investigation of compacted hydrophilic and hydrophobic colloidal silicon dioxides as glidants for...

www.elsevier.com/locate/powtec

Powder Technology 141 (2004) 31–43

Investigation of compacted hydrophilic and hydrophobic colloidal silicon

dioxides as glidants for pharmaceutical excipients

S. Jonata, S. Hasenzahlb, M. Drechslerb, P. Albersb, K.G. Wagnera, P.C. Schmidta,*

aDepartment of Pharmaceutical Technology, Eberhard-Karls-University of Tubingen, Auf der Morgenstelle 8, D-72076 Tubingen, GermanybDegussa AG, D-63403 Hanau, Wolfgang, Germany

Received 20 August 2003; accepted 27 January 2004

Available online 1 April 2004

Abstract

The glidant properties of compacted hydrophilic and hydrophobic colloidal silicon dioxides were compared with respect to mixing time

and mixer type using microcrystalline cellulose, pregelatinized starch and a-lactose-monohydrate as model excipients. Flowability studies,

including angle of repose measurements and a novel dynamic conveyor belt method, showed differences in the flow enhancement between

the colloidal silicon dioxide types. An influence of mixing conditions on flowability was also observed for hydrophilic colloidal silicon

dioxide. These differences could be explained by SEM and XPS investigations, indicating the importance of a uniform coverage of the

glidant particles on the excipient’s surface. The influence of size and distribution of the colloidal silicon dioxide particles on the surface of the

excipient, mixing time, mixer type are described in detail. In addition, moisture studies showed that colloidal silicon dioxide protects the

excipients against a flowability decrease caused by humidity.

D 2004 Elsevier B.V. All rights reserved.

Keywords: Colloidal silicon dioxide; Flowability studies; Mixing process; Moisture studies; Scanning electron microscopy; X-ray photoelectron spectroscopy

1. Introduction layer on the powder or granulate particles causing them to

Glidants are incorporated into solid dosage forms to

improve the flow properties of cohesive powders and

granulates. When applied to tableting operations, glidants

improve flow into the hopper and the die cavities of the

tablet press. They increase the tablet weight, decrease the

weight variation and minimize the tendency of powder or

granule components to separate or segregate due to exces-

sive vibrations. For capsules, addition of glidants ensures a

high accuracy of metering and a uniform distribution of

active ingredients. The mode of action is not yet well

understood, but two theories have been described [1]. The

fine glidant particles adhere to the surface of the powders or

granulates, increasing the distance between two particles,

which in turn leads to a reduction of the attraction forces

between them. The second theory can be described as a ball

bearing type of action. Glidants form a monoparticulate

0032-5910/$ - see front matter D 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.powtec.2004.01.020

* Corresponding author. Tel.: +49-70-71-29-24-62; fax: +49-70-71-59-

55-31.

E-mail address: [email protected]

(P.C. Schmidt).

roll over one another. The rough surface is thus smoothed

out, reducing the frictional and adhesive forces that operate

between the surfaces.

Glidants used in pharmacy include talc, colloidal silicon

dioxide, calcium phosphates and various metallic stearates

[2]. Several groups have investigated the addition of gli-

dants to a variety of powders and noted that the silica-type

glidants are the most efficient because of their small particle

size [1–3]. AEROSILR 200 (Degussa, Dusseldorf, Ger-

many) is a highly disperse or colloidal silicon dioxide (also

referred to as fumed silica) that is commonly used to

improve flowability [4]. This conventional colloidal silicon

dioxide has low bulk and tapped densities and can produce

dust if handled improperly. Furthermore, it requires consid-

erable storage space and is relatively complex to process. In

order to improve the handling of colloidal silicon dioxide,

special mechanical processes were developed for the homo-

geneous compaction of colloidal silicon dioxide [5]. As a

result, two new products have been recently introduced:

AEROSILR 200 VV Pharma, which is especially designed

for the pharmaceutical industry, and AEROSILR R 972 V.

AEROSILR 200 VV Pharma is hydrophilic and chemically

Table 1

Comparison of the physicochemical properties of hydrophilic and

hydrophobic colloidal silicon dioxide types used in the study

Properties AEROSILR200 (hydrophilic,

non-compacted)

AEROSILR200 VV Pharma

(hydrophilic,

compacted)

AEROSILRR972 V

(hydrophobic,

compacted)

Average primary

particle size (nm)

12 12 16

BET surface

area (m2/g)

200F 25 200F 25 110F 20

Tapped density

(g/cm3)

approx. 0.05 approx. 0.12 approx. 0.09

Silanol group

density (nm� 2)

approx. 2 approx. 2 approx. 0.75

S. Jonat et al. / Powder Technology 141 (2004) 31–4332

identical to AEROSILR 200. It differs from conventional

colloidal silicon dioxide only in its higher tapped density

and its larger secondary agglomerates. The compacted

product AEROSILR R 972 V is hydrophobic, a result of

dimethyl silyl groups chemically bound to the silica surface

[6]. The hydrophobic nature of AEROSILR R 972 V gives

it superiority over water-wettable conventional colloidal

silicon dioxide in a range of pharmaceutical applications

including slow-release formulations and W/O emulsions.

Gstirner and Pick first compared hydrophobic and hydro-

philic AEROSILR products in the late 1960s, evaluating

their effect on the flowability of zinc oxide and their

influence on the water absorption of various powders

[7,8]. They showed that AEROSILR R 972 enhanced the

flowability of zinc oxide and reduced the water uptake of

powders to a greater extent than did the hydrophilic type

AEROSILR 200. With the introduction of the compacted

hydrophobic and hydrophilic AEROSILR products, their

effects on the flowability and water absorption of typical

tableting excipients seemed to be an interesting topic.

Physical test methods for powder flow characterization

described in the literature are numerous and exist in many

variations [9–11]. Amidon et al. [12] have recommended

several procedures for the measurement of flow properties.

However, individual tests are not always able to measure

small differences in flow between similar powders and to

rank their flow properties [13]. Generally, it is not possible

to characterize powder flow by one single test methodology.

Thus the combination of various tests is a better approach to

achieve reliable data. Two methods were chosen for this

work: the classical static angle of repose and a dynamic

conveyor belt method.

The primary particles of colloidal silicon dioxide are

linked into relatively stable aggregates that can range up to

several hundred nanometers in size. The aggregates in turn

form larger agglomerates, which can be broken down by the

shear forces that occur during powder processing. The

mixing time and the mixer type can therefore influence

the size of the agglomerates and as a result the flowability.

The effect of mixer type and mixing time on pharmaceutical

mixtures has been described in the literature [14–16]. The

purpose of this work was to compare the flow-enhancing

properties of conventional colloidal silicon dioxide to the

new compacted hydrophilic and hydrophobic types with

respect to mixing time, mixer type and relative humidity.

2. Experimental procedures

2.1. Materials

All colloidal silicon dioxide types were obtained from

Degussa. Their physico-chemical properties are summarized

in Table 1 [6].

Three excipients used as filler/binders for direct compres-

sion of tablets were selected for their different structures, flow

properties and compressibility behavior: Microcrystalline

cellulose (AvicelR PH 101, FMC Biopolymer, Cork, Ire-

land), pregelatinized starch (Starch 1500R, Colorcon, Kent,England) and agglomerated a-lactose-monohydrate (Tab-

lettoseR 80, Meggle, Wasserburg, Germany).

Potassium acetate, magnesium chloride, potassium car-

bonate, sodium bromide, sodium nitrite, sodium chloride,

potassium chloride and potassium nitrate (Merck, Darm-

stadt, Germany) were of analytical grade and used as

saturated salt solutions.

2.2. Methods

2.2.1. Mixing process

To assess the influence of the mixing time and the mixer

type on the flowability of the mixtures, different mixers and

mixing conditions were chosen, as shown in Fig. 1 using

Starch 1500R as an example: a free-fall mixer (Turbula, W.A.

Bachofen, Basel, Switzerland), with a 2 l vessel volume, a

maximum filling degree of 75% and a rotational speed of 42

rpm; a high speed mixer based on the plowshare principle

(SW 1/S, Erweka, Heusenstamm, Germany), with a 5 l drum,

a filling degree of 70% and a rotary speed of 2400 rpm; and a

pin mill (160 Z, Alpine, Augsburg, Germany) with a rotary

speed of 14,000 rpm. Based on previous investigations the

amount of colloidal silicon dioxide was set at 0.5 wt.%

AEROSILR based on the total formulation.

2.2.2. Determination of flowability

2.2.2.1. Angle of repose. The angle of repose was mea-

sured using a sieve-cone-method according to DIN ISO

4324. The distance between the sieve and the metal cone

was 10 cm. First, 100 g of material were sieved through an

800 Am sieve onto a metal cylinder with a radius (r) of 25

mm to determine the exact distance between the cone and

the sieve. The sieve was then fixed 3 cm above the top of

the powder cone and 100 g material were sieved through an

800 Am sieve onto the metal cylinder. The height (h) of the

powder cone was measured and the angle of repose (a) wascalculated using the following equation: tana = h/r. The

Fig. 1. Mixing process of Starch 1500R with 0.5% AEROSILR.

S. Jonat et al. / Powder Technology 141 (2004) 31–43 33

mean, the standard deviation and the 95% confidence

interval of six samples were calculated.

2.2.2.2. Conveyor belt. A novel method was designed to

measure the flowability of the different mixtures using a

conveyor belt (Fig. 2). The conveyor belt consists of a

circular strip of plastic material 77 cm long and 10 cm wide

(Reiff-Technische Produkte, Reutlingen, Germany). A mo-

tor (model 3557K024CR, Dr. F. Faulhaber, Schonaich,

Germany) rotates the strip at a speed of 2.28 cm/s. Mixtures

were passed through the funnel to the belt and conveyed to a

balance (model PM 6100, Mettler-Toledo, Giessen, Ger-

many). The accumulated mass of powder versus time was

plotted at 0.5 or 1-s intervals. The diameter of the funnel’s

orifice was 1.9 cm. The gap between the funnel and the belt

was set to 2 mm for Starch 1500R and TablettoseR 80 and

to 3 mm for AvicelR PH 101 using a micrometer screw

(0.01 mm of accuracy, Mitutoyo Messgeraete, Neuss, Ger-

many). The weight of the powder conveyed was plotted

Fig. 2. Schematic diagram of the conveyor belt.

automatically (HP VEE program, version 5.0). Mixtures 2

and 4 of Starch 1500R were investigated for 2.5 min,

Mixtures 2 and 4 of TablettoseR 80 for 2 min, and Mixtures

2 and 4 of AvicelR PH 101 for 6 min. All measurements

were repeated three times. In order to compare the different

flow behavior of the mixtures, the slopes of the linear

regression of the mass accumulated versus time curves were

calculated.

2.2.3. Moisture studies

Pyrex desiccators containing appropriate saturated salt

solutions (Table 2) in distilled water were prepared to create

chambers with different levels of relative humidity. All

desiccators were kept at 20F 0.5 jC to maintain the desired

relative humidity level.

Mixture M4 of each excipient without AEROSILR,containing 0.5 wt.% AEROSILR 200 VV Pharma or

containing 0.5 wt.% AEROSILR R 972 V were placed

Table 2

Saturated salt solutions used for different relative humidity levels

Saturated salt solution Relative humidity

at 20 jC (%)

Potassium acetate 20

Magnesium chloride 33

Potassium carbonate 44

Sodium bromide 59

Sodium nitrite 65

Sodium chloride 75

Potassium chloride 85

Potassium nitrate 93


in each of the eight desiccators (24 samples total) for 18

days. The study was divided into two parts: water uptake

and flowability. For the water uptake, 90 ml samples of

each mixture were accurately weighed (model PG 5002-S,

Mettler-Toledo, Giessen, Germany) into 14 cm diameter

Petri dishes and then placed in the dessicators. After 3, 7,

12 and 18 days, the samples were again accurately

weighed. The water uptake was calculated and expressed

in %. For the flowability, the angle of repose was mea-

sured before and after 18 days storage in the dessicators.

For Starch 1500R, the experiment was performed three

times in order to evaluate the reproducibility.

2.2.4. Particle size analysis

The particle size distributions of AvicelR PH 101, Starch

1500R and TablettoseR 80 were measured by laser diffrac-

tion spectrometry (Mastersizer 2000, Malvern Instruments,

Herrenberg, Germany) using the dry-dispersing system

Scirocco 2000 (Malvern Instruments). The dispersing air

pressure was 3 bar for AvicelR PH 101, no dispersing air

pressure was used for Starch 1500R and TablettoseR 80.

Data were directly collected by means of the system

software (Malvern Instruments). The mean value of three

measurements was calculated.

2.2.5. Scanning electron microscopy (SEM)

The powder mixture was examined using a Zeiss DSM

940 A SEM (Carl Zeiss, Oberkochen, Germany) equipped

with a Contax M 167 MT camera (Yashica-Kyocera,

Hamburg, Germany). Each mixture was fixed on an

aluminium sample holder using a double-adhesive tape

(Tempfix) and coated with a thin gold layer prior to

examination using a Sputter Coater E 1500 (Bio-Rad,

Munich, Germany). The samples were sputtered four

times for 60 s and exposed to 20 mA current and 2.1

kV acceleration voltage in a vacuum of 2–3 Pa. The

micrographs were taken at 5 kV and at a magnification of

5000 and 20,000.

2.2.6. X-ray photoelectron spectroscopy (XPS)

The samples were measured as loose powders without

any pretreatment such as compressing. They were sup-

ported by a tantalum sample container and were transferred

into the XPS instrument by means of a differential pump-

ing stage. After evacuation, a sample was transferred into

the main spectrometer chamber and was measured at about

5� 10� 8 mbar. Broad area XPS conditions were adjusted

to obtain information on the surface properties of about 0.5

cm2 of the material in a single spectrum. The XPS

measurements were performed using MgKa radiation at a

power of 150 W. The electron energy analyzer (Leybold

EA11A, Cologne, Germany) was operated at a pass-energy

of 75 eV in the fixed analyzer transmission mode. Addi-

tional details are given in Ref. [17]. The spectra were

treated as follows: subtraction of the X-ray satellites,

smoothing by polynomial fits, Shirley-type background

subtraction [18], peak integration and correction by sensi-

tivity factors according to the elemental ionization cross

sections and the transmission function of the electron

energy analyzer. The elemental sensitivity factors for the

quantitative evaluation were: Si 2p 0.4, O 1s 0.78 and C 1s

0.34. The binding energy scale of the spectrometer was

referenced to the Au 4f7/2 signal at 84.0 eV.

3. Results and discussions

3.1. Flowability studies

3.1.1. Angle of repose method

The dependence of the angle of repose of AvicelRPH 101 and Starch 1500R mixtures on the AEROSILR-type, the mixing time and the mixer type is shown in

Fig. 3. As expected, each type of colloidal silicon

dioxide clearly improved the flowability of AvicelR PH

101 and Starch 1500R for all mixtures. According to the

classification of Carr [12], which relates the flow prop-

erties to the angles of repose, AvicelR PH 101 became

by the addition of 0.5% AEROSILR a fairly (36–40j)or a passably (41–45j) flowing material and Starch

1500R showed good (31–35j) or fairly good (36–40j)flow properties. A statistical analysis (ANOVA and

Newman–Keuls test) was used to rank the increase in

flowability: for mixtures M1, M2 and M3, AEROSILRR 972 V improved the flowabilty most, followed by

AEROSILR 200 VV Pharma and AEROSILR 200. For

mixture M4, there was no statistically significant differ-

ence between AEROSILR 200 VV Pharma and AERO-

SILR R 972 V. Under these mixing conditions, the two

AEROSILR types improved the flowability to the same

extent, whereas AEROSILR 200 showed again the

lowest flowability enhancement.

The flowability of mixtures containing AEROSILR 200

or AEROSILR 200 VV Pharma were strongly influenced

by the mixing conditions. Flowability increased with mix-

ing time and energy, reaching an optimum at mixture M4.

The results of the ANOVA test for AvicelR PH 101

mixtures with AEROSILR R 972 V, however, indicated

that there were no differences in the angles of repose of

mixtures M1, M2, M3 and M4 (Fig. 3a). The flowability

increase obtained with 0.5% AEROSILR R 972 V was

influenced neither by the free-fall nor by the high speed

mixer. This means that the flowability optimum was

obtained directly after gentle mixing (M2) and remained

constant until mixture M4.

For Starch 1500R, from step 1 to step 4, the angle of

repose of AEROSILR 200, AEROSILR 200 VV Pharma

and AEROSILR R 972 V mixtures decreased from 39.2j to

35.1j, from 35.2j to 32.7j and from 34.1j to 32.4j,respectively (Fig. 3b). Again, the highest flowability was

obtained with and the mixing conditions had the lowest

influence on AEROSILR R 972 V. As a glidant, hydropho-

Fig. 3. The effect of AEROSILR-type, mixer type and mixing time on the angle of repose of AvicelR PH 101 (a), Starch 1500R (b) and TablettoseR 80

(c) mixtures. Error bars indicate the 95% confidence interval of six measurements.


bic AEROSILR R 972 V seems to be less sensitive with

respect to the mixing steps.

Fig. 3c depicts the angle of repose of TablettoseR 80

mixtures as a function of colloidal silicon dioxide type and

the mixing steps. TablettoseR 80 is already a free-flowing

material, nevertheless its flowability was further improved

by the addition of 0.5% AEROSILR. However, the angle ofrepose increased with increasing mixing steps for all mix-

tures. This observation can be explained by the reduction of

the particle size during the mixing process, as shown in

Table 3

The influence of mixer type and mixing time on the mean particle diameter

of AvicelR PH 101, Starch 1500R, and TablettoseR 80

Sample AvicelR PH

101 (Am)

Starch 1500R(Am)

TablettoseR80 (Am)

Bulk 48.8 73.3 159.9

M2 49.5 74.4 149.0

M4 49.7 75.4 130.6

M5 28.8 49.5 23.8

Table 4

Analysis of mass accumulation curves of AvicelR PH 101, Starch 1500Rand Tablettose 80 containing 0.5% AEROSILR R 972 V, 0.5% AEROSILR200 VV Pharma, 0.5% AEROSILR 200, and without AEROSILR

AEROSILRR 972 V

AEROSILR200 VV

Pharma

AEROSILR200

No

AEROSILR

AvicelR M2 0.3813 0.3500 0.3397 0.2878

PH 101 slopeF s 0.0017 0.0027 0.0027 0.0010

M4 0.3947 0.3971 0.3810 0.2661

slopeF s 0.0068 0.0093 0.0057 0.0228

Starch M2 0.3915 0.3566 0.3377 –

1500R slopeF s 0.0038 0.0098 0.0040 –

M4 0.3872 0.3707 0.3696 –

slopeF s 0.0065 0.0050 0.0032 –

TablettoseR M2 0.3545 0.3576 0.3707 0.3182

80 slopeF s 0.0042 0.0107 0.0044 0.0064

M4 0.3623 0.3623 0.3737 0.3265

slopeF s 0.0035 0.0035 0.0029 0.0046


Table 3. The agglomerates of TablettoseR 80 were

destroyed in the plowshare-mixer and in the pin mill,

leading to an increase in the adhesion forces between lactose

particles, which decreased the flowability and overshad-

owed the action of AEROSILR [19]. Moreover, the ag-

glomerate structure of the TablettoseR 80, which has many

irregularities and cavities, leads to a reduction of the

AEROSILR action in general due to the settling of AERO-

SILR particles into the cavities. The ANOVA results

revealed that there were no statistically significant differ-

ences between the AEROSILR-types.For each preparation, mixture M5 resulted in the highest

angle of repose, that is to say it had the poorest flowability.

The pin mill decreased flowability by reducing the particle

size of the materials (Table 3).

3.1.2. Dynamic conveyor belt method

The flow properties of each mixture were also studied

using the dynamic conveyor belt method. This method was

chosen to simulate the filling process of a die on a rotary

tablet press, where the vertical powder flow is superimposed

by a horizontal movement by the moving die table. The

powder mass accumulating on the balance was plotted

versus time for mixtures M2 and M4. The flow profiles of

AvicelR PH 101 mixtures were linear within 6 min indicat-

Fig. 4. Mass accumulated vs. time for mixtures M2 (left) and mixtures M4 (rig

AEROSILR 200 VV Pharma (2), 0.5% AEROSILR 200 (3) and without AERO

ing that the sample flowed continuously out of the funnel

(Fig. 4). Starch 1500R without AEROSILR blocked the

orifice of the funnel and therefore no slope was calculated.

A statistical analysis (ANOVA and Newman–Keuls test) of

the slope of the curves of AvicelR PH 101 and Starch

1500R confirmed the angle of repose results with respect to

both the influence of the AEROSILR-type and the mixing

conditions. For mixtures M2 the flowability improvement

ranking for AvicelR PH 101 and Starch 1500R was: AERO-

SILR R972 V>AEROSILR 200 VV Pharma>AEROSILR200>no AEROSILR (Table 4). For mixtures M4 the slopes

were significantly higher compared to mixtures M2 for

AEROSILR 200 VV Pharma and AEROSILR 200, while

no further increase in flowability was observed for AERO-

SILR R 972 V. This reveals again that the flowability of

mixtures with AEROSILR 200 VV Pharma and AEROSILR200 were strongly influenced by the mixing conditions, a

ht) of AvicelR PH 101 containing 0.5% AEROSILR R 972 V (1), 0.5%

SILR (4).

Fig. 5. SEM images of AvicelR PH 101 containing 0.5% AEROSILR R 972 V M2 (a), 0.5% AEROSILR 200 M2 (b), 0.5% AEROSILR 200 VV Pharma M2

(c and d) and AEROSILR 200 M4 (e), and 0.5% AEROSILR 200 VV Pharma M4 (f). The bars represent 500 nm and 2 Am for images a, b, c, e, f and for

image d, respectively.


Table 5

XPS-Spectra of AvicelR PH 101 alone and of mixtures containing 0.5% of

AEROSILR R 972 V, AEROSILR 200 VV Pharma, and AEROSILR 200

AvicelRPH 101

containing

AEROSILRR 972 V

AEROSILR200 VV

Pharma

AEROSILR200

No

AEROSILR

M2 M4 M2 M4 M2 M4 Bulk

C 1 s 29.8 39.5 37.7 27.2 37.0 25.8 43.4

O 1 s 63.0 50.2 59.6 64.6 59.8 64.9 56.6

Si 2p 7.22 10.3 2.76 8.25 3.25 9.21 0.0


longer mixing time resulting in better powder flow, whereas

AEROSILR R 972 V is almost independent of mixing time

and energy. As with the angle of repose method, no

differences between the three AEROSILR types were

found in the conveyor belt experiments for TablettoseR80 (Table 4).

3.1.3. Surface analysis by SEM/XPS

In order to elucidate the differences in the influence of

the AEROSILR-types on powder flowability, scanning

electron microscopy (SEM) studies were performed to

visualize the degree of AEROSILR particle coverage and

distribution on the AvicelR PH 101 surface. Fig. 5 depicts

the corresponding SEM images. Samples were carefully

checked to make sure that the differences observed are

representative for each particular mixture. The comparison

of Fig. 5a, b and c shows that the surface of the substrates

was covered with small particles of colloidal silicon dioxide

(aggregates and agglomerates), however, to varying

degrees. The distribution of AEROSILR R 972 V on the

surface of AvicelR PH 101 was regular and homogeneous

without enrichment of AEROSILR particles at edges or in

cavities of the substrate (Fig. 5a). Although every sample

contained the same amount of AEROSILR, the coverage ofAvicelR PH 101 was less extensive and the distribution less

homogeneous for mixtures containing AEROSILR 200

Fig. 6. Correlation of angle of repose and XPS Si 2p signa

(Fig. 5b) and AEROSILR 200 VV Pharma (Fig. 5c). In

addition, large AEROSILR particles of up to 500 nm in size

could be detected on the surface of the substrate (Fig. 5d)

indicating that not all particles were sufficiently broken up

in M2. As Fig. 5e and f shows, higher energy mixing

conditions 4 are necessary to achieve a homogeneous

distribution of AEROSILR 200 and AEROSILR 200 VV

Pharma particles, as it is already achieved by the more

gentle mixing conditions 2 for AEROSILR R 972 V. The

degree of coverage qualitatively correlated to the flow-

enhancement ranking.

While SEM studies only allow for a qualitative assess-

ment, X-ray photoelectron spectrometry (XPS) enables a

quantitative measure of the degree of dispersion of AERO-

SILR on AvicelR PH 101 by detecting the surface coverage.

The evaluation of Si 2p photoelectron signals allowed a

quantitative comparison on the shielding of the substrate by

AEROSILR particles [17].

Table 5 shows the results of the quantitative evaluations

of the XPS spectra. In spite of the given bulk concentra-

tion of SiO2 (0.5%), the values for 0.5 cm2 of powder

surface ranged between 2.76 and 10.3 at.%. Due to the

high surface selectivity of XPS, the samples showing the

best dispersion of the colloidal silicon dioxide particles to

the substrate could be identified. All M4 samples and the

samples with hydrophobic AEROSILR R 972V showed

much higher concentrations of Si than the M2 samples.

The qualitative impressions and trends as suggested by

SEM (Fig. 5) were confirmed by quantitative XPS meas-

urements (Table 5).

Although there was no linear correlation between the

angle of repose and the Si 2p values, two clusters were

formed when correlating the two parameters as shown in

Fig. 6. One cluster contained the M2-mixtures with AERO-

SILR 200 and AEROSILR 200 VV Pharma at high angles

of repose and low Si 2p values, while the other represented

the M4-mixtures and the M2-mixtures of AEROSILR R

l intensity of AvicelR PH 101/AEROSILR mixtures.

Table 6

Results of Gaussian/Lorentzian line shape analyses of the C1s signal of

AvicelR PH 101 containing 0.5 wt % of AEROSILR R 972 V and without

AEROSILR

AvicelR PH 101 No AEROSILR AEROSILR R 972 V

containingEv % eV %

C1s peak 1 285.3 7 284.5 46

C1s peak 2 286.8 83 285.9 41

C1s peak 3 288.8 10 288.0 13


972 V at low angles of repose and increased AEROSILRcoverage (higher Si 2p values).

The actual degree of surface coverage was even higher

than shown by the XPS Si 2p signals in Table 5. It is also

necessary to take into consideration the corresponding

inorganic oxygen of the SiO2 particles covering the

oxygen- and carbon-containing structures of the organic

substrate. This means that in Table 5 the inorganic oxygen

signal contribution of the SiO2 particles and the organi-

cally bound oxygen of the AvicelR PH 101 added up to

the complete O1s signal. The same holds for the carbo-

naceous species. Since XPS was not only able to identify

and to quantify the presence of different elements but also

to identify the different types of chemical bonds, addi-

tional information was obtained. Fig. 7 compares the C1s

signals of the pure cellulose substrate and of a sample

covered with AEROSILR R 972 V (Si value: 10.3%,

Table 5). Table 6 shows the corresponding binding energy

values.

With respect to reference data on various materials

[20,21], it was evident that the binding energy value of

the C1s signal of AvicelR PH 101 was dominated by the

peak around 286.8 eV according to the –C–O–C– and

Fig. 7. Binding energy values of the carbon atoms of AvicelR PH 101 cont

–C–O–C–O–C– functions with different substitutional

patterns. Furthermore, and with respect to high resolution

XPS data reported by Beamson [22] for different kinds of

cellulose, it follows that the signal at about 285/286 eV was

due to CHx-type functions in side chains such as �CH2 �CH3 as known from ethylcellulose or the O� CH2 � CH2–

CH2–OH in hydroxypropylcellulose, the signal at about

286/287 eV was due to the C atoms in the ring which are

directly connected to the OH-groups in cellulose (or the

alkoxy functions of ethylcellulose) and the signal at about

287/288 eV showed the bridging oxygen atoms between the

rings. This indicates that the binding energy values for the

aliphatic carbon species in cellulose were enhanced due to

aining 0.5 wt.% of AEROSILR R 972 V (a) and no AEROSILR (b).


the numerous adjacent electronegative oxygen entities. On

the other hand, for the case of the AEROSILR R 972 V-

covered material, the aliphatic components which were

enriched in the topmost atomic layers of the silica particles

that shield the AvicelR PH 101 substrate were selectively

detected by means of XPS. The binding energy values of

these carbon atoms were detected at lower binding energy

(Fig. 7). The Gaussian/Lorentzian line shape analyses, as

summarized in Table 6, show that 46% of the detected

surface-carbon on the AvicelR PH 101/AEROSILR R 972

V sample was purely aliphatic (–CH3).

XPS could not only detect the varying degree of surface

coverage of AvicelR PH 101 by measuring the Si 2p signal

but could also differentiate between aliphatic carbon entities

of the cellulose substrate and the aliphatic chains in the

topmost atomic layers of the finely dispersed particles of

silanized silica such as AEROSILR R 972 V. It can be

measured how much of the aliphatic carbon at the surfaces

of the modified AvicelR PH 101 samples may contribute to

hydrophobic conditions (CH2-chains without adjacent oxy-

gen atoms, at 285 eV binding energy and less) and to

hydrophilic conditions (CH2-entities of the cellulose rings).

So the difference between the C1s peak 2 of AvicelR PH

101 and AvicelR PH 101 containing 0.5% AEROSILR R

972 V gave the exact surface coverage of AEROSILR R972

V on AvicelR PH 101, i.e. 42%, corresponding to a 10.3%

surface coverage given by Si 2p (Table 5) and to the

qualitative impression of Fig. 5a.

SEM and XPS studies provide an explanation for the

flow-enhancement and the sensitivity of the various AERO-

SILR-types to the mixing steps. AEROSILR R 972 V

resulted in the best flow-enhancement of AvicelR PH 101

compared to AEROSILR 200 VV Pharma and AEROSILR200 due to its high degree of dispersion. A higher degree of

Fig. 8. Water uptake of mixtures M4 of Starch 1500R, AvicelR PH 101 and Table

VV Pharma or AEROSILR R 972 V, after 18 days storage at different (constant) le

confidence interval of three measurements.

coverage allowed a better action of the glidant particles.

AvicelR PH 101 containing 0.5% AEROSILR R 972 V

showed equal angle of repose values for Mixture M2 and

Mixture M4, as the degree of coverage of AEROSILR R

972 V in both mixtures was the same. Under mixing

conditions 2, the agglomerates of AEROSILR R 972 V

have reached their final size and optimal distribution; hence

the high speed mixer (Mixture M4) showed no increase in

the degree of coverage. The hydrophobic agglomerates were

rapidly broken down and a stable distribution of AERO-

SILR R 972 V resulted.

3.2. Moisture studies

One of the advantages of hydrophobic colloidal silicon

dioxide, such as AEROSILR R 972 V, is its low and

relatively constant water content at all humidity levels,

while the hydrophilic types adsorb water at increasing

relative humidity [6]. Because of the low AEROSILRconcentration of 0.5%, this property could not be dem-

onstrated in this investigation (Fig. 8). Rather, the water

uptake was characteristic for each excipient and ranged

from � 0.5% to 9.5% for Starch 1500R, from � 0.5% to

7.3% for AvicelR PH 101 and from � 0.2% to 0.2% for

TablettoseR 80, independent of the addition of AERO-

SILR or the AEROSILR type used.

In spite of this, the angle of repose revealed an interesting

positive effect for all AEROSILR types with respect to

humidity (Fig. 9). For each excipient, mixtures without

AEROSILR always showed a higher angle of repose after

18 days in the humidity chamber, i.e. the flowability

decreased. Moreover, the higher the relative humidity, the

larger the increase in the angle of repose. This can be

correlated to the variation of interparticulate forces with

ttoseR 80, without AEROSILR or containing 0.5 wt.% of AEROSILR 200

vels of relative humidity and room temperature. Error bars indicate the 95%

Fig. 9. The effect of relative humidity on the angle of repose of AvicelR PH 101 (a), Starch 1500R (b), and TablettoseR 80 (c), without AEROSILR or

containing 0.5 wt.% AEROSILR 200 VV Pharma or AEROSILR R 972 Vafter 18 days at different levels of humidity. Error bars for Starch 1500R indicate the

95% confidence interval of three measurements.



humidity. Adhesion forces increase as humidity increases

resulting in larger angles of repose [23–25]. On the other

hand, the angles of repose of mixtures containing 0.5%

AEROSILR were hardly influenced by increasing the rela-

tive humidity, even though their water uptake was the same

as of mixtures without AEROSILR. In fact, AEROSILRreduced or avoided, depending on the type of excipient, the

increase in interparticulate forces. For AvicelR PH 101 and

Starch 1500R, there was no difference before and after 18

days in the chambers, even at high humidity (Fig. 9a and b).

For TablettoseR 80, angles of repose of mixtures containing

0.5% AEROSILR were mostly larger after 18 days, but this

increase was smaller when compared to the mixtures with-

out AEROSILR (Fig. 9c). These results confirmed that

AEROSILR protects the three excipients from a flowability

decrease under humid conditions [26]. Except for the best

flow enhancement, the hydrophobic colloidal silicon diox-

ide did not show further advantages over the hydrophilic

types under the experimental conditions chosen. Smaller

angles of repose were observed for Starch 1500R and

TablettoseR 80 mixtures containing the hydrophobic

AEROSILR R 972 V at some relative humidity levels, but

the differences were not statistically significant.

The decreasing flowability of the pure excipients at

increasing humidity can possibly be explained as the result

of the formation of a water film or an increase in the

capillary condensation between the excipient particles. The

addition of colloidal silicon dioxide increases the surface

roughness of the exipient particles and quite likely also the

average distance between them. Therefore capillary conden-

sation would not play an important role even at high relative

humidity levels, and the mixtures containing colloidal

silicon dioxide would retain good flow properties indepen-

dent of atmospheric moisture.

4. Conclusions

The results of our investigation of the glidant properties

of compacted hydrophilic and hydrophobic colloidal silicon

dioxides using microcrystalline cellulose, pregelatinized

starch and a-lactose-monohydrate as model excipients can

be summarized as follows:

(1) The novel types AEROSILR R 972 V and AEROSILR200 VV Pharma are efficient glidants. Apart from their

handling advantages, compacted colloidal silicon diox-

ides are even superior to the non-compacted AERO-

SILR 200 with respect to their powder flow enhancing

properties.

(2) Among the colloidal silicon dioxide types investigated

here, AEROSIL R 972 V is the most efficient glidant.

While already gentle mixing conditions are sufficient to

achieve high flowability enhancement with hydrophobic

AEROSILR R 972 V, the flowability increase obtain-

able with hydrophilic AEROSILR 200 VV Pharma and

AEROSILR 200 strongly depends on the mixing

conditions. Generally, the flowability of microcrystal-

line cellulose and pregelatinized starch mixtures con-

taining hydrophilic colloidal silicon dioxide increases

with higher mixing time and energy. Furthermore, the

angles of repose correlate well with the mass accumu-

lation curves obtained by the novel conveyor belt

method.

(3) SEM and XPS analyses of mixtures of microcrystalline

cellulose and AEROSILR show a clear correlation

between the degree and uniformity of coverage of the

colloidal silicon dioxide particles on the excipient’s

surface and the flow enhancement exerted by the

glidant. While the soft agglomerates of AEROSILR R

972 V reach their final size and optimal distribution

already under gentle mixing conditions, higher mixing

energies are required to break up the agglomerates of

both AEROSILR 200 and AEROSILR 200 VV Pharma

and to achieve a uniform coverage. Mixing conditions

generally need to be adjusted in a way to ensure an

optimum dispersion of the glidant particles on the

excipient surface. Severe mixing conditions must be

avoided since they may reduce the excipient particle

size, resulting in decreased flowability.

(4) At low glidant concentrations, the moisture uptake of the

excipients is not influenced by any of the colloidal

silicon dioxide types used. Rather the water adsorption

is characteristic of the pure excipients. However, even

after equilibrating at high humidity levels, all excipient

mixtures containing AEROSILR maintained a good

flowability, while the flowability of the pure excipients

turned strongly cohesive. This proves that the com-

pacted colloidal silicon dioxide types are at least as

efficient anti-caking agents as their non-compacted

counterparts.

Acknowledgements

We would like to thank Ann Gray for carefully reviewing

this text and giving valuable advice.

References

[1] B.B. Sheth, F.J. Bandelin, R.F. Shangraw, in: Pharmaceutical Dosage

Forms: Tablets, vol. 1. Marcel Dekker, New York, 1980, pp. 141–143.

[2] S. Varthalis, N. Pilpel, The action of colloidal silicone dioxide as a

glidant for lactose, paracetamol, oxytetracycline and their mixtures,

J. Pharm. Pharmacol. (1977) 37–40.

[3] W.A. Ritschel, A. Bauer-Brandl, Die Tablette, Editio-Cantor-Ver-

lag, Aulendorf, 2002, pp. 144–146.

[4] Technical Information No. 1237, AEROSILR 200 Pharma: A Versa-

tile Excipient for the Pharmaceutical Industry, Degussa, Dusseldorf,

2001, p. 3.

[5] Technical Information No. 1247, AEROSILR 200 VV Pharma: A

New Excipient for the Pharmaceutical Industry, Degussa, Dusseldorf,

2002, p. 5.


[6] Technical Bulletin Fine Particles No. 11, Basic Characteristics of

AEROSILR, Degussa, Dusseldorf, 2003, p. 12.[7] F. Gstirner, C. Pick, Der Einfluß von AEROSILR und ‘‘Fullstoff R

972’’ auf die Fließeigenschaft von Zinkoxid, Arch. Pharm. 800 (1967)

757–765.

[8] F. Gstirner, C. Pick, Einfluß von AEROSILR und AEROSILR R 972

auf das Wasseraufnahmevermogen von Puderrohstoffen, Arch.

Pharm. 800 (1969) 590–604.

[9] R.L. Carr, Evaluating flow properties of solids, Chem. Eng. 18 (1965)

163–168.

[10] D. Schulze, Zur Fließfahigkeit von Schuttgutern—Definition und

Meßverfahren, Chem. Eng. Technol. 67 (1995) 60–68.

[11] W.J. Ullrich, Powder flow measurement techniques. What’s new?

Adv. Powder Metall. Part. Mat. 1 (1998) 107–127.

[12] G.E. Amidon, L.L. Ausburger, H.G. Brittain, S.R. Byrn, C.D. Fox,

G.E. Peck, D.E. Wurster, Physical test methods for powder flow

characterization of pharmaceutical materials: a review of methods,

Pharmacop. Forum 25 (3) (1999) 8298–8308.

[13] M.K. Taylor, J. Ginsburg, A.J. Hickey, F. Gheyas, Composite method

to quantify powder flow as a screeningmethod in early tablet or capsule

formulation development, AAPS PharmSciTech 1 (3) (2000) article 18.

[14] M. Poux, P. Fayolle, J. Bertrand, Powder mixing: some practical rules

applied to agitated systems, Powder Technol. 68 (1991) 213–234.

[15] W. Muller, H. Rumpf, Das mischen von pulvern in mischern mit

axialer mischbewegung, Chem. Eng. Technol. 39 (1967) 365–374.

[16] U. Sindel, A. Schweiger, I. Zimmermann, Determination of the opti-

mum mixing time for a mixture of lactose and colloidal silicon diox-

ide, J. Pharm. Sci. 4 (67) (1998) 524–526.

[17] P. Albers, K. Seibold, T. Haas, G. Prescher, W.F. Holderich, SIMS/

XPS study on the deactivation and reactivation of B-MFI catalysts

used in the vapour-phase Beckmann rearrangement, J. Catal. 176 (2)

(1998) 561–568.

[18] D.A. Shirley, Phys. Rev. B5 (1972) 4709.

[19] F. Podczeck, Particle–Particle Adhesion in Pharmaceutical Powder

Handling, Imperial College Press, London, 1998, pp. 111–114.

[20] G. Beamson, D. Briggs, High Resolution XPS of Organic Polymers,

Wiley, Chichester, 1992.

[21] C.D.Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, G.E. Muilenberg,

Handbook of X-ray Photoelectron Spectroscopy, Perkin Elmer, Phys-

ical Electronics Division, Eden Prairie, MN, USA, 1978.

[22] G. Beamson, D. Briggs, High Resolution XPS of Organic Polymers,

Wiley, Chichester, 1992, p. 158.

[23] P.L. Stephenson, W.J. Thiel, The effect of humidity on the production

of ordered mixtures, Powder Technol. 25 (1980) 115–119.

[24] H. Schubert, Haftung zwischen Feststoffteilchen aufgrund von Flus-

sigkeitsbrucken, Chem. Eng. Technol. 46 (1974) 333–334.

[25] A.D. Zimon, Adhesion of Dust and Powder, Plenum, New York,

1969, p. 84.

[26] R.K. Chang, M. Leonzio, M.A. Hussain, Effect of colloidal silicon

dioxide on flowing and tableting properties of an experimental, cross-

linked polyalkylammonium polymer, Pharm. Dev. Technol. 4 (2)

(1999) 285–289.

investigation of compacted hydrophilic and hydrophobic colloidal silicon dioxides as glidants for...

Documents