J. Pharm. Pharrnacol. 1986, 38: 489493 01986 J. Pharm. Pharmacol.

Received November 20, 1985

A study of powder adhesion to metal surfaces during

compression of effervescent pharmaceutical tablets
F. E. J . SENDALL* A N D J . N . STANlFORTHt

School of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath BA2 7AY, UK

Effervescent tablets were produced using four different formulations containing citric
and/or tartaric acid and sodium bicarbonate with povidone and macrogol 6000. The same
formulations were prepared with the addition of 1% sucrose ester powder. The adhesion of
each formulation to the metal faces of tableting machine punch tips was determined using
electron microscopy, surface roughness measurements and quantification of punch weight
variations during tablet production. The basic formulations were inherently adhesive and
produced tablets with a weak, porous structure which were ualitatively and quantitatively
1
rougher than conventional, non-effervescent compresse tablets. Both formulations
containing tartaric acid produced tablets with a lower surface roughness and with less
tendency to stick to tablet punch faces than the two formulations containing citric acid
alone. The addition of a water-soluble sucrose ester had a beneficial effect especially on the
formulations with inherently high adhesive tendencies.
The commercial production of effervescent phar- compounds. Such components may have low intrin-
maceutical tablets requires careful selection of for- sic anti-adherent properties when their moisture
mulation components and processes to obtain tablets content is low but they are hygroscopic and may take
of the required specification (Aumonier 1970; Mou- up moisture to produce intrinsically adhesive gran-
lin 1972; Aiache 1974, 1975; Faguet 1976; Faguet et ules. This situation can be exacerbated during tablet
a1 1977; Boymond 1979; Lachman & Lieberman compression, with the result that material adheres to
1980). Formulations which provide optimal disinte- the compression tool faces and there is ‘picking out’
gration or dissolution times, or produce pleasant of large areas of the compact surface on completion
tasting solutions for oral administration, frequently of a compression cycle.
have poor compression properties. Specialized equipment and formulations can effect
As most effervescent tablets are formulated to large changes in the appearance of effervescent
react chemically with water, it is necessary that the tablets (Sendall et al 1983). To find the effectiveness
residual moisture content of the granules for com- of these changes requires some quantitative evalua-
pression is minimal (usually below 0.5% w/w). As a tion of tablet surface properties coupled with a
result the bonds formed between and within the qualitative assessment. In the present study, the
granules are weak when compared with conventional effect of varying the formulation of an effervescent
pharmaceutical tablets. granulation, and the incorporation of a novel anti-
Official compendia usually require effervescent adherent was examined by different techniques.
tablets to disintegrate completely within 1 or 2 min. Surface roughness profiles were obtained for 25 mm
This is achieved by using a tablet with a relatively diameter flat-faced tablets and low-resolution con-
large diameter, resulting in an increased specific tours were produced by digital processing of the
surface area compared with most pharmaceutical profiles. Arithmetic mean roughness (R,) values
tablets, making a larger surface available to the were also obtained and used to characterize surface
reacting liquid. The combination of large diameter, quality (Rowe 1979).
thin compacts prepared from weakly bound granules The surfaces of tablets of all the formulations were
of low moisture content frequently results in porous, examined by scanning electron micrography and
brittle compacts that can only be formed at relatively quantitative measurement of adhesion to compres-
high compression pressures. To achieve a clear sion tool faces in a single-punch machine was
solution from the formulation, all the components monitored.
must either be soluble or react to produce soluble
M A T E R I A L S A N D METHODS
Effervescent tablet formulations
* COX Pharmaceuticals, Barnstaple, Devon, UK.
Present address: Maresty Machines Ltd., Liverpool, UK. Granulations were prepared from mixtures of
t Correspondence. organic acids and sodium bicarbonate with povidone
490 F. E. J . SENDALL AND J. N . STANIFORTH

(Table 1) in a small planetary mixer. After mixing, Preliminary experiments showed that large varja-
the materials were granulated with anhydrous indus- tions in R, values could be expected because of the
trial methylated spirits. The wet granules were then nature of the surface, the short traverse length and
tray dried at 45 "C, milled through a 2.5 mm aperture the large maximum and minimum values of the
screen using a small mill (Apex Equipment Ltd, UK) height of the surfaces being examined. The appa-
and then blended with Macrogol 6000 passed ratus was first used in the R, mode with a cut-off of
through a 0.5mm screen in an 8litre cube mixer 0.8 mm and the standard stylus of 5 ym radius giving
(type KB20 Erweka, GmbH, FRG). an R, determination over a distance of 4.5 mm.
Mean R, values were calculated making 30 deter-
Table 1. Composition of effervescent formulations used in minations on each formulation by traversing the
the study. pick-up across three dimensions of each of ten
tablets. The tablets were mounted in a Perspex
Formulation (YOw/w) holder to eliminate movement. Then profiles of
Ingredient 1 2 3 4 surface roughness were obtained with the Surtronic 3
Sodium bicarbonate 55.5 61.1 50.4 51.7 in traverse mode. In this mode the pick-up traverses
Citric acid 42.3 36.7 31.6 - 25 mm at 0.25 mm s-1 the result being an output
Tartaric acid - - 15.8 46.1 voltage proportional to the vertical displacement of
Povidone 0.5 0.5 0.5 0.5
Macrogol6OOO 1.7 1.7 1.7 1.7 the stylus (1 V representing a displacement of
100 ym) which was recorded by an XY chart recorder
operating at a chart speed of 2 mm s-1. The ratio of
For the studies on punch tip adhesion, 1 part of chart and traverse speeds provided a compromise
food grade sucrose ester (TAL 160, Contract Chem- between good resolution of the stylus heights with
icals, Liverpool, UK) was dry blended with 99 parts the necessity of obtaining a trace length of approxi-
of tablet granulation in a cube mixer (Erweka mately 250 mm.
GmbH, Frankfurt, FRG). The lubricated granula- To obtain a complete picture of the tablet surface,
tions were compressed using a Manesty type RD3 the diametral traverse recording was followed by a
rotary press (Manesty Machines, Liverpool, UK) 2mm shift of the tablet under the stylus, perpen-
equipped with 25 mm diameter flat-punches with dicular to the traverse direction. Thirteen parallel
bevelled edges, or a single punch machine (Manesty chords about 2 mm apart were therefore traversed
type E2) equipped with 12.5 mm diameter flat faced and recorded for a 25 mm tablet of each formulation
punches. sampled after about 100 compression cycles.
Before compression, punch faces were cleaned The analogue traces obtained were then digitized
using industrial methylated spirits, tablets were by a microcomputer (BBC model B, Acorn Compu-
sampled after 100-300 compression cycles. ters Cambridge, UK) connected to a PL Graphics
system digitizer (B. S. Dollamore Ltd, Burton-on-
Electron microscopy Trent, UK). The 'peak and trough' positions of the
Samples of each formulation compressed into 25 mm trace were thus translated into digital values and
diameter tablets were prepared for electron micro- stored, 250-300 points representing a typical dia-
scopy (Type 35C, JEOL, Japan). metral traverse. The resolution of the profile was
such that ten changes in surface height could be
Surface roughness studies identified per millimetre of tablet surface. A com-
Surface rugosity was studied using a Surtronic 3 puter program was subsequently developed to clas-
display/traverse unit (Rank Taylor Hobson Ltd, sify the surface heights into bands of 20 ym enabling
Leicester, UK) which interprets data obtained from the percentage of points at any height to be
a pick-up arm that traverses a stylus across the determined (see Fig. 1).
surface to be characterized, in terms of the arith-
metic mean roughness (R,). The unprocessed infor- Adhesion studies
mation can also be passed to a microcomputer. The Adhesion of material to the tips of punches was
apparatus is designed to characterize smooth (in monitored by compression of each formulation,
powder technology terms) metal or ceramic surfaces prepared with and without antiadherent at a force of
and great care was required in setting it to obtain about 15 kN on the single punch tablet machine
reproducible results with the rough surfaces encoun- equipped with 12.5 mm diameter flat-faced pun-
tered in the present study. ches. The punches were cleaned as described earlier
 POWDER ADHESION TO METAL SURFACES DURING COMPRESSION OF EFFERVESCENT TABLETS 491
Formulation 4 which contains tartaric acid and
sodium bicarbonate, gave tablets with over 75% of
the surface within a 40 pm height range, while at least
40% of the surface of other tablets prepared with
formulations was outside a 60 pm height range
(Table 3). The formulations containing citric acid
(1,2,3) had similar surface height ranges indicating
that the percentage composition does not affect
2 5 rnrn surface roughness as much as the inclusion or
exclusion of specific components.
FIG. 1. Typical diametral scan of a 25 mm tablet after
digitization and reconstruction by microcomputer. The
data from the scans is summarized in Table 3. The height
range within which most points occur is seen to be Table 3. Percentage of tablet surface above and below a
narrowest for those formulations which exhibit the lower reference line for four 25 mm diameter effervescent tablets
R, values. of different formulation.

Formulations
Height of
profile/pm 1 2 3 4
and the upper punch weighed before loading the
hopper. After completing a pre-selected number of <-lo0 13.383 2.753 11.973 5.272
- 100 7.495 1.733 4.367 1.620
compression cycles, the upper punch was removed -80 11.242 4.333 7.907 2.763
and weighed to within 0.1 mg. The data obtained -60 16667 7.086 11.596 5.240
-40 16.488 11.547 13.743 28.072
were tabulated and displayed graphically. -20 13.954 13.765 14.006 32.899
0 9.172 13.994 13.47.9 16.577
20 5.639 11.445 9.450 5.049
40 2.819 9.534 6.212 1.302
60 1.392 7.443 4.104 0.349
RESULTS 80 0.750 5,149 1.996 0.127
Average surface roughness (R,) 100 O@OO 0.280 0.OOO O.Oo0
>lo0 0.999 10.936 1.167 0.730
Table 2 shows the mean of the average surface
roughness (R,) values determined. The R, values for
formulations without anti-adherent are much greater
than those with anti-adherent, and those formu- For each formulation, irrespective of the degree of
lations containing tartaric acid (3,4) have a lower R, roughness, there were proportionally more surface
than those containing citric acid alone. The mean R, cavities than asperities, both in overall terms and at
values are all less than 10 pm while maximal peak individual profile contours (Table 3). This is con-
values lie between ten and twenty times the R, value sistent with a 'picking out' mechanism being re-
(Table 3). It thus appears that R, values reflect the sponsible for major alterations in surface profiles
overall surface quality, values of 3 pm or less such as are found in effervescent tablets.
representing the surfaces of conventional tablets,
while values of 5 pm or greater indicate a surface of Adhesion studies
poor appearance and large variations in surface Table 4 and Fig. 2 show that the addition of anti-
height over the tablet surface (see note, Table 2). adherent to the formulations tested after 100 com-

Table 2. Surface roughness (R,) values for effervescent Table 4. Increase in weight in mg of 12.5 mm up er punch
formulations with and without anti-adherent (mean f against number of com ression cycles for four efkrvescent
s.d.). tablet formulations witgout and with (*) anti-adherent.

Formulation number Number of cycles

1 2 3 4 Formula. 0 1 5 10 20 40 60 80 100
R, values (rm) 1 0 0 0.7 1.6 2.8 7.0 9.1 9.7 7.0
Without anti- 1' 0
adherent
With 1% w/wanti-
*
6.99? 1.63 5.35 0.83 4.50 f 1.39 4.04 0 4 4 * 2
2'
0 2.4 3.8 2.1 2.1 1.9 1.1 1.1

adherent * *
3.49 0.89 2.99 ? 0.27 2.46 0.39 2.69 ? 0.52 3
3' 0 0.2 0.2 0.5 0.5 0.3 0.7 1.0
1.9
8.1
0.7
4 0 0.1 0.7 1.7 1.9 2.0 2.1 1.9 2.1
Non-effervescent tablet produced from starch-lactose granules under 4' 0 0.3 0.3 0.5 0.3 0.1 0.2 0.7 1.1
similar conditions: mean 3.13 f 0.62.
492 F. E J . SENDALL AND J . N . STANIFORTH

10.0

-8.0
vF
z
P6.0
5
.c
U

5Q L.0
L
bl
Q
32 0

0 100 200 300

No. of c y c l e s
FIG.2. (A) Plot showing change in punch weight against
number of corn ression cycles for formulation 3 without
(A)and with my anti-adherent.
(B) Plot showingchange in
punch wei ht over 100 compression cycles for formulation 1 FIG.3. Scanning electron photomicrogra hs showing sur-
without (8)
and with (0)anti-adherent. face detail of tablets compressed from (As Formula 1; (B)
Formula 2; (C) Formula 3; (D) Formula 4.

pression cycles reduced the adhesion of material to open structure than formulation 4. In general, the
punch faces to less than 2 mg in all cases. The largest micrographs correlated well in observed surface
improvement was effected in formulation 1, while quality with the observed R, values and the rough-
there was minimal improvement with formulation 4 ness profiles obtained.
which was the least adhesive initially. This formula-
tion also had the lowest R, value. Therefore it can be CONCLUSIONS
inferred that increased surface roughness is caused Effervescent pharmaceutical tablet formulations are
by the removal of material from the surface and that inherently adhesive and the tablets produced have a
the adhesive forces between the punch face and the weak, porous structure. Tablet surfaces have a much
granulation are greater than the bonding forces higher average roughness than conventional tablets
within the compressed tablet. The rate at which and the surface height may change by over 100 pm
material is added to the punch face is also within a 100 pm distance.
formulation-dependent, and is most rapid with the Formulations containing tartaric acid produced
two formulations containing citric acid alone. tablets having a lower surface roughness, and less
The compression forces appear to cause material adherence to tablet tools than those containing citric
to leave the tool surface once a film of material has acid.
bonded to the metal. This caused the punch weight to R, values may be used to quantify tablet surface
vary in a cyclical fashion over periods of 20-60 quality. Great care must be taken to establish test
compression cycles, the compression and adhesive conditions, and correlate observed quality with
forces acting in opposition as compression conti- numerical R, determinations. Low values (<3pm)
nued. indicate surfaces approaching the quality of conven-
Scanning electron micrographs of tablet surfaces tional (non-effervescent) pharmaceutical tablets.
from formulations 1-4 showed highly porous, open The water soluble sucrose ester tested at 1% wiw
structures with little intergranular bonding. At low significantly decreased adhesion to tablet punch
magnification the surface of tablets with low R, faces of the more adhesive formulations. Decreasing
values (formulation 4) had a fine, uniform structure; the adhesive properties of the granulation should
at higher magnification (Fig. 3) the surfaces of decrease significantly the surface roughness of the
tablets containing citric acid (formulations 1-3) were tablet, and consequently improve the quality of the
observed t o have many more cavities and a more finished product.
 POWDER ADHESION TO METAL SURFACES DURING COMPRESSION O F EFFERVESCENT TABLETS 493
REFERENCES Faguet, J . P., Puisieux, F., Duchene,
Pharma-Problemes et Techniques 25:
Aiache, J . M. (1974) Pharm. Acta Helv. 49: 169-178
Lachman, L., Lieberman, H. A. (eds) (1980) in: Phar-
Aiache, J. M. (1975) in: Proceedings of Direct Compres- maceutical Dosage Forms: Tablets, Vol. 1 . Marcel
sion Symposium. Kingsley and Keith, London Dekker Inc.. New York DD 225-258
L 1

Aumonier, P. (1970) Actualites Pharmaceutiques 54: 4-7 Moulin, J. (1972) Sciences et Techniques Pharmaceutiques,
Boymond, C. (1979) Labo-Pharma-Problemes et Tech- 1: 325-329
niaues 25: 987-995 Rowe. R. C. (19791 J. Pharm. Pharmacol. 31: 473-474
Fa uLt, J . P. (1976) DoctoralThesis, Universite Paris-Sud, Sendall, F. E.'J., Staniforth, J . N., Rees, J. E., Leatham,
.
Ehatenay-Malabry Paris M. J . (1983) Pharm. J . March 12: 289-294