APPLICATION OF FOURIER TRANSFORM INFRARED SPECTROSCOPY TO SILICA DIAGENESIS: THE OPAL-A TO OPAL-CT TRANSFORMATION S.B. RICEI*, H. FREUND~**, W.-L. HUANGI, J.A. CLOUSE1, ANDC.M. ISAACS~ i Exxon Production Research Co., P.O. Box 2198, Houston. Texas 77252 U.S.A. z U.S. Geological Survey. 345 Middlefield Road. Menlo Park, Cal~omia 94025 U.S.A.

and Thompson 1984), and both cristobalite and tridymite frameworks have low (a) and high (I3) structures, adding more complexity. Because of the low temperatures it is likely that the low forms are involved in the opalCT intergrowths. However, the difference in framework connectivity that distinguishes idealized cristobalite from idealized tridymite is the most important consideration for the present work. For the porcellanite-rich sections of the Miocene Monterey Formation, Murata and Nakata (1974) and Murata and Larson (1975) showed that the reaction progress depends on burial depth. They showed using XRD that during diagenesis the structure of opal-CT changes, resulting in decrease of the (101) planar spacing (d) of cristobalite, and narrowing of the peak. Increased time and temperature result in quartz crystallization.Considerable variation in temperatures (35-51 °C) for the opal-A to opal-CT reaction has been obtained from downhole DSDP measurementsin the Bering Sea (Hein et al. 1978) and from calculations of oxygen isotope equilibration temperatures (41-56°C) for the Temblor Range in California (Murata et al. 1977). Using a thermal model, Pisciotto (1981) estimated a minimum 9-27°C transformationtemperature for the Monterey Formation in the Santa Maria Valley. His diagenetic temperatures calculated from oxygen isotopes were INTRODUCTION higher, 37-50°C. Matbeney and Knauth (1993), using a stepwise fluoriSeveral important basins, particularly those rimming the Pacific Ocean, nation technique, recently reported oxygen isotopic temperatures from the contain large thicknesses of biogenic silica. During burial diagenesis, dia- same area, and suggested that opal-CT can form at temperatures 10-30°C tom frustules (opal-A) react to form a disordered crystalline phase called lower than previously assumed. Botz and Bohrmann (1991) reported opalopal-CT, which constitutes dense stratigraphicunits that are prominent seis- CT formation in Antarctic deep-sea sediment at 0-4°C, also calculated from mic reflectors in some deep-sea surveys (Yhein and yon Rad 1987; Hein oxygen isotopes. Variability in opal-CT transformation temperatures might et at. 1978). Although its fine-grained nature and the inherent disorder in be expected from host-rock lithology and details of burial history. In adits silica framework have prevented a full description of opal-CY, it is most dition, there is uncertainty in the temperature estimates because of calibralikely a cristobalite-like structure incorporating randomly intergrown trid- tion problems and lack of equilibrium, especially for the shallow Antarctic ymite (FI6rke 1955; Graetsch et al. 1987). This interpretation, included by sediments. Keller and Isaacs (1985) proposed a model for estimating temJones and Segnit (1971) in their definition of the opal phases, is the one peratures in clay-rich diatomaceous rocks; see their paper for a discussion tentatively accepted by most investigators. In addition to X-ray diffraction of the factors influencing temperature estimates. Here we accept that opal-CT formation is a temperature-activatedpro(XRD) evidence, the cristobalitic nature of opal-CT is implied by the fact that when heated to 1000°C, many opal-CT specimens are converted either cess. Hydrothermal products from experiments to develop a rate expression to low cristobalite or opaI-C (Jones and Segnit 1971; Tada and Iijima 1983; were used for this study. This transformation has potential for paleotherElzea et al. 1993). Jones and Segnit defined the various opal phases mainly mometry in the temperature range of interest for petroleum generation, and on the basis of XRD patterns, but aim included that opals contain at least accurate, precise methods for measuring the progress of opal-CT precipi1 wt % water. Spectroscopic studies have concentrated on the forms of tation are needed. water and OH in opals (e.g., Langer and Fl6rke 1974; Adams et al. 1991). XRD is typically used to follow the reaction of opal-A to opal-CT, and 2~Si nuclear magnetic resonance (NMR) spectroscopy has proven effective although it is a standard, reliable technique, suitable for relatively small in the determination of framework coordinations in silicates, but has not amounts of material and unaffected by most impurity phases, it is not ideal proven useful for opal characterization. Adams et al. (1991) and de Jong for two reasons. (1) The small particle size and lack of perfect ordering in et al. (1987) were unable to distinguish opal-CY from opal-A by NMR. opal-CT mean that the diffraction peaks are often quite broad and therefore The small particle size may be partly responsible. difficult to measure. (2) Because the starting material is noncrystalline silLippincott et al. (1958) and Plyusnina et al. (1971) studied the structure ica, either volcanic glass or biogenic silica such as diatoms or radiolaria, of infrared features in several silica phases and discussed the interpretations XRD cannot determine the structure of the material during the early stage of various frequencies. Assignments were difficult for cristobalite and trid- of the reaction. ymite because the structures were not well known at the time, and even In contrast, Fourier transform infrared spectroscopy (FTIR) does not denow the extremely complex structure of tridymite is controversial. Tridy- pend on long-range order. Instead, its response to molecular-scaleenergetmite is complicated by the existence of several polymorphs (Wennemer ics and symmetries provides a glimpse of a variety of atomic vibrations, such as stretching and bending, whose frequencies depend only on shortrange interactions and can therefore be applied equally well to crystalline * Present address: McCroneAssociates,Inc., 850 PasquinelliDrive, Westmont, or noncrystalline solids. Chester and Elderlield (1968) used infrared specIL 60559 **Present address:ExxonResearchand EngineeringCo.. Route 22 East, Annan- troscopy to determine the opal contents of deep-sea sediment, and Fr0hlich (1989) proposed molecular models for biogenic silica on the basis of IR dale, NJ 08801

ABSTRACT: An important goal in silica diagenesis research is to understand the kinetics of opal transformation from noncrystalline opalA to the disordered silica polymorph opaI-CT. Because the conventional technique for monitoring the transformation, powder X-ray diffraction (XRD), is applicable only to phases with long-range order, we used Fourier transform infrared spectroscopy (FTIR) to monitor the transformation. We applied this technique, combined with XRD and TEM, to experimental run products and natural opals from the Monterey Formation and from siliceous deposits in the western Pacific Ocean. Using a ratio of two infrared absorption intensities (w = 147~ cm '/15~ cm '), the relative proportions of opaI-A and opaI-CT can be determined. The progress of the transformation is marked by changes in slope of w vs. depth or time when a sufficient stratigraphic profile is available. There are three stages in the opai-A to opaI-CT reaction: (1) opaI.A dissolution; (2) opal.CT precipitation, whose end point is marked by completion of opaI-A dissolution; and (3) opaI-CT ordering, during which tridymite stacking is eliminated in favor of eristobalite stacking.

Journal OF Seel[menxarYrEsEarcH, VoL. A65, NO. 4, OctoBer, 1995, P. 639-647

Copyright© 1995.SEPM(Societyfor SedimentaryGeology) 1073-1318/95/0A65-639/$03.00

640

S.B. RICE ET AL.

0.8-

TABLEl.--Mineralogy of Monterey porcellanites in the San Joaquin subsurface, Chevron APC-65 well* OpaI-A

O.7-

Sample depth Opal (m) CT

Quartz Ca(cite Dolomite

Siderite

Plagio- Clay + erase Zeolite

d loll IA)

I(472)I I[500)

4.110 4100 4.089 4.102 4.097 4.106 4.097 4.11)6 4.114 4 102 4.093

1.52 1.54 1.28 t.22 1.25 1.20 1.20 1.17 1.14 1.17 1.13

0.6-

i

0.4-

407 413 439 449 464 523 550 565 591 623 653

0.4

0.3-

o.2 I I , 7 600 500 400 300

, ~ 0.2-

/

.~,

o ~ ' ' 4000 3600 I

'

'

'

i ' ' '

3200

cm-1

i

'

'

2800

'

i

'

'

2400

St-O bands

'

i

'

2000

'

'

i

'

'

1600

'

I

'

1200

•

'

i

'

'

I

400

Wavenumber (era"1) Fic. I.--FrIR diffuse reflectance spectrum fur opal-A (Monterey diatomite). Inset shows the region of interest for monitoring the opal-Mopal-CT transition.

for

data. We investigated the potential using FTIR to monitor the transformation of opaI-A to opal-CT, and to gain insight into the reaction by the combined use of FHR. XRD, and transmission electron microscopy (TEM) applied to experimental run products and natural opals. SAMPLES AND TECHNIQUES

Suites of specimens were obtained from experimental runs reacting a Monterey Formation diatomite, from diatomaceous shales and porcellanites recovered from the Chevron APC-65 well in the San Joaquin subsurface, California, and porcellanites from Ocean Drilling Program (ODP) Leg 129 (Behl and Smith 1992), western Pacific Ocean. Diatomites are shales rich in biogenic opaI-A from diatoms and radiolaria. Porcellanites are rich in opal-CT, and are derived from diatomites by diagenesis. Therefore, the experimental runs were intended to produce opals similar to those in Monterey porcellanite. Pure synthetic cristobalite (CB-25) and tridymite (TY-

0,8-

0 0 1 0 0 1 5 5 5 2 2

0 0 2 0 !!

0 I ~ace trace ~ace 0

0 0 ~ace I 0 0 trace 0 0 0 10

0 8 11 5 8 5 6 6 5 2 2

2 28 29 24 28 20 20 28 30 30 22

27) from N.I.O.S.H. (National Institute for Occupational Safety and Health) were used as reference materials for both XRD and FHR. Experiments were carried out by hydrothermally reacting acid-washed Monterey diatomite from Thousand Oaks, California at 300°C and 0.5 kb in seawater (I.A.P.S.O. standard seawater 1'99 (Institute of Oceanographic Sciences, Wormley, England) sampled offshore Copenhagen), between 0 and 180 days. Subsurface samples are from the Monterey Shale in the Midway-Sunset oil field in the San Joaquin basin. From these diatomite and porcellanite samples, containing quartz, feldspar, and calcite, 11 samples with the lowest A1203 content (4.3-5.9 wt %) were chosen for analysis (Table 1). Carbonate phases are in low abundance, and most of the nonopal material is other silicates. ODP Leg 129 recovered Middle Jurassic through Lower Miocene porcellanite and chert from the Pigatetta and East Mafianas Basins, west-central Pacific. From among these samples, the purest porcellanite specimens were chosen for analysis, as well as several containing 20-50% of quartz or calcite to assess the effects on the IR spectra (Table 2). Samples from Sites 800 and 801 (Pigafetta Basin) were combined with Site 802 (East Marianas Basin) to obtain a depth profile. For details of the petrology and geology of this setting, see Lancelot et al. (1990); for additional opal-CT characterization, see Behl and Smith (1992). Powder XRD and FTIR were carried out on all samples. The XRD patterns were collected on a Siemens D5000 diffractometer, using Cu Ka radiation at 45 mA and 25 kV with a step size of 0.02 ° 20. Equal weights of the run products were prepared on glass slides, whereas rock samples

TrT

0.6-

Cristobaliie

~.

~,,

/2!

0

02-

j,ll

.

2200

7 11 13 9 I0 12 7 8 3 6 10

* Values are in volume %, 5 10% relative error: trace values are less than - 1%.

'

800

91 54 44 60 53 61 60 53 57 57 53

2000

18(X~

i t

.i~t, :" ~t

...~ "~

/3"

1600

1400

12'00

1000

Wnvenumbers (era "11

i

i

800

600

400

FIG. 2.--FHR diffuse reflectance spectra for Monterey opal-CT, synthetic tridymite (NIOSH, TY-27), and synthetic cfistobalite (NIOSH, CB25).

TIR IN SILIC4 DIAGENESIS

CT

4A

4.25~TJ~ L !1

641

r

~

/

CT

~

co 19 ¢: ....

I

I , ,

0

I,

I

f

I

15

I

I

I,

20

,

I,

,,

,

25

I

,

,

30

,

,

I

,

,

35

,,

40

Two theta, degrees

F=c. 3.--XRD patterns for experimental run products of Monterey Thousand Oaks diatomite, heated in seawater at 0.5 kb and 300°C for 5, 35, 60, and 180 days. The peaks are labeled for cristobalite (C), tridymite (T), and quartz (Q).

CT 4A

c

Q

i

CT

4.25A "'TEl

2042ft, APC-65 J . . . .

10

I

15

. . . .

I

20

. . . .

I

25

'

' ' I ....

Two theta, degrees

30

I ....

35

40

F]c. 4.--Comparison of XRI3 patterns of the 180-day run product with a Monterey opaI-CT (625 m, APC-65 Chevron well). The experimental opal resembles the opal-C of Jones and Segnit (1971).

642

S.B. RICE ET AL.

FIG. 5.--TEM images of experimental run products. Thousand Oaks diatomite starting material, upper left; 35-day run product; 60-day mn product; 180-day run product. The opal particles coarsen with run duration.

were mounted in powder pack holders. Phase compositions (Tables 1, 2) were determined from diffraction intensities compared to mineral standards. Errors in these percentages are 5-10% relative. FTIR spectra were taken in a Bomem MB102 spectrometer. Rock samples were crushed and ground into fine powder to avoid absorption band broadening. Powdered samples were mixed with an inert matrix (1% sample in either KBr or CsI) and inserted into SpectraTech's diffuse reflectance attachment. Spectra were collected at 4 cm ~ resolution; 20 scans were coadded and averaged. Transmission spectra were generated by referencing the acquired spectra to the pure matrix spectrum and converted to absorption spectra using Galactic Software's Spectra Calc data acquisition software.

Several experimental samples were examined by TEM, using the JEOL 2000FX at Exxon Production Research Co. and a Philips 420ST at Exxon

Research and Engineering Co. in Annandale, New Jersey. Powders were dispersed in methanol and pipetted onto holey carbon support films. RESULTS AND DISCUSSION

Figure 1 shows a typical absorption spectrum of a Monterey diatomite (opal-A) sample. The broad band at 2600-3800 cm- J is a response to OH absorption. Hydrocarbons would show an absorption at 2700--3100 cm ~, but these are absent from this material. The principal Si-O stretch is around 1100 cm-~, and there are several other Si-O bands that represent different modes of the Si-O system. Of particular interest are the bands near 470 cm ~ (inset of Figure 1) and 500 cm -~, which develop in the opal-CT (Fig. 7). Lippincott et al. (1958) assigned absorptions at 485 and 515 cm 1 to Si-O bending modes in cristobalite. The features we used, men-

FTIR I]V SILICA DIAGENESIS

643

TABLE2.--Mineralogy of ODP Leg 129 porcellanites. WesternPa~(fic Ocean Sample ID 80~)A-6r-l-53-55 811~)A-Sr-l.2S-33a 801)A-Sr-l-28-33b 801A-Tr-4-40-~ 8flllA-12r-l-25-29a g01A-12r-l-17-19u 801A-12r-l-17-]9b 8