasomatism, such as the transformation of oligoclase into potash feldspars. The oligoclase surrounded by the rims of reverse zoning in the mottled porphyritic syenite indicates that it is ascending to higher levels. The presence of peristerites formed by the sub-microscopic intergrowth of sodiumrich and calcium-rich phases is observed in the mica syenite porphyries. The compositional zoning indicates that with a decrease of the temperature, the sequence solidified from the margin inwards. However, such metasomatism appears to take place at sub-solidus phases of the phenocrysts through the escape of volatile constituents carrying the required ions in the vapour or liquid phases. Sheets and plugs emplaced at the inner portions appear to be derived from a deep-seated magmatic column. The solidification was interrupted repeatedly by surges of fluid core magma. The emplacements of the various plutons are structurally controlled. The isoclinal folding of the flow bands of the carbonatitic or syenitic layers were developed, indicating the stretching and spreading of the carbonatitic or syenitic magma along the narrow zones of the emplacements. The residual magmas formed through the differentiation and fractionation of a shonkinitic parent magma resulted in decreasing silica

magmatic alkaline rocks. Larger syenite bodies show imperceptible gradation to the adjacent syenite bodies. The syenites and carbonatites are composed of both early-formed and late-formed minerals (Saravanan and Ramasamy, 1995). All of these rocks are classified based on their wide variations in their textural features and mineralogical compositions. The mapping was carried out based on these features in the field. The anorthite content in plagioclase feldspars in the syenites varies between An28% and An35%, and often oligoclase is the predominant in the speckled syenite and it decreases to An4% in the places from the mottled porphyritic syenite which imperceptibly grade from that speckled porphyritic syenite with an oligoclase of An18%. Plagioclase growth twins are evident in the syenites. Carlsbad (twin axis {001}, penetration twin), acline(composition plane {001}) Baveno (contact twin, twin plane {021}) and Manebach (contact twin, twin plane {001} twins as well as periclines (twin axis {010}) and albite (twin plane {010}) twins are common in the feldspars of the syenites. Normal zoning (progressively sodic towards the rim) is common and reverse zoning (progressively more calcic towards the rim) is also found, occasionally, in the syenites. In places, the oscillatory normal zonings ranging between An34% and An24% are seen in the speckled syenites. Patchy zonings are found in the biotite syenite porphyries. The anorthite contents of the phenocrysts in the porphyritic syenites, vogesites and the syenite porphyries are slightly higher than those of the plagioclases in the matrix. The potash feldspars have the following mean variations of orthoclase and albite components: speckled hornblende syenite (Or77 Ab23), speckled hornblende porphyritic syenite (Or87Ab13), mottled augite syenite (Or50 Ab50), mottled augite porphyritic syenite (Or87 Ab13), shonkinite (Or61 Ab39), garnetiferous syenite (Or62 Ab38) and riebeckite syenite (Or43 Ab57). Microcline cryptoperthite is commonly found in mottled porphyritic syenite and in syenite porphyries. The plot of the percentages of or-ab-an in a trilinear diagram shows that most of alkali feldspars fall within the field of sanidine, anorthoclase, albite and oligoclase. The Al/Si ratios (0.11 in albitite, -0.49microcline-melanite-syenite) in the alkali feldspar appear to be related to the history of the unmixing of the alkali feldspar (Deer et al., 1965). These components are plotted in a diagram constructed for the experimental data under 5000 bars H2O pressure; they indicate that the individual components fall within the range of 800oC to 950oC (Yoder et al., 1957). Microcline, microperthite, anorthoclase, sanidine and orthoclase are present in various syenite members and show a high degree of triclinicity (ranging between 0.6 and 1.0), indicating its disorder states and their high temperature formations and the emplacements (Ramasamy, 1986) of the syenite bodies. The microclinemicroperthite, potassic host ranges from 48% to 67%, while the sodic guest ranges from 33% to 52%, indicating their high disorder relationships and high temperature formations. The feldspars display clouding owing to the presence of very fine dusty inclusions of iron oxides in some varieties of shonkinite and syenites. Such types of clouding are present in calcite, dolomite and in ankerite, with varieties of textural patterns in the carbonatites of Sevvattur. It is interpreted that dissociation of ankerite into calcite and magnetite during the course of the ascension of the magma during successive emplacements of cone sheets of carbonatites from deep-seated sources through the centre of the magmatic body at the late magmatic stages would have produced such textural variations (Ramasamy et al., 2001).

The diopside in pyroxenite and feldspar syenites adjacent to the carbonatite exposures are transformed into calcite. The ultramafic body in the Jogipatti basin is extensively deformed and carbonatized. the proportions of plagioclase content decrease towards the centre of the mottled porphyritic syenite. The colour index (Rittmann, 1973) for the syenites varies widely in the alkaline rocks, depending upon the degree of fractionation and the separation of the

magmatic alkaline rocks. Larger syenite bodies show imperceptible gradation to the adjacent syenite bodies. The syenites and carbonatites are composed of both early-formed and late-formed minerals (Saravanan and Ramasamy, 1995). All of these rocks are classified based on their wide variations in their textural features and mineralogical compositions. The mapping was carried out based on these features in the field. The anorthite content in plagioclase feldspars in the syenites varies between An28% and An35%, and often oligoclase is the predominant in the speckled syenite and it decreases to An4% in the places from the mottled porphyritic syenite which imperceptibly grade from that speckled porphyritic syenite with an oligoclase of An18%. Plagioclase growth twins are evident in the syenites. Carlsbad (twin axis {001}, penetration twin), acline(composition plane {001}) Baveno (contact twin, twin plane {021}) and Manebach (contact twin, twin plane {001} twins as well as periclines (twin axis {010}) and albite (twin plane {010}) twins are common in the feldspars of the syenites. Normal zoning (progressively sodic towards the rim) is common and reverse zoning (progressively more calcic towards the rim) is also found, occasionally, in the syenites. In places, the oscillatory normal zonings ranging between An34% and An24% are seen in the speckled syenites. Patchy zonings are found in the biotite syenite porphyries. The anorthite contents of the phenocrysts in the porphyritic syenites, vogesites and the syenite porphyries are slightly higher than those of the plagioclases in the matrix. The potash feldspars have the following mean variations of orthoclase and albite components: speckled hornblende syenite (Or77 Ab23), speckled hornblende porphyritic syenite (Or87Ab13), mottled augite syenite (Or50 Ab50), mottled augite porphyritic syenite (Or87 Ab13), shonkinite (Or61 Ab39), garnetiferous syenite (Or62 Ab38) and riebeckite syenite (Or43 Ab57). Microcline cryptoperthite is commonly found in mottled porphyritic syenite and in syenite porphyries. The plot of the percentages of or-ab-an in a trilinear diagram shows that most of alkali feldspars fall within the field of sanidine, anorthoclase, albite and oligoclase. The Al/Si ratios (0.11 in albitite, -0.49microcline-melanite-syenite) in the alkali feldspar appear to be related to the history of the unmixing of the alkali feldspar (Deer et al., 1965). These components are plotted in a diagram constructed for the experimental data under 5000 bars H2O pressure; they indicate that the individual components fall within the range of 800oC to 950oC (Yoder et al., 1957). Microcline, microperthite, anorthoclase, sanidine and orthoclase are present in various syenite members and show a high degree of triclinicity (ranging between 0.6 and 1.0), indicating its disorder states and their high temperature formations and the emplacements (Ramasamy, 1986) of the syenite bodies. The microclinemicroperthite, potassic host ranges from 48% to 67%, while the sodic guest ranges from 33% to 52%, indicating their high disorder relationships and high temperature formations. The feldspars display clouding owing to the presence of very fine dusty inclusions of iron oxides in some varieties of shonkinite and syenites. Such types of clouding are present in calcite, dolomite and in ankerite, with varieties of textural patterns in the carbonatites of Sevvattur. It is interpreted that dissociation of ankerite into calcite and magnetite during the course of the ascension of the magma during successive emplacements of cone sheets of carbonatites from deep-seated sources through the centre of the magmatic body at the late magmatic

stages would have produced such textural variations (Ramasamy et al., 2001).

The diopside in pyroxenite and feldspar syenites adjacent to the carbonatite exposures are transformed into calcite. The ultramafic body in the Jogipatti basin is extensively deformed and carbonatized. the proportions of plagioclase content decrease towards the centre of the mottled porphyritic syenite. The colour index (Rittmann, 1973) for the syenites varies widely in the alkaline rocks, depending upon the degree of fractionation and the separation of the clinopyroxenes or mafic minerals during the course of magmatic differentiation. According to the volume proportion and mineral structure of the mafic minerals present in the various alkaline members, they are subjected to transformations of clinopyroxenes into amphibole, biotite or magnetite. There exists continuity in the compositional variations of the mafic minerals from one syenite member to another (Ramasamy, 1986a). The clinopyroxenes from the youngest alkali syenites fall into the fields of ferrosilite and the calcium-rich hedenbergite end. The clinopyroxene from the melanite-orthoclase syenite contains a high content of CaO owing to the partial transformation of the clinopyroxene into grossulariteandradite garnet (Saravanan and Ramasamy, 1995). During late magmatic stages, due to increase of agpaitic coefficient {(Na+K)/Al} with progressive depletion of Al, orthoclase transforms into highly disordered microcline incorporating ferric iron in its lattice. Release of Tiiv and Feiii in the appropriate sites of orthoclase and grossularite-andradite garnet as per the exchange of TiivFeiii ↔ TilvAliii from subsolidus garnetiferous orthoclasite transforms to melanite-microcline syentite (Ramasamy, 1986). Owing to increasing crystallization and the transformation of mafic minerals into alkaline mafic minerals, it is also dependent on the mineral proportion of mafic minerals present in the individual alkaline rocks. Replacing Ca ions from the amphiboles of Ca76Na15K9 by the substitution of (Na+K) in the M4 sites of the amphiboles to Ca27Na65K8 in the younger generation of syenites caused the crystallization of the riebeckite. Furthermore, the ratio of 100 Mg/(Mg+Fe+Mn) of the calciferous amphiboles decreases from 85 to 30% because of the changing composition from pargasite to ferrohastingsite in the younger syenites. The katophorite and richterite are restricted to vogesite and hybrid carbonatites respectively. The eckermannite, magnesio-arfvedsonite, arfvedsonite, magnesioriebeckite and riebeckite are widely distributed in potash rich syenites and carbonatites. Both the clinopyroxenes and amphiboles are deficient in Si in their tetrahedral sites and are compensated by Al and Ti ions. However, under increasing PH2O in the late magmatic syenites, Si is saturated and Na replaces Ca with increasing silica activity. The "m" value of biotite Mg/(Mg+Fe) decreases from 0.91 to 0.29 in the late magmatic syenites. The Cs (4-10 ppm), Rb (30-345 ppm) and Li (1.5-7.5 ppm) contents increase as K increases in this complex.

The emplacement of younger alkali syenite bodies at the contacts of older syenite units plays a critical role in the transformation and metasomatism of mafic minerals into alkali-rich minerals. Again, feldsic minerals are also subjected to metasomatism, such as the transformation of oligoclase into potash feldspars. The oligoclase surrounded by the rims of reverse zoning in the mottled porphyritic syenite indicates that it is ascending to higher levels. The presence of peristerites formed by the sub-microscopic intergrowth of sodiumrich and calcium-rich phases is observed in the mica syenite porphyries. The compositional zoning indicates that with a decrease of the temperature, the sequence solidified from the margin inwards. However, such metasomatism appears to take place at sub-solidus phases of the phenocrysts through the escape of volatile constituents carrying the required ions in the vapour or liquid phases. Sheets and plugs emplaced at the inner portions appear to be derived from a deep-seated magmatic column. The solidification was interrupted repeatedly by surges of fluid core magma. The emplacements of the various plutons are structurally controlled. The isoclinal folding of the flow bands of the carbonatitic or syenitic layers were developed, indicating the stretching and spreading of the carbonatitic or syenitic magma along the narrow zones of the emplacements. The residual magmas formed through the differentiation and fractionation of a shonkinitic parent magma resulted in decreasing silica

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

zoned alkali carbonatite complex of Tiruppattur, Tamil Nadu.

SiO2 (Yagi, 1953).

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 553

on slow cooling with release of volatile constituents into the residual magma, inducing magmatic pressure by incorporating CaO. The crystallization of the clinopyroxene in the place of the olivine, the residual magma, is depleted in the silica. Under increasing magmatic pressure and in order to meet the silica deficiency, more Na2O and Al2O3 are incorporated with the development of aegirine augite. With the impoverishment of SiO2 and Fe2O3, jadeite crystallizes from the residual magma along the peripheral portion of the olivine or the clinopyroxene. The jadeite develops along the peripheral portions of the olivine within the temperature range of 600 to 800oC with a pressure ranging from 10 to 20 kbar under progressive enrichments of Al2 O3 and Na2O and with decreasing silica activity in the residual magma (Ramasamy et al., 2010). The crystallization of jadeite (jd) in the place of albite (Ab) or nepheline (Ne) compensates the silica deficiency in the residual magma.

2jd ↔ Ne + Ab The further crystallization of jadeite - the residual magma - is enriched with K2O, Al2O3 and

Thus, the development of jadeite produces as end products silica under-saturated nepheline-free shonkinite with the crystallization of potassium feldspars from the residual magma. With an increase of the vapour pressure during the late magmatic processes, particularly under high PCO2, the mafic cumulates become unstable and release more CaO into the melt. The Ca-enriched residual magma immiscibly separates into silicate and carbonate magmas under high PCO2. The further differentiation of these magmas produces a series of alkali syenites and carbonatites, successively emplaced in a sequential order, first in the Sevvattur basin and then in the Jogipatti basin, and resulting in the formation of the

Normally, a carbonatite complex is associated with alkaline rocks accompanied with significant amounts of alkali-rich ferro-magnesium minerals. Often, a fenitized aureole is present around the carbonatite complex owing to the alteration of country rocks by its reaction with alkali fluids escaping from the carbonatite body. The country rock may be changed into silica-impoverished nepheline-bearing fenites. In this complex, such a development of fenites is absent. As magmatic differentiation takes place within a closed magmatic chamber for a prolonged period of crystallization and differentiation, the alkaline fluid evolves during the course of magmatic evolution, reacting with the minerals and crystallizing from the magma and increasing the alkali constituents up to the development of ultrapotassic syenites. The transformation of diopside into acmite and jadeite during the course of magmatic evolution is adjusted with the development of acmite (Fig. 2) and the ensuing silica deficiency is compensated for by the reaction of albite with silica undersaturated residual magma forming a jadeite component. Under high oxygen fugacity and a volatile concentration, biotite and magnetite form in the place of jadeite. The distribution of volume proportions of alkali feldspar against quartz (Fig. 3a) and alkali feldspar against plagioclase is calculated on the basis of the Rittmann norm using the chemical analyses listed in Tables 1 and 2 for the plot in the Q-A-P-F (Rittmann, 1973) diagram, indicating a linear differentiation trend for co-magmatic alkali series of rocks (Fig. 4 a & b). In Fig. 5a-j, it is indicated that the distribution of major elements in binary variations indicate overlapping smooth linear trends for syenites, ultramafics, carbonatites and dolerites. The dolerites in

this area also show a trend of alkali-enrichment during its magmatic evolution.

activity and steadily increasing oxygen fugacity and agpaicity due to the progressive enrichment of alkalis, ferric iron and volatile constituents in a closed magmatic chamber (Saravanan and Ramasamy, 1995). The relative rate of the development of agpaicity in the residual liquid and the crystallization of femic and felsic minerals also played a critical role in the alkaline magma's evolution. The silica deficiency developed in the residual magma due to the early crystallization and fractionation of the clinopyroxene and alkali feldspar, which may have been compensated for through the formation of amphibole, biotite, carbonates, phosphates and magnetites (Yagi, 1953). Therefore, all of these magmatic bodies exhibit compositional zonings. The compositional zoning indicates that, with a limited decrease in the temperature, the sequence solidified from the margin inwards. Sheets and plugs emplaced at the inner portions appear to be derived from a deep-seated magmatic column. The emplacements of the various plutons are structurally controlled.

The early-formed clinopyroxene in some of the syenites reacted with the residual magma and transformed into calcic-rich diopside and then into augite and aegirine-augite to acmite. Similarly, the acmite transformed into garnet which, again, transformed into melanite with an enrichment of titanium. the calcite reacted with silica-forming wollastonite in the garnetiferous syenite and it reacted with plagioclase forming scapolite, taking the required amount of SO3 from the volatile phase enriched with this constituent (Tables 2 and 3), as revealed from the chemical analyses by the enrichment of SO3 in bulk rock compositions. The quartz content decreases from the rocks outcropping at the outer shell to those rocks occurring towards the centre of the Sevvattur basin. On the other hand, such variation is remarkable given the formation of riebeckite, biotite, magnetite, wollastonite, garnet, melanite, scapolite and calcite by the heteromorphic transformation of mineral assemblages to compensate for the silica deficiency of the residual magma impoverished in the silica owing to the crystallization of sodaclase anorthoclase, sanidine and high temperature microclines in the syenites of the Jogipatti basin. The enrichment of volatile constituents like H2O, CO2, SO3, P2O5 and F in the late magmatic residual magmas developed different varieties of oxide minerals, like magnetite, rutile, ilmenorutile, perovskite and zircon, carbonate minerals like calcite, dolomite, para-ankerite, benstonite, sulphate minerals like barite, gypsum and scapolite, phosphate minerals like apatite and monazite, and sulphide minerals like galena, chalcopyrite, pyrite and pyrrhotite. The very presence of these minerals indicates that the residual magma was enriched in these volatile constituents, which compensates for any silica under-saturation owing to the extensive development of feldspathic constituents. Most of the rocks in this complex have insufficient alumina to form adequate alkali feldspars. The emplacements of carbonatites and its co-magmatic alkali syenites are in a sequential order due to a series of pulses occurring during intermittent tectonic disturbances.

Olivine is present in significant proportions in ultramafic rocks, shonkinite and carbonatites. The compositions of olivine in the ultrabasic rock represent an intermediate position in the course of the magmatic evolution of the olivine occurring in the shonkinite. The olivine in the shonkinite exhibits peripheral zonal variation augite, aegirine-augite and jadeite, indicating a prolonged period of crystallization under the liquid stage. The high concentration of Ca-Na-K-Al in the olivine from the ultramafics indicates its kimberlitic affinity. These features indicate that shonkinite magma is the parent magma for the ultramafics. The early fractionated olivine from the parental shonkinite magma reacts with the residual magma and the peripheral portion of the olivine is transformed into diopside

activity and steadily increasing oxygen fugacity and agpaicity due to the progressive enrichment of alkalis, ferric iron and volatile constituents in a closed magmatic chamber (Saravanan and Ramasamy, 1995). The relative rate of the development of agpaicity in the residual liquid and the crystallization of femic and felsic minerals also played a critical role in the alkaline magma's evolution. The silica deficiency developed in the residual magma due to the early crystallization and fractionation of the clinopyroxene and alkali feldspar, which may have been compensated for through the formation of amphibole, biotite, carbonates, phosphates and magnetites (Yagi, 1953). Therefore, all of these magmatic bodies exhibit compositional zonings. The compositional zoning indicates that, with a limited decrease in the temperature, the sequence solidified from the margin inwards. Sheets and plugs emplaced at the inner portions appear to be derived from a deep-seated magmatic

The early-formed clinopyroxene in some of the syenites reacted with the residual magma and transformed into calcic-rich diopside and then into augite and aegirine-augite to acmite. Similarly, the acmite transformed into garnet which, again, transformed into melanite with an enrichment of titanium. the calcite reacted with silica-forming wollastonite in the garnetiferous syenite and it reacted with plagioclase forming scapolite, taking the required amount of SO3 from the volatile phase enriched with this constituent (Tables 2 and 3), as revealed from the chemical analyses by the enrichment of SO3 in bulk rock compositions. The quartz content decreases from the rocks outcropping at the outer shell to those rocks occurring towards the centre of the Sevvattur basin. On the other hand, such variation is remarkable given the formation of riebeckite, biotite, magnetite, wollastonite, garnet, melanite, scapolite and calcite by the heteromorphic transformation of mineral assemblages to compensate for the silica deficiency of the residual magma impoverished in the silica owing to the crystallization of sodaclase anorthoclase, sanidine and high temperature microclines in the syenites of the Jogipatti basin. The enrichment of volatile constituents like H2O, CO2, SO3, P2O5 and F in the late magmatic residual magmas developed different varieties of oxide minerals, like magnetite, rutile, ilmenorutile, perovskite and zircon, carbonate minerals like calcite, dolomite, para-ankerite, benstonite, sulphate minerals like barite, gypsum and scapolite, phosphate minerals like apatite and monazite, and sulphide minerals like galena, chalcopyrite, pyrite and pyrrhotite. The very presence of these minerals indicates that the residual magma was enriched in these volatile constituents, which compensates for any silica under-saturation owing to the extensive development of feldspathic constituents. Most of the rocks in this complex have insufficient alumina to form adequate alkali feldspars. The emplacements of carbonatites and its co-magmatic alkali syenites are in a sequential order due to a series of pulses occurring during intermittent

Olivine is present in significant proportions in ultramafic rocks, shonkinite and carbonatites. The compositions of olivine in the ultrabasic rock represent an intermediate position in the course of the magmatic evolution of the olivine occurring in the shonkinite. The olivine in the shonkinite exhibits peripheral zonal variation augite, aegirine-augite and jadeite, indicating a prolonged period of crystallization under the liquid stage. The high concentration of Ca-Na-K-Al in the olivine from the ultramafics indicates its kimberlitic affinity. These features indicate that shonkinite magma is the parent magma for the ultramafics. The early fractionated olivine from the parental shonkinite magma reacts with the residual magma and the peripheral portion of the olivine is transformed into diopside

column. The emplacements of the various plutons are structurally controlled.

tectonic disturbances.

on slow cooling with release of volatile constituents into the residual magma, inducing magmatic pressure by incorporating CaO. The crystallization of the clinopyroxene in the place of the olivine, the residual magma, is depleted in the silica. Under increasing magmatic pressure and in order to meet the silica deficiency, more Na2O and Al2O3 are incorporated with the development of aegirine augite. With the impoverishment of SiO2 and Fe2O3, jadeite crystallizes from the residual magma along the peripheral portion of the olivine or the clinopyroxene. The jadeite develops along the peripheral portions of the olivine within the temperature range of 600 to 800oC with a pressure ranging from 10 to 20 kbar under progressive enrichments of Al2 O3 and Na2O and with decreasing silica activity in the residual magma (Ramasamy et al., 2010). The crystallization of jadeite (jd) in the place of albite (Ab) or nepheline (Ne) compensates the silica deficiency in the residual magma.

$$\mathbf{2jd} \leftrightarrow \mathbf{N} \mathbf{e} + \mathbf{A} \mathbf{b}$$

The further crystallization of jadeite - the residual magma - is enriched with K2O, Al2O3 and SiO2 (Yagi, 1953).

Thus, the development of jadeite produces as end products silica under-saturated nepheline-free shonkinite with the crystallization of potassium feldspars from the residual magma. With an increase of the vapour pressure during the late magmatic processes, particularly under high PCO2, the mafic cumulates become unstable and release more CaO into the melt. The Ca-enriched residual magma immiscibly separates into silicate and carbonate magmas under high PCO2. The further differentiation of these magmas produces a series of alkali syenites and carbonatites, successively emplaced in a sequential order, first in the Sevvattur basin and then in the Jogipatti basin, and resulting in the formation of the zoned alkali carbonatite complex of Tiruppattur, Tamil Nadu.

Normally, a carbonatite complex is associated with alkaline rocks accompanied with significant amounts of alkali-rich ferro-magnesium minerals. Often, a fenitized aureole is present around the carbonatite complex owing to the alteration of country rocks by its reaction with alkali fluids escaping from the carbonatite body. The country rock may be changed into silica-impoverished nepheline-bearing fenites. In this complex, such a development of fenites is absent. As magmatic differentiation takes place within a closed magmatic chamber for a prolonged period of crystallization and differentiation, the alkaline fluid evolves during the course of magmatic evolution, reacting with the minerals and crystallizing from the magma and increasing the alkali constituents up to the development of ultrapotassic syenites. The transformation of diopside into acmite and jadeite during the course of magmatic evolution is adjusted with the development of acmite (Fig. 2) and the ensuing silica deficiency is compensated for by the reaction of albite with silica undersaturated residual magma forming a jadeite component. Under high oxygen fugacity and a volatile concentration, biotite and magnetite form in the place of jadeite. The distribution of volume proportions of alkali feldspar against quartz (Fig. 3a) and alkali feldspar against plagioclase is calculated on the basis of the Rittmann norm using the chemical analyses listed in Tables 1 and 2 for the plot in the Q-A-P-F (Rittmann, 1973) diagram, indicating a linear differentiation trend for co-magmatic alkali series of rocks (Fig. 4 a & b). In Fig. 5a-j, it is indicated that the distribution of major elements in binary variations indicate overlapping smooth linear trends for syenites, ultramafics, carbonatites and dolerites. The dolerites in this area also show a trend of alkali-enrichment during its magmatic evolution.

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 555

Rock No SiO2 Al2O3 Fe2O3 FeO MnO MgO CaO Na2O K2O TiO2 P2O5 CO2 SO3 H2O Bio oligoclasite 723 48.35 12.97 3.78 3.26 2.68 12.91 3.54 5.00 0.75 1.06 2.76 3.15

Bihbpor syenite 428 51.53 18.22 7.53 2.79 4.00 7.50 3.85 2.70 0.65 0.22 0.39 0.44 Vogesite 733 53.07 13.03 4.28 5.09 3.75 11.75 4.00 3.93 0.69 0.48 0.36 Aughb syemite 464 53.17 14.06 5.32 5.42 0.35 3.62 9.06 3.80 2.25 1.30 1.50 0.68 Ac albitite 468 54.37 5.30 19.84 1.95 0.52 0.98 2.50 11.75 0.80 1.50 0.36 0.46 Bihbpor syenite 370 54.44 17.80 4.63 3.82 0.34 2.71 5.90 3.80 3.60 0.65 0.85 0.95 Augpor syenite 459 57.72 18.18 1.75 2.24 2.38 7.90 4.15 4.10 0.28 0.46 0.34 Augpor syenite 476 58.16 17.98 2.88 2.70 1.87 5.98 3.69 4.70 0.65 0.82 0.39 kato syenite 18 58.18 17.19 2.64 3.19 0.17 1.36 6.36 5.40 4.08 0.69 0.29 0.20 Augpor syenite 485 58.58 11.14 3.47 3.01 0.30 2.97 8.37 4.84 3.50 0.70 1.48 0.21 1.20 S.concave syenite 34 59.59 20.67 0.74 0.28 0.01 0.01 4.86 8.66 0.01 0.01 0.29 0.07 Albitite 461 60.87 14.09 0.32 2.62 0.40 2.51 4.14 8.00 2.57 0.20 1.81 0.51 1.58 Sev gar peg syenite 43 61.25 20.76 1.78 0.28 0.01 1.98 10.62 0.65 0.01 0.01 1.08 0.07 Kakangarai syenite 7 61.26 20.91 0.89 0.53 0.01 1.10 4.67 8.84 0.01 0.01 0.25 0.10 Sevvattur syenite 10 62.22 20.98 1.14 0.28 0.01 1.44 4.54 9.07 0.01 0.01 0.48 0.08 SevNE syeite 36 62.68 19.41 1.17 0.13 0.01 0.93 9.07 3.15 0.01 0.01 0.62 0.03 Kunnattur syenite 3 63.05 22.39 0.43 0.28 0.01 0.67 8.55 4.00 0.01 0.01 0.44 0.05 Koratti syenite 44 63.53 20.60 0.59 0.27 0.01 0.40 7.77 6.96 0.01 0.01 0.40 0.02 Karapattu syenite 6 63.54 22.27 0.74 0.27 0.01 0.75 8.55 2.86 0.01 0.01 0.30 0.09 PinkBihbporsy 463 64.24 14.74 2.84 2.02 0.83 1.13 2.42 4.93 3.34 0.62 0.52 0.44 1.43 Bi-ab-peg syenite 635 67.01 14.06 3.99 1.98 0.52 2.38 6.77 1.40 0.15 0.81 0.34 0.20

Bio fol aplite 1122 50.28 23.25 5.54 2.98 0.36 3.82 5.60 2.08 1.80 0.90 0.60 0.60

Bi hb syenite 369 70.30 13.35 1.94 2.63 0.20 1.17 4.05 3.36 1.58 0.45 0.62

Olaipatti syenite 78 51.94 22.27 3.53 0.41 0.14 0.01 4.50 0.92 13.54 0.01 0.01 0.93

Olaipatti syenite 75 53.43 19.31 6.94 0.28 0.01 7.58 2.35 7.72 0.01 0.01 0.93 Olaipatti syenite 77 54.25 22.99 2.83 0.56 0.01 2.12 2.44 12.05 0.01 0.01 1.04 Olaipatti syenite 79 54.60 23.21 4.05 0.55 0.16 0.01 5.70 1.93 8.84 0.01 0.01 1.04

Jogipatti syenite 51 58.92 19.14 1.89 0.56 0.01 1.20 1.09 14.68 0.01 0.01 1.00

Garigaipalli syenite 85 62.69 20.15 0.21 0.14 0.01 0.01 1.43 15.06 0.01 0.01 0.19

Richteite ultramafic 321 21.02 1.64 5.48 2.87 0.63 6.39 31.68 1.96 1.44 0.45 23.65 0.69 0.65 pl px ultramafic 319 41.26 3.65 10.18 8.26 0.29 9.67 20.18 0.41 1.20 1.48 0.73 0.41 0.34 1.15 bi-fels px ultramafic 320 44.12 9.10 3.64 1.26 0.06 14.20 21.78 1.15 1.24 0.88 0.10 1.12 0.60 gar pegmatite 572 44.84 11.35 12.14 0.50 0.40 0.70 19.01 3.95 4.98 0.40 1.20 0.40 px hb bi ultramafic 203 45.82 15.53 1.50 5.59 0.19 3.95 9.62 5.27 5.16 0.66 4.66 0.14 1.80 hb oligo ultramafic 204 49.16 21.03 2.57 5.16 0.08 5.43 12.37 2.43 0.83 0.14 0.16 0.30 Shonkinite 561 49.20 7.61 6.40 5.58 0.36 10.43 10.03 3.44 4.29 0.76 0.82 0.80

gar aplite 39 52.28 19.59 2.82 1.73 0.13 0.25 5.15 0.14 17.16 0.56 0.13 0.44 bi px ultramafic 318 52.92 12.98 5.31 1.80 0.11 3.21 6.18 2.03 11.76 1.09 0.46 2.24 0.16 0.44

Bi-mus oligoglacite 219 55.02 18.75 2.17 0.36 0.03 1.24 4.47 3.31 10.32 0.28 2.54 1.30 1.30 hb syeite 361 56.84 15.69 2.37 3.02 0.17 1.73 7.73 3.24 6.96 0.50 0.47 0.44 0.76 0.82 aplite 201 58.50 18.65 3.11 0.85 0.07 0.49 2.75 6.48 6.24 0.44 0.30 0.13 1.00

anorthoclasite 349 58.98 16.20 2.51 2.72 0.12 2.22 5.15 6.75 3.48 0.56 0.17 0.22 0.74 0.50 Rieb gar pegmatite 360a 59.78 18.27 1.32 1.14 0.07 1.21 5.61 5.40 6.78 0.20 0.17 0.10 Bi pegmatite 340 61.82 16.90 0.61 0.34 0.02 0.49 2.06 0.27 16.80 0.09 0.42 0.25 0.32

bihb oligoclasite 206 63.18 15.02 2.73 4.39 0.14 2.72 5.50 3.98 1.32 0.66 0.30 0.38 rieb pegmatite 301 63.24 17.74 0.88 0.72 0.03 0.49 2.41 4.73 10.12 0.28 0.15 rieb pegmatite 85 63.75 13.01 1.98 0.65 2.59 0.89 5.21 9.98 0.10 0.40 1.10 rieb aplite 541 64.01 15.44 2.44 0.25 0.40 1.62 0.21 4.04 9.15 0.14 0.22 1.58 Rieb diorite 80 64.97 16.75 2.61 1.31 0.40 0.53 3.75 5.04 2.70 0.30 0.48 0.82 Rieb sy pegmatite 36a 65.01 17.66 1.25 0.57 0.03 0.49 2.41 7.02 4.92 0.16 0.05 0.20 Pink granite 12 69.72 14.92 1.40 0.71 0.07 0.71 3.15 4.79 4.32 0.31 0.48 Pink granite 1 70.20 14.42 1.94 0.57 0.04 0.47 3.09 4.19 4.74 0.40 0.16 0.62 Rieb granite 78 71.40 13.60 2.53 0.63 0.50 0.72 2.67 3.80 2.93 0.15 0.32 0.52 ap-mt rock 498 0.48 0.73 54.17 18.07 0.49 0.73 6.66 0.14 0.42 13.20 0.77 barite rock 452 0.50 0.10 0.05 0.50 0.05 0.53 33.40 0.80 apatite rock 450 2.06 6.91 2.89 2.06 0.07 1.46 41.36 0.47 0.10 10.01 31.77 0.26 Crocidolite rock 60 6.50 1.09 0.49 0.26 0.09 0.85 50.84 0.01 0.01 1.68 0.02 37.86 0.00 Ultramafic 460 18.84 0.91 28.94 17.60 0.19 5.42 15.42 0.06 0.14 3.15 3.00 4.50 0.30 1.30

Sevvattur syenites No723-369; Jogipatti syenites No 321-36a; Granites No12-78; Ultramafics No 498-47;

Ultramafic 492 27.72 0.80 16.25 7.15 0.78 6.37 19.09 4.04 0.28 0.60 0.21 15.26 0.23 0.36

Table 1. Chemical composition of the alkaline rocks in the carbonatite complex of

Feore rock 13 20.36 0.07 27.60 37.85 0.46 0.01 11.16 0.71 0.34 1.57 0.04

Carbonatites No 310-495; Dolerites No 1176-1224

Tiruppattur

Fig. 2. The chemical compositions of alkaline rocks calculated on the basis of Acmite+Jadeite, Diopside+Enstatite and Hedenbergite+Ferrosalitse compositions, indicating a magmatic differentiation trend moving towards the Ac+Jd end members. The diagram indicates the enrichment of alkalis and iron during the course of differentiation

Fig. 3. a. & b. Streckeisen (1967) QAPF double triangle plots are presented in Qz-alk-felds and Alk-felds – Pl variation diagrams showing linear variations in co-magmatic series (The C.I. for syenites varies from 5-30; for hybrid rocks 30-70; for ultramafics 70-100).

Fig. 4. a., b. Streckeisen graph and double triangle showing the linear trends of the magmatic evolution of alkaline rocks in the Tiruppattur carbonatite complex

#### Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 555

554 Crystallization – Science and Technology

Fig. 2. The chemical compositions of alkaline rocks calculated on the basis of

indicates the enrichment of alkalis and iron during the course of differentiation

Acmite+Jadeite, Diopside+Enstatite and Hedenbergite+Ferrosalitse compositions, indicating a magmatic differentiation trend moving towards the Ac+Jd end members. The diagram

a) b)

a) b)

Fig. 3. a. & b. Streckeisen (1967) QAPF double triangle plots are presented in Qz-alk-felds and Alk-felds – Pl variation diagrams showing linear variations in co-magmatic series (The

C.I. for syenites varies from 5-30; for hybrid rocks 30-70; for ultramafics 70-100).

Fig. 4. a., b. Streckeisen graph and double triangle showing the linear trends of the magmatic evolution of alkaline rocks in the Tiruppattur carbonatite complex


Sevvattur syenites No723-369; Jogipatti syenites No 321-36a; Granites No12-78; Ultramafics No 498-47; Carbonatites No 310-495; Dolerites No 1176-1224

Table 1. Chemical composition of the alkaline rocks in the carbonatite complex of Tiruppattur

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 557

system is a superimposed structure over which a series of older block-faulted horsts and graben structures are comprised of a number of magmatic emplacements of charnockites, anorthosites, alkaline rocks, carbonatites and volcanic effusives. It is also favourable for the genesis of under-saturated alkaline magma charged with relatively more anhydrous volatiles (such as CO2, SO3, P2O5, F, Cl and CH4 etc.) with respect to H2O vapour, which is the main constituent in the volatile phase derived from a deep-seated source. The Mg/(Mg+Fet) ratios of the magmatic melts are frequently used as an indicator of whether a melt could be a partial melting product of a mantle material. The high mg values between 0.76 and 0.86 in this area depend on the TPX and the volatile constituents by which they were formed; moreover, these ratios of the starting material of the partial melts from the mantle may fall between the ratios of 0.9-0.3 (Mysen, 1975). Accordingly, all of the rocks in this complex can be derived from the partial melts from the mantle horizons. The propagation and penetration of the Eastern Ghats Paleo-rift system by reactivations during the subsequent tectonic episodes from the Early Proterozoic Period may extend to different depth levels, causing certain low degrees of partial melting which have been attained only after some prolonged period of tectonic deformation. According to Ramasamy (1982), a span of charnockitic activities extends from 3100 Ma to 2600 Ma, with anorthositic events extending from 2000 Ma to 1100 Ma and alkali syenitic- carbonatitic activities commencing from 1200 Ma onwards (Ramasamy, 1981). It seems that successively younger magmatic episodes were formed under more anhydrous- and alkali-rich environments and at deeper levels than the older ones, owing to the deeper penetration of the rifted continental plates into the mantle. The emplacement of most of the carbonatites from the Proterozoic to recent periods occurring in various parts of the world is restricted along deep crustal fractures that were controlled by regional structures and tectonics. According to Macintyre (1975), the ages of carbonatite complexes younger than 200 Ma indicate that many of them are intimately related with major changes in plate motion which were globally synchronous. The continental plate formed by the propagation and penetration of the Eastern Ghats Paleo-rift system led to the later separation of Gondwana Land from Peninsular India. The wide-spread Deccan Trap volcanic activity covered more than 500000 km2 and the thickness of the lava flows exceeds 2000 m in some places (Krishnan, 1962), with younger eruptions of olivine-tephrite, soda-trachyte and carbonatite eruptions at Kudangulam near Cape Comorin indicating that Peninsular India is prone to repeated volcanic and magmatic activities (Ramasamy, 1987) and are also bearing evidence of Indian plate movement towards the north after the break of Antarctica from Peninsular India. The ascension of magma and its rate of cooling, pressure and the volume of volatile constituents played a critical role in the magmatic evolution of the residual magma under a specific geological setting and tectonic movement. The influence of local variations in the TPX conditions during the course of the crystallization of minerals creates complexities in tracing the trends of magmatic evolution in a spatial order. There exists a compositional relationship between these new discrete minerals in younger plutons with the minerals in older plutons in this area. The distribution of incompatible HFSE of K, Ti, P, Zr, Nb, Ba, Sr (Table 4) in these rocks and high ratios of HFSE such as Ti/P, K/(K+Na), Nb/Ta, Zr/Hf, Sr/Ba, LREE/HREE and the presence of minerals like allanite, zircon, apatite, monazite, pyrochlore, niobian rutile, magnetite, galena and feldspars, indicate that the low degree of partial melts from the mantle horizon and the parent magma


Table 1. Chemical composition of the alkaline rocks in the carbonatite complex of Tiruppattur (continued)

### **9. Geological setting and tectonomagmatic evolution**

The location of the Precambrian carbonatite complex of Tiruppattur in the Indian peninsula along the Eastern Ghats Paleo-rift System (Ramasamy, 1982, 1987), extending over a stretch of 3000 km x 200 km from Cape Comorin / Palghat Gap to the Brahmaputra valley (Eastern Syntaxes of the Himalayas), is favourable site for a low degree of the partial melting of alkali-enriched upper mantle rocks from a low velocity zone (Schleicher et al., 1998). The rift

Rock No SiO2 Al2O3 Fe2O3 FeO MnO MgO CaO Na2O K2O TiO2 P2O5 CO2 SO3 H2O Peridotite 31 30.09 2.00 19.09 6.95 0.79 9.91 15.68 3.67 0.01 0.63 1.29 8.93 Biofels pyroxinite 9 36.88 1.47 15.49 10.72 0.10 8.00 17.76 1.53 1.88 1.74 1.97 1.10 Kimberlite 29 37.71 0.12 8.46 3.29 0.15 8.71 21.10 0.32 2.93 0.75 0.09 16.60

Samalpatti ultramafics S3 38.09 5.17 1.05 14.10 0.18 14.18 15.52 0.23 3.44 1.66 3.69 1.90

Riebcc ultramafic 491 42.36 1.27 7.54 1.98 0.19 12.97 15.65 4.59 0.90 0.44 0.10 9.75 2.15 Olaugoligo ultramafic 40 42.58 4.62 6.32 5.25 0.15 10.62 25.08 0.34 0.06 0.93 3.10 0.60 0.33 0.60 apatite rock 451 42.82 1.27 21.86 2.51 0.04 14.76 1.95 0.06 1.56 7.90 2.46 1.65 0.24 Samalpatti ultramafic S4 45.21 1.28 0.90 4.32 0.09 14.82 20.00 2.34 0.28 0.23 0.05 9.68 1.10 Ultramafic nodule 497 45.92 6.12 5.52 5.33 0.18 9.60 22.46 1.08 0.18 1.26 0.19 1.02 0.50 Rieb ultramafic 402 46.40 10.41 9.31 2.87 0.22 8.32 12.42 3.04 4.38 0.84 0.44 0.17 0.55 Sam ultramafics S5 46.41 2.09 1.64 2.90 0.10 14.46 18.90 2.55 0.74 0.17 0.74 8.73 1.08

wollastonite rock 491 49.66 1.22 2.01 0.63 0.67 1.52 41.86 0.34 0.30 0.07 1.15 0.30 0.30 0.20 Biopx ultramafic 736 52.40 12.26 2.99 6.75 0.02 6.01 8.48 3.11 4.26 0.99 1.53 0.49 0.00

Sev carbonatite 310 0.20 0.18 3.60 2.97 0.55 17.40 29.54 0.10 0.60 0.10 0.05 43.56 0.50 Icelandspar 499 0.28 0.08 0.14 0.25 55.03 0.21 0.12 0.14 43.04 0.50 Sev.carbonatite 11 0.30 0.12 0.06 3.84 52.12 0.10 0.05 43.10 0.20 Sev. Carbonatite 59 0.50 0.29 1.25 1.38 0.48 5.00 45.59 0.46 0.07 0.40 1.30 41.45 1.10 Sev carbonatite 389 0.60 0.79 1.23 3.29 15.28 32.28 0.50 0.10 0.10 2.33 42.35 0.40 Sev carbonatite 250 1.00 0.70 4.08 0.60 0.06 0.45 50.46 0.34 0.22 0.20 0.82 39.28 0.50 Sev.carbonatite 410 1.26 0.42 5.00 1.33 4.52 45.46 0.68 0.43 0.15 1.80 39.10 0.28 Jogi. carbonatite 313 1.88 0.65 3.55 2.36 0.62 11.25 36.04 0.20 0.09 0.15 0.02 40.02 1.66 0.40 Jogi.carboatite 315 2.06 0.60 3.90 2.93 0.60 13.23 32.51 0.37 0.33 0.10 0.04 40.24 1.35 0.45 Jogi.carbonatiate 311 2.64 0.50 1.10 0.42 0.50 2.45 48.28 0.20 0.13 0.20 0.04 40.46 0.57 0.32 Jogi. carbonatite 990 3.24 0.94 2.30 0.41 1.06 48.43 2.30 0.50 0.20 1.48 38.52 0.52 Sev.carbonatite 369 4.90 2.95 4.62 0.95 12.29 32.24 0.73 0.08 0.20 1.66 38.14 0.44 Sev. carbonatite 20 9.97 5.06 0.88 2.35 3.12 46.90 1.10 0.57 0.30 2.01 26.32 1.00 Onna Bens carbonatite 500 12.51 5.73 2.05 0.13 0.16 24.39 1.05 0.98 27.17 0.59 Jogi carbonatite 321 21.02 1.64 5.48 2.87 0.63 6.39 31.68 1.96 1.44 0.45 23.65 0.69 0.65 Jogi. carbonatite 1010 24.70 2.01 2.54 2.00 0.30 7.06 33.51 2.92 0.90 0.20 3.68 19.45 1.20 Onna carbonatite 496 27.26 1.92 16.11 7.09 2.46 4.91 17.16 3.58 0.24 1.23 1.26 16.34 0.95 Onna.carbonatite 492 27.72 0.80 16.25 7.15 0.78 6.37 19.09 4.04 0.28 0.60 0.21 15.56 0.23 0.36 Onna.carbonatitae 495 41.36 1.45 11.16 5.30 0.32 8.32 13.61 5.20 0.24 0.98 0.52 10.95 0.20 dolorite 1176 47.22 18.44 1.30 9.76 0.38 7.19 11.09 1.68 0.68 0.75 0.40 0.80 dolorite 1155 48.08 19.06 2.30 9.41 0.15 5.42 10.01 2.52 0.90 1.25 0.50 0.38 dolorite 1159 48.58 15.49 1.60 12.80 0.40 4.68 7.80 2.92 0.90 2.60 0.93 0.84 dolorite 33 49.50 15.25 1.67 9.01 7.23 11.85 2.60 0.20 0.53 0.68 0.33 0.62 dolorite 1186 54.18 12.76 1.80 8.80 0.15 8.96 8.20 2.08 0.90 0.85 0.23 0.63 dolorite 1187 55.13 8.46 3.32 9.98 0.30 7.21 9.04 3.36 1.33 1.65 0.12 0.20 dolorite 1224 56.04 8.26 3.21 9.62 0.20 4.61 8.41 6.02 1.58 1.10 0.31 0.55

Feore rock 21 39.30 0.07 49.84 10.36 0.02 0.01 0.01 0.16 0.17 0.33 0.14 Kimberlite 7a 40.85 0.74 10.33 0.57 0.13 19.08 7.00 0.40 0.01 0.01 0.01 20.65

Sev Ultramafics S1 46.77 2.17 3.43 12.58 0.13 14.02 19.76 0.19 0.35 0.66 0.33

Sam ultramafics S2 54.65 1.01 0.85 4.12 0.15 17.17 21.37 0.44 0.12 0.22 0.02 Fe ore ultramafics 47 62.14 0.17 33.58 1.22 0.01 0.01 1.07 0.01 0.34 1.30 0.01 0.39

Table 1. Chemical composition of the alkaline rocks in the carbonatite complex of

The location of the Precambrian carbonatite complex of Tiruppattur in the Indian peninsula along the Eastern Ghats Paleo-rift System (Ramasamy, 1982, 1987), extending over a stretch of 3000 km x 200 km from Cape Comorin / Palghat Gap to the Brahmaputra valley (Eastern Syntaxes of the Himalayas), is favourable site for a low degree of the partial melting of alkali-enriched upper mantle rocks from a low velocity zone (Schleicher et al., 1998). The rift

**9. Geological setting and tectonomagmatic evolution** 

Tiruppattur (continued)

system is a superimposed structure over which a series of older block-faulted horsts and graben structures are comprised of a number of magmatic emplacements of charnockites, anorthosites, alkaline rocks, carbonatites and volcanic effusives. It is also favourable for the genesis of under-saturated alkaline magma charged with relatively more anhydrous volatiles (such as CO2, SO3, P2O5, F, Cl and CH4 etc.) with respect to H2O vapour, which is the main constituent in the volatile phase derived from a deep-seated source. The Mg/(Mg+Fet) ratios of the magmatic melts are frequently used as an indicator of whether a melt could be a partial melting product of a mantle material. The high mg values between 0.76 and 0.86 in this area depend on the TPX and the volatile constituents by which they were formed; moreover, these ratios of the starting material of the partial melts from the mantle may fall between the ratios of 0.9-0.3 (Mysen, 1975). Accordingly, all of the rocks in this complex can be derived from the partial melts from the mantle horizons. The propagation and penetration of the Eastern Ghats Paleo-rift system by reactivations during the subsequent tectonic episodes from the Early Proterozoic Period may extend to different depth levels, causing certain low degrees of partial melting which have been attained only after some prolonged period of tectonic deformation. According to Ramasamy (1982), a span of charnockitic activities extends from 3100 Ma to 2600 Ma, with anorthositic events extending from 2000 Ma to 1100 Ma and alkali syenitic- carbonatitic activities commencing from 1200 Ma onwards (Ramasamy, 1981). It seems that successively younger magmatic episodes were formed under more anhydrous- and alkali-rich environments and at deeper levels than the older ones, owing to the deeper penetration of the rifted continental plates into the mantle. The emplacement of most of the carbonatites from the Proterozoic to recent periods occurring in various parts of the world is restricted along deep crustal fractures that were controlled by regional structures and tectonics. According to Macintyre (1975), the ages of carbonatite complexes younger than 200 Ma indicate that many of them are intimately related with major changes in plate motion which were globally synchronous. The continental plate formed by the propagation and penetration of the Eastern Ghats Paleo-rift system led to the later separation of Gondwana Land from Peninsular India. The wide-spread Deccan Trap volcanic activity covered more than 500000 km2 and the thickness of the lava flows exceeds 2000 m in some places (Krishnan, 1962), with younger eruptions of olivine-tephrite, soda-trachyte and carbonatite eruptions at Kudangulam near Cape Comorin indicating that Peninsular India is prone to repeated volcanic and magmatic activities (Ramasamy, 1987) and are also bearing evidence of Indian plate movement towards the north after the break of Antarctica from Peninsular India. The ascension of magma and its rate of cooling, pressure and the volume of volatile constituents played a critical role in the magmatic evolution of the residual magma under a specific geological setting and tectonic movement. The influence of local variations in the TPX conditions during the course of the crystallization of minerals creates complexities in tracing the trends of magmatic evolution in a spatial order. There exists a compositional relationship between these new discrete minerals in younger plutons with the minerals in older plutons in this area. The distribution of incompatible HFSE of K, Ti, P, Zr, Nb, Ba, Sr (Table 4) in these rocks and high ratios of HFSE such as Ti/P, K/(K+Na), Nb/Ta, Zr/Hf, Sr/Ba, LREE/HREE and the presence of minerals like allanite, zircon, apatite, monazite, pyrochlore, niobian rutile, magnetite, galena and feldspars, indicate that the low degree of partial melts from the mantle horizon and the parent magma

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

Sev sovite Sub21 5225 5180 2690 2930 1900 45 140 450 Sev beforsite Sub22 3570 9700 220 350 1500 450 90 10 110 Sev ferrocarbonaite Sub23 4250 635 40 180 1860 570 130 30 40 Jogi sovite Sub24 35 300 10 25 40 10 10 10 10 Gari skarn Sub25 220 350 20 15 30 10 10 10 Onna ferrocarbontite Sub26 1159 7795 60 50 290 760 240 105 10 25 40

Jogi ultramafic Sub3 900 30 80 180 50 35

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 559

Rock No Ba Sr Nb Zr La Ce Nd Y Th U V Cu Zn Cr Ni Co Pb Sc

Sev syenite Sub11 270 740 120 200 650 430 200 50 10 20 20 15 120 Sev syenite Sub12 2780 1120 15 120 60 30 10 25 104 15 100 Sev syenite Sub13 955 35 260 70 120 45 30 15 10 20 50 25 170 Sev ultramafic Sub1 375 12 10 10 10 10 510 90 Jogi ultramafic Sub2 275 15 10 25 10 10 10 40 55

Gari ultramafic Sub4 217 20 10 10 90 40 Sev granite gneiss Sub32 2321 1424 60 155 55 25 10 10 45 70 10 20 Jogi syenite Sub33 1726 900 60 170 55 30 30 10 20 50 20 35 Jogi syenite Sub34 1945 1120 35 240 60 20 15 10 20 60 15 180 Jogi albitite Sub35 220 30 180 350 20 10 250 100 10 15 35 10 Jogi syenite Sub36 470 80 10 25 50 50 10 20 20 10 880 Jogi shonkinite 8 1000 2000 600 15 2000 1000 1000 1000 300 200 10 300 Jogi Riebeckite R 600 60 30 10 200 9 100 2000 30 10 10 20 Jogi hy carbonatite 319 200 200 10 8 100 100 100 8 20 30 10 Jogi hy carbonatite 320 200 20 60 60 200 8 8 8 8 Jogi hy carbonatite 321 2000 2000 8 8 60 30 8 8 8 20 Jogi Rieb carbonatite 401 200 300 60 8 8 30 8 200 600 30 20 20 20 Jogi Rieb carbonatite 402 600 100 20 6010 8 100 30 30 10 Jogi Rieb carbonatite 490 100 10 8 10 8 30 10 10 Jogi Wollastonite rock 491 300 300 8 60 300 100 8 Jogi ultramafic nodule 497 2000 10 300 100 300 100 60 30 10 Jogi ultramafic 40 200 100 20 30 60 300 300 600 100 60 8 60 Sev ultramafic 201 3000 300 50 60 10 60 20 100 8 8 100 Sev ultramafic 203 6000 2000 100 50 60 100 8 200 100 30 20 100 Sev hy carbonatite 204 300 30 20 200 30 200 200 200 60 20 Sev hy carbonatite 206 600 600 60 60 100 100 100 600 100 60 20 60 Sev oligoclasite 219 1000 1000 60 100 10 300 60 30 10 10 20 Jogi carbonatite 318 4000 1000 20 60 60 100 100 100 200 30 30 30 30 Apatite magnetite rock 460 100 100 8 200 300 100 300 30 30 8 Sev pyroxenite 789 300 400 10 100 60 60 60 60 Jogi ultramafic 493 790 170 145 90 950 50 100 175 75 675 Jogi ultrapot syenite 39 475 2300 880 380 790 2630 315 280 670 320 375 560 260 390 Jogi melanite sovite 990 120 4913 1115 250 165 123 733 78 73 110 73 73 160 133 610 Jogi shonkinite 561 322 5960 708 136 142 76 534 76 24 64 76 70 50 262 74

Table 3. Trace elements' distribution in carbonatites and associated alkaline rocks

The dolerites are hypabyssal equivalents of gabbro and basalts, which are common products of the partial melting of mantle / crustal rocks at depth. The overlapping petrographic

**10. Conclusion** 

separated from the melt were derived from HFSE-enriched portions of the mantle during the folding and up-arching of the mantle rock at a depth there by enriching the above characteristic elements just prior to the initial stage of the rifting of the diverging continental plates of the Eastern Ghats Mobile Belt.


The chemical compositions of the rocks are listed in Tables 1 and 2 on the basis of rock types and their places of occurrence. The chemical compositions of these rocks vary widely in their silica, alkalis and carbonate contents. CO2-rich rock carbonatites, Alkali syenites vary in SiO2 between 35% and 65%, Ultramafic rocks vary in SiO2 between 40% and 52%. Hybrid rocks have both SiO2 and CO2 in an intermediate position.

Table 2. Compositions of various minerals occurring in the carbonatite complex of Tiruppattur



Table 3. Trace elements' distribution in carbonatites and associated alkaline rocks

## **10. Conclusion**

558 Crystallization – Science and Technology

separated from the melt were derived from HFSE-enriched portions of the mantle during the folding and up-arching of the mantle rock at a depth there by enriching the above characteristic elements just prior to the initial stage of the rifting of the diverging

Mineral SiO2 Al2O3 Fe2O3FeO MnO MgO CaO Na2O K2O TiO2 BaO SrO P2O5 CO2 SO3 F H2O H20 Total Wollastonite 49.66 1.22 2.01 0.63 0.67 1.52 41.86 0.34 0.30 0.07 0.00 0.00 1.15 0.20 0.00 0.00 0.00 0.00 99.63 Epidote 37.72 17.61 4.67 0.18 0.06 2.29 34.34 0.41 0.84 0.96 0.00 0.00 0.07 0.58 0.00 0.00 0.00 0.00 99.73 Apatite1 0.05 0.18 0.00 0.00 0.00 0.52 54.25 0.30 0.00 0.00 0.00 0.80 39.80 0.00 0.00 0.00 0.00 0.00 95.90 Apatite 2 0.22 0.06 0.00 0.00 0.00 0.17 54.15 0.29 0.00 0.00 0.00 1.51 40.03 0.00 0.00 0.00 0.00 0.00 96.43 apatite 1a 0.00 0.00 0.00 0.00 0.00 1.51 55.73 0.00 0.00 0.00 0.00 0.00 40.95 0.00 0.00 1.56 0.00 0.00 99.75 apatite2a 0.00 0.00 0.00 0.00 0.00 1.51 53.28 0.00 0.00 0.00 0.00 0.00 37.96 0.00 0.00 0.80 0.00 0.00 93.55 apatite 450 2.06 6.91 2.89 2.06 0.07 1.46 41.36 0.47 0.00 0.00 0.00 0.00 41.50 0.00 0.00 1.70 0.26 0.00 100.74 Magnetite 498I 0.10 0.18 67.64 22.57 0.39 0.25 1.05 0.00 0.00 8.10 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 100.28 magnetite 498ii 0.72 0.73 54.17 18.07 0.49 0.73 6.66 0.14 0.42 0.00 0.00 0.00 0.00 0.44 0.77 0.00 0.22 0.00 83.56 Iceladspar 0.36 0.00 0.14 0.00 0.00 0.25 55.03 0.21 0.12 0.14 0.00 0.00 0.00 43.04 0.34 0.00 0.00 0.00 99.63 calcite cc250 2.00 0.00 5.40 0.50 0.06 0.25 51.36 0.24 0.12 0.00 0.00 0.00 0.00 40.28 0.00 0.00 0.00 0.00 100.21 dolomite cc310 0.38 0.00 3.80 2.87 0.55 18.40 30.54 0.10 0.60 0.00 0.00 0.00 0.00 46.06 0.00 0.00 0.00 0.00 103.30 calcite cc311 3.74 0.00 1.40 0.32 0.50 2.25 48.52 0.10 0.03 0.00 0.00 0.00 0.00 40.46 0.57 0.00 0.00 0.00 97.89 Carbonatite cc3136.40 0.00 2.00 0.25 0.37 1.50 31.93 0.17 0.18 0.00 0.00 0.00 0.00 26.62 0.41 0.00 0.00 0.00 99.83 Mg calcite 313 2.78 0.00 4.05 2.16 0.62 11.65 36.44 0.10 0.09 0.00 0.00 0.00 0.00 41.02 1.66 0.00 0.00 0.00 100.57 carbonatite 314 35.30 0.00 1.70 0.36 0.49 1.12 32.72 0.07 0.11 0.00 0.00 0.00 0.00 26.34 0.67 0.00 0.00 0.00 98.88 ferrocarbonate c 3.26 0.00 4.30 2.73 0.60 13.63 33.31 0.27 0.23 0.00 0.00 0.00 0.00 41.84 1.35 0.00 0.00 0.00 101.52 ferrocarbonate 4 53.02 0.00 6.25 1.15 0.20 3.37 19.09 0.04 0.08 0.00 0.00 0.00 0.00 15.56 0.29 0.00 0.00 0.00 99.05 ferrocarbonatecc49.74 0.00 2.25 0.00 0.42 1.00 23.25 0.04 0.11 0.00 0.00 0.00 0.00 19.92 0.60 0.00 0.00 0.00 97.33 ferrocarboate49449.82 0.00 2.50 0.00 0.23 1.25 23.95 0.04 0.11 0.00 0.00 0.00 0.00 20.78 0.39 0.00 0.00 0.00 99.07 Benstonite 12.51 5.73 2.03 0.13 0.16 0.00 24.39 1.08 0.98 0.00 20.69 4.33 0.00 27.17 0.00 0.00 0.59 0.00 99.79 *Biotite* 32.96 15.64 15.09 0.00 0.05 16.10 3.51 0.41 6.42 3.42 0.00 0.00 0.32 0.00 0.00 1.68 5.96 0.00 101.56 vermiculite1 30.92 8.17 23.56 0.00 0.00 10.15 6.60 4.95 4.14 2.88 0.00 0.00 0.00 0.00 0.00 0.00 7.54 1.09 100.00 vermiculite2 29.88 7.33 23.15 0.00 0.00 11.18 8.50 4.93 3.25 3.17 0.00 0.00 0.00 0.00 0.00 0.00 8.04 0.57 100.00 vermiculite3 32.45 24.47 4.09 0.00 0.00 9.96 11.20 1.27 4.08 2.19 0.00 0.00 0.00 0.00 0.00 0.00 7.97 1.61 99.29 vermiculite4 31.58 14.24 24.75 0.00 0.00 6.61 3.15 3.19 4.11 2.41 0.00 0.00 0.00 0.00 0.00 0.00 7.94 2.02 100.00 vermiculite5 31.64 10.47 21.36 0.00 0.00 12.26 3.15 2.39 5.05 3.17 0.00 0.00 0.00 0.00 0.00 0.00 8.18 2.33 100.00 vermiculite6 30.14 11.64 13.00 0.00 0.00 12.92 8.69 0.20 1.33 2.27 0.00 0.00 0.00 0.00 0.00 0.00 11.50 7.05 98.74 vermiculite7 22.44 13.99 3.42 0.00 0.00 9.02 23.97 2.39 0.88 1.22 0.00 0.00 0.00 0.00 0.00 0.00 19.93 2.82 100.08 vermiculite8 26.25 20.76 3.63 0.00 0.00 12.53 15.38 2.54 0.88 1.33 0.00 0.00 0.00 0.00 0.00 0.00 14.54 2.04 99.88 vermiculite9 26.41 20.48 3.76 0.00 0.00 8.24 20.46 2.27 0.82 1.61 0.00 0.00 0.00 0.00 0.00 0.00 13.72 1.88 99.65 vermiculite10 34.10 25.49 4.51 0.00 0.00 12.27 4.97 3.27 1.15 1.97 0.00 0.00 0.00 0.00 0.00 0.00 9.30 2.93 99.96 vermiculite11 21.82 13.70 3.68 0.00 0.00 7.39 24.93 1.41 0.60 0.99 0.00 0.00 0.00 0.00 0.00 0.00 22.01 3.83 100.36 vermiculite12 30.92 18.20 17.16 0.00 0.00 13.65 4.88 0.00 0.00 1.04 0.00 0.00 0.00 0.00 0.00 0.00 14.47 0.00 100.32 vermiculite13 32.15 21.58 12.51 0.00 0.23 4.36 4.89 0.18 1.52 3.01 0.00 0.00 0.00 0.00 0.00 0.00 18.12 0.00 98.55 Barite452 0.50 0.00 0.10 0.00 0.00 0.05 0.60 0.05 0.00 0.00 62.54 1.30 0.00 0.56 33.40 0.00 0.70 0.00 99.80 BariteAlan 4.56 0.17 0.06 0.00 0.00 0.00 1.00 0.00 0.00 0.00 58.74 4.54 0.00 0.05 30.66 0.00 0.08 0.00 99.86 Wo wollastonite from wollastonite-calcite-garnet syenite Epi Epidote from skarn rock from Garigaipalli Aps1, Aps2 and ap1 apatite from Sevvattur carbonatites; ap2,ap450 apatite from Onnakarai carbonatites mt498 magnetitae from Sevvattur carbonatite Ice Icelandspar from skarn rock of Garigaipalli Cc250 Calcite from carbonaitte of Sevvattur Cc310, cc311. cc312, cc313, cc314,

The chemical compositions of the rocks are listed in Tables 1 and 2 on the basis of rock types and their places of occurrence. The chemical compositions of these rocks vary widely in their silica, alkalis and carbonate contents. CO2-rich rock carbonatites, Alkali syenites vary in SiO2 between 35% and 65%, Ultramafic rocks vary in SiO2 between 40% and 52%. Hybrid rocks have both SiO2 and CO2 in an

Carbonates from carbonaites of Jogipatti Cc492, cc493, cc494 carbonates from Carbonatites of Onnakarai cc315 Ben Benstonite from Benstonite carbonatite from Onnakarai Bi Biotite from Biotite pyroxenite verm 1-13 Vermiculite from

Sevvattur carbonatites Ba452 Barite from Onnakarai BaAlan Barite from Alangayam

Table 2. Compositions of various minerals occurring in the carbonatite complex of Tiruppattur

intermediate position.

continental plates of the Eastern Ghats Mobile Belt.

The dolerites are hypabyssal equivalents of gabbro and basalts, which are common products of the partial melting of mantle / crustal rocks at depth. The overlapping petrographic

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 561

Crystallization, Fractionation and Solidification of Co-Magmatic Alkaline

help of detailed petrographic and field investigations.

Edn, New York, p. 416

Longmans, London

India, v. 12, pp. 56-62

Acta, v. 31 (11), pp. 2253-2254

v. 205, p. 587

**11. References** 

Series Sequentially Emplaced in the Carbonatite Complex of Tiruppattur, Tamil Nadu, India 563

variations of all these rocks indicate that they are all derived from the same mantle source by different degrees of partial melting. The composition of the parent magma for this complex appears to be very close in composition to that of shonkinite magma, which might have been derived by liquid fractionation and separation from the low degree of the partial melt of the mantle material. The partial melt might have initially been a crystallized olivine and calcic plagioclase; therefore, the residual shonkinitic parent magma so evolved is impoverished in alumina and in silica and acts as a primary magma in a closed magmatic chamber under volatile enriched conditions. Later on, with early crystallizations of calciumrich clinopyroxenes, the residual magma is impoverished in silica with the enrichment of volatile constituents such as H2O and CO2. In an ascending convection current which motivated a low viscous and high temperature shonkinite magma, clinopyroxenes are crystallized with the enrichment of Ca, Mg, Fe, Tiiv and Aliv and a depletion in Si, Alvi, Na and K in low pressure zones towards the top of the magmatic column. On the other hand, from the top of the magmatic column, sinking clinopyroxenes liberate Ca, Mg, Fe, Tiiv and Aliv and accumulate Na and Alvi in their subsolidus crystal lattices during descending convection currents and towards the bottom of the chamber with the crystallization of subsolidus aegirine. Thus, both salic and mafic constituents in the form of ions are concentrated at the top of the magmatic column in different portions. In this closed magmatic chamber under high PH2O and PCO2, an immiscible separation of camafic carbonate liquids and alkali silicate liquids are derived at depth (Saravanan and Ramasamy, 1995). The separated alkali silicate magmas are emplaced in sequences of co-magmatic bodies of syenites, first in the Sevvattur Basin and subsequently emplaced in the Jogipatti basin followed by carbonate magmas in both (Fig. 1 ad). It is quite a complex matter to study the impact of all these processes by tracing the history of the magmatic evolution inscribed on the rocks over the course of millions of years with the

Barth, T.F.W. (1952) Theoretical Igneous Petrology, Ms Book and Mineral Company, 2nd

Borodin, I.S., Gopal, V., Maralev, V.M. and Subramanian, V. (1971) Precambrian carbonatites of Tamil Nadu, South India, Journal of the Geological Society of India, v. 12, pp. 101-112

Dickson,J.A.D. (1965) A modified staining technique for carbonates in thin sections, Nature,

Grady, J.C. (1971) Deep main faults in South India, Journal of the Geological Society of

von Eckermann, H. (1966) Wollastonite in carbonatite rocks, Geochimica et Cosmochimica

Schieicher, H., Kramm, U., Pericka, E., Schidlowski, M., Schmidt, F., Subramanian, V., Todt,

Macintyre, R. (1975) Age and significance of carbonatites and alkaline intrusive, Mem.

carbonatites, Journal of Petrology, v. 39, (10)m, pp. 1765-85 Krishnan, M.S. (1962) Geology of India and Burma, Higgin Bothams, Chennai

Noticicas. Mus. Lab. Min. Geol. Univ. Coimbra, v. 80, pp. 99-103

W. and Viladkar, S.G. (1998) Enriched sub-continental upper mantle beneath Southern India: Evidence from Pb, Nd, Sr and C-O Isotopic studies on Tamil Nadu

Bowen N.L. (1956) The Evolution of Igneous Rocks, Dover, New York, 2nd Edn, p. 332 Deer, W.A., Howie, R.A. and Zussman, J. (1965) Rock Forming Minerals, v. 1 to v. 5,

Fig. 5. a-j show linear trends of the magmatic evolution of different types of rocks from common parent magma

variations of all these rocks indicate that they are all derived from the same mantle source by different degrees of partial melting. The composition of the parent magma for this complex appears to be very close in composition to that of shonkinite magma, which might have been derived by liquid fractionation and separation from the low degree of the partial melt of the mantle material. The partial melt might have initially been a crystallized olivine and calcic plagioclase; therefore, the residual shonkinitic parent magma so evolved is impoverished in alumina and in silica and acts as a primary magma in a closed magmatic chamber under volatile enriched conditions. Later on, with early crystallizations of calciumrich clinopyroxenes, the residual magma is impoverished in silica with the enrichment of volatile constituents such as H2O and CO2. In an ascending convection current which motivated a low viscous and high temperature shonkinite magma, clinopyroxenes are crystallized with the enrichment of Ca, Mg, Fe, Tiiv and Aliv and a depletion in Si, Alvi, Na and K in low pressure zones towards the top of the magmatic column. On the other hand, from the top of the magmatic column, sinking clinopyroxenes liberate Ca, Mg, Fe, Tiiv and Aliv and accumulate Na and Alvi in their subsolidus crystal lattices during descending convection currents and towards the bottom of the chamber with the crystallization of subsolidus aegirine. Thus, both salic and mafic constituents in the form of ions are concentrated at the top of the magmatic column in different portions. In this closed magmatic chamber under high PH2O and PCO2, an immiscible separation of camafic carbonate liquids and alkali silicate liquids are derived at depth (Saravanan and Ramasamy, 1995). The separated alkali silicate magmas are emplaced in sequences of co-magmatic bodies of syenites, first in the Sevvattur Basin and subsequently emplaced in the Jogipatti basin followed by carbonate magmas in both (Fig. 1 ad). It is quite a complex matter to study the impact of all these processes by tracing the history of the magmatic evolution inscribed on the rocks over the course of millions of years with the help of detailed petrographic and field investigations.

## **11. References**

562 Crystallization – Science and Technology

Fig. 5. a-j show linear trends of the magmatic evolution of different types of rocks from

common parent magma


Mysen, B. (1975) Partitioning of iron and magnesium between crystals and partial melts in

Mysen, B.O., Virgo, D. and Seifert, F.A. (1985) Relationships between properties and structure of aluminosilicate melts, Am. Mineralogist, v. 70, pp. 88-105 Naidu, P.R.J. (1958) 4-Axes Universal Stage, Dept. of Geology, Madras University

Ramasamy, R. (1973) Geology of the area South west of Tiruppattur, Madras State (Tamil

Ramasamy, R. and Shapenko, V. (1980) Fluid inclusion studies in carbonatites of Tiruppattur, Tamil Nadu , India (Abst) Int. Geol. Cong. Paris , v. 1, p. 79 Ramasamy, R. (1981) Tectonomagmatic Episodes of the Peninsular India, Sympet, Jaipur,

Ramasamy, R. (1982a) A possible paleo-rift system of the eastern Ghats in the Peninsular India, J. Moscow State Univ., Geology, Ser. 4, No. 2, pp. 32-37 ( in Russian) Ramasamy, R. (1982) Structure and tectonics of carbonatite complex of Tiruppattur, Tamil

Ramasamy, R. (1986a) Ca-rich pyroxenes from the carbonatite complex of Tiruppattur,

Ramasamy, R. (1987) Reactivation of the Eastern Ghats Paleorift system during Tertiary and other periods, Proc. National, Seminar Tertiary Orogeny, pp. 109-127. Ramasamy, R., Gwalani, L.G. and Subramanian, S.P. (2001) A note on the occurrence and

Nadu, Southern India Journal of Asian Earth Sciences, v. 19, pp. 297-304 Ramasamy, R., Subramanian, S.P. and Sundaravadivelu, R. (2010) Compositional variations

Saravanan, S. and Ramasamy, R. (1995) Geochemistry and petrogenesis of shonkinite and

Schairer, J.F. and Yoder, H.S.J.R., (1960) The nature of residual liquids from crystallization

Udas, G.R. and Krishnamurthy, P. (1970) Carbonatites of Sevvattur and Jogipatti, Madras State, India, Proc. Of the Indian National Academy of Science, 36A, pp. 331-343 Verwoerd, W.J. (1966) South African Carbonatites and their probable mode of origin, Ph.D.

Viladkar, S.G. and Subramanian, V. (1995) Mineralogy and geochemistry of the Sevvattur

Winchell, A. N. (1945) Elements of Optical Mineralogy, part II Optical properties of Minerals Yagi, K. (1953) Petrochemical studies on the alkaline rocks of the Morotu District, Sakhalin,

Yoder, H.S., Stewart, D.B. and Smith, J.R. (1957) Ternary feldspars, Carnegie Inst.

of Tiruppattur, Tamil Nadu, Current Science, v. 55, pp. 1026-29

Tamil Nadu, Current Science, v. 55, pp. 981-984

Nadu, India, Mineralogical Magazine, v. 38, pp. 376-377

of the Geological Society of India, v. 46, pp. 235-243

Volume, v. 258A, pp. 273-83

Thesis, Univ. of Stellenboch

v. 45, pp. 505-518

GSA Bull, v. 64, 769-810

Washington Year Book, pp. 206-214

Nadu In: Bhattacharya, A./K. (Ed.) Current Trends in Geology, IV Indian Geological Congress, Today and Tomorrows Publishers, New Delhi, pp. 119-136 Ramasamy, R. (1986) Titanium-bearing garnets from alkaline rocks of carbonatite complex

formation of magnetite in the carbonatites of Sevvattur, North Arcot district, Tamil

of olivine in shonkinite and its associate ultrabasic rock from the carbonatite complex of Tiruppattur, Tamil Nadu, Current Science, v. 99(10), pp. 1428-1433 Rittmann, A. (1973) Stable Mineral Assemblage of Igneous Rocks, Springer, Berlin, p. 251 Saravanan, S. and Ramasamy, R. (1971) A magnesio-riebeckite from Samalpatti area, Tamil

associated alkaline rocks of Tiruppattur carbonatite complex, Tamil Nadu, Journal

data on the system Nepheline-diopside-silica, American Journal of Science, Bradley

and Samalpatti complexes, Tamil Nadu, Journal of the Geological Society of India,

Peridotites Upper Mantle, Contrib, Mineral Petrol., v. 52, pp. 69-76

Nadu), India, Ph.D Thesis, p. 226, University of Madras

Geol. Surv. India, Abstract L2-L3.

## *Edited by Marcello Rubens Barsi Andreeta*

Crystallization is one of the most ancient and interdisciplinary topics of research known to mankind. Crystals can be organic or inorganic and may be produced from melts, liquid solutions, vapors or even in solid state. Notwithstanding its inherently high complexity, the crystallization process is part of our everyday lives, from ice making in our homes to the most state-of-the-art chemical and electronic industry. In this book, our purpose was to present new insights to the reader, as well as crucial and very useful information for researchers working in this field, while simultaneously creating a comprehensive text about crystallization processes which may serve as a starting point for people with different backgrounds.

Photo by MarinaMariya / iStock

Crystallization - Science and Technology

Crystallization

Science and Technology