CYROGENICS WITH CEMENT MICROSCOPY REDEFINES CEMENT BEHAVIOR

Sudhir Mehta, Richard Jones
ARCO Exploration & Production Technology
Plano, Tex.

Bill Caveny
Halliburton Energy Services
Duncan, Okla.

Cement microscopy (CM), cryogenics, environmental scanning microscopy (ESM), scanning electron microscopy (SEM), and other technologies are leading investigators to change their views on cement gelation, hydration, and retardation.

The process of drying samples fundamentally changes the chemistry and/or morphology so that what the viewer sees in the laboratory may not represent what is actually happening during the setting process or what is being applied in cementing operations.

Cement samples frozen in a nitrogen slush and viewed with an SEM present a more accurate picture of the setting process. Observations made through this technique have revolutionized ARCO Exploration & Production Technology's and Halliburton Energy Services' oil field cement procurement and slurry design.

In one case, a compound long believed to control gelation was found to be an artifact of sample preparation. Besides this, analysis shows that other compounds will require different retarders, accelerators, etc. to establish control of the gelation process.

The study also contradicts the well-accepted belief that all void spaces in cement are caused by air trapped during hydration.

Findings from this joint study are expected to lead to:

• Optimized waiting on cement (WOC) times

• Reduced planning and design time

• Optimized slurry retarder additions

• Optimized gel times to fit given situations; especially applicable to squeeze operations

• Improved cement selection (from vendors) for peak performance

• Improved cement manufacture.

CEMENT VOIDS

Voids, or bubbles in cement, as shown in Fig. 1a, have historically been accepted to be air or gas pores. This conclusion was based on cement samples that had been air-dried, polished, and examined under a microscope. But a sample from the same slurry (Fig. 1b) viewed with low-temperature scanning electron microscopy (CRYO-SEM) shows that voids are sometimes water droplets encapsulated by calcium hydroxide. This water could significantly affect the cement job because these "water voids" might be large enough to undermine cement integrity by allowing corrosive gases and fluids to permeate the matrix.

The CRYO-SEM process allows the observation of encapsulated water droplets by:

• Flash freezing cement samples in liquid nitrogen slush (cryogen) at about -210 C. after selected time periods such as 10, 30, and 60 min, and 24 hr.

• Arresting the hydration process.

• Giving a snapshot of the process without drying.1 A true picture in real time can then be studied.

The environmental scanning electron microscope (ESEM) allows scientists to examine in situ hydration as well as wet-hydrated products of Portland cement to characterize early cement hydration reactions. ESEMs have characterized morphology and phase chemistry (over periods from 1 to 150 hr of hydration) of the four principal constituents of cement: C3S, C2S, C4AF, and C3A.2 The work showed unique hydration characteristics associated with each of the four phases and enabled isolation of competing reactions that occur during early hydration of commercial cements.

EQUIPMENT

The recently developed environmental scanning electron microscope (ESEM), although it can observe and record dynamic reactions in real time, is somewhat limited in characterizing relationships between liquids and solids (Fig. 2a). During examinations, excess liquid can mask the surfaces of the reacting solids. The ESEM cannot "see" through a film of water, although grains still remain in contact with water. For example, unbound water may move freely and flood the surface of the specimen, while the water continues to react with the cement grains.

To overcome ESEM limitations in setting cements, the CRYO-SEM technique can characterize solids, liquids, and reaction products at the desired moment during hydration. The hydration reaction is arrested by quick-freezing cement pastes in liquid nitrogen. Then the frozen pastes are examined in a conventional scanning electron microscope (SEM) equipped to analyze frozen samples (Fig. 2b).

In biological examinations the CRYO-SEM technique typically assesses the true nature of ultrastructural cell details. In nonbiological systems, the techniques can determine oil/water distribution in reservoir rock and in clay and latex suspensions.34 The current work is the first known application of CRYO-SEM to oil field cement analysis. The box lists the steps for preparing a wet cement sample for viewing under SEM.

CEMENT GELATION

Gelation control is critical to oil well cementing because:

• During pumping, high displacement and friction pressures from early gelation can break down producing formations.

• During pumping, high pressure from early gelation can rupture tubulars.

• In the well bore, gelation can contribute to poor annular fill and zonal isolation.

• In sensitive formations, late gelation can allow excessive fluid loss that can damage the formation.

• In the well bore, late gelation can allow gas migration.

Conventional cement microscopy, such as viewing dried (totally dehydrated) slurry samples mounted on slides, indicates that ettringite formation is a prime component of the gelation process. Fig. 3a shows ettringite crystals in an air-dried sample.

But a frozen sample from the same cement batch (Fig. 3b) reveals no ettringite; however, incipient crystallization of calcium hydroxide crystals is clearly evident. Fig. 3c shows both hexagonal and amorphous calcium hydroxide.

An X-ray diffraction analysis (Fig. 4) on the same samples viewed by the CRYO-SEM process compared the following three different sample preparations:

1. Solvent extraction-Acetone was added to the frozen samples to remove water.

2. Oven dried-Sample was heated to evaporate water.

3. Freeze dried-Sample (still frozen) was dried under vacuum.

Examination of the samples revealed:

• Solvent extraction resulted in formation of a small amount of ettringite.

• Air drying produced more ettringite.

• Freeze drying showed no ettringite formation. This supported the CRYO-SEM finding that ettringite is not the major gelation mechanism in this set of cement samples.

• Calcium hydroxide crystallization is evident in all three samples and is believed to be the cause of gelation.

Magnitude of the peak intensity is not indicative of the percentage of calcium hydroxide formed.

BACKGROUND

In cement research, a general acceptance is that as cement hydrates gelation occurs when:5

• Colloidal products form

• Crystalline products form

• Both colloidal and crystalline products form

Gelation commonly is attributed to tobermorite gel, a colloidal product, and ettringite crystals. The importance of calcium hydroxide is usually considered minimal.

In some cements, both colloidal and crystalline products seem to contribute significantly to hardening. In high-sulfate cements, the compressive strength at early ages seems to be due mainly to ettringite formation, whereas at later ages, tobermorite gel plays an increasingly important role. Hardening can thus be associated with formation of colloidal products, crystalline products, or both.

Fig. 3 shows the thickening time curve for a cement sample prone to premature gelation. Increases occur during formation of calcium hydroxide. Decreases appear to correspond to dissolution of calcium hydroxide crystals.

DISCOVERY IMPACT

The discovery that ettringite is sometimes an artifact of cement sample preparation rather than a true chemical reaction that causes cement to gel means that slurry designers have concentrated on a nonexistent problem. Designers have not dealt with the detrimental effects of calcium hydroxide because they have not known that it is the prime constituent of some slurries undergoing gelation.

With ettringite diminished in the design equation, designers might now need to change practices related to control of early gelation by:

• Stopping further expenditure of time, resources, and money toward ettringite formation control for preventing early gelation.

• Analyzing problem cements to determine economical treatments for preventing gelation.

• Ensuring that cement does not contain a high percentage of free lime because a high free-lime level in a slurry contributes a proportionately high calcium hydroxide level.

• Determining new procedures for controlling gelation during cement manufacturing.

This knowledge should enable development of chemistry that controls the formation of calcium hydroxide during formation, thus allowing accurate prediction of waiting-on-cement (WOC) time.

RETARDATION MECHANISMS

Detailed study of cement hydration and retardation processes, using the CRYO-SEM technique, has revealed methods of cement and retarder selection that optimize job design by:

• Obtaining better cement displacement, casing protection, and zonal isolation.

• Allowing more control in reducing WOC time.

• Optimizing retarders and other slurry additives.

• Reducing required planning and laboratory testing time.

• Lowering operating costs by improving primary cementing successes and reducing remedial cementing.

To devise an optimum mix of additive and predict required setting time, hydration products of neat (no additives) cement were mixed with three popular oil well cement retarding agents: lignosulfonate (for low temperatures), polymer (for medium temperatures), and tartaric acid (for high temperatures).

EXPERIMENTS

Classes A and H oil well cements were studied. Slurry samples first were mixed with retarders and allowed to hydrate in consistometers until they reached 70 Bc (Bearden consistency units). At 70 Bc, cement slurry no longer can be pumped. Then to arrest hydration reactions, the samples were frozen in liquid nitrogen slush at -210 C. before being examined in the CRYO-SEM.

A base line study was made on neat (no additives) cements to establish a comparison base. Thickening times and CRYO-SEM analysis from the base line study were compared with a study of samples with retarders added.

Tables 1 and 2 summarize the results. Significant findings are as follows:

• Neat Class H cement slurry reached 70 B, in 2 hr, 38 min, showing no ettringite but having numerous crystals of calcium hydroxide (Fig. 6a). The primary gelation mechanism is therefore linked to calcium hydroxide.

• Class H cement with 0.2% lignosulfonate retarder required 3 hr and 19 min longer to gel, with a total time of 5 hr and 57 min to reach 70 Bc. This is attributed to the capability of lignosulfonate to block crystallization of calcium hydroxide. No ettringite was seen nor was lignosulfonate coating found on silicate grains.

• Class H cement with 0.5% tartaric acid did not gel at all, reaching only 5 Bc consistency after 32 hr; however, with 0.2% concentration, consistency reached 70 Bc within 6 hr (Fig. 6b).

In the 0.5% sample, there is a conspicuous lack of calcium hydroxide crystals. It appears that tartaric acid keeps the mix water pH low enough to prevent calcium hydroxide crystallization and creation of a network of calcium hydroxide crystals, thus inhibiting gelation. No ettringite was observed, indicating that it is not a factor in early gelling of Class H cement with or without the addition of tartaric acid or other organic retarders.

• Class H cement with 0.2% tartaric acid plus silica flour (35%) achieved compressive strength in 9 hr and 30 min (Fig. 6c). With 10% silica fume, strength was attained in 4 hr and 46 min (Fig. 6d). Gelling mechanisms differed for silica fume and flour.

Addition of silica flour produced the needle-like ettringite. This event is attributed to a virtual drying effect created as the calcium-rich water was soaked up by silica particles, allowing ettringite to form (drying samples for viewing causes ettringite to form).

Adding fumed silica produced no visible ettringite and had shorter gel time, probably because of lower concentration and increased contributions of amorphous calcium-silicate-hydrate (C-S-H) gel and calcium hydroxide. Note that only 10% silica fume was added relative to 35% silica flour.

• Class H cement with 0.3% polymer reached compressive strength in 10 hr and 4 min (Fig. 6e). Adding polymer appeared to suppress free mixing of various ions in the mix-water, as evidenced by the globularized gel, retarding early gelation processes. Adding 0.2% tartaric acid to this mix shortened gelation time by 3 hr and 30 min.

• Class A cement with 0.2% tartaric acid gelled in 28 min, attributed to extensive precipitation of calcium hydroxide and gypsum crystals (Fig. 60.

• Class A cement without retarder developed early compressive strength in 1 hr and 2 min. Calcium hydroxide and C-S-H gel appeared to be principal factors in early gelation time. Again, no ettringite was observed.

• The addition of 0.3% lignosulfonate to Class A cement raised the gelation time to 1 hr and 55 min, with primary components of the gelation being calcium hydroxide crystals and C-S-H gel (Fig. 6g).

• Class A cement with 0.5% tartaric acid did not gel after 36 hr. This is attributed to suppression of calcium hydroxide formation. No ettringite was observed (Fig. 6h).

IMPACT ON OPERATIONS

Research is continuing. Based on observations made to date, the following operational changes are expected by ARCO and Halliburton.

• Cements that have in the past been difficult to control may now be conditioned for practical use in oil field cementing operations.

• Differences in raw material and manufacturing practices globally have contributed to variations in cement gelation characteristics. New manufacturing processes and improved understanding of the hydration process can result in a more uniform, higher quality cement.

• Designing the optimum cement slurry for given well conditions challenges the creativity of the oil field cementing engineer. Gelation control will ease this process and reduce engineering and laboratory testing time.

• WOC times should be shortened since the amount of retarder added to slurries to assure pumpability can be reduced.

• Real-time investigation of other hydration reaction mechanisms has led to improvements in additive research and development, such as conditioning aids to ease slurry mixing, control premature gelation, and improve cement displacement.

• Better cement quality will improve primary cementing success rates and provide zonal isolation for maximum productivity and well life.

• Scientific improvements in the analysis of cement hydration will have a major impact on oil field economics, and may also benefit the construction industry.

REFERENCES

1. Mehta, S., Jones, R., Caveny, W.J., Chatterji, J. and McPherson, G., "Cement Hydration During the First 24 Hours Examined by Cryo-scanning Electron Microscopy," 15th International Conference on Cement Microscopy, April 1993.

2. Mehta, S., Jones, R., Caveny, W.J., Chatterji, J., and McPherson, G., "Environmental scanning electron microscope (ESEM) examination of individually hydrated Portland cement phases," 16th International Conference on Cement Microscopy, April 1994.

3. Mehta, S., "Imaging of Wet Specimens in Their Natural State Using ESEM," Paper No. SPE 22864, 66th Annual Technical Conference and Exhibition, Dallas, Oct. 6-9, 1991.

4. Sutanto, E., Davis, H.T., Scriven, L.E., "Liquid Distribution in Porous Rock Examined by Cryo Scanning Electron Microscopy," Paper No. SPE 20518, 65th Annual Technical Conference and Exhibition, New Orleans, Sept. 23-26, 1990.

5. Taylor, H.F.W., Cement Chemistry, Academic Press, 1990.