Homeopathy (2014) 103, 143e146

Ó 2013 The Faculty of Homeopathy

http://dx.doi.org/10.1016/j.homp.2013.11.005, available online at http://www.sciencedirect.com

ORIGINAL PAPER

The mathematics of dilution

Barun Kumar Chatterjee*

Department of Physics, Bose Institute, 93/1 A.P.C. Road, Kolkata 700009, India

The major objection to homeopathic medicine is that the doses of medicine prescribed in
some cases are too dilute for any active ingredient to be present. The medicines would
hence be rendered inactive, necessitating novel explanations for the action. A further ex-
amination of dilution in the light of the Langmuir equation shows that homeopathic
medicines may not be as dilute as a simplistic application of Avogadro’s Principle sug-
gests, due to surface effects. Homeopathy (2014) 103, 143e146.

Keywords: Homeopathy; Ultramolecular dilution; Langmuir equation; Surface

effects

Introduction monolayer of molecules is adsorbed in dynamical equilib-

Homeopathic medicines are in use in large parts of the
world and evidence for their effects is well proven in-vitro.1
The major criticism to homeopathic medicine is that it em-
ploys dilutions like 60X (60D or 30C) where the concentra-
tion of the active ingredient is 10
60 times the original
concentration. If one assumes the presence of 1 M of the
active ingredient in the mother tincture then the probability
that one molecule of the active ingredient will be present in
a liter of 30C solution would be 6  10
37, which is fantas-
tically dilute. Arnica 30C, the alcoholic extract of Arnica
Montana (containing helenalin amongst other compounds
as an active ingredient) diluted 1060 times is used for its
anti-inflammatory and anti-tumor properties. Assuming
that mother tincture contained 262.3 g of helenalin in a liter
of alcohol, the probability to find one molecule of helenalin
in 1 ml of Arnica 30C would be 6  10
40.
Various attempts have been made to reconcile the action
of the drugs at such super-Avogadro dilutions.2e12
Recently however, experimental evidences12,13 show that
more than the expected amount of active materials are
present at extreme dilutions in solutions.
Here, a minimal mathematical model of dilutions has
been presented, where the role of interfaces on dilution
has been taken into account. Avogadro’s law is not ques-
tioned, simply the role of surfaces on dilution is stressed.
The Langmuir adsorption model14 is invoked, where a
*Correspondence: BK Chatterjee, Department of Physics, Bose
Institute, 93/1 A.P.C. Road, Kolkata 700009, India.
E-mail: barun_k_chatterjee@yahoo.com
Received 10 May 2013; revised 6 November 2013; accepted 28
November 2013
rium with those in solution. The equilibrium between the
adsorbed state and the dissolved state results in an initial
linear rise in the adsorbed number with the concentration
of the dissolved molecule, later reaching a saturation.
The model is not specific to the type of the molecule of
the active ingredient or the solvent. The goal of this model
is to help reconcile some recent observations of unexpected
higher concentrations at high dilutions.
Dilutions
When a solution is recursively diluted, it involves the
mixing of v liter of the solution with concentration cn
1
moles/liter and mixing V liter of the solvent to give about
v + V liter of the solution with a concentration cn. This pro-
cess assumes the additivity of volumes, which is strictly
valid for ideal solutions, is roughly valid for small concen-
trations cn
1 of solute.
The molecules of the active ingredient would not only be
dispersed through the bulk of the solvent (both homoge-
neously and/or by nanostructures) but would also be present
on the free surface (solventevapor interface) and the sol-
ventesolid interface. Molecules of the active ingredient
are physiadsorbed onto the surface in equilibrium.14 The
iterative-dilution could proceed in the following two
ways: (a) the mixing container is not reused (Hahnemannian
method of dilution), where a v amount of solution is poured
into V volume of solvent, succussed and from it a volume v is
poured into the next container of V volume of solvent and
this process is repeated, or (b) the mixing container is reused
(Korsakovian method of dilution), and in it Vamount of sol-
vent is added to an initial volume v of the solution, succussed
and the a volume V is poured off, and this process is
 Mathematics of dilution
BK Chatterjee
144
repeated. Equilibrium is assumed at all stages. The two pro- which gives
cesses are treated in the following two subsections.
cn ¼ c0 mn : (8)
(a) Situation where the mixing container is not reused
(Hahnemannian method of dilution) If f < Ql0 /(Ql + K(a + b)) then m > f and the dilution is
slower than perceived. If one uses v = 1 cm3, V = 9 cm3,
In this process, where the glassware is changed at each
a = 50.8 cm2, A = 4.5 cm2, K = Q = 0.07 (value of Q calcu-
dilutional step, assuming that the surface density of the
lated from the data by Chang et al.15 for sodium dodecyl sul-
active ingredient for the solventesolid interface is sn and
fate (SDS) in 0.5 M NaCl solution; it is assumed that K = Q),
that on the free surface it is tn, we have,
a0 = 50.3 cm2, a00 = 0.02 cm2, a = 5.1 cm
1, b = 0.45 cm
1,
cn ðv þ VÞ þ t n a0 þ sn ða þ AÞ ¼ cn
1 v þ t n
1 ða0 þ a00 Þ: l = 5.027 cm
1, l0 = 5.028 cm
1 one gets m = 0.26. This
(1) would imply that a 30C dilution is effectively a 17C dilution.
(b) Situation where the mixing container is reused (Kor-
where, a is the original liquidesolid interface area, A is the sakovian method of dilution)
liquidesolid interface area after the solvent is added, a0 is the
solution-vapor surface area when at rest and a00 is the in- In this process, where a single piece of glassware is used
crease in the solution-vapor surface area when being poured. throughout the process of dilution, assuming that the sur-
Ignoring s and t one gets the usual exponential decay of face density of the active ingredient for the solventesolid
concentration as interface is sn and that on the free surface it is tn, we have,
cn ¼ c0 f n (2)
cn ðv þ VÞ þ t n a0 þ sn ða þ AÞ
which for f = 0.1 gives cn = c010
n as expected. ¼ cn
1 v þ t n
1 ða0 þ a00 Þ þ sn
1 a (9)
One simple model for s and t would be to assume that
the molecules on the surface are in dynamic equilibrium where, a is the original liquidesolid interface area, A is the
with those in the bulk of the solvent. When the solute liquidesolid interface area after the solvent is added, a0 is the
molecules adsorbed on the surface are in dynamic equi- solution-vapor surface area when at rest and a00 is the in-
librium with those in solution, the adsorbed number den- crease in the solution-vapor surface area when being poured.
sity (s) of solute molecules and the dissolved This equation can be rewritten as
concentration (c) are related by the Langmuir adsorption
isotherm14 v t n
1 ða0 þ a00 Þ t n a0 a
cn ¼ cn
1 þ
þ sn
1
vþV vþV vþV vþV
smax bc ða þ AÞ
s¼ (3)
sn
1 þ bc vþV
which is dependent on the temperature T. At very low con- (10)
which can be written as
centrations (which is true for the dilutions discussed here)
the Langmuir equation can be reduced to s = smaxbc/ cn ¼ cn
1 f þ t n
1 l0
t n l þ sn
1 a
sn ða þ bÞ (11)
(1 + bc) z Kc, where b is a function of temperature. The
equilibrium constants can hence be written as where

sn a A a0
¼ KðTÞ (4) a¼ ; b¼ ; l¼ ;
cn vþV vþV vþV
and a0 þ a00 v
l0 ¼ .l and f ¼ \1:
tn vþV vþV
¼ QðTÞ (5)
cn
One may rewrite Eq. (11) as
where K and Q are equilibrium constants, which are func-
X
n
1
tions of temperature. Eq. (1) can hence be rewritten as cn ¼ c0 f n þ t 0 l0 f n
1 þ ðl0
f lÞ t k f n
1
k
t n l
k¼1
cn þ Qcn l þ Kcn ða þ bÞ ¼ cn
1 f þ Qcn
1 l0 : (6)
where X
n
1
þ s0 af n
1 þ ða
ða þ bÞf Þ sk f n
1
k
0
a A a k¼1
a¼ ; b¼ ; l¼ ;
vþV vþV vþV
sn ða þ bÞ (12)
a0 þ a00 v
l0 ¼ .l and f ¼ \1: which can be solved, if the values of s and t are known.
vþV vþV
From the Langmuir adsorption isotherm at high dilutions
This enables one to write as in the previous case, one gets


f þ Ql0 cn ðv þ VÞ þ Qcn a0 þ Kcn ða þ AÞ
cn ¼ cn
1 ¼ cn
1 m (7)
1 þ Ql þ Kða þ bÞ ¼ cn
1 v þ Qcn
1 ða0 þ a00 Þ þ Kcn
1 a (13)

Homeopathy
 Mathematics of dilution
BK Chatterjee
145
which gives which would have similar mathematical ramifications).

 The main thrust of this work is to show that the effect of
f þ Ql0 þ Ka the surface on dilution would may result in concentrations
cn ¼ cn
1 ¼ cn
1 g (14)
1 þ Ql þ Kða þ bÞ much larger than what is naively perceived. One may also
which gives note here that the equilibrium constant K would depend on
cn ¼ c0 g n : (15) the mixing container (whether it is glass, stainless steel,
quartz) as different materials and different active ingredi-
If f < (Ql0 + Ka)/(Ql + K(a + b)) then g > f and the dilu- ents and solvents would have different values of K and
tion is slower than perceived. If one uses v = 1 cm3, Q. For multicomponent mixtures, unless the equilibrium
V = 9 cm3, a = 50.8 cm2, A = 4.5 cm2, K = Q = 0.07 constants are comparable, the ratio of the mole fraction
(from the data by Chang et al.15 as discussed in previous of the components would keep changing with dilutions.
section), a0 = 50.3 cm2, a00 = 0.02 cm2, a = 5.1 cm
1, The material of the containers also plays a role in the gen-
b = 0.45 cm
1, l = 5.027 cm
1, l0 = 5.028 cm
1 one eration of nanoparticles.16
gets g = 0.46. This would imply that a 30C dilution is effec- A proper and accurate measurement of concentration of
tively a 9C dilution. active ingredient is needed before considering a certain
dilution beyond the Avogadro limit. The function of suc-
cussion, its role in establishing equilibrium or in the forma-
Dilutions of multicomponent mixtures tion of nanostructures,13,17,18 and conformational changes,
For multicomponent mixtures, the effect of all surfaces if any, of macromolecules at large dilutions need to be
would be to change the composition of the mixtures unless understood separately.
the equilibrium constants of the individual components are
equal. If the concentration of the Ith component in the nth
dilution is cIn and the IIth component in the nth dilution is
cIIn . It can be seen, that for the second mixing scheme


References
f þ QI l0 þ KI a
cIn ¼ cIn
1 ¼ cIn
1 gI : (16) 1 Endler PC, Thieves K, Reich C, Matthiesen P, Bonamin L, Scherr C,
1 þ QI l þ KI ða þ bÞ et al. Repetitions of fundamental research models for homeopathi-
and
cally prepared dilutions beyond 10e23. Homeopathy 2010; 99:


f þ QII l0 þ KII a 25e36.
cn ¼ cn
1
II II
¼ cIIn
1 gII ; (17) 2 Benveniste J. Human basophil degranulation triggered by very
1 þ QII l þ KII ða þ bÞ dilute antiserum against IgE. Nature 1988; 333: 816e818.
3 Maddox J, Randi J, Stewart WW. “High-dilution” experiments a
where the equilibrium constants are sIn =cIn ¼ KI ðTÞ, delusion. Nature 1988; 334: 287e290. Benveniste J. Dr Jacques
t In =cIn ¼ QI ðTÞ, sIIn =cIIn ¼ KII ðTÞ, and t IIn =cIIn ¼ QII ðTÞ. Benveniste replies. Nature 1988; 334: 291.
This gives, 4 Thomas Y. The history of the memory of water. Homeopathy 2007;

n 96: 151e157.
cIn cI0 mI 5 Beauvais F. Memory of water and blinding. Homeopathy 2008; 97:
¼ II (18)
cIIn c0 mII 41e42.
6 Montagnier L, Aissa J, Ferris S, Montagnier J-L, Lavallee C. Elec-
and for the Korsakovian scheme, it is tromagnetic signals are produced by aqueous nanostructures
derived from bacterial DNA sequences. Interdiscip Sci Comput

n Life Sci 2009; 1: 81e90.
cIn cI0 gI
¼ II : (19) 7 Elia V, Baiano S, Duro I, Napoli E, Niccoli M, Nonatelli L. Perma-
cIIn c0 gII nent physico-chemical properties of extremely diluted aqueous so-
lutions of homeopathic medicines. Homeopathy 2004; 93:
Here we see that unless KI = KII and QI = QII, the ratio of 144e150.
the mole fractions of the components will keep varying 8 Anick DJ, Ives JA. The silica hypothesis for homeopathy: physical
with successive dilutions. chemistry. Homeopathy 2007; 96: 189e195.
Assuming KI = QI = 0.07 cm and KII = QII = 0.026 cm 9 Private communication with Dr. Parimal Banerjee.
(along with the geometry parameters as in the 10 Anick DJ. The octave potencies convention: a mathematical model
previous two sections) one gets mI/mII = 0.26/ of dilution and succussion. Homeopathy 2007; 96: 202e208.
11 Chikramane PS, Kalita D, Suresh AK, Kane SG, Bellare JR. Why
0.18 = 1.44 and gI/gII = 0.46/0.28 = 1.63 and hence extreme dilutions reach non-zero asymptotes: a nanoparticulate hy-
the ratio of the concentrations would evolve as given pothesis based on froth flotation. Langmuir 2012; 28:
by Eqs. (18) and (19). 15864e15875.
12 Gopal B. Some studies on homeopathic medicines. Masters thesis.
Indian Institute of Technology (Kanpur), 1992.
Discussion 13 Chikramane PS, Suresh AK, Bellare JR, Kane SG. Extreme homeo-
pathic dilutions retain starting materials: a nanoparticulate perspec-
The effect of surface should be considered when calcu- tive. Homeopathy 2010; 99: 231e242.
lating dilution. Here two schemes of dilutions are shown 14 Langmuir I. The constitution and fundamental properties of
and more schemes could be envisaged (e.g. using the solids and liquids. Part I. Solids. J Am Chem Soc 1916; 38:
same lid for succussion or draining in place of pouring 2221e2295.

Homeopathy
 Mathematics of dilution
BK Chatterjee
146
15 Chang CH, Franses EI. Modified Langmuir-Hinshelwood kinetics
for dynamic adsorption of surfactants at the air/water interface. Col-
loids Surf 1992; 69: 189e201.
16 Ives JA, Moffett JR, Arun P, Lam D, Todorov TI, Brothers AB, et al.
Enzyme stabilization by glass-derived silicates in glass-exposed
aqueous solutions. Homeopathy 2010; 99: 15e24.
17 Anagnostatos GS, Vithoulkas G, Garzonis P, Tavouxoglou C. A
working hypothesis for homoeopathic microdiluted remedies.
Berl J Res Homoeopath 1991; 1: 141e147.
18 Lyell D. An NMR study of the effect of succussion on parallel
potencies of Natrum muriaticum. Masters thesis. Department of
Homoeopathy at Durban Institute of Technology, 2004.

Homeopathy