|
Items in cart [1] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 120
The Mechanism of Masonry Decay
Through Crystallization
SEYMOUR Z. LEWIN
One of the most common and extensive sources of deterioration of stone,
brick, mortar, plaster, and concrete is the consequence of crystallization phe-
nomena that take place in pores, channels, and cracks at and near exposed
surfaces. Liquid water deposits dissolved matter wherever evaporation occurs.
The site of this crystallization is determined by the dynamic balance between
the rate of escape of water from the surface and the rate of resupply of solution
to that site. The former is a function of temperature, air humidity, and local
air currents. The latter is controlled by surface tension, pore radii, viscosity,
and the path length from the source of the solution to the site of the evapo-
ration.
The detailed nature of this balance determines the form that the decay will
take. If the rate of resupply of solution to the surface is sufficient to keep pace
with the rate of evaporation, the solute deposits on the external surface and
is characterized as an efflorescence. If the rate of migration of solution through
the pores of the masonry does not bring fresh liquid to the surface as rapidly
as the vapor departs, a dry zone develops just beneath the surface. Solute is
then deposited within the stone at the boundary between the wet and dry
regions, generating spells, flakes, or blisters.
The site of crystal deposition can be predicted by applying the physical-
chemical laws governing capillarity, viscous flow, and diffusion. These con-
siderations disclose the quantitative relationship between the porosity of the
masonry and the dimensions of the flakes, blisters, or spells that develop, as
well as the manner in which the decay progresses.
Data from controlled experiments in which salt decay is induced in labo-
ratory specimens, together with measurements on examples of salt decay in
buildings and monuments in a variety of environments, confirm the validity
of these insights.
Seymour Z. Lewis Professor, Department of Chemistry, New York University.
120
.
OCR for page 121
Mechanism of Masonry Decay Through Crystallization
121
Exposed stone and other masonry materials are subject to a number
of deteriorating influences, chief among which are the effects of crys-
tallization, freezing, acidic attack, and mechanical erosion. The reality
and ubiquity of the phenomenon termed "salt decay" are recognized
by many of those concerned with the conservation of buildings and
monurnents,~-3 but the detailed- mechanism by which the crystalli-
zation of waterborne substances can break up the surface of a some-
what porous solid has not hitherto been objectively demonstrated.
When water at 0° C changes into ice, there is a volume increase of
9 percent. If liquid water is confined in a pore or crack, and this phase
transformation takes place, it is evident that the resulting expansive
force can damage the host solid.
It is also clear that susceptible materials can be dissolved by acidic
substances generated from air pollutants {e.g., fossil-fuel combustion
products!, microorganisms, associated minerals (e.g., sulfides that undergo
oxidation), or vegetation. Such attack can destroy surface modeling
and sculpted details and weaken internal induration that binds the
grains of the solid together. Similarly, the manner in which mechanical
abrasion {e.g., the sandblasting effect of wind-driven dust) erodes a
surface is readily visualized.
However, it is not irnrnediately evident why the deposition of a
solute from a solution into a pore or crack at the surface of a solid
should damage the latter. Consider, for example, the evaporation of a
sodium chloride solution at-a stone surface. When, as a consequence
of the escape of water vapor, the solution reaches saturation, it contains
at ordinary temperatures about 26 percent solid matter by weight.
Hence, a pore filled with such a solution con have only about one-
quarter of its volume taken up by the residue left when evaporation
is complete. Each repeated imbibition of salt solution can be expected
to reduce the remaining free volume of the pore by one quarter of that
value, until the pore is filled with deposited solute. But there is no
analogy in this process to the expansive force that develops when water
filling a pore transforms into ice or when certain types of solid phases
filling a pore recrystallize into higher hydrates (as, for example, when
sodium sulfate (Na2SO4) transforms at -high humidity into Na2SO4 10
H2O).
Nevertheless it is a fact that the deposition of a simple, nonhydrat-
able solid, such as sodium chloride, during the evaporation of its so-
lution in the pores of stone and masonry, can disrupt the solid. The
external manifestations of this disruption are similar to those produced
by the freezing of water in the pores of the surface of the solid scaling,
flaking, and blistering and/or crumbling of the surface.
..
OCR for page 122
22
CONSERVATION OF HISTORIC STONE BUILDINGS
It is the..purpose of this paper toiinvestigate whether the so-called
salt decay is due solely to the deposition of solute at a stone surface,
to determine the conditions under which salt decay occurs, and to
establish the quantitative relationships between the type of decay ob-
served and the physical properties of the liquid and solid phases in-
volved.
THEORY
The experimental section of this paper demonstrates that salt decay
. . . . . . .. . . . . . . . ...
occurs only when solute.is deposited within the pores ot the solid
that is, a certain distance beneath the external surface (usually a frac-
tion of a millimeter to a few millimeters). This can occur when- the
rate at which~water departs from the surface of the solid via evaporation
is equal to the rate at which fresh solution is brought to the surface
via. migration through the internal capillary system of the solid.
If migration of solution to the surface is faster than the rate of cIrying,
then liquid oozes out onto the exposed surface, and solute is deposited
on top of that external surface. This corresponds to the formation of
visible efflorescences.4 5 Although they may be unsightly, . and usually
indicate that subsurface crystallization is occurring elsewhere, they
-are not, per se, damaging to the stone.
If the migration of solution toward the exposed surface is very slow,
then very little deposition of solutes takes place. Whatever.deposition
does occur is deep within the stone and does not manifest itself in the
form of surface decay.
It is proposed herein that the necessary condition for surface decay
is the establishment of a steady state in which the rate of diffusion of
water through a thin layer of the porous solid at the surface is balanced
by the rate of replenishment of water to that site from.the source
freservoir) of the solution. The principle of this mechanism is depicted
schematically in Figure 1.
Evaporation by Diffusion
The ~ying-out of solution within a pore opening at the surface occurs
by diffusion of water vapor through a layer of thickness ~ centimeter
of the porous solid. The rate of diffusion, ~ grams per square centimeter
per second, is expressed by Fick's first law:
[ = D(dC/~, { 1 )
where D is the diffusion coefficient in cm2 see-i, and dC/dX is. the
concentration gradient across the diffusion layer.6
OCR for page 123
Mechanism of Masonry Decay Through Crystallization
r
I'
is'
FIGURE 1 Parameters involved in the proposed mechanism for masonry
deterioration due to deposition of solute from solution. The masonry, M, is
in contact with a reservoir of solution, S. Solute is deposited a distance ~
inside the stone at the height h above the reservoir. The radius of the pore
opening at the site of deposition is r; the average radius of the channel through
which solution migrates is R; the length of the migration path is L.
123
OCR for page 124
24
CONSERVATION OF HISTORIC STONE BUILDINGS
Because the air at the surface of exposed masonry is generally in
motion, the aqueous tension at the surface of the solid tends to be
constant, and the concentration gradient can be expressed in terms of
the difference in vapor pressures of the water at the solution surface,
Ps' and in the ambient air, Pa' divided by the diffusion-layer thickness,
a:
~ = D(PS — Pa)(M W./Nk~/~. (2)
In the steady state the rate of escape of water from 1 cm2 of exposed
surface of the porous solid is equal to the diffusion rate, I, times the
fraction of open area at the solid surface, Fs The latter is related to
the porosity of the solid and is typically between 0.05 and 0.40 for
natural stone and other masonry materials.7
Replenishment by Capillary Migration
Solution is drawn to the surface of the porous solid by capilIarity. The
interfacial tension, By, at the free surface of the liquid provides the
driving force that draws the liquid to the surface through the capillary
network from the source. The equilibrium pressure difference at the
liquid surface in a circular pore as a result of the interfacial tension is
given by the Laplace equation:
Ups = 2 ~ cos sir,
(3)
where ~ is the contact angle of wetting of the meniscus at the walls
of the pore, and r is the radius of the pore at the liquid surface.8 If there
is a distribution of pore sizes in the solid, an effective radius can be
adopted that represents the weighted average of the contributions of
the various pores to the resultant surface Strivings force.
This driving force draws solution to the surface to replace that which
departs via evaporation. The flow of liquid through the capillary net-
work under this driving force is governed by Poiseuflle's law:
AV ~ R4 AP
_ = . _
At 8~ L'
(4)
where R is the effective radius averaged over the total length L of the
capillary network through which the flow is occurring, V is the volume
of solution passing through 1 cm2 of pores in the time t, and ~ is the
viscosity in poise {g see-i cm-.9 The term AP/L is the total gradient
of pressure from the solution reservoir to the evaporation site.
OCR for page 125
Mechanism of Masonry Decay Through Crystallization
The driving force, UP, in an empty, uniform capillary of radius r is
given by equation 3 above. As the liquid rises in the capillary, the
driving force diminishes, since part of the surface pressure must provide
the hydrostatic pressure to support the column of liquid of height h:
125
~ ~ cos ~
Apnea = r —hPg.
1,5)
If there were no evaporation occurring, the liquid would rise in the
capillary until the surface pressure and hydrostatic pressure became
equal, i.e., apnea would be zero. When evaporation is occurring, a steady
state tends to be established in which Apnea has that value which
produces a Poiseuille flow just sufficient to balance the rate of escape
of liquid at the evaporation site.
The driving force in the Poiseuille equation involves the effective
radius r at the height h. The frictional force limiting the rate of flow
involves a different parameter, R. which describes an effective radius
for the entire length of capillaries through which the liquid moves to
get from the source to the evaporation site.
It is shown in the experimental section that, operationally for ma-
sonry, these two parameters generally will have quite different values.
The reason is as follows. Whereas the surface (driving) force involves
the inverse first power of the pore radius, the viscous "opposing) force
involves the capillary radius raised to the fourth power. Thus, for small
capillaries The condition for laminar flow is R ~ 1), the rate of Poi-
seuille flow falls very rapidly as the radius decreases. This has the
important practical consequence that in a porous solid containing a
range of pore sizes, only the upper part of the pore-size-distnbution
curve contributes significantly to the rate of flow of liquid to the
evaporation site during the times involved in the wet-to-dry cycling
of masonry in buildings and monuments. On the other hand, the sur-
face force increases inversely as the radius decreases. Therefore, pores
too small to participate significantly in the viscous flow do neverthe-
less make an important contribution to the net driving force drawing
the liquid through the capillary network.
The Steady State
In the steady state the rate of escape of water is equal to Fs times J.
The rate of replenishment is equal to the Poiseuille flow rate, ~V/At,
times the fractional area of the solid that consists of contributing
OCR for page 126
126
CONSERVATION OF HISTORIC STONE BUILDINGS
capillaries, Fp, times the liquid density, p, and weight fraction, Fw, of
water in the solution. Thus:
FsJ=FppFw(/\V//~15.
Substituting equations 2, 4, and 5 into equation 6 yields:
FSD(Ps — Pa)(M.W./Nk~
Fp pFw=R4 (25/cosU _ h pa)
\ r
and
(6)
17)
8 ~ L
8FsD~L(Ps — Pa) (M.W./Nk 1)
=-
Fp p Fw ~ R4 { ~ cos ~ - h p g}
r
/l8)
If, because of evaporation of water, solute crystallizes a distance ~
beneath the exposed surface of the porous solid, and if this is the source
of the deterioration of the surface, then equation 8 permits the quan-
titative prediction of the extent of the surface decay (i.e., the thickness
of the flake, blister, or powder layer). Such prediction can be made
directly and rigorously on the basis of the properties of the solid (po-
rosity, pore-size distribution), the solution (concentration, vapor pres-
sure, interfacial tension, viscosity, density), the solvent (diffusion coef-
ficient, molecular weight), and the environment (temperature, relative
humidity).
Detenoration from NaC! Crystallization
One of the common types of salt decay is that caused by deposition
of sodium chloride in stone and brick.3 The source of the salt may be
seawater or groundwater, deicing practices, or aerosol particles.
In this case the solution just below the exposed surface tends to be
a saturated sodium chloride solution; the temperature is the ambient
temperature, and the solution migrating within the solid is dilute. The
following values will be taken as fairly representatively of this system:
D = 0.22 cm2 see-i (for water vapor in air at 1 atm and 20° Call
Ps = 19 tort = 0.025 atm (for 5.3 M NaCI)
OCR for page 127
Mechanism of Masonry Decay Through Crystallization
Pa = 10 tort = 0.0125 atm (for air at 60% relative humidity)
M.W.=l~gmol-i
Nk = 82.06 cm3 atm mol- ~ deg-i
T
p
cos ~ = 1.00
= 293K
1.00 g cm-3
= 82.0 dyne cm-1 (for 5.3 M NaCl)
127
g = 980 dyne g- 1
h p g, the hydrostatic pressure term, will generally be negligible relative
to the surface pressure term (2 my cos Or.
Substituting these values into equation 8 yields the following result:
= 3.17 x 10 · L F {Fp R4)
{91
which permits the prediction of the thickness of surface decay that
will result from crystallization of sodium chloride. The prediction
employs no arbitrary, empirical parameters; it is based on data on the
location of the decay zone {L), the concentration of the salt in the
reservoir of solution (~ /Fw), and the pore characteristics of the solid
(Fs r/Fp R4~.
EXPERIMENTAL TEST OF THEORY
Design of the Experiment
The validity of equation 8 has been tested by a series of laboratory
experiments in which the conditions during the deposition of sodium
chloride at an exposed stone surface were controlled and measured.
The experimental arrangement is shown schematically in Figure 2.
A rectangular sandstone column, 60 cm x 2.5 cm x 5 cm, was
mounted in a glass vessel inside a Plexiglas box with its lower 5 cm
immersed in a sodium chloride solution. The neck of the glass vessel
was sealed with a plug of paraffin wax 1 cm thick so that liquid could
not migrate up the external surface of the stone column. This served
to confine all liquid migration to the internal capillary network of the
stone. Salt solutions of known concentrations were fed into the glass
vessel at the rate necessary to maintain the liquid there at constant
level. A constant, uniform flow of air at 60 percent relative humidity
OCR for page 128
128
.
CONSERVATION OF HISTORIC STONE BUILDINGS
;
~ 1
NaCI
Sol 'a
Glass Beads
for Support
1 1
_. ..
it'=
~wc
~ ~ ~ ~ _~ _
J I
Stone
Column
Paraffin Wax
Seal
Air at 20 C and 60% R.H.
1 ~
FIGURE 2 Experimental arrangement for producing salt decay.
Lower end of masonry column is immersed in a salt solution, which
is able to reach the exposed surface only by capillary rise through
the interior of the stone. A uniform flow of air at controlled tem-
perature and relative humidity is maintained over the exposed sur-
face of the stone.
and 20° C was maintained over the exposed surface of the stone col-
umn.
Under these conditions a steady state was established within six
to eight hours in which the solution migrated upward through the
interior of the stone column to the exposed surfaces where the water
evaporated, depositing the sodium chloride. The lower part of the
OCR for page 129
Mechanism of Masonry Decay Through Crystallization
column's surface received solution faster than it dried out, and a
heavy deposit of salt formed on the stone. With increasing distance
from the reservoir of solution, the thickness of the salt deposit di-
minished until at a certain height the surface of the stone appeared
to be darker than the part above it. This indicated that the pores in
the darker surface contained some liquid, but~very little salt was
visible on the outer surface. The typical appearance of the stone
column after a two-week run with a saturated salt solution is shown
in Figure 3a and with a half-saturated solution in Figure 3b.
The stone was damaged only in the region where the external
deposition of salt had diminished to minute amounts that is, where
the rate of arrival of solution at the exposed surface was approxi-
mately equal to the rate of evaporation of the water, so that salt
deposited within the surface rather than on ton of it. The damaged
129
~ . . , ~
~ ~ ~ . ~ ~ ~ . .' ~ .1 1 r - . 1
surface layer ot stone was nelo togetner ny tne sunsurtace crystals
of sodium chloride. However, when the salt deposit was washed
away, the area of decay became readily apparent as can be seen in
Figure 4.
This type of experiment was conducted on the same sandstone
column employing solutions of sodium chloride at three different
concentrations. Before each run, the stone column was removed
from the apparatus, washed free of deposited salt, soaked for two
weeks in daily changes of distilled water to remove any remnants
of the previously imbibed salt solution, and dried. After each run,
the site and depth of the surface decay were measured. The location
and thickness of the decayed zone were different for each of the
different salt concentrations, as can be seen in Figure 5.
Experimental Data
Effective PoiseuiRe Rachus
The data recorded in these controlled salt-deposition experiments rel-
ative to the surface decay are summanzed in Table 1. To use these
results to test the theory, it is necessary to evaluate the porosity of
the stone. A type of measurement proposed here as particularly useful
in this respect is the record of the rate of water imbibition as a function
of the time of immersion of a test block. The data for the Longmeadow,
New Hampshire, ferruginous sandstone employed in this work are
shown in Figure 6.
The total porosity is estimated from (a) the volume of water imbibed
after extremely long room-temperature immersion, or, equivalently,
(b) that which is taken up during five hours of immersion in boiling
OCR for page 130
~ - -
- -
~ l ~——
- -
O a'
. ~ ~ ~ - -
e~
it ~ ~
~ ~ o o ~
; _ ~ y m-
~ o ~ o
~4
_ ~~ -3 o,= o
~~.~ +, I ~ ~: )? ~ ~~ ~ ~ .
c - 8 ~
° ~ ~ Am.
°
3 ~-
OCR for page 134
134
8
7
6
5
hi:
m
m
CONSERVATION OF HISTORIC STONE BUILDINGS
4
Maximum Water Absorption
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
-
n
1
a I I I '` ~ I I o
o 5 10 15 20 35 40
[TIME, hours] 1/2 ,
o 5 10 15
FIGURE 6 The rate of imbibition of water by a test block of New Hampshire sandstone
yields information about the effective porosity for liquid flow through the internal
capillary network. There is an initial rapid gain in weight due to the filling of the larger
capillaries, then a very slow, diffusion-controlled process that requires marry months
(in fact, more than three years) to fill all the interior spaces. Immersion of the same test
block in boiling water for five hours, or in room-temperature water after exhaustive
evacuation, yields the weight increase shown as "maximum water absorption." The
fully water-saturated test block, allowed to air-dry, loses water much more rapidly and
completely than it had imbibed-the water, and by a different mechanism, showing that
evaporation occurs from most of the pores at the surface, whereas liquid migration
occurs mainly through the larger capillaries. The test block was 7.2 x 5.5 x 15.4 cm;
its dry weight was 1338.8 g; its dry density was 2.20 g/cm3.
imbibition per 100 g of stone. Ibis value yields an effective fraction
of intemal cross-sectional area participating in Poiseuflle transport of:
Fp= {~2.20~0.042~2/3
= 0.20 cm2 capillary area per 1 cm2 of {11)
stone cross section.
From the comparison of total pore area with the area participating
in Poiseuille flow (0.20/0.32), it follows that the upper 0.63 of the pore-
OCR for page 135
Mechanism of Masonry Decay Through Crystallization
135
size-distribution curve should be considered in evaluating the effective
radius, R. to be employed in equation 8. For the present study, this
part of the curve has been estimated from the scanning electron mi-
crographs, representative examples of which are reproduced in Fi-gure
7.
The frequencies of occurrence in 3-,um-wide pore-size intervals per
1 cm2 of cross section in fracture surfaces have been estimated; these
frequencies, multiplied by the average radius in the interval, have been
raised to the fourth power; and the weighted average has been com-
puted. The result is that the effective pore radius for Poiseuille flow
in this stone is estimated as:
R = {1.6 + 0.21 x 10-3 cm.
(12)
In this averaging technique the smallest pores those whose di-
mensions were less than 0.1 Am—were not included. This does not
seriously affect the validity of the resulting average, since, as has been
shown, such small pores do not contribute significantly to the observed
flow rate.
An alternative method of estimating the effective radius for Poi-
seuille flow would be to measure the rate of effusion of liquid through
a plug of the stone of known dimensions under a controlled driving
force and divide by the number of capillaries contributing to the total
flow. This approach would also involve microscopic detection and
counting of pores in cross sections of the stone and does not appear
to offer any advantages of precision or convenience over the technique
adopted in the present work.
Effective Laplace Radius
It remains now to estimate the effective pore radius at the stone surface
that determines the surface (driving) force. This value is most reliably
derived from the rate of advance of liquid through the capillary network
of the stone when no evaporation is taking place, and when the hy-
drostatic (retarding) force is negligible. Under these conditions the
Washburn equational applies:
~ ~ '
1,13)
where x is the distance that the interfacial tension, A, draws the liquid
of viscosity, A, through a capillary of radius, r, in time, t. Figure 8
OCR for page 136
136 CONSERVATION OF HISTORIC STONE BUlEDINGS
-
FIGURE 7 Scanning electron micrographs showing the internal pore character of t
New Hampshire sandstone. Magnification employed to estimate frequencies of occur-
rences of pores of average radius between: 7a 0.05 and 0.005 mm; 7b 10.0 and 1.0 ,um;
7c 5.0 and 0.5 ~m; 7d 1.0 and 0.1 ,um.
OCR for page 137
Mechanism of Masonry Decay Through Crystallization 137
d
FIGURE 7 Continued
OCR for page 138
38
CONSERVATION OF HISTORIC STONE BUILDINGS
Sat'd. Vapor Enclosure
- Ru ler
Column Support
T-- -1 _
\ — — _ I
\ ~ ,
Liquid ~ n—~ umn
Constant
Flow Device :]
. _
_
Constant ~ , _ _
Level ~ _~ ~ <, ___
,.....
_
- Ruler
Holder
1
FIGURE 8 Experimental arrangement for determining the Laplace radius effective in
generating the surface pressure in a masonry specimen. The height, x, to which liquid
has risen in the stone column is observed visually by means of the darkening effect of
wetting as a function of time, t, of contact with the bulk liquid. The inner bent tube
in the constant flow device ensures that the position of contact of the liquid source
with the stone column remains constant.
shows a convenient arrangement for carrying out this type of mea-
surement,~3 and Figure 9 shows the data obtained for the sandstone
under study.
The slope, m, of the graph of x versus tile is given by:
~ i/2
m= 2Y ~ (rcos0~/2,
(14)
OCR for page 139
J
10
8
6
4
f
, ~
~~ m = A/ yr/271
O At. I
0 2
-
4 6 8 10 12 14 16 18 20
a,
FIGURE 9 Graph of height of capillary rise, x, versus tl'2 of contact of a New Hampshire
sandstone column with a reservoir of 5.3 M NaC1 solution. The slope of the straight
portion of the curve yields the effective Laplace radius, r = 0.28 ,um Contact angle taken
to be zerol.
OCR for page 140
140
CONSERVATION OF HISTORIC STONE BUILDINGS
chloride, since these properties refer to the solution that is crystallizing
at the stone surface and not to the solution in the interior, which may
be more dilute. The internal path length, L, has been taken as equal
to the vertical distance from the source of the solution to the decay
zone (i.e., equal to hi. That is, the tortuosity factor is taken as equal
to unity for this rather porous solid.~4 The contact angle for the aqueous
solutions against the polar x-quartz surfaces of this stone's pores is
taken as zero degrees, and the interfacial tension is taken as that of a
saturated NaCl-glass interface.
The remaining relevant data are collected in Table 1, which also
compares the values predicted by equation 8 with those observed ex-
perimentally. The largest source of uncertainty in this test of the theory
is the estimation of the effective Poiseuflle radius, R. It will be noted
that the experimental results agree very satisfactorily with the pre-
dicted values, within the precision of the data.
In these experiments the order of magnitude of the thickness of the
deteriorated surface layer resulting from these salt solutions in this
particular sandstone proves to be between a fraction of a millimeter
and one or two millimeters. Our observations, and those of others, in
studies of the decay of exposed stone and masonry in buildings and
monuments have disclosed that the natural decay of many other sand-
stones and other types of natural stone and masonry results in surface
losses of the same order of magnitude. That is, when initial salt decay
is indicated i.e., when a single layer of stone has been lifted up in
the form of a blister, spell, or flake—the thickness of the deterioration
is in the vicinity of a millimeter. Thus, it appears that the parameters
characteristic of the present experimental setup are similar to those
commonly encountered in practical instances of crystallization-in-
duced deterioration of masonry.
CONCLUSIONS
The present work demonstrates that the mechanism of the so-called
salt decay of exposed stone and masonry consists in the deposition of
solutes from solution within the pores of the solid close to the surface.
This is characteristically manifested in the form of a thin layer of the
surface that lifts up in the form of a blister, peels outward as a spell,
flakes off, or powders away. The initial thickness of this surface decay
is of the order of a millimeter. When this thickness of surface has
separated, the decay process may be initiated again in the underlying,
still sound stone, resulting in a second such decay layer under the first.
The processes can then proceed again beneath these layers, and so on.
OCR for page 141
Mechanism of Masonry Decay Through Crystallization
141
In some cases, many successive layers of decay can be recognized, all
of them similar in character, with thicknesses from a fraction of a
millimeter to 1 or 2mm. An example of the occurrence of blisters 1
mm thick on exposed granite is shown in Figure 10; examples of the
multiplication of decay layers, progressing from the outer surface of
the exposed stone toward the interior, are shown in Figure 11.
The necessary condition for the occurrence of this type of decay
is the development of a steady state at the exposed surface, wherein
FIGURE 10 Granite surface stone in the lower course
of a New York City landmark building has been sub-
jected to the action of salt used in de-icing the adjacent
street. The surface has lifted up In numerous places,
forming blisters with a layer thickness that ranges from
0.5 to 1.5 mm.
OCR for page 142
142
_
FIGURE lla Cross sec-
tion of the surface layers
of a salt-decayed sand-
stone.
CONSERVATION OF HISTORIC STONE BUILDINGS
_> _
_
FIGURE llb A sandstone sculpture in the sculpture garden of the Brooklyn Museum,
New York City, showing the development of multiple layers of surface decay, resulting
from successive salt-decay processes proceeding from the outside inward.:
the rate of evaporation of water via diffusion through a layer of the
porous solid is balanced by the viscous flow of solution from the
reservoir to that site through the internal capillary network. The
quantitative relationship between the thickness of surface deterio-
ration and the characteristics of the liquid and solid media are de-
rivable from classical physical chemistry via the Fick and Poiseuille
laws. The parameters needed to describe the porous nature of the
OCR for page 143
Mechanism of Masonry Decay Through Crystallization
143
solid are obtainable from water imbibition, capillary rise, and pore-
size-distribution measurements. Laboratory experiments conducted
under controlled conditions yield results that are in good agreement
with the predictions of this theory.
These considerations, and the related experimental observations,
establish beyond reasonable doubt that the deposition from solution
of a simple, nonhydrated salt, such as sodium chloride, in the pores
at the surface of a stone generates pressures sufficient to break down
the induration. We are convinced of the reality of the phenomenon
and can now account for it in detail and predict where and under
what conditions it will occur. We do not yet understand how the
requisite disruptive pressures can be developed in the pores of the
solid. The fundamental question that remains to be addressed is:
How do crystals that have grown from a solution until they fill the
volume of a pore continue to grow at the areas of direct contact
between crystal and pore wall?
REFERENCES AND NOTES
1. S.Z. Lewin and A.E. Charola, Scanning Electron Microscopy in the Diagnosis of
"Diseased" Stone, Scarming Electron Microscopy, 1978, vol. I, pp. 695-703, SEM, AMP
O'Hare, Ill.
2. A.E. Charola and S.Z. Lewin, Examples of Stone Decay Due to Salt Efflorescence,
Third International Congress on the Deterioration and Preservation of Stones, Venice,
24 October 1979, in press.
3. S.Z. Lewin and A.E. Charola, The Physical Chemistry of Deteriorated Brick and
Its Impregnation Technique, Congress for the Brick of Venice, 22 October 1979, Venice,
Proceed~ngs, pp. 189-214, University of Venice, Italy.
4. S.Z. Lewin and A.E. Charola, Aspects of Crystal Growth and Recrystallization
Mechanisms as Revealed by Scanning Electron Microscopy, Scanning Electron M~cros-
copy, 1980, vol. I, pp. 551-558, SEM, AMP O'Hare, Ill.
5. A.E. Charola and S.Z. Lewin, Efflorescences on Building Stones; SEM in the
Characterization and Elucidation of the Mechanisms of Fonnation, Scarming Electron
Microscopy, 1979, vol. I, pp. 379~87, SEM, AMP O'Hare, Ill.
6. W. lost, Diffusion in Solids, Laquids, Gases, Academic Press, N.Y., 1952, 8ff.
7. A.E. Scheidegger, The Physics of Flow Through Porous Media, Univ. of Toronto
Press, 1960, p. 13.
8. F.A.L. Dullien and V.K. Batra, Determination of the Structure of Porous Media,
In Flow Through Porous Media, American Chemical Society, Washington, D.C., 1970,
17ff.
9. T.L. Poiseuille, Gompt. Rend., 11, 961 jl8401; 12, 112 {1841~; 15, 1167 {18421.
10. There is some uncertainty concerning the appropriate value of the contact angle,
0, for this system. The data of D.D. Eley and D.C. Pepper, Trans. Faraday Soc., 42, 697-
702 {19461, suggest that ~ = 0 degrees for water In plugs of powdered Pyrex glass. The
data of B.V. Deryagin, M.K. Melr~ikova, and V.I. Krylova, Colloid I. USSR, 14, 459 (19521,
suggest ~ = 60 to 70 degrees for water spreading through a packing of quartz sand.
OCR for page 144
144
CONSERVATION OF HISTORIC STONE BUILDINGS
However, advancing contact angles tend to be different from receding angles, which
tend to be zero; of. W. Rose and R.W. Heins, I. Colloid Sci., 17, 39 {19621.
11. A. Winkelmann, Wied. Ann., 22, 1, 152 ~18841; 23, 203 {1884i; 26, 105 {18851; 33,
445 ~18881; 36, 92 {18891.
12. E.W. Washburn, Phys. Rev., 17, 27~83 jl9211; see also V.G. Levieh, Physico-
chemical Hydrodynamics, Prentiee-Hall, Englewood Cliffs, NJ., 1962, pp. 382-383.
13. J.N. Chan, Gypsum Plaster as a Prototype in the Study of the Physical Chemistry
of Solid Porous Media, Ph.D. thesis, New York University, October 1980.
14. The tortuosity factor, T. is defined empirically as the correction factor needed to
make the calculations for certain theoretical models of pore structure agree with ex-
perimental data. F.A.L. Dullien calculates, based on his particular model of porosity,
that the tortuosity for sandstone would be in the range of 1.5 to 1.7; see AIChE [. 21,
299 {1975i. However, he points out that "using different models of pore structure, widely
different values may be obtained for T. some of which completely lack any physical
meaning." {See Porous Media Fluid Transport and Pore Structure, Academic Press, N.Y.,
1979, p. 22,71. It may be noted that if L is significantly greater than h (i.e., T > 1l, the
calculated value of ~ will be too large by that tortuosity factor. This effect is in the
opposite direction and would tend to offset any overestimation of ~ due to the contact-
angle factor Cf. ref. 101.
Representative terms from entire chapter:
exposed surface
