CEN Tech.J., vol. 5(1), 1991, pp.

81–90

The Formation and

Cooling of Dykes
DR ANDREW SNELLING

INTRODUCTION a distance of transport of only about 30 kilometres.

In order to define the combination of variables com­
When molten (igneous) rock forces its way, or is patible with dyke formation by magma intrusion, Szekely
squeezed up, into a vertical or near-vertical crack in other and Reitan considered a simplified physical picture, as
rocks, be they flat-lying sedimentary rocks such as sand­ shown in Figure 1. A fissure through the earth’s crust is
stone or shale, or folded metamorphic rocks such as schist represented by a rectangular slot of width W, which
or gneiss, or other igneous rocks, the crack is usually connects a magma chamber or source (with the magma at
forced apart and the molten rock cools in the space to form h = 0 when t = 0) to the free surface at h = H. At time t =
a tabular igneous intrusion cutting across the surrounding 0, which corresponds to the opening of the fissure, the
rocks that is known as a dyke (or dike, depending on which static pressure acting on the magma ΔΡ will force the
country you live in!). The molten rock (magma) crystal­ magma upward through the fissure.
lizes as it cools, the most common dyke rocks being As the magma rises it encounters wall rock of progres­
dolerite (or diabase to North Americans), basalt and fine-
grained granitic rocks. Although dyke formation and
magma cooling in dykes have not been observed and
timed, prolific studies have still been directed towards
understanding the mechanism of magma intrusion to form
dykes and the timing of magma cooling in them.
Issues such as the rate of cooling of magmas are
critical to young-earth creationists, since the timescales
usually envisaged for the cooling of such large bodies of
magma as granite batholiths are usually quoted in the
order of from hundreds of thousands of years to tens of
millions of years. Anti-creationists and theistic evolution­
ists have pointed to such apparent contradictions with the
young-earth timescale, thus insisting that young-earth
creationists and Flood geologists are ‘obviously’ wrong
and so their viewpoint is untenable. It is thus clearly
appropriate that the data on these topics be reviewed.
Dykes perhaps represent the simplest example of a cool­
ing magma body because of their limited dimensions, and
so are an appropriate place to start in our investigations.

MAGMA INTRUSION

Szekely and Reitan1maintain that the width of a dyke

places a limit on the possible vertical distance of magma
transport (that is, the distance from the source of the dyke
material), provided that the magma viscosity, and the
temperature and thermal properties of the surroundings
and the magma, are known or can be estimated. In many
cases dyke compositions indicate that the source of the Figure 1. Physical model for the intrusion of a dyke into a fissure from
magmas is the lowermost crust or uppermost mantle. The a magma chamber at depth below the crust (after Szekely
geological environment of emplacement may then dictate and Reitan1).
81
Formation and Cooling of Dykes

sively lower temperature. The resultant heat loss causes Within the framework of these assumptions and their
a substantial increase in viscosity and at least partial physical model, Szek e ly and Reitan concluded that the
crystallization. The increase in viscosity, the partial position of the top of the magma within the slot is readily
blockage of the slot, and the reduction in the effective described by a first-order differential equation which
driving force will hinder upward flow. Should crystalli­ relates the rate of rise of the magma to the width of the
zation be completed before the free surface (or any other dyke, the static pressure due to the overburden load, the
elevation of h above the magma chamber that is of magma density, the acceleration due to gravity, the dis­
interest) is reached, the mechanism is ruled out for the tance risen by the magma, and the position-dependent
given geometry. viscosity. The variation of viscosity as a function of
A rigorous formulation of the problem, where allow­ position in the crust is dependent on the following facts:
ance is made for the two-dimensional nature of the (1) The viscosity is an exponential function of tempera­
velocity and temperature fields, and simultaneously, ture;
account is taken of the interaction of the velocity field with (2) The ambient temperature increases with depth; and
rate of crystallization and inward-moving (from the walls (3) Viscosity is dependent on pressure, which increases
of the fissure) solidification front, results in a rather with depth.
complex set of partial differential equations. Thus Szek­ After integration of their first-order differential equa­
ely and Reitan considered it preferable, at least as a first tion, Szekely and Reitan derived a simplified equation
approximation, to examine the asymptotic behaviour of which they then used to perform computations, substitut­
the system. They first considered the rise of magma in a ing in appropriate property values:-
rectangular slot without crystallization and developed • The static pressure, ΔΡ = 8 x 109 dynes cm-2
expressions relating the rate of rise of the magma toward • The magma density, ρ = 2.7g cm-3
the free surface to other parameters, such as width of the • The viscosity at h = 0, μ0 = 105g cm-1sec-1
slot, viscosity, etc. Then they determined the time re­ • The viscosity at h = Η, μΗ= 1010g c m -1sec-1
quired for complete crystallization as a function of dyke Their computations were then used to construct
width, for given values of the latent heat of fusion and the curves relating the time of rise of the magma to the width
difference of temperature between the wall rock and the of a dyke for six different heights above the top of the
solidus temperature of the magma (that is, that tempera­ magma chamber (see Figure 2). Also shown in this figure
ture at which the originally liquid magma has completely are curves relating time required for complete crystalliza­
solidified). tion to the width of a dyke, for two different values of the
Using the parameters already depicted in Figure 1, at latent heat liberated by crystallization.
t = 0 magma enters the slot from the magma chamber and In their computation of the time of crystallization,
if: forced upward owing to the pressure of the overburden Szekely and Reitan assumed that heat transfer is by
acting on the walls of the magma chamber. To describe conduction only and solely in the direction perpendicular
the rise of the magma within the slot, Szek e ly and Reitan to the dyke in a semi-infinite medium. Parameters used
made the following assumptions: were:-
(1) The flow is laminar throughout; • The thermal diffusivity of the magma
(2) Entrance and inertial effects can be neglected; 4 x 10-3 cm2 sec-1
(3) There are no radial gradients within the slot; • The thermal diffusivity of the surrounding rock
(4) For the effect of gradual cooling of the magma 6 x 10-3 cm2 sec-1
throughout its rise, the viscosity can be considered as • The initial temperature of the magma, taken to repre­
a given function of the vertical position h, varying sent the liquidus temperature (the lowest temperature
over a specified permitted range; and at which the magma is still totally liquid)
(5) Frictional heat generated by viscous flow can be ne­ 1200°C
glected. • The temperature at which crystallization is assumed
Although restrictive, the assumptions represent a rea­ to be completed
sonable compromise between faithful representation of 1000°C
reality and the desired mathematical simplicity. Assump­ • The latent heat of fusion liberated uniformly over the
tions 1, 2 and 5 can be justified in hindsight from their temperature range 1200°C to 1000°C
computed results, which show that the rise time is only 150 or 100 cal g-1
significantly affected by the passage of the magma in the • The initial temperature of the surrounding rock
upper portion of the slot, where the assumptions would 600°C
appear to be valid. Assumption 3 is very reasonable for These assumptions made by Szek e ly and Reitan avoid
most cases, in view of the fact that the ratio between H (the underestim ating the time required for complete crystal­
height the magma rises from the magma chamber to the lization.
free surface) and W (the width of the dyke) is very much Assessing the results as plotted on Figure 2, it should
greater than 1. be noted that the region to the lower left of the lines
82
 Formation and Cooling of Dykes

Figure 2. Graph showing the time of rise o f magma versus dyke width W, with height of rise of the magma through the earth’s crust as the parameter
h, and also time of complete crystallization versus dyke width, with the latent heat o f crystallization as the parameter L (after Szekely and
Reitan1).

83
Formation and Cooling of Dykes

Figure 3. Graph showing the locus of intersections of height of rise of the magma h versus width of the dyke W in Figure 2, defining an ‘impossible’
region above and/or to the left of the curves and a ‘p ermitted’ region below and/or to the right of the curves (after Sze
k e l y and Reitan1).

labelled with values of h is ruled out, as the magma would dyke width for values of L, is shown in Figure 3. The
not have reached the height in the time available. Simi­ curves define an ‘impossible’ region to the upper left and
larly, the region to the upper left of the lines labelled with a ‘feasible’ region to the lower right.
values of L (the latent heat) is ruled out, since complete Of particular interest are the timescales involved in
crystallization of the dyke will have occurred. Critical the rise of the magma from the considerable depths of the
limiting values of dyke width for any combination of lowermost crust and for crystallization to occur. Since the
latent heat fusion and height above the magma chamber region to the left of the lines labelled with values of latent
are given by the intersections of the lines. The locus of heat L is ruled out, this imposes a m aximum time for the
such intersections, relating height of rise of magma to rise of a magma into a fissure and subsequent crystalliza­
84
 Formation and Cooling of Dykes

Figure 4. Graph showing the locus o f intersections between time in years for rise and cooling of the magma versus width of the dyke W as derived
from Figure 2, again defining an ‘impossible’ region above and/or to the left of the curves and a ‘p ermitted’ region below and/or to the right
of the curves.

tion. Taking the critical limiting values of dyke width for for the magma to rise from a depth of around 29 kilometres
any combination of latent heat fusion again (but this time and then crystallize is definitely less than five or six years,
the intersections of the lines give the timespans involved) and probably less than three years. Significantly, most
the locus of such intersections, relating time of rise and dykes are narrower than 12 metres, so this represents a
crystallization to dyke width for values of L, is shown in limiting timespan for dyke formation. Furthermore,
Figure 4. Again, the curves define an ‘impossible’ region Szekely and Reitan concluded that their approach, and
to the upper left and a ‘feasible’ region to the lower right. their selection of values for their computations, tended to
Comparing Figures 2 and 4, it is thus possible to underestimate.
calculate that for a dyke of width 12 metres, the time taken At dyke widths greater than 12 metres, Szekely and
85
Formation and Cooling of Dykes

Figure 5. Graph showing the family o f curves for the cooling history of a dyke of thickness D metres for which the intrusion temperature is 700°C, the
solidification temperature 500°C, and the latent heat of solidification is 80 cal g -1. The labels on the curves are distances from the centre of
the dyke measured as fractions of the thickness D (after Jaeger2).

Reitan’s analysis appears to be inadequate. On Figure 2 temperature, which, for simplicity, will be taken to be
it can be seen that it is impossible to relate the height of the the liquidus temperature appropriate to its composi­
magma rise parameter to the time and width parameters, tion. (There is good geological evidence that this is
but even a dyke as wide as 100 metres, ignoring the height frequently the case.)
of rise parameter, can be seen, according to the graph, as (5) The thermal properties of the country rock and solidi­
taking less than 300 or 400 years to form and crystallize. fied magma are the same and are independent of
temperature. (This assumption was again made for
COOLING TIMES simplicity, but it is a reasonable one in view of the
uncertainty of these quantities.)
In an earlier study, Jaeger2 modelled the cooling by (6) Heat removed by escaping volatiles need not be con­
conduction of an intrusive sheet of magma — a dyke (or sidered. (Since the effects of volatiles are dispersed
sill, in the case of a horizontal sheet) — taking into ac­ over a region whose thickness is many times that of
count the effect of heat of solidification. As in any the magma sheet, it is reasonable to assume that the
analysis of this type, a number of general assumptions heat conveyed by them is also dispersed over a large
must be made. Jaeger assumed that: region and so will make little contribution to the
(1) The intrusive body is in the form of a flat sheet temperature in the narrow region which is heated by
extending indefinitely in the perpendicular direction conduction from the cooling sheet.)
— a reasonable approximation to most dykes. (7) The effects of convection in the magma can be ne­
(2) The intrusion takes place rapidly, that is, in a time glected and cooling is by conduction only.
which is very short compared to the time of solidifi­ After developing a numerical solution from first prin­
cation. ciples, Jaeger presented his results as families of curves of
(3) The country rock is at constant temperature which temperature against time. Solutions for three typical cases
will be taken to be zero. (If it is not zero, Jaeger are presented in Figures 5, 6 and 7. In each case the
concluded that the results of his analysis should be temperatures are plotted as functions of the time in years
added to the initial country rock temperature.) for selected points of the dyke, measured from the centre
(4) The magma is supposed to be intruded at a constant of the dyke as distances 0 , 0.1D, 0.2D, 0.3D and 0.4D (in
86
 Formation and Cooling of Dykes

Figure 6. Graph showing the family of curves for the cooling history of a dyke of thickness D metres for which the intrusion temperature is 1000°C,
the solidification temperature 800°C, and the latent heat of solidification is 100 cal g -1. The labels on the curves are distances from the centre
of the dyke measured as fractions of the thickness D (after Jaeger2).

the dyke), 0.5D (which is the edge of the dyke and the 500°C, and can be read directly from the graph as 0.016D2
contact with the country rocks), and then 0.6D, 0.7D, years, where D is the thickness of the dyke. Thus a granitic
0.8D, 0.9D and 1.0D (in the country rock), where D is the dyke 1 metre wide would take 0.016 years to crystallize,
thickness of the dyke in metres. The timescale is ex­ a 2 metre wide granitic dyke would take 0.064 years, a 5
pressed as a factor of D2. metre wide granitic dyke would take 0.4 years, and a 10
Figure 5 represents the family of cooling curves for a metre wide granitic dyke would take 1.6 years to crystal­
typical granitic dyke, where the temperature of the lize, etc.
magma at the time of intrusion T2= 700°C, the tempera­ However, this does not take into account complete
ture of solidification T1= 500°C, the latent heat of solidi­ cooling to ambient temperatures from the temperature of
fication (or fusion) L =80 cal. g-1, and the density of the crystallization of 500°C. Once the granitic dyke reached
magma ρ = 2.7 g cm -3. The cooling time for complete that crystallization temperature at its centre, it would then
crystallization (or solidification) is the time taken for the take still further time for the dyke and the surrounding
temperature of the magma at the centre of the dyke to fall country rock to cool right down again to ambient tempera­
from its initial temperature T2 to the temperature of tures. This of course is depicted in the diagram by the
solidification T 1. In this case it is the time taken for the partial extension of the family of curves beyond the time
magma at the centre of the dyke to fall from 700°C to of crystallization. If the curves are projected (extrapo­
87
Formation and Cooling of Dykes

Figure 7. Graph showing the family of curves for the cooling history of a dyke of thickness D metres for which the intrusion temperature is 1100°C,
the solidification temperature 800°C, and the latent heat of solidification is 100 cal g -1. The labels on the curves are distances from the centre
of the dyke measured as fractions of the thickness D (after Jaeger2).

lated) even further until the dyke is completely cooled, the take 0.05 years, a 5 metre wide dyke would take 0.3 years,
time taken would appear to be about 0.09D2years. In the and a 10 metre wide dyke would take 1.2 years, etc.
case then of a 10 metre wide granitic dyke, complete Similarly, a dolerite dyke 1 metre wide intruded at 1100°C
cooling would appear to take about 9 years. would take 0.014 years to cool (crystallize completely) to
Figures 6 and 7 correspond to a typical dolerite (dia­ 800°C, a 2 metre wide dyke would take 0.056 years, a 5
base) dyke. In Figure 6 the initial temperature of the metre dyke would take 0.35 years, and a 10 metre wide
magma T2is 1000°C, while in Figure 7 it is 1100°C. Oth­ dyke would take 1.4 years, etc.
erwise, the other parameters — the temperature of Again, the dykes need to then cool from 800°C to
solidification T 1= 800°C, the latent heat of solidification ambient temperatures, and so some further time is needed
L = 100 cal g -1, the density of the magma ρ = 2.8g cm-3 — for complete cooling. If the families of curves are again
are the same in both examples. The times for complete projected until the dykes are completely cool, the time
solidification (crystallization) are 0.012D2 and 0.014D2 taken in both the examples of Figures 6 and 7 would again
years respectively. Thus, a dolerite dyke intruded at a be about 0.09D2years. Thus a 10 metre wide dolerite dyke
temperature of 1000°C which is 1 metre wide would take would also appear to take about 9 years to completely
0.012 years to cool to 800°C, a 2 metre wide dyke would cool.
88
 Formation and Cooling of Dykes

The results set out in Figures 5 , 6 and 7 apply to deeply but yet others consisting of a composite granite-basalt
buried dykes, where the magma has not reached the mixture. The latter are unusual, since basaltic and granitic
surface and there is a considerable amount of rock cover magmas are not normally known to co-exist like this. The
above the top of the dyke. Of course, nearer the surface geological setting of these composite dykes allows rea­
cooling will be more rapid, so the results of this analysis sonable estimates to be made of the time required for so­
represent an upper time limit on crystallization. In any lidification of each magma. Indeed, the solidification
case, Jaeger concluded that his results should be appli­ time had to have been short, otherwise the basaltic and
cable at depths below the surface greater than the thick­ granitic magmas in these composite dykes would have
ness of the dyke. begun to mix, which the field observations, petrographic
In a subsequent study, Jaeger3 confirmed his earlier observations and geochemical analyses indicate has only
calculations and results. His additional comments are of just begun on a very small scale.
importance here too. For example, he states that if Wiebe observed that at least one of these composite
volatiles (principally gases) are liberated on solidification dykes varied gradually over a distance of 25 metres along
and escape rapidly (for example, through cracks), the heat its length from homogeneous basalt, to pillowed basalt
carried by them away from the dyke will have the effect (rounded blobs of basalt that look like pillows) with
of accelerating the process of cooling of the dyke. Addi­ granitic veins between the pillows, to granite with less
tionally, if the dyke has intruded into sediments or meta­ than 50% pillow-like inclusions of basalt. In all the
morphic rocks that are wet, because of either ground or composite dykes, mixing of the two magmas appears to
pore water, the effect will be to lower the contact tempera­ have formed by a progression from simple curved con­
tures by amounts of up to 50°C and slightly shorten the tacts to intimate granite veining of the basalt, to ultimate
times of solidification and complete cooling. subdivision of basalt into smaller pillows and lenses.
Jaeger also reported a simplified expression for cool­ These and many small-scale features suggest that the
ing time, which encompasses all the relevant parameters:- basalt and granite were magmas when they initially made
τ = κt/a2 contact with each other in the composite dyke. Because
where the granite clearly remained mobile longer than the basalt,
τ is a Fourier number or dimensionless time; the latter must have cooled first, even though the basalt
κ is the thermal diffusivity (and accounts for thermal magma may have been intruded at a higher temperature
conductivity, magma density and the specific than the granite.
heat); Temperature and pressure in the country rock at the
t is the time in years; and time of dyke emplacement were difficult for Wiebe to es­
a is half the width of the dyke (D = 2a). timate, the chilled margins of most dykes suggesting low
Substituting in what Jaeger described as a reasonable temperatures and probably low pressures. The existence
value for κ (0.01 cm2se c -1), and rearranging the equation of the mineral hornblende in the basalt suggested load
to calculate cooling time, he found that pressures greater than 1 kilobar (kb) were likely. Wiebe
thus estimated the temperature of the country rock to have
been 200°C and the pressure 2kb. Initial temperatures of
Furthermore, he suggested that for τ<0.01 cooling is the basaltic and granitic magmas would have been 1100°C
only superficial; for τ = 0.1 it will have penetrated to the and 750°C respectively, such high temperatures in the
centre of the dyke; for τ = 1 there is substantial cooling at dykes thus ensuring that even if there was a variation of
the centre; and for τ >10 cooling is practically complete. as much as 200°C in the temperature of the country rock,
Consequently, if a value of τ for complete cooling is set at this would affect the cooling times of these magmas by
12, then a 10 metre wide dyke (a=5) would cool com­ less than 50%. Mineralogical observations suggested that
pletely in approximately 9.5 years, a similar estimate to at the time of injection the temperatures of the magmas
the extrapolation of Jaeger’s families of curves. were slightly below these liquidus temperatures for each
rock.
A FIELD EXAMPLE Wiebe then proceeded to use the standard cooling cal­
culations of Jaeger as already outlined above. In so doing,
The calculations and results reviewed here, derived as he adopted Jaeger’s values for the other properties of the
they are from theoretical considerations and first prin­ magma and country rock. Other parameters used were an
ciples, have successfully been used in practical situations average dyke thickness of 50cm (some portions of these
where they correlate well with other field observations. dykes were over 1 metre wide), an average dyke compo­
Perhaps one of the most unusual field examples of sition of about 70% granite and 30% basalt, an initial
dykes are those studied by Wiebe4 along the northern average temperature of the dyke segments of approxi­
coast of Cape Breton Island, Nova Scotia, Canada. At this mately 850°C (even though the initial temperature of the
locality are found a number of dykes, some consisting of basalt and granite magmas would have been 1100°C and
homogeneous basalt, and some of homogeneous granite, 750°C respectively), solidification temperatures of the
89
Formation and Cooling of Dykes

basalt and granite of 900°C and 700°C respectively, a so­ completed in about 9 years. It is also good to see that these
lidification temperature therefore of the dyke as a whole estimates are confirmed by field studies such as the
of 700°C, an overall density of 2.6g cm-3, a thermal con­ example cited. Since the intrusion of dykes into other
ductivity of 0.005 cal cm-1 se c -1d e g -1, a specific heat of rocks is an event which can be much later than the
0.28, and a latent heat of crystallization of 80 cal g -1. formation of the rocks into which the dyke intrudes, dyke
Based on the 70:30 ratio of granite and basalt, and the formation in many instances could still have occurred
average initial temperature estimated at 850°C, solidifica­ after Noah’s Flood, so these timeframes greater than the
tion of the composite (or average) dyke is thus assumed to duration of the Flood year do not generate an additional
have been complete when the 700°C isotherm reached the problem. We have also seen that even the time taken for
centre of the dyke, that is, when all the granitic component a magma to rise from depths as much as 30 kilometres
was solid. Using the methods of Jaeger and the conditions down in the earth’s crust and crystallize is also only a
already specified, Wiebe thus calculated that the compos­ matter of a few years.
ite dyke would have solidified completely in 27.8 hours, While such thin bodies of igneous rock would appear
although he cautioned that this time must be considered to be a simple solvable challenge for creationists, they
only as an estimate value for a dyke of homogeneous provide the easiest place to start in this whole investiga­
average composition and width, since it cannot be applied tion of the cooling of igneous rocks. Further studies will
specifically to any point in the real dyke. Portions of need to delve into the cooling times for larger bodies of
basaltic composition would have solidified more rapidly, igneous rock, such as 1 kilometre thick sills and granite
and granite very near to the basalt would have solidified batholiths up to tens of kilometres wide.
more slowly than this average value. Thus the calculated
time of solidification based on the average initial tempera­ REFERENCES
ture is probably a moderate overestimate of the actual
time. 1. Szekely, J., and Reitan, P. H., 1971. Dyke filling by magma intrusion
The basalt pillows particularly would have become and by explosive entrainment of fragments. Journal of Geophysical
Research, 76(11):2602–2609.
solid much more rapidly than the granite host, and because 2. Jaeger, J. C., 1957. The temperature in the neighborhood of a cooling
of their discrete sizes a cooling history timeframe can be intrusive sheet. American Journal of Science, 255:306–318.
calculated. A basalt pillow would have solidified when 3. Jaeger, J. C., 1968. Cooling and solidification of igneous rocks. In:
the 900°C isotherm reached its centre, so in contact with H.H. Hess and A. Poldevaart (Eds), Basalts, vol. 2, Interscience Pub­
a granite magma at 750°C the approximate time necessary lishers, New York, pp. 503–536.
4. Wiebe, R. A., 1973. Relations between coexisting basaltic and granitic
for solidification of a pillow with a 5cm radius would have magmas in a composite dike. American Journal of Science,
only been 35 minutes. Similarly, a pillow with a 10cm 273:130–151.
radius would have cooled in about 140 minutes. Thus
Wiebe concluded that the 900°C isotherm would have
reached the centre of a typical pillow within less than 1
hour. He also found that these short timeframes for Dr Andrew Snelling is a geologist with a B.Sc.(Hons)
cooling of both the basalt and granite magmas were from The University of New South Wales and a Ph.D.
supported by the geochemical data on the transfer during from The University of Sydney. He has worked in the
magma cooling of various elements between the basalt mining industry and is still a consultant geologist, but now
and granite components of these composite dykes. also works full-time with the Creation Science Founda­
tion where he contributes to Creation Ex Nihilo maga­
CONCLUSIONS zine and edits the Creation Ex Nihilo Technical Jour­
nal. He resides in Brisbane, Australia.
For creationists the cooling times for fine-grained
igneous rocks such as basalt and dolerite (diabase) have
never been considered to be a real problem, although the
timeframes involved have never been really documented
in the relevant creationist literature. This review therefore
has endeavoured to supply this documentation of the
cooling times involved in such thin igneous bodies as
dykes.
We have seen that by the application of standard
cooling theory for solid bodies, the timeframes involved
are indeed relatively short, and that even for a granitic
dyke 10 metres wide, which is fairly rare, the cooling time
for crystallization (solidification) to occur is only of the
order of 1.6 years, and cooling to ambient temperatures is
90