The influence of hydrothermal synthesis conditions on

3 downloads 0 Views 2MB Size Report
Mar 25, 2011 - area SBET = 38.28 m2/g, the radius of dominant plate pores rp = 30–40 ... a-C2S hydrate and C–S–H(II) prevail in the ... pure gyrolite forms after 16 h of isothermal curing at. 200°C. ... The surface area, total pore volume and pore size ..... synthesis (h). Sample mass, m (g). BET equation constants. Capacity ...
The influence of hydrothermal synthesis conditions on gyrolite texture and specific surface area K. Baltakys, A. Eisinas, T. Dizhbite, L. Jasina, R. Siauciunas & S. Kitrys

Materials and Structures ISSN 1359-5997 Volume 44 Number 9 Mater Struct (2011) 44:1687-1701 DOI 10.1617/s11527-011-9727-8

1 23

Your article is protected by copyright and all rights are held exclusively by RILEM. This eoffprint is for personal use only and shall not be self-archived in electronic repositories. If you wish to self-archive your work, please use the accepted author’s version for posting to your own website or your institution’s repository. You may further deposit the accepted author’s version on a funder’s repository at a funder’s request, provided it is not made publicly available until 12 months after publication.

1 23

Author's personal copy Materials and Structures (2011) 44:1687–1701 DOI 10.1617/s11527-011-9727-8

ORIGINAL ARTICLE

The influence of hydrothermal synthesis conditions on gyrolite texture and specific surface area K. Baltakys • A. Eisinas • T. Dizhbite L. Jasina • R. Siauciunas • S. Kitrys



Received: 23 June 2010 / Accepted: 15 March 2011 / Published online: 25 March 2011  RILEM 2011

Abstract Influence of hydrothermal synthesis conditions on the gyrolite specific surface area, dominant pore size and their differential distribution by the radius were determined. The synthesis of gyrolite has been carried out in unstirred suspensions within 32, 48, 72, 120, 168 h at 200C temperature from a stoichiometric composition (the molar ratio of CaO/ SiO2 was equal to 0.66 where water/solid ratio of the suspension was equal to 10.0) of the initial CaO and SiO2nH2O mixture. It was found that the structure of gyrolite and the shape of dominated pores (from pores between parallel plates to cylindrical pores) changes prolonging the duration of hydrothermal synthesis. The stable gyrolite crystal lattice was formed only after 120 h of isothermal curing. Its specific surface area SBET = 38.28 m2/g, the radius of dominant plate ˚ , the cumulative pore volume pores rp = 30–40 A 3 RVp = 0.08 cm /g. It was determined that the pores K. Baltakys (&)  A. Eisinas (&)  R. Siauciunas Department of Silicate Technology, Kaunas University of Technology, Radvilenu 19, LT-50270 Kaunas, Lithuania e-mail: [email protected] A. Eisinas e-mail: [email protected] T. Dizhbite  L. Jasina Latvian State Institute of Wood Chemistry, Dzerbenes 27, Riga LV-1006, Latvia S. Kitrys Department of Physical Chemistry, Kaunas University of Technology, Radvilenu 19, LT-50270 Kaunas, Lithuania

with 4.0–5.0 nm radius were dominated in gyrolite structure after 168 h of synthesis. It was estimated that the ion exchange between gyrolite with less orderly structure in Zn(NO3)2 ? NH4OH alkaline solution (cZn2þ —0.3 g/dm3) proceeds more faster and effectively. Keywords Gyrolite  Ion exchange reactions  Calcium silicate hydrate  BET analysis

1 Introduction Calcium silicate hydrates are highly multiplex system with over 30 stable phases. Most of them occurring in nature as hydrothermal alteration products of calcium carbonate rocks and as vesicle fillings in basalts, include many chemically and structurally distinct phases [17]. From a theoretical and practical point of view, the synthesis, properties and structure of the main calcium silicate hydrates—C–S–H(I), 1.13 nm tobermorite, xonotlite, a-C2S hydrate have been analyzed in detail [13, 15, 18, 24, 25, 27, 29, 33]. Majority of these compounds are occurring naturally or may be synthesized in the laboratory. Recently, the interest of gyrolite group compounds (gyrolite, Z-phase, truscottite, reyerite) increases because the new possibilities of application were found: it may be used to educe heavy metal ions and remove them from wastewaters [5, 12, 19]. It was show that gyrolite sorption properties is greater than

Author's personal copy 1688

tobermorite group minerals [7, 8, 28]. Of specific interest is the purported ability of gyrolite to separate supercoiled plasmid, open circular plasmid, and genomic DNA [34]. Gyrolite can be synthesized from CaO and various forms of SiO2 with the molar ration CaO/SiO2 (C/ S = 0.66) in aqueous suspension at temperatures of about 200C. Kalousek and Nelson [11], and also Stevula and Petrovic [29] found that gyrolite could likewise be prepared by interacting dicalcium silicate (2CaOSiO2) with SiO2 in aqueous suspension under hydrothermal conditions. Stevula et al. were found that over the temperature range of 200–300C under hydrothermal conditions, both the natural and synthetic gyrolite behave analogously. Above this temperature, both natural and synthetic gyrolite decompose, forming the stable phases truscottite and xonotlite [30]. Okada et al. using lime and amorphous silica as the starting materials, hydrothermally prepared gyrolite with the C/S molar ratio of 0.66 and 0.50 at 200C for 0.5–128 h [22, 23]. Jauberthie et al. determined that the tobermorite gel, formed by hydrothermal reaction of silica and lime, is transformed either into Z-phase if the quantity of lime is less than 37% or into 1.0 nm tobermorite if the quantity of lime is between 37 and 42%. The 1.0 nm tobermorite is stable in the presence of gyrolite, whereas Z-phase is metastable [10]. Baltakys and Siauciunas [3, 27] were determined the influence of SiO2 modification on crystallization process of gyrolite. Authors showed that gyrolite does not form even during a week in the mixtures of CaO and amorphous SiO2 at 150C under the saturated steam pressure. The temperature increase positively affects synthesis of this compound—pure gyrolite is produced at 175C after 72 h, and after 32 h—at 200C. It should be underlined that in the mixtures with quartz the mechanism of compound formation is quite different. Due to a low quartz solubility rate at temperature range from 150 to 200C, neither Z-phase, nor gyrolite is formed even after 72 h of hydrothermal curing. a-C2S hydrate and C–S–H(II) prevail in the beginning of the synthesis and gradually recrystallizes into 1.13 nm tobermorite and xonotlite. Shaw et al. [26] having used the synchrotron X-ray radiation source of high energy have explored the mechanical, kinetic, and energetic processes that are

Materials and Structures (2011) 44:1687–1701

proceeding during the formation of gyrolite. The formation of gyrolite in the temperature range from 190 to 240C in the pure calcic system involves a three stage process: amorphous gel ? C–S–H gel ? Z-phase ? gyrolite. Although often it is stated that the crystal lattice of the gyrolite found in nature always has both sodium and aluminium ions [9, 16]. There is a little data in the references about the influence of Al2O3 and Na2O additives on the synthesis of low base calcium silicate hydrates (in contrast to 1.13 nm tobermorite) [19–21, 32]. It should be noticed that Miyake et al. [19] successfully synthesized (Al ? Na)-substituted gyrolite (Ca8Si11.32Al0.68Na0.44O30(OH)46.6H2O) and used it for the ions exchange reactions (K? and Cs?) in aqueous solutions. Stumm et al. [31] have indicated that zinc incorporation into synthetic gyrolite is also possible up to Zn/(Zn ? Ca) = 1/6, corresponding to approximately 6 wt%. Increasing zinc content led to a gradual diminishing of the basal reflection (001) of gyrolite, as for the nanocrystalline phases. Baltakys and Siauciunas [1] proved that the formation of gyrolite is accelerated by mixing suspension because pure gyrolite forms after 16 h of isothermal curing at 200C. Stirring affects the sequence of intermediate compounds: gyrolite crystallizes together with Z-phase. Also, 5% Na2O additive in the stirred suspensions significantly accelerates the synthesis of gyrolite: this compound dominates already after 6 h at 200C. The structure, optical properties and chemistry of natural gyrolite were studied by many scientists [2, 4, 16, 26]. However, their opinions differ. A full structural solution for gyrolite was proposed by Merlino. However, some properties (like sorption capacity) depend not only on the crystal lattice of a porous body but also on that of the surface microstructure and specific surface area, as well as on the dominant pore size and their differential distribution in the compound according to the radius. In the case of gyrolite crystal lattice, these properties usually depend on the propor2 ; S2 Þ and amorphous (X tion of crystalline ðS1 ; S2 ; S sheet) parts. However, no data were found in references about the influence of synthesis conditions on gyrolite crystal lattice. The aim of this work was to determine the influence of hydrothermal synthesis conditions on

Author's personal copy Materials and Structures (2011) 44:1687–1701

the gyrolite specific surface area, as well as dominant pore size and their differential distribution according to the radius in this compound. Also, the application of gyrolite in the ion exchange reaction is discussed.

2 Materials and methods 2.1 Materials and manufacture of synthetic gyrolite In this work the following reagents were used as starting materials: fine-grained SiO2nH2O (Reachim, Russia, ignition losses 21.43%, specific surface area Sa = 1,560 m2/kg by Blaine’s); CaO (97.6% Reachim, Russia) additionally was burned at 950C for 0.5 h. The synthesis of gyrolite has been carried out in unstirred suspensions in the vessels of stainless steel. Pure gyrolite was synthesized within 32, 48, 72, 120, 168 h at 200C temperature from a stoichiometric composition (the molar ratio of CaO/SiO2 was equal to 0.66 where water/solid ratio of the suspension was equal to 10.0) of the initial CaO and SiO2nH2O mixture. These synthesis conditions were chosen according to previously published data [3, 27]. The products of the synthesis have been filtered, rinsed with ethyl alcohol to prevent carbonization of materials, dried at the temperature of 50 ± 5C and sieved through a sieve with a mesh width of 50 lm. 2.2 Analytical techniques The X-ray powder diffraction data were collected with a DRON-6 X-ray diffractometer with Bragg– Brentano geometry using Cu Ka radiation and graphite monochromator, operating with the voltage of 30 kV and emission current of 20 mA. The stepscan covered the angular range 2–60 (2h) in steps of 2h = 0.02. The computer program X-fit was used for calculation of crystallite size and for diffraction profile refinement under the pseudoVoid function and for a description of the diffractional background under the third degree of Tchebyshev polynom, we have used fundamental parameters peak profiling (a computer program X-fit) [6]. The surface area, total pore volume and pore size distribution of the synthesis products were performed by a BET surface area analyzer ‘‘KELVIN 1042 Sorptometer’’ (Costech Instruments).

1689

2.3 Methodology 2.3.1 Specific surface area from the BET equation The specific surface area of gyrolite was calculated by the BET equation using the data of the lower part of N2 adsorption isotherm (0.05 \ p/p0 \ 0.35): X



1 p0 p

1



C1 p 1 ;  þ X m  C p0 X m  C

ð1Þ

where X is the mass of adsorbate, adsorbed on the sample at relative pressure p/p0, p the partial pressure of adsorbate, p0 the saturated vapor pressure of adsorbate, Xm the mass of adsorbate adsorbed at a coverage of one monolayer, C is a constant which is a function of the heat of the adsorbate condensation and heat of adsorption (CBET is a constant). BET equation yields a straight line when 1/X[(p0/ p) - 1] is plotted versus p/p0. The slope of (C - 1)/ XmC and the intercept of 1/XmC was used to determine Xm and C: S = slope = (C - 1)/XmC and I = intercept = 1/XmC. Solving for Xm yields Xm = 1/(S ? I). BET plot is usually found to be linear in the range p/p0 = 0.05–0.35. The total surface area of the sample St was determined by the equation St = XmNAcs/M, where M is the molecular mass of the adsorbate, N the Avogadro’s constant, Acs the crosssectional area occupied by each nitrogen molecule (16.2 9 10-20 m2). The specific surface area was given by the equation SBET = St/m, where m is the mass of gyrolite sample. 2.3.2 Classification of hysteresis loops It is widely accepted that there is a correlation between the shape of the hysteresis loop and the texture (e.g., pore size distribution, pore geometry, connectivity) of a mesoporous materials. An empirical classification of hysteresis loops was given by the IUPAC [14], which is based on an earlier classification by de Boer. The IUPAC classification is shown in Fig. 1. According to the IUPAC classification type H1 is often associated with porous materials consisting of well-defined cylindrical-like pore agglomerates of compacts of approximately uniform spheres. It was found that give rise to H2 hysteresis are often disordered and the distribution of pore size and shape is not well defined. Isotherms revealing type H3

Author's personal copy 1690

Materials and Structures (2011) 44:1687–1701

ln

p cVm cosh ¼ 2 ; p0 RTrk

ð2Þ

where p is the saturated vapor pressure in equilibrium with the adsorbate condensed in a capillary or pore, p0 the normal adsorbate saturated vapor pressure, c the surface tension of nitrogen at its boiling point (c = 8.85 ergs/cm2 at (-195.8C)) the molar volume of liquid nitrogen (Vm = 34.7 cm3), h the wetting angle (usually taken 0 and cosh = 1), R the gas constant (R = 8.134 9 107 ergs deg-1 mol-1), T the absolute temperature (T = 77 K) and rK the Kelvin radius of pore. When nitrogen is used as adsorbate Kelvin equation can be rearranged: 4:146 rk ¼ : ð3Þ lg pp0

Fig. 1 IUPAC classifications of hysteresis loops

hysteresis do not exhibit any limiting adsorption at high p/p0, which is observed with non-rigid aggregates of plate-like particles giving rise to slit-shaped pores. The desorption branch for type H3 hysteresis contains also a steep region associated with a forced closure of the hysteresis loop, due to the so-called tensile strength effect. Similarly, type H4 loops are also often associated with narrow slit pores. 2.3.3 Pore volumes and size distributions Vapors in equilibrium with liquid contained in fine capillaries or pores will have depressed vapor pressures. In fact, if the pore is adequately small in diameter, the vapor will condense at pressures far below normal. As indicated by the Kelvin equation, nitrogen gas will condense into all pores with radius less than 150 nm at a relative pressure of 0.99. By measuring the volume of nitrogen adsorbed at a relative pressure of 0.99 and with prior knowledge of the surface area, the average pore radius can be calculated. The total pore volume and pore size distribution were calculated according to the corrected Kelvin equation and Orr et al. scheme using entire N2 desorption isotherm at 77 K [14]. The Kelvin equation relates the adsorbate vapor pressure depression to the radius of a capillary which has been filled with adsorbate:

The Kelvin radius rK is not the actual pore radius because some adsorption takes place on the wall of the pore prior to the occurrence of condensation in the pore. During desorption an adsorbed layer remains on the wall when evaporation takes place. Therefore, the true pore radius rp was calculated by the equation rp = rK ? t. Theoretically t value is, by assuming that for any value of relative pressure the thickness of the adsorbed film on pore walls is the same as the thickness of the adsorbed layer on a plane surface, the thickness of the adsorbed layer is given by equation: Va t¼ s; ð4Þ Vm where t is the thickness of the adsorbed layer on the pore wall (mm), Va is the volume of adsorbed gas (mm3 adsorbate per gram of adsorbent), s is the thickness of a monomolecular layer of adsorbent (mm). t was calculated according to the Halsey equation [14], which for nitrogen can be written as: " #1=3 5 t ¼ 3:54 : ð5Þ 2:303 lg pp0 The values of 5 and 3.54 in equation are empirical, where 3.54 is the thickness of one adsorbed nitrogen ˚ is somewhat less that the layer. This value of 3.54 A diameter of a nitrogen molecule based on the cross˚ 2. sectional area of 16.2 A If the pores are assumed to be cylindrical and if the relative pressure is changed from (p/p0)2 to (p/p0)1, then pores between radii r2 and r1 will empty (p2 \ p1,

Author's personal copy Materials and Structures (2011) 44:1687–1701

1691

r2 \ r1). When p2 is lowered to p1 the thickness of the adsorbed film on previously emptied pores changes from t2 to t1. The volume of liquid (VL) evaporated due to emptying of pores and the decrease in thickness of adsorbed layers in previously empty pores was calculated by the following equation: X  2 VL ¼ p r p  t2 L þ ðt2  t1 Þ A; ð6Þ where r p is the average pore radius in the interval r2 to r1, L the total length of all pores in the range r2 to r1, A the area of adsorbed film remaining in the pores after evaporation out of the pores has occurred. The mean volume of pores with r p is Vp ¼ pr 2p L: Then  2  2 X rp rp Vp ¼ VL  ðt2  t1 Þ A: r p  t2 r p  t2 ð7Þ Cylindrical using rp = rK ? t:  2  X  rp Vp ¼ VL  Dt A : rK

ð8Þ

The volume of liquid desorbed in any interval of desorption isotherm is related to the volume of gas by the equation: DVL (cm3) = DVgas (1.54 9 10-3). For cylinders A was given by   2DVp Ac m 2 ¼  104 ; rp

ð9Þ

˚. with Vp in cm3 and rp in A For the calculations of pores between parallel plates we used the following equation: dp = rk ? 2t; rk = l, where l is a distance between plates: dp ¼ rK þ 2t;

ð10Þ

˚. dp —distance between two plates, A X  dp  DVL  2Dt A ; rK   2Vp Ap m 2 ¼ : dp

Vp ¼

ð11Þ ð12Þ

The analysis of desorption isotherms was finished when (DtRA) exceeded DVL, because the desorbed gas is coming from an adsorbed layer rather than from evaporation of the liquid out of the pore center. RA is the cummulative surface area and is obtained by summing A in each radius interval.

3 Results and discussion It was determined that a stable monolayer of absorbed N2 was formed on the surface of gyrolite. The BET equation gives a linear plot in the range of relative pressures 0.05 B p/p0 B 0.30 (Figs. 1, 2). Straight lines were obtained for all gyrolite samples in BET coordination ð1=ðX ½ðp0 =pÞ  1ÞÞ  ðp0 =pÞ: A straight line correlation coefficient R2 of synthesized after 32 h hydrothermal treatment gyrolite was equal to 0.9986 (Fig. 2a), the same as after 48 h (Fig. 2b). A straight lines correlation coefficients R2 remains very close to the unit, i.e. 0.9991, when hydrothermal synthesis was prolonged up to 168 h (Fig. 3). It was noted that duration of hydrothermal synthesis has influence on both gyrolite structure and specific surface area (Table 1). It was determined that SBET was equal to 82.07 m2/g after 32 h of isothermal curing of gyrolite. Prolonging the hydrothermal synthesis duration to 48 h, SBET of the product also increased to 91.52 m2/g. It should be noted that SBET characteristic for gyrolite decreased more than two times to 43.54 m2/g already after 72 h of synthesis. Besides, this tendency was observed also after 120 h because SBET was equal to 38.28 m2/g. Meanwhile, after 168 h of isothermal curing SBET value of gyrolite slightly increased (46.00 m2/g). It is likely that one of the reasons was gyrolite crystal structure, where water content can varies from 12 to 20% weight [2, 26, 31]. Also, depending on the synthesis conditions crystallite size of this compound varies ˚ (Fig. 4). from 1,066 to 2,156 A Calculated CBET constant values showed the structural changes of gyrolite during hydrothermal synthesis. Certainly, CBET values varied in the range from 50 to 250 when experimental conditions were ideal, while lower constant values meant that adsorbate was condensed into pores and calculated SBET should be higher than the real one. In contrarily, when CBET [ 250 chemical reaction between adsorbent surface and adsorbate might occur without formation of adsorbate monolayer. Obtained results confirmed Merlino [16] suggested gyrolite lattice structure which is consisted of octahedral Ca–O layers, interlayers (Ca2? and H2O) and silicate layers. Based on the ‘‘OD’’ theory, Merlino determined that gyrolite is consisted of different layers (Fig. 5):

Author's personal copy 1692 Fig. 2 The isotherm of N2 adsorption by gyrolite at 77 K in BET plot. Duration of hydrothermal synthesis at 200C temperature, h: a 32, b 48

Materials and Structures (2011) 44:1687–1701

a

b

500 y = 1635.3x - 1.9796 R2 = 0.9986

400

1 ⎡⎛ p 0 ⎞ ⎤ 200 X ⎢⎜⎜ ⎟⎟ − 1⎥ ⎢⎣⎝ p ⎠ ⎥⎦ tg α

tg α

100

0

0 0.00

a

y = 1188.7x - 0.7088 R2 = 0.9986

300

1 300 ⎡⎛ p 0 ⎞ ⎤ X ⎢⎜⎜ ⎟⎟ − 1⎥ ⎢⎣⎝ p ⎠ ⎥⎦ 200 100

Fig. 3 The isotherm of N2 adsorption by gyrolite at 77 K in BET plot. Duration of hydrothermal synthesis at 200C temperature, h: a 72, b 120, and c 168

400

0.06

0.12

0.18 p p0

0.24

0.30

600 500

400 1 ⎡⎛ p 0 ⎞ ⎤ 300 ⎟⎟ − 1⎥ X ⎢⎜⎜ ⎣⎢⎝ p ⎠ ⎦⎥

0.00

b

0.18 p p0

0.24

0.30

y = 1867.3x + 4.9717 R2 = 0.9992

400 1 ⎡⎛ p 0 ⎞ ⎤ 300 ⎟⎟ − 1⎥ X ⎢⎜⎜ ⎣⎢⎝ p ⎠ ⎦⎥

200

200

tg α

tg α

100

100

0

0 0.00 0.06 0.12 0.18 0.24 0.30

0.00 0.06 0.12 0.18 0.24 0.30

p p0

p p0

c

0.12

600 500

y = 1789.2x + 11.426 R2 = 0.9994

0.06

700 600 500

y = 2095.7x + 7.4048 R2 = 0.9991

1 400 ⎡⎛ p 0 ⎞ ⎤ ⎟⎟ − 1⎥ 300 X ⎢⎜⎜ ⎣⎢⎝ p ⎠ ⎦⎥ 200

tg α

100 0 0.00 0.06 0.12 0.18 0.24 0.30

p p0







A centrosymmetric layer S1, which composition is (Si8O20)8-, consists of two types of six-membered rings; Double tetrahedra layer S2, which composition is (Si14Al2O38)14-. There are the other two types of six-membered rings in the layer; Octahedral layer (O), which composition is [Ca7O10(OH)4]4-, consists calcium octahedra combined by the edge;



X interlayers which are consisted of H2O and O2-, coordinated Ca and Na octahedra (coordinate bond).

It was proved that structural elements of gyrolite lattice after 32 and 48 h of synthesis were combined with undefined composition (Ca2?, H2O) of X interlayer. Due to this it isn’t coincidentally that CBET constant values are equal to 827.08 and 1678.06

Author's personal copy Materials and Structures (2011) 44:1687–1701

1693

Table 1 Calculated parameters of gyrolite specific surface area (SBET) Capacity of monolayer 1 Xm ¼ SþI ðgÞ

SBET (m2/g)

CBET ¼ IX1 m

Reliability coefficient R2

5

6

7

8

1.9796

0.000611

82.07

827.08

0.9986

0.7088

0.000841

91.52

1678.06

0.9986

1789.20

11.4260

0.000555

43.54

157.59

0.9994

0.0490

1867.30

4.9717

0.000534

38.28

376.59

0.9992

0.0363

2095.70

7.4048

0.000475

46.00

284.02

0.9991

Duration of synthesis (h)

Sample mass, m (g)

BET equation constants

1

2

Slope S = tga 3

32

0.0262

1635.30

48

0.0323

1188.70

72

0.0448

120 168

Intercept I 4

Crystallite size, Å

2200

1800

1400

5 1000

24

48

72

96

120

144

168

Duration of synthesis, h Fig. 4 Dependence of gyrolite crystallite size on duration of hydrothermal synthesis

Intensity, a. u.

600

4

3

2

1

2

3

4

5

6

2 θ, deg. Fig. 6 X-ray diffraction curves of gyrolite. Duration of hydrothermal synthesis at 200C temperature, h: 1 32, 2 48, 3 72, 4 120, and 5 168

Octahedral Ca-O layers Interlayers (Ca2+ and H2O) Sicate layers

Fig. 5 The crystal lattice structure of gyrolite (by Merlino [16])

(Table 1). Meanwhile CBET constant values of the product after 72, 120 and 168 h synthesis decreased 4–5 times. This allows to state that after sufficiently long synthesis duration the structural elements of gyrolite lattice were connected to more orderly structure of the interlayer X. This assumption is

Author's personal copy 1694

Materials and Structures (2011) 44:1687–1701

and overlap in a slightly higher pressure p/p0. The isoterms of adsorption–desorption coincided when relative pressure p/p0 & 0.70, which seems to be more typical to the cylindrical slits (Fig. 8). It allows to state that during the isothermal curing changes both the structure of synthesized products and the form of pores, i.e. cylindrical pores started to dominate. Before calculation of distribution of pores to diameters it is necessary to identify the common model of a dominating form of pores. These models are selected according to the form of hysteresis loop (Figs. 1, 7, and 8). Theoretically, SBET value is equal to the value of RA when the form of pore is uniform. Meanwhile, during hydrothermal synthesis the particles of various diameters are formed in the products. For this reason, in polydisperse systems are determined only dominant pores. The most suitable model of pore distribution is the one whose experimentally defined SBET value is the closest to the value of the calculated specific surface RA. Later on, according to model of dominate pores is calculated the dominant pore volume, size and their differential distribution according to the radius. Therefore, the distribution of pores to the diameters was calculated according to two models (cylinder pores and pores between parallel plates) (Tables 2, 3). The obtained results showed that none of the selected models suits for a description of the pores of the product which was synthesized after 32 h of isothermal curing. These data show that the structure of pores

confirmed by XRD analysis: prolonging the isothermal curing duration the basic diffraction maximum intensity and area increases (Fig. 6). The results showed that all samples were characterized by gyrolite hysteresis phenomenon: adsorption and desorption isotherms do not coincide, desorption isotherm was more in the left than adsorption isotherm (Figs. 7, 8). It is characteristic to mesoporous solids when the radius of pores was in ˚ . Besides, it has a clear bend a range from 15 to 500 A in a lower p/p0 area (point A). Microporous sorbents do not have point A in the interval 0.05 B p/ p0 B 0.30. Thus, it should be stated that during all experimental conditions formed gyrolite is mesoporous compound (Figs. 7, 8). Assumption about shape of the pores was made on the basis of nature of N2 adsorption–desorption isotherm (hysteresis) and taking into account the abundant information on the structure of the gyrolite. Right side of the hysteresis loop would correspond to the isotherms that are obtained in the case of materials with pores formed between parallel crystal planes. However, the isotherms of adsorption and desorption must coincide in the range of relative pressures p/p0 & 0.30, moreover the curve of desorption isotherm falls very steeply (Fig. 7). This indicates that in the product structure was dominated by disordered, the plate form pores. Meanwhile, prolonging synthesis duration to 1 week, the isotherms were much narrower

a

120

b 320

100

280 240

Vads, cm3/g

80

Vads, cm3/g

Fig. 7 The isotherm of N2 adsorption–desorption by gyrolite at 77 K. Duration of hydrothermal synthesis at 200C temperature, h: a 32, b 48

60

200 160 120

40

80

A

20 A

40 0

0

0.0

0.2

0.4

p/p o

0.6

0.8

1.0

0.0

0.2

0.4

p/p o

0.6

0.8

1.0

Author's personal copy Materials and Structures (2011) 44:1687–1701 Fig. 8 The isotherm of N2 adsorption–desorption by gyrolite at 77 K. Duration of hydrothermal synthesis at 200C temperature, h: a 72, b 120, and c 168

a

1695

80

b 50

70

60

40

Vads, cm3/g

Vads, cm3/g

50

40

30

30

20

20

A A

10

10

0

0 0.0

0.2

0.4

0.6

0.8

1.0

c

0.0

0.2

0.4

0.6

0.8

1.0

p/p o

p/p o 60

50

Vads, cm3/g

40

30

20 A

10

0 0.0

0.2

0.4

0.6

0.8

1.0

p/p o

is not well formed. For this reason, we performed calculations of products after 48–168 h of synthesis. It was determined that the pores of gyrolite after 48 h of synthesis were formed between parallel planes, because SBET = 91.52 m2/g and RA = 99.33 m2/g. Meanwhile, by using a cylindrical model the error between SBET and RA exceeds higher value (SBET = 91.52 m2/g and RA = 150.72 m2/g) (Table 2). Prolonging the synthesis duration (168 h) pores shape was better described by a cylindrical model, because

calculated RA values was closer (Table 3). This phenomenon was confirmed also by the shape of isotherm which narrowed (Fig. 8). It was determined that the pore volume of the gyrolite samples with characteristic orderly crystal structure decreased from 110 to 80 mm3/g (Fig. 9). Also, the obtained results showed that the stable gyrolite crystal lattice was formed only after 120 h of isothermal curing because cumulative pore volume is not influenced by synthesis conditions.

N2 volume adsorbed on the Vad (ncm3/g)

Kelvin Thickness True radius of of N2 pore ˚ ) radius the pores, layer, t (A ˚) ˚) rK (A (A

52.29

46.69

34.83

32.66

31.84

0.81

0.78

0.75

0.70 0.65

0.60

0.55

3

4

5

6 7

8

9

10 0.50

11 0.45

12 0.43

13 0.40

10.51

11.42

12.07

13.90

16.14

18.89

27.06 22.40

33.58

38.84

45.83

55.51

98.25

59.41 52.29

46.69

34.83

32.66

31.84

0.81

0.78

0.75

0.70

0.65

0.60 0.55

3

4

5

6

7

8 9

10 0.50

11 0.45

12 0.43

13 0.40

67.51

79.79

95.21

106.52

10.51

11.42

12.07

13.90

18.89 16.14

22.40

27.06

33.58

38.84

45.83

55.51

98.25

6.25

6.43

6.54

6.86

7.60 7.21

8.04

8.57

9.21

9.66

10.21

10.88

13.17

6.25

6.43

6.54

6.86

7.21

7.60

8.57 8.04

9.21

9.66

10.21

10.88

13.17

16.76

17.85 10.97

18.61 11.75

20.76 12.98

26.48 17.51 23.35 15.02

30.44 20.64

35.63 24.73

42.79 30.32

48.50 36.21

56.04 42.33

66.40 50.67

111.42 76.88

23.01

24.27 10.97

25.16 11.75

27.62 12.98

30.56 15.02

34.08 17.51

44.19 24.73 38.49 20.64

51.99 30.32

58.18 36.21

66.26 42.33

77.28 50.67

124.58 76.88

Average Kelvin radius of the pores, rK ˚) (A

17.30

18.23

19.69

24.92 22.06

28.46

33.03

39.21

45.64

52.27

61.22

88.91

23.64

24.72

26.39

29.09

32.32

41.34 36.28

48.09

55.08

62.21

71.77

100.93

0.18

0.12

0.32

0.39 0.35

0.44

0.52

0.64

0.46

0.55

0.67

2.28

0.18

0.12

0.32

0.35

0.39

0.52 0.44

0.64

0.46

0.55

0.67

2.28

Average Change, ˚) true pore Dt (A radius, rp ˚) (A

0.82

2.17

11.86

7.11 5.60

8.11

12.27

15.42

11.31

17.12

23.29

80.04

0.82

2.17

11.86

5.60

7.11

12.27 8.11

15.42

11.31

17.12

23.29

80.04

1.27

3.34

18.27

10.95 8.63

12.49

18.90

23.75

17.42

26.36

35.86

123.26

1.27

3.34

18.27

8.63

10.95

18.90 12.49

23.75

17.42

26.36

35.86

123.26

2.65

1.74

3.54

3.49 3.58

3.47

3.20

2.76

1.47

0.95

0.01

0.00

1.75

1.16

2.43

2.53

2.54

2.46 2.60

2.19

1.19

0.79

0.01

0.00

Change, Change of DtRA DVad evaporated liquid (9103 (ncm3/g) adsorbate volume, cm3/g) DVL (9103 cm3/g)



3.86

33.86

15.11 10.88

17.14

28.03

35.09

25.35

38.74

52.33

164.83



2.15

27.25

6.92

10.85

23.36 12.82

30.70

22.87

36.43

50.76

161.81

True volume of the pores, Vp (9103 cm3/g)



4.23

34.40

12.12 9.87

12.04

16.97

17.90

11.11

14.82

17.10

37.08



1.74

20.65

4.76

6.71

11.30 7.06

12.77

8.31

11.71

14.15

32.06

Surface of pore walls, A (m2/g)



150.72

146.49

102.23 112.10

90.11

78.06

61.09

43.20

32.09

17.27

0.17



99.33

97.59

76.94

72.18

58.40 65.47

47.10

34.33

26.03

14.32

0.17

Total surface, RA (m2/g)

1696

123.64

146.92

0.84

2

226.96

0.91

1

Calculations using cylindrical pore model

59.41

79.79 67.51

95.21

106.52

123.64

146.92

0.84

2

226.96

0.91

1

Calculations using between parallel plates pores model

No. Relative pressure, p/p0

Table 2 The data of RA calculations and pore size distribution of gyrolite (48 h)

Author's personal copy Materials and Structures (2011) 44:1687–1701

N2 volume adsorbed on the Vad (ncm3/g)

Kelvin radius of the pores, ˚) rK (A

22.58

21.98 21.12

20.63

0.87

0.84

0.81

0.78

0.75

0.70

0.65

3

4

5

6

7

8

9

10 0.63 11 0.60

12 0.55

16.16

20.90 18.91

22.42

27.12

33.64

38.91

45.99

55.61

69.81

92.68

134.53

22.58

21.98

21.12

20.63 18.21

17.03

0.87

0.84

0.81

0.78

0.75

0.70

0.65

3

4

5

6

7

8

9

10 0.63

11 0.60

12 0.55 13 0.50

14 0.45

24.35

26.88

29.13

31.76

35.04

39.42

45.48

0.90

2

55.49

0.93

1

14.62

6.55

7.21 6.87

7.60

7.86

8.05

8.57

9.21

9.67

10.22

10.89

11.75

12.91

19.90 17.53

21.66

24.77

30.38

36.27

42.45

50.80

62.71

81.25

18.63

23.38 20.80

26.51

28.75

30.47

35.69

42.85

48.58

56.21

66.51

81.56

105.60

11.31

15.05 13.01

17.53

19.90

21.66

24.77

30.38

36.27

42.45

50.80

62.71

81.25

149.15 113.60

30.59

7.21 14.62

36.61 34.11

38.52

44.26

52.06

58.25

66.43

77.40

93.31

118.51

7.86 7.60

8.05

8.57

9.21

9.67

10.22

10.89

11.75

12.91

17.71

22.09 19.72

24.94

27.63

29.61

33.08

39.27

45.71

52.39

61.36

74.03

93.58

127.37

35.36 32.35

37.56

41.39

48.16

55.15

62.34

71.91

85.35

105.91

141.14

Average true pore radius, rp ˚) (A

0.29

0.35 0.32

0.39

0.26

0.19

0.53

0.64

0.46

0.55

0.67

0.86

1.16

1.71

0.26 0.39

0.19

0.53

0.64

0.46

0.55

0.67

0.86

1.16

1.71

Change, ˚) Dt (A

1.24

2.41 1.19

0.49

0.86

0.60

1.77

2.53

2.25

2.63

3.28

4.38

6.06

10.01

0.86 0.49

0.60

1.77

2.53

2.25

2.63

3.28

4.38

6.06

10.01

1.91

3.72 1.83

0.76

1.32

0.92

2.73

3.90

3.46

4.05

5.05

6.75

9.34

15.42

1.32 0.76

0.92

2.73

3.90

3.46

4.05

5.05

6.75

9.34

15.42

0.79

0.65 0.78

0.72

0.45

0.31

0.76

0.75

0.44

0.42

0.35

0.24

0.02

0.00

0.33 0.52

0.23

0.58

0.60

0.36

0.34

0.30

0.21

0.02

0.00

Change, Change of DtRA DVad evaporated (9103 (ncm3/g) liquid cm3/g) adsorbate volume, DVL (9103 cm3/g)

2.73

6.62 2.40

0.09

1.69

1.14

3.52

5.26

4.80

5.54

6.85

9.07

12.36

19.38

1.18 –

0.79

2.61

4.29

4.17

4.94

6.30

8.61

12.12

19.16

True volume of the pores, Vp (9103 cm3/g)

3.09

5.99 2.43

0.07

1.23

0.77

2.13

2.68

2.10

2.12

2.23

2.45

2.64

3.04

0.67 –

0.42

1.26

1.78

1.51

1.59

1.75

2.02

2.29

2.71

Surface of pore walls, A (m2/g)

30.10

24.57 27.01

18.58

18.51

17.28

16.52

14.39

11.71

9.61

7.49

5.26

2.81

0.17

13.46 –

12.79

13.37

11.11

9.33

7.81

6.23

4.48

2.46

0.17

Total surface, RA (m2/g)

Materials and Structures (2011) 44:1687–1701

12.08

16.16 13.94

18.91

20.90

22.42

27.12

33.64

38.91

45.99

55.61

69.81

92.68

134.53

Calculations using cylindrical pore model

24.35

26.88

29.13

31.76

35.04

39.42

45.48

0.90

2

55.49

0.93

1

Average Kelvin radius of the pores, ˚) rK (A

163.77 113.60

Thickness True of N2 pore ˚ ) radius layer, t (A ˚) (A

Calculations using between parallel plates pores model

No. Relative pressure, p/p0

Table 3 The data of RA calculations and pore size distribution of gyrolite (168 h)

Author's personal copy 1697

Author's personal copy

[mm3/g]

Cumulative pore volume, Vp

– –

1

90

3

70

2

50 30 10

1

10

100

– 0.82 0.80 0.52 0.26 13.06 7.48 13.72

12.40 5.45

5.71

13.99 17 0.25

6.95

15.50 16 0.30

8.01

1.20 0.92 1.64 1.98 1.29 0.55 15.25 9.27 16.78 6.25

110

Fig. 9 Gyrolite total pore volume. Duration of hydrothermal synthesis at 200C temperature, h: 1 72, 2 120, and 3 168

15.79

10.53

130

Pore diameter, D p [nm]

15 0.40

Surface of pore walls, A (m2/g) True volume of the pores, Vp (9103 cm3/g) Change, Change of DtRA DVad evaporated (9103 (ncm3/g) liquid cm3/g) adsorbate volume, DVL (9103 cm3/g) Change, ˚) Dt (A Average true pore radius, rp ˚) (A Average Kelvin radius of the pores, ˚) rK (A Thickness True of N2 pore ˚ ) radius layer, t (A ˚) (A Kelvin radius of the pores, ˚) rK (A N2 volume adsorbed on the Vad (ncm3/g) No. Relative pressure, p/p0

Table 3 continued

31.30

Materials and Structures (2011) 44:1687–1701 Total surface, RA (m2/g)

1698

It was determined that the pores with 2.0–3.0 nm radius were dominated in the gyrolite structure after 72 h of synthesis (Fig. 10). Meanwhile, the pores with 3.0–4.0 nm and even bigger radius started to form after 120 h of hydrothermal treatment. This tendency of pores growth remains by continuing isothermal curing (up to 168 h), because the pores with 3.0–5.0 nm radius is dominated in the pore distribution by radius curve (Fig. 10c). In the next stage of experiment there was done 2? Zn ion adsorption by using gyrolite samples with characteristic different structural properties (specific surface area, crystal size, pore shapes, etc.). It was determined that already after 30 s of adsorption more than half zinc ions were intercalated to gyrolite structure, i.e. *71% (21.36 mg Zn2?/g), when the initial solution concentration was 0.3 g/dm3 (Fig. 11a). It should be noted that ion exchange reaction ended after 15 min when adsorbent was substituted by *99% (28.63 mg Zn2?/g) of zinc ions. Meanwhile, the rapidity of these ions intercalation to orderly structure of gyrolite was much slower (synthesis duration—168 h) because only *55% (16.67 mg Zn2?/g) of zinc ions were intercalated after 30 s of reaction. Besides, the equilibrium was reached after 5 min of adsorption—91% (27.25 mg Zn2?/g), because the concentration of zinc ions remains almost the same continuing the reaction (Fig. 11a). Thus, the ion exchange between gyrolite with less orderly structure and Zn(NO3)2 ? NH4OH alkaline solution (cZn2þ —0.3 g/dm3) proceeds more faster and effectively. The exchange reaction mechanism (Ca2? $ Zn2?) in gyrolite may take place from edge and planar surface sites as well as from interlayer Ca2? sites (Fig. 5).

Author's personal copy Materials and Structures (2011) 44:1687–1701

First derivative, dV p /dDp [mm3/g*nm]

a

b 25.0 20.0 15.0 10.0 5.0 0.0 1

10

100

First derivative, dV p /dDp [mm3/g*nm]

Fig. 10 Gyrolite differential pore volume plot when duration of hydrothermal synthesis at 200C temperature, h: a 72, b 120, and c 168

1699

9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0 1

10

100

Pore diameter, Dp [nm]

Pore diameter, Dp [nm]

First derivative, dV p /dDp [mm3/g*nm]

c 12.0 10.0 8.0 6.0 4.0 2.0 0.0 1

10

100

Pore diameter, Dp [nm]

30

b

30 1

29.5

2

29

3 4

28.5 28

5

27.5 0

7.5

15

22.5

30

37.5

45

τ, min

R2 = 0.9716

29.5

Σ X, mg/g

Σ X, mg/g

a

29 28.5 28 27.5 32

52

72

92

112

132

152

τ, h

Fig. 11 Integral kinetic curves (a) and total amount (b) of Zn2? ions adsorption from Zn(NO3)2 solution when the initial concentration of Zn2? ions is 0.3 g/dm3 at 25C. Duration of hydrotermal synthesis, h: 1 32, 2 48, 3 72, 4 120, and 5 168

The results presented here show that the gyrolite cation exchangers are interesting family of exchangers and further research is needed to understand better their exchange and selectivity properties.

4 Conclusions It was determined that the structure of gyrolite and the shape of dominated pores (from pores between parallel plates to cylindrical pores) changes prolonging the

duration of hydrothermal synthesis. It should be underlined that the stable gyrolite crystal structure was formed only after 120 h of isothermal curing. Its specific surface area SBET = 38.28 m2/g, the radius of ˚ , the calculated dominant plate pores rp = 30–40 A total pore volume RVp = 0.08 cm3/g. Calculated CBET constant values confirmed the structural changes of gyrolite during hydrothermal synthesis. After sufficiently long synthesis duration the structural elements of gyrolite structure were connected to more orderly structure of the interlayer X. This phenomenon is

Author's personal copy 1700

confirmed by X-ray diffraction analysis too: prolonging the duration of isothermal curing crystallite size of ˚ . It was this compound varies from 1,066 to 2,156 A estimated that the ion exchange between gyrolite with less orderly structure in Zn(NO3)2 ? NH4OH alkaline solution (cZn2þ —0.3 g/dm3) proceeds more faster and effectively because already after 30 s of adsorption more than half zinc ions were intercalated to gyrolite structure, i.e. *71% (21.36 mg Zn2?/g). Acknowledgments This work was supported by the partners of Action COST MP0701 ‘‘Composites with novel functional and structural properties by nanoscale materials’’ and financially by the Agency for International Science and Technology Development Programmes in Lithuania.

References 1. Baltakys K, Siauciunas R (2006) The influence of c-Al2O3 and Na2O on the formation of gyrolite in the stirring suspension. J Mater Sci 41:4799–4805 2. Baltakys K, Siauciunas R (2009) Physically and chemically bound H2O in the gyrolite structure. Mater Sci Pol 27:255–263 3. Baltakys K, Siauciunas R, Baltusnikas A (2004) SiO2 atmainos ˛itaka girolito susidarymo eigai. Chemine technologija 1:45–52 (in Lithuanian) 4. Baltakys K, Siauciunas R, Kitrys S (2008) Surface microstructure and specific surface area of pure and Nasubstituted gyrolites. Mater Sci Pol 26:633–645 5. Bankauskaite A, Baltakys K (2009) The sorption of copper ions by gyrolite in alkaline solution. Mater Sci Pol 27: 899–908 6. Cheary RW, Coelho AA (1996) Programs XFIT and FOURYA, deposited in CCP14 Powder Diffraction Library, Engineering and Physical Sciences Research Council, Daresbury Laboratory, Warrington, England. http://www. ccp14.ac.uk/tutorial/xfit-95/xfit.htm 7. Coleman NJ (2006) Interactions of Cd(II) with wastederived 11 angstrom tobermorites. Sep Purif Technol 48:62–70 8. Coleman NJ, Brassington DS, Raza A, Mendham AP ˚ (2006) Sorption of Co2? and Sr2? by waste-derived 11 A tobermorite. Waste Manag 26:260–267 9. Elton NJ, Hooper JJ, Holyer VAD (1998) Al-free gyrolite from the Lizard, Cornwall, England. Mineral Mag 62: 271–272 10. Jauberthie R, Temimi M, Laquerbe M (1996) Hydrothermal transformation of tobermorite gel to 10 angstrom tobermorite. Cem Concr Res 26:1335–1339 11. Kalousek GL, Nelson EB (1978) Hydrothermal reactions of dicalcium silicate and silica. Cem Concr Res 8:283–289 12. Kasperaviciute V, Baltakys K, Siauciunas R (2008) The sorption properties of gyrolite for copper ions. Ceram-Silik 52:95–101

Materials and Structures (2011) 44:1687–1701 13. Lin KL, Chang J, Chen GF, Ruan ML, Ning CQ (2007) A simple method to synthesize single-crystalline beta-wollastonite nanowires. J Cryst Growth 300:267–271 14. Lowell S, Shields JE, Thomas MA (2004) Characterization of porus solids and powders: surface area, pore size and density. Kluwer Academic Publisher, New York 15. Meller N, Hall C, Phipps JS (2005) A new phase diagram for the CaO-Al2O3-SiO2-H2O hydroceramic system at 200C. Mater Res Bull 40:715–723 16. Merlino S (1988) Gyrolite: its crystal structure and crystal chemistry. Mineral Mag 52:377–387 17. Merlino S, Bonaccorsi E, Armbruster T (1999) Tobermorites: their real structure and order-disorder (OD) character. Am Mineral 84:1613–1621 18. Mitsuda T, Chan CF (1977) Anomalous tobermorite in autoclaved aerated concrete. Cem Concr Res 7:191–194 19. Miyake M, Iwaya M, Suzuki T, Kakehi H, Mitsuda T (1990) Aluminum-substituted gyrolite as cation exchanger. J Am Ceram Soc 73:3524–3527 20. Nocum-Wczelik W (1997) Effect of some inorganic admixtures on the formation and properties of calcium silicate hydrates produced in hydrothermal conditions. Cem Concr Res 27:83–92 21. Nocum-Wczelik W (1999) Effect of Na and Al on the phase composition and morphology of autoclaved calcium silicate hydrates. Cem Concr Res 29:1759–1767 22. Okada Y, Fang Y, Ishida H, Nishido H (1994) Dehydration processes of gyrolite. J Ceram Soc Jpn 102:452–458 23. Okada Y, Masuda T, Ishida H (1995) Behavior of silicate anions on the formation processes of gyrolite followed by Si-29 NMR. J Ceram Soc Jpn 103:124–127 24. Sasaki K, Masuda T, Ishida H, Mitsuda T (1997) Synthesis of calcium silicate hydrate with Ca/Si = 2 by mechanochemical treatment. J Am Ceram Soc 80:472–476 25. Shaw S, Clark SM, Henderson CMB (2000) Hydrothermal formation of the calcium silicate hydrates, tobermorite (Ca5Si6O16(OH)24H2O) and xonotlite (Ca6Si6O17(OH)2): an in situ synchrotron study. Chem Geol 167:129–140 26. Shaw S, Henderson CMB, Clark SM (2002) In situ synchrotron study of the kinetics, thermodynamics, and reaction mechanisms of the hydrothermal crystallization of gyrolite. Am Mineral 87:533–541 27. Siauciunas R, Baltakys K (2004) Formation of gyrolite during hydrothermal synthesis in the mixtures of CaO and amorphous SiO2 or quartz. Cem Concr Res 34:2029–2036 28. Siauciunas R, Janickis V, Palubinskaite D, Ivanauskas R (2004) The sorption properties of tobermorite modified with Na? and Al3? ions. Ceram-Silik 48:76–82 29. Stevula L, Petrovic J (1983) Formation of an intermediate C-S-H phase during the hydrothermal synthesis of gyrolite. Cem Concr Res 13:684–688 30. Stevula L, Harman M, Horvath I, Putyera K (1990) The mineral gyrolite and its stability under hydrothermal conditions. Ceram-Silik 34:315–330 31. Stumm A, Garbev K, Beuchle G, Black L, Stemmermann P, Nuesch R (2005) Incorporation of zinc into calcium silicate hydrates, part I: formation of C-S-H(I) with C/S = 2/3 and its isochemical counterpart gyrolite. Cem Concr Res 35:1665–1675

Author's personal copy Materials and Structures (2011) 44:1687–1701 32. Suzuki K, Nishikawa T, Ikenaga H, Ito S (1986) Effect of NaCl or NaOH on the formation of C-S-H. Cem Concr Res 16:333–340 33. Taylor HFW (1993) Models for the composition and structure of calcium silicate hydrate (C-S-H) gel in hardened tricalcium silicate pastes and the incorporation of minor and trace-elements into calcium silicate hydrate

1701 (C-S-H) gel in hardened cement pastes—discussion. Cem Concr Res 23:995–998 34. Winters MA, Richter JD, Sagar SL, Lee AL, Lander RJ (2003) Plasmid DNA purification by selective calcium silicate adsorption of closely related impurities. Biotechnol Prog 19:440–447