Carrageenans: Analysis of Tablet Formation and Properties

Carrageenans are natural polymers that have been successfully applied to different purposes in tabletting technology. They can be used as controlled release agents;^1-11 they can reduce the inactivation of α-amylase during tabletting;¹² and they can reduce the recrystallization of amorphous indomethacin during tabletting,¹³ thus, they enable soft tabletting.¹⁴ Therefore, it is important to know the powder technological and tabletting properties in detail.

Carrageenans are polysaccharides extracted from red algae¹⁵ and are similar to alginates (extracted from brown algae) and agar (extracted from red algae). The most important members used for extraction are Chondrus crispus and Gigartina stellata, which originate from Ireland, Brittany, Denmark, the US and the Philippines. The carrageenans are named according to their origin, the Irish coastal village Carragheen, and have been used since 1945. In the US Pharmacopeia they have a monograph.¹⁶ Chemically they are galactanes, which consist partially of anhydrogalactose, are substituted by sulfate groups (Figure 1) and are anionic. They were first isolated in 1953;¹⁷ and their structure was analysed in 1955.¹⁸ Three basic types exist: ι-, κ- and λ-carrageenans (Figure 1), which differ in their sulfate content. The sulfate content increases in the following order: κ-carrageenan (25–30%), ι-carrageenan (28–35%) and λ-carrageenan (32–39%).

Figure 1: Chemical structure of carrageenans.

For carrageenan production, the algae are washed and dried, and treated with alkali.¹⁵ The raw extract is cleaned, concentrated and the carrageenans are precipitated in alcohol. The raw carrageenan is produced by drying. Following this, the product is milled and by mixing fractions with different substitution or different potassium content, a range of commercially available types of carrageenan can be produced. Further particle size can vary.

To study the process of tablet formation of the carrageenans, the physicochemical properties of these polymers are important. All types of carrageenan show a broad distribution of molecular weight between 100000 and 500000. The crystallinity of these polymers has not been studied until now. The glass transition temperature (Tg) was only determined for lyophilized products. It varies between –10 °C and –80 °C depending on the humidity.¹⁹

3D modelling is a method that has been successfully applied to study the deformation properties of different materials during tabletting. The utility and validity of this technique has previously been documented.^20–23 In particular, this technique is unique as it allows the simultaneous evaluation of the three most important tabletting process variables (normalized time, pressure and density). To these data a twisted plane is fitted, which is characterized by three parameters, d, e and ω. Time plasticity (d) describes the plastic deformation with respect to time. Increasing time plasticity indicates faster deformation during tabletting. Pressure plasticity (e) describes the relationship between density and pressure. Increasing pressure plasticity indicates greater deformation at lower pressures. The angle of torsion (ω) is a measure of the material's elasticity. With increasing ω, elasticity decreases. Using this method, brittle fracture, plastic deformation and elasticity can be clearly distinguished.

The aim of this study was to analyse the powder technological and compaction properties completely for six types of carrageenan and to study the process of tablet formation and the properties of the final tablets in detail.

Materials

The excipients used were ι-carrageenan Gelcarin GP-379 NF, Gel 379 (Lot # ZA 502); two κ-carrageenans which differ in their potassium content,

²⁴

Gelcarin GP-812 NF, Gel 812 (Lot # ZB 502) and Gelcarin GP-911 NF, Gel 911 (Lot # ZC 502); and two λ-carrageenans, Viscarin GP-109 NF, Vis 109 (Lot # ZH 712) and Viscarin GP-209 NF, Vis 209 (Lot # ZI 502). There was also one carrageenan containing both κ- and λ-carrageenans: Viscarin GP-328 NF, Vis 328 (Lot # ZJ 714). All carrageenans were obtained from FMC Corp., (Princeton, NJ, USA). For comparison MCC, Avicel PH 101 (Lot # 6911C) and Avicel PH 102 (Lot # 7808C, FMC Princeton NJ, USA]) were used.

Methods

Test conditions

All materials and tablets were equilibrated, produced and stored between 35% and 45% relative humidity (RH). Tabletting was performed in a special climate-controlled room, which was set to 23 ±1 °C and 45% ±2% RH. To study the influence of humidity the substances were equilibrated at 30, 45, 60 or 75 RH; the climate room was set to these conditions at 23 ±1 °C and humidity remained constant after tabletting.

Sorption isotherms

Sorption isotherms (Figure 2) were recorded gravimetrically after equilibration over saturated salt solutions for 7 days in triplicate.

²⁵

The powder was equilibrated at a specific RH. After equilibration, the powder was weighed and transferred to the next higher RH for equilibration. This procedure started at 32% RH and was performed up to 90% RH. Then the powder was weighed and, after equilibration, moved to the next lower RH up to 0% RH (phosphorous pentoxide).

Figure 2: Sorption isotherms of six different carrageenans compared with MCC.

Water content

The water content was determined by thermogravimetric analysis using TGA 209 (Netzsch Gerätebau GmbH, Germany) in triplicate. The powder was heated with 10 K min

^-1

up to 150 °C and water loss was determined.

Particle size determination

Particle size distribution was determined by laser light diffractometry using a dry feeder (Sympatec Rodos 12 SR; Sympatec, Germany). The settings of the feeder were

• pressure 4 bar

• injector beneath pressure 60 mbar

• focal distance 200 mm

• measuring time 25–35 s in triplicate.

The mean volume particle size distribution was calculated and medium particle size determined.

Scanning electron microscopy

Both powder and tablets (upper surface and breaking surface) were analysed by scanning electron microscopy (SEM [JSM 6400; JEOL, Tokyo, Japan]) at an accelerating voltage of 5 keV depending on the sample. Previously, they were mounted onto a sample holder and coated with coal/gold/coal.

Apparent particle density

The apparent particle density of all materials was determined by Helium pycnometry (Accupyc 1330; Micromeritics, Norcross, GA, USA) in triplicate. The equilibrated materials were analysed to determine the apparent particle density of the materials containing some moisture (Tables 1 and 2). The method is described by Picker and Mielck.

²⁶

Bulk and tap density

Bulk and tap density were determined with two repetitions in a weighed 250 mL cylinder using a volumeter (Erweka GmbH, Germany). Determinations were performed according to the

European Pharmacopoeia

.

²⁷

Carr-index

To analyse flowability the Carr-index

²⁸

was calculated using bulk and tap density (Equation 1 in Figure 3).

Figure 3 Equations.

Glass transition temperature

The Tg was determined using DSC 200 (Netzsch Gerätebau GmbH) and hermetically closed pans. All determinations were performed for the materials equilibrated at 45% RH. For Gel 911, analysis was also performed at 30% RH. Sample size varied between 5 and 10 mg. Heating rate was 40 K min

^-1

. Weak transitions can only be determined with a high heating rate.

²⁹

The temperature interval was set to –50 °C to 150 °C. The Tg was determined by calculating the temperature of the half step height during the first heating. To verify the results the maximum of the first derivative was also determined.

X-ray diffraction studies

The crystallinities of the powders were compared using a Roentgen diffractometer (URD 63; Freiberger Präzisionsmechanik, Germany). The radiation was copper and a nickel filter was used. Bragg's angle was analysed between 3 and 50 2θ.

Tabletting

Tabletting was performed on an instrumented eccentric tabletting machine (EK0/DMS, No. 1.0083.92; Korsch GmbH, Germany) with 11 mm diameter flat faced punches (Ritter GmbH, Germany). Equal volumes of the substances based on apparent particle density were tabletted to different graded maximum relative densities (ρ

_{rel, max}

) of the tablets (precision 0.001) between 0.70 and 0.90. The tablet height at maximum densification under load was held constant at 3 mm.

Displacement of the punch faces was measured using an inductive transducer (W20 TK; Hottinger Baldwin Messtechnik, Germany) and corrected for elastic deformation of the punches. The depth of filling was held constant at 13 mm. The production rate was 10 tablets/min. No lubricant was used. The amount of material necessary for each tablet with a given ρ _{rel, max} was calculated. The powder was manually filled into the die and one compaction cycle was performed.

Ten single tablets were produced at each condition. Data acquisition was performed by a DMC-plus system (Hottinger Baldwin Messtechnik) and data were stored by BEAM-Software (AMS-Flöha, Germany). Force, time and displacement of the upper punch were recorded for each compaction cycle.

Data analysis

For analysing tabletting data, only data >1 MPa were used. For five compaction cycles of each material, normalized time, pressure and

ln

(1/1-

D

_rel

) according to Heckel

³⁰

were calculated.

3D model

For applying the 3D modelling technique,

²²

all three measured values were presented in a 3D data plot, to which, a twisted plane was fitted by the least-squares method according to Levenberg-Marquardt (Matlab) using Equation 2 in Figure 3. The plane is twisted at t=

t

_max

.

d, e, and ω of the five compaction cycles at each tabletting condition (material and a given ρ _{rel, max}) were averaged, and means and standard deviations were calculated. The mean standard deviation for d was 0.02, for e it was 0.0001, and for ω it was 0.0004.

Heckel function

To apply the Heckel function

³⁰

a porosity pressure plot is necessary. Heckel describes the decrease of porosity with pressure by first order kinetics. The slope of the Heckel equation (Equation 3 in Figure 3) gives, as calculated, information on the total deformation of the powder.

Pressure-time function

The pressure-time function

³¹

is a repeatedly modified Weibull equation. In the present form it is able to describe the normalized pressure-time curve of the tabletting process (Equation 4 in Figure 3).

The parameter γ indicates the symmetry of the plot and thus, is a measure for the elasticity of the powder. With increasing γ elastic deformation increases. The parameter β is the time difference between maximum pressure and the lifting of the upper punch from the tablet. Information on the elastic recovery is given.

γ and β can be presented in a γ-β-diagram, which gives information on the deformation behaviour of the powders. The mean standard deviation for γ was 0.02 and for β it was 0.40.

Force-displacement profiles

Force-displacement profiles can be used to calculate the energy of compaction.

³²

This includes the compression and decompression of the powder to the tablet. The area between compression and decompression is the area of the energy of the plastic deformation. The area between maximum displacement and decompression is the energy of the elastic deformation. Both these areas were calculated for five compaction cycles.

Elastic recovery

Elastic recovery after tabletting was calculated using Equation 5 in Figure 3.

³³

The calibrated inductive transducer (W 20 TK; Hottinger Baldwin Messtechnik) was used to measure the position of the upper punch and thus, the axial expansion in the die. The height of the tablet after 10 days was measured by a micrometer screw (Mitotuyo, Japan). Ten tablets were analysed, and the means and standard deviations were calculated.

Additionally, elastic recovery was determined in dependence on time. For this purpose, directly after tabletting, out-die measurements were performed. Either thermomechanical analysis (TMA; Netzsch Gerätebau GmbH [precision 0.0005–0.001 mm]) or the automatic micrometer screw (University of Halle, precision 0.005 mm) were used to measure the height changes continuously.^34,35 The use of the micrometer screw was validated by measuring the six steel tablets of different heights (1, 2, 3, 4, 5 and 6 mm) each 50 times at three independent days on the rotary table.

For calculating the total height changes, in-die and out-die height changes were added. Additionally, these were confirmed against each other by measuring the tablet shaped parallel plates of 1, 2, 3, 4, 5 and 6 mm height inside and outside the die 20 times each. The measured values were corresponding.

Environmental scanning electron microscopy (ESEM)

To analyse changes in the microstructure continuously, a tablet was analysed by ESEM (XL 30 FG; Philips, Germany) at 40% RH and 2.9 torr for 12 h using video analysis. Accelerating voltage was 2 keV. The images were also edited using Corel Photo paint as a black and white image.

Freeze fracturing and transmission electron microscopy

Tablets were frozen in a freeze fracturing apparatus (BAF 400; Balzers, Switzerland) at –210 °C. The frozen tablets were broken inside the apparatus. The fractured surface was coated with a mixture of platinum and coal (2 nm). The platinum coating was stabilized with coal (20 nm). The replica were detached using concentrated sulfuric acid. Afterwards, the replica were analysed by TEM (EM 300; Philips, Germany).

Crushing force

The crushing force of the tablets was determined with the crushing force tester (TBH 30; Erweka GmbH). In all cases, five tablets were analysed 10 days after tabletting, and the means and standard deviations were calculated.

Results and Discussion

Material properties

Figure 2 shows the sorption isotherms of the carrageenans compared with MCC. All the carrageenans exhibit a higher water sorption than MCC. Water sorption is more than threefold. Up to 60% RH 20% (w/w) water were sorbed. The RH during production and storage should be controlled, as it was in this study. In Table 1, the water content of the materials is given. It corresponds with the sorption isotherms at the RH used in this study.

Figure 4: Particle size distribution of six different carrageenans and MCC determined by laser diffraction.

Figure 4 shows the cumulative particle size distribution as analysed by laser diffraction. Particle size is similar to that of Avicel PH 101. Thus, when comparing the materials deformation behaviour, the influence of particle size can be neglected. Also, the medium particle size is similar. Only Vis 209 exhibits slightly bigger particles.

Figure 5a: SEM of powders (first column); upper tablet surfaces (second column) and breaking surfaces (third column) for (a) Gelacarin GP-379 NF (b) Gelcarin GP-911 NF and (c) Gelcarin GP 812 NF [magnification 1500x].

The particle shape (Figure 5) is different for the different carrageenan types. Gel 911, Vis 109 and Vis 328 exhibit long threads, whereas Gel 379, Gel 812 and Vis 209 show flat particles composed of several small threads similar to MCC. The ι-carrageenan Gel 379 has a structure different to that of the κ- and λ-carrageenans. The potassium content in Gel 812 seems to influence particle structure and Vis 209 consists of bigger particles.

Figure 5b: SEM of powders (first column); upper tablet surfaces (second column) and breaking surfaces (third column) for (d) Viscarin GP-328 NF; (e) Viscarin GP-109 NF; (f) Viscarin GP-209 NF; (g) and Avicel PH 101 (magnification: 1500x).

In Table 1 the apparent particle, tap and bulk density of the materials is given. Apparent particle density is higher than for MCC. It is similar for κ- and λ-carrageenan, and significantly higher for ι-carrageenan containing anhydrogalactose groups. For the materials that were stored and tabletted at different RH the apparent particle densities are given in Table 2.

Table 1: Powder technological properties of six different types of carrageenans and MCC (mean6SD).

Bulk and tap density can provide information on the flowability of powders and, by using both these values, the Carr index was calculated. The higher the Carr index, the better the compressibility of the powder, but the flowability is worse. The following order for flowability can be set up: 209<109<379<328<911 to 812<MCC. The bigger particles of Vis 209 showed better flowability and the order is λ<ι<κ-carrageenan.

Table 2: Apparent particle density of Gelcarin GP-911 NF and Avicel PH 102 at different relative humidities, calculated according to reference 26.

The deformation of materials is dependent on their Tg. The results of Tg analysis are shown in Table 1. Tg is for no carrageenan higher than 2.5 °C (45% RH). For Gel 911, Tg at 30% RH did not significantly differ from that at 45% RH. Thus, the carrageenans are in the rubbery state at room temperature. They should be easily deformable and show some elasticity. Contrary to this, MCC with a Tg of 60–80 °C²⁹ is in the glassy state.

Additionally, the carrageenans are predominantly amorphous. X-ray diffractograms showed peaks at 28 and 36 2θ which belong to the calcium and potassium salts in the carrageenans.

This article provides an overview of the current knowledge regarding carrageenans and tabletting. The materials and methods of this study have been described and the material properties of the carrageenans have been discussed. Part II will continue with the results and will discuss tabletting, final formation of the tablets and tablet properties. Finally, general conclusions on the tabletting of carrageenans will be drawn.

Katharina M. Picker-Freyer is associate professor (Private Docent) at the Institute of Pharmaceutical Technology and Biopharmacy, Martin-Luther-University Halle-Wittenberg, Germany.

References

All references will be listed in Part II of this article, which will appear in the September issue of

Pharmaceutical Technology Europe

.