Pethig Kell pmb87 PDF

Phys. Med. Biol., 1987, Vol. 32, No 8, 933-970.
Printed in the UK
Review article
The passive electrical properties of biological systems:

their significance in physiology, biophysics and biotechnology
Ronald Pethig'r and Douglas B KellS
t Institute of Molecular and Biomolecular Electronics, University College of North Wales,
Dean Street, Bangor, Gwynedd LL57 IUT, U K
3 Department of Botany and Microbiology, University College of Wales, Aberystwyth,
Dyfed SY23 3DA, UK
This review was completed in February 1987
Contents
1. Introduction and scope 933
2. A summary of dielectric theory 935
3. Amino acids, peptides, proteins and DNA 940
3.1. Aminoacids 940
3.2. Polypeptidesandproteins 94 1
3.3. Deoxyribonucleicacid ( D N A ) 946
4. Bound water in biological systems 947
5. Biological electrolytes 949
6. Membranes and cells 952
7. Tissues 958
7.1. Skin 96 1
7.2. Othertissues,includingtumours 962
1. Introduction and scope

There is much current interest in the electrical properties of biological materials, not
least because of an increasing awareness of the possible physiological effects elicited
by the absorption by tissues of non-ionising electromagnetic radiation (e.g. Presman
1970, Sheppard and Eisenbud 1977, Adey 1981, Illinger 1981, Konig et a1 1981, Becker
and Marino 1982, Frohlich and Kremer 1983, Adey and Lawrence 1984, Chiabrera er
a1 1985, Marino and Ray 1986, Polk and Postow 1986). Studies of the ways in which
electromagnetic energy interacts with tissues are also of importance to the continuing
development of the use of R F and microwave hyperthermia (Hahn 1982, Storm 1983),
in impedance plethysmography (Hill1979, Mohapatra 1981, Wheeler and Penney1982,
Brown 1983, Anderson 1984, Penney 1986) and pneumography (Henderson and Web-
ster 1978, Baker 1979), in electrical impedance tomography (Pryce 1979, Barber and
Brown 1984), in the gentle thawing of cryogenically preserved tissue (Burdette et a1
1980), in the use of pulsed electromagnetic fields to aid tissue and bone regeneration
and healing (Barker and Lunt 1983, Pilla et a1 1983) and in a variety of existing or
projected biosensing devices (Kell 1986a). The use of dielectric measurements as a
tool for studying molecular and cellular parameters is a well established technique of
continuing importance (e.g. Grant era1 1978, Pethig 1979, 1984).
0031-9155/87/080933+38$02.50 0 1987
Publishing
IOP Ltd 933
934 R Pethig and D B Kell
Historically, measurements of the electrical properties of biological materials hold

a pre-eminent position in several areas of physiology and biophysics. Thus, Hobner
(1910) measured the electrical impedance of suspensions of erythrocytes up to frequen-
cies of 10 MHz and, finding that their impedance decreased with increasing frequency,
concluded that the cells were composed of a poorly conducting membrane surrounding
a cytoplasm of relatively low resistivity. Amongst the first indications of the ultra-thin,
molecularthicknessofthismembranewerethoseprovidedbyFricke(1925),who
obtained, from a theoretical and experimental analysis of the dielectric properties of
redblood cell suspensions, a value of 0.81 pFcm” for the erythrocyte membrane
capacitance. By assuming a value of three for the relative permittivity of the membrane,
Fricke derived a value for its thickness of 3.3 nm, a value within a factor of two of
thosecurrentlyaccepted(see,e.g.,Cole 1972, SchanneandCeretti1978, Kell and
Harris1985bandseelater).Similarly,theobservationofaninductance(negative
capacitance) by ColeandBaker(1941)duringmeasurementsofthe AC‘ electrical
properties of squid axons led directly to the conceptof voltage-gated membrane pores,
as embodied in the celebrated Hodgkin-Huxley (1952) treatment (Cole 1972, Jack er
a1 1975), as the crucial mechanism of neurotransmission.
Quantitative details of the molecular size, shape and extentof hydration of protein
molecules were provided by the dielectric measurements of Oncley (1943), and these
were extended in the laboratories of Hasted (Haggis et a1 1951), Grant (1965) and
Schwan(1965a)toprovideimportantdetails of thephysicalnatureandextentof
protein hydration. As examples of how electrical and dielectric studies continue to
provide new and unique knowledge in the biophysical and physiological sciences, we
may mention measurements of the diffusional motions of proteins and lipids in cell
membranes by Harris and Kell (Kell 1983, Kell and Harris 1985a, b, Harris and Kell
1985),studies of themechanism of field-induced cell fusionand cell rotation in
Zimmermann’s laboratory (Zimmermann 1982, Arnoldet a1 1985) andby others (Fuhr
er a1 1985,1986),studiesoftheinfluenceofhydrationonenzymeactivityand
intramolecularmobility(BoneandPethig1985,FinneyandPoole 1984, Neumann
1986) and on protonic and ionic charge transport processes in proteins (Careri et a1
1985, Morgan and Pethig 1986).
Themajorpurposeofthisreviewthen is to document those factors which are
believedmoststronglytoinfluencethewayinwhichnon-ionisingelectromagnetic
( E M ) radiationinthefrequencyrange1Hzto100GHzinteractswithbiological
materials. Now, since E M radiation is composed of both electrical and magnetic fields,
we need in principle to consider not only the electrical properties such as permittivity
andconductivitybutalsothemagneticpermeability of thesubstancesinvolved.
However, apart from ferromagnetic materials such as magnetite, which is known to
be associated with the orientational and navigational abilities of certain bacteria, insects
and mammals (Gould 1984), most biological materials have a magnetic permeability
equaltothatofavacuum.Equally,whilstmagneticfieldsarealsoabletointer-
actwithcertainnuclearmagneticmomentsorwithunpairedelectronsassociated
with paramagnetic ions or free radicals, such interactions are highly specific and in
generalhavelittleinfluenceonthebulkelectricalproperties.Thus we shallbe
primarily concerned with the (complex) dielectric properties of biological materials,
namely with the frequency-dependent magnitudes of the conductivity (T and relative
permittivity E ‘ .
We do not here attempt togive a full historical account of the development of this
topic, an overview of which may be found elsewhere (Schwan 1957, Cole 1972, Grant
Passive
electrical
properties of biological
systems 93 5
et a1 1978, Schanne and Ceretti 1978, Pethig 1979). In addition, we shall be concerned
mainlywiththeso-calledlinearpassiveelectricalproperties,i.e.thosewhichare
independent of the magnitude of the fields used to assess them; relevant aspects of
the non-linear electrical properties of cells and tissues in the stated frequency range
may be derived (mainly for neuromembranes) from Cole (1972), Jack et a1 (1975),
Almers (1978) and Schanne and Ceretti (1978). Similarly, we shall not devote much
space to methodological aspects, discussions of which may be found, for instance, in
the review by Schwan (1963) and the books by Grant et a1 (1978) and Pethig (1979).
Finally, tabulations of the dielectric properties of various tissues and biomaterials have
beengivenbyGeddesandBaker(1967),SchwanandFoster(1980),Stuchlyand
Stuchly(1980)andPethig(1984).Thepresentreview is designedtoextendthese
surveys and to provide an overview of the more persuasive theories which have been
formulated in an effort to understand the observable electrical properties in terms of
the underlying biophysical processes.
2. A summary of dielectric theory

The passive electrical properties of a material held between two plane-parallel elec-
trodes of area A separated by a distanced are completely characterised by the measured
electrical capacitance C (units Farads) and conductance G (units Ohm" or Siemens),
as defined in the following two equations:
G = Auld C = Aee,,/d. (1)
The conductivity U is the proportionality factor between the electric current density
and the electric field, and is a measure of the ease with which 'delocalised' charge
carriers can move through the material under the influence of the field. For aqueous
biological materials, the conductivity arises mainly from the mobility of hydrated ions
and other processes to be described later. The factor eo is the dielectric permittivity
of free space, and has the value 8.854 X lo-" F m", whilst e is the permittivity of the
material relative to that of free space. F is (erroneously, but for historical reasons)
sometimes referred to as the dielectric constant. The permittivity is proportional to
the ratio of the charge to the electric fields, and reflects the extent to which 'localised'
charge distributions can be distorted or polarised under the influence of the field. For
biological materials, such charges are mainly associated with electrical double layers
occurring at membrane surfaces or around solvated macromolecules, or with polar
molecules which (by definition) possess a permanent electric dipole moment.
Figure 1 shows somewhat diagrammatic examples of electrical double layers at the
surface of a membrane and around a globular protein, and the constitution of an
electricdipole.Thesimplestmoleculardipoleconsists of apair of opposite,unit
electric charges (of magnitude +q and - 4 ) separated by a vector distance S ; in this
case,themoleculardipolemoment m (whichisanintensiveproperty) is givenas
m = qs and has units of C m. For a complex protein molecule, such as that illustrated
in figure 1, positive and negative charges arise from the presence of ionisable acidic
and basic amino acid sidechains (see later) in the protein structure, and due to the
large size of protein molecules these will give rise to a comparatively large permanent
dipole moment whose value will varywith
pH, molecularconformation and
intramolecular mobility.
Each type of polar or polarisable entity will exhibit its own characteristic response
to an imposed electric field; to describe this, we write the relative permittivity as a
-S-
IC1
Figure 1. An illustration of theelectricaldoublelayersformedatthesurface of achargedbiological
membrane ( a ) and around a charged, aqueous globular protein( b ) . ( c ) A simple polar molecule, consisting
in this case of a pair of opposite unit charges +q and - 9 separated by a distance S and possessing a dipole
moment of m = qs C m.
complex function of the form (Debye 1929)

E*(w)=E,+(E,-E~)/(~+~wT) (2)
in which E , is the permittivity measured at a frequency sufficiently high that the polar
orpolarisableentity is unabletorespondtotheelectric field, E , is thelimiting
low-frequency permittivity (sometimes called 'static' permittivity) where the polarisa-
tion is fully manifest, W is the angular frequency of the (sinusoidal) electrical field (in
rad S"), i is (-l)"2 and T is the characteristic response time or relaxation time. The
real and imaginary partsof the complex permittivity may also be expressed in the form
E* = &"iE't (3)
in which the real part E', corresponding to the permittivity described in equation ( l ) ,
is given by
E'(W)=E,+(E,-Ee,)/(l+w2T2). (4)
The imaginary componentE " , known as the dielectric
loss, corresponds to the dissipative
loss associated with the movement of polarisable charges in phase with the electric
field and is given by
&"=(&,-E,)(W7.)/(1+W2T2). (5)
It takes the form of a peak (loss peak) such as that shown for water in figure 2 . The
loss factor E" may also be defined in terms of a frequency-dependent conductivity as
E"=(T(W)/OEg=((+O+(+d(W))/WEO (6)
where go is the steady state conductivity arising (predominantly) from mobile ions
and u d ( w ) is the frequency-dependent conductivity arising from dielectric polarisation.
Passive electrical properties of biological systems 937
U'
>
W
il
601
2o t
80-
60-
-
B 431
: 0 20 0.1 1 Log f (GHzl

10 100
Figure 2. The dielectric dispersion exhibited by pure water at 20°C, illustrated in terms of the change i n
( a ) the real ( E ' ) and imaginary ( E " ; dielectric loss) parts of the permittivity and( b )the frequency dependence
of the Conductivity. The low-frequency conductivity at neutral pH, due to the presence of H + and OH-
ions, has a value of some 5 pS m".
By defining the magnitude of a dielectric dispersion (figure 2 ( a ) ) as
A&)=E:-&Ei,
we obtain, by combination with equations (4) and ( 5 ) ,
and
in which fc is the relaxation frequency or 'characteristic frequency' ( f c= 1 / 2 m ) . The

factor U , is the low-frequency limit of the conductivity that includes the steady state
(DC) conductivity and dielectric losses associated with polarisation processes having
relaxation frequencies significantly lower than that defined by fc above. By putting
f > > f c ,a conductivity increment ( A a ) is obtained (figure 2 ( b ) ) given by
AU = U= - U , = ~ T ~ ~ E O A B ' . (10)
This shows that, as a dielectric dispersion (with a single relaxation time) is traversed
by changingthefrequency of measurement,thechange in conductivity is directly
proportional to the change in permittivity. This follows from the fact that the total
energyin the field is constant (for a given voltage) and musteitherbestored(as
reflected in & ' ( W ) ) or dissipated (as reflected in & " ( W ) ) by the system with which it
interacts. This allows an additional check upon the validity of the experimental data.
Equation (10) may therefore be written in the form

r = Ae’eo/hu.
Equations (10) and (1 1) hold strictly only for a process with a single relaxation time;
however, they are obeyed reasonablywell provided that any spread of relaxation times
is not too large.
It is useful to note that by using frequency as the parameter a ‘complex permittivity’
plot of e ” against F ’ may be obtained; this has the form of a semi-circle whose centre
lies on the abscissa and which intersects the e ‘ axis at the points EL a n d e,:, and such
a plot may be used to derive values forE & and E : when, for technical or other reasons,
the frequency range over which one can measure is restricted in some way. This is
known as a Cole-Cole circle and it is named after the brothers K S and R H Cole
who first derived it (Cole and Cole 1941). A similar plot, the complex conductivity
plot (Grant 1958), may be obtained by plotting U ” ( = w e O ( e ’ - e:)) against W ’ ,and
usedtoderivethevalues of W, and w ~ When . thesystemexhibitsonlyasingle
relaxation time, the frequency at which e “ a n d W ” take their maximum values is equal
to the characteristic frequency.
For real systems, and especially in biological work, it is commonly observed that
the centre of the circular locus traced by the data in these complex plane diagrams
lies below the abscissa. This is conventionally interpreted (but cf Jonscher 1983) in
terms of the existence of an ensembleof processes possessing a significant distribution
of relaxation times and contributing independently to the macroscopically observable
dispersion of interest, although the appearanceof these plots as arcs of a circle is only
barely influenced by the exact type of relaxation time distribution function invoked
(Schwan 1957). Nonetheless, a particular analysis of the foregoing (Cole and Cole
1941) is widely used in dielectric studies of biological materials since, though empirical,
it allows an economy in the expression of the degree to which a particular relaxation
exhibits heterogeneity in its relaxation times. We therefore briefly discuss the salient
feature of the Cole-Cole analysis.
What Cole and Cole (1941) realised was that a simple empirical modification of
the Debye equation could be derived which had the property that, whilst a plot of e ”
against e’ would give a locus that traced the ofarc a circle, the semi-circle so extrapolated
would have a centre that both lay below the abscissa and which would make an angle
a ~ / rad2 with the points ( E ; ) or ( e & ) ,thus still permitting the estimation of F : and
E:. This Cole-Cole a appears in the modified Debye equation as follows:

F*(w)=E,+(F,-~~)/[~+(~wT)”~]. (12)
Whilst this Cole-Cole a is an entirely phenomenological variable (it lacks a proper
molecular or physical basis), and Boyd (1980) surveys many other modifications of
the Debye equation which may also be used to fit or to express dielectric data, the
Cole-Cole formalism remains overwhelmingly the most common means of describing
datafromexperimentsonthepassiveelectricalpropertiesofbiologicalsystems.
Furthermore, whilst the actual distribution of relaxation times embodied in the Cole-
Cole formula is rather complex (e.g. Cole and Cole 1941, Hasted 1973), a variety of
plausiblerelaxationtimedistributions(e.g.Gaussian,rectangular)givesbehaviour
that is forpracticalpurposesindistinguishablefrom‘true’Cole-Colebehaviour
(Schwan1957). It is alsotruethatthesum of twoDebyeprocessescannotbe
distinguished from a small distribution of relaxation times (Sheppard and Grant 1974).
We may therefore characterise any dielectric dispersion observed in biological materials
in terms of three parameters only: the dielectric increment A E ' , the relaxation time ( r )
or characteristic frequency f,, and the Cole-Cole a.
Asimplephysicalmodel,whichonemayusetogainamentalpictureofthe
processes occurring during the dielectric relaxation of a molecule with a permanent
dipole moment, is one that considers the dipoles to besphereswhoserotation in
resplonse to the field is opposed by frictional interaction with the surrounding viscous
medium. The relevant relaxation time for the orientation of such a sphere is given by
r =x/2kT (13)
where x is amolecularfrictioncoefficientwhichrelatesthetorqueexerted on the
dipolar molecule by the applied electric field to the molecule's angular velocity, k is
Boltzmann's constant and T is the absolute temperature. Assuming the dipole to be
equivalent to a rigid sphere of radius a turning in a Newtonian hydrodynamic fluid
of macroscopic viscosity 0, the Stokes-Einstein relation gives x = 8 d a 3 so that the
relaxation time is
T =4 d a 3 / k T . (14)
Considering that the model described is virtually as simple as one may conceive,
equation (14) often gives remarkably good results, even for the case of water molecules
rotating in bulk water. In bulk water, the distance between adjacent oxygen molecules
is 0.28 nm, i.e. a =0.14nm. At293 K,theviscosityofwater is kg m" S" (i.e.
1 mPa S), so that the value for the relaxation time of water at this temperature that
one would derive from equation (14) is 8.5 PS, in excellent agreement with the experi-
mentally observed value of 9.3 PS. However, it should be added that this relationship
between viscosity and dielectric relaxation can also be accounted for using a model
of water basedon directed hydrogen bonds (Grant 1957). The relaxation time of 9.3 PS
is equivalent to a characteristic frequency of some 17 GHz, and the frequency-depen-
dent values of E ' and E" for bulk water between 100 MHz and 100 GHz are given in
figure 2(a). The corresponding increase in U for 'pure' water as this frequency range
is scanned is from about 5 X S m-' to some 70 S m-', and is displayed in figure
2 ( b ) . Thelow-frequencyconductivityoftheelectrolytesolutionstypicalofhigher
euk.aryotes (usually taken as equivalent to that of a 150 mM NaCl solution) is approxi-
mately 2 S m", so that above 2 GHz o r so their conductivities are dominated by the
water present (at a concentration approaching 55.5 M), and thus exhibit approximately
the same frequency dependence as that shown in figure 2 ( b ) .
Following Kirkwood (1932, 1939), we may relate the dielectric increment(a macro-
scopic observable) to the dipole momentm and the molecular weightM (both intensive
properties) of the polar molecule according to the relationship
As'= N ~ g m ~ / 2 & ~ M k T (15)
where N is Avogadro's number, c the concentration in kg m-3 of the polar molecule
in the solvent and g a parameter introduced by Kirkwood to account for molecular
associationsandcorrelationeffectsbetweenthemotionsofsoluteandsolvent
molecules. Such effects are particularly noteworthy for hydrogen-bonded liquids, where
the rotation of one molecule seriously disrupts the local hydrogen bonding and requires
the correlated or cooperative reorientation of as many as four neighbouring water
molecules to compensate for this (Hasted1973, Hallenga et ai 1980). For four-bonded
water at room temperatures, g has a value of about 2.8, dropping to 2.5 at 100 "C. For
zero-bonded water molecules the value of g is, of course, unity. Evidently, a value
for g of unity implies no intermolecular associations, or that the net value of such
associations cancels out, as for a protein molecule in an aqueous environment where
there will be numerous solute-solvent forces acting over a wide range of directions
(South and Grant 1972). We shall see shortly that g values for a-amino acids are of
the order of 1.2.
3. Amino acids, peptides, proteins and DNA
3.1. Amino acids
a-amino acids have the general formula R-C,(H)(NH:)-COO-, and thus exist in
two forms of opposite optical activity due to the fact that, unless the sidechain R = H
as in glycine, C, constitutes a chiral centre. These sidechains, of which twenty are
particularly widespread in nature, determine in large measure the dielectric properties
of proteins and polypeptides. Given the presence of Goth amino and carboxyl groups,
the degree of ionisation of a-amino acids is obviously strongly pH dependent; the
doubly charged form shown above is known as a zwitterion and is predominant at
neutral pH. The zwitterionic nature of amino acids has the consequence that their
solvation by water is accompanied by a large negative change in volume, resulting
from the strong electrostatic interaction between the polar water molecules and the
two charged groups. Similarly, since the zwitterion represents a large dipole, neutral
solutions of amino acids (which may have a negligibleDC conductivity) exhibit a high
permittivityandabsorbinfra-redradiationat1580 cm", anabsorptionband(in
wavenumbers) characteristic of the carboxylate ion, and not at 1720cm" as would be
the case for an uncharged "COOH group.
As mentioned above, the simplest a-amino acid, and the only common one to lack
a chiral centre, is glycine, in which R = H. The distance between the centres of the
positive ammonium group and the negative carboxyl group should be about 0.32 nm,
so that the effective dipole moment should have a value given by
m = q s = ( 1 . 6 x 1 0 ~ ' 9 ) x ( 3 . 2 x 1 0 " 0 ) = 5 . 1 x l O ~ 2 9 C m = 1 5 . 3 D e b yunits
e (D) (16)
(1 D = 3.33 X 10"' C m). (The displacement of one electronic charge through 10"' m
gives a dipole moment of 4.8 D.) Thisvalue of the dipole moment calculated for
glycine solutions compares reasonably with that of 20 D obtained by Wyman (1934)
fromdielectricmeasurements.DunningandShutt(1938)showedfurtherthatthe
permittivity of glycine solutions is constant from pH 4.5 to pH 7.5 but falls sharply on
either side of these pH values. The interpretation of this is that at the extremesof p H
glycine possesses a single net charge only, so that the more dipolar, zwitterionic form
disappears in acid or alkaline solutions.
The dipole moment per unit volume of a zwitterionic a-amino acid is larger than
that of water, so that an amino acid solution exhibits a greater low-frequencyor static
permittivity than does water, as is illustrated for glycine in figure 3.At room tem-
perature, the characteristic frequency of the dielectric dispersion due to the rotation
of glycine is 3.3 GHz (Grant et al 1978), so that the dispersion overlaps with that due
towater(figure3).Actually,because of theelectrostaticinteractionsbetweenthe
amino acid and the water molecules, the simple dipole model used in the derivation
of equation (14) (and which would predict f c = 12.56 GHz for glycine) is invalid.
Qualitatively, it is not surprising that the experimental finding is that the rotation of
glycine is slower than that calculated on the basis that there no aremolecular interactions
between the solute and solvent save those frictional forces to be expected for a solid
Passive electrical properties of biological systems 94 1
1 10 100
log frequency IGHzI
Figure 3. The dielectric dispersion exhibited by a 1 M aqueous solution of glycine in its zwitterionic form.
It may be observed that the low-frequency permittivity exceeds that of pure water (figure2 ) , whilst, as more
evident in the loss peaks, the dispersion of the glycine overlaps that of the water molecules.
sphererotating in anisotropicfluid.Quantitatively, no straightforwardtheoretical

treatment of the dielectric behaviour of amino acid solutions exists. Thus, even in the
case of the simplest amino acid, we lack a precise, quantitative molecular interpretation
of the observable dielectric behaviour. We shall have further cause to ponder such
difficulties when we consider the dielectric properties of proteins and more complex
ensembles of biological molecules.
It is possible to express the dielectric behaviour of amino acids in aqueous solution
in the form
&I=&jSSc (17)
where E ' a n d E ; are the permittivities of the solution and pure solvent, respectively. c
is the molar concentration of the solute and S is a constant which is related to the
total dielectric increment A s ' by the expression
CS =A&" As,. (18)
In this equation, A&= is the high-frequency decrement and is equal to the amount that
the static permittivity of pure water is lowered by the presence of the solute (Grant et
a1 1978, p 175, Young et a1 1982). S therefore quantifies the increase in polarisation
due to the presence of the amino acid of interest. Bearing in mind that pure water
has a molarity of 55.5, it is perhaps not surprising that the measured permittivity of
amino acid solutions varies more or less linearly with the concentration of solute (i.e.
that S is constant),evenforwhatare,fromabiologicalstandpoint,ratherhigh
concentrations of amino acid. For aqueous solutions of a-amino acids at 25 "C, 6 has
a valueofsome 26-28 permittivityunits M" forfrequenciesupto1 GHz and
concentrations up to 2.5 M. If we take the value for the dipole moment of glycine
(15.3 D) as being typical, a value for the Kirkwood g of approximately 1.2 may be
obtained from equation (15) for amino-acid solutions.
3.2. Polyveptides and proteins

Thewayinwhichpolypeptidechainsarebuiltup is shown in figure 4 ( a ) . These
peptide bonds take the form of an amide linkage resulting from the condensation of
thea-aminoanda-carboxylgroupsofadjacentaminoacids. As thepolypeptide
increases in length, the distance between the terminal amino and carboxyl charges also
(c l
Figure 4. The construction of an a-helical structure in a protein. ( a ) The formation of a peptide bond, via
the condensation of the a-amino and a-carboxyl groups of a-amino acids. ( 6 ) The magnitude and direction
of the dipole moment possessed by an unmodified peptide bond. ( c ) The arrangementof the dipole moments
of individual peptide bonds in the classical a-helix structure of a protein.
tends to increase, and thus the effective dipole moment and observed dielectric incre-
ment. For polyglycine chains up to the heptapeptide, the valueS increases
of according
to the relation
S = 1 4 . 5 1-
~ 5.87 i19)
where n is the number of chemical bonds between the terminal amino and carboxyl
groups (Pethig 1979, p 83).
Protein molecules are composed of one or more polypeptide chains folded in a
complex, fractal geometry (Creighton 1983). The static and dynamic three-dimensional
structure of proteins is controlled at different levels of organisation by non-covalent
interactions between the peptide bonds, amino acid sidechains and solvent molecules,
to form regions of a-helix or P-pleated sheet (Salemme 1983), for example (Kabsch
and Sander 1983). The N-C bond in the peptide units has a partial double-bond
character, so thatthesixatoms C,,NHCOC,, arecoplanar.Inaddition,the C=O
bond is itself polar, so that the peptide bond possesses a permanent dipole moment.
A relatively straightforward quantum mechanical calculation (Pethig 1979, pp 44-9)
gives the magnitude of this dipole moment as approximately 3.6 D, directed at an
angle of 46.7" to the C-N bond axis, as illustrated in figure 4(b). Since each peptide
unit possesses a permanent dipole moment, polypeptide chains take the form of strings
of connected dipoles.
Passive electrical
properties of biological systems 943
The contribution that an individual dipoleof moment m makes to the polarisability

of a medium is proportional to m'. For a completely rigid linear polymer chain of n
regularly spaced dipoles of unit moment m, fixed and directed normally to the chain,
the overall contribution to the polarisability will lie between zero and approximately
nm2, depending upon whether the dipoles are additive or cancel vectorially. I f the
dipoles are perfectly free to rotate, the contribution will be nm'. For most proteins,
n is a large number, typically 100-1000, so that the contribution of the dipoles of the
peptide units to the overall dipole moment of the protein molecule, and hence to the
permittivity of an aqueous protein solution, is extremely sensitive even to the 'static'
polypeptide configuration.
A good example (Wada 1976) in which the peptide group moments are purely
additive is the extended a-helix configuration depicted in figure 4 ( c ) . The axis of a
rigid polypeptide a-helix is directed at approximately 56" to the C-N bond of the
constituent peptide units and, since the peptide dipole moment is itself directed at an
angle of 46.7" to the C-N bond, the peptide contribution to the total moment parallel
to the axis of the helix will be
cos(9.3")
m = 3.6n
(20) D. = 3.6n
The 'classical' a-helix configuration (Pauling et al 1951) contains 3.65 peptide units
per helix turn, so that even the relatively small helix of not quite three turns shown
in figure 4 ( c ) will have a dipole moment of some 34 D. Such a helix dipole moment
is equivalent to there being half a positive electronic charge at the N terminus and
half a negative electronic charge at the C terminus. However, dipole-dipole interactions
between neighbouring a-helical and other segments of the main polypeptide backbone
influencetheoverall'static' or 'average'polypeptideconformation(Sheridan et a1
1982) and vice versa (Blundell et a1 1983). Indeed, the subtle interplay between short-
andlong-rangefunctionallinkagesbetweendifferentpartsofaproteinmolecule
represents a primary focus of many studies of the effects of site-directed mutagenesis
on protein structure and function (Ackers and Smith 1985), and opens up a number
of interesting possibilities for the improvement of our interpretation of the dielectric
behaviour of protein molecules.
We saw above that, even in the case of glycine, uncertain degrees of interaction
between the solute and the solvent gainsaid our ability to properly interpret the dielectric
behaviourobserved.Thisproblem is compounded in thecaseofproteinsfortwo
particular reasons: ( i ) whilst x-ray and other measurements can give us a good idea
ofthe static structureofproteins, we still knowratherlittleabouthowmutual
interactions between different groups and domains of the protein serve to control the
dynamics of the constituent groups (Welch 1986), and (ii) the number of atoms in
typical proteins is so large that we cannot expect, euen in principle, to gain a full picture
of the dynamic organisation of a protein molecule and how this affects its functional
behaviour (Welch et a1 1982, Somogyi et al 1984, Kell 1986b). In the latter case, the
argument runs asfollows(e.g.Jaenicke1984, Kell 1986b).Aproteinofmolecular
weight 20 kDalton might contain 200 amino acids and possess, or be able to explore,
some 10'" conformationalstates.Sincetheuniverseitself is 'only' some10'' S old
(Barrow and Silk 19831, even if we allowed o u r protein to explore these states at a
rate of 10'j S-', it cannot conceivably explore all of those available on any relevant
timescale. I f we therefore consider, for instance, the transition from the unfolded to
the folded state, the above reasoning alone means that we cannot state whether or not
a protein in a particular conformation is residing in a local o r a global free energy
minimum. These considerations themselves raise difficultiesof a philosophical nature,

some of which are explored elsewhere (Kell 1986b, Welch and Kell 1986).
The value of the effective dipole moment for small globular proteins, as obtained
from measurements of the dielectric increment and the use of equation (15),is typically
of the order of several hundred D or 5-20 kDalton" (e.g. Takashima 1969, Gerber
et a1 1972, Grant et a1 1978, Pethig 1979). This corresponds to a measured dielectric
increment in the range of 1-15 relative permittivity units for a concentration of 1g
(100ml)".Examples of thedielectricdispersionsexhibited by solutions of non-
enzymic proteins are given in figure 5 for bovine serum albumin ( BSA) and myoglobin
( M B ) . The fall in E ' from a value of about 90 to about 70is due mainly, if not entirely,
totherotationalrelaxationoftheproteinmolecules,and is referredtoasthe p
dispersion. Consistent with this interpretation, but not conclusively so (since changes
in the bulk solvent viscosity can be transmitted to motions of mobile groups on the
protein surface and/or within the protein and vice versa (Beece et a1 1980)), is the
finding that the relaxation time of the p dispersion is proportional to the medium
viscosity(Takashima1962,Laogun et a1 1984).However, BSA is alargermolecule
than MB, so that its relaxation time (cf equation (14)) is greater. Further, given the
size of protein molecules (diameter 1-10 nm), the dipole moments observed correspond
to those expected if the dipole is constituted by a few electronic charges only (say less
than five) separatedbyamoleculardiameter,suggestingthattheorganisation of
proteins is such that the majority of the vector dipoles of which they are constituted
cancel to a high degree.
l
l I l
4 5 6 7 8 9
l g f (Hz1
Figure 5. The dielectricpropertiesofaqueousglobularproteins. ( A ) Bovine serumalbumin ( B S A ) at a

concentration 100.8 mg ml" (data after Essex er a / 1977). ( B ) Myoglobin at a concentration of 10% w / v
(data after Grant er a / 1978).
For frequencies below 10 kHz (for BSA) and below 1 MHz (for M B ) , the rotational
motion of the protein molecules can contribute fully to the polarisabilityof the solutions
and, as the polarisability (per unit volume) of a proteinis greater than that of a water
molecule,theresultant(static)permittivity is greaterthanthat of purewater. In
contrast, as the frequency is raised above the rate at which the protein molecules can
reorientatebyrotation,theproteinmoleculescontributethroughtheiratomicand
electronic polarisation only, and thus the permittivity drops below the value of around
80 expected for bulk water at these frequencies. As mentioned above, this dielectric
decrement should be added to the S defined in equation (17) in order to obtain the
dielectric increment to be used in equation (15).
In an important collaborative study of the dielectric properties of BSA, Essex et a1
(1977) noted two components of the relaxation process underlying the p dispersion,
arising from the fact that the BSA molecule is not spherical but possesses an axial ratio
of some 3 : 1, so that it is capable of rotating about each of the axes of the equivalent
ellipsoid of revolution. Additionally, these workers established that the small decrease
in permittivity (the S dispersion) in the frequency range from about 10 to 200 MHz
could also be resolved into two subprocesses; the process with a shorter relaxation
time was deemed to arise from the relaxation(s) of water molecules bound to the
protein. Figure 5 also shows a 6 dispersion centred at around 100 MHz for M B , which
has recently been discussed in full by Grant et a1 (1986). The effects of fluctuations
in the position of protons on the protein structure were considered to be the two most
probable contributions to the other component of the S dispersion observed in BSA.
As discussed elsewhere (Kell and Harris 1985b), the use of intramolecular cross-linking
reagents represents a favourable experimental opportunity for assessing the contribu-
tion of amino-acid sidechain motions. That the motions of amino-acid sidechains, and
entire domains of the protein, are both asymmetric (Bialek and Goldstein 1985, Glover
et a1 1985) and collective (i.e. non-linear) (see e.g. Somogyi et a / 1984, Welch 1986,
Kell 1986b) not only contributes to the difficulty in refining such interpretations (Kell
and Hitchens 1983) but provides some of the more interesting and advanced possibilities
for the direct detectionof specific protein molecules (Kell 1986a). In this latter regard,
it is worth stating that we know of no study in which the dielectric behaviour of an
enzyme has been studied under turnover and non-turnover conditions; in view of the
possibility that ligand bindingor enzymatic activity might (transiently) be accompanied
by very significant (resonant) changes in the dipole moment of the enzyme (Frohlich
1975), it is to be hoped that such studies, aided by technical advances capable of the
rapid registration of dielectric behaviour, will soon be forthcoming. Finally, we note
that the addition of proteins to glycerol-water mixtures actually leads to the preferential
exclusion of glycerol from proteins (Gekko and Timasheff 1981). Thus the addition
ofglyceroltoaproteinsolutionshouldnotsignificantly affect thatpart of the S
dispersion which is due to the rotation of bound water (since such bound water will
not experience the macroviscosity of the glycerol relative to the forces exerted on it
by the protein). In contrast, the ability of glycerol to effect viscous damping on at
least some of the intramolecular motions of the protein should thus lead to a shift in
frequency of that part of the S dispersion due to the motions of protein sidechains.
Obviouslytherelaxationtimeoftheprotein’s p dispersion will beincreasedin
proportion to the microviscosity. Such an approach would seem to offer interesting
and novel possibilities for the dissection of the mechanistic bases of the S dispersion
of proteins.
Other processes that have been considered to influence the dielectric behaviour of
protein solutions include relaxation of the diffuse double layer surrounding the protein
molecule and ‘surface’ conduction processes associated with the movements of protein-
bound ions. Nowadays, however, such effects are considered to contribute little to the
dielectric behaviour of protein solutions (Grant et a1 1978, Pethig 1979). Finally, the
dispersion centred around 20 GHz for both the BSA a n d M B solutions, and indeed for
946 R Pethig
D and B Kell
all aqueous biological material, is due to the relaxation of ‘bulk’ water molecules.
It is generally found that the relaxation times for proteins calculated on the basis
of equation (14) lead to molecular radii that are larger than the values obtained from
x-ray diffraction experiments. The interpretation of this is that there is a layer or two
ofwater,molecules(thehydrationsheath) so stronglyassociatedwiththeprotein
surface (and indeed visible as such in high-resolution (<0.15 nm) x-ray work) that
theyremainboundtotheproteinmoleculeas it rotates.Obviously,thesewater
molecules experience a strong (electrostatic or dipolar) interaction with the protein
and will tend qualitatively to relax at a frequency much lower than that of bulk water.
This is discussed in more detail in a later section.
3.3. Deoxyribonucleicacid (DNA)
In its native configuration, D N A comprises two anti-parallel helical chains of nucleo-

tides. It may take the form of linear rods or closed circles with varying degrees of
supercoiling, and may have a molecular weight in the range lo6-10’ kDalton. Since
the two helical chains point in different directions, the dipole moments of each backbone
cancel one another. Although the D N A molecule has no net permanent dipole moment
due to the presence of the helices, early dielectric measurements (e.g. Takashima 1963,
1966, Hanss1966)showedconclusivelythat D N A moleculespossessalargedipole
moment directed along the axis of the double helix. This dipole moment arises from
the fact that the D N A molecule bears a net negative charge when in neutral aqueous
solution, so that it will be surrounded by an atmosphere of counter-cations. These
counter-cations will tend to be displaced along the surface of the macromolecule under
the influence of an applied electricfield and thereby give riseto a large induced dipole
moment. The relaxation time of the resulting dielectric dispersion will depend upon
the effective mobility of the ions along the macromolecule’s ‘surface’, and for rod-
shaped macromolecules will be given by
7= ? r e ,L’/2uzqZ (21)
where E , is the effective permittivity of the surrounding ionic atmosphere of z ions
per unit length, U is the (two-dimensional surface) counter-ion mobility, q is the charge
ontheionsand L isthelengthofthe D N A molecule.Theexpectationthat T is
proportional to L’ is indeed borne out in practice (Takashima 1967). Typical dielectric
relaxation times for D N A molecules as received are of the order of 1 ms, whilst values
of the dielectric increment are large and of the order of 1000.
Apart from a dielectric dispersion ( a dispersion) arising from the induced dipole
moment associated with the electrical double layer around the macromolecule, DNA
solutions (colloids) exhibit one or more smaller dispersions in the frequency range
1-50 MHz (e.g. Mandel 1977, Takashima et a1 1984). The relative permittivity of a
1% solution of D N A is shown for the frequency range 0.2 M H z - l 0 G H z in figure 6.
Thehigh-frequencytailofthe a dispersion is clearlyevident,as well as the more
rapidly relaxing dispersion centred in this case around 20 MHz; the latteris considered
to arise from motions of polar groups within the D N A molecule (Takashima et al 1984).
As with the protein solutions, dielectric dispersion above 1 GHz is dominated by the
relaxation of bulk water, and does not differ appreciably from that observed in pure
water.
The finding of Takashimaer a1 (1984) that the high-frequency (1-10 GHz) dielectric
behaviour of an aqueous solution of D N A does not differ appreciably from that of pure
I 1 I I I I 1
5 6 7 8 9 10
l g f [Hz)
Figure 6. The dielectric properties of a lo!, solution of calf thymus oivA (data after Takashima er a / 1984).
water is in contrast to recent reports that large, resonant absorptions can take place
in D N A solutions at these frequencies (Swicord and Davis 1982, 1983, Edwards et al
1984, 1985). These latter workers have provided detailed evidence for the possibility
that aqueous solutions of monodisperse, plasmid D N A at room temperature can res-
onantly absorb microwave energy, and that the mechanism for this absorption is related
tothecouplingoftheelectric field toacousticmodes(standingwaves) in the
macromolecule. If suchresonanceeffects,whicharereallyonlyvisibleagainstthe
‘background’ dielectric properties of the solvent when the D N A is very homogeneous,
can be confirmed (and an extensive effort by Grant and colleagues has failed to do
this (see Gabrielet al 1987)), theywill have significant implications. These implications
includethepossibilitythatbiochemicalprocessesmightbeinfluenced by highly
selective microwave frequencies of low intensity (Frohlich 1980, Frohlich and Kremer
1983) and that acoustic modes (phonons, solitons, etc) could provide a dissipation
less mechanism for transporting free energy over large molecular distances (see Kell
andHitchens1983,Kelland Westerhoff 1985, Kell 1986b andreferencestherein).
Obviously, confirmation of such findings would similarly stimulate the search, using
dielectric methods, for resonant modes in proteins as well.
Since they may form a variety of secondary structures, nucleic acid molecules might
beexpectedtoexhibitadielectricdispersionat very lowfrequencies (f,< 10 Hz).
Whilst this is not likely to be easily observed directlyby dielectric means (for technical
reasons), the hysteretic behaviour of such structural transitions has led to the develop-
ment of an extremely useful orthogonal pulsed-field gel-electrophoretic technique for
the separation of intact chromosomalD N A (Schwartz and Cantor 1984, I’ownson 1986).
Such developments provide an excellent illustration of the use of dielectrically-based
methods in analytical and preparative biochemistry and biophysics.
4. Bound water in biological systems

Although most (but not all) of the hydrophilic amino-acid residuesof a protein molecule
will be situated on the outside ‘surface’ (Connolly 1983) of the proteinin contact with
the aqueous environment, significant areas of the protein surface are likely to be made
u p ofhydrophobicregions.Inlysozyme,forinstance,approximatelyhalfofthe
exposedproteinsidechainsarenon-polar(LeeandRichards1971),although func-
tionally, as judged by hydrogen-exchange kinetics, ‘full’ hydration is attained when
less than half the surface residues have actually been hydrated (Schinkel et a1 1985).
The water molecules surrounding such hydrophobic groups will be forced to form
networks of hydrogen bonds with each other in a way that differs from those characteris-
tic of normal, bulk water; computer simulations indicate (Lee et a1 1984) that such
modifications in water structure can extend at least 1 nm into the bulk liquid from the
hydrophobic surface. Indeed, dielectric studies of the effect of hydrophobic solutes
on water have shown (Hallenga et a1 1980)thatthedielectricrelaxationtimesare
increased,suggestingthatthere is anincreaseddegreeofinteraction(presumably
hydrogen bonding) in the immediate neighbourhood of the hydrophobic solutes.
In a protein solution, one may expect that some water molecules will form hydrogen
bonds with the protein structure and otherswill experience strong electrostatic interac-
tionswithchargedgroupssuchasthoseinglutamateandlysineresidues.Water
moleculesnearchargedmembranesurfaces will besimilarlyaffected. All ofthese
factorsresultintheformationofa‘layer’ or twoofwatermoleculesnearbio-
macromolecular surfaces, characterised by the fact that they have physical (including
dielectric)propertieswhichdifferfromthoseofnormalbulkwater.Suchwater
molecules are usually referred to as ‘bound’ water, the dielectric behaviour of which
has recently been reviewed for several types of biological molecule (Grant et a1 1986).
The existence of protein-associatedor bound water, which as a result its of hindered
rotational mobility exhibits dielectric properties which differ from bulk water, was first
indicatedbystudiesofthedielectricbehaviourofproteinsolutionsatradiowave
frequencies (Oncley 1943) and later confirmed by measurements at microwave frequen-
cies (Buchanan et al 1952). On extrapolating the data observed to frequencies below
that where the relaxation of bulk water occurs, Buchanan et a1 found that the value
of the relative permittivity was lower than expected. This was ascribed to the existence
of water ‘irrotationally’ bound to the protein molecules, in an amount corresponding
to 0.2-0.4 g g” of the protein. The identification of the S dispersion in terms of the
relaxation of such bound water was made by Schwan (1965a, b) and Grant (1965)
during studies of haemoglobin and albumin solutions.
By making measurements of the dielectric properties of protein powders of various
degreesofhydration,effectsassociatedwiththerotationoftheproteinmolecules
themselves and with the relaxation of ions in electrical double layers are largely avoided,
so that the dielectric properties of the protein-bound water may be investigated directly.
Measurements by Bone et a1 (1977, 1981) of the microwave dielectric properties of
protein powders as a function of the degree of hydration indicate that some 7-8% of
the water molecules associated with a fully hydrated protein are so bound as to be
unable to contribute to the dielectric polarisation at lOGHz, a frequency at which
normal bulk water exhibits a marked dielectric absorption. Subsequent investigations
(Bone and Pethig 1982) of the S dispersion for hydrated lysozyme powders indicated
that approximately 36 water molecules are tightly incorporated into each lysozyme
molecule and as such form an integral part of the overall protein structure. I t was
also concluded that vibrational motions within the protein structure contribute to the
S dispersion, especially at the higher levels of hydration, where the water appears to
act as a molecular plasticiser or lubricant for the protein (Bone and Pethig 1985), a
pointofsubstantialsignificanceto our understandingoftherelationshipbetween
Passive electrical
protein flexibility and enzyme kinetic properties (Rupley et a1 1983). Such consider-
ations may also lead to a re-evaluation of the relative significance of the motions of
bound water and protein sidechains in contributing to the S dispersion in aqueous
solutions of globular proteins.
The physical state of water associated with D N A molecules has also been investigated
bydielectricmeans.Microwavestudies(CrossandPethig1983) on herring-sperm
D N A in the temperature range 90-300 K have shown that a considerable proportion of
the bound water, corresponding to some 280 molecules per D N A helix turn, exhibits
properties significantly different from those of bulk water, especially in terms of their
dielectric relaxation below 273 K. Indeed, an effective depression of the freezing point
of this water of no less than 138 K was observed in this study (Cross and Pethig 1983),
indicating that the loss of rotational mobility with decreasing temperature is a much
more gradual process than occurs in the normal (and cooperative) water-ice transition.
Computer simulations by Clementi (1983) suggest that, for levels of hydration u p to
some 270 water molecules per helix turn, the ensemble of water molecules rather
accurately mirrors the electrostatic field generated by the D N A backbone.
Thestrengthofionpairsandhydrogenbondsbetweenaminoandcarboxylate
groups in proteins will obviously be a strong function of the degree of hydration of
the protein. As the hydration is increased, the ability of protons to move between such
groups will similarly be increased. The presence of fluctuating populations of protons
will be of great relevance to the dielectric properties of structural and membrane-bound
proteins, and also of significance in the analysis of proton-conducting pathways in
general (see Westerhoff et a1 1984). I n an important series of measurements, Careri
et a1 (1980) obtained significant details of the protein-water interaction for lysozyme
(see Finney and Poole 1984). By measuring the infrared absorbance at 1580 cm” (an
absorbance characteristic for the carboxylate anion) and the specific heat as a function
of hydration, these workers found that the first water molecules to bind to lysozyme
interactwiththeionisablecarboxylateandammoniumgroups. At around 5 wt0h
hydration (0.05 g water g-’ lysozyme), a redistributionof the proton population occur-
red, together with transitions in the protein-water and water-water hydrogen-bonding
networks. Rates of proton transport as a function of the degree of hydration of protein
powders have been described by Gascoyne et a1 (1981), Behi et a1 (1982), Careri et
a1 (1985) and Morgan and Pethig (1986), and these measurements indicate that protons
are able to migrate relatively freely in protein structures.In particular, the dependence
of the time constant for proton motions on the extent of hydration, and the effect of
a substrate analogue, strongly suggested that such motions are collective in nature and
directed towards the enzyme’s active site (Careri et a1 1985). They may appropriately
be treated by percolation theory (Careri et a1 1986). Cyclodextrins have found some
use as model enzymes, and studies by Bone and Pethig (1983) have shown that the
flip-flop hydrogen bond networks that occur in the cyclodextrin hydrates also provide
a useful model system for the study of proton conductivity.
5. Biological electrolytes
Mammals have a total water content amounting to some 65-70% of their body mass,
and apart from the effects of dissolved macromolecules and membrane surfaces (Clegg
1984), the dielectric propertiesof this water are influencedby the presence of dissolved
ionic salts. The physical effects of such dissolved ions on the dielectric permittivity
of biological fluids arise from more than just the volumeeffect of replacing polar water
molecules by charged but non-polar ionic particles. In particular, the strong electric
field around each ion has the effect of orienting the water molecules, reducing the way
in which they can rotate in response to an applied electrical field. As for the case of
amino-acid solutions, an equation of the form of equation (17) may be used to describe
the permittivity F ' of dilute electrolytes and that ( F ; ) of the pure aqueous solvent in
terms of a dielectric decrement, as follows:
F'= E{ -6C (22)
where 6 is the sum of the decrements arising from the cation and from the anion and
is given by
6=6*+K. (23
Values for 6' and 6 - for various ions in water are given in table 1 for concentrations
( c ) up to 1 molar, a value sufficient to cover all biologically relevant cases except that
of halophilic micro-organisms. Thus, to estimate the extent to which, say, KC1 will
reduce the permittivity of water, we note that the total decrement value for K' and
Cl is S = 11 (table 1 ) . Thus, since the relative permittivity of pure water F ; = 79 at
23 "C, the relative permittivity of a 1 M aqueous KC1 solution at this temperature is 68.
Table 1. Dielectricdecrement values for some ions i n

aqueous solution.
Na' 8 Cl 3
K- 8 F 5
Li 11 I 7
H' 17 so: 7
Mg'+ '4 OH l3
As well as the effect of lowering the relative permittivity of the aqueous solvent,
dissolvedions will generallydecreaseitsrelaxationtime. ( A n ion for whichthis is
not the case is the proton.) To a first approximation, this may be considered to result
from the disruption by the solvated ions of the normal hydrogen-bond structure of
purewater. For concentrations of dissolvedionslessthan 1 molar,thismaybe
expressed in terms of a relaxation frequency increment by the equation
f =.f,+ C V (24)
where
Sf = Sf" + Sf" (25)
represents the effect of the cation and the anion. Values for Sf and Sf are givenin
table 2, and may be used to estimate how far the relaxation frequencies for various
electrolytes differ from the value of 17.1 GHz accepted for pure water at 20 "C.
The overall AC' conductivity of a biological ( o r other) material is given by
(T( W ) = 0
1f W&,rF" (26)
where (T, is theconductivityarisingfromtheelectric-fieldinducedmotions of the
various ions in the electrolyte. I t has the approximate value:
(TI =q c z,n,u, (271
Passive electrical properties of biological systems 95 1
Table 2. Relaxationfrequencyincrementvaluesforsomeions i n aqueous

solution.
H+ -0.34 OH 0.24
Li+ 0.34 Cl 0.44
Na + 0.44 F 0.44
K+ 0.44 so: 1.20
Mg" 0.44 I 1.65
where i denotes an ionic species with valencyz, concentration n and electrical mobility
U. The mobility values (extrapolated to infinite dilution) for some biologically relevant
ions are listed in table 3.
Since a pH value of 7 corresponds to there being 10" mol of H' and OH- ions in
a litre of water, i.e. to a concentration of 6.03 X 10" H+ and OH- ionsI-', the conduc-
tivity of pure water at, say, 24°C may then be calculated from equation ( 2 7 ) , using
the mobility values given in table 3, as 5.4 $3 m". In fact, H+ ions rarely exist as such
in water, since they are rapidly hydrated to the H,O+ ion and hydrated versions thereof.
The mobility of the H30f ion is given in table 3, where it may be observed that it is
the same as that of the proton. The explanation for this is that this apparent mobility
of the hydronium ion is dominated by, and hence equal to, the rate at which protons
transfer between neighbouring hydrogen-bonded water molecules.
It is usually more convenient to deal with the effective molar conductivities of ions;
thesearegivenintable4 for dilutesolutions, so that the conductivity of a dilute
electrolyte can then be obtained from the equation
U = 1 miam,
Table 3. The electrical mobility of some ions i n dilute aqueous solution at
25 "C.
Mobility
(lo-x "-1
Cation Anion (lO-Xm'V - ' S-')
H+, H,O+ 36.3 OH- 20

K+ 7.6 C H , 1- 7.7
Na' 5.0 F -- 5.4
NH; 7.6 Br ~ 7.8
6.2 Ca'+ so:- 8.3
Table 4. Molar conductivity of some ions in dilute aqueous solution at 25 "C.
Conductivity Conductivity
Cation S m'mo~-l) Anion S m' mol")
H+, H,O+ 350 OH- 193

K+ 73 cl-, 1- 74
Na+ 48 F- 52
NH: 73 75 Br-
Ca'+ 119 so:- 160
-
Table 5. Watercontentvaluesforvarioustissuesandorgans.Valuesderivedfrom
Altman and Dittmar (1972) except those for fat (unpublished data of first a u t h o r ) a n d
those for ocular tissue (Gabriel el a / 1983).
Wt 010 Wt Yo
Tissue water content Tissue water content
Bone 44-55 Muscle 73-78

Bone marrow 8-16 Ocular tissues
Bowel 60-82 choroid 78
Brain cornea 75
white matter 68-73 iris 77
grey matter 82-85 lens 65
Fat 5-20 retina 89
Kidney 78-79 Skin 60-76
Liver 73-77 Spleen 76-81
Lung 80-83
where m,is the molar concentration of ions of type i having a molar conductivity g,,,i.
The dominant ions in the extracellular fluidsof human tissue are those of sodium and
chlorine, each at a concentration of some 150 mM. From tables 4 and 5 and equation
(28), the conductivity of tissues such as liver and muscle should be approximately
1.4 S m”, and indeed, as described later (see table 7 ) , this accurately predicts the
conductivity at frequencies around 1 GHz where the electrical properties of tissues are
determined not by the tissue structures and organisation themselves but by the elec-
trolyte content. However, it should be stressed that the electrical conductivity at lower
frequencies (and hence over a longer distanceof ion movement) might be significantly
lower, possibly reflecting frictional interactions between ions and dissolved or cytoskel-
eta1 materials. Thus, for those studying this problemof the organisation of the cellular
cytoplasm, it may be quite misleading to state that the conductivity of a tissue or cell
suspension at l G H z is the same as that to be expected on the basis of its ionic content
andhencethatthereisnospecialorganisationoftheaqueouscytoplasm.This is
because, as discussed in this context by Clegg (1984) for water molecules, the relevant
conductivities are those which are observed over a longer range, and hence a longer
timescale, than that equivalent to 1 GHz. At least in erythrocytes (Pauly and Schwan
1966) and prokaryotes (Marquis and Carstensen 1973) it is found that the electrical
conductivities estimated in situ at frequencies lower than 1GHz are smaller by a factor
of approximately three than those to be expected on the basis of the number and
nature of the ions present and indeed observed upon disruption of the cell membrane
and dilution of the cellular contents. The extent to which this ‘conductivity decrement’
is afunction of thephysiologicalstateofagivenorganism,andvaries between
organisms, is not known, but there are reasons to suppose (Kell and Westerhoff 1985)
that the greater degree of cytoplasmic organisation to be expected in thermophilic
micro-organisms might be reflected in the conductivity decrement of such organisms.
6. Membranes and cells

Cell membranes are generally structures with a thickness of about 5 nm, composed
mainly of a bilayer of long-chain lipid molecules in, on and through which are dispersed
proteins. An approximate value for the effective capacitance per unit area (C,) of
such a membrane can be found using the equation for a slab dielectric:
where d is the thickness of the membrane and E , the permittivity of the membrane-
forming material. Values of C, as a function of d are plotted, for various values of
E , , in figure 7 ( a ) , where it may be seen thatC, does not exceed0.5 p F cm-2 if E , = 2.5
and d = 5 nm. Whilst correct within a factor of two or so, this is in disagreement with
the widely held textbook view (Cole 1972) that biological membranes have a static
capacitance of 1 p F cm-2, and it is of interest to enquire as to why this may be so.
dinm) lb)
Figure 7. ( a ) The relationship between the static capacitance of a biological membrane of thickness d and
thepermittivityvaluesgiven,baseduponequation (29). ( b ) Organisationandelectricalequivalent of a
phospholipid bilayer. Because the capacitance C, of the hydrocarbon region is the lowest of those given,
and is in series with thern, it dominates the observable electrical properties.
In the case of a simple bilayer membrane, composed of a single type of phospholipid,

an estimate of d may be obtained from a measurement of C, and vice versa. In most
cases, the membrane capacitance is largely determined, to within 1 or 2%, by the
hydrophobic phospholipid core (seefigure 7( b ) )(Hanai et a1 1965, Everitt and Haydon
1968, Coster and Smith 1974, Almers 1978, Laver et a1 1984) and the value of d is
reasonably well establishedfrommolecularmodelsandfromx-ray,electronand
neutron diffraction measurements (the latter also gives estimates of the hydration of
thehydrocarboninterior(KnottandSchoenborn1986)).Avalueof 2.0-2.2 has
therefore become accepted for the relative permittivity of black lipid membranes( B L M )
(e.g., see Hanai era1 1964, Tien 1974, Tien and Diana 1968, Benz et a1 1975, Laver et
al 1984, Dilger and Benz 1985, Kell and Harris 1985b), depending upon the phos-
pholipid composition and the solvent used in their preparation (Fettiplace et a1 1975).
It is reasonable to assume that the surface area of an energy-coupling biological
membrane is composed of approximately 30% protein (Hackenbrock 1981, Capaldi
1982), that these proteins are arrangedin complexes of about 4 nm diameter such that
each complex contains one 'permeant' aqueous pore of diameter 1 nm (Edmonds 1981),
and that the permittivityof the hydrophobic core is 2.2. For a spherical microbial cell
of radius 0.55 pm there will then be approximately 14 500 aqueous pores with a total
surface area of 6 of the geometrical surface area of the cell. If each of these pores is
assumed to consist of water with a permittivity equal to that of bound water (which
we may take, from what is known of the capacitance of electrode/electrolyte interfaces,
to have a value of 10 (Bockris and Reddy 1970)), then since the pores are in parallel
with the hydrophobic core the total membrane permittivity will be raised to a value
of about 2.32 (=2.2 +g;). This value is hardly different from that of a pure phospholipid
membrane given above, even when, as here, the assumptions that give rise to it are
rather extreme. Thus the presence of aqueous pores, even transient ones (Weaver et
a1 1984a, b), cannot underly the paradox that the permittivity of biological membranes
appears to be somewhat greater than that of E L M . Since the interior of transmem-
branous, integral proteins is likely to be no less hydrophobic than that of the hydrocar-
bon core of the membrane, it wouldalsoseemimplausible to invokeasignificant
contribution of the proteinsper se to the bulk permittivity and thus to the macroscopic,
staticelectricalcapacitanceofbiologicalmembranes,unlessoneassumesthatthe
intramembranalprotein/lipidinterfacesareextensivelyhydrated,apossibilityfor
which no present evidence exists.
Theearlymeasurementsofbiologicalmembranecapacitanceswhichledtothe
widely accepted value for C, of 1 p F cm-’ (see Cole 1972, Schanne and Ceretti 1978)
werecarriedout on cell suspensions, using extravesicular electrodes. I n thecaseof
sphericalcellswithanegligibletransmembraneconductanceatthefrequenciesof
interest, the following equations are used to describe the electrical properties of the,B
dispersion arising from the presence of the cell membrane (see, e.g., Schwan 1957,
Kell and Harris 1985b):
E; = E L +9PrC,/4~, (30)
a ; = a;)(1 - P ) / [1 + ( P / x ) ]
a; = a;)[1+ 3 P ( v ;+ ab)/(a;+ 2 4 1
T= rCm[(l/a;)+(1/2ab)] (33)
In these equations, the subscripts 1 and CO indicate that the measurements are made
at (or extrapolated to) frequencies that are very low and very high relative to the
characteristic frequencyf, ( = 1 / 2 m ) , P is the volume fraction of the suspended phase,
r the cell radius, the subscripts i and o refer to the (isotropic) inner and outer phases
which the cell membrane separates, x is a shape-dependent ‘form factor’ ( x = 2 for a
sphere) and the other symbols are as above. A nomograph has been published to aid
the rapid calculation off, and hence T from equation (33) (Kell and Burns 1986).
Equation (30) shows that (as did Fricke (1925)) one may estimate the value of C,
from measurements of the dielectric increment of the p dispersion, provided that the
charging of a ‘static’ membrane capacitance is the only mechanism contributing to
dielectric relaxation in the relevant (radio-) frequency range, and translate this into a
value for the membrane permittivity via equation (29). However, an examination of
the available literature suggests that the majority of membrane capacitances found
using extracellular electrodes to be greater than 0.9 p F cm-2 appear to correlate with
an overestimation of the membrane permittivity (e.g. Asami et a1 1976,1980, 19841,
although in one of these studies (Asami et a1 1980) a plot of A E ’ against P does not
extrapolate to the origin as required by equation (30) (Harris 1985). However, it is
worth drawing atttention to the fact that a judicious use of measurements of the R F
permittivity of cellular suspensions has allowed us to develop a novel and convenient
method for the real-time estimation of (microbial or other) biomass in laboratory and
industrial fermentations, a particularly thorny and long-standing problem (Harris et

a1 1987, Kell et al 1987).
Insomeneurophysiologicalstudies,measurementsof C, havebeenmadewith
transmembrane electrodes (Takashima 1976, Haydon et a1 1980). In this case, whilst
the transmembrane capacitance may exceed p F1cm-' at low frequencies, the sensitivity
of this value to inhibitors of the opening of ion channels, and the fact that the value
drops below 0.5 p F cm-? at higher frequencies (which are still low compared with the
value of f c to be expected for a classical p dispersion caused by a Maxwell-Wagner
type of mechanism (see later)), suggests again that the true value for C, i n biological
membranes is probably not greater than perhaps 0.6 pF cm". In the caseof neuromem-
branes monitored with transmembrane electrodes, it is evident that the larger values
are associated with (the highly non-linear) gating currents and the consequent trans-
membrane movement of ions in response to the large transmembrane electrical fields
which are set up under the conditions of this type of measurement.
Inthecaseofmeasurementswith extracellular electrodes,thetransmembrane
electrical potentials (V,,,) which are set up in response to the applied field are miniscule
and given (e.g., Zimmermann 1982) by the equation
V,=1.5rE,cos~/[l+(u~)~]~~~ (34)
where E,, is the macroscopic electrical field between the electrodes and 4 is the angle
between the plane of the membrane and the direction of the electric field. Thus gating
currents and field-induced transmembrane current flow asoccurring in neuromem-
branes are not likely to underlie the high values for C, invoked in some studies. One
possibility at least is the (partially) restricted lateral electrophoresis of double layer
ions and membrane lipids and proteins, as has been discussed in detail elsewhere (Kell
1983, Harris and Kell 1985, Kell and Harris 1985a, b, Kell and Westerhoff 1985). The
lateral electrophoresis of membrane components has also been observed under the
microscope (Po0 1981,SowersandHackenbrock1981)and is likelytocontribute
tofield-inducedmodulationofbiochemicalactivities(ChiabreraandRodan 1984,
Grattarola et a1 1985, Kell and Harris 1985b).
Electrically, we may representtheresistiveandcapacitivepropertiesofa cell
membrane as an equivalent electrical circuit of the form of figure 8 ( a ) , wherethe
membrane capacitance C, is shown in parallel with the membrane resistance R,. A
feature of such a circuit is that with increasing frequency the membrane resistance is
increasingly short-circuited by the reactance (l/uC,) of C,. The consequence of this
in terms of the electrical properties of the cell is shown in figure 9 ( a ) a n d 9 ( h ) . At
low frequencies (figure 9(a)) the resistance of the cell membrane insulates the cell
interior (cytoplasm) from an external electrical field and no current is induced within
the cell interior. Thus the cell appears as an insulating spheroid and decreases the
effective conductivity of the aqueous suspension, as embodied in equation 1 ) . (3
Indeed,
measurements of the low-frequenq conductivity of a suspension and the fluidin which
it is suspendedhavebeenexploitedtoprovidearapidestimationofthebiomass
content of immobilised cell suspensions, an otherwise extremely difficult task (Lovitt
et a1 1986). At higherfrequencies(figure9(b)),theshort-circuitingeffectofthe
membrane capacitance allows the electric field to penetrate into the cell untilata
sufficientlyhighfrequencytheeffectivemembraneresistancebecomesvanishingly
small and the cell appearsdielectricallyasaspheroidcomposedofthecytoplasm
dispersed in the suspending electrolyte. Thus the effective permittivity and conductivity
of a cell suspension will respectively fall and rise with increasing frequency, leading
d
!b l
Figure 8. ( a ) The equivalent electrical circuit for a cellular membrane, as assessed using transmembrane
electrodes. R , and C, arethemembraneresistanceandcapacitancerespectively. ( 6 ) An equivalent
electrical circuit for a cell or vesicle (suspension) as observed using external electrodes. R , represents the
resistanceoftheextracellularfluidwhilst R , represents the access admittance bywhichthemembrane
capacitance C, may be charged. The resistors may additionally be drawn in series with another capacitor
to represent the geometric (high-frequency) capacitance of the space between the electrodes. The circuit as
drawn possesses a single relaxation time (i.e. dispersion) corresponding to the p dispersion observed i n
vesicular, cellular or tissue material.
"
(C)
Figure 9. The flow ofcurrentatdifferentfrequenciesrelativetothecharacteristicfrequencyofthe p

dispersion of a cellular system. ( a ) At low frequencies, the membrane resistance shields the cell interior
from the applied electrical field. ( b ) At higher frequencies, the membrane resistance is progressively shorted
out by the membrane capacitance, such that the field enters the cell, current flows through the cell interior
and the measured conductivity is increased. (For further details, see e.g. Zimmermann1982, Kell and Harris
1985b.) ( c ) The effect of a small conducting pore in the cell membrane is to change the voltage drop across
the membrane so that i t is inhomogeneous and less than that in the siruation described in ( a )
to the existence of a dielectric dispersion, the p dispersion, of roughly the same form
as thep dispersion exhibited by tissues (figure 10). The existence of a large, frequency-
dependent dielectric increment in heterogeneous materials is generally known as the
Maxwell-Wagner effect.
The representation of a completely insulating membrane, as depicted in figures
9( a ) and 9( b ) is obviously only an approximation. Voltage-dependent, ion-conducting
Figure 10. An idealised representation of the way in which the permittivity of a biological tissue such as
skeletalmusclemightvarywithfrequency,showingthethreemajordispersions ( a , p and y ) generally
observed.
channels are known to exist in cell membranes and these could act as localised areas
of lowmembraneresistance.Whilstthemagnitudeofthe p dispersion is rather
insensitive to the macroscopic conductivity of the membrane (Schwan 1957, Harris
and Kell 1985), Klee and Plonsey (1974)have shown that such areas can alter the
field pattern across a cell membrane, as indicated for example in figure 9(c). The exact
roleofsuchbehaviourinmodifyingthequantitativedielectricpropertiesof cell
suspensions (Casaleggio et a1 1984, 1985, Marconi et a1 1985) is as yet uncertain, but
is of obvious relevance to studies of electroporation and electric-field-induced cell-to-
cell fusion (Zimmermann 1982) and of the ionic currents implicated in developmental
processes by the aficionados of the vibrating probe approach (Nuccitelli 1986). The
facts that (i) the characteristic frequency of the p dispersion is dependent upon the
cell radius and (ii) the existence of conducting ion pores is both voltage and frequency
dependent means that no hard rules can be drawn up to assess the contribution of
such behaviour to the p dispersion in the absence of reasonably detailed electrophysio-
logical knowledge of the system of interest. One might well imagine that patch clamp
studies (Sakmann and Neher 1983,1984,Owyer 1985)could be ofbenefit here, although
they are not without their pitfalls (Kell 1986~).
A simpleequivalentelectricalcircuitwhich is often used to represent a cell o r
vesicle suspension is shown in figure 8(6), where R , represents the resistance of the
extracellularmedium, C, themembranecapacitanceand R ? , theso-calledaccess
impedance, is a composite function of the membrane and cytoplasmic resistances.
As with aqueous globular proteins, cell membranes contain ionisable acidic and
basic groups; additionally, because of the predominance of acidic phospholipids, most
cell membranes bear a net negative charge at physiological pH values. As depicted
infigure l(a), therefore, there will be an electrical double layer at the membrane-
solution interface on either side of the membrane. We may therefore expect, as with
D N A colloids,adielectricdispersionduetorelaxation effects withinatleastthe
extracellular electrical double layer. Such a dispersion is known as the a dispersion
a n d is similar to that shown for a 'typical' muscle tissue in figure 10. I n micro-organisms
it is much more evident in Gram-positive than Gram-negative cells, a consequence of
the very different organisation of the cell envelope in the two types of prokaryote
(Harris and Kell 1985), althoughno exact theory exists due to the molecular roughness
of the cell surface.
7. Tissues
Thedielectricpermittivityofbiologicaltissuestypicallydecreaseswithincreasing
frequency in three major steps (i.e. dispersions) which are designated the a , p and y
dispersions, an idealised and widely cited representation of which (Schwan 1957) is
giveninfigure 10.It should be mentioned, however, that in some cellular systems,
the a and dispersions are by no means as well separated as suggested in figure 10.
The a dispersion is generally considered to be associated with the relaxation of ions
tangential to charged membrane surfaces, while the p dispersion is shown as a doublet
in figure 10 so as to illustrate that it is composed of both a Maxwell-Wagner effect at
cell membranesandtherotationofproteinaceaousmaterial(andprobablybound
water). The y dispersion arises mainly from the relaxation of free water within the
tissues. Not the least of the problemsof interpreting the dielectric behaviour of tissues
arises from the fact that the volume fraction of the suspended phase approaches one,
so that existing exact biophysical or mechanistic explanations, which hold only for
cell suspensions of well defined geometries present at a volume fraction less than 0.2,
are inapplicable.
For technical reasons connected with the measurement of very small phase angles
in the presence of potentially large artefacts due to electrode polarisation (even in
four-terminalmeasurements;see, e.g., Schwan 1963, Grant et al 1978,Kell 1986a),
rather little work has been reported concerning the dielectric properties of tissues in
the frequency range up to a few kHz. The early work of Schwan (1954) showed that
skeletal muscle possessed a significant a dispersion which decreased with time follow-
ing the excision of the tissue, a finding qualitatively consistent with the view that the
a dispersion depends upon the physical integrity of cell membranes. Singh er al(1979)
described a study of the low-frequency dielectric properties of freshly excised kidney
and in vivo measurements, using external electrodes, of normal and malignant breast
tissue. A representative selection of these data is shown in figure 11, where it may be
seen that malignancy appears to influence the observed dielectric behaviour rather
markedly. Another clear example of the a dispersion in tissues may be found in the
work of Kosterich et a1 (1983) on excised rat femurs.
The dependence of the p dispersion upon the integrity of cellular membranes was
indicated by the work of Pauly and Schwan (1964), in which the effect of the detergent
digitonin in lysing the fibre membrane of bovine eye tissue was studied (figure 12).
Nonetheless, these and related experiments on erythrocytes (Pauly and Schwan 1966)
used rather high concentrations of detergent, and it is also possible that the structural
basis for any other type of mechanism underlying the p dispersion (e.g. the diffusion
of membrane components) was also destroyed by the addition of the detergent. In
view of current controversies regarding the interpretation of the effects of much lower
concentrations of detergent on the organisation of the cytoskeleton and microtrabecular
lattice (Schliwa et a1 1981), one might be prudent to regard these early studies as less
definitive than they might firstat sight appear, althoughit must be added that disruption
of the cell structure is certain to have a major effect on the p dispersion. Thus far we
have considered the classical mechanism of the p dispersion in terms of the short-
circuiting of the membrane resistance by the membrane capacitance and, whilst this
yY """" cB
log f (Hz)
Figure 11. Dielectric permittivity of human kidney ( A ) , normal breast tissue ( B ) and breast tissue containing
a malignant tumour ( C ) , as measured using a four-electrode technique (data after Singh er a / 1979).
I I I 1
5 6 7 8 9
log f (Hz1
Figure 12. Dielectric permittivity of a suspension of bovine eye lens fibres ( A ) , the same, after lysis of the
cell membrane using digitonin ( B ) , and a 0.9% saline solution at 2 7 ° C ( C ) .
is now the accepted interpretation,it does not altogether emphasise that the underlying
physicalmechanismresides in theinterfacialpolarisationsassociatedwiththefact
that the membrane-solution boundary marks the interface between two heterogeneous
structures. As indicated above, there is a build-up of charge at the interface between
two dissimilar dielectrics, so that this gives rise to an interfacial or Maxwell-Wagner-
Sillars type of polarisation. The equations for the magnitude and relaxation time of
the dispersion in the case of a spherical shell membrane vesicle have been given above;
othercasessuchassolidparticlesandoil-in-waterandwater-in-oilemulsionsare
discussed, for instance, in Pethig (1979, ch 4).
Figure 12 also shows the frequency dependence of the permittivity of an aqueous

0.9% w/v KC1 solution, a concentration corresponding to the ionic strength of extracel-
lular fluid. It may be observed that the dielectric properties of the tissue are rather
similar to that of the ionic electrolyte above a frequency of some 50-100 MHz, although,
due to the presence of non-polar membranous and other materials, the permittivity of
the tissue between 0.1 and 1 GHz is slightly lower.
At these high frequencies, let us say above 100 MHz, it is to be expected that the
dielectric properties of the tissues will largely reflect the properties of the intra- and
extracellular electrolytes, and in particular will exhibit a dielectric dispersion corre-
sponding to the relaxation of water dipoles, as in figure2 . To assess this, we give data
for the high-frequency permittivity and resistivity of several tissues and a 0.9% saline
solutioninfigure13.Resistivityvaluesareusedthebettertoindicatedifferences
between the tissues and the solution. It is clear that the dielectric behaviour of tissues
in this frequency region is greatly influenced by the tissue water content, so that muscle,
with a typical water content of 75% w/w, exhibits a much higher permittivity and
conductivity than does fat (which typically has a water content of some 5-20% w/w).
Sincethetissuedielectricpropertiesareratherindependentofthestructureand
organisation of membranes at these frequencies, it is also to be expected that theywill
be little changed as a function of time post mortem. This expectation is in contrast
to the lower-frequency behaviour, and is borne out in practice (Kraszewski er a1 1982,
Surowiec et a1 1985, 1986). Taking these facts together, then, it is rather obvious that
dielectric methods might be suitable for the rapid non-invasive determination of the
water and/or fat contents of all sorts of tissues and biological (and other) materials,
7 0
8
c
9
Log f (Hz1
10
0
8 9 10
log f ( H z )
Figure 13. The high-frequency permittivity ( a ) and resistivity ( b ) of rat brain ( A ) , rat muscle ( B ) and rat
adipose tissue ( C ) (at one fifth scale) obtained using an in ~ i v oprobe (data after Burdette er a / 1980). ( D )
a 0.9% saline solution (data from Schwan and Li 1953 and Schwan et al 1976).
Passive electrical
and indeed microwave dielectric methods have long been used to assess the moisture
content of stored grain(e.g., see Pande 1975, Pyper 1985), a variable of great economic
importance. The water content of various tissues is tabulated in table 5.
Schwan and Foster (1977), studying the microwave dielectric properties of liver,
muscle and skin, concluded that these properties could be ascribed to the bulk water
content with an appropriate correction being made for the presence of bound water
andsolidmaterial.Thesamegroupcametoasimilarconclusionconcerningthe
microwave dielectric properties of canine brain tissue (Fosteret a1 1979) and a variety
of normal and malignant tissues (Foster et a1 1980, Schepps and Foster 1980). More
recent studies have confirmed (Foster et a1 1982b, Stuchly et a1 1982, Gabriel er a1
1983, Smith and Foster 1985) that high-water-content tissues in particular, and also
ionic and non-ionic microemulsions (Foster er a1 1982a, Epsteiner a1 1983) and polymer
solutions (Foster er a1 1984), exhibit dielectric properties that can be accounted for
by a simple mixture theory provided that a contribution from bound wateris included.
A similar conclusion has been arrived at by Clegg er a1 (1982), who also point out ( i )
that the estimation by physical methods of the degree to which tissue water is normal
bulk water is extremely sensitive to the model used in the interpretation of the data,
and (ii) that the degree of motional freedom ascribed to tissueis water highly dependent
upon the time- (and hence distance) scale considered. As discussed above, this last
point is of some significance to the estimation of intracellular conductivity values.
7.1. Skin
Many therapeutic and diagnostic techniques rely upon the application of electrical
fields or the measurement of electrical properties. Since skin tissue often constitutes
the interface between the biological and electronic parts of the system, its dielectric
properties are of some interest and importance (Salter 1979, Gabriel et a1 1986). The
dielectric properties of skin are largely determined by the stratum corneum, which has
a thickness (in non-calloused areas) of some 15 p m a n d consists largely of dead cells.
These dead cells are formed mainly of keratin and membranous matter and tend to
wear off, to be replaced by underlying epidermal cells.
The dielectric properties of skin show considerable variability over different parts
of the body, the macroscopic electrical admittance being greatest in those parts, such
asthepalms,whicharemostfreelysuppliedwithsweatducts.Rosendal(1945)
measured the dielectric properties of piecesof wet, freshly excised skin approximately
1 mm thick, and obtained values for the effective capacitance and resistance at 1 and
10 kHz of 4.6 n F cm-2 and 34.9 kf2 cm2, while a value of 6.2 kR cm' was obtained for
theeffectiveseriesresistanceoftheskinplustheunderlyingtissue. I f therelative
permittivityofdry,keratinousandmembranousmaterial is takenforthesakeof
argument to have a value of ten, these results may be interpreted to show that the
capacitive, resistive element of skin, which provides the uppermost layer of the protec-
tive barrier between the body tissues and the environment, has a thickness of the order
of 2 pm.
The average in vivo electrical properties of skin in the range 1 Hz-l MHz, deter-
mined by Schwan (1965b) and by Yamamoto and Yamamoto (1976), are shown in
figure 14. An interpretation of these properties was approached via a consideration
of the inhomogeneous structure and composition of skin and of the way in which this
varies from the skin surface to the underlying dermis and subcutaneous tissues. It
0 2 4 6 8 10
log f ( H z )
Figure 14. The dielectric permittivity and conductivity of skin at 37 “C. Data based on those of Yamamoto
and Y a m a m o t o ( 1 9 7 6 ) a n d S c h w a n ( 1 9 6 5 b ) .
maybenotedthatskinpossesses a relativelyweak a dispersion(atleastdownto

1 Hz), and this relative lack of a significant dispersion in the frequency range 1 Hz-
10 kHz is plausibly ascribed to the dead nature and low conductivity of the stratum
corneum. Clar er a1 (1982) have found that the dispersion exhibited by normal skin
inthefrequencyrange 0.5 Hz-l0kHzcanbedescribedinterms of twoseparate
relaxations centred at frequencies of approximately 80 Hz and 2 kHz. It was tentatively
suggestedthattheoriginofthesedispersionslayinthe stratumcorneum and was
associated with the relaxation of ions surrounding the corneal cells. The dielectric
response of psoriatic skin was found, as might be expected, to differ significantly from
that of normal skin. Finally, it is worth drawing attention to the fact that the electric
impedance of those parts of the skin at which the points and meridians of acupuncture
are located is significantly less than that of the surrounding tissue (see Becker and
Marino 1982, Jakoubek and Rohlicek 1982), a fact which may be presumed to be of
diagnosticvalueandmayplausiblyunderliethemechanism by whichsignalsare
transmitted around the body by means of the meridians of acupuncture. Whilst the
relevant current carriers are not known, the fact that the locations of the meridians d o
not correlate with any known anatomical features might be interpreted to mean that
the current carriers are protons.
7.2. Othertissues,includingtumours
The relative permittivity and conductivity of various tissues at 37 “C are summarised
in tables 6 and 7, respectively. These data, whilst considered reliable, are derived from
the references cited, and might most suitably be taken to represent ‘average’ or ‘typical’
values (Lin 1975), since the physiological conditions and the history of the organisms
from whence the tissues were derived are very rarely given. Indeed, it does not seem
out of place to remark that the advent of rapid, digital impedimetric instrumentation
means that one can now contemplate many more systematic studies than have hitherto
been possible of the effect of various treatments and environments on the dielectric
properties of biological systems, with a view to answering biologically relevant ques-
tions. The ranges of tissue water contents given in table 5 may be used to estimate
the likely spread of electrical properties to be expected for the various tissues.
The results shown in figure 11 for breast tumours indicate that the permittivity of
cancerous tissue may be significantly greater than that of normal tissues, over a wide
Passive
electrical
properties of biological
systems 963
Table 6. The relative permittivity of biological tissues at 37 “C for various frequencies commonly used for
therapeutic purposes.
Material 13.56 M H z 27.12 MHz 433 M H z 915 MHz 2.45 G H z Xeference
Artery - - - - 43 Brddy er a / i 19x1 l

Blood 155 1 10 66 62 6n Schwan 1 1965bl
Bone
marrow)
(with l1 9 5.2 4.9 4.8 Pethig (1979)
(inHank’ssolution) 28 24 - - - Schwan i 1954)
Bowel
( p l u s contents) 73 49 - - - Hahn er a / i 19801
Brain
iwhite matter) 182 123 48 41 35.5 Foster el a / (1979)
(grey matter) 310 186 57 so 43 Foster er a / 11979)
Fat 38 7?
” 15 15 12 Burdette er a /
i 1980)
Kidney 402 229 60 55 50 Burdette er a /
(1980). Stoq e / a /
i 1982)
Liver 288 182 47 46 44 Burdette er a /
i 19801, Stoy er a /
i 1982)
Lung
(inflated) 42 29 - - - Schwan (1985b),
Hahn er a / I 1980)
(deflated) 94 57 35 33 - Schwan (1965b).
Hahn er a / (1980)
Muscle 152 112 57 55.4 49.6 Schepps and Foster
(19801
Ocular tissues
(choroid) 240 144 60 55 52 Gabriel er a /
i l983 1
icornea 132 100 55 51.5 49 Gabriel er a /
i 1983 1
iirisl 240 150 59 55 52 Gabriel er a /
(1983)
(lens cortex) 175 107 55 52 48 Gabriel er a /
i 1983)
ilens nucleus) 50.5 48.5 31.5 30.8 26 Gabriel er a /
i 1983 1
(retina) 464 250 61 57 56 Gabriel er a /
( 1983 )
Skin 120 98 47 45 44 Schwan i1965b),
figure 13
Spleen 269 170 - - - Stoy er a / i 1982)
frequency range, a trend first noted by Fricke and Morse in 1926. This feature is also
indicated by the data presented in table 8 and is particularly clearly drawnin the work
of Bottomley and Andrew (1978)on rat liver and that of Rogers et a1 (1983) on mouse
muscle. The conductivity of tumours also seems to be greater than that of the normal
tissue from which they are derived, and this property (Foster and Schepps 1981) might
be of use both in the further development of the R F and microwave hyperthermic
treatment of tumours and in impedance imaging systems, muchas N M R imaging systems
rely upon the fact that the water in tumours exhibits markedly different transverse and
Table 7. The conductivity ( S m") of biological tissues at 37 "C.
Material 13.56 MHz 27.12 MHz

MHz
MHz
433
915 2.45 GHz
Reference
Artery - - - - 1.85 Brady et a / ( 1981 )

Blood 1.19 1.27 1.41 2.04 Schwan ( 1965bl
Bone
marrow)
(with
0.03 0.04 0.1 1 0.15 0.2 1 Stoy er a / (19821
( i n Hank'ssolution)0.021 0.024 - - - Stoy er a / (1982)
Brain
(white matter) 0.27 0.33 0.63 0.77 1.04 Foster et a / (1979)
(grey matter) 0.40 0.45 0.83 1.0 1.43 Foster er a / (1979)
Fat 0.21 0.21 0.26 0.35 0.82 Burdette er a /
(1980)
Kidney 0.72 0.83 1.22 I .4l 2.63 Burdette et a /
( 1980). Stoy er a /
(1982)
Liver 0.49 0.58 0.89 1.06 I .79 Burdette et a/
11980), Stoy et a /
i 1982)
Lung
(inflated) 0.1 1 0.13 - - - Schwan (196%).
Hahn er a / (1980)
(deflated) 0.29 0.32 0.71 0.78 - Schwan (1965b3,
Hahn et a/ (1980)
Muscle 0.74 0.76 1.12 1.45 2.56 Schepps and Foster
(1980)
Ocular tissues
(choroid) 0.97 1 .o 1.32 1.40 2.30 Gabriel et a /
( l983 1
(cornea) 1.55 1.57 1.73 1.90 2.50 Gabriel et a /
(1983)
(iris) 0.90 0.95 1.18 1.18 2.10 Gabriel et a /
( l983 l
(lens cortex) 0.53 0.58 0.80 0.97 1.75 Gabriel er a /
i 1983)
(lens nucleus) 0.13 0.15 0.29 0.50 1.40 Gabriel et a /
(1983)
(retina) 0.90 1 .o 1S O 1.55 2.50 Gabriel et a /
(1983)
Skin 0.25 0.40 0.84 0.97 - Schwan (1965b1,
figure 13
Spleen 0.86 0.93 - - - Stoy et a/ (1982)
spin-lattice relaxation times from that of the surrounding tissues (Hazlewood et a1

1974). N M R results have also shown u p the fact that the water content and sodium
concentration of tumour tissues are higher than in normal cells (Damadian and Cope
1974, Pool et a1 1981). The low-frequency dielectric results may also be related to the
fact that cancerous cells have a reduced (less negative) membrane potential (Cone
1969, ConeandTongier1971)andanalteredabilitytotake up positiveions
(Ambrose et a1 1956, Purdom et a1 1958). By incorporating these known differences
inthewatercontentandioniccompositions of thetwotypesoftissue(normal
againsttumorous),GrantandSpyrou(1985)haverecentlyandsuccessfulybeen
able to model the
interfacial
polarisations so as to
reproducethe
dielectric
differences that were observed by Bottomley and Andrew (1978).
L A
e,
U
- M
--ft.
g 8 g
L Q Z Z
U 5%-
e, 2 2Z-G
E
2
g”g3z
eC322c
:
b
P
B $m m5 aP Z4 %I S
&
“ - G 2
“ “ 2 1 1 c??
P 0 0 1 0 0
2
E
9 % Z Z % l l b m
c; z
- e ,m-
’-
.m E
h
m
v1
gz
l z k2$
e,
2
.-
W
P
**c’: l
N t ”
0 3 0
2 s a s
O E
f
2E y
m
5 2 m o m
W
l l$:3%
zS m e m Q m
E2
L
S
P m - m m m
099
-M= S P 22s 1 I 0 -
F
- 2
E *
c:
P
: : 4
r-Nb
m i o m I l -\o
QUI
W
.-x
W
.->
W
m
t-
30 m y 0 -
! I“ p s - - - - & - II
- m m m m
5 =
-
y 2
G e
Q m * 3 0 0 , ~x ~ ~ * o
0“z“- - m
.->X
W
N 1 1
.-
.-
W
I
E
a
U W Q m m m
.- c’:-?q“q*Y
-; I “ P O ” - N m - l l
2 5
052
U EEEsz%%
I I
3 l
F
-J
g 2E + <-
m g
i
g mE - :
:,g:
‘ K E D b E
.- m 2 a 2 W
3 - z
- nzggzg
. ? - c m G
S €
yz
-Z0
a
-
.-m
b
i
“2.2 a;
S’S E B € , S
Zg Zu E CZ i; Z; bUS
e,
5g
SE
g
*G
It is worth considering the fact that tumour cells are more electronegative than
normal cells (Ambrose et a1 1956) and that tumour tissues are similarly more elec-
tronegative (i.e. have a more negative surface potential) than normal tissues (Schaubel
and Habal 1969, 1970). Such results would suggest that structurally less differentiated
tissues are more electronegative than normal tissues, as observed also for regenerating
tissue by Becker (1961). In conclusion, it is evident that dielectric methods can usefully
complement more traditional approaches, and, particularly in concert with modern
molecular biological techniques, might provide new insights into the structural and
electrical differences between normal and transformed cells.
Acknowledgments
RP thankstheScienceandEngineeringResearchCouncil,theMedicalResearch
CouncilandtheNationalFoundation for CancerResearch,andDBKthanksthe
Science and Engineering Research Council and IC1 Biological Products Business for
financial support. We are extremely grateful to Professor E H Grant for many useful
and stimulating discussions.
References
Ackers G K and Smith F R 1985 Ann. R e t . Biochem. 54 597-629
Adey W R 1981 Physiol. Rec. 61 435-514
Adey W R and Lawrence A F (ed) 1984 Non-linear Electrodvnamics ofBiologicalSj.stemr (New York: Plenum)
Almers W 1978 Rev. Physiol. Biochem. fharmacol. 82 96-190
Altman P L and DittmarD S 1972 Biology Data Handbook (Bethesda, MD: Federationof American Societies
for Experimental Biology) 2nd edn pp 392-8
Ambrose E J , James A M and Lowick J H B 1956 Nature 177 576-7
Anderson F A Jr 1984 Ann. Biomed. Eng. 12 79-102
Arnold W M, Wendt B, Zimmermann U and Korenstein R 1985 Biochim. Biophyy. Acta 813 117-31
Asami K, Hanai T and Koizumi N 1976 J. Membr. Bid. 28 169-80
~ 1980 Biophys. J. 31 215-25
Asami K, Irimajiri A, Hanai T, Shiraishi N and Utsumi K 1984 Biochim. Biophys. Acta 778 559-69
Baker L E 1979 Non-invasice Physiological Measurements e d P Rolfe (New York: Academic) pp 65-94
Barber 0 C and Brown B H 1984 J. Phys. E: Sri. Instrum. 17 723-33
Barker A T and Lunt M J 1983 Clin. Phys. Physiol. Meas. 4 1-27
Barrow J D a n d Silk J 1983 The Left Hand of Creation: the Origin and Ecolurion o f t h e Expanding Uniuerse
(London: Unwin)
Becker R 0 1961 J. Bone J t Surg. 43A 643-56
Becker R 0 and Marino A A 1982 Electromagnetism and Life (Albany, NY: State University of New York
Press)
Beece D, Eisenstein L, Frauenfelder H, Good D, Marden M C, Reinisch L, Reynolds A H, Sorensen L B
and Yue K T 1980 Biochem. 19 5147-57
Behi J, Bone S , Morgan H and Pethig R 1982 In/. J. Quantum Chem. Quantum B i d . Symp. 9 367-74
Benz R, Frohlich 0, Lauger P and Montal M 1975 Biochim. Biophys. Acta 394 323-34
Bialek W and Goldstein R F 1985 Biophys. J. 48 1027-44
Blundell T, Barlow D, Borkakoti N and Thornton J 1983 Nature 306 281-2
Bockris J 0 M and Reddy G K N 1970 Modern Electrochemistry v01 2 (New York: Plenum)
Bone S, Eden J and Pethig R 1981 Int. J. Quantum. Chem. Quantum Biol. Symp. 8 307-16
Bone S, Gascoyne P R C and Pethig R 1977 J . Chem. Soc. Faraday Trans. I 73 1605-11
Bone S and Pethig R 1982 J. Mol. Biol. 157 571-5
- 1983 Inr. J . Quantum Chem. Quantum Biol. Symp. I O 133-41
-1985 J. Mol. Biol. 181 323-6
Bottomley P A and Andrew E A 1978 fhys. Med. Biol. 23 630-43
Boyd R H 1980 Methods ofExperimental Physics v01 16C ed R A Fava (New York: Academic) pp 379-421
Brady M M, Symons S A and Stuchly S S 1981 l € € € Trans. Biomed. Eng. BME-28 305-7
Brown B H 1983 Imaging with non-ionising radiation ed D F Jackson (Guildford: Surrey University Press)
pp 85- 110
Buchanan T J, Haggis G H, Hasted J B and Robinson B G 1952 Proc. R. Soc. A 213 379-91
Burdette E C, Cain F L and Seals J 1980 I E E E Trans. Microwaoe Theory Techno/. MIT-28 414-27
Capaldi R A 1982 Biochim. Biophys. Acta 694 291-306
Careri G, Garaci M, Giansanti A and Rupley J A 1985 Proc. Natl Acad. Sci. U S A 82 5342-6
Careri G, Giansanti A and Rupley J A 1986 Proc. Nut/ Acad. Sei. U S A 83 6810-4
Careri G, Gratton E, Yang P H and Rupley J A 1980 Nature 284 572-3
Casaleggio A, Marconi L, Morgavi G, Ridella S and Roland0 C 1984 J. Bioelec. 3 305-28
- 1985 Bioelectrochem.Bioenerg. 14 13-21
Chiabrera A, Nicolini C and Schwan H P (ed) 1985 Interactions Between Electromagnetic Fields and Cells
(New York: Plenum)
Chiabrera A and Rodan G A 1984 J. Bioelec. 3 509-21
Clar E J, Carnbrai M and Sturelle C 1982 Cosmet. Toilet. 97 33-40
Clegg J S 1984 Am. J. Physiol. 246 R133-51
Clegg J S , McClean V E R, Szwarnowski S and Sheppard R J 1984 Phys. Med. B i d . 29 1409-19
Clegg J S, Szwarnowski S, McClean V E R, Sheppard R J and Grant E H 1982 Biochim. Biophys. Acta 721
458-68
Clementi E 1983 Structure and Dynamics: Nucleic Acids and Proteins ed E Clementi and R H Sarma (New
York: Adenine) pp 321-64
Cole K S 1972 Membranes, Ions and Impulses (Berkeley, CA: University of California Press)
Cole K S and Baker R F 1941 J. Gen. Physiol. 24 771-88
Cole K S and Cole R H 1941 J. Chem. Phys. 9 341-51
Cone C D 1969 Trans. N Y Acad. Sci. 31 404-27
Cone C D and Tongier M 1971 Oncology 25 168-82
Connolly M L 1983 Science 221 709-13
Coster H G L and Smith J R 1974 Biochim. Biophys. Acta 373 151-64
Creighton T E 1983 Proteins.StructuresandMolecularProperties (San Francisco: Freeman)
Cross T E and Pethig R 1983 In/. J. Quantum Chem. Quantum B i d . Symp. 10 143-52
Damadian R and Cope F W 1974 Physiol. Chem. Phys. 6 309-22
Debye P 1929 Polar Molecules (New York: Dover)
Dilger J €'and Benz R 1985 J. Membr. Biol. 85 181-9
Dunning W J and Shutt W J 1938 Trans. Faraday Soc. 34 479-85
Dwyer T 1985 J. Electrophysiol.Techniques 12 15-29
Edmonds D T 1981 Trends Biochem. Sci. 6 92-4
Edwards G S , Davis C C, Saffer J D and Swicord M L 1984 Phys. Rea. Lett. 53 1284-7
- 1985 Biophys. J. 41 799-807
Epstein B R, Foster K R and Mackay R A 1983 J. Col/. Interface Sci. 95 218-27
Essex C G, Symonds M S, Sheppard R J, Grant E H, Lamote R, Soebewey F, Rosseneu M Y and Peeters
H 1977 Phys. Med. Biol. 22 1160-7
Everitt C T and Haydon D A 1968 J. Theor. B i d . 18 371-9
Fettiplace R, Gordon L G M, Hladky S B, Requena J, Zingsheim H P and Haydon D A 1975 Methods in
Membrane Biology v01 4 ed E D Korn (New York: Plenum) pp 1-75
Finney J L and Poole P L 1984 Comment. Mol. Cell. Biophys. 2 129-51
Foster K R, Cheever E, Leonard J B and Blum F D 1984 Biophys. J. 45 975-84
Foster K R, Epstein B R, Jenin P C and Mackay R A 1982a J. Col/. Interface Sci. 88 233-46
Foster K R and Schepps J L 1981 J. Microwave Power 16 107-19
Foster K R, Schepps J L and Epstein B R 1982b Bioelectromagnefics 3 29-43
Foster K R, Schepps J L and Schwan H P 1980 Biophys. J. 29 271-82
Foster K R, Schepps J L, Stoy R D and Schwan H P 1979 Phys. Med. Biol. 24 1177-87
Fricke H 1925 J. Gen. Physiol. 9 137-52
Fricke H and Morse S 1926 J. Cancer Res. I O 340-6
Frohlich H 1975 Proc.Narl Acad. Sci. U S A 72 4211-5
- 1980 Ado. Electron. Electron Phys. 53 85-152
Frohlich H and Kremer F (ed) 1983 Coherent Excitarions in Biological Systems (Berlin: Springer)
Fuhr G, Gimsa J and Glaser R 1985 Studia Biophvs. 108 149-64
Fuhr G, Glaser R and Hagedorn R 1986 Biophys. J. 49 395-402
Gabriel C, Grant E H, Tarta R, Brown T R, Gestblom B and Norland E 1987 Nature to be published
Gabriel C, Grant E H and Young I R 1986 J. Phys. E: Sci. Instrum. 19 843-6
Gabriel C, Sheppard R J and Grant E H 1983 Ph.vs. Med. Biol. 28 43-9

Gascoyne P R C, Pethig R and Szent-Gyorgyi A 1981 Proc. Natl Acad. Sei. U S A 78 261-5
Geddes L A and Baker L E 1967 Med. Biol. Eng. 5 271-93
Gekko K and Timashetf S N 1981 Biochemistry 20 4667-76
Gerber B R, Routledge L M and Takashima S 1972 J. Mol. Biol. 71 317-37
Glover I D, Moss D S, Tickle I J, Pitts J E, Haneef I , Wood S P and Blundell T L 1985 Adc. Biophys. 20 1-12
Could J L 1984 Ann. Rev. Physiol. 46 585-98
Grant E H 1957 J. Chem. Phys. 26 1575-7
- 1965 Ann. N Y Acad. Sci. 125 418-27
Grant E H, McClean V E R, Nightingale S R V, Sheppard R J and Chapman M J 1986 Bioelectromagnetics
7151-62
Grant E H,SheppardRJandSouthGP 1978 DielectricBehaoiour of BiologicalMolecules in Solution
(Oxford: Clarendon)
Grant F A 1958 J. Appl. Phys. 29 76-80
Grant J P and Spyrou N M 1985 J. Bioelec. 4 419-58
Grattarola M, Chiabrera A, Viviani R and Parodi G 1985 J. Bioelec. 4 211-25
Hackenbrock C R 1981 Trends Biochem. Sci. 6 151-4
Haggis G H, Buchanan T J and Hasted J B 1951 Nature 167 607-8
Hahn G M 1982 Hyperthermia and Cancer (New York: Plenum)
Hahn G M , Kernahan A, Martinez A, Pounds D and Prionas S 1980 Ann. N Y Acad. Sei. 335 327-46
Hallenga K, Grigera J R and Berendsen H J C 1980 J. Phys. L'hem. 84 2381-90
Hanai T, Haydon D A and Taylor J 1964 Proc. R. Soc. A 281 377-91
- 1965 J. Theor. Biol. 9 278-96
Hanss M 1966 Biopolymers 4 1035-42
Harris C M 1985 PhD thesis University of Wales
Harris C M and Kell D B 1985 Eur.Biophys. J. 13 11-24
Harris C M, Todd R W, Bungard S J, Lovitt R W, Morris J G and Kell D B 1987 Enzyme Microbial Techno/.
9 181-6
Hasted J B 1973 Aqueous Dielectrics (London: Chapman and Hall)
Haydon D A, Requena J and Urban B W 1980 J. Physiol. 309 229-45
Hazlewood C R, Chang D C, Nichols B L and Woessner D E 1974 Biophys. J. 14 583-606
Henderson R P and Webster J G 1978 IEEE Trans. Biomed. Eng. BME-25 250-4
Hill D W 1979 Non-inoasioe Physiological Measurements ed P Rolfe (New York: Academic) pp 95-112
Hobner R 1910 Arch. Gesamr. Physiol. Mensch 133 237-59
Hodgkin A L and Huxley A F 1952 J. Physiol. 117 500-44
lllinger K H (ed) 1981 Biological Efects of Non-ionising Radiation (Washington, DC: American Chemical
Society)
Jack J J B, Noble D and Tsien R W 1975 Electric Current Flow in Excitable Cells (Oxford: Clarendon)
Jaenicke R 1984 Angew. Chem. Inr. 23 395-413
Jakoubek B and Rohlicek V 1982 Physiol. Bohem. 31 143-9
Jonscher A K 1983 Dielectric Relaxation in Solids (London: Chelsea Dielectrics)
Kabsch W and Sander C 1983 Biopolymers 22 2577-637
Kell D B 1983 Bioelectrochem.Bioenerg. 11 405-15
-1986a Biosensors: Fundamentals and Applications ed A P F Turner, I Karube and G S Wilson (Oxford:
Oxford University Press) pp 429-70
- 1986b Collectiue Phenomena in Physicsand Biophysics ed T W Barrett ( S e w York: Plenum) in press
-1986c Merh. Enzymol. 127 538-57
Kell D B and Burns A 1986 J. Bioelec. 5 139-44
Kell D B and Harris C M 1985a Eur. Biophys. J. 12 181-97
- 1985b J. Bioelec. 4 317-48
Kell D B and Hitchens G D 1983 Coherent Excitations in Biological Systems ed H Frohlich and F Kremer
(Berlin: Springer) pp 178-98
Kell D B, Samworth C M, Todd R W, Bungard S J and Morris J G 1987 Studia Biophys. in press
Kell D B and Westerhoff H V 1985 OrganizedMultienzymeSystems:CatalyticProperties ed G R Welch
( S e w York: Academic) pp 63-139
Kirkwood J G 1932 J. Chem. Phys. 2 351-61
- 1939 J. Chem. Phys. 7 911-9
Klee M and Plonsey R 1974 IEEE Trans. Biomed. Eng. BME-21 452-60
Knott R B and Schoenborn B P 1986 Meth. Enzymol. 127 217-29
Konig H L, Krueger A P, Lane S and Sonnig W 1981 BiologicEfects of Encironmental Electromagneti,rrn

(Berlin: Springer)
Kosterich J D, Foster K R and Pollak S R 1983 IEEE Trans. Biomed. Eng. BME-30 81-6
Kraszewski A, Stuchly M A, Stuchly S S and Smith A M 1982 Bioelectromag. 3 421-32
Laogun A A, Sheppard R J and Grant E H 1984 Phys. Med. Biol. 29 519-24
Laver D R, Smith J R and Coster H G L 1984 Biochim. Biophys. Acta 772 1-9
Lee B and Richards F M 1971 J. Mol. Bid. 55 379-400
Lee C Y, McCammon J A and Rossky P J 1984 J. Chem. Phys. 80 4448-55
Lin J C 1975 IEEE Trans. Biomed. Eng. BME-22 74-6
Lovitt R W, Walter R P, Morris J G and Kell D B 1986 Appl.Microbiol.Biotechnol. 23 168-73
Mandel M 1977 Ann. N Y Acad. Sci. 74 303-1 1
Marconi L, Morgavi G, Ridella S and Rolando C 1985 J. Bioelec. 4 237-49
Marino A A and Ray J 1986 The Electric Wilderness (San Francisco, CA: University of San Francisco Press)
Marquis R E and Carstensen E L 1973 J. Bacteriol. 113 1198-206
Mohapatra S N 1981 Non-invasive Cardiouascular Monitoring by Electrical Impedance Technique (London:
Pitman)
Morgan H and Pethig R 1986 J. Chem. Soc.FaradayTrans. I 8 2 143-56
Neurnann E 1986 Modern Bioelectrochernistry ed F Gutmann and H Keyzer (New York: Plenum) pp 97-175
Nuccitelli R (ed) 1986 Ionic Currents in Development (New York: Liss)
Oncley J L1943 Proteins, AminoacidsandPeptides ed E J Cohn and J T Edsall (New York: Reinhold)
pp 543-68
Pande A 1975 Handbook ofMoisture Determination and Controlv01 2 (New York: Marcel Dekker) pp 267-344
Pauling L, Corey R B and Branson H R 1951 Proc. Natl Acad. Sci. U S A 37 205-1 1
Pauly H and Schwan H P 1964 IEEETrans.Biomed.Eng. BME-11 103-9
- 1966 Biophys. J. 6 621-39
Penney D C 1986 CRC Crit. Rev. Biomed. Eng. 13 227-81
Pethig R 1979 Dielectric and Elecrronic Properties of Biological Materials (Chichester: Wiley)
- 1984 IEEE Trans. Elect. Insul. El-l9 453-74
Pilla A A, Sechaud P and McLeod B R 1983 J. Bid. Phys. 11 51-8
Polk C and Postow E (ed) C R C Handbook ofBiological Efects ofElectromagnetic Fields (Boca Raton, FL:
CRC)
Po0 M-M 1981 Ann. Rev. Biophys. Bioeng. IO 245-76
Pool T B, Cameron I L, Smith N K R and Sparks R L 1981 f i e Transformed Cell ed I L Cameron and T
B Pool (New York: Academic)
Presman A S 1970 Electromagnetic Fields and Lije (New York: Plenum)
Pryce L R 1979 IEEE Trans. Nucl. Sci. NS-26 2736-9
Purdom L, Ambrose E J and Klein G 1958 Nature 181 1586-7
Pyper J W 1985 Anal. Chim. Acta 170 159-75
Rogers J A, Sheppard R J, Grant E H , Bleehen N M and Honess D J 1983 Er. J. Radio/. 56 335-8
Rosendal T 1945 Acta Physiol. Scand. 9 39-46
Rupley J A, Gratton E and Careri G 1983 Trends Biochern. Sci. 8 18-22
Sakmann B and Neher E 1983 Single-channel Recording (New York: Plenum)
- 1984 Ann. Rev. Physiol. 46 455-72
Salemme F R 1983 Prog. Biophys. Mol. Biol. 42 95-133
Salter D C 1979 Non-invasive Physiological Measurements v01 1 ed P Rolfe (New York: Academic) pp 21-68
Schanne D F and Ceretti E R P 1978 Impedance Measuremenrs in Biological Cells (New York: Wiley)
Schaubel M K and Habal M B 1969 J. Surg. Res. 9 513-20
-1970 Arch. Pathol. 90 41 1-5
Schepps J L and Foster K R 1980 Phys. Med. B i d . 25 1149-59
Schinkel J E, Downer N W and Rupley J A 1985 Biochemistry 24 352-66
Schliwa M, van Blerkom J and Porter K R 1981 Proc. Natl Acad. Sci. U S A 78 4329-33
Schwan H P 1954 Z. Nuturf: 9b 245-51
-1957 Adu. Biol. Med. Phys. 5 147-209
- 1963 PhysicalTechniques in BiologicalResearch v01 VIB, edWLNastuk(NewYork:Academic)
pp 323-407
- 1965a Ann. N Y Acad. Sci. 125 344-54
- 1965b Therapeutic Heat and Cold ed S Licht (Baltimore: Waverly) pp63-125
Schwan H P and Foster K R 1977 Biophys. J. 17 193-7
- 1980 Proc. I E E E 68 104-13
Schwan H P and Li K1953 Proc. I R E 41 1735-40

Schwan H P, Sheppard R J and Grant E H 1976 J. Chem. Phys. 64 2257-8
Schwartz D C and Cantor C R 1984 Cell 37 67-75
Sheppard A R and Eisenbud M (ed) 1977 Biological Efects of Electric and Magnetic Fields of Extreme/>,
Low Frequency (New York: New York University Press)
Sheppard R J and Grant E H 1974 Adv. Mol. Relax. Proc. 6 61-7
Sheridan R P, Levy R M and Salemme F R 1982 Proc. Narl Acad. Sci. U S A 79 4545-9
Singh B, Smith C W and Hughes R 1979 Med. Biol. Eng. Comput. 17 45-60
Smith S and Foster K R 1985 Phys. Med. Biol. 30 965-73
Somogyi B, Welch G R and Damjanovich S 1984 Biochirn. Biophys. Acta 768 81-112
South G P and Grant E H 1972 Proc. R. Soc. A 328 371-87
Sowers A E and Hackenbrock C R 1981 Proc. Narl Acad. Sci. U S A 78 6246-50
Storm F K (ed) 1983 Hyperthermia in Cancer (Boston: Hall)
Stoy R D, Foster K R and Schwan H P 1982 Phys. Med. Biol. 27 501-13
Stuchly M A, Kraszewski A, Stuchly S S and Smith A M 1982 Phys. Med. Biol. 27 927-36
Stuchly M A and Stuchly S S 1980 J. Microwave Power 15 19-26
Surowiec A, Stuchly S S and Swarup A 1985 Phys. Med. Biol. 30 1131-41
-1986 Bioelectromagnetics 7 31-43
Swicord M L and Davis C C 1982 Biopolymers 21 2453-60
- 1983 Bioelectromagnetics 4 21-42
Takashima S 1962 J. Polymer Sci. 56 257-63
- 1963 J. Mol. Biol. 7 455-8
_. 1966 J . Phvs. Chem. 70 1372-7
- 1967 Adv. Chem. Ser. 63 232-44
-1969 Physical Principles and Techniques ofprotein Chemistry part A, ed AJ Leach (New York: Academic)
pp 291-333
- 1976 J. Membr. B i d . 27 21-39
Takashima S , Gabriel C, Sheppard R J and Grant E H 1984 Biophys. J. 46 29-34
Tien H T 1974 BilayerLipidMembranes ( E L M ) . TheoryandPractice (New York: Marcel Dekker)
Tien H T and Diana A L 1968 Chem. Phys. Lipids 2 55-101
Townson H 1986 Lab. Equip.Digest 24 119-23
Wada A 1976 Adc. Biophys. 9 1-63
Weaver J C, Powell K T, Mintzer R A, Ling H and Sloan S R 1984a Bioelectrochem. Bioenerg. 12 393-404
Weaver J C, Powell K T, Mintzer R A, Sloan S R and Ling H 1984b Bioelectrochem. Bioenerg. 12 405-15
Welch G R ( e d ) 1986 7'he Fluctuating Enzyme (New York: Wiley)
Welch G R and Kell D B 1986 7'he Fluctuating Enzvme ed G R Welch (New York: Wiley) pp 451-92
Welch G R, Somogyi B and Damjanovich S 1982 Prog. Biophys. Mol. Biol. 39 109-46
Westerhoff HV, Melandri B A, Venturoli G, Azzone G F and Kell D B 1984 Biochim. Biophys. Acta 768 257-92
Wheeler H B and Penney C 1982 Non-invasive Diagnostic Techniques in Vascular Disease ed E F Bernstein
(St Louis: Mosby) pp 104-16
Wyman J 1934 J . Am. Chem. Soc. 56 536-44
Yamamoto T and Yamamoto Y 1976 Med. Biol. Eng. 14 494-9
Young R C, Ganellin C R, Graham M J and Grant E H 1982 Terrahedron 38 1493-7
Zimmermann U 1982 Biochim.Biophys.Acta 694 227-77

Pethig Kell pmb87 PDF

Uploaded by

Copyright:

Available Formats

Pethig Kell pmb87 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pethig Kell pmb87 PDF

Uploaded by

Copyright:

Available Formats

Phys. Med. Biol., 1987, Vol. 32, No 8, 933-970.

The passive electrical properties of biological systems:

This review was completed in February 1987

1. Introduction and scope

Historically, measurements of the electrical properties of biological materials hold

2. A summary of dielectric theory

complex function of the form (Debye 1929)

: 0 20 0.1 1 Log f (GHzl

By defining the magnitude of a dielectric dispersion (figure 2 ( a ) ) as

we obtain, by combination with equations (4) and ( 5 ) ,

in which fc is the relaxation frequency or 'characteristic frequency' ( f c= 1 / 2 m ) . The

Equation (10) may therefore be written in the form

E:. This Cole-Cole a appears in the modified Debye equation as follows:

3. Amino acids, peptides, proteins and DNA

3.1. Amino acids

sphererotating in anisotropicfluid.Quantitatively, no straightforwardtheoretical

3.2. Polyveptides and proteins

The contribution that an individual dipoleof moment m makes to the polarisability

minimum. These considerations themselves raise difficultiesof a philosophical nature,

Figure 5. The dielectricpropertiesofaqueousglobularproteins. ( A ) Bovine serumalbumin ( B S A ) at a

3.3. Deoxyribonucleicacid (DNA)

In its native configuration, D N A comprises two anti-parallel helical chains of nucleo-

4. Bound water in biological systems

Table 1. Dielectricdecrement values for some ions i n

Table 2. Relaxationfrequencyincrementvaluesforsomeions i n aqueous

H+, H,O+ 36.3 OH- 20

Table 4. Molar conductivity of some ions in dilute aqueous solution at 25 "C.

H+, H,O+ 350 OH- 193

Tissue water content Tissue water content

Bone 44-55 Muscle 73-78

6. Membranes and cells

In the case of a simple bilayer membrane, composed of a single type of phospholipid,

industrial fermentations, a particularly thorny and long-standing problem (Harris et

Figure 9. The flow ofcurrentatdifferentfrequenciesrelativetothecharacteristicfrequencyofthe p

Figure 12 also shows the frequency dependence of the permittivity of an aqueous

maybenotedthatskinpossesses a relativelyweak a dispersion(atleastdownto

Material 13.56 M H z 27.12 MHz 433 M H z 915 MHz 2.45 G H z Xeference

Artery - - - - 43 Brddy er a / i 19x1 l

Table 7. The conductivity ( S m") of biological tissues at 37 "C.

Material 13.56 MHz 27.12 MHz

Artery - - - - 1.85 Brady et a / ( 1981 )

spin-lattice relaxation times from that of the surrounding tissues (Hazlewood et a1

Gabriel C, Sheppard R J and Grant E H 1983 Ph.vs. Med. Biol. 28 43-9

Konig H L, Krueger A P, Lane S and Sonnig W 1981 BiologicEfects of Encironmental Electromagneti,rrn

Schwan H P and Li K1953 Proc. I R E 41 1735-40

You might also like