Evaluation of Lubrication Methods How to Generate a Comparable Lubrication for Dry Granules and Powder Material for Tableting Processes 2014 Powder Te

Powder Technology 266 (2014) 156–166

Contents lists available at ScienceDirect

Powder Technology

j ourna l homepage: www.e lsev ie r .com/ locate /powtec

Evaluation of lubrication methods: How to generate a comparablelubrication for dry granules and powder material for tableting processes

Johanna Mosig, Peter Kleinebudde ⁎Institute of Pharmaceutics and Biopharmaceutics, Heinrich-Heine-University, Universitaetsstrasse 1, 40225 Duesseldorf, Germany

⁎ Corresponding author. Tel.: +49 211 81 14220; fax: +E-mail addresses: [email protected] (J. Mosig), k

(P. Kleinebudde).

http://dx.doi.org/10.1016/j.powtec.2014.06.0220032-5910/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t

Article history:Received 17 December 2013Received in revised form 10 June 2014Accepted 14 June 2014Available online 23 June 2014

Keywords:LubricationExternal lubricationExternal surface areaRoll compaction/dry granulationCompressionTablet

Lubrication can strongly influence tablet characteristics like strength, disintegration and dissolution and thusshould be performed appropriately taking properties of the materials into account. As magnesium stearatecoats the particles, determination of the specific external surface area could be an approach to find the optimallubricant amount to prevent sticking, to obtain aminimal reduction in strength and to generate a comparable lu-brication between different dry granules and the starting powdered material. Therefore, in this study a surfaceproportional lubrication was investigated in comparison to a mass related internal and an external lubricationfor five different materials (two types of microcrystalline cellulose, powder cellulose, magnesium carbonateand lactose). The specific external surface areas of different dry granules as well as of the starting powderedma-terialswere determined. Prior to tableting an amount ofmagnesium stearate proportional to the specific externalsurface area (2.5 μg/cm2) was added. For eachmaterial specific external surface area of the granules was smallerthan that of the rawmaterial. Hence, the amount of magnesium stearate added to the granules was much lower(0.1–0.25 (w/w) %) compared to the rawmaterial (0.76–2 (w/w) %). Compaction curves for the granules showeda decrease in tensile strength of the corresponding tablets with increasing specific compaction force during rollcompaction. Based on this work-hardening phenomenon, compression of the raw material should lead to thestrongest tablets. In contrast to this, compactibility of direct compressed microcrystalline cellulose and powdercellulosewasmuch lower compared to the granules. As bothmaterials are lubricant sensitive, the added amountofmagnesium stearatewas too high and prevented particle bonding. Formagnesiumcarbonate, compactibility ofthe rawmaterial was still higher compared to the granules in spite of the tenfold lubricant quantity. Further ex-periments with equal amounts of magnesium stearate for powder and granules of magnesium carbonate wereproblematic, indicating an insufficient lubrication of the powdered magnesium carbonate in this case.External lubrication was practicable for all materials, brittle or plastically deforming, and resulted in the highesttablet tensile strength. As lubrication was in the same magnitude for different materials and lubrication propor-tional to the specific external surface area was not practicable for plastically deforming materials, external lubri-cation should be the method of choice.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

Tablets are one of the most popular solid dosage forms, as produc-tion of large quantities is feasible and cost-effective and patients acceptthis dosage form well. The starting material, e.g. powder or granules, isfilled into a die and pressure is applied by punches, compressing it into atablet. Tableting is usually not possible without small quantities of a lu-bricant in order to reduce wall friction during compression and ejectionout of the die and to prevent sticking of the material to the punches.

Different lubricants are available on the market like fatty alcohols,fatty acids, metal salts of fatty acids and mono-, di- and triglycerides.Magnesium stearate is most commonly used, as it is more effective

49 211 81 [email protected]

than other lubricants and reduces friction coefficients already in smallamounts [1,2]. Strickland et al. showed [3], thatmagnesium stearate ad-hered on the surface of the particles and formed a lubricant film aroundthe particles during mixing. This resulted in lower friction as well as ininterference in particle bonding and in an increase of the hydrophobiccharacter. Due to the interferences in particle bonding, lubricantsweakened the tablets, resulting in an inferior tensile strength [3,4].The coating of the particles with magnesium stearate increased thewater repellant character, resulting in slower disintegration [3,5]and prolonged dissolution time [2,6]. Increasing amounts of lubri-cant or prolonged mixing [7–11] enhanced these adverse effects asfilm formation is promoted.

Materials differ in susceptibility for these adverse effects, in particu-lar with respect to the sensitivity for the reduction in strength. De Boeret al. [12] showed a dependence between the compression behavior andthe effect of magnesium stearate on the tablet strength. Materials

Fig. 1. Schematic figure of the Friedrich-manometer.

157J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

undergoing a complete plastic deformation were most influenced bythe addition of lubricants. Brittlematerialswere less sensitive tomagne-sium stearate as strength did not decrease significantly with an increas-ing amount of lubricant [13]. Moreover, factors other than thefragmentation behavior can affect the lubricant sensitivity of excipients.Vromans et al. showed [14], thatmaterialswith highly irregular surfacesoffered a lower susceptibility for magnesium stearate as film formationis unlikely to become complete. Almaya and Aburub [15] could relatethe lubricant sensitivity to the excipient particle size. According tothese publications, smaller particles, which have a higher specific sur-face area, will be less affected by magnesium stearate. Furthermore,lubricant efficacy as well as the degree of the adverse effects dependson the type of magnesium stearate. Johansson [16,17] investigated theinfluence of granular magnesium stearate on tableting. Compared topowder material a higher amount of lubricant was required to preventsticking, whereas at the same time a less adverse effect on the tabletpropertieswas observed. The film forming behavior ofmagnesium stea-rate could explain these observations. The granular type will be lessefficient in surface covering than the powder material. Frattini andSimioni [18] confirmed these findings by correlating the used quantityof magnesium stearate with the specific surface area of different typesofmagnesium stearate. The usage of equal amounts ofmagnesium stea-rate of the different types resulted in varying tablet characteristics,whereas for the usage of amounts leading to equivalent lubricatingareas tablet properties were almost identical. Further studies affirmedthe approach of lubrication dependence on the specific surface area ofmagnesium stearate [19,20]. Rao et al. [19] demonstrated, that particleproperties like size and specific surface area of the lubricant exertedthemost important impact on the lubrication efficacy. Due to this, mag-nesium stearate types with smaller particle sizes and higher specificsurface areas will be more efficient in lubrication [20] as distributionon particle surface of the excipient is higher. Since the addition of mag-nesium stearate is a crucial aspect for tablet characteristics, it should bewell considered. Depending on the varying specific surface areas ofmagnesium stearate and the tableting materials, surface coverage willdiffer between different types of lubricants as well as for the lubricationof different materials. Resulting from this, Duberg and Nyström [21]used amounts of magnesium stearate to generate a fixed surface ratiobetween the calculated surface area of the lubricant and the material.

As many materials are not suitable for direct compression often agranulation step takes place prior to tableting. Roll compaction andsubsequently dry granulation of the compacted material is beside wetgranulation one commonway to produce granules, which gained in im-portance for the pharmaceutical industry. It can be applied to moistureand heat sensitive materials, the process is environmental friendly, asno removing of solvents is necessary, and scale up is easily feasible[22]. Nevertheless, roll compaction is also associated with drawbackslike the accruing high amount of fines and a loss of strength after a re-compression step. Malkowska and Khan [23] showed, that tabletsmade from dry granules exhibited a lower strength as those from directcompression and strength decreased with increasing force during thefirst compaction step. They explained it by awork-hardening of thema-terial, which resulted in an increased resistance to the deformation. Asproperties of granules will vary with changes in the production setupduring roll compaction/dry granulation, e.g. differences of the specificcompaction force, it is not useful to perform the addition of a constantamount of lubricant when comparing different types of dry granulesor dry granules and the starting material. Moreover granule properties,which are crucial for the lubrication efficacy, as the specific surface area,should be considered.

The aim of this study was to investigate different lubricationmethods in order to find a method for a uniform lubrication for dif-ferent types of dry granules as well as for the raw powder material.Specific external surface area as tool to generate a comparable inter-nal lubrication was tested as well as the usage of an external lubri-cation method.

2. Materials and methods

2.1. Materials

Granules and tablets were prepared from five different excipients.Two qualities of microcrystalline cellulose (MCC), normal (Vivapur102, Batchno. 5610206849,JRS Pharma GmbH & Co, Holzmühle,Germany) and high density (Vivapur 302, Batchno. 5630290339,JRS Pharma GmbH & Co, Holzmühle, Germany), powder cellulose(Arbocel P290, Batchno. 74817101129, JRS Pharma GmbH & Co,Holzmühle, Germany), α-lactose monohydrate (Granulac 200, Batchno.2019, Meggle Excipients and Technology, Wasserburg, Germany) andmagnesium carbonate (Magnesia 18, Batchno. 241118, MagnesiaGmbH, Lüneburg, Germany)were used as received.Magnesium stearate(Parteck LUB, Batchno. K42017563, Merck Millipore, Darmstadt,Germany) served as lubricant. Calibration of the Friedrich-manometerwas performed with glass spheres (3 sizes: 170–180 μm; 100–110 μm;1–60 μm).

For equilibration all materials were stored at 21 °C and 45% relativehumidity. First experimentswere performed after a storage time at leastof one week.

2.2. Methods

2.2.1. Roll compaction/dry granulationRoll compaction/dry granulation was performed with an instru-

mented roll compactor (Minipactor 250/25, Gerteis Maschinen +Prozessengineering AG, Jona, Switzerland). Ribbons were compactedat a gap width of 2 mm, a roll speed of 3 rpm and specific compactionforces of 2, 8 and 12 kN/cm. Afterwards theywere directly dry granulat-ed through a 1mm sieve with a star granulator, rotating 120° clockwiseand 180° counterclockwise and a rotor speed of 40 rpm clockwise and60 rpm counterclockwise. Granules were sieved in portions of nearly100 g for 5 min (Retsch Vibrio AS 200 control, Retsch GmbH, Haan,Germany) to obtain a particle fraction between 315 and 630 μm forevery batch. Further experiments were performed with granules ofthis sieve fraction.

2.2.2. Particle densityParticle densities were measured with a helium pycnometer

(AccuPyc, Micromeritics, Norcross, USA). Measurements were per-formed at 25 ± 0.1 °C.

158 J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

2.2.3. Scanning electron microscopySEM images were acquired by the microscope Phenom G2 pro

(Phenom-World, Eindhoven, Netherlands). If necessary sampleswere sputtered with a 14 nm layer of gold (Automatic Sputter CoaterMSC 1T, Ingenieurbüro Peter Liebscher, Wetzlar, Germany).

2.2.4. Specific external surface areaSpecific external surface area for each granule fraction between 315

and 630 μm as well as for the raw materials was determined by airpermeametry measurements in a Friedrich manometer [24] (self-construction of Evonik Industries, Darmstadt, Germany) equippedwith a sample holder according to Gupte [25] (Fig. 1). Water was usedas test fluid. The fluid level was raised by a vacuum pump (Pipetus,Glaswerk,Wertheim, Germany) tomark A. Then air permeated throughthe powder bed and lowered the water level. Measuring the time for aconstant decrease in the water level enabled measuring of the time aconstant volume of air needs to permeate the powder bed. Therefore,the time interval was measured to decrease the water level from markB to mark C by a light barrier system (self-construction of EvonikIndustries).

The square of the volume specific external surface area can be calcu-lated by Eq. (1) measuring the time a constant volume needs to perme-ate a defined powder bed. Themass specific external surface area can bederived by Eq. (2).

SV2 ¼ 1

k� ρm � g � AL �

Z−dV

h

� ε3

1−εð Þ2 �tη

cm−2h i

ð1Þ

Sm ¼ SVρ

cm2

g

" #ð2Þ

SV volume specific surface areak particle shape factorρm density of manometer fluidg gravitational constantA cross-sectional area of the powder bedL length of the powder bedV volume of manometer fluid in one arm from starting to

endpositionh height difference of fluid level in the manometer armsε porosity of the powder bedt permeation timeη air viscositySm mass specific surface areaρ powder density.

The detected flow time has to be reduced by the running timewithout any testing material to calculate the pure permeation timet. The blank value of the instrument was determined to be 3.92 ±0.086 (n = 100). Furthermore, values for the viscosity of air (η)(0.01819–0.01829 mPa*s) and the density of water (ρm) (0.998203–0.997295 g/cm3) were used according to the measured room tempera-ture. Due to the graded powder container, bed porosity (ε)was calculat-ed by measuring the weight of the testing material, the height (L) and

Table 1Helium density of the materials (n = 3, mean).

MCC High density MCC

Density [g/cm3] 1.594 1.596

the use of the cross-sectional area (A) of the sample holder and thedetermined material density (Table 1). For the particle shape factor(k) an often in literature proposed value of five was used [26]. As thesolving of the integral − dV

h is not possible, a calibration step for theFriedrich manometer has to be performed, to determine an instrumentconstant. For this, the specific BET surface area of three types of glassspheres was determined (Tristar, Micromeritics Instrument Corpora-tion, Norcross, USA), as well as the flow time in the Friedrich manome-ter (Table 2). Using the experimental derived values, Eq. (1) can be

solved for∫−dVh

(Table 2). For further experiments the instrument con-

stant (∫−dVh) was chosen to be 2.27, as the mean of the results for the

three different glass sphere types.For determination of the specific external surface area approximate-

ly 100 g of material was filled in the graduated powder container, togain material heights of 30 ± 2 cm. Pre-tests showed an experimentalsensitivity for different powder bed porosities. Due to this, the materialwas tapped 1250 times within the powder container (volumetricanalyzer, J. Engelsmann AG Apparatebau, Ludwigshafen, Germany) tokeep the porosities comparable. Flow time was determined in triplicatefor each sample preparation and each material was measured threetimes.

Specific external surface area was calculated according to Eqs. (1)and (2) using the determined instrument constant.

2.2.5. Particle size distributionThe particle size of the starting materials was determined by laser

diffraction (Helos H1402+, Sympatec, Clasuthal-Zellerfeld, Germany).Measurements were performed with the dry dispersing unit (Rodos,Sympatec, Clausthal-Zellerfeld, Germany) and a dispersing pressure of1 bar. Starting materials were measured three times and the distribu-tions were evaluated by the instrument software. The particle size dis-tributions of granules from each specific compaction force weredetermined in triplicate by digital image analysis (Camsizer XT, RetschGmbH, Haan, Germany). The X-Jet module was used and a dispersingpressure of 0.3 bar applied to avoid agglomeration of the particles aswell as a destruction. The quantiles of the particle size distributionswithin the fraction 315 to 630 μmwere calculated using the instrumentsoftware.

2.2.6. LubricationGranule fractions and raw powder materials were lubricated

with different amounts of magnesium stearate (MgSt). Two differentmethods were used for the internal lubrication. In one case, mate-rials were lubricated with an amount of magnesium stearate propor-tionally to the determined specific external surface area. For this,2.5 μg magnesium stearate per square centimeter external surfacearea of the material was added to the granules as well as to the rawmaterials. 2.5 μg/cm2 was the lowest amount, which allowed tabletingfor all components including lactose andmagnesiumcarbonate. A higheramount would result in even larger quantities for direct compression.Furthermore, the raw materials were lubricated with concentrations of0.1–0.5% (w/w) of MgSt. Powder and granule material were blendedfor 2 min with magnesium stearate in a Turbula mixer (42 U/min, T2C,Willy A. Bachofen AG, Basel, Switzerland) for the internal lubricationand tableted. External lubrication was achieved by manually lubricatingpunches and die by an eye shadow applicator.

Powder cellulose Magnesium carbonate Lactose

1.559 2.305 1.544

Table 2BET surface of glass spheres (mean ± sd; n = 3) and the corresponding calculatedintegral of Eq. (1) for test measurements with the Friedrich-manometer (mean ± sd;n = 5).

Glass sphere size[μm]

BET surface area[m2/g]

∫−dVh

1–60 0.073 ± 0.001 2.34 ± 0.34100–110 0.043 ± 0.004 2.29 ± 0.32170–180 0.034 ± 0.006 2.17 ± 0.08

159J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

2.2.7. TabletingDirect compression (DC) and granule compression were performed.

Flat-faced 8 mm tablets of 200 mg mass were produced on an instru-mented rotary die press (PressIMA, IMA Kilian, Cologne, Germany)with a rotation speed of 10 rpm. Seven compression pressures between60 and 418 MPa were applied. Tablets were stored for 48 h afterproduction at 21 °C and 45% relative humidity and characterized dueto their weight (CP 224S, Sartorius AG, Göttingen, Germany) heightand diameter (Digimatic Caliper, Mitutoyo, Hoshima, Japan) andcrushing force (TBH 210, Erweka GmbH Apparatebau, Heusenstamm,Germany). The tensile strength was calculated according to Fell andNewton [27]. Compactibilitywas defined as the ability to form compactsof a specific tensile strength under pressure.

3. Results and discussion

3.1. Specific external surface area

Granulation decreased the specific external surface area. For all pow-dered materials, specific surface area was in the range between 3000and 8400 cm2/g, whereas granules offered a specific surface area be-tween 400 and 1000 cm2/g (Table 3). The raw material is declaredwith 0 kN/cm, as the material was not subjected to a roll compactionprocess and no pressure was applied prior to tableting. Measuredspecific surface areas were smaller as BET surface areas reported inliterature. BET measurements detected the total surface area, inter-nal and external. Measurements of the specific external surfacearea acquired neither pores nor surface roughness and resultedtherefore in the smallest detectable surface area and could be per-formed by calculations out of the particle size distribution [26].Permeatry measurements detected primarily the external surfacearea, whereby surface structures (e.g. cavities) are included to a cer-tain degree.

Comparing the five different raw materials, magnesium carbonate(Mg) offered the highest specific surface areawhereas the other four ex-cipients revealed comparable specific surface areas. Specific externalsurface area for both types ofMCC, normal density (M) and high density(Ms), as well as for lactose (L) and powder cellulose (Pc) was approxi-mately half of these of magnesium carbonate. This huge differencecould not be explained by the particle size. Although magnesiumcarbonate offered amuch smaller x50 value compared toMCC and pow-der cellulose, lactose presented a comparable small mean particle size(Table 4). Even if the interquartile range is smaller as for lactose (64.1

Table 3External surface area [cm2/g] of the different powder materials and granules (mean ± sd; n =

Specific compactionforce [kN/cm]

MCC MCC (high density)

0 4174 ± 91 3489 ± 2592 991 ± 11 838 ± 128 617 ± 7 529 ± 612 526 ± 3 436 ± 11

to 88.1 μm), the differences in the specific surface area of 5000 cm2/gwill not be a result of this, but may be caused by the particle structureof the magnesium carbonate.

Freitag [28] pointed out that the particles were agglomerated and byparticle size measurements just the diameter of these agglomerateswere detected. SEM pictures of the magnesium carbonate showed theagglomerate structure of the particles (Fig. 3d). Detected particles arecomposed of much smaller particles. Due to the granular structure, spe-cific surface area of the magnesium carbonate powder is higher com-pared to the other four materials, even if the measured particle size iscomparable.

Increasing the specific compaction force within the dry granulationstep resulted in a decrease in the specific external surface area of thegranules (Fig. 2). According to Jaminet and Hess [29], this is caused byan increase in particle size. They demonstrated, that with increasingcompaction force stronger ribbons were produced which resulted incoarser particles after the dry granulation step. Measurements of theparticle sizes within the fraction confirmed this suggestion. Table 5showed the x50 values for the different batches and an increase in theparticle size with increasing roll compaction force is observable forMCC, powder cellulose and magnesium carbonate. Lactose revealedthe lowest decrease and also the overall lowest specific surface area.In contrast to the other investigated materials, granule size of lactosedid not increase over an increase in the specific compaction force(Table 5). Parrot [30] also observed a lower increase in particle sizeafter roll compaction for the brittle behaving lactose compared to theother investigated materials. This low tendency for particle growthwith increasing compaction force could implicate the observed low de-crease of the specific surface area with increasing force.

The sharpest reduction of the specific surface area was found forboth types of MCC due to the plastically deforming behavior of micro-crystalline cellulose under pressure. Specific surface area as well as thedecrease over an increasing forcewas higher for the normal MCC. Parti-cle size measurements (Table 5) for the granules showed, that the in-crease in size is more pronounced for the normal density MCC. As theincrease in size was about 6% compared to 3% for the high densityMCC, the stronger decrease in the specific surface area with increasingroll compaction force can be attributed to this. The overall higher sur-face area of the normal density MCC can be explained by differencesin the granule structure. SEM pictures (Fig. 4a,b) showed, that the fi-brous structure of the primary particles persisted in the granules in con-trast to the high density MCC. Granules from powder cellulose andmagnesium carbonate exhibited overall the highest specific surfacearea and, comparable with lactose, a small decrease with increasingcompaction force during granulation. According to the starting mate-rials, these differences could not be explained solely by the particlesize of the granules. SEM pictures (Fig. 4c,d) showed that the particleshape properties of the starting materials survived the compactionstep and lead to the observed higher specific surface area compared tothe other investigated materials. Particularly the fibrous shape of pow-der cellulose granules, which is more pronounced compared to the nor-mal density MCC will result in larger particle surfaces compared toround particles. Just for the lowest specific compaction force of 2 kN/cm the specific external surface area of MCC surmounted those of pow-der cellulose and magnesium carbonate.

3).

Magnesium carbonate Powder cellulose Lactose

8320 ± 407 3715 ± 233 3022 ± 346883 ± 52 972 ± 8 578 ± 37675 ± 15 758 ± 3 479 ± 15620 ± 14 752 ± 33 398 ± 43

Table 410%, 50% und 90% quantiles of the particle size distribution of the starting materials [μm] (mean ± s; n = 3).

MCC MCC(high density)

Magnesiumcarbonate

Powder cellulose Lactose

x10 29.36 ± 0.18 24.67 ± 0.32 5.76 ± 0.25 26.55 ± 0.08 3.77 ± 0.03x50 103.37 ± 0.27 102.37 ± 0.98 31.18 ± 0.36 66.59 ± 0.54 26.93 ± 0.15x90 219.57 ± 0.06 207.95 ± 0.58 69.83 ± 0.30 134.53 ± 2.56 91.82 ± 0.28

160 J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

3.2. Tableting of powder and granules

3.2.1. Surface proportional lubricationAdding an amount of MgSt proportional to the specific external sur-

face area was used for surface proportional lubrication. To work with afeasible lubrication, the correlation factor was scheduled to be2.5 μg/cm2. This resulted in amounts of 0.76–2% MgSt for the directcompression of the powder material and in 0.1–0.25% MgSt for thegranule compression (Table 6). For the same material, differences ofthe added amount of MgSt to dry granules produced with differentcompaction forces were small, as the variations were between 0.12and 0.03% of magnesium stearate.

The whole tableting process could be performed without stickingand friction problems for all five materials. Hence, lubrication propor-tional to the specific external surface area was considered to be suitablefor the direct compression as well as for the compression of drygranules.

Fig. 5a to e shows the compaction curves for each of the five excipi-ents. Compactibility for the granules decreased with increasing specificcompaction force during roll compaction according to thework harden-ing phenomenon, described by Malkowska and Khan [23]. Thus, directcompression should lead to the strongest tablets, but contradictoryresults were found within the tableting experiments.

Both types of MCC and powder cellulose revealed similar behaviorcomparing direct and granule compression (Fig. 5a, b and c). As describedbefore, tensile strength of granule tablets decreased with increasing spe-cific compaction force during dry granulation. Compactibility of the directcompression was much lower for these three materials compared to thegranule compression, especially for granules produced with 2 kN/cm.This unexpected behavior was most pronounced for powder cellulose.Compressing the powder with a nearly fourfold higher amount of MgStcompared to the granules caused a dramatic decrease in compactibility(Fig. 5c). Tensile strength of direct compressed tablets was even lowerthan these of tablets made from granules compacted at the highest com-paction force.

Although the added amount of MgSt was with 0.9% in a usual range,the resulting tablets offered an unacceptable strength with a tensilestrength below 1 MPa. For MCC, the decrease in strength was also

Fig. 2. Specific external surface area [cm2/g] of the different granules of MCC, normal (M)and high density (Ms), powder cellulose (Pc), magnesium carbonate (Mg) and lactose (L)(mean ± sd; n = 3).

obvious comparing direct and granule compression, but here ahigher strength (M: 4 MPa; Ms: 3 MPa) was achieved than in thecompression of powder cellulose. Tensile strength of direct com-pressed tablets of MCC with high density was for compression pres-sures above 200 MPa lower than the strength of all granule tablets,also for those of granules produced with 8 and 12 kN/cm (Fig. 5b).Although the compactibility of the DC tablets of the normal MCCwas much lower compared to those of the 2 kN/cm tablets, strengthwas still in the region of those tablets consisting of granules pro-duced with 8 and 12 kN/cm (Fig. 5a).

As the amount of MgSt was lower in the direct compression experi-ments for Ms and Pc (0.9%) compared to the MCC of normal density(1%), differences in the reduction of strength could not be explainedby differences in the amount of MgSt. Moreover, different lubricantsensitivities were responsible for the overall reduction in strength forthe direct compressed tablets and the differences of the loss in strengthbetween the three materials.

Compression of magnesium carbonate showed the expected behav-ior. An increase in the compaction force, fromzero at direct compressionto 12 kN/cm, led to a decrease of the tensile strength (Fig. 5d). Accord-ing to the work-hardening phenomenon, direct compression revealedthe strongest tablets, whereas those produced fromgranules compactedwith 12 kN/cm offered the lowest strength. For magnesium carbonate,compactibility of the raw material was still higher compared to thegranules in spite of the tenfold lubricant quantity.

For lactose, no differences between the different granule tablets aswell as between DC and granule compression were found (Fig. 5e).The decrease in tensile strength for the direct compression comparedto the granule compression at compression pressures above 350 MParesulted from the tableting process. As the ungranulated materialoffered a poor flowing behavior, filling of the dies was problematicand caused the observed deviations.

Lubricating proportional to the specific external surface area result-ed in different behaviors for elastically, plastically and brittle deformingmaterials. Due to the plastically deforming behavior ofMCC and powdercellulose, they are highly sensitive to the addition of magnesium stea-rate. According to De Boer [12], no new clean surfaces were createdand thus, the former generated films of MgSt around the particlesavoided bindings. Even if the amount of MgSt is calculated to the avail-able surface of the particles (granules and powder) and therefore thepossibilities to form bonds should be equal for the direct compressionand the granule compression, DC revealed a lower binding tendency.As the creation of new lubricant free surfaces will not differ betweenthe direct compression and the granule compression, it can be assumed,that the here performed lubrication will in fact result in differences ofthe surface occupation. This resulted in varying amounts of free surfacescausing the observed lower binding tendency during direct com-pression for the plastically deforming materials. For brittle behavingmaterials as lactose or magnesium carbonate, surface proportionallubrication did not cause any problems. Although the amounts of lu-bricant were much higher, occurring fracture offered new bindingpositions. Jarosz and Parrot [13] showed for the brittle behavingdicalcium phosphate that tensile strength was not changed by theaddition of magnesium stearate, even the amount was up to 2%,which is comparable to the added amount for the direct compressionof magnesium carbonate. For materials undergoing brittle fracture,this way of lubrication seems to be promising.

Fig. 3. SEM pictures of the starting materials, MCC with normal density (a), with high density (b), powder cellulose (c), magnesium carbonate (d) and lactose (e).

Table 5Mean particle size [μm] of the granules between 315 and 630 μm (mean ± s; n = 3).

Specific compaction force MCC High densityMCC

Powdercellulose

Magnesiumcarbonate

Lactose

2 kN/cm 453.4 ± 0.8 463.3 ± 0.7 486.0 ± 4.0 478.8 ± 1.7 488.2 ± 4.04 kN/cm 470.9 ± 1.1 472.7 ± 1.0 488.6 ± 1.1 490.1 ± 1.3 483.4 ± 2.08 kN/cm 478.4 ± 2.3 477.0 ± 1.0 490.7 ± 0.3 493.1 ± 1.5 484.9 ± 1.710 kN/cm 479.2 ± 1.2 479.0 ± 0.9 490.9 ± 1.1 490.1 ± 0.9 485.0 ± 0.512 kN/cm 481.1 ± 0.6 481.0 ± 1.2 489.8 ± 0.5 493.7 ± 1.3 485.4 ± 0.5

161J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

Fig. 4. SEM pictures of granules from produced with a specific compaction force of 12 kN/cm, MCC with normal density (a), with high density (b), powder cellulose (c), magnesiumcarbonate (d) and lactose (e).

Table 6Added amount of MgSt in % for the surface proportional lubrication to the dry granules produced with different specific compaction forces.

Specific compaction force MCC MCC(high density)

Magnesiumcarbonate

Powdercellulose

Lactose

0 kN/cm 1.04 0.87 2.08 0.93 0.762 kN/cm 0.25 0.21 0.20 0.24 0.158 kN/cm 0.16 0.13 0.17 0.19 0.1212 kN/cm 0.13 0.11 0.16 0.19 0.10

162 J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

Fig. 5. Compactibility curves of MCCwith normal (a), and high density (b) powder cellulose (c), magnesium carbonate (d) and lactose (e) for the usage of surface proportional lubrication(mean ± sd; n = 10).

Fig. 6.Magnesium carbonate tablets compressed from powder with 0.2% MgSt.

163J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

Duberg and Nyström [21] used a similar approach to standardize theaddition of magnesium stearate aiming a constant surface ratio of ninebetween lubricant and material surface. In this study the same type ofMgSt was used for all experiments. Therefore, specific surface area ofthe magnesium stearate did not change during the experiments andthe addition of MgSt by the correlation constant of 2.5 μg/cm2, willalso result in a constant ratio of the surface areas. In contrast to thisstudy, no lubrication problems occurred by Duberg and Nyström [21].They also compressed particles of different size fractions, but the sizedifferences were much smaller compared to this study. For five mate-rials, fractions of 90–250 μm and 355–500 μm were investigated,whereas for the lubricant sensitive starch only the smaller fractionwas investigated. Due to the smaller differences in size, the amount ofMgSt varied just between 0.05 and 0.27%. Resulting from this, no lubri-cation problems occurred for most of thematerials. The compression ofstarch was problematic, as no coherent compacts could be formed withan addition of 0.23% of MgSt. This also indicates an overlubrication inthe case of starch. Therefore both studies were not contradictory but

the actual study rather investigated the surface proportional lubricationin more detail. Overall, standardizing the lubrication to the specific ex-ternal surface area will not work as a universal method. Although itcould be expected that for finer materials a bigger surface has to be

164 J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

occupied for a sufficient lubrication, correlation with the specific exter-nal surface area of the material will not work for all materials.

3.2.2. Internal lubrication vs. external lubricationExternal lubrication with manually lubricating die and punches was

possible for all five excipients. No materials stuck to the machine parts,which allowed a continuous process without any cleaning stops. Com-paring the materials, different behaviors were observed during the in-ternal lubrication with different amounts of magnesium stearate. Asbefore, the materials could be divided in two groups (see Section 3.2.1).

For the brittle behaving materials, magnesium carbonate and lac-tose, problems occurred during tableting with lower amounts of MgStwithin the internal lubrication. Powder compression with amounts of0.1 and 0.2% of MgSt was not possible for lactose as material stuck tothe die and the punches. For magnesium carbonate, tableting processcould be performed with an amount of 0.2% of MgSt, however 0.1%MgSt was insufficient to avoid sticking. In these cases, the surfaces ofthe tablets were not smooth (Fig. 6) and the tableting process had to

Fig. 7.Compactibility curves of the direct compression ofMCCwith normal (a), andhigh densityof MgSt (mean ± sd; n = 10).

be interrupted. Due to the sticking problems, just the surface propor-tional lubrication, with an amount of 0.76%MgSt, served as comparisonfor lactose to the external lubrication. A slight increase in tensilestrength could be observed for the external lubrication (Fig. 7e). In thecase of magnesium carbonate the increase in strength for the externallubrication was even smaller, for the tablets containing 2% MgSt(surface proportional lubrication) as well as for those containing a ten-fold lower amount (Fig. 7d). Due to the creation of lubricant free sur-faces during the fragmentation described by de Boer [12], lubricantsensitivity is small for magnesium carbonate and lactose. Especially formagnesium carbonate, an increase in lubricant amount caused almostno loss in strength. However, using too small amounts of lubricantperforming an internal lubricationwill offer limits, as a tabletingprocesswithout problems was impossible.

MCCs and powder cellulose formed the second group. For thesethree materials, tableting with internal lubrication was possible downto 0.1% MgSt. All compaction curves showed the same behavior. In-creasing the amount of lubricant resulted in a decrease in tensile

(b), powder cellulose (c)magnesium carbonate (d) and lactose (e)with different amounts

165J. Mosig, P. Kleinebudde / Powder Technology 266 (2014) 156–166

strength. External lubrication led always to tablets with the highestcompactibility. Tablets containing MgSt proportional to the specificsurface area showed the lowest strength caused by the highestamount of lubricant (Fig. 7a, b, c). Differences between the varyingamounts of MgSt during the internal lubrication as well as betweeninternal and external lubrication were more pronounced for thesematerials compared to lactose and magnesium carbonate. Powdercellulose presented the strongest increase switching from the inter-nal lubrication with the highest amount to the external lubrication(Fig. 7c). Considering differences in strength between lubricationwith 0.1 and 0.2% MgSt and external MgSt, increase is comparablefor both types of MCC and powder cellulose. Loss in strength wasabove 30% for a usage of 0.1% MgSt and nearly 40% for 0.2% MgSt.Even if the reductions were similar, high density MCC offered aslightly higher decrease than both other materials of this group.Compared to this remarkable reduction of strength for both typesof MCC and powder cellulose decline for magnesium carbonate isjust about 5%, changing from an external lubrication to a lubricationwith 0.2% MgSt. These differences occurred due to differing lubricantsensitivity. As MCC and powder cellulose present a plastic and elasticdeforming behavior, no lubricant free surfaces will be generatedduring the compressing step (see 3.1.1).

4. Conclusion

Tabletingwith amounts ofMgSt proportional to the specific externalsurface area was feasible as no sticking problems occurred for all of thefive materials, for the raw materials as well as for the granules. Theamount of MgSt was sufficient, even the relative low amounts of 0.1to 0.2% MgSt for the granule compression of lactose and magnesiumcarbonate. With such amounts, tableting of these raw materials wasnot possible. Hence, surface proportional lubrication seemed to providean indication for sufficient lubrication for brittle behaving materials.However, for the direct compression of MCC and powder cellulose, aloss in strength occurred due to the high amount of lubricant. Further-more, for these three materials direct compression was also possiblewith low amounts of MgSt, comparable to the concentrations added tothe granule compression. Although the specific surface area is muchhigher for the powdered material, the amount of lubricant needed forthe compression will not differ between the raw material and thegranules in such dimensions as it was expected due to the differencesin the specific surface area. In this case, specific surface area is not thecrucial part for the lubrication decision and a lubrication proportionallyto the specific external area resulted in an overlubrication of the pow-der. Considering these results, the mechanism behind the lubricationwith magnesium stearate offered new questions. The effect of lubrica-tion will not just occur due to film forming around the particles. If thiswas the case, surface proportional lubrication would not cause anyproblems for plastically deforming materials as well as for brittledeforming materials. Lubricant free surfaces for new bindings duringcompression should be equal for the different granule compressions aswell as for the direct compression. Thus, even for the lubricant-sensitive materials, strength of the direct compressed tablets shouldbe comparable to those of granule compression, if not actually be higherin the case of dry granules.

Overall, the results showed that the addition of magnesium stearateproportional to the specific surface area was not suitable as a universaltool for a comparable lubrication. For a comparable internal lubricationmore factors in addition to the specific external surface area have to betaken into account. Lubrication phenomena are more complex as justthe coverage of the external surface area of the material with MgSt,that considerations of the particle morphology and the lubricant distri-bution might be useful for further investigations on the lubricantaddition.

The results for the external lubricationwere promising. The tabletingprocess was practicable without sticking for all five different materials

and the compactibility was higher or comparable to tablets with thelowest concentration of MgSt.

Even if tableting with equal amounts for direct compression andgranule compression will be practicable, critical examination of the re-sults should be performed. Otherwise lubrication within the tabletingprocesses is not comparable and the impact of the lubricating mightbe underestimated, even neglected. Hence, especially literature aboutthe reduced compactibility of dry granules should be reconsideredwith respect to lubrication phenomena. In conclusion, external lubrica-tion should be the first choice to receive truly comparable results fortableting processes. Using this method, there is no risk for an over- orunderlubrication aswell as for complicationswithin the tablet strength.

Acknowledgments

The authors would like to thank Evonik Industries for providingthe Friedrich manometer to the Institute of Pharmaceutics andBiopharmaceutics in Duesseldorf. This applies in particular toMr. Weisbrod for the support in organization and startup.

References

[1] A.W. Hölzer, J. Sjögren, Evaluation of some lubricants by the comparison of frictioncoefficients and tablet properties, Acta Pharm. Suec. 18 (1981) 139–148.

[2] N.-O. Lindberg, Evaluation of some tablet lubricants, Acta Pharm. Suec. 9 (1972)207–214.

[3] W.A. Strickland, E. Nelson, L.W. Busse, T. Higuchi, The physics of tablet compressionIX. Fundamental aspects of tablet lubrication, J. Am. Pharm. Assoc. 45 (1956) 51–55.

[4] E. Shotton, C.J. Lewis, Some observations on the effect of lubrication on the crushingstrength of tablets, J. Pharm. Pharmacol. 16 (1964) 111T–120T.

[5] G.K. Bolhuis, A.J. Smallenbroek, C.F. Lerk, Interaction of tablet disintegrants andmagnesium stearate during mixing I: effect on tablet disintegration, J. Pharm. Sci.70 (1981) 1328–1330.

[6] T.A. Iranloye, E.L. Parrott, Effects of compression force, particle size, and lubricantson dissolution rate, J. Pharm. Sci. 67 (1978) 535–539.

[7] G.K. Bolhuis, C.F. Lerk, H.T. Zijlstra, A.H. De Boer, Film formation bymagnesium stea-rate during mixing and its effect on tableting, Pharm. Weekbl. 110 (1975) 317–325.

[8] C.F. Lerk, G.K. Bolhuis, S.S. Smedema, Interaction of lubricants and colloidal silicaduring mixing with excipients. I. Its effect on tableting, Pharm. Acta Helv. 52(1977) 33–39.

[9] C.F. Lerk, G.K. Bolhuis, Interaction of lubricants and colloidal silica during mixingwith excipients. II: Its effect on wettablility and dissolution velocity, Pharm. ActaHelv. 52 (1977) 39–44.

[10] A.C. Shah, A.R. Mlodozeniec, Mechanism of surface lubrication: influence of durationof lubricant-excipient mixing on processing characteristics of powders and proper-ties of compressed tablets, J. Pharm. Sci. 66 (1977) 1377–1382.

[11] J. Bossert, A. Stains, Effect of mixing on the lubrication of crystalline lactose bymagnesium stearate, Drug Dev. Ind. Pharm. 6 (1980) 573–589.

[12] A.H. De Boer, G.K. Bolhuis, C.F. Lerk, Bonding characteristics by scanning electronmi-croscopy of powders mixed with magnesium stearate, Powder Technol. 20 (1978)75–82.

[13] P.J. Jarosz, Effect of lubricants on tensile strengths of tablets, Drug Dev. Ind. Pharm.10 (1984) 259–273.

[14] H. Vromans, G.K. Bolhuis, C.F. Lerk, Magnesium stearate susceptibility of directlycompressible materials as an indication of fragmentation properties, PowderTechnol. 54 (1988) 39–44.

[15] A. Almaya, A. Aburub, Effect of particle size on compaction of materials with differ-ent deformation mechanisms with and without lubricants, AAPS PharmSciTech 9(2008) 414–418.

[16] M.E. Johansson, Granular magnesium stearate as a lubricant in tablet formulations,Int. J. Pharm. 21 (1984) 307–315.

[17] M.E. Johansson, Granular magnesium stearate as a lubricant, Acta Pharm. Suec. 24(1987) 45.

[18] C. Frattini, L. Simioni, Should magnesium stearate be assessed in the formulation ofsolid dosage forms by weight or by surface area? Drug Dev. Ind. Pharm. 10 (1984)1117–1130.

[19] K.P. Rao, G. Chawla, A.M. Kaushal, A.K. Bansal, Impact of solid-state properties on lu-brication efficacy of magnesium stearate, Pharm. Dev. Technol. 10 (2005) 423–437.

[20] S. Patel, A. Kaushal, A. Bansal, Lubrication potential of magnesium stearate studiedon instrumented rotary tablet press, AAPS PharmSciTech 8 (2007) 57–64.

[21] M. Duberg, C. Nyström, Studies on direct compression of tablets. VI. Evaluation ofmethods for the estimation of particle fragmentation during compaction, ActaPharm. Suec. 19 (1982) 421–436.

[22] P. Kleinebudde, Roll compaction/dry granulation: pharmaceutical applications, Eur.J. Pharm. Biopharm. 58 (2004) 317–326.

[23] S. Malkowska, K.A. Khan, Effect of re-compression on the properties of tabletsprepared by dry granulation, Drug Dev. Ind. Pharm. 9 (1983) 331–347.

[24] W. Friedrich, Gerät zur Messung der spezifischen Oberfläche empfindlicher Güter,Chem. Ing. Tech. 29 (1957) 104–107.

166 J. Mosig, P. Kleinebudde / Powder Te

[25] A.R. Gupte, Messung der spezifischen Oberfläche grober Granulate und dermittleren Porengröße von Tabletten, Acta Pharm. Technol. 22 (1976) 153–168.

[26] B. Koglin, K. Leschonski, W. Alex, Teilchengrößenanalyse. 7. Oberflächenmessung,Chem. Ing. Tech. 46 (1974) 984–987.

[27] J.T. Fell, J.M. Newton, Determination of tablet strength by the diametral-compressiontest, J. Pharm. Sci. 59 (1970) 688–691.

[28] F. Freitag, Walzenkompaktieren und Trockengranulieren zur Verbesserung desTablettierverhaltens anorganischer Hilfsstoffe am Beispiel von MagnesiumcarbonatundCalciumcarbonat, (Dissertation) Institute of Pharmaceutics andBiopharmaceutics,2004.

[29] F. Jaminet, H. Hess, Untersuchung über Kompaktierung und Trockengranulierung,Pharm. Acta Helv. 41 (1966) 39–58.

[30] E.L. Parrott, Densification of powders by concavo-convex roller compactor, J. Pharm.Sci. 70 (1981) 288–291.

Johanna Mosig studied Pharmacy at the Heinrich-Heine-University Duesseldorf, Germany from 2005 to 2009. Since2011 she is a PhD student at the Institute of Pharmaceuticsand Biopharmaceutics at the Heinrich-Heine-UniversityDüsseldorf under supervision of Prof. Peter Kleinebudde.Her research focuses on granulation methods, especially

Peter Kleinebudde is a full professor for PharmaceuticalTechnology at the University Duesseldorf since 2003. Peterwas the president of the International Association forPharmaceutical Technology (APV) from 2002 to 2010.Since 2010 he is the head of the APV focus group SolidDosage Forms. He is a member of the editorial boards ofthe Eur J Pharm Biopharm, J Pharm Sci, AAPS PharmSciTech,and Pharm Dev Tech. In 2004 he was designated as AAPSFellow and 2013 he received the Doctor honoris causa fromthe University of Szeged. His main research interests aresolid dosage forms and pharmaceutical processes like rollcompaction/ dry granulation, extrusion and coating.

chnology 266 (2014) 156–166