ELECTRONIC / PROTONIC CHARGE
CARRIER RATIOS IN SOLVATED
BIOMACROMOLECULES

Thesis for the Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
MICHAEL ROBERT POWELL
1969'

   
     

THE‘EU

  

LIBRARY ‘

Michigan State
University

 

This is to certify that the
thesis entitled
ELECTRONIC/PROTONIC CHARGE CARRIER
RATIOS IN SOLVATEU BIOI‘IACROEIOLECULLS
presented by

Michael Robert Powell

has been accepted towards fulfillment
of the requirements for

‘ 1311.1). degree inJinMcs

Wee

0.mi-iZ/Major professor

 

D m Rwafglgeg,

 

  
 

  

800K BRIBERY INC.

 

m .‘fi-‘v
abut-IV-..

 

 

ABSTRACT
ELECTRONIC/PROTONIC CHARGE CARRIER RATIOS IN
SOLVATED BIOMACROMOLECULES
By Michael Robert Powell

The change in the ratio of electronic to protonic
charge carriers in solvated systems has been investigated
for hemoglobin. cytochrome-c. lecithin, melanin. DNA, and
collagen. The electrical properties in the solid state
of these biomacromolecules follow the Operational defini-
tion of a semiconductbr, i.e.. the conductance is given by
the equation 0" = 0'6 eXp (-E/ZkT), where O" is the con-
ductivity, and E is the activation energy.

These compounds exhibit multilayer adsorption and
show a conductivity increase when adsorption occurs. The
adsorption process follows Roginsky-Zeldovich kinetics
which indicates that the water is penetrating into the
crystal; the kinetics of the current increase during the
hydration process is rapid (does not follow the Elovich
Equation).

The current is; hydration curves show a saturation of
the current after 2 to 2% BET monolayers have been adsorbed.
It is at this point that the capacitance of the samples
(hemoglobin. cytochrome-c, and lecithin) shows a further

rapid increase upon additional hydration.

glotin a:
:ared to
water.
Soli
errant e

t. .,

Potential

 

1:th hael KO DGI‘TZ I’OWGLL

The adsorption process can be eXpressed by either
the Bradley or the BET adsorption isotherm. The former
suggests the possibility of ordered (polarized) multilayers.

Tracer studies using tritiated water adsorbed on hemo-
globin and collagen showed small yields of tritium as com-
pared to the same number of coulombs passed through liquid
water.

Solid state electrolysis eXperiments determining the
current efficiency for hydrogen generation allows one to
determine the electronic/protonic charge carrier ratios.
The change in this ratio, for most materials, is found to
be linear with increasing hydration.

Electrolysis of hemoglobin at one hydration state of
28% water showed that protonic conductivity decreases at
potentials of less than 50 volts and is constant from 50
to 300 volts. Large deviations from Ohm's law are found
at applied potentials of less than 30 volts.

There does not appear to be a sharp onset of protonic
conduction at 2 BET monolayers (except for water on DNA
and methanol on hemoglobin), but rather this species of
carrier seems to be intrinsic in a manner similar to elec-

tronic conduction.

ELECTRONIC/PROTONIC CHARGE CARRIER RATIOS IN
SOLVATED BIOMACROMOLECULES

By
Michael Robert Powell

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Biophysics

1969

 

at no . c .. CC - .
.r e e C. C w."
Y” I l n- . Flv
Dy 8 ad" C n ... w ..
a: .l S nu 1.5..
av .f... I .. . . . Ab
.JA ,. a 3 in {In .C

ACKNOWLEDGEMENTS
I wish to eXpress my gratitude to Professor Barnett
Rosenberg whose patience and encouragment has aided in
this effort. I wish also to thank Drs. J. 0. Williams,
D. w. Boniface, and E. Postow for their many helpful
discussions. The financial assistance of the United States
Atomic Energy Commission (AT 11-1 -1714) is gratefully

acknowledged.

iii

TO MY WIFE AND PARENTS

I

V

TABLE OF CONTENTS

ACKE‘IO‘I‘JLEDGEIeIEI'I TS . . . . . . . . . . . . . . . . . . . i i 5.
LIST OF TABLES . . . . . . . . . . . . . . . . . . . Vi
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . vii
SECTION

I. INTRODUCTION. . . . . . . . . . . . . . . . 1

O O O O O O 1

Organic Semiconductors. . . .
Electrical Conductivity and Hydration . . 12
Adsorption Isotherms. . . . . . . . . . . 17

II. Tl‘EORY. O O O O O O O O O I O O O O O O O O 22

Electrical Conductivity . . . . . . . . . 22
Adsorption Isotherms. . . . . . . . . . . 25
Adsorption Kinetics o o o o o o o o o o o 28
Electrolysis Rate . . . . . . . . . . . . 29
III. EXPERII’EI‘T'PAL I'IETi'iODS . o o o o 0 o o o o O 31
Samples 0 o o o o o o o o 31

Conductivity Measurements
Adsorption Isotherms. . .
Adsorption Kinetics . . .
Dielectric Measurements .
Denaturation. . . . . . .

o o. o oo o o
o. o o. o o.
o. o o. o o.
.- o a. o o.
o. o oo o oo
o. o oo o o.
oo o o. o .0
o o o. o o. o

b)

) O\

Electrolysis EXperiments. 3O
Tracer EXperimemts. . . . 43
IV. EXPERIMENTAL RESULTS. 0 o o o o o o o o o o “7
Adsorption Studies. . . . . . . . . . . . 47
Kinetic Studies . . . . . . . . . . . . . 52
Conductivity Measurements . . . . . . . . 62
Resistance-Voltage Measurements . . . . . 74
DieleCtriC StUdieSo o o o o o o o o o o o 82
Electrolysis EXperiments. . . . . . . . . 86
Tracer EXperimentS. o o o o o o o o o o o 95

iv

V. DISCUSSION. 0 o o o o o o o o o o o o o o 0 98

Biological Systems and Semiconduction. . 98

Future Improvements in the Electrolysis
System. C O I O O O O I O O O O O O 104

Charge Carriers and Solvation. . . . . . 106

VI. BIBLIOGRAPHY. . O O O I O O C O O O O O O O 119

 

l.

O
A/h

Table
1.

LIST OF TABLES
Page

The BET monolayer coverage for the materials
tested in this work 0 a o o o o o o o o o o 57

The tritium yield for the electrolysis of
collagen and hemoglobin . . . . . . . . . . 96

vi

Figure

2.

3.
4.

10.

11.
12.
13.
1h.
15.
16.
17.

LIST OF FIGURES

The six adsorption isotherms for physical
and chemical adsorption. . . . . . . . . . .

Schematic of vacuum-microbalance apparatus .
Diagram of the electrolysis apparatus. . . .

Diagram of the apparatus used in the tritium
tracer eXpeI‘iment. I I I I I I I I I I I I I

Adsorption isotherm of water on melanin,
collagen. and salmon sperm DNA at 26.00 C. .

lecithin,
26.0 0 0.. .

on
at

Adsorption isotherm of water
cytochrome-c, and hemoglobin

lecithin,
26.00 c. . .

on
at

BET plot of water adsorption
cytochrome-c, and hemoglobin

collagen,

BET plot of water adsorption on
at 2600.. .

melanin, and salmon Sperm Na-DHA

Bradley plot of water adsorption
andleCithinIIIIIIOOII

Bradley plot of water adsorption
globin and cytochrome-c. . . . .

Bradley plot of water adsorption

Bradley plot of water adsorption

on collagen

on hemo-

on DNA. . .

on melanin.

Adsorption kinetics of water on hemoglobin .

Adsorption kinetiqsof water on collagen. . .

Adsorption kinetics of water on lecithin . .

Adsorption kinetics of water on melanin. . .

Conductivity of collagen as a function of

adsorbed water . . . . . . . . .

vii

Page

19
35
39

Uh

49

51

54

56

58

59
60

61
63
6n
65
66

67

e

r; . . -
14 C. u fla/ HIJ
le l ,0 li. H/~
II a

fin.

 

Figure

18.

19.

20.

21.

22.

23.

24.

25.
26.

27.

28.

29.

3o.
31.
32.
33.
3a.

35-

36-

Conductivity of melanin as a function of
adSOrbed water“. I I I I I I I I I I I I I I I

Conductivity of lecithin as a function of
adsorbed water. I I I I I I I I I I I I I I I

Conductivity of cytochrome-c as a function of
adsorbed water. . . . . . . . . . . . . . . .

Conductivity of hemoglobin as a function of
adsorbed water. . . . . . . . . . . . . . . .

Conductivity of DNA as a function of
adsorbedwa'ter................

Conductivity of hemoglobin as a function of
adsorbed methanol I I I I I I I I I I I I I I

Resistance-voltage plots for hydrated
leCithinI I I I I I I I I I I I I I I I I I I

Resistance-voltage plots for hydrated melanin

Resistance-voltage plots for hydrated
cytOChrome-CI I I I I I I I I I I I I I I I I

Resistance-voltage plots for hydrated
hemoglObinIIIIIIIOOOOOIIIOII

Resistance-voltage plots for methanol and
hemoglObinI I I I I I I I I I I I I I I I I I

Resistance-voltage plots for ethanol and
hemoglObinI I I I I I I I I I I I I I I I I I

Resistance-voltage plots for hydrated DNA . .
Capacitance Kg; hydration for hemoglobin. . .
Capacitance fig; hydration for cytochrome-c .
Capacitance ng hydration for lecithin. . . .

The yield of hydrogen as a function of
hydration for collagen strips«.-. . . . . . .

The yield of hydrogen as a function of
hydration for melanin . . . . . . . . . . . .

The yield of hydrogen as a function of
hydration for lecithin. . . . . . . . . . . .

viii

Page

68

69

7O

71

72

73

75
76

77

78

79

80
81

83
84

85

87

88

89

n I. .. a
58/ A4 I

A.‘

are
D

‘ I

to

ad

. HY

Figure

37-

38-

39-

40.

41.

42.

43.

Page

The yield of hydrogen as a function of
hydration for cytochrome-c. . . . . . . . . .

The yield of hydrogen as a function of
hydration for hemoglobin . . . . . . . . . .

The yield of hydrogen as a function of
Solvation (methanol) for hemoglobin . . . . .

The yield of hydrogen as a function of
hydration for DNA . . . . . . . . . . . . . .

The yield of hydrogen as a function of
voltage for one hydration state of hemoglobin

The resolution of the current increase upon
hydration into the ionic and the electronic
componentSIIIIIIOOIIIIIIIIII

The conductivity of the biomacromolecules

used in this study as a function of adsorbed
Water (in EAST). I I I I I I I I I I I I I I I

ix

9O

91

92

93

94

109

112

I
Cgpgw I A
.

I.
A-
presse:

similar

l‘.
”0'

f‘ .-
Jne‘r CE

ex:q¢

‘Vbafir

 

I. INTRODUCTION

Organic Semiconductors

Albert Szent-Gyorgyi (1941), taking up an idea ex-
pressed to him by one of his graduate students about the
similarity in order between organic and inorganic crystals,
proposed that electronic "energy bands" could be found in
biological compounds. This similarity was based upon the
periodic array found amongst the amino acids in the hydro-
gen-bonded crystals.

A method of energy transfer and charge transport
became possible which did not involve the diffusion of
molecules. This was particularly important for energy
transfer in solid structures such as the cytochromes of the
mitochondria.

The war intarupted the testing of this valuable idea
for several years. It was not until later that combinations
of proteins and dyes showed that energy migration could
take place in these photoactivated systems.

Not until 1949, when Evans and Gergley (1949) published
their calculations, was anything known about the possible

existance of these electronic energy bands. The calculation

 

2

was based upon the hydrogen-bonded repeat structure

I I IH‘N/
\c=o . . . .H-N\/ >C=O . . .
/ /C=O I I I IH-Pz\

and gave a band gap of 3 e.v. between the highest filled
and lowest empty band.

Measurements of the electrical conductivity of viol-
anthrone, iso-violanthrone, and pyranthrone by Akamatu gt
gl&(1950) showed that they were semiconductors with an
activation energy E of 3/4 to 1 electron volt, and ova-
lenen (Akamatu gt gl;, 1951) was found to have E= 1.13 e.v.
Bayliss (1948) calculated the Spectra of conjugated poly-
enes using the assumption that the 17 electrons were "free",
as in a metal, and he suggested a role for them in photo-
conduction.

An extensive study of organic semiconductiong compounds
has been carried out in the laboratory of D. D. Eley.
Pressed powders were studied by Eley gt g1;(1953). and
they observed that the resistances and activation energies
decreased with increasing compression; the presence of water
also lowered the activation energy. They did not find a
direct correlation between the number of 1T electrons and
the conductivity but did conclude that '1inear polyenes are
less effective semiconductors than polynuclear systems
with a similar number of IT electrons": they associated the

conductivity with the 7T electrons of the aromatic molecules.

Using an a.c. method Eley and Parfitt (1958) found that the

, F
cyanlne
mfl 1.1:

C 5 e.

\L)

174 e.

5- pic

(a.c.)

"

101' the

ES Uti

ductio:

 

 

activation energies were intermediate between the pressed
(80 kg./cm.2) and unpressed samples of metal-free pthalo-
cyanine and isodibenzanthrone. The values were, pthalo-
cyanine: 1.49 e.v. (a.c.), 2.39 e.v. (d.c., zero compression),
and 1.11 e.v. (d.c., 80 kg./cm.2); for isodibenzanthrone:
0.96 e.v. (a.c.), 1.54 e.v.(d.c., zero compression) and

0.74 e.v. (d.c., 80 kg./Cm.2). The radical N - a diphenyl
@— picryl hydrazyl gave an activation energy of 0.26 e.v.
(a.c.) and 1.49 e.v. (d.c., light compression). To account
for the low activation enengrof the radical, a band picture
was utilized where the odd electron was already in the con-
duction band so that E became the energy needed for inter-
molecular transfer.

The studies of hemoglobin by Cardew and Eley (1959)
showed an activation energy of 2.75 e.v. for dry hemoglobin
and 2.97 e.v. for dry globin(some dependence on pressure
was noted for both compounds). The activation energy
difference between compressed powders and single crystals
(a.c.) amounted to only 0.2 e.v. It was reasoned that the
charge carriers were electronic in the dry state because
(i) steady state values of the current were reached within
one minute, (ii) the effect of decomposition was to give a
lower conductivity -- higher conductivities would be ex-
pected if the carriers were ionic, and (iii) the (To values
were 103 to 105 times smaller than polyamides -- suSpected
protonic conductors. They agreed that adsorbed water would

induce a protonic conductivity in the solid proteins.

h

CO.

to

hug

n71

' I
.q‘lI‘av

‘
I

or de
h

 

SEER.
‘
D

de cn
artifac

“

O

QHI'VKY"
Sl

IJ'e
the si-

;:
un“‘

.
.1 fly.“
hv~-“.

-~I
an

if»

Studies of water adsorption on hemoglobin by Cardew
and Eley (1958) gave values for the BET monolayer which
differed for the freeze-dried and alcohol-denatured hemo-
globin. For native material, they calculated vm, the BET
monolayer, as 5.76 g/100 g. protein (300 C.) and 5.72 g/lOOg.
(40° C,), and for the denatured material they found 5.52
g./lOOg. (300 C.) and 5.48 g./100g. (40°C,). These values
supported the idea that the water was adsorbed by the polar
side chains on the surface of the molecule; about 73% of
the side chains were involved in the adsorption process.

Water was found by Eley and Spivey (1960 a) to greatly
increase the conductivity of hemoglobin; because of an
artifact, this increase was not as large as is generally
seen. It was postulated that water donated electrons into
the conduction band of the hemoglobin. Eley and Spivey
(1960 b) found that alcohol denaturation changed the value
of E from 2.36 e.v. for native hemoglobin to 2.88 e.v.
for denatured material: there was only a small change in
the value of 0’0.

The possibility of electronic carriers in slightly
hydrated hemoglobin (7.5% water) being dominant over protonic
carriers was tested by Rosenberg(1962 a) in an eXperiment
where a current decrease with time was checked for. The
absence of this decrease indicated that the water was not
being electrolysed, and therefore charge transport by the
water was negligible.

A study by Rosenberg (1962 b) of water and hemoglobin

1221*
lca‘

 

 

indicated that the water served only to reduce the activation
energy for charge carrier generation, and it did not affect
the value of O" .

Rosenberg (1964 a) made an attempt to measure the
amount of protonic charge carriers utilizing a traCer tech-
nique with tritiated water. At a hydration state of about
30%. he found 44% protonic conduction.

On the basis of the few measurements which had been
made, Rosenberg (1964 b) suggested that the charge carriers
were electronic in the range where the current increases
exponentially with hydration. In the region where satura-
tion of current occurs, the carriers would be mixed ionic
and electronic, and in the range where the amount of adsorbed
water is high, the carriers would be primarily protonic.

Eley and Leslie (1964) found that the electrical con-
ductivity for denatured hemoglobin was greater than native
hemoglobin for the same hydration state. They also found
that there was no saturation point in the current at the
first BET monolayer as they had earlier found; this finding
was in agreement with the finding of Rosenberg (1962 b).
They prOposed that the energy gap Ed decreases with hydra-

tion with the form
1 o 1
Ed = Ed " a. nd /3 (l)

where nd is the amount adsorbed, and the eXponent 1/3
is derived on the basis that the interaction energy between

impurity molecules is inversly proportional to the distance

.

ectrt

 

“O
‘
‘

”wee?”
m~re
e

bV'

%.
ab

aa

.iE‘.’ Ti"
do.‘
d.a or

O
Q»
l

.‘ .s
4%

between them. By plotting log conductivity Kg; (water
adsorbed)1/3, they obtained a straight line which showed
current saturation at high regain.

The adsorption of water on hemoglobin was found to
follow Roginsky-Zeldovich kinetics (Eley and Leslie, 1966).
This implies an increase of activation energy for adsorp-
tion with increasing value of regain. They eXplained this
by an electron injection mechanism where the water donates
electrons to the protein. Thus the adsorption follows the
process:

. . + . -
H20 + Protein ,_._’ (H20. . .Prote1n)vw_, (H20 —7 Protein )CT

and the rate-limiting step is the transition from the Van
der Waals state (VW) to the charge transfer state (CT).

They modified their original equation (Equation 1) such that
the exponent of nd was now unity (1).

Mariéié, Pifat and Pravdié (1964) attemped an electro-
lysis on hemoglobin at various hydration levels. Their
proceedure was to measure the increase of volume of the
electrolysis cell during the release of hydrogen (this was
accomplished by noting the movement of a slug of mercury in
a capillary). Their data indicated protonic conductivity
only at 18% hydration or above, but the scatter in their
data on the amount of protonic conductivity (20% to 94%)
made interpretation very difficult indeed.

The hydration of hemoglobin produces changes in the
dielectric constant. Brausse gt_g;; (1968) found that the

dielectric constant slowly increased with hydration, and at

anut 1
and her:
was atf
change

a n I
“'6‘ f
01. --

 

about 16% regain (2% BET monolayers) it increased rapidly
and became constant at about 35% regain. The diSpersion

was attributed to Maxwell-Wagner polarization. A large
change in the loss tangent at 15 % hydration was noted by
MariEié and Pifat (1966). Measurements of the dielectric
constant of hemoglobin with adsorbed water,ethanol, methanol,
and ammonia by Postow (1968) showed that the increase in
conductivity could be explained by the increase in the
dielectric constant according to the equation of Rosenberg
(1962 b).

Similar effects of capacitance increase with hydration
have been noted by Rosen (1963) on bovine serum albumin. He
noted a large change in the capacitance at about 20% regain.
Takashima and Schwan (1965) found similar effects by water
on the capacitance of ovalbumin; the large capacitance change
was seen at about 12% for this protein.

The porphyrin prosthetic group of hemoglobin, ferriheme,
was found by Cardew and Eley (1959) to have an activation
energy of 1.74 e.v., considerably below the value of 2.74
e.v. found by them for hemoglobin. A change of the central
metal atom effected only small changes in the activation
energy (Eley and Spivey, 1962).

The investigation of DNA was undertaken more from the
theoretical point of view of solid state biophysics than
for any definite implication in living systems, although
there may be some relation between conduction and carcin-

ogenesis or radiation damage. Lowdin has also produced

a theory to eXplain mutations based upon the movement of
protons from one base to another; these movements are gener-
ally of short distance, however.

Studies by Eley and Spivey (1962) on thymus DNA gave
an B= 2.42 e.v. for vacuum dried samples: the energy gap
for yeast RNA was 2.42 e.v. In that polarization effects
were small, they predicted that dry nucleic acids were
electronic semiconductors.

Eley and Leslie (1963) studied the activation energies
of nucleic acid componentsand found the following values
for E :(a) nucleoside ( base + ribose) 4.5 - 5.2 e.v.,

(b) nucleotide (base + ribose + phosphate) 2.0 - 2.2 e.v.
This last value is comparable to DNA and RNA, each of
which have an activation energy of about 2.4 e.v. They
suggested that the conduction state would correSpond to a
TT-7T* excitation ( or ne1T*), using a band theory
approach for DNA; for the nucleoside, aZU-7T** excitation
would be necessary.

Solid-state electrolysis eXperiments were made on
Na-DNA by MariEic’ and Pifat (1966) with the finding that
hydrations of 37% and higher gave very large values for the
amount of protonic conductivity. Unfortunately, the poor
degree of precision in the eXperimental values was not
amenable to a good interpretation of their results. It
was their contention, however, that at hydrations of less
than 37%. the amount of protonic conductivity was small or

even non-existant.

Ir,

 

 

Burnel g3,gl;(1969)reéxamined the semiconduction para-
meters of the nucleic acids and found an E= 2.36 e.v. and
a log 6'0 of 2.35 for DNA and RNA. They also found that Poly-
A and Poly-C had dry-state activation energies of 2.2 and
2.4 e.v., reSpectively. The similarity between the elec-
trical prOperties of these two compounds and DNA has an
important consequence. Both Poly-A and Poly-C retain their
crystalline character even when desicated as opposed to DNA
which becomes disordered upon drying.Thus the similarity
of semiconductivity properties of dried material suggests
that base-stacking is of a small consequence for electronic
conduction.

It was postulated,then, that the activation energy for
charge carrier generation represented the energy needed to
remove an electron from the bases and place it on the
phoSphate of the backbone: the stacked bases would then be
hole conductors and the backbone would be an electronic
conductor. The excited state was then an optically inactive,
charge-transfer state.

J. Ladik (1968) believes that the high frequency
activation energy (0.2 e.v.) found by O'Konski gt gl;(l964)
represents the energy gap for a single DNA molecule, and
the higher d.c. values of E = 2.34 e.v. represent the inter-
molecular energy barrier. Ladick calculates a band gap of
about 3.7 e.v. ,and he concludes that DNA is an inhomogen-
eous impurity semiconductor.

A-form DNA contains large amounts of water (Falk, at

w., 1663 2

’

than. 0.7.
of water a:
the hand 5‘.
water.

A ot‘lat

"rrent in;
The differ
atsence of

Studi
Easer and
in a hydra
0 to 230 v
"as ohmic,
photocOnd,‘
that dry I
Samples 1

cond‘dctim

Eley

(“J

= 2.60

lO

g1;, 1663 a, b)and is stable only at values of p/po greater
than 0.7. Calculations of Sklenar. Ladik.and Biczo (1967)
of water and Poly G-C showed only minor modifications of
the band structure with respect to systems not containing
water.

Making use of proton-injecting electrodes (hydrogen-
saturated palladium), Thomas gt_gl; (1969) found large
current increases with the pyrimidine base isocytosine.

The difference between the current in the presence and the
absence of hydrogen was greater than 1010.

Studies of the d.c. conductivity of Na-DNA by O'Konski,
Moser and Shirai (1964) showed regions of non-ohmic behavior
in a hydrated sample (p/po = 0.31) with potentials of from

1

0 to 200 v. cm.’ at 25° 0.: at larger potentials the current

was ohmic. From various observations (the effect of 02,
photoconduction, and space-charge effects), they concluded
that dry DNA was an electronic conductor. and that hydrated
samples likely. possessed a large complement of ionic
conduction.

Eley and Spivey (1960) measured the semiconduction
activation parameters for dry cytochrome-c and found an
E = 2.60 e.v. and a log 0'0 = 4.8: these values being
slightly smaller than those which they found for hemoglobin
(cf. E = 2.66 e.v. and log 0'0 = 5.0). In a manner similar
to hemoglobin, cytochrome-c displayed ohmic behavior up to

field strengths of 2.000 V. cm.’1 at 1150 C. in the dry

state.

fer, the
Quantum.-
nt’dTL. .811

there is

11

Cytochrome-c is of interest to the biologist by
virtue of its presence in the oxidative phosphorylation
chain. Furthermore, this material appears to be bound to
the cristae of the mitochondria making electron transfer
by diffusion of the cytochrome to the substrate a rather
improbable event.

Two possible mechanismsexist for this electron trans-
fer, the first being semiconduction and the second being
quantum-mechanical tunnelling. although in cases where
"tunnelling" is the rate limiting step in semiconduction.
there is little difference between the two. This latter
case might be seen in inter-molecular transfer of charge.

At present. there remains a difference of opinion amongst
workers in the field as to which is the correct method (if
there is indeed a difference), and no irrefragable evidence
exists for either side. The present investigation was not
undertake to resolve the controversy, but rather to show
that, at high degrees of hydration and consequently lower
resistances, the component of electronic conductivity was
large and could conceivably be used to eXplain the transfer
of electrons. In short, it lends support to, but not proof
of, electronic semiconduction in cellular organelles.

A large number of enzymes are present in living organisms
which contain transition metal ions. Cytochrome-c. an iron-
containing molecule, is one example. A rewiew of this topic
has been prepared by R. J. P. Williams (1968).

A calculation by Cardew and Eley (1959) showed that the

c ytochro'
arrived a'
possible '

m‘rin e;

12

resistance of cytochrome-c was too large to adequately
account for the observed resipiration rate in the sea
urchin egg. They used, however. the dry state activation
energy. and from this calculation, they found a discrepancy
of 1016 between theory and eXperiment. Using the hydrated
crytochrome-c resistances, Rosenberg and Postow (1969)
arrived at the conclusiaithat semiconduction was indeed a
possible mechanism to eXplain respiration in the sea
urchin egg.
At present, there is some evidence that the movement
of electrons may proceed by a fairly long range tunnelling
process (about 60 R). DeVault, Parkes, and Chance (1967)
showed that the oxidation of cytochrome-c in the purple
sulphur bacterium could occur at liquid helium temperatures
with a time constant of 2.3 msec. and no indication of an
activation energy at temperatures of less than 1200 K.
Measurements of conductivity and dielectric prOperties
by Taylor (1960) lead him to the conclusion that conductivity
was mediated either by anions or by protons (in the dry state)

in cytochrome-c.

Electrical Conductivity and Hydration

The history of the conductivity increase in hydrated
biomacromolecules is rather extensive, and it can be traced
back to the investigations of Evershed (1914) who studied
the effect of moisture on materials which were used to
insulate electrical wires, such as cotton. He came to the

conclusion that practically all of the current increase was

effected

srfaces
Eur:

increase

{A “
A Ju. .1 thz

£1 for‘

‘4‘

at for

13

effected by water condensation on the internal and external
surfaces of the fiber.

Murphy and Walker (1928) studied the d.c. conductivity
increase of cotton, silk, and wool with adsorbed water and

found that the resistance for cotton was given by

log R = -9.3 log M + B (2)
and for silk by

log R = -16.0 log M + C (3)
and for wool

log R = -l6.4 log M + D (4)

It was conjectured that the water flowed in water channels,
and the fiber resistance depended upon the number of ‘re-
strictions in these channels; the number of restrictions
was reduced with the increase of adsorbed water.

Murphy (1929 a) studied the a.c. conductiviy of cotton
and silk and found that the parallel conductance increased
with frequency. but it became frequency independent at high
humidities. He found that the a.c. conductivity could be

represented by

C)
II
C)
Pb

o n (5)

where Go and n are constants.and f is the frequency;
n was found to be an approximately linear function of the

humidity. In a later paper, Murphy (1929 b) discussed the

material

clusicn

l4

Evershed effect, i.e.. the decrease of resistance of hydrated
materials with increasing applied potential. with the con-
clusion that it was caused by a combination of a back-emf
at low applied potentials and the Wein effect at higher
potentials.

Marsh and Earp (1933), working with single wool fibers,
found a decrease of resistance with increasing hydration,
but the currents were found to be ohmic with no evidence
of polarization. It was their opinion that the current was
carried in water filled pores in the fiber.

Baxter (1934), however, concluded that the water in
the pores could not be of the same form as water in the
liquid state in that the activation energy for conduction
was different in water than in wool. He obtained the

following equation for the conductance:

6': A eXp (~E/ kT) (6)

where 6‘ is the conductivity. and A is a constant. He
proposed that conduction was the result of electron
tunnelling between theadsorbedvmter molecules. The
resistance would then vary eXponentially with the water-
water spacing in that the probability of electron tunnelling
from one water molecule to another is proportional to r ,

the width of the barrier, and this probability is given by
P = B eXp (hr) (7)

where B and h are constants. If one assumes a uniform

' + i'rsw“ .
dis wriuiul

15
4/3,

distribution of water, then r is proportion to m
where m is the amount of adsorbed water. Thus the

resistance could be given by the equation

R = 0 exp (k m'1/3) (8)

where C and k are constants.

Fuoss pr0posed an ionic conduction mechanism based
upon the idea that plasticizing agents facilitate ion
migration. This idea was tested by King and Medley
(1949 a) with the conclusion that substances such as
water and formic acid not only lower the diffusion resis-
tance of the ions in the polymer matrix, but that increased
dissociation of the ions also occurs. They electrolysed
a keratin-water sample and found that,at 18% adsorbed water,
the yield of hydrogen was 92% of the theoretical amount.
From this. King and Medley (1949 b) concluded that electro-
lysis was taking place as was predicted by the ionic theory.
albeit, the amount of oxygen found was only 185 of the
theoretical amount.

E. J. Murphy (1960) used a variation Of the theory
proposed earlier concerning the completion of water channels.
In the revised picture. the water was adsorbed on specific
sites, and these water adsorption sites were prOposed to
lie between "ion-generating‘ sites". These latter were
places where the dissociation energy ,or the ‘ionlzation
energy,for the formation of H+ was less than for the

. + - . .
formation of H30 and 0H from water. The ion-generation

‘3
r0-
:1"

D
U)

occu; ie-i
aisorcti
PTO'Ea‘sil

ENG)

16

sites consisted of protons which had exchanged for metal
ions in an ion-exchange process. The anion to which the
proton was bound was a part of the macromolecude (cellulose
in the Specific case treated by Murphy).

The water molecules were envisioned to connect the
easily ionized protons and thus provide a conduction
pathway. If the amount of water adsorbed at saturation is

CXO . and the amount of water present at any given hydration
state is OK . then the probability of any site being
occupied is ( O! /O( 0). If there are n sites for water
adsorption between any two ion-generating sites, then the
probability that all of the sites will be filled is given
by ( 0! /oc o)n.

The conductivity is then calculated empirically from
this probability. The conductivity at saturation is given
by 0'; . Thus, for any state less than saturation, the

conductivity is given by
0‘: 0—8 (a/qo)n. (9)

The theory is applicable only to protonic conductors. in
so far as the idea of ion-generating sites is utilized.
However, in the case of cellulose, to which Murphy
directed his theory, the primary conductors were found by
him to indeed be protons (Murphy 1963, 1965).

Several studies by Eley have led him to the conclusion
that the adsorbed water increases the conductivity of

proteins by injecting either electrons or holes into the

cgnd‘dCtil
the 67.81"?
general 4
for borc:
over the
in Squat

gap woul

Accordir.

then 'C e c

o:

01‘

17

conduction bands of the material. As discussed earlier.
the energy gap is decreased according to Equation 1. This
general equation was found by Pearson and Bardeen (1959)
for boron doped silicon. If the water is distributed

over the surface of the spherical molecules, the eXponent
in Equation 1 would be i, and the equation of the energy

gap would then be

ED = EDo - 0K nD‘g (10)

According to Eley, the total equation for the conductivity

then becomes

 

 

 

0’: ep(2nD)% (Zykai/u eXp - (ED -(Xan)
hB/Z 02kT (11)
or
_ i x
log d'- Constant + log nD -+P1%) (12)

where x is a function of the distribution of the phya
Sically adsorbed water on the molecule and of the forces
between the adsorbate molecules; in the case of inorganic
compounds, x is equal to 1/3.

Another theory proposed by Rosenberg (1962 b). based
upon the change of the dielectric constant of the conducting

medium will be discussed in the next chapter.

Adsorption Isotherms
The adsorption of gases on solids is a large problem

in itself, but for this study. it was used only as a tool

, r

18

and not as an end in itself. Much discussion in the past
has centered around the change in the carge carriers as the
amount of adsorbate increases: this change was often thought
to occur in the hydration range of from two to three Brunauer-
Emmett-Teller (BET) monolayers. This prediction was inves-
tigated as a part of this study.

Brunauer (1943) has distinguished several different
types of adsorption isotherms. and these are shown in Figure 1.
The first is Langmuir or monolayer adsorption and is known

as Type I. It follows the equation
v = (kbp)/(1 * bp) (13)

where v is the amount of gas adsorbed , p is the pressure,
and k and b are constants Specific to the system under
investigation. An example of this type is the adsorption
of 02 or N2 on carbon or silica gel at low temperatures.
Type II isotherms are commonly referred to as BET or
multilayer adsorption. In this case, the adsorbate adsorbs
first on the primary sites, and this is later followed by
the buildup of molecule upon molecule until several layers
result. These isotherms are a very common case of physical

adsorption, and follow the equation

v (1”)

 

= CQVHL
(pO-p) [1 + (c-l)(p7po)]

where v is the amount adsorbed, p is the pressure, pO
is the saturation vapor pressure, vm is the amount of

adsorption for the first monolayer, and c is a constant

.cofiuQHOmpm HmoflEmno paw HmUHmanm How mEHmnuOma coflDQMOmpm xam one .H ousmflm

mesmmmcd

<3
0.
C)
O.

 

----—————J

 

 

““"""""1&

———-————----q

 

 

 

 

 

 

 

 

 

A
W
_ m
1 _
V
_ . _
_ _ . w.
_ _ _ 0
_ _ _ J
_ _ _ 0.
_ - _ a
. _ _ D.
. u _
_ . _
- P
_ u u
m HH 3»... m HH 2;... u H on»...
- P —

 

given 33’

where 3.
layer, a:

"1 .
‘ 1

1
N.
\r

'd

forces w?

than be tw

20

given by the equation
C = el'{p[(E1 -' L)/ RT] (15)

where E is the heat of adsorption of the first mono-

1
layer, and L is the heat of vaporization.

Type III are rare and are the result of molecular
forces which are stronger between the adsorbate molecules
thanbetween the adsorbate and the adsorbent. That is to
say, E1< L. An example of such a system is bromine on
silica gel at 790 C. or water on graphite.

Capillary condensation is thought to be responsible
for Type IV isotherms. In this type of system, a signifi—
cant amount of vapor is believed to condense in small
pores. When these pores fill. saturation is reached.

In Type IV isotherms, E1 is greater than L.

Similar to Type IV, the Type V isotherm is the result
of capillary condensation. The distinction between the
two is the result of the difference between the binding
force between adsorbent and adsorbate such that for Type V,
E141 L.

Type VI is characteristic of chemisorption processes,
that is, those instances where covalent or ionic bonds are
formed as contrasted with physical adsorption where only
weak forces are involved. In chemisorption. the heat of
adsorption is large as compared with physical adsorption.

By an appropriate arrangement of the BET equation, it

is possible to graph the eXperimental results and obtain

21

vm , the monolayer coverage.

Adsorption isotherms for very polar adsorbates may
also be calculated by means of the equation of Bradley
(1936). It is a modification of a theory originally
pr0posed by DeBoer and Zwicker (1929) in which the more
polar surface causes an induced dipole to be formed in
the adsorbed layer. This layer polarizes the layer above
it g3 cetera.

The equation, if lacking certain theoretical veri-
similitude, is a valuable empirical tool for describing
the eXperimental results over values of p/pO from 0.05
to 0.9. This is in contrast to the BET theory which is
linear only from p/pO of 0.05 to 0.5 in general.

The range of the Bradley equation, more than Specific
molecular or physical information about the adsorption
process,was the reason for its use. This was true for the
methanol-hemoglobin system where mechanical problems with
the balance made it impossible to solvate the sample at
a p/p0 of greater than 0.75. For all other studies, the
BET equation, with its value of vm, proved to be the most

valuable.

Electr:

R3
electr‘
the pre
be inc:
needed

w'r, s'
”ich .
c

II. THEORY

Electrical Conductivity

Rosenberg (1962 b) has pr0posed a theory of the
electrical conductivity increase upon hydration using
the premise that the effective dielectric constant will
be increased. This will lead to a reduction in the energy
needed to separate the positive and negative charges
which account for the current.

Many biomacromolecules follow the operational defi-
nition of a semiconductor, that is, they follow the
equation

0' = GB eXp (-E/ 2kT) (16)

Their conductivity increases with increasing temperature.
This is contrasted with metallic conductors which show
a resistance increase with increasing temperature. The
number 2 in the denominator of Equation 16 is customary
in the field of organic semiconductors and is the legacy
of band theory. While there is no proof that energy bands
exist in these organic compounds as they do in inorganic
semiconductors, the formalism of band theory is still
followed, and the "2" persists.

Of interest is the fact that organic semiconductors
change their conductivity upon hydration; this change can

22

he reCT'

where

coestan

Ah-

placir

90nd 1

no
(7)

23

be represented by the equation

O—m = O‘Dry exp (0: m) (17)

where m is the amount of water adsorbed, and 11 is a
constant. Rosenberg (1962 b) has shown that 0‘0 does
not change upon hydration, but instead it is E which is
a function of hydration. It is therefore possible to

combine Equations 16 and 17 and obtain
O"(T,m) = 0'0 eXp[(-ED / 2kT) + OLm] (18)

where ED is the dry state activation energy. The

activation energy E is then

E = ED - 2kTOkm (19)

The question is how to calculate E , the activation
energy of the solvated system, using a phenomenological
approach. without recourse to¢Xm , so that one accrues a
better physical picture of the processes involved.

If one desires to move free changes in the presence
of an applied field, the first prerequisite is the gener-
ation of the charges themselves. This process can be thought
of as simply removing an electron from the molecule and
placing at a point where the coulombic attraction is now
less than kT. The energy for this process would corres-
pond to the ionization potential of the molecule. If the
charge is placed on another molecule, extra energy, the

"electron affinity”, is returned. In addition. the lattice

.v- r
.

For the
this he‘d

24

will relax in such a fashion that the dipoles will orient

themselves to further lower the energy of the system; this
energy is termed the "polarization energy", The polariza-
tion energy is multiplied by two in that both the negative
and positive charges are included. The energy for the

total process in the dry state would be

E=I-A-2P 20
D g g ()

where the ionization potential, Lg, and the electron
affinity, Agg are taken as the gas phase values. This
polarization energy is comparable to the orientation of
solventmolecules around a dissolved ionic compound.

The amount of the polarization energy for Spherical

symmetry is given by
2
P=(e/2R)(1-1/K) (21)

where e is the electron charge, R is the radius of
polarization, and VK is the dielectric constant of the
medium. The actual value of VC can not be calculated
without a knowledge of the positions of all of the dipoles
involved, so the bulk dielectric constant is used as a first
approximation. If Equations 20 and 21 are combined. the

following equation results:

.. __ _ 2 .-
ED _ 1g Ag (e /R) (1 1/u) (22)

For the hydrated systems, the value of V1 will change;

this new value is represented by' K'. The gas phase

aaaa

and ’r

-‘

5(T.k'

.
Miser“-

&

Covera;

‘3

MEET}. f

 

25

ionization energy and electron affinity are not functions
of hydration. The new energy for hydrated systems is then

_ _ 2 _ .
E— Ig’Ag (e /R) (1 l/K) (23)

Combining Equations 22 and 23 yields
a 4 2 ,
r. = ED - (e /R>[(1/k> - <1/x )] (21+)
and from Equations 19 and 24 we have
_ 2 .
2mm - (e /R)[(1/u) - (1/w )] (25)
Substituting (25) into (19) gives the final result

0(T,k') = 0'2) eXp(-ED/2kT). eXp{(e2/2kTR)[(1/\A) - (1/k')]}
(26)

Adsorption Isotherms

These were used only to determine the monolayer
coverage, but a few words, at least, Should be said about
them. Specifically, the two systems used to analyse the
data were the BET and the Bradley isotherms.

In the BET theory, it is assumed that a Specific heat
of adsorption exists for the first monolayer; all subse-
quent layers will yield a heat equal to the heat of
vaporization of the compound being adsorbed.

The BET model assumes that there exist particular
adsorption Sites, and, at any one time, S0 of these will
be free, 81 will have one molecule adsorbed, 82 will
have two molecules adsorbed and so on. When the system is

in a state of equilibrium, a certain probability of coverage

will exist on each Site. For the first monolayer, the

av

w’°”

r
A...
“u.

AJ
C»
(F;

n:
a

”a?

O
‘
‘

F.

83‘."

ii

Onlfi

h

in w

lne w

l

err»

t“ i.
1e,

.

‘vl

m0?!

ts

26
ratio of covered to uncovered Sites will be
so/s1 = (bl/a1p> exp <-nl/ RT) (27)

where a1 is a constant from kinetic theory, p is the
vapor pressure, b1 is a constant (a function of the

frequency of molecular oscillation), E1 is the heat of
adsorption of the first monolayer, and R and T are the
gas law constant and the absolute temperature, respectively.

For each subsequent layer, the adsorption is
8.14/81 = (bi/aip) eXp (-bi/ RT) (28)

If x represents the total amount of vapor adsorbed for

any given pressure p , and x is the monolayer coverage,

 

m
it is possible to Show that
_ CM
x/xm - (1. h (1. y + Cy) (29>
where C is given by
c = eXp.[(E1 - L)/ RT ] (30)

in which L is the heat of vaporization, and y is p/po.

This yields the familiar eXpression for the BET isotherm

1) -_-_1_ ______..C-1 B
X(pO-p7 x0 +[ ’Snc ]po, (31)

m
Thus a plot of p/ x(po - p) vs. p/po yields a straight

 

line whose Slope is (C - 1)/me and whose intercept is
1/me. From this the monolayer coverage xm may be
determined in that there are two equations for the two

unknows (xm and C).

VN‘Here

35.801“!

and

27

The isotherm develOped by Bradley has its main use
in cases where the adsorbate possesses a large permanent
dipole moment. This is true for water and methanol (1.85
and 1.70 Debye units, reSpectively), although gases such as
nitrogen and argon (not tested here) are not believed to
follow this theory so closely. This isotherm was used
here to predict the solvation state of the methanol—hemo-
globin system.

The derivation of the equation is not given here in
that the theory on which it was developed was not correct
for the general case (of. Brunauer gt al;, 1938). Thus
the Bradley equation is mainly empirical. In its general

form, the isotherm is described by the equation

A

2
3 (3 )

T = K K
10810 (pO/p) 1

where T, p, and p0 have their usual significance in

adsorption theory, and K1 and K A are defined by

2 2 2 6
_ NE ;_ _ u - k (1 -Ak )
K1 ’ 2.3R [2a a131 [(1 - 2k2;2][ k‘ ] (33)

 

 

k 2
K2 =[n - 1.2)] (3”)
where N is Avogadro's number, p is the dipole induced by
the crystal in the first layer of adsorbed vapor, R is
the gas constant, k is a constant of zero dimensions,
a is the polarizability of the adsorbed molecule. and a1
is the Spacing of the dipoles of the adsorbent. The

exponent A in the K3 term is the amount of adsorbed gas.

28

The term K is equal to bi, where b is a constant.

3
For use, Equation 32 is put into a log form and A is
plotted against log log (pO/p), and the result is a

straight line.

Adsorption Kinetics

The adsorption of gases on solids may proceed at a
fast rate which indicates surface adsorption on a crystal.
or it may proceed at a much Slower rate indicating penetra-
tion into the crystal. The adsorption of water vapor onto
proteins is of the latter type, as has been shown for one
protein, hemoglobin, by Eley and Leslie (1966). This study
extended the observations to other systems with the same
conclusions.

Studies of gas adsorption, particularly inorganic
compounds, has Shown that the process can not be described
by simple maagaction laws. The first investigators to
study the problem quantitatively were Roginsky and Zeldo-
vich (1934); they worked with the carbon monoxide - manganese
dioxide system. They found that the kinetics could be

represented by the equation

dq/dt = K exp (-bq) (35)

where q is the amount of gas adsorbed, and K and b are
constants. The equation states that the rate of adsorption
possesses an activation energy which increases as the amount
of the gas q adsorbed increases. This could be eXpressed

as

dQ/dt = K exp (-O(Q/ RT) (36)

 

,. +1
ecdac.
“e

.

W29

u v y are . .. I! r

+|U CU 1M“ V x d “HIM Hm.‘ ‘% 5‘. Q. \YJ'I Ob U§ Qt» A a
an AA,“ HI“ VJ; 0. s. 0‘ ‘ X \l‘ 3% he firu “I; QiU MUs

.N a .I O a k e 3: e .3 l 0 e 0 .1 r .

29

where 0(/ RT is the constant b in Equation 35. This
equation is integrated to eXpress q as a function of

time, 111;.
q = (R'MO [1n{ t + (RT/qk')} - 1n (RT/Ozk')] (37)
where k' is a constant. This can also be written
q = (RT/CC) ln (1: + to) + Constant (38)

Plotting q against ln (t + to) Should give a straight
line: tO must be chosen empirically. If the value of tO
is too small, the line will be convex to the ln (t + to)
axis, and if tO is too large, the curve will be concave.
The number of adsorption Sites alone is the limiting
factor in the rate of gas adsorption: the amount of gas
only determines the initial rate of the reaction. Taylor
and Thom (1952) eXplain the kinetics by a mechanism of
bimolecular site-Site interactions. This kinetic law is
given a different interpretation by Eley; his view was

eXplained in the Introduction of this paper.

Electrolysis Rate

The evolution of hydrogen was calculated from Faraday's
law of electrolysis, i.e., one Faraday of charge produces
one equivalent of compound; diatomic hydrogen contains two
equivalents so two Faradays are needed. From the fact that
one ampere equals one coulomb/second, and that one Faraday
is equal to 96,500 coulombs, one can calculate that 1 uamp.

produces 5.18 x 10"12 (moles HZ/sec.).

'11»

4M
"-4. .

sun-r

‘rr

so tia'

T.

as.
3.“ u
e.

[V

..‘+

'u

v- '0
..--_‘- U
a

30

Using the equation PV = nRT , one can now calculate
the pressure change since n , in (moles HZ/see), is now
known from Faraday's law. The system must be calibrated
so that the volume V is known, of course.

The system used here had a volume of 422 cm.3 so that
the generation of hydrogen correSponded to 1.34 x 10-
millitorr/flamp-min. The apparatus, when well out-gassed,

9

had a sensitivity of about 10- moles of hydrogen.

Samples

III. EXPERIMENTAL METHODS

The materials used in this study were obtained

commercially, with the exception of melanin, and they

were not further purified. Each compound is described in

more detail below.

(i)

(ii)

(iii)

(iV)

Hemoglobin was purchased from Gallard-
Schlesinger and processed by Servac: the
material was dialysed and twice recrystall-
ized.

Cytochrome-c (horse heart) was obtained from
the Sigma Chemical Company and was free of
ammonium sulphate and sodium chloride. It
was primarily in the oxidized form.

Collagen was obtained in the form of strips
approximately 0.2 cm. wide and 5 x 10“3 cm.
thick. The material was a special metal-free
preparation from the Ethicon Corporation and
was extracted from bovine Achilles tendon.
Experimental cells of this material consisted
of several strips Side by Side.

Melanin was prepared by an auto-oxidation

polymerization of DOPA (dihydroxyphenylalanine).

This material was extensively dialyzed. and

31

32

it is the highest purity obtainable, and
was prepared by Dr. Elliot Postow.

(v) Lecithin was purchased from Nutritional
Biochemicals as synthetic B,Y’-dipalmitoyl-
DL- 0( -lecithin .

(vi) Deoxyribonucblc acid was purchased from the
Sigma Chemical Company in the form of the
sodium salt. It was extracted from salmon

Sperm.

Conductivity Experiments

Conductivity measurements were made on samples pressed
into thin pellets. The die was pressed with the palm of
the hand so that the pressure would not be great enough
to denature the sample. These small pressures were found
to be sufficient to provide samples with sufficient mechanical
strength. The pellets were approximately one millimeter
thick and were placed between two platinium foil electrodes.
These were affixed to a Teflon block with screws. Before
assembling the cell, the block was washed sequentially with
ethanol, water. and again with ethanol: it was then only
handled with forceps. The area of the electrodes was about
0.25 cm.2. This sandwich cell was placed in a glass tube
in the vacuum line. Electrical connections to the outside
were made with tungsten feed-throughs sealed in quartz to
minimize electrical leakage.

The potentials were varied from 1 volt applied (10

volts/cm.) to 300 volts applied (3.000 volts.cm.). All

D 9. a
.r . LUV nu .filt
2w .

in A u

a n

50'

33

measurements were d.c. In the case of some highly

hydrated samples, polarization effects were large at the
higher voltages, and it was necessary to follow the electro-
meter readings with a strip chart recorder to determine

the ”equilibrium current" - this often required 3 hours.

Adsorption Isotherms

Powder samples were placed in the Cahn Vacuum Micro-
balance (Model RG) which was counterweighted so that the
sensitivity was increased. The hang-down tube of the
balance was thermostated by circulating water in an outer
jacket. The temperature was kept at 26.00 i 0.1 0C.

All samples were heated prior to the start of the
eXperimentS with either a heating tape or a heat lamp to
determine the dry weight. The temperature was about 750C.
At this temperature, the sample could be brought to constant
weight (less than a change of 0.02%) within two hours. On
some samples, pyrolysis would occur if heated to higher
temperatures - this was very true of collagen.

While the samples were being heated, the system was
evacuated to a pressure of approximately 5 millitorr as
measured on an ionization gauge. While the system was
being pumped on, the traps were always cooled with dry ice
to prevent back streaming of oil which could coat the
samples and affect the adsorption.

Water vapor was introduced from Side tubes containing
vacuum degassed water. The vapor pressure was determined

by means of a Dubrovin ("cartesan diver") gauge which had

water

W88

to de

1 a u
.l t.
nu.

C. 0
e C.»

W e

34

a least count of 0.1 torr and a range of from 0.0 to 10
torr. Above this limit, the temperature of the water in
the side tube was varied to control the pressure. The
water temperature could be measured to i 0.1 C.0 and the
vapor pressure could be found to three significant figures
from the Handbook g: Chemistry and Physics. This proceedure
was found to be superior to the measurement of pressure
directly with the mercury manometer as this device could
only be read to i 1.0 torr.

The Cahn BC was connected to a strip-chart recorder
to determine more easily the adsorption kinetics and the
equilibrium weight. The balance pans were counter-weighted
so that the full-scale measurement was 10.0 mg. The
sensitivity of the balance at this setting was such that
a weight change of five micrograms could be detected; this
was 0.01% of the sample weight.

Bouyancy effects could be calculated from the equation

P: P m we R T (39)

where P is the gas pressure (atm), M is the molecular
weight of the gas, w is the weight of the sample. P is
the density of the sample, and R and T are the gas
constant and the absolute temperature, respectively. It
was found, however, that even at saturation vapor pressure,
the bouyancy correction for water was only one microgram.

Therefore,in this work, no bouyancy corrections were used.

“1‘.“ 1“ (C(n"

35

 

 

 

 

.msumummmm mocmamnouoflelsssom> mo oeumsmsom .m musmflm

 

 

 

 

 

 

 

 

 

 

 

 

sass
43.25.82
7
. same
055329
a. «52.32.
.25 ass...
38 542. m<o
usa<o
sand 20.
2053.15 iifimo D 3
$1 -2528 5d
‘

m.o>¢mmu¢ m<o

mwhmloz<3
>¢30¢u2

 

 

 

 

 

 

 

 

 

 

uoz<4<a loco-2

tel

36

Equilibrium was generally reached in less than Six
hours for a p/pO of from 0.0 to 0.4; equilibrium was
defined as no detectable weight change after three hours
(note that weight changes of 0.01% were detectable). The
higher values of p/pO always required longer periods to
reach equilibrium, and the measurements were made after
a period of 24 hours. The time to reach equilibrium was
less in the instances where powder samples were used as
compared to pressed pellets. This is, no doubt, the result
of easier diffusion of the water vapor. In that equilibrium
weights were not affected by the sample form, powder samples
were used to reduce the time needed for each weight deter-

mination.

Adsorption Kinetics

Using the above mentioned Cahn RC microbalance and a
recorder, the adsorption kinetiascould be studied. In these
instances, it was necessary to bring the sample to its dry
weight beforecommencing the adsorption. Sample weights

and times were read directly from the strip chart.

Dielggtric Measurements

The measurements were made on pressed pellets 1 cm.
in diameter and about 1 mm. thick. The samples were slowly
pressed to avoid denaturation to a pressure of 103 Kg./cm.2.
They were placed in a Teflon insulated stainless steel
dielectric cell (Balsbaugh Laboratories - Model LD-3).

The dielectric cell was placed in a large, brass vacuum

Naif

as I

('0'

37

tight chamber (grounded) and attached to the vacuum line.
Water vapor pressure could be determined in the same way
as described under "Adsorption Isotherms".

In order to insure that equilibrium had been reached,
twenty-four hours intervened between the introduction of
water vapor and the dielectric measurements. This was
found to be sufficient as equilibrium times for these
eXperiments was similar to water adsorption equilibrium
times.

Measurements were made in the frequency range 100 Hz.
to 105 Hz. with a General Radio 1610-8 Capacitance Measuring
Assembly which consisted of:

(i) A Schering Bridge (716) which reads from 30 Hz.
to 105 Hz.

(ii) A Guard Circuit (716-P4)

(iii) An Oscillator (1302-A) with a range of 10 Hz.
to 105 Hz.

(iv) An Amplifier and Null Detector (1231-B)

Denaturation

Hemoglobin was the only compound which was tested for
denaturation, however, the proceedure could only discover
the grossest types. Samples of hemoglobin were weighed,
dissolved in distilled water. and filtered through Millipore
filters. The filter was dried in an oven and weighed to
determine the water-insoluable protein.

The average of the two samples tested was 5.8%

denaturation.

na

I‘..

38

Electrolysis Experiments

The samples for electrolysis were prepared and mounted
in a manner similar to that described under "Conductivity
Measurements". The samples were suspended in glass sample
chambers, care being taken to avoid contact with the walls.
This was easily accomplished when the chambers were attached
to the vacuum line, Figure 3, by simply tilting them. The
platinium electrodes were blocking for protons so all protonic
carriers were released as hydrogen gas.

The glass sample chambers were made without glass joints
and were closed with a torch when the samples were added.
Thus no hydrogen was able to leak into or out of the system
through the vacuum grease.

The system was pumped with a fore pump and a glass
three-stage oil diffusion pump, and the system pressure
(exclusive of the mercury vapor from the McLeod gauge) was
less than 5 x 10..6 torr. All samples were pumped continu-
ously from the time they were sealed to the line except for
the time of electrolysis and measurement. In no case were
they again eXposed to atmospheric pressure.

Many designs were tried before the present one was
arrived at, and it represented the best compromise of
sensitivity, precisidn , and accuracy. One other design
is described more fully in a later section.

The volume of the electrolysis manifold was measured
using a mercury manometer, a calibrated volume, and Boyle's

law. In this way, the volume was determined to 1%.

39

.mSumumem mam>a0uuowam ecu mo Emumwao .m wusmflm

moao oowqoi

% rmmomoommfl
H
Q Tmmzomsbmmgl

 

 

 

 

.25 N2 .0: $962. $23

mammzslo :23 \m \

a ram.

C 223.35

 

 

 

 

 

 

 

40

The number of moles of hydrogen generated could be

determined from a measurement of the pressure with the
McLeod gauge and a knowledge of thenmufijkflxivolume and
the equation PV = nRT; the gas laws were assumed to hold
at the low pressures involved, yigL, 0.1 to 50 millitorr.

The sample chambers were immersed in a water bath
held at the same temperature at which the adsorption
isotherms were made (26.00 C.). The samples were hydrated
for about three hours (sufficient for electrical measure-
ments) by opening the stopcock to the water reservoir.

The vapor pressure was changed by adjusting the temperature
of the water in the reservoir: this was done with a separate
circulating water bath. The hydration state was then deter-
mined from the adsorption isotherm. The samples used in

the electrolysis were from the same bottle as those used

in determining the adsorption isotherm; this was necessary
as it was found that there were small, but significant,
differences between adsorption isotherms of samples from
different lots.

Potential differences for the electrolysis eXperiments
were kept between 50 and 200 volts, as it was found that
erronious results could arise from the application of
higher applied potentials, eSpecially with the drier
samples. The currents passed with these potentials varied
from 5 x 10"8 amperes to 5 x 10.“ amperes depending on the
sample used and its hydration state. Hydration states

producing currents of less than 10'8 amperes (at 200 v.)

41

could not be studied in this apparatus as they produced
to little hydrogen to be detected against the "background"
pressure. This "background"was the result of out-gassing
of the walls of the sample chamber, the sample itself,

and the tubing of the McLeod gauge, none of which could
be heated; the "background" pressure was on the order of
0.1 millitorr.

The number of coulombs passed was determined graphic-
ally with an electrometer and a strip-chart recorder. In
those cases where polarization effects produced a rapid
change of the current with time, a mechanical integrator
attached to the recorder was used. The number of coulombs
passed could be determined to about 2%.

With a knowledge of the number of coulombs passed,
the theoretical yield of hydrogen could be determined,
assuming 100% ionic conduction, from the laws of electro-
lysis. Thus, two equivalents of charge are needed to
produce one mole of diatomic hydrogen.

The amount of hydrogen evolved during electrolysis
was measured by first measuring the pressure of hydrogen
plus residual gas (nitrogen and oxygen); and then heating
the palladium tube to allow the hydrogen to leak out.
Palladium is Specifically permeable to hydrogen. The
hydrogen removal generally took about 20 to 30 minutes.
The second pressure measurement then determined the remaining
residual gas. The hydrogen pressure was the difference

between the first and second measurement. Because of the

42

Specific permeability of palladium, no chemical analysis
for hydrogen was needed.

It was necessary to use a cold trap during the
measurements to remove the water vapor. If this was not
done, it would condense in the McLeod gauge upon compression
of the gas with the mercury. Using liquid nitrogen helped
to reduce the residual gas pressure somewhat.

Unlike ionization gauges, a McLeod gauge does not
need a Specific calibration for hydrogen thus all measure-
ments were "absolute". To insure that there was no loss
of hydrogen from the system, preliminary tests were made
using oxalic acid dihydrate. This compound was pressed into
pellets and mounted in a similar manner to the biological
samples and hydrated. It is strongly suspected that these
crystals are protonic conductors from the work of Pollack
and Ubbelohde (1956) on the activation energy and conduc-
tivity. Two measurements on hydrated samples (p/po = 0.82)
gave values of 98% and 112%. The average was 105%, and the
difference between the two values. 14%, served as an
indication of the reproducibility in the other measurements.

A sheet of Teflon was also placed into the sandwich
cell. and the chamber was fully hydrated (p/Po = 1.0).

The leakage current was 10.10 amperes at 500 volts: this
was 10' of the current which would pass through a protein
sample at this same hydration and potential.

Graphite, in the form of Aqua Dag was placed in the

sample holder, hydrated, and electrolysed. A current of

43

300 milliamperes was passed for 20 minutes with no
detectable release of hydrogen. This was to be eXpected

in that graphite is an electronic conductor.

Tracer EXperiments

The amount of protonic conduction in hydrated
compounds was examined by a tracer technique using tritiated
water. The apparatus is shown in Figure 4. It was used
both for the calibration of the water activity and the
electrolysis of the biomacromolecules.

The device consisted of a sample holder and chamber
similar to that described earlier, and an evacuated mani-
fold. For the calibration of the activity of the water,

a small Teflon Spacer was placed in the sandwich cell to
prevent the electrodes from Short circuiting and to hold
the water sample.

The tritiated water was calibrated by determining
the counts per minute produced per microampere-minute
(epm/pamprmin.) for a series of dilutions. The change
in activity was found to be linear with concentration.

By extrapolating back to zero dilution, it was possible
to determine the activity of the original tritiated water
sample. Undiluted water was too radioactive to determine
directly.

The activity of the water was found to be independent
of the pH of the water used for dilution.

The activity measurements, and electrolysis measurements

were made by evacuating the entire manifold, and then

44

diam mun—n5...

mmmE<Io
uni—24$ I (

 

 

 

m<o

 

.ucmsflummxm Houmup Seaway» mcu CH poms msumummmm 0:» mo EmuwMfiQ .: wusmflm

way—b.2025).

o. .d

 

as

 

 

awkzaoo
Ago—FmOdoma

 

45

closing off the sample chamber while hydration and electro-
lysis were occuring. At the conclusison of the electrolysis,
the tritium evolved was separated from the tritiated water
vapor by pumping the gas and vapor through a liquid nitrogen
cooled trap with a Toepler pump. The tritium was pumped

into the counter tube, and after ten cycles of the pump,

more than 99% of the tritium was in the prOportional counter
tube.

With the tritium in the counter tube, one atmOSphere
of P-10 gas (argon, 10% methane) was admitted rapidly to
prevent the back diffusion of tritium. The activity of the
sample (in counts per minute) was then measured with the
sealer.

The counter tube was of an all glass design for the
outer tube. The inner surface was grounded while the
inner tungsten wire (2 mil) was at a potential of +2,100
volts. The plateau region of the counter was from 2,000
volts to 2,300 volts.

At the end of the counting period, the eXperiment was
terminated by pumping out all of the gas. All counting
eXperiments were bracketed by measurements of the back-
ground. This was typically of from 100 cpm. to 300 cpm.

If there was no leakage of the tritiated water vapor past
the cold trap, the background after the eXperiment was the
same as before the electrolysis began. However leakage,

when it occurred, could always be easily seen in that the

background after electrolysis would be several thousand cpm.

46

In addition, it would not fall unless there was protracted
pumping for two days with heating. In measurements where
no leakage occurred, the background was the same as initially
found with only short pumping times - on the order of one
to two minutes.

The same proceedure was used for the electrolysis of
the sample as was followed in the calibration of the water
activity. Thus if any loss occurred through leakage or
adsorption, this would be true for the calibration and would
simply result in a lower calibrated activity.

The loss in activity of the radioactive water from the
exchange of tritium with the labile protons of the protein
was measured by two methods. In the first, the sample
(hemoglobin) was hydrated in a U-tube (water on one side
and protein on the other) for five days. At the end of
this period, the water was desorbed from the protein by
placing one side of the tube in a dry ice bath; this water
was then diluted and analysed for its activity. The loss
in activity represented the amount of radioactivity taken
up by the protein by exchange of pntium for tritium.

In the second method, the sample (collagen) was
hydrated and then dehydrated after a period of about six
hours, the time needed for electrolysis. The sample was
then placed in a large volume of water so that the tritium
on the collagen would exchange for protium of the water.
The water was then analysed for activity.

These effects are mentioned in that the adsorptionand

exchange are the most'bothersmneparts of this method.

IV. EXPERIMENTAL RESULTS

Adsorption Studies

It can be seen from the adsorption isotherms in Figures
5 and 6 that all of the compounds form the BET (Type II)
isotherm. This is common for compounds diSplaying physical
adsorption as contrasted with chemisorption.

With the exception of melanin, the compounds did not
exhibit a hysteresis when dried and rehydrated. It was
found that two states of melanin appear to exist, viz.. a
"low adsorption" and a "high adsorption" state. If the
melanin was dried at 750 C. and hydrated at 26°C., a given
hydration would be found. Then if the samples were pumped
"dry", but not heated, it was found that upon hydration at
260 C. to the same value of p/po' a larger amount of water
was adsorbed. This effect was eXperimentally disconcerting
at first, and the reason for the "activation" is not known.

Excessive heating of the sample will alter the adsorption
isotherm to a large degree. This was found to be true for
collagen; the amount of water adsorbed was considerably
much less when the sample had been heated so that slight
pyrolysis had occurred as evidenced by a slight darkening

of the strips.

47

Figure 5. Adsorption isotherm of water on melanin,
collagen, and salmon sperm DNA at 25.00 C. The
solid circles are melanin, the open circles are
DNA, and the squares are collagen.

 

1+9

 

cubi’ i

 

FIG-N3

Figure 6. Adsorption isotherm of water on lecithin,
cytochrome-c, and hemoglobin at 26.00 C. The solid
circles are the hemoglobin, the open circles are the
lecithin, and the squares are the cytochrome-c.

 

51

 

 

cmh<3 * bro-N3

/P

52

Sample characteristics are also changed if the samples
are hydrated to such an extent that they dissolve in their
water of hydration. Thus a sample of cytochrome-c was
found to diSplay entirely different adsorption character-
istics after excessive wetting and redrying in the balance.

It is necessary to follow the same precautions in the
electrolysis cells as followed in the adsorption isotherm
cells as the latter must serve as a calibration for the
former. This is particularly true of heating effects.

BET plots were made of the adsorption data, and these
are shown in Figures 7 and 8. They are linear to a value
of p/po of about 0.6. From these plots, a value may be
found for vm , the monolayer coverage. These values of
the monolayer coverage are shown in Table 1. In addition,
the values found by Postow (1968) and Gardew and Eley (1958)
for hemoglobin are given here fOr comparison.

Bradley adsorption isotherms were plotted in the hOpe
that they would provide a method of determining the amount
of adsorption with a minimum number of measurements. These
are Shown in Figures 9 to 12. Unfortunately the isotherms
are not linear over the entire range of p/pO for half of
the materials tested, and breakpoints were noted. Its
range of linearity, however. is considerably wider than

the BET plots.

Kineti§»Studies
The adsorption kinetics of four of the compounds were

studied, with the results that all could be fitted to the

Figure 7. BET plot of water adsorption on lecithin,
cytochrome-c, and hemoglobin at 26.00 C. The solid
circles are hemoglobin, the Open circles are lecithin.
and the triangles are cytochrome-c.

54

m

an

 

4
5.3.2..

Figure 8. BET plot of water adsorption on collagen,
melanin, and salmon Sperm DNA at 26.00 C. The solid
circles are collagen, the open circles are DNA, and

the triangles are melanin.

 

 

4
.cucsxxc

 

I.O

57

Table 1. The BET monolayer coverage for the materials
tested in this work.

 

 

 

Material BET Monolayer

 

I. By Powell

Collagen (@260 C.) + water
Salmon Sperm DEA "
Hemoglobin "
Melanin "
Lecithin "
Cytochrome-c "

H H
O\-(:'(\)0\N\l
O O O O O O

\1) O‘x-F'H (:0
6x

II. By Postow (1968)

Hemoglobin (@ 230 C.) + water
Hemoglobin + methanol 1
Hemoglobin + ethanol

\ono
LnUIUI
0‘

III. By Cardew and Eley (1958)

Hemoglobin (@ 300 C.) + water 5
Hemoglobin (@ 400 C.) + water 5

\3\)
NO\
\Q. \‘a
cxox

 

N
O

WEIGHT % WATER

5

Figure 9.
lecithin

58

    

COLLAGEN

2.4

.8 LG
L06 [L06( Pe/P)]+2

Bradley plot of water adsorption on collagen and

death-“1T7? ._“ ._'.I.

Immi‘i;ujrt '3'"

59

40
CYTOCHROME' C
30
HEMOGLOBIN
20

WEIGHT % WATER

Io \

\\

\

\

.8 LG 2.4
LOGl'LOG(P°/P)] + 2

Figure 10. Bradley plot of water adsorption on hemo-
globin and cytochrome-c.

-‘.'.’.l‘ m m Saris-"f“?!
'

80

60
5
g; ‘40
3'
g.-
.5.
u
3
20
0

Figure 11.

/
o O

7 L4 ZJ
LOGILOGH’e/ P)“ 2

Bradley plot of water adsorption on DNA.

21!

“1

tunam-m—ma am).- ‘ . a _

(40

30

20

WEIGHT % WATER

IO

Figure 12.

61

 

7 L4 ZJ 28

LOG! LOG( Po/PU o 2
Bradley plot of water adsorption on melanin.

62

Roginsky—Zeldovich, or Elovich, equation. This data is
Shown in the Figures 13 to 16. The free parameter, to, is
indicated at the side of each line. It will be seen that
all required a small value of to. This value must always
be positive.

It was observed that the conductivity increase does
not follow Roginsky-Zeldovich kinetics. If one compares
the conductivity at any one given point in time during
the adsorption process, the "equilibrium weight" can be
found from a graph of conductivity yg; equilibrium weight.
The values from the graph correspond to conductivities
and weights at equilibrium, i.e., after 24 hours.

It can be seen from Figure 14 that the"current weight"
does not follow the Elovich equation. A curve which is
concave to the right can not be made linear as only

positive values of tO are allowed.

Conductivity Measurements

All compounds tested were found to give conductivity
increases with hydration. This effect is shown in Figures
17 to 23. The conductivity increases could be fitted to

the equation

ln0"=0(m+ln0" (40)
dry

Additionally, all samples showed a saturation of current
with increasing hydration with the exception of hemoglobin+
water and hemoglobin+methanol. Saturation occurred after

the adsorption of about three BET monolayers. This was as

WEIGHT % WATER

Figure 13.

5 IO 50
L06( If 10)

Adsorption kinetics of water on hemoglobin.

 

IOO

A wgnag—H

—...".‘ —

I E r .EZE'IFTI.‘

00.

FNUWIIIII:

 

.commaaoo co umum3 mo moauocflx COAumuompm

1...: 00..

 

are»; 35.8..

   

.22»; 22.8 5.5.35.

.ea owsmas

N.

HBLVM % 1H9|3M

m.

cm

65

00.

on

.canufloma so umum3 mo moaumcflx coflumuompa .mH musmflm

3.300..
o. - I m ._
mud. mus. no
0 . N
c
M
B
m
m 1
%
M
w.
o 3
8
o.
N.

00

on

.CHcmHmE so Hmum3 mo mUfluwCHx :oHumHOmva .wH musmflm

3150..

o.

O.

HBLVM % .LH9I3M

67

Nil

(amps)

LOG CURRENT

 

'/

BET 2851' 3BET

A A A
IO 20 30 4O
WEIGHT 96 WATER

Figure 17. Conductivity of collagen as a function of
adsorbed water.

   

Fa .lw . .Trlb}br4§m.guair

.uwumB UmQHOmUm mo coHuucow m mm CowHmE mo aufl>fluosvcoo .wH musmflm
cabs, i 2.333

 

 

ov On ON 0.
_ m _ 4 _ f— P
P” n PuON hum.
IE:
[w .I
0
9
n
a
a
3
R4 1 F m.
In ‘
mu:
m
M
l m ‘
1m
C

 

 

 

 

-_ i-1

~fl

 

c U R RENT (amp!)
0|

LOG

 

 

    
 

 

ZB‘ET L3‘BET i
I2

IS 24

 

WEIGHT % WATER

Figure 19. Conductivity of lecithin as a function of
adsorbed water.

70

ml

:7
CL
E
s
5...
me
a:
a:
D
o-
9
(.9
O
.l
76

       

20 40 so
WEIGHT 96 WATER

Figure 20. Conductivity of cytochrome-c as a function of
adsorbed water.

g‘ 1
(r

 

fiilWiu l .. -

mm

.1?—

LO 6 CURRENT (amps)

Figure 21.
adsorbed water.

IS

71

24

 

32

The conductivity of hemoglobin as a function of

 

CURRENT (amps)

LOG

72

 

3O 60 90 I20
WEIGHT 9‘ WATER

Figure 22. The conductivity of DNA as a function of
adsorbed water.

‘;P—-r-h_I!-v.fwii7km.r- T1

ml «I an

LOG CURRENT (amps)
col

Figure 23.
of adsorbed

  

O

73

      

IBET 2§ET 386T
20 40 so

W EIGHT 96 METHANOL

The conductivity of hemoglobin as a function
methanol.

74

predicted by the Dielectric Constant theory as described
earlier.

The curve in Figure 23 (hemoglobin+methanol) is
calculated from an extrapolation of the Bradley adsorp-
tion isotherm for this system. Direct measurement of the
entire range of p/pO is not possible (only measurements to
p/pO = 0.75 have been done) with the present equipment as 1V
the Cahn Microbalance eXperiences problems in the presence

of large amounts of methanol vapor.

Resistance - Voltage Measurements

 

It can be seen from the data in Figures 24 to 30 that
Ohm's law is not strictly followed at low applied voltages.
This is indicated by the non-linearity of the data for field
strengths of less than 300 volts/cm. (30 volts applied).
Above this value, the results indicate an almost ohmic
behavior; this is characterized by an almost horizontal line.
For low states of hydration, the samples diSplayed the
Evershed effect discussed in the Introduction. Hemoglobin
was the only compound examined which showed this effect
over the full range of solvation. The various states of
solvation for all the compounds are given as per cent
solvations by the small numbers at the right side of the
figure.

The possible variations from Ohm's law at the higher
hydrations may not be real but only the result of obtaining
incorrect measurements of the conductance in the presence

of very large polarization effects.

RESISTANCE (fl/cm.)

r’; r"
r’ 3

 

 

 

 

I3
I2
5% i
ll i
E
l
E.
61%
9
_. 9.4%
8
IO.4%
L/k new.
7
6
|00 I000 3000

FIELD STRENGTH (v./cm)

Figure 24. Resistance-voltage plots for hydrated lecithin.

76

I3

n. “J..-
U.
1.:

 

111' ‘

 

1P . tu'MAG m‘fifi‘

   
   

5

 

RESISTANCE (fl/cm.)
co

 

 

I00 I000 3000
FIELD STRENGTH (v./cm.)

Figure 25. Resistance-voltage plots for hydrated melanin.

 

 

 

RESISTANCE (ll/cm.)

 

 

 

 

 

'00 IOOO 3000

FIELD STRENGTH (v. lcm.)

Figure 26. Resistance-voltage plots for hydrated
cytochrome-c.

RESISTANCE (II/cm.)

 

 

 

 

 

 

 

I00 |000 3000

FIELD STRENGTH (VJCM)

Figure 27. Resistance-voltage plots for hydrated hemoglobin.

 

 

 

 

 

 

+ _‘
8.8%
I
I2
' ‘r
5: “ l8%
m
o
3 IO
p.
Q
m
m N
“ 3w.
9 I
+
38%
49%
8
7
I00 I000 3000

F IE LD STRENGTH (£an

Figure 28. Resistance—voltage plots for methanol and
hemoglobin.

—I-.T
(.1

‘filéifi 92:..«73I.'t-.’—‘“ks““1“-r - ._ ‘

RESISTANCEM/cm.)

80

 

 

 

 

 

 

I00

Figure 29.

 

|000 3000

FIELD STRENGTH IV./Cm.I

Resistance-voltage plots for ethanol and hemoglobin.

 

31

l2

I0

 

(0

RESISTANCE (fl/cm.)

 

I00 I000 3000

FIELD STRENGTH “1ch
Figure 30. Resistance-voltage plots for hydrated DNA.

82

There appears to be no correlation between the line
shapes of the Voltage-Resistance curves and the amount of

electronic or protonic conduction.

Dielectric Studies
For the three materials studied, all produced curves

of capacitance vs. hydration which showed a rapid rise

_ -afi
_ J"
. ___M:i—'

after a hydration of about two BET monolayers. This is
shown in the Figures 31 to 33.
3

Frequencies lower than 10 Hz. were not measured as .

it was observed by Postow (1968) that the capacitance had

 

{’51 - .. _.- ,

a large temperature coefficient in these ranges. In that
a temperature coefficient is observed only in the term

eXp (-E/ 2kT) of the semiconduction equation, that is. the
"activation energy" for charge carrier generation, the low
frequency measurements do not apply to this theory. The
increase of capacitance at low frequencies has been
ascribed to Maxwell-Wagner polarization. This is produced
by inequalities in the ratio of the dielectric constant
and conductivities of the surfaces and the bulk volumes of
the sample, or different components of the sample — water
and protein, for example.

The dissipation factors of the samples measured were
all characterized by a decrease in value with increase in
frequency. This is at variance with the results found by
other workers (Rosen, 1963: Brausse gt gig, 1968; Takashima
and Schwan, 1965) but has been found consistently in this

laboratory.

30

CAPACITANCE

20

IO

.Figure 31.

IOHz

IO Hz.

   
  

5

.—4

I0 20
WEIGHT ’- WAT ER

Capacitance vs. hydration for hemoglobin.

——/

__‘

 

30

u

 

N
O

CAPACITANCE

I0

 

8 I6 24
WEIGHT 96 WATER

Figure 32. Catacitance y§_._ hydration for cytochrome-c.

85

25

   

8

CAPACITANCE
6.

I0

   

I0 IS 20 25

WEIGHT % WATER

Figure 33. Capacitance y_s_._ hydration for lecithin.

 

Electrolysis EXperiments

The results for the electrolysis of compounds with
varying amounts of adsorbed water are shown in Figures
3b to 40. The results are as eXpected, that is, increasing
hydration results in increased protonic conduction. This
is true for all sampleswith the exception of collagen, a

triple helix with a high degree of hydrogen bonding. rm

_’_'.-'I

The error bars on the graphs represent the 15% vari-
ation found in the electrolysis of oxalic acid dihydrate
and appear to be quite realistic in estimating the precis-

ion of the measurements;small figures are eXperiment numbers.

 

Collagen, as mentioned, appeared to start out as a
protonic conductor and become more of an electronic
conductor with hydration. The bound water and the hydrogen
bonding may be the cause of this large protonic mode.

Small changes in.capacitance with increased hydration may
then result in a small changewhich allows the increase of
electronic conduction. It must be remembered that the total
amount of protonic conduction does not decrease, only the
gatig of protonic to electronic carriers.

Melanin exhibited no change in the ratio of charge
carriers through the entire hydration range of one to three
BET monolayers. It was found to be the only compound to
show this constancy.

Lecithin was tested in hydration ranges between 0.8
and 3.5 BET monolayers. It exhibited a gradually increaéing

protonic component starting from an even ratio.

 

.mmfluuw commfiaoo mom coflumucwn mo cofluucsm m mm :mm0upxc mo pflmflm one

«mph; $ #1 033
on ON 0.

 

.zm madman

NSSOHOAH

ma. I.

I . r h. ...1H.:."lallt..p

.CHCmHoE How Coflumuvmn mo cofluucsm m mm ammonpxn mo paoax one
«3.2? .8 .2..ng
on ON 0.

 

.mm musmflm

C3
(0

CD
¢

NBSOUOAH 96

89

 

1"“1

 

.cacuflooa now ceaumup>n mo :ofiuocsm m mm cmm0u0>5 mo pamfla one

musk; $ P1933
N. 0

.mn musmflm

ON

NBSOHOAH 96

OO.

90

m1-
Fl. . I Ll

.olmEOucUOumo new :oHumupmn mo cofluucsw m mm cmm0u0>: mo ©H0H> one

”$5.43 o\o From;
on ON 0.

 

.sn ousmgm

om
%
9. m
0
8
O
m
00 N
0m

00.

91

 

.CfiQOHmOEwL no“ coflumupxc mo cofluocsM m mm comoupxz mo Gama» one .mn gunman

CHER; 3 .5533
0» ON O.

  

O Q d
Pmon .PUON hum.
ON

O¢ an

H

.A

.. a: m

00 0

9

3

N

92

.CHQOHmOEmn How “Hocmnumev

On O¢

 

coHum>H0m mo cofluocsm m we somoucan mo Gama» $28

40315:: 29...:
on 8 ; 9

.mm oudmflm

%.

NSSOHOAH

.420 How coHumup>: mo COauucsm m mm somehow: mo papa» one .C: madman

«5.23 1. Prom;
GO. ON Ow . V ON

. d
hum» hum N .50 _

.+.

O
N

r.

O
V

93

O
‘D
NSSOHOAH 96

O
O

= v

00.

.cflQOHmOEmz mo h&mwv mumum
coHumupxn oco HON mmmuao> mo CONDUCSM m we com0u©>n mo nama> one .H: wusmflm

A333 44:2 m Pom
00m OON OO.

 

92.

ON

 

OO.

N39080AH ‘36

95

Cytochrome—e was the only compound found to change
from electronic conduction as the majority mode to protonic
as the main form and yet vary linearly with hydration. The
hydration varied from two BET monolayers to about six.

Hemoglobin exhibited a pattern of change which was
very similar to cytochrome-c. This similarity was also
seen in their adsorption isotherms and BET monolayers
(6.1% for hemoglobin and 6.3% for cytochrome-c). An
undefined hydration state for hemoglobin (p/po = 0.98),
probably about 60%, gave a hydrogen yield of 9 %; this is
indicated at the side of the figure.

When hemoglobin was allowed to adsorb methanol
instead of water, the change of carriers was not linear
with solvation as shown in Figure 39. The cause for this
is not yet known with certainty.

DNA showed a change similar to that for the methanol-
hemoglobin system. Notice that the adsorbate was water
and not methanol. This sharp change camein.the range of

two to three BET monolayers.

Tracengxperiments

Tritium exchange (tritium for protium) was measured
for the two materials tested, hemoglobin and collagen, but
only at one hydration state. Nevertheless, the figures
for the exchange indicate an "order of magnitude" feeling
for the process

For the hemoglobin system, the sample was studied at

8% to 10% hydration with the finding that the exchange was

 

96
about 545. That is, of the initial activity of the water,
5b% of it could now be found on the protein. (This should
not be confused with the idea that 54$ of the protons on
the protein were now tritium.)

For the collagen system, one measurement was made at
about 12% hydration; the exchange was about 38%. For these
two systems, the hemoglobin was hydrated to 1% BET mono- Pޤ
layers, and the collagen was hydrated to 1% BET monolayers. I

The results of the electrolysis were uneXpected in
their small yield of tritium. The results are given in

Table 2. ,

 

Table 2. The tritium yield for the eleirolysis of collagen
and hemoglobin.

L— 4 T -
L w;-:— J —

 

 

 

Hydration State Yield of Tritium
Hemoglobin:
12.5% 23
13-33 3%
18.5% 1%
about 35 % 0.1%
" 35 % 0.5%
Collagen:
13 % 4.3%
13 % 5.4%

very wet 0.23%

 

97

All values indicated a smaller protonic/electronic
charge carrier ratio than given by the eXperiments
measuring the change in pressure upon the evolution of
hydrogen. This discrepancy is, at present, unaccounted

for.

 

V. DISCUSSION
Biological Systems and Semiconductivity

It might now be asked, "What systems have been

postulated in the past to involve the movement of electrons pg]
It

by semiconduction?". There are numerous examples, a few
of which will be discussed here although it must be
remembered that none have been proven. ‘

Enzyme systems are among the most interesting

 

possibilities for this semiconduction process. The
activation energies (now figured on a l/kT basis)are often
of the order of magnitude of 0.5 to 1.0 e.v. (12 to 23
kcal./mole) which is the value for the semiconduction
activation energy for hydrated proteins. The idea is
theoretically appealing but is, unfortunately, difficult
to prove on a system as small as one molecule - semiconduc-
tion is, at present, a macrosc0pic phenomenon. Additionally
there are many theories of catalysis, all with their
Specific "proofs", and it would be difficult to relate many
theories with one, simple mechanism. However, oxidation-
reduction enzymes are a likely starting point.

LuValle and Goddard (1948) published a theory which
could be considered as the forerunner of a semiconduction
mechanism. Semiconduction was not known to them, but rather

they used resonance over unsaturated bonds. These were

98

.1 9‘“

99

postulated to exist between electron donor and acceptor
sites on different parts of the enzyme molecule.

Weber (1955) proposed a semiconduction mechanism for
catalysis but later rejected it since the activation energy
for charge carrier generation was larger than that for
catalysis. He however used the dry state activation energy.

Cope has derived quantitative kinetic rate equations fa
for oxidation-reduction enzymes. He however only uses ET}
the empirical fact that proteins are able to conduct an :

electronic current. The reaction rates are governed by

.. . ‘nl'f—E' "' n.‘ '. "

diffusion and over-voltage and not by any semiconduction

 

T [.11

parameters. The current carrying ability is impressed,
without modification, onto single-molecule systems.

In the realm of inorganic catalysis, Vol'kenstein
(1963) has worked out a theory utilizing the semiconductive
properties of many catalysts.

Rosenberg and Postow (1969) have called attention
to the fact that changes in the molecular confermation of
an enzyme could result in hydration changes and concomitant
changes in activation energy. The activation energy
change could then effect a change in the catalytic rate.

Such changes in activation energy at increased temp-
erature was first noted by Crozier (1925). He postulated
a series of reactions, one of which was rate-limiting, the
"master reaction", and which governed the entire process.
These effects were noted in biological systems. Among these

were the chirping of crickets, the creeping of ants,

lOO

ciliary activity, the flashing of fireflies, oxygen
consumption, bacterial growth, and the production of carbon
dioxide. In many of these cases, there was an abrupt break
in the Arrhenius plot, and this was frequently at about 15°C.

lg_yitgg effects have also been obserVed involving a
change in the activation energy, and Dixon and Webb (1958)
have prOposed the following eXplainations of the effect:

(1) a phase change in the solvent,
(ii) a change from one rate-limiting enzyme to
another,
(iii) a parallel series of reactions, each with
different activation energies,
(iv) reversible inactivation of the enzyme,
and (v) one enzyme, existing in two forms, with
different activation energies for each.
It is this last eXplanation which is of interest to the
theory of semiconducting enzymes. It is one however which
is difficult to demonstrate with regard to semiconducting
mechanisms.

From the time that Szent-Gyargi proposed the idea of
biological semiconduction, workers have been looking for
instances to apply this fine idea. Arnold and Sherwood
(1957) proposed that chlorOplast grana were semiconductors,
and, by this process, were able to transfer the light energy
to the active, enzymatic site. The light was postulated
to form electrons and holes which migrated separately to

their respective "trapping centers“ where they were utilized

 

101

in oxidations and reductions. Resonance transfer is more
in vogue at this particular moment.

Arnold and Sherwood used the semiconductor model to
eXplain their data on thermoluminescence. Bradley and
Calvin (1955) proposed that, if the chlorophyll were to be
positioned between lipoprotein layers, one of which is

n-type and and the other of Which is p—type, the photo-

w . 9' “filming
_.
J

‘ ,q.

generated electrons would be able to diffusato their
reSpective sites of chemical activity.

ESR work at low temperature by Calvin and Androes

 

‘mm- -‘—:

(1962) and Calvin, Kurtz, and Ruby (1965) has given some
support to the semiconductive hypothesis, but it has also
been applicable to quantum-mechanical tunnelling. The
hypothesis is that exciton migration brings the electron-
hole pair to the enzymatic site; there triphosphopyridine
nucleotide is oxidized and the hole migrates to another
site where it either recombines with another electron
which has been photoexcitéd, or it combines with an electron
from oxidized cytochrome-c. These processes would occur in
green plants and photosynthetic bacteria, respectively.

Eley and Snart (1965) have found that charge separa-
tion occurs in complexes of protein-chlorOphyll-carotene-
B-methyl napthoquinone. These complexes exhibited different
semiconductive activation energies than the components.

The chemiosmotic hypothesis of Mitchell (1961)
proposes that the link between metabolic energy and oxi-

dative, or photosynthetic, phosphorylation is the membrane-

102

based, reversible ATPase. The membrane must be impermeable
to protons and possesselectron-transporting components.
The nature of this conduction is not known, but it could
be a semiconductive one.

E. J. Conway (1953) has suggested an ion tranSport
mechanism based on the movement of electrons in the membrane
components. These electrons, or negatively charge "carriers? !
would then combine with cations to form a neutral Species {*1
capable of migration through the membrane. At the opposite
side of the barrier, the electron would be removed, and the

complex would dissociate thus releasing the ion.

 

Cytochrome compounds (cytochrome-b5) have been detected
in the membranes of endoplasmic reticulum (Remmer and
Merker, 1965). These cytochromes have been implicated in
ion pumps by Davies (1960) and also in functions such as
the terminal oxidation reactions in the synthesis of
proteins. These latter reactions would also involve the
shuttling of electrons from NADH to various molecular acceptors.
Additionally, biotransformation of pharmacological compounds
takes place at the endoplasmic reticulum.

Again in the area of membranes and boundries, is the
finding by Digby (1965) that the cuticle (quinone-tanned
protein) of various crustacea is capable of conducting
electrons. This was demonstrated by electrodepOSItion of
copper on the surface of the cuticle. Digby has proposed
that a cathodic reaction with the salt water (the animal

is normally negative with respect to the water) results

103

in the formation of an alkaline pH in the area. This
promotes the calcification of the animals exoskeleton.

Recent eXperiments by Rosenberg and Pant (1969) may
indicate that membranes (bimolecular membranes of cholesterol-
oxidized choloesterol) are electronic conductors at low
potentials and ionic conductors at higher potentials.

The research is in a preliminary stage at the time of this
writing.

Rosenberg (1966) has demonstrated a similarity between
the photoconductive and photovoltaic responses in model
systems and those observed in the eye - this is particularly
with respect to the S-potentials. Misra g3,gl;(1968) have
prOposed a mechanism of olfaction based upon semiconduction.

Energy migration in the mitochondrion has long been
an area of active interest as it is known that the electron
tranSport chain is made up of enzymes which are rigidly
fixed to the cristae. This means that electron transfer
between the enzymes can not occur by diffusion. Movement
by molecular rotation, in some parts of the chain, is not
possible as reduction of cytochrome can take place at
liquid helium temperatures as discussed earlier in this work.
Semiconduction is one alternative.

COpe (1965) has develOped a kinetic treatment of the
electron transfer question and has tested it in yitgg. For
living systems, there is still some debate.

In the systems prepared by Chapman and Fast (1968),
it was found that chlorophyll in aqueous glycolipid and

phospholipid diSpersions were able to reduce an aqueous

10h

solution of cytochrome-c in the presence of light. This
reaction will not take place in an ethanol solution of the
components. The recent eXperiments of Tien (1968) have
domonstrated charge separation in bimolecular lipid
membranes (ELM) containing chlorophyll in the presence of

light.

Future Improvements in the Electrolysis System fit)
The primary question to which this study has been
directed dealt with the nature of the charge carriers at

various levels of solvation. For this purpose, various

 

1r. ~

facets of the adsorption process have been studied, and
the electrolysis eXperiments have been carried out. At
'nopoint in this sequence of eXperiments has it been
possible to study the very dry macromolecules, and therefore
the nature of the charge carriers in this region can only
Ixainferred from extrapolation. From the theoretical point
of view, this can be a dangerous proceedure. But from a
position of relevance to the biological realm, it is only
the solvated systems which are of importance for it is
these alone which posseasthe necessary low resistance to
make electron tranSport rates significant.. The penetration
of the low solvation region is then the "new frontier"
and awaits the deve10pment of a more sensitive device.

An attempt was made to improve the sensitivity of the
equipment by replacing the McLeod gauge with an ionization
gauge. This arrangement suffers from the problem of not

being able to detect the hydrogen against the background

105

pressure of nitrogen, oxygen, and water, the principal

gases released during the out-gassing of the walls of the

vacuum chamber. Since the ionization gauge reSponds in

a different manner for each Species of gas (the current

measured by the gauge controller is a function of the

ionization potential of the gas), unless only one species

of gas is present, an accurate determination of the pressure _

of that gas is not possible. Hydrogen would have to be the ,hI'

major component by far.
To separate the hydrogen from the other gaseous .

components of the system, a palladium tube was placed such

 

that the hydrogen gas passed into a previously well-pumped
ionization gauge instead of to the atmosphere. Unfortunately,
the reaction of palladium and hydrogen (as yet not known)
prevented all but a small fraction of the hydrogen to
penetrate the palladium tube and enter the ionization
gauge. This problem was also found by Young and Whetten
(1960) for their hydrogen measuring device based on the
palladium tube. It became necessary to generate such
large amounts of hydrogen to compensate for this loss, that
one did not achieve any more sensitivity over the sensitivity
of the device used for this work. In addition, one no
longer had an absolute gauge as possessed by utilizing
the McLeod gauge. After five months of testing, this
system was rejected.

Another method, making use of a partial pressure gauge,

will overcome the problem of determining the hydrogen pressure

106

against the background pressure in that the pressure of
each Species can be measured with this instrument. It
will be necessary,however, to bake-out the manifold so
that the systems can be sealed off from the pumps and

the pressure still not rise above 10-4 torr.

Charge Carriers and Solvation

The charge carrier change with solvation was found
to follow, in a rough way, what had been predicted by
earlier workers in the field; that is, the amount of
protonic conductivity increased with increasing amounts
of adsorbed water. However, the general thought was
directed along the line that it was not until two to three
BET monolayers had formed that protonic conduction would
become appreciable.

0n the basis of the present study, the question of the
onset of protonic conductivity is not quite settled for all
of the compounds which were tested. For the samples of
melanin, lecithin, and collagen, the current was of a
sufficiently high level that it was possible to extend the
electrolysis measurements to the one BET monolayer hydration
level. If these materials were to be called "pure electronic
conductors" in the dry state, the transition from mixed to
electronic conduction would have had to occur in a very small
hydration range. There does not seem to be, at this time,
any theoretical justification for believing that less than
one BET monolayer could alter the conduction properties to

such a degree. Indeed, in the case of collagen, the material

 

107

has a negative lepe of percent protonic conduction Kg;
hydration, and thus when dry would be a protonic conductor.
(Collagen, though, is never completely dry: high temperatures
used for a through drying produce a change in the material
as evidenced by a different adsorption isotherm.) For
these eXperiments, collagen would contain some residual water,
as is true for all systems, in the folds of the triple
helix which could contribute to its observed protonic
conductivity in the "dry" state.

Only in the cases of DNA + water and hemoglobin +

methanol was there an eXperimentally verifiable large change

Hr. . I 1'5 P77"? (BET _ fr—rquFJ-F .mmmiq
..
I

over a small solvation range (in units of BET monolayers).
This was a rapid change from electronic carriers to protonic
ones.

The basic premise of the electrolysis eXperiments is
that protonic conductivity must result in the evolution of
hydrogen. This is a consequence of the fact that metals
are "blocking electrodes" with reSpect to protonic flow.

The hydrogen evolution would not have occurred if another
type of electrode, potassium chloride in agar, for example,
had been used. Indeed, this latter type is "blocking"

for electrons.

It is believed that the evolution of hydrogen reflects
the protonic flow, and that this flow is the only ionic
Species. Protons are the only ions possessing a sufficiently
high mobility, and being present in sufficient concentration,

to be responsible for the currents observed.

108

It is found for hydrated proteinsthat the current as
a function of hydration can be given by Equation 17. It
has now been found that the current is a function of two
currents, yigg, the protonic and the electronic. It is
possible to calculate the reSpective protonic and electronic
currents as a function of hydration. This is shown for the
case of cytochrome-c (which even shows a crossover point
at 2% BET monolayers) in Figure 42. As can be seen, both
the electronic and the ionic components follow Equation 17.

That is to say

0'”) = 07W“) up (pm) (.1)
and
-) -
0‘ = ogry( ) up (1 m) (L12)

where 8 and Y" are constants, and m is the amount of
adsorbed water, and. O-(+) represents the protonic current
and (7"(') represents the electronic current. The total

current is then the sum of these two, or

0": 0"(‘I'I d- O—(-) = 021- (+)

W eXp (p m) + qry(-)eXp(V m)

(43)
These equations are applicable in the region from zero to
about 2% BET monolayers, that is, before current saturation
occurs.

It has been shown (Postow, 1968) for the systems of
hemoglobin (with water, methanol, and ethanol), melanin

(with water) and collagen (with water) that the activation

109

 

LOG CURRENT (claps)

Tl

/
[I‘LL—ELECTRONIC
/’

 

PROTONIC

 

6 9 I2 I5
WEIGHT 9‘ WATER

Figure 42. The resolution of the current increase upon
hydration into ionic and electronic components.

 

110

energy changes upon solvation, and that the value of

G}, in.Equation 16 is a constant. This indicates that
charge carriers are Egg added in the solvation process ,
and that it is only the value of E which changes.

It can also be seen for the case of melanin, that the
ratio of charge carriers is a constant for the range of from
one to three BET monolayers. From unpublished work of Postow, ‘_7
it is known that the activation energy changes upon hydra- ‘ 1
tion in a manner Similar to all other materials tested.
Thus, one should not associate the change of activation

energy, then, with a Shift from one carrier to another

 

with each possessing a different value of E . The energy
needed to overcome the coulombx:attraction of the charge
carriers and their "parent" molecule appears to be the
prinCipalenergy barrier and is the same for protons and
electrons.

One can then take Equations 41 and 42 and from them

proceed in a manner similar to Equation 17.

QT,m)(+) = 0‘0“) exp[(-ED/2kT) +P m] (44)
and
GET,m)(-) = 0;“) eXp[(-ED/2kT) +'Vm] (45)

so that one gets for each Species of charge carrier

, (+) ‘
O—(T,K)(+) = 0.0 exp (-ED/2kT) exp [(ez/ZKTR1)(% " 3;);
(46)

and mutatis mutandi for the electronic carriers.

 

111

The total current is then given by the equation
O'(T,W) =[O':)(+) eXp{(e2/2kTR1)(-3-c- - %.)} +
0”0(-) eXp{(e2/2kTR2)(%-+ é») eXp(-ED/2kT) (47)

If the values of the radii of polarization R1 are about

the same, the equation will reduce to

0'(T.K') = (a‘o(+)+ 0'0“) ) eXp (-ED/2kT) eXp{(e2/2kTR)(kl 1435-8))

in J;

and is the form usually found.
In the case of some materials, such as hemoglobin +

methanol and DNA + water, the current does not exhibit a

 

saturation with large amounts of adsorbate (about two BET g 1
monolayers). It can be seen from Figure 43 that saturation
usually occurs within one order of magnitude from a deviation
from the linear region in a log current Kg; hydration plot;
the end of the linear region is indicated by the dotted
line. For the DNA and methanol-hemoglobin systems, the
change of charge carrier species is not linear with
increasing solvation, but instead it seems to show a large
change at the point where saturation would normally occur.
No doubt, some other effect is taking place; this may be a
conformation change as is most likely the case with DNA.
This is the "disorder to A-form" transition which occurs
at 75% relative humidity (Franklin and Gosling, 1953; Falk
31; gl_._, 1963a. 1963 b).

The saturation of current which occurs from two to
three BET monolayers is the result of the rapid change in

the bulk dielectric constant of the material which makes

112

 

 

 

3 7 I V V U
Mde1 0.

z —

ammo

Cytochrm- .

§__
3.. . Humqwmh
7

 

C U RRE NT (amps)

 

 

 

A J 4 l
I 2 3 4 5 6

SET MONOLAYERS

Figure 43. The conductivity of the biomacromolecules
used in this study as a function of adsorbed water(in BET).

SI

113

l/PK' very small and reduces the exponent in Equation 26
to a constant. Figures 31, 32 and 33 show this increase
in capacitance for hemoglobin, cytochrome—c. and lecithin.
respectively.

The adsorbed water is most likely quite polarized
with regard to its orientation on the macromolecule in that

the adsorption process is found to obey the Bradley isotherm.

 

That is to say, the theory is based on the polarization of
one adsorbed layer by the substrate beneath it. This

theory has been criticized by Brunauer g1 al. (1938),but

 

 

they do state that "...if the adsorbed gas has a large
permanent dipole, it is possible that many layers may be
successively polarized..." It may be that after two to
three BET monolayers, the rotation of the molecule may
increase to allow the large increase in the bulk dielectric
constant which is observed.

The difference in the observed ratios of the electronic
to protonic charge carriers in the different materials
tested illustrates the large diversity of types in the
conduction process. The range was one of from pure protonic
(collagen) to an almost even mixture (melanin) to pure
electronic (DNA).

In the materials tested with tritiated water (collagen.
and hemoglobin), the yield of tritium was small. This
could possibly be the result of a different process for the
release of hydrogen in the solid-liquid system as compared

with the liquid (calibration) system, or it could reflect

114

the fact that the hydrogen was evolved only from the

protons of the adsorbent. Since the mole per cent of tritium
in the labeled water was small, it would be possible to

label only a small mole per cent of the protons of the
protein. The exchanged protons would also be required to

be current carrying protons if one were to eXpect the same
yield of tritium/coulomb as from the liquid water used for a}
the calibration: all protons in water are potentially
"current carrying protons".

The change in per cent yield of hydrogen at a constant

 

state of hydration also gives some indication that the i
electrolysis is not the simple decomposition of water found

in liquid systems. This change, shown in Figure #1, occurs

at about fifty volts, the hydrogen yield falling off below

this voltage. If "simple electrolysis" was occurring,

the change should come at the decomposition potential of

water, 1.35 volts.

In the voltage-resistance plots, it was seen that the
resistance showed large changes at potentials of less than
thirty volts applied ( field strength = 300 v. cm.-1). This
is also at a higher potential than the decomposition voltage
of water.

In an earlier study by Powell (1966), it was found that
the current increasing capability of adsorbates was a
function of the dielectric constant of the adsorbate. Thus
a compound such as dimethyl sulfoxide (DMSO) which has no

ionizable hydrogens, but which has a large dielectric

115

constant, 49, could effect a large current increase when
adsorbed on hemoglobin. If the protonic current is
significant, as has been found for hemoglobin / methanol.
then these protons must assuredly come from the polymer
itself. It is obvious, however, from the data on the
hemoglobin / methanol system, that not all adsorbates are
equally effective in promoting protonic conduction if it
is intrin ic to the adsorbent. But protonic conduction,
when it does occur for the before mentioned system and
for the DNA / water system, sets in sharply. and there is
no evidence of current saturation with increasing salvation.
It was hoped that some information concerning the
nature of the charge carriers could be obtained from the
resistance-voltage plots thus allowing an inference to be
made about the other materials whose high resistances
prevented electrolysis experiments. This has however seemed
to be a futile endeavor. The plots are more characteristic
of the material than of the charge carriers. Except for
high solvation states in the methanol { hemoglobin system,
the plots are similar to the water / hemoglobin system.
Thus little can appear to be learned from the resistance-
voltage plots of the ethanol { hemoglobin system, Figure 30.
There even seem to be differences between cytochrome-c,
Figure 24, and hemoglobin, Figure 28, yet their pattern and
degree of charge carrier changes are quite similar.
Last of all, one should note that the difference

between the kinetics of the current increase from dryness

 

(n.2,! ..-I.
.

116
and the weight increase from dryness as a function of
time (illustrated in Figure in for one case) indicates
that the current increase is mainly effected by
conduction in the more superficial layers of the cry-
stallites, a previously suspected idea.

It must be pointed up, of course, that many questions
still remain, and the case has not been proved that
protonic conduction is intrinsic. Several methods by
which this could be demonstrated are possible. One such
method would involve the resolution of the activation
energy plots into the electronic and protonic components
and then shown that there indeed exists a U; + and a

63 '. This type of calculation has heretofore been
impossible to do because of the amount of variation in
the slopes of the lines in the activation energy plots
leading to poor values of the y-intercept. This is an
experimental difficulty and is not easily resolvable;
large extrapolations of the data are involved and small
errors become greatly magnified.

If attempted, the proceedure would simply involve
hydrating a sample to a known value‘of p/pO and then
finding its amount of regain from the adsorption isotherm
and the amount of protonic and electronic conduction
from the graphs in this thesis. The small errors in the
activation energy plots are most likely the result of
variations in the amount of adsorbed water during the
heating cycle. This evaporation of the water could partially

be reduced by increasing the pressure of the ambient (non-

 

 

_____...

 

ll?

adsorbed) gas surrounding the sample. One could pressurize
the chamber to two atmospheres of nitrogen, for example.
This would answer the question of the "intrinsic"
nature of the protonic charge carriers but would not
answer the question if the protons were being replaced
by the adsorbed vapor. This would be required for the
case where currents of a relatively large magnitude are
passed for long periods of time (milliamps for a few
minutes) because "intrinsic protons", those present in
the biomacromolecule, are not present in sufficient quantity
for sustained currents. By making use of a partial pressure
gauge, and an isotope label in the adsorbate (or adsorbent),
this question could be resolved by an examination of the
electrolysis products. This work is in progress at the
time of this writing.
Another unanswered question is the nature of the
charge carrier at low applied potentials. That unusual
effects are occurring (with respect to protonic conduction)
is evidenced by the non-ohmic behavior below 30 volts
applied and the low hydrogen evolution as measured in the
electrolysis apparatus. The high over-voltage for hydrOgen
production could result in the current carriers being either
electrons or some other ion, as yet unidentified. Ionic
carriers, however, present theoretical problems because of
their low concentratio n and their supposedly low mobility.
Again, the partial pressure gauge could prove useful if a

gaseous product was involved.

Fe

II.

 

118

Interesting questions arise with respect to the
temperature dependence of the charge carrier species.
Does the amount of protonic conduction increase in the
biomacromolecules with temperature as was found for nylon-
66? Higher temperatures would also increase the conductance
and would thus allow measurements to be made at lower
solvation ranges. If the activation energy for the T]
production of protons and electrons is equal ( or approx- =

imately so), as is suggested here, there should not be

""1 -“.l.“.'§.‘£‘.l'id§z.. ‘ux. '; .

found any change in the electronic/protonic charge carrier

 

ratio with temperature, only with solvation. This could be
an important point for poikilotherms if electronic processes

are to be important in biological systems.

-——.,

jig

BIBLIOGRAPHY

 

120

BIBLIOGRAPHY

Akamatu, H., and Inokuchi, H, H., (1950).J. Chem. Phys.
l8, 60.

Akamatu, H, Inokuchi, H., and Handa, T.,(1951). Nature.

g. 52.. ”.1
L

Arnold, W., and Sherwood, H. K., (1957). Proc. Natl. Acad. 3
8010 Us So a g 1050

Baxter, S., (1943). Trans. Faraday Soc., 39, 207.
Bayliss, K. S., (1948). J. Chem. Phys., lg, 287.

 

1."Klu'm.).'~“_I -*P f...

Bradley, R. S., (1936). J. Chem. Soc., 1&67.

Bradley, D. F. and Calvin, M., (1955). Proc. Natl. Acad.
Sci. U. 3., El, 563.

Brausse, G., Meyer, A., Nedetzka, T., Schecht, H., and
Vogel, H., (1968). J. Phys. Chem., 2;, 3098.

Brunauer, S.,(19h3). The Adsorption of Gases and Vapors,
Princton University Press, Princton.

Brunauer, S., Emmett, P. H., and Teller, E., (1938). J.
Am. Chem. Soc., éQJ 309.

Burnel, M. E., Eley, D. D., and Subramanyan, V.,(l969),
Proc. N. Y. Acad. Sci., 158, 191.

Calvin, M., and Androes, G. M., (1962). Science, 138, 867.

Calvin, M., Kurtz, D., and Ruby, R. H., (1965). Recent
Progress in Photobiology, (E. J. Brown, ed.) Blackwell,
Oxford, 225.

Cardew, M. H., and Eley, D. D., (1958). Fundamental Aspects
of the Dehydration of Foodstuffs. (Society of Chemical
Industry). Macmillan, New York, 2h.

Cardew, M. H., and Eley, D. D., (1959). Faraday Soc. Dis-
cussions, 21, 115.

Chapman, D., and Fast, P. G., (1968). Science., 160, 188.

121

Conway, E. J., (1953). Inter. Rev. Cytology., g, 419.
COpe, F. w., (1963). Arch. Biochem. BiOphysics. 103, 352.
COpe, F. W., (1965). Oxidase and Related Redox Systems, Vol.1,
(T. S. King, H. S. Mason, and m. Morrison, eds.),
Wiley, New York, 51.
Crozier, w. G., (1925). J. Gen. Physiol., 2, 189.

Davies, R.E., (1960). Membrane Transport and Metabolism.
Proc. Sym. Prague, Academic Press, New York, 320.

DeBoer J.H. and Zwikker, C., (1929). Z. Physik. Chem., B3,
407.

DeVault, D., Parkes, J. H., and Chance, B., (1967). Nature,
215, 642.

Digby, P. J. B., (1965). Proc. Roy. Soc. (London), B161, 504.

Dixon, M. and Webb, E. G., (1958). The Enzymes, Chapter IV.
Academic Press, New York.

Eley, D. D. and Leslie, R. B., (1963). Nature. 192, 898.

Eley, D. D. and Leslie, R. B., (1964). Adv. Chem. Physics.
(J. Duchesne, ed.). 2; 238.

Eley, D. D. and Leslie, R. B. , (1966). Trans. Faraday Soc.,
6;, 1002.

Eley, D. D. and Parfitt, G. D., (1958). Trans. Faraday Soc.
§_1_, 1529.

Eley, D. D., Parfitt, G. D., Perry, M. J. and Taysum, D. H.,
(1953). Trans. Faraday Soc., 49, 79.

Eley, D. D. and Snart, R. S. (1965). Biochem. BiOphys. Acta,
102. 379-

Eley, D. D. and Spivey, D. I., (1960 a). Nature, 188, 725.

Eley, D. D. and Spivey, D. I., (1960 b). Trans. Faraday Soc.,
56. 114‘32.

Eley, D. D., and Spivey, D. I., (1962). Trans. Faraday Soc.,
1, 188.

Evershed, S., (1914). J. Inst. El. Eng. (London). 2, 51.

Falk, M., Hartman Jr., K. A., and Lord R. G., (1963 a), J.
Am. Chem. Soc., 85, 387.

 

122
Falk, M., Hartman Jr., K. A., and Lord, R. C., (1963 b),
J. Am. Chem. Soc., 85, 391.
Franklin, R.E., and Gosling, R.G., (1953), Nature, £12, 156.
King, G. and Medley, J. A., (1949 a). J. Colloid Sci., 3.9.
King, G. and Medley, J. A., (1949 b). J. Colloid Sci., 4,1.
Ladik, J. (1968). Studia Biophysica, 8. 87.

LuValle, J. E., and Goddard, D. B., (1948). Quart. Rev.
Biol., 23, 197.

Maricic, s.. and Pifat, G., (1966). Electrochemische Methoden
und Prinzipien in der Molekular-Biolo ie, III Jena
Symposium, Akademic Verlang, Berlin, 3.

Maricioz S., Pifat, G. and Pravdid, V., (1964). Biochem.
BiOphys. Acta., 29, 293.

Marsh, M. C., and Earp, K., (1933). Trans. Faraday Soc.,

32. 173-

Misra, T. N., Rosenberg, B., and Switzer, R. C., (1968).
Jo Chemo PhySo’ fl. 2096.

Mitchell, P., (1961). Nature, ;9_, 144.

Murphy, E. J., (1929 a). J. Phys. Chem., 33, 200.
Murphy, E.J., (1929 b). J. Phys. Chem., 33, 509.
Murphy, E. J., (1960). Phys. Chem. Solids, $6, 115.
Murphy, E. J., (1963). Canadian J. Physics, 4;, 1002.
Murphy, E. J., (1965). Ann. M. Y. Acad. Sci., ;;§, 725.

Murphy, E. J., and Walker, A. C., (1928). J. Phys. Chem.,
3_2_, 1761.

O'Konski, C. T., Moser, P., and Shirai, M., (1964). BiOpolymers,
Symposium No. 1, 465.

Pearson, G. L., and Bardeen, J., (1959). Phys. Rev., 15, 865.

Pollack, J., and Ubbelohde, A., (1956). Faraday Soc. Trans.,
5;, 11120

Postow, E., (1968). Ph. D. Thesis, Michigan State University.

Powell, M. R. (1966). Master's Thesis, Michigan State
University.

 

123

Remmer, H., and Merker, (1965). Ann. M. Y. Acad. Sci., 123,
79-

Roginsky, S. 2., and Zeldovich, J., (1943). Acta Physico-
chemica U. S. S. R., l, 449.

Rosen, D., (1963). Trans. Faraday Soc., 59, 2178.

Rosenberg, B., (1962 a). Nature, 193, 364.

Rosenberg, B., (1962 b). J. Chem. Phys., 6, 816.

Rosenberg, B., (1964 a), BiOpolymers, Symposium Mo. 1, 453. !

Rosenberg, B., (1964 b), Oxidases and Related Redox Systems, hi
Vol. I, (T. S. King, H. S. Mason, and M. Morrison, eds.) '
1964, John Wiley and Sons, New York, 72.

Rosenberg, B., (1966). Advan. Radiat. Biol., g, 193. g

Rosenberg, B., and Pant, H., (1969). unpublished results.

 

Rosenberg, B. and Postw, B., (1969). Ann. N. Y. Acad. Sci.,
158, 161.

Sklenar, H., Ladik, J., and Biczo, G., (1967). Studia
BiOphysica, 6, 45.

Szent-Gyorgyi, A., (1941). Nature, 148, 157.

Takashima, S., and Schwan, H. P., (1965). J. Phys. Chem.,
62, 4176.

Taylor, C. P. S., (1960). Ph. D. Thesis, University of
Pennsylvania.

Tayloi,6H. A., and Thom, N., (1952). J. Am. Chem. Soc.. 21.
1 9.

Thomas, J. M., Evans, J. R. N., Lewis, T. J., and Seeker,
P., (1969). To be published.

Tien, H. T., (1968). J. Phys. Chem., 2;, 4512.

Vol'kenstein, F. F., (1963). The Electronic Theory of
Catalysis on Semiconductors. Pergamon Press, Oxford.

Weber, G., (1955). Disc. Faraday Soc., 29, 156.
Williams, R. J. P., (1968). R. C. I. Reviews, A, 13.

Young, J. R., and Whetten, N. R., (1960). Rev. Sci. Instr.
1;, 1112.