ORIGINAL PAPER

Ti diffusion in quartz inclusions: implications for metamorphictime scales

Frank S. Spear • Kyle T. Ashley • Laura E. Webb •

Jay B. Thomas

Received: 3 April 2012 / Accepted: 28 July 2012

� Springer-Verlag 2012

Abstract Quartz inclusions in garnet from samples col-

lected from the staurolite zone in central New England are

zoned in cathodoluminescence (CL). The CL intensity is

interpreted to be a proxy for Ti concentration and the

zoning attributed to Ti diffusion into the quartz grains

driven by Ti exchange between quartz and enclosing garnet

as a function of changing temperature. The CL zoning has

been interpreted using a numerical diffusion model to

constrain the time scales over which the diffusion has

occurred. Temperature–time histories are sensitive to the

presumed peak temperature but not to other model

parameters. The total time of the metamorphic heating and

cooling cycle from around 450 �C to the peak temperature

(550–600 �C) back to 450 �C is surprisingly short and

encompasses only 0.2–2 million years for peak tempera-

tures of 600–550 �C. The metamorphism was accompanied

by large-scale nappe and dome formation, and it is sug-

gested that this occurred as a consequence of in-sequence

thrusting resulting in a mid-crustal ductile duplex structure.

Keywords Cathodoluminescence � Quartz � Ti diffusion �Vermont metamorphism � TitaniQ

Introduction

Spear and Wark (2009) described the appearance of quartz

grains in cathodoluminescence (CL) images in rocks

crystallized at metamorphic grades ranging from the biotite

to the migmatite zone. They observed that in rocks of the

staurolite and staurolite–kyanite zones, quartz inclusions

within garnet (and staurolite) were typically zoned with

increasing CL intensity from the quartz core to the quartz–

garnet interface. They attributed this zoning to have been

caused by diffusion of Ti in quartz, driven by Ti–Si

exchange between quartz and enclosing garnet. Quartz

inclusions from all 15 of the staurolite-zone samples

examined revealed this type of zoning. In contrast, quartz

inclusions from lower-grade samples were unzoned, and

quartz inclusions from higher-grade samples displayed

either complex internal CL patterns, or were zoned with

decreasing CL intensity (and Ti contents) toward the

quartz–garnet interface.

The purpose of this paper is to present observations on

additional staurolite-zone samples and describe the results of

diffusion modeling of the zoning profiles. The objective of

this study is to constrain the time scales of the metamorphism

of these samples and to evaluate possible tectonic processes

that may have been responsible for these time scales.

Analytical methods

CL imaging and Ti analyses were performed on the

Cameca SX-100 electron microprobe at Rensselaer

Communicated by T. L. Grove.

F. S. Spear (&) � J. B. Thomas

Department of Earth and Environmental Sciences,

Rensselaer Polytechnic Institute, 110 8th Street,

Troy, NY 12180, USA

e-mail: spearf@rpi.edu

K. T. Ashley

Department of Geosciences, 4044 Derring Hall, Virginia Tech,

Blacksburg, VA 24061, USA

L. E. Webb

Department of Geology, University of Vermont,

180 Colchester Avenue, Burlington, VT 05405, USA

123

Contrib Mineral Petrol

DOI 10.1007/s00410-012-0783-z

Polytechnic Institute following the methods described in

Spear and Wark (2009). CL images were collected with a

Gatan Mono-CL detector equipped with red, green, and

blue filters. The blue filter was used exclusively for the

images used in this study because of the linear relationship

between Ti concentration and CL emission at blue wave-

lengths (e.g. Wark and Spear 2005; Spear and Wark 2009;

Leeman et al. 2012). Ti analyses were collected on four

PET crystals simultaneously, yielding a precision of around

6–7 ppm for each spot analysis (Spear and Wark 2009).

Sample selection and modeling approach

Samples were selected for analysis from the authors’ col-

lections from east-central Vermont and west-central New

Hampshire after routine CL imaging of dozens of samples

from different metamorphic grades (e.g. Spear and Wark

2009). With the exception of sample TM-549, all are from

the staurolite or staurolite–kyanite zones (Table 1) and all

samples contain garnet with abundant quartz inclusions.

The approach to be followed in this paper is to present in

detail the zoning character and diffusion modeling results

from a single sample from eastern Vermont (79-149D) and

then to apply this method to diffusion profiles from the

entire suite of samples. Table 1 contains the complete list

of samples to be discussed and their locations.

Sample 79-149D is a metapelite from the Orfordville

belt of eastern Vermont and western New Hampshire

(Fig. 1. Also see Fig. 3 of Spear and Rumble 1986, for a

map showing the location of this sample). The assemblage

is garnet ? biotite ? staurolite ? kyanite ? muscovite ?

quartz ? plagioclase ? ilmenite ? monazite ? zircon.

Garnet preserves growth zoning of, from core to rim,

decreasing Mn and Ca and increasing Fe/Mg. The P–T path

calculated for this sample from garnet zoning shows an

episode of nearly isothermal loading followed by heating

with decompression (Spear and Rumble 1986, Fig. 7) with

a peak P–T conditions of 580 �C, 5.1 kbar.

Figure 2 shows a CL image of the garnet from this

sample at different scales (see also Spear and Wark 2009,

Figs. 1c, d, e; 6a, b, c). Quartz in the matrix is relatively

unzoned with Ti concentrations of 5–14 ppm. In contrast,

all of the quartz inclusions in the garnet are zoned with

increasing Ti content from core to rim. In addition, quartz

inclusions inside of staurolite are zoned similarly to those

in garnet, although not as cleanly and quartz inclusions in

kyanite from this sample are unzoned.

The Ti contents in one inclusion measured on the

electron probe range from 3 ppm to 14 ppm (Figs. 3, 6c of

Spear and Wark 2009) and correlate linearly with CL

intensity in the blue region, although the uncertainty in

these measurements is on the order of ±7 ppm (1 sigma).

Additional data supporting the correlation of CL intensity

in quartz in the blue region (ca. 415 nm wavelength) with

Ti concentration have been presented by Spear and Wark

(2009), Wark and Spear (2005), and Leeman et al. (2012).

Based on these studies, Ti will be assumed to be a linear

function of CL intensity in the blue region, thus enabling

rapid semi-quantitative assessment of the Ti zoning profiles

in quartz.

In contrast, CL images of quartz inclusions in garnet in

samples from other metamorphic grades display distinctly

different zoning. In garnet zone samples, quartz inclusions

are relatively unzoned (e.g. Fig. 2 of Spear and Wark

2009), and in sillimanite-zone and higher-grade samples,

quartz inclusions display complex internal textures and are

typically zoned with decreasing Ti contents toward the

garnet interface (e.g. Fig 8. of Spear and Wark 2009).

Two causes of the observed CL (i.e. Ti) zoning in quartz

are considered. In the first, prior to entrapment by garnet,

the quartz in the matrix is assumed to be zoned with lower

Ti cores and higher Ti rims. In the extreme, the rims might

have been step functions, although no such zoned quartz

Table 1 Sample locations and metamorphic grade

Sample Latitude Longitude Grade Peak T (C)a Peak P(Kb)a Assemblage

79-149 43.77335 -72.23648 St–Ky 580 ± 25 5–6 Qtz ? Pl ? Grt ? Bt ? Ms ? St ? Ky ? Ilm ? Mnz

TM-549 43.79511 -72.28250 Grt 480 ± 25 4–5 Qtz ? Grt ? Bt ? Ms ? Chl

TM-543 43.78847 -72.35205 St 555 ± 25 8–10 Qtz ? Pl ? Grt ? Bt ? Chl(r) ? Rt

TM-828 43.84327 -72.34288 St 575 ± 25 8–10 Qtz ? Pl ? Grt ? Bt ? Ms ? Rut ? Gra ? Tur

TM-732 43.79472 -72.39863 St 595 ± 25 7–9 Qtz ? Pl ? Grt ? Bt ? Ms ? Rut ? Ap ? Tur

TM-763 43.79333 -72.44711 St 550 ± 25 7–9 Qtz ? Pl ? Grt ? Bt ? Ms ? Chl ? Rt ? Ap

09SD08A 43.82659 -72.44454 St 550 ± 25 7–9 Qtz ? Pl ? Grt ? Bt ? Ms ? Rt

TM-675 43.80938 -72.50372 St 550 ± 25 7–9 Qtz ? Pl ? Grt ? Bt ? Ms ? Chl(r) ? Rt ? Gra ? Tur

BF-38B 43.18159 -72.35856 St 550 ± 16 3–5 Qtz ? Pl ? Grt ? Bt ? Ms ? St ? Ilm

a Peak P and T sources: Samples TM-xxx are from Menard and Spear (1994); sample 79-149D from Spear and Rumble (1986); sample BF-38B

from Spear and Wark (2009). Mineral abbreviations after Kretz (1983)

Contrib Mineral Petrol

123

grains are now observed in the matrix. Mechanisms for

producing quartz zoned in this fashion were discussed by

Spear and Wark (2009) and include (a) prograde reactions

and (b) recrystallization due to fabric reorientation. It is not

expected that pre-garnet prograde reactions could have

produced quartz zoned with increasing Ti concentrations

toward the rim because matrix quartz in garnet- and sub-

garnet-grade samples shows no such zoning and the first

major quartz-producing reaction experienced by these

rocks is the staurolite-in reaction (garnet ? chlorite ?

muscovite = staurolite ? biotite ? quartz ? H2O). It is

still possible that quartz recrystallization as a result of

strain could produce pre-garnet rims of higher Ti concen-

trations, but again evidence for this in lower-grade samples

is lacking.

The second cause considered is that the pre-garnet

quartz grains were relatively homogeneous with low Ti

concentrations on the order of 3–5 ppm, similar to the

Fig. 1 Geologic sketch map of

a part of eastern Vermont and

western New Hampshire

showing location of samples

studied with the exception of

BF-38B, which is described by

Spear and Wark (2009) and

located in Spear et al. (1995).

Diamonds with numbers are

locations of samples from Wing

et al. (2003) discussed in the

text with monazite ages shown.

Inset shows general location of

study area

Fig. 2 CL images of sample

79-149D. a Low-magnification

image showing garnet (black)

and distribution of quartz

inclusions. Box shows location

of (b). b CL image showing

zoning in quartz inclusions.

Box shows location of Fig. 3a

Contrib Mineral Petrol

123

cores of the quartz inclusions in garnet, with the Ti con-

centration governed by the local TiO2 activity. Sample

79-149D contains ilmenite in the matrix and as inclusions

in garnet, so it is expected that the activity of TiO2 was

close to 1.0 (Ghent and Stout, 1984). The interior of the

quartz at the time of garnet overgrowth was not necessarily

in equilibrium with the rim as garnet nucleation is believed

to have occurred at around 500 �C (Spear and Rumble

1986), whereas the Ti concentration of the quartz interiors,

which ranges from 3 to 5 ppm in both matrix and inclusion

grains, reflects a temperature of approximately 350–400 �C

at around 5 kbar (calibration of Thomas et al. 2010). Once

garnet completely overgrew a quartz grain, equilibrium

with the matrix TiO2 activity is no longer relevant and is

replaced by exchange equilibria between garnet and quartz.

The substitution mechanism for the incorporation of Ti into

garnet is not known, so it is not possible to write a specific

exchange reaction and associated equilibrium constant.

However, garnet contains more Ti than coexisting quartz.

Analyses for Ti contents in garnet from sample 79-149D

using LA-ICPMS reveal concentrations ranging from ca.

60 ppm near the rim to 300 ppm near the core. Microprobe

analyses of garnet crystals examined in this study range up

to a thousand ppm Ti. Regardless of the nature of the

exchange reaction between garnet and quartz, the equilib-

rium constant for the reaction should approach 1.0 (equal

partitioning) with increasing temperature with the result

that the quartz rim should become more Ti-rich and the

adjacent garnet more Ti-poor with increasing metamorphic

grade. This exchange would set up a gradient in the quartz,

which would drive diffusion. The quartz rim would con-

tinue to become increasingly Ti-rich up to the metamorphic

peak, and the exchange would reverse on cooling. The final

zoning profile in the quartz would thus be a function of the

time-dependent composition of the quartz rim, the diffu-

sivity of Ti in quartz, and the temperature–time history.

Of these two mechanisms, the second is preferred in part

because of the reasons given above that argue for homo-

geneous quartz at the time of garnet growth, and in part

because of the symmetry of the Ti zoning in the inclusions.

In any case, the results of this study are not strongly

dependent on whether the pre-garnet quartz grains had step

functions in Ti concentration or were modified by the

exchange of Ti with the host garnet. In either case, the

present-day zoning profile is modified by diffusion over

length scales of less than 10 lm, which, as will be dis-

cussed below, requires quite rapid heating and cooling.

Ideally, the zoning profiles in quartz should be modeled

as Ti concentration versus distance. However, the linear

relationship between CL intensity in the blue spectrum and

Ti concentration provides a more sensitive, both spatially

and compositionally, approach. As can be seen in Fig. 3,

the linear correlation between CL intensity and Ti con-

centration permits the use of CL intensity as a proxy for Ti

concentration and permits the modeling of CL intensity

profiles directly. Although the scatter in Fig. 3b is

12,1011

97 5

3 4 4

7

11

25 µm

y = 0.1232x + 2.0649R2 = 0.80

0

2

4

6

8

10

12

14

0 10 20 30 40 50 60 70 80 90

0

2

4

6

8

10

12

14

0 10 20 30 40 50

CL Intensity (arbitrary scale)

Ti (

ppm

)T

i (pp

m)

Distance (µm)

(a)

(b)

(c)

Fig. 3 a CL image of quartz inclusion in garnet (sample 79-149D).

Bright spots are electron microprobe analytical spots with Ti (ppm)

indicated for each. Garnet host is anomalously bright (it should be

dark) because of secondary luminescence of oil on the CL mirror.

Rectangle shows location of zoning traverse in (c). b Plot showing CL

intensity (gray scale value) against measured Ti concentration. c Plot

of Ti zoning in the quartz inclusion in a based on calibration shown in

(b)

Contrib Mineral Petrol

123

relatively large, more detailed justification for this corre-

lation has been given by Wark and Spear (2005), Spear and

Wark (2009), and Leeman et al. (2012). The observed CL

intensity profile in many quartz inclusions is somewhat

asymmetric (Fig. 3c), as is the measured Ti profile, for

unknown reasons. Both sides of the profile are systematic

with respect to the garnet–quartz interface, and both sides

show an increase from core to rim to a maximum value,

then a decrease right at the rim. It should be noted in

Fig. 3a that garnet is the brightest phase in the image, even

though the luminescence is almost nil. This has occurred

because, at the time this CL image was collected, a thin

film of oil had contaminated the parabolic light focusing

mirror on the CL light finger. This oil luminesces in the

blue region and garnet is bright in this image because

backscattered electrons are activating the oil causing it to

luminesce (ilmenite luminesces even brighter in similar

images). This artifact was corrected by cleaning the mirror

but has the fortuitous advantage of showing exactly the

position of the garnet–quartz interface. It can be clearly

seen in Fig. 3a that a dark border, indicative of a decrease

in Ti concentration, mantles the edge of the quartz at the

garnet interface. This decrease is important in the modeling

as it is the change in concentration due to diffusion during

cooling.

Results

Figure 4 shows an example of the results of a diffusion

model of a CL zoning profile from a quartz grain from

sample 79-149D based on the assumptions inherent in the

second mechanism discussed above (initial conditions of

homogenous quartz of low Ti concentration included in

garnet). CL intensities were calculated as average pixel

intensities using ImageJ (Rasband 1997–2011) software

and a traverse width of typically 10 pixels. The diffusivities

measured by Cherniak et al. (2007) were incorporated into

a finite difference code using a linear geometry, fixed

boundary position, and boundary composition that changed

with temperature. The initial condition was that of an un-

zoned quartz grain with the composition of the core of the

grain. The CL intensity at the quartz–garnet interface was

modeled using an expression to mimic a partitioning

expression, namely:

Log(CL intensity) ¼ aþ b=T ð1Þ

The parameters a and b were fit by a two-point fit using

the core CL intensity with an assumed temperature of

quartz core formation (typically 450 �C) and the estimated

peak metamorphic temperature (Table 1) and an assumed

CL intensity at the quartz–garnet interface at the peak

temperature. This ‘‘peak CL intensity’’ is not known

because of diffusional modification on cooling, but was

adjusted as a model fit parameter. For example, for sample

79-149D with an assumed peak temperature of 600 �C

(Fig. 4a), the initial and ‘‘peak’’ CL intensities were taken

to be 28 CL units, 450 �C and 250 CL units, 600 �C,

respectively. In practice, it was discovered that the fit of the

model profile to the observed profile was quite sensitive to

the choice of the ‘‘peak’’ CL intensity, but the overall T–t

history was not. On the other hand, the overall T–t history

was found to be quite sensitive to the choice of the peak

metamorphic temperature. The fit of CL intensity with

temperature (Eq. 1) was adjusted depending on the

assumed peak metamorphic temperature. For example,

for the model in which the assumed peak temperature was

575 �C (Fig. 4b), the ‘‘peak CL intensity’’ was taken to be

250 CL units, 575 �C.

The three diffusion models in which the peak meta-

morphic temperature was assumed to be 600, 575, and

550 �C display equally good fits (Fig. 4). The only major

difference is the times scales for the metamorphic episode

(heating and cooling) of approximately 0.25 m.y., 0.5 m.y.,

and 1.0 m.y., respectively, for the three peak temperatures.

That is, the time scale approximately halves for every

25 �C increases in peak metamorphic temperature. Spear

and Rumble (1986) report the peak metamorphic temper-

ature for the southwest Orfordville quadrangle (NH and

VT) where sample 79-149D was collected as 580 ± 25 �C,

5.1 kbar, so the metamorphic time scale consistent with this

peak condition is on the order of 0.5 m.y.

Considerations regarding the diffusion model

Sources of possible error in the diffusion modeling are not

simple to evaluate. It is not believed that the diffusivities

measured by Cherniak et al. (2007) are significantly in

error unless a different mechanism operates in the natural

samples compared with the experimental study. The

Cherniak et al. (2007) experiments were conducted dry,

and it is possible that diffusion under hydrous conditions

such as those found in an evolving schist might be faster.

This possibility is difficult to evaluate, but if diffusivities

were faster it would only serve to shorten the time scales of

metamorphism. It is expected that Ti diffusion in garnet is

slower than that in quartz, especially if a coupled substi-

tution in garnet is required. If Ti diffusion in garnet is

infinitely slow, this raises the question whether equilibrium

between garnet and quartz inclusion could be maintained.

Clearly, it is based on the observed pervasiveness of zoned

quartz inclusions. What is required is that the flux out of

garnet equals the flux into quartz. Inasmuch as garnet may

contain several hundred ppm Ti, compared with 5–20 in

quartz, the flux balance would require only a very small

shell of garnet to exchange with the quartz, so it is not

Contrib Mineral Petrol

123

likely that diffusion in garnet has limited the diffusion

observed in quartz. Even if it had to some extent, the dif-

fusion models are not dependent on a priori assumptions

about the partitioning of Ti between quartz and garnet, but

rather on the geometric shape of the zoning profile in

quartz (see ‘‘peak’’ lines in Fig. 4). The major determining

factor in generating a good fit to the diffusion profile is the

penetration distance of the diffusion profile (e.g. around

5–6 lm in Fig. 4), which, coupled with the peak meta-

morphic temperature, determines the time scale for profile

development.

The alternative mechanism for generating the zoning

profiles (mechanism 1 discussed above) has as the initial

condition a step function in the quartz inclusion. The initial

400

450

500

550

600

T C

Time (m.y.)

Peak T = 550 C

400

450

500

550

600

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

T C

Time (m.y.)

Peak T = 575 C

0.00

50.00

100.00

150.00

200.00

250.00

Radius microns

CL

inte

nsity

79-149d

Peak T = 550 C

0.00

50.00

100.00

150.00

200.00

250.00

Radius microns

CL

inte

nsity

79-149d

Peak T = 575 C

400

450

500

550

600

T C

Time (m.y.)

Peak T = 600 C

0 0.2 0.4 0.6 0.8 1 1.2 1.40.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0

0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0

0 0.05 0.1 0.15 0.2 0.25 0.30.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0

Radius microns

CL

inte

nsity

79-149dPeak T = 600 C

0.00

50.00

100.00

150.00

200.00

250.00(a)

(c)

(e)

(b)

(d)

(f)

Fig. 4 Results of diffusion models for Ti diffusion in quartz from

sample 79-149D assuming peak temperatures of 600 �C (a, b),

575 �C (c, d) and 550 �C (e, f). (a, c, e) Plots of CL intensity versus

distance. Dashed line shows assumed profile at the peak temperature.

Solid line shows the final calculated profile. b, d, f Temperature–time

plots for each of the diffusion models. Note the scale difference along

the time axis

Contrib Mineral Petrol

123

step would need to be symmetrical around a quartz grain

(not a likely possibility), and the length of the step would

need to have been on the order of the midpoint of the

diffusion profile (e.g. around 3–5 lm). Although diffusion

models involving this geometry were not attempted, the

time scale for the development of the final profile would

have to be shorter than with the preferred model because

the length scale for diffusion is only around half of that

with the preferred model.

It should also be noted that the linear temperature–time

histories used in the model experiments are unrealistic and

considerations of heat flow would suggest that any realistic

T–t history should be curved concave downward somewhat

like a parabola. Models with this type of T–t history were

not attempted, but it is clear that such models would result

in even shorter time scales for the same peak temperature

because the time the rock would spend near the peak

temperature would be larger than with the linear models.

Thus, the results presented here utilize the more conser-

vative temperature–time paths.

Other samples

CL zoning profiles from quartz inclusions in garnet from

samples collected along a traverse across the Strafford

Dome in eastern Vermont (see Fig. 1 for locations) and

from a single staurolite-grade sample from New Hampshire

are shown in Fig. 5, and temperature–time plots from the

diffusion modeling are shown in Fig. 6. With the exception

of the single garnet zone sample (TM-549), which shows

little or no zoning, all samples have similar length scales

for diffusion of around 5–10 lm. The temperature–time

plots indicate metamorphic time scales of 0.1–2 m.y., and

the major reason for the difference in time scales is the

differences in peak metamorphic temperatures, which

range from around 550–600 �C (Table 1). The single

sample from New Hampshire (BF-38B) has a similar time

scale of 1.5 m.y. If the results of the diffusion modeling are

taken at face value, then the time scales for metamorphism

across this segment of Vermont must have varied by an

order of magnitude for all diffusion profiles to have similar

length scales. The shortest duration, as well as the highest

peak metamorphic temperature, is recorded by the sample

from the core of the Strafford Dome (TM-732), and it is

possible that this sample did, in fact, experience a shorter

metamorphic heating–cooling cycle by virtue of its struc-

tural setting. For example, it may have experienced rapid

burial and subsequent exhumation during dome formation,

whereas samples on the flanks of the dome were heated and

cooled more slowly as limited by heat conduction. Alter-

natively, it is certainly possible that the estimates of

the metamorphic temperature, which were made using

garnet–biotite thermometry and are only accurate to ca.

±25–30 �C, actually are much more similar than reported.

Further work refining the peak temperatures should help

resolve this issue.

Discussion

Metamorphic heating and cooling cycles of 2 m.y or less

are rapid and somewhat unanticipated for this Barrovian

terrane where the models of England and Thompson (1984)

for continental collision would suggest a time scale of tens

of millions of years. However, other recent studies have

reached similar conclusions based on independent meth-

ods. Dachs and Proyer (2002) concluded from a study of

diffusional relaxation of Mn zoning profiles in garnet that

the total time elapsed following formation of the garnet

overgrowth (below the metamorphic peak of 570 �C) to

around 450 �C was 1 m.y. or less, based on their finite

difference diffusion modeling. They conclude that this

required very rapid exhumation rates of 4.6–7.4 cm/year,

but it also required rapid heating to the metamorphic peak

as well. A similar rapid heating–cooling cycle was reported

by Ague and Baxter (2007) for samples from the classic

Barrovian terrane in the Dalradian of Scotland based on

diffusion modeling of Sr zoning in apatite, and Viete et al.

(2011) argue for short (a few million years or less) thermal

pulses to explain Mn zoning profiles in garnet. Ague and

Baxter (2007) attribute the short thermal pulse to advective

heat transfer by magmas and associated fluid flow, and

Viete et al. (2011) also suggest that shear zones may play

a critical role in providing significant advective heat

transport.

Geochronologic studies have also suggested rapid time

scales for metamorphism. Exhumation at a rate of several

cm/year is required by data reported by Harrison et al.

(1997) and Catlos et al. (2001) for the Main Central Thrust

of the Central Himalaya, and similar rapid rates are

required for exhumation of eclogites in the Alps (Rubatto

and Hermann 2001; Smye et al. 2011) and the Papua New

Guinea eclogites (Baldwin et al. 2004).

Although the time scales inferred from our diffusion

studies are similar to those obtained for Barrovian meta-

morphism in the Dalradian (e.g. Ague and Baxter 2007;

Viete et al. 2011), thermal pulses due to magmas or fluids

only are not likely in Vermont. No evidence of such plu-

tons is evident in the immediate geology of the study area,

and those that do exist along strike are 10–30 m.y. older

than the presumed time of metamorphism based on mon-

azite ages (e.g. Wing et al. 2003). Furthermore, the P–T

paths of the Vermont samples display an episode of nearly

isothermal loading (e.g. Spear and Rumble 1986; Menard

and Spear 1994), which requires tectonic thickening for a

significant part of the P–T history. This thickening must

Contrib Mineral Petrol

123

have been rapid and involved placement of hot rocks above

cooler rocks to cause the rapid heating. Exhumation and

cooling must have followed immediately and may have

been caused by the rocks under study having been placed

onto cooler rocks. A possible kinematic scenario is a series

of in-sequence ductile thrust faults or shear zones in the

general form of a mid-crustal duplex structure.

Significantly, geochronologic studies in Vermont do not

provide any inklings about the rates of tectonic processes

because the duration of the heating and cooling revealed by

the diffusion study is shorter than the typical uncertainty

associated with age determinations. For example, monazite

ages from this area of Vermont from SIMS analyses have

been presented by Wing et al. (2003) and are shown in

Fig. 1. Well-crystallized monazites are found only above

the staurolite isograd and are interpreted to have formed

from the breakdown of allanite. Ages of three samples from

the staurolite–kyanite zone are 350–359, 335–383, and

347–367 Ma (Fig. 1). In contrast, the ages reported for

monazite from a chlorite zone sample are 309–338 Ma and

that monazite was poorly crystallized. The ages of the

staurolite–kyanite zone samples are all within error, and

even the best of the samples, which gave an average age of

352.9 ± 8.9 (2 sigma), has an error that is outside of the

entire metamorphic history recorded by the diffusion

modeling in this study. The younger ages recorded by the

chlorite zone sample are difficult to interpret without

additional study, but possibly indicate that a distinct

metamorphic heating–cooling cycle has affected these

rocks and may suggest that a significant post-metamorphic

fault separates them from the higher-grade rocks. In

another study, Janots et al. (2009) report ages of allanite

and monazite from the central Alps and interpret the age

difference of around 13 m.y. (31.5 and 18.5 Ma, respec-

tively) in terms of the duration of the metamorphic event.

However, their results are also consistent with a T–t history

that involves a short thermal spike at around 19 Ma

(to produce the monazite) imposed on a background

0

20

40

60

80

100

120

140

79-149D

0

20

40

60

80

100

120

140

160

180

200

0 5 10 15 20 25 0 5 10 15 20 25

TM-549

0

20

40

60

80

100

120

140

160

180

TM-543

0

50

100

150

200

250

0 5 10 15 20 25

0 5 10 15 20 25

0 5 10 15 20 25 0 5 10 15 20 25

TM-828A

0

50

100

150

200

250

0 5 10 15 20 25

TM-732

0

50

100

150

200

250

TM-763

0

50

100

150

200

250

0 5 10 15 20 25

TM-675

0

50

100

150

200

250

300

0 5 10 15 20 25

09SD08A

0

50

100

150

200

250

BF-38B

CL

Inte

nsity

(ar

bitr

ary

scal

e)

Distance (µm) Distance (µm) Distance (µm)

Fig. 5 Plots of CL intensity versus distance for the samples studied.

Horizontal scales are equal to facilitate comparison, whereas vertical(intensity) scales are not comparable. Note the similarities in length

scales for the CL zoning in all samples from the staurolite zone (e.g.

around 5 lm)

Contrib Mineral Petrol

123

metamorphic grade in the allanite stability field. The

implication is that, whereas accessory mineral geochro-

nology can serve to place the metamorphic event in an

absolute time scale and may also serve to differentiate

distinct metamorphic heating/cooling events, it cannot be

expected to resolve the relationships between metamor-

phism and tectonics at the same resolution possible from

diffusion studies. Obviously, there is considerable room for

substantial additional work using this potentially powerful

new method to estimate T–t histories in metamorphic

rocks.

Acknowledgments The authors wish to thank helpful discussions

with Daniele Cherniak, Bruce Watson, Michael Ackerson, Ben Hal-

lett, Kenny Horkley, and the able assistance of Dan Ruscitto with the

microprobe and LA-ICPMS. Thoughtful reviews by J. M. Ferry and J.

J. Ague are also gratefully acknowledged. This work was funded in

part by grants from the National Science Foundation (EAR-0948530

to Spear and Thomas and EAR-0948529 to Webb) and the Edward P.

Hamilton Distinguished Professor Chair at Rensselaer Polytechnic

Institute.

References

Ague JJ, Baxter EF (2007) Brief thermal pulses during mountain

building recorded by Sr diffusion in apatite and multicomponent

diffusion in garnet. Earth Planet Sci Lett 261:500–516

Baldwin SL, Monteleone BD, Webb LE, Fitzgerald PG, Grove M,

Hill EJ (2004) Pliocene eclogite exhumation at plate tectonic

rates in eastern Papua New Guinea. Nature 431:263–267

Catlos EJ, Harrison TM, Kohn MJ, Grove M, Ryerson FJ, Manning

CE, Upreti BN (2001) Geochronologic and thermobarometric

constraints on the evolution of the Main Central Thrust, central

Nepal Himalaya. J Geophys Res 106(B8):177–204

Cherniak DJ, Watson EB, Wark DA (2007) Ti diffusion in quartz.

Chem Geol 236:65–74

Dachs E, Proyer A (2002) Constraints on the duration of high-

pressure metamorphism in the Tauern Window from diffusion

modelling of discontinuous growth zones in eclogite garnet.

J Metamorp Geol 20:769–780

England PC, Thompson AB (1984) Pressure—temperature—time

paths of regional metamorphism, part I: heat transfer during the

evolution of regions of thickened continental crust. J Petrol

25:894–928

Ghent ED, Stout MZ (1984) TiO2 activity in metamorphosed pelitic

and basic rocks: principles and applications to metamorphism in

southeastern Canadian Cordillera. Contrib Mineral Petrol

86:248–255

Harrison TM, Ryerson FJ, Le Fort P, Yin A, Lovera OM, Catlos EJ

(1997) A late Miocene-Pliocene origin for the central Himalayan

inverted metamorphism. Earth Planet Sci Lett 146:1–7

Janots E, Engi M, Rubatto D, Berger A, Gregory C, Rahn M (2009)

Metamorphic rates in collisional orogeny from in situ allanite

and monazite dating. Geology 37:11–14

Kretz R (1983) Symbols for rock-forming minerals. Am Miner

68:277–279

Leeman WP, MacRae CM, Wilson NC, Torpy A, Lee C-TA, Student

JJ, Thomas JB, Vicenzi EP (2012) Quantitative application of

0

20

40

60

80

100

120

140

160

180

200

220

240

Time (m.y.)

Time (m.y.)

Distance (µ m)

Time (m.y.)

CL

Inte

nsity

(ar

bitr

ary

scal

e) TM-543

0 2 4 6 8 10 12 14

0 2 4 6 8 10 12 14

0

40

80

120

160

200

240

280

320

360

400

Distance (µ m)

CL

Inte

nsity

(ar

bitr

ary

scal

e) TM-828A

0.0 2.0 4.0 6.0 8.0 10.050.0

90.0

130.0

170.0

210.0

250.0

290.0

330.0

Distance (µ m)

CL

Inte

nsity

(ar

bitr

ary

scal

e)

T (

C)

T (

C)

T (

C)

T (

C)

T (

C)

T (

C)

TM-732

350

400

450

500

550

600

650

350

400

450

500

550

600

650

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

350

400

450

500

550

600

650

0 2 4 6 8 10 12 14 16 18 20 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20 2 4 6 8 10 12 140

40

80

120

160

200

240

280

320

360

400

Distance (µ m)

CL

Inte

nsity

(ar

bitr

ary

scal

e) 09SD08A

350

400

450

500

550

600

650

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 340

20

40

60

80

100

120

140

160

180

200

220

240

Distance (µ m)

CL

Inte

nsity

(ar

bitr

ary

scal

e) TM-763

350

400

450

500

550

600

650

0 0.5 1 1.5 2 2.5 3 3.5 4

0

40

80

120

160

200

240

280

320

360

400

Distance (µ m)

CL

Inte

nsity

(ar

bitr

ary

scal

e) BF-38B2

350

400

450

500

550

600

650

Time (m.y.)

Time (m.y.)

Time (m.y.)

Fig. 6 Plots showing results of diffusion modeling of CL zoning for all samples from this study. a Observed and modeled CL versus distance

plots. Horizontal scales are equivalent. b Temperature–time plots. All results have been scaled the same to facilitate comparison

Contrib Mineral Petrol

123

cathodoluminescence (CL) to natural quartz with application to

geothermobarometry. Microsc Microanal (in press)

Menard T, Spear FS (1994) Metamorphic P–T paths from calcic

pelitic schists from the Strafford Dome, Vermont. J Metamorp

Geol 12:811–826

Rasband W (1997–2011) ImageJ. National Institutes of Health,

Bethesda, Maryland

Rubatto D, Hermann J (2001) Exhumation as fast as subduction?

Geology 29:3–6

Smye AJ, Bickle MJ, Holland TJB, Parrish RR, Condon DJ (2011)

Rapid formation and exhumation of the youngest Alpine

eclogites: a thermal conundrum to Barrovian metamorphism.

Earth Planet Sci Lett 306:193–204

Spear FS, Rumble D III (1986) Pressure, temperature and structural

evolution of the Orfordville Belt, west-central New Hampshire.

J Petrol 27:1071–1093

Spear FS, Wark DA (2009) Cathodoluminescence imaging and

titanium thermometry in metamorphic quartz. J Metamorp Geol

27:187–205

Spear FS, Kohn MJ, Paetzold S (1995) Petrology of the regional

sillimanite zone, west-central New Hampshire, USA with

implications for the development of inverted isograds. Am

Miner 80:361–376

Thomas JB, Watson EB, Spear FS, Shemella PT, Nayak SK,

Lanzirotti A (2010) TitaniQ under pressure: the effect of

pressure and temperature on Ti-in-quartz solubility. Contrib

Miner Petrol 160:743–759

Viete DR, Hermann J, Lister GS, Stenhouse RI (2011) The nature and

origin of the Barrovian metamorphism, Scotland: diffusion

length scales in garnet and inferred thermal time scales. J Geol

Soc Lond 168:115–132

Wark DA Spear FS (2005) Ti in quartz: cathodoluminescence and

thermometry. In: 15th V.M. Goldschmidt Conference, Moscow,

p A592

Wing BA, Ferry JM, Harrison M (2003) Prograde destruction and

formation of monazite and allanite during contact and regional

metamorphism of pelites: petrology and geochronology. Contrib

Miner Petrol 145:228–250

Contrib Mineral Petrol

123