International Journal of Pharmaceutics, 92 (1993) 81-88 81

0 1993 Elsevier Science Publishers B.V. All rights reserved 0378-5173/93/$06.00

IJP 03040

The use of the Washburn method in determining

the contact angles of lactose powder

Juha Kiesvaara a and Jouko Yliruusi b

a Orion Corporation, Orion-Farmos Pharmaceuticals, R&D, Department of Physical Pharmacy, P.O. Box 65,
SF-02101 Espoo (Finland) and ’ University of Helsinki, Department of Pharmacy, Pharmaceutical Technology Divtiion,
Fabianinkatu 35, SF-001 70 Helsinki (Finland)

(Received 9 June 1992)

(Modified version received 1 September 1992)
(Accepted 10 September 1992)

Key words: Powder contact angle; Lactose; Liquid intrusion method; Liquid intrusion equipment;
Washburn technique

Summary

This paper studies the contact angles of lactose powder determined by the Washburn method based on liquid intrusion into the
powder bed. The liquids used were water, methanol, 1-propanol, 1-butanol, 1-octanol and n-hexane. Liquid intrusion was studied
with saturated solutions. Liquid flow to the powder column was controlled with a technique allowing constant liquid flow with zero
hydrodynamic pressure. Packing of the lactose powder into glass tubes was done manually by tapping in a vertical motion. The
packing of the powder columns was studied with different packing times (3, 6 and 12 min) and different size fractions (> 212 pm,
106-212 pm and < 106 Frn) of lactose. Packing time was not found to be critical with optimal shaped lactose particles and the
shortest packing time, 3 min, was adequate. Liquid intrusion rate diminished with diminishing particle size. The intrusion rates of
the liquids varied significantly. The slope values of intrusion distance squared vs intrusion time were 0.798 (n-hexane), 0.414
(methanol), 0.179 (water), 0.123 (l-propanol), 0.099 (l-butanol) and 0.036 (l-octanol). n-Hexane was found to be the best wetting
liquid. Estimating the n-hexane contact angle to be O”, the following values for the other liquids were obtained: water 61.7”,
methanol 38.7”, propanol 39.2”, butanol 33.4” and octanol 36.2”.

Introduction 1990). Wetting of a solid is required in granula-

tion, pelletization and tablet coating. Wetting is
The wettability of a solid material is important also important in disintegration and dissolution
in pharmaceutical technology when considering of solid dosage forms. The contact angle has
the manufacturing processes and the properties usually been used in characterizing the wettability
of the end product (Nogami et al., 1966; Rowe, of pharmaceuticals (Buckton and Newton, 1986a;
Lippold and Ohm, 1986).
Typically the contact angles have been deter-
Correspondence to (present address): J. Kiesvaara, Orion Cor-
mined from systems having a more or less smooth
poration, Farmos, R&D, P.O. Box 425, SF-20101 Turku, surface by the sessile drop method (Zografi and
Finland. Tam, 1976; Ohm and Lippold, 1985). In practice
82

this requires compression of the powder to a Solutions saturated with lactose were prepared
tablet. These results cannot be used to character- and used in intrusion studies in order to avoid
ize the wettability of the powders as the compres- breaking of the powder column structure due to
sion process can dramatically alter the physica dissolution.
properties of the particles resulting in a different To minimize differences between the powder
type of surface (Buckton and Newton, 1986b; particle surfaces in different measurements, lac-
Kiesvaara and Yliruusi, 1991b). Because powders tose powder was stored in a standard humidity
are the usual raw materials in pharmaceutical atmosphere (20 i 2”C, 53 _t l%RH) for 24 h prior
technology it is important that the analysis result to liquid intrusion studies. The standard atmo-
has an emphasis on powder characteristics. sphere was formed in an exsiccator with saturated
Washburn (1921) introduced the basic theory sodium dichromate solution (Na,Cr,O, *2H,O,
for determining the contact angle of a liquid and J.T. Baker).
a solid from liquid intrusion data. After that
many different methods for assessing the wetta-
bility of a powder have been presented and dis-
cussed in literature (Lerk et al., 1976; Stamm et The morphological characterization of the ma-
al., 1984; Buckton et al. 1986, 1987; Buckton terial was made by scanning electron microscopy
1988, 1990; Buckton and Beezer, 1988; Kaya and (Jeol JSM-840A, Japanese Electron Optical Ltd,
Koishi, 1988). Despite the limitations of Wash- Tokyo, Japan). Crystal properties of lactose were
burn method (Carii and Simioni, 1979; Buckton studied by X-ray powder diffraction (XRPD)
and Newton, 1985, 1986a; Yang and Zografi, (Siemens 500D, Siemens AG, Karlsruhe, Ger-
1986) it is used extensively in the assessment of many). The measuring parameters were: tute
wettability, being low-cost and the measuring sys- voltage 40 kV and CuKtv radiation (1.541 A).
tem easy to assemble. The powder column can be Measuring speed was l”/min. The particle size
placed either into a vertical or a horizontal posi- distribution was determined by the laser diffrac-
tion. In the vertical position liquid intrusion can tion method (Series 2600~ Droplet And Particle
occur either upwards or downwards (Studebaker Sizer, Malvern Instruments, Malvern, U.K.); the
and Snow, 1955; Hansford et al., 19801. specific surface area was measured by the nitro-
In this study we will describe the use of the gen gas adsorption method (BET, single point)
Washburn method in contact angle determina- (Flowsorb II 2300, Micromeritics Instrument Cor-
tions, and generally evaluate the suitability of the poration, Norcross, GA, U.S.A.). The surface
method in pharmaceutical powder technology. tensions of the liquids were determined with a
surface tension balance (Digital Tensiometer
Kriiss KlOT, Kriiss GmbH, Hamburg, Germany)
using the ring method. The kinematic viscosities
Materials and Methods
of the liquids were measured by capihary vis-
cosimetry (Ubbelohde Microviscosimeter, Schott
GerHte GmbH, Germany). The densities of the
The material used was lactose (Ph. Eur.). A liquids were determined by accurately weighing
fraction of 100-300 pm (Orion-Farmos Pharma- 25.0 ml of the liquid. The dispersion and polar
ceuticals) was used in packing and liquid intru- components of the liquids were calculated with
sion studies. Three sieve fractions; greater than the geometric-mean method (Wu, 1973; Ohm and
212 urn, 106-212 pm and smaller than 106 pm, Lippold, 1985) using a computer program de-
were used to study the liquid intrusion rate into scribed elsewhere (Kiesvaara and Yliruusi, 1991a).
the powder columns. The int~sion was deter- The contact angles of the liquids were measured
mined with six liquids. Of the liquids, methanol, on a solid paraffin surface with an experimentally
1-propanol, 1-butanol, 1-octanol and n-hexane assessed surface free energy of 24.3 mN/m
were p.a. grade. The water was triple distilled. (Nurminen et al., 1991).
 83

motion using an amplitude of 10 f 2 mm and a

frequency of 190 k 5 taps per min. Packing times
of 3, 6 and 12 min were used. The number of
repetitions was six.

Powder contact angle measurements

Liquid penetration in the powder bed was
monitored by recording the length of the liquid
front intrusion as a function of time. The basic
Fig. 1. Diagram of the equipment. See text for corresponding equation of the Washburn technique is (Wash-
numbers. burn, 1921):

Washburn equipment yXcosOXrXt

12 =
A diagram of the Washburn equipment is pre- 2X7l
sented in Fig. 1. The equipment has two major
parts: the horizontal sample tube packed with where 1 is length of liquid intrusion, y denotes
powder, and the liquid reservoir with flow con- liquid surface tension (mN/m), 0 is the contact
trol. The Washburn tube (1) made of glass with angle, r represents the pore radius and n is
length of about 200 mm and internal diameter of liquid viscosity.
10 mm was heat siliconized in order to make the Assuming the contact angle of the best wetting
surface hydrophobic. A millimeter scale from 0 to liquid is known, the contact angles of the other
150 mm was engraved on the tube. The other end liquids with respect to the best wetting liquid can
of the tube was closed with a glass sinter (2), pore be calculated with the following modification of
size 160-200 pm (No. 0). A silicone rubber tube Eqn 1 (Studebaker and Snow, 1955):
(3) with a diameter of 5 mm connected the Wash-
burn tube to the liquid reservoir (4). The glass
reservoir was closed with a rubber stopper (5) (2)
penetrated by a glass capillary tube (6). The lower
end of the glass capillary tube was adjusted to the
where k is the slope of the 1* vs t linear intru-
level of the upper side of the powder column. As
sion curve. Subscripts 1 and 2 refer to the best
the liquid flows from the reservoir, the glass tube
wetting liquid and to the liquid under examina-
supplies air to compensate for the volume of the
tion, respectively.
used liquid. This arrangement reduces the hydro-
static pressure to zero at the level of the lower
end of the glass capillary tube. Thus the liquid
flow into the powder bed is free and controlled Results and Discussion
only by the capillary forces in the powder bed.
A valve (7) was added to the tube to allow for The morphological characterization of lactose
the rapid formation of the equilibrium hydro- powder made from SEM micrographs showed
static pressure right from the beginning of the that the particles are nodular corresponding to
liquid intrusion. A laboratory stand (8) was used the typical shape of lactose. According to XRPD
to hold the Washburn tube. The liquid reservoir analysis the material is well crystallized. The spe-
was placed on an adjustable laboratory tray (9) to cific surface area of lactose powder (fraction
facilitate the vertical alignment. 100-300 pm> is 1.38 m’/g. The relative volume
size distribution of the material is presented in
Packing of the columns Fig. 2. The median of the distribution is 220 ,um.
The packing of lactose powder into the glass The distribution is log-normal by shape, being
tubes was made by manually tapping in a vertical optimal for this type of experiment, also repre-
84

Particle volume (%I TABLE 1

Contact angles (0) (n = 10) of the liquids on solid paraffin
80 surface with measured surface tensions (yL) (20 f 0.2”C) and
dispersion (y,“) and polar (y[) components
60
Liquid 0 (“I

40 Mean SE

Water 110.4 0.8 72.8 23.1 49.7

20. Methanol 33.5 1.4 22.8 18.0 4.8
Propanol 29.1 0.7 23.9 20.6 3.3
Butanol 27.6 1.1 24.6 22.1 2.5
10 100 1000 10000 Octanol 26.7 1.2 28.5 (30.0) (-1.5)
Particle elm (urn) Hexane 0 0 18.7 (14.4) (4.3)
Fig. 2. Particle size distribution of lactose powder.
Parentheses indicate an inconsistent result.

senting the form of distribution often encoun-

tered in pharmaceutical powders. The size distri- 3 exhibit more irregular shapes due to sieving.
butions of the lactose fractions presented in Fig. The median values differ significantly, being 355,
220 and 50 pm for size fractions greater than
212, 106-212 and less than 106 pm, respectively.
Particle volume (%)
The contact angles on solid paraffin surface,
i
surface tensions and calculated dispersion and
polar components of the liquids are presented in
Table 1. The values are very close to those given
in literature (Weast, 1982; Ohm and Lippold,
1985). Water has the highest polar component.
Starting with methanol, the polar component de-
creases as the length of the hydrocarbon chain
increases.
The surface tension, the cosine of the contact
100 ,000
angle and the surface free energy value of the
Particle size (urn)
Fig. 3. Particle size distributions of lactose powder sieve
solid paraffin surface include an error so that the
fractions. > 212 pm (dot), 106-212 pm (solid), < 106 pm mathematical treatment yields inconsistent re-
(dash). sults. This manifests as a negative polar compo-

TABLE 2
The densities and dynamic viscosities at 24 f 0.2”C, and contact angles (0) (n = 10) on solid paraffin surface, the surface tensions (ye)
(20 f 0.2”C) and dispersion (yf) and polar fyi) components of the liquids saturated with lactose

D P
Liquid Density Dynamic 0 (“I
k/cm3) viscosity TkN/m) ?kN/m) TkN/m)
Mean SE
(mN s rn?

Water 1.09 1.96 107.6 0.7 58.2 17.0 41.2

Methanol 0.78 0.55 39.9 0.8 22.8 16.7 6.1
Propanol 0.80 1.86 26.8 1.2 23.9 21.1 2.9
Butanol 0.81 2.66 28.6 1.0 24.6 22.0 2.6
Octanol 0.82 7.86 30.4 0.9 27.3 26.6 0.7
Hexane 0.66 0.29 0 0 18.7 (14.4) (4.3)

Parentheses indicate an inconsistent result.

 85

Polar component (mN/m) component of both pure and lactose saturated

IDOL
liquids and the relative permittivity (Weast, 1982)
(Fig. 4). Liquids with non-existent polarity, oc-
tanol and hexane, are not presented in the graph
(Tables 1 and 2) because of the logarithmic scale.
In rank correlation level the relationship is well
understood as the relative portion of the elec-
tronegative hydroxyl group determines the polar-
ity of the liquid.
The densities and dynamic viscosities of the
liquids used in contact angle calculations are pre-
1 sented in Table 2. The density values are slightly
0 20 40 60 80
Relative permittivity
higher compared to the pure liquid values due to
dissolved lactose (Weast, 1982). The effect of
Fig. 4. Polar component vs relative permittivity of both pure
and lactose saturated water, methanol, propanol and butanol. dissolved lactose on the viscosity of the liquids is
significant, with water for example, the viscosity is
twice as great as with pure water.
nent with octanol, hexane has a polar component Table 2 also lists the contact angles on solid
although the contact angle on the paraffin sur- paraffin, the surface tensions and the calculated
face is zero degrees. It can be assumed, though, dispersion and polar components of the liquids
that as the contact angle of hexane is zero de- saturated with lactose. Only slight differences,
grees, the polar component is also negligible. The apart from the significant decrease of the surface
inconsistent values are given in parentheses in tension and the components of water, are ob-
Tables 1 and 2. It is evident that a similar devia- served when comparing with the values in Table
tion is present also in the values calculated for 1. This is due to the great solubility of lactose in
the other liquids. water and smaller solubility in the other liquids.
It is interesting to observe a linear correlation The effect of particle size of the powder on
between the logarithm of the measured polar liquid intrusion is presented in Fig. 5. The intru-

Time (seconds)
Fig. 5. The slope of distance squared vs time of different size fractions of lactose; > 212 pm (a), 106-212 Km (b) and
< 106 pm Cc).
86

sion into the two greatest fractions is clearly TABLE 3

faster than into the smallest fraction. This can be Linear regression parameters of water intrusion into lactose
attributed to a small particle size and a continu- powder bed for different packing times
ous particle size distribution in the lower end
Packing k (cm’s_‘1 b (cm’) r
enabling a more effective degree of packing. The
time
fraction of particles less than 10 pm in diameter Mean SE Mean SE
(min)
amounts to up to six percent according to Fig. 3.
3 0.179 0.005 4.62 1.51 0.999
Because the liquid properties remain the same,
6 0.187 0.005 -2.24 1.41 0.998
the mathematical treatment results in different 12 0.184 0.008 3.53 1.92 0.999
contact angles for the size fractions; the greater
the intrusion rate the smaller the contact angle. k, slope; b, intercept; r, regression coefficient; n = 12. The
number of data points in regression analysis is 11.
The beginning of the liquid intrusion did not
generally yield a straight line when l2 was plotted
against time. The packing method generates an
impulse that mostly affects near the sinter in the sion coefficients for the curves are by and large
bottom of the tube causing heterogeneous and equally good and thus all packing times studied
loose packing of the particles. Therefore, a steady can be considered acceptable. The packing times
flow of the liquid is reached only after some can be classified by the slope values, ,however,
distance. The data used in calculations is ob- and thus it seems necessary to determine the
tained from the linear part of the penetration optimal packing time for individual powder sys-
curve. tems.
The data in Table 3 show that with different A typical example of intrusion data for contact
packing times the resulting powder column ex- angle determination is presented in Fig. 6. The
hibits a similar structure. The shape of the parti- graph contains all data points with least squares
cles in terms of packing properties is close to fits from six parallel studies. Although the points
optimal. The most compact powder bed and of each parallel study deviate from each other to
smallest capillaries were, surprisingly enough, ob- some extent, the numerical slope values used in
tained with the shortest packing time. As a result, calculations are relatively near one another. The
a packing time of three minutes was used in the parameters of the least squares fit of liquid intru-
final contact angle determinations. The regres- sion data and the calculated contact angle values

240

220

200

_ 160

*$ 160

- 140
N, 120

100

60

60

Time (seconds)
Fig. 6. Methanol intrusion into lactose (100-300 pm) powder bed; data of six parallel studies with linear least-squares fits.
 87

TABLE 4 native excipients, for instance, for a drug formu-

Linear regression parameters of liquid intrusion into lactose lation, where the interface should remain stable
powder bed and the calculated dynamic contact angles of the for the entire shelf-life of the drug product.
liquids
The water contact angles previously measured
on lactose tablets were about lo” (Kiesvaara and
Liquid k (cm* s-‘) b (cm*) r 0C) n
Yliruusi, 1991b). Although the results from a
Mean SE Mean SE Mean SE
static and a dynamic method are not directly
Water 0.179 0.005 -1.31 1.51 0.999 61.7 0.7 12 comparable, the difference in the contact angle
Methanol 0.414 0.010 0.02 0.97 0.995 38.7 1.5 6
values implies a difference between the powder
Propanol 0.123 0.005 0.49 0.77 0.994 39.2 2.4 6
Butanol 0.099 0.003 -1.43 1.08 0.999 33.4 2.1 6 particle and tablet surfaces.
Octanol 0.036 0.001 -7.26 3.02 0.995 36.2 2.4 6 Technically the Washburn method for deter-
Hexane 0.798 0.020 - 11.2 2.09 0.999 0 0 6 mining the relative rate of wettability and contact
The number of data points in regression analysis is 11. Hex-
angles of a powder is applicable. The use of a
ane is the best wetting liquid with estimated contact angle of horizontal powder column simplifies the mathe-
0”. k, slope; b, intercept; r, regression coefficient. matical treatment by making the use of the grav-
ity term in the Washburn equation unnecessary.
The intrusion rate constant k is a good mea-
are presented in Table 4. All the liquids studied sure of powder wettability because it includes
behaved quite well during the intrusion; the ad- liquid viscosity. All the liquids studied can be
vancement of the liquid front was easily moni- ranked by this parameter. The contact angle is
tored. The difference in penetration rate with the already more difficult to determine, requiring
fastest and the slowest liquid is noticeable. It knowledge of the best wetting liquid and its con-
took octanol 90 min to advance 14 cm, while tact angle.
hexane needed only 4 min. The y-ariance in intru- It is evident that the results are very much
sion rate is greatest with the fastest liquids. The dependent on both the powder properties; pack-
good correlation coefficients indicate steady in- ing and contact angle, and liquid viscosity and
trusion of all of the liquids. surface tension.
Water has clearly the highest dynamic contact The repeatability of the packing technique
angle value with lactose powder (Table 4). The must be confirmed as constant structure of inter-
alcohols all have approximately equal contact an- particle space of the powder bed is determining
gles about two thirds of the water contact angle. factor in obtaining valid results. This means that
Although there is no clear trend among the alco- when comparing results from previous batches,
hols, it can be observed that the higher the liquid for instance, the particle morphology and size
polarity, the higher is the contact angle value. distribution must be taken into account. Pro-
This results from a greater dissolution into a cesses like milling affect the wettability by alter-
polar liquid increasing the viscosity of the liquid. ing the morphology, size distribution and the
Thus the resulting numerical contact angle is surface properties. The method also yields differ-
greater. ent contact angles for the different size fractions
Water has the poorest wettability judged by of a powder because the intrusion rate is differ-
the contact angle, but it still wets lactose faster ent. The liquid intrusion method can thus be
than propanol, for instance, although the differ- used to measure and quantify changes caused by
ence between the contact angles is 13”. This is a processes.
most important advantage of the dynamic liquid
intrusion method compared with the static sessile
drop method, that does not reveal the rate of Acknowledgement
wetting. The rate of wetting is important in study-
ing of manufacturing processes, whereas the static The authors would like to thank the University
contact angle can be utilized when ranking alter- Pharmacy for contributing financially to this work.
88

References Kiesvaara, J. and Yliruusi, J., The effect of compression force

and compression pressure on the surface free energy of
Buckton, G. and Newton, J.M., Assessment of the wettability tablets. Actu Phurm. Nord., 3 (1991b) 171-177.
and surface energy of a pharmaceutical powder by liquid Lerk, CF., Schoonen, A.J.M. and Fell, J.T., Contact angles
penetration. J. Pharm. Pharmacol., 37 (1985) 605-609. and wetting of pharmaceutical powders. J. Phurm. Sci., 65
Buckton, G. and Newton, J.M., Liquid penetration as a (1976) 843-847.
method of assessing the wettability and surface energy of Lippold, B.C. and Ohm, A., Correlation between wettability
pharmaceutical powders. J. Pharm. Pharmacol., 38 (1986a) and dissolution rate of pharmaceutical powders. Int. J.
329-334. Phurm., 28 (1986) 67-74.
Buckton, G. and Newton, J.M., The significance of contact Nogami, H., Nagai, T. and Uchida, H., Studies on powdered
angles measured on surfaces which have undergone plastic preparations. XIV: Wetting of powder bed and disintegra-
deformation. In Gorrod, J.W., Gibson, G.G. and Mitchard, tion time of tablet. Chem. Phurm. Bull., 14 (1966) 152-158.
M. (Eds), Deu. Drugs Mod. Med., Ellis Horwood, Nurminen, K., Kiesvaara, J. and Yliruusi, J., The surface free
Southampton, 1986b, pp. 421-424. energy of solid paraffin. Actu Phurm. Fenn., 100 (1991)
Buckton, G., Beezer, A.E. and Newton, J.M., A vacuum 160.
microbalance technique for studies on the wettability of Ohm, A. and Lippold, B.C., Surface tension of solid drugs
powders. J. Phurm. Phurrnucol., 38 (1986) 713-720. and it’s correlation with solubility. Int. J. Phurm. Tech.
Buckton, G., Dissado, L.A., Hill, R.M. and Newton, J.M., The Prod. Mfr., 6 (1985) l-6.
potential value of dielectric response measurements in the Rowe, R.C., Correlation between predicted binder spreading
assessment of the wettability of powders. Inr. J. Phurm., 38 coefficients and measured granule and tablet properties in
(1987) l-7. the granulation of paracetamol. Int. J. Pharm., 58 (1990)
Buckton, G. and Beezer, A.E., A microcalorimetric study of 209-213.
powder surface energetics. Int. J. Phurm., 41 (1988) 139- Stamm, A., Gissinger, D. and Boymond, C., Quantitative
145. evaluation of the wettability of powders. Drug Deu. Ind.
Buckton, G., The assessment, and pharmaceutical impor- Pharm., 10 (1984) 381-408.
tance, of the solid/liquid and the solid/vapour interface: Studebaker, M.L. and Snow, C.W., The influence of ultimate
a review with respect to powders. Int. J. Phurm., 44 (1988) composition upon the wettability of carbon blacks. J. Phys.
l-8. Chem., 59 (1955) 973-976.
Buckton, G., Contact angle, adsorption and wettability - a Washburn, E.W., The dynamics of capillary flow. Phys. Reu., 8
review with respect to powders. Powder Technol., 61 (1990) (1921) 273-283.
237-249. Weast, R.C., Handbook of Chemistry and Physics, 63rd Edn.,
Carli, F. and Simioni, L., Limitations of the Washburn equa- CRC Press, FL, 1982.
tion in quantifying penetration rates. J. Pharm. Phurmucol., Wu, S., Polar and nonpolar interactions in adhesion. J. Adhe-
31 (1979) 128. sion, 5, 39-55 (1973).
Hansford, D.T., Grant, D.J.W. and Newton, J.M., The Influ- Yang, Y.W. and Zografi, G., Use of the Washburn-Rideal
ence of processing variables on the wetting properties of a equation for studying capillary flow in porous media. J.
hydrophobic powder. Powder Technol., 26 (1980) 119-126. Pharm. Sci., 75, 719-721 (1986).
Kaya, N. and Koishi, M., The wetting properties of powders Zografi, G. and Tam, S.S., Wettability of pharmaceutical
and devices for measuring them. Konu, 6 (1988) 86-97. solids: Estimates of solid surface polarity, J. Phurm. Sci.,
Kiesvaara, J. and Yliruusi, J., A Turbo Pascal program for 65 (1976) 1145-1149.
determinations of surface free energy. Lab. Microcomput.,
10 (1991a) 14-18.