Set retardation of well cements

1 Historical development

The industrial exploitation of deep and ultra-deep oil and natural gas reservoirs has also increased the difficulties in bore hole cementing. Elevated borehole temperatures call for set-retarded cements, i.e., the cement slurry must remain pumpable until it has reached the location of the final cement mantle. Until 1930 the setting reactions in well cementing were usually accelerated, i.e., the aim was to shorten the “waiting-on-cement (WOC)” time, but later it became necessary to achieve precisely the opposite effect.

This did not represent a problem, particularly as organic admixtures for retarding the setting of Portland cement suspensions had already been known for a long time. In 1882, for example, Tronc [1] recommended the addition of rye flour for extending the setting time and in 1908
Gresley [2] patented dextrin and starch as set retarders. Between 1908 and 1927 the wide variety of natural resins and dextrins had already been used occasionally as set retarders without making any changes to the traditional binder (Portland cement/PC) itself [3].

It was not until 1935 that the first set-retarded PCs for oil well cementation were offered in the USA under the brands of UNAFLOW^® and STARCOR^®. The UNAFLOW^® PC (Universal-Atlas Portland Cement Comp., USA) was retarded with borax, boric acid and gum arabicum [4], while the setting characteristics of STARCOR^® (Lone Star Cement Corp., USA) were controlled by a variation of the raw materials and raising the C₄AF content and/or lowering the C₃A content. Specific changes were also made to the grinding of the Portland cement clinker – the cement raw meal became coarser and the specific surface area was reduced. Despite of these substantial changes there was an increasing demand for ever more effective set-retarding admixtures.

The effectiveness of casein for retarding the setting of Portland cements was patented in 1942 [5].

Ten years later Hansen [3] was able to list more than 20 different admixtures that were said to extend the setting time.

According to Smith [6] the market in 1975 was dominated by the following set-retarding admixtures:

Lignin derivatives (Ca/Ca-Na lignosulfonates),

Carboxymethyl hydroxyethyl cellulose (CMHEC), and

NaCl in concentrations > 15 % by weight of cement (BWOC)

So far the inexpensive lignin derivatives have proved particularly successful. In addition to retarding the set, they also improve the degree of dispersion (plasticizing effect) of binder suspensions. Combinations of two or more products usually produce a synergistic effect.

Up to a static bore hole temperature of about 140 °C, lignin derivates or CMHECs can still be used successfully at concentrations of 0.1-1.5 % (BWOC). However, at higher temperatures the retarding effect drops sharply in spite of increasing additive concentrations. Sodium carboxymethyl cellulose (Na-CMC) is also used as a set retarder in Russia instead of CMHEC, which is not always available. The deficient thermal stability of the carboxymethyl cellulose can apparently be increased by simultaneous addition of antioxidants (phenols, etc.), so that they are said to be usable up to 200°C. The well-known sensitivity of CMC to Ca²⁺-ions does not appear to be a problem [7].

Numerous organic admixtures (such as hydrolyzed pentosans, sulfite spent liquor, whey and whey proteins [8], mustard and linseed oil pressed cake) as well as inorganic salts (such as various phosphates, silicofluorides, calcium nitrate and -bromide, aluminium chloride) are also described as set retarders in the technical literature and in patents. The use of zinc and lead oxides and of metallic zinc beads with diameters of 0.1-1.5 µm [9] as set retarders seems strange.

It is not always clear which components of a complex chemical compound are actually responsible for retarding the setting of cements, because, for cost reasons, it is normal to describe and use technical products with numerous impurities.

The limit for using natural organic raw materials for set retarders lies at borehole temperatures around 100°C, corresponding to a depth range of 3000 to 4000 m, and only those inorganic or synthetic organic substances - with proven temperature stability - are normally used.

Because of their great importance in oil well cementation, set retarders are given particularly high priority among all cement admixtures.

All suitable set retarders should be highly compatible with all standardized API oil well cements (API = American Petroleum Institute) [42], with other admixtures and additives, auxiliary materials and the components contained in reservoirs.

Set retardation of oil well cements (borehole cements/Portland cements) should more precisely be designated as an “Inhibition of Cement Hydration”.

Results of the set retardation of oil well cements  (acc. to API specifications) can usually also be applied to many other inorganic binders.

2 Types of set retarders

Basically, most of the wide variety of set retarders used industrially for oil well cements can be subdivided into the following main product groups:

Starch and cellulose derivatives (I),

Lignosulfonates and lignin-based derivatives (II) [30, 31],

Sugar and sugar derivatives (III) [29-33 and 35-40],

Modified phosphonic acids and phosphonates (IV) [25-27],

Acrylic acid derivates, styrene-acrylate copolymers (V),

Hydroxy acids (with one or more OH-groups) and their salts (VI),

Proteins and their derivatives (VII),

Inorganic salts (VIII) [33, 34], and

Special compounds (IX).

For stricter classification it is possible to work out common chemical structural features for set retarders. These would then just become:

Polyhydroxy and polycarboxy compounds (I, II, III, V, VI),

Amines, polyamino- and polyamido-compounds (VII) [33],

Inorganic salts (VIII), and

Complexing agents (IV).

The mechanisms of action of set retarders can also be explained by these chemical and structural features.

2.1 Polyhydroxy- and polycarboxy-compounds

The polyhydroxy- and polycarboxy-compounds that affect the setting processes of the cements through their OH- or COOH-functions in the molecule, include:

Saccharides and polysaccharides as well as their modifications and secondary products, such as:

– Wheat and rye flour,

– Starches, oxidized and sulfonated starches,

– Dextrins and their decomposition products,

– Natural resins, such as gum arabicum gum, tragacanth, etc.,

– Cellulose derivatives, such as methylhydroxy-ethyl cellulose (MHEC), hydroxyethyl cellulose (HEC), hydroxypropyl cellulose (HPC), carboxymethyl hydroxyethyl cellulose (CMHEC), carboxymethyl cellulose (CMC), methylhydroxypropyl cellulose (MHPC), oxidized cellulose,

– Hydrolyzed pentosans,

– Pressed cake of various oils (with restrictions),

– Oxyaldehydes and oxyketones,

– Various sugars and their derivatives,

Polyphenols, polyhydroxy phenols, hydroxycarboxylic acids and their derivatives as:

– Tannins and their salts,

– Humic acids and their salts,

– Gallic acids and their salts,

– Lignin derivates, lignosulfonates and their salts,

– Tartaric acid and its salts.

Polyacrylates and their derivatives as well as their copolymers.

The technologically most important products from this group are polyacrylates, cellulose derivatives (especially CMHEC), lignosulfonates and various sugars, that can all be produced economically and on a large scale with virtually uniform quality.

However, low-molecular polyacrylates and their copolymers with acrylamides are sensitive to shear stress (mechanical degradation in turbomixers, pumps and at valves); they are easily hydrolyzed at moderate temperatures and react with Ca²⁺-ions, causing flocculation. These properties usually limit the range, in which these products can be used, to low depths with moderate temperatures (< 60 °C) and to those cement slurries, which are prepared with water of low salinity. Undesirable increases in viscosity rarely occur with low-molecular products.

Industrial cellulose derivatives are available in a large number of modifications and molecular weight distributions. The most favourable and cost-effective use of these products as set retarders can only be determined through extensive tests. While anionic types of CMC are rarely suitable for cement suspensions, modified cellulose derivatives are predominantly stable to shear, temperatures up to about 140°C and are not sensitive to Ca²⁺-ions. Any hydrolysis products that appear are usually monosaccharides and short-chain polysaccharides.

When compared with other cellulose derivatives, CMHEC has proved to be the most frequently used product as a set retarder in concentrations of 0.1-0.3 % BWOC. However, cellulose derivatives are often accompanied by unwanted increases in viscosity, so dispersants and plasticizers are frequently applied with these products. The influence of these macromolecules on the early setting processes in cements is based on the same principles as that of the various low-molecular mono-, di- or polysaccharides.

The properties of industrial lignosulfonates (usually Na- or Ca-salts) differ very widely from manufacturer to manufacturer due to the natural variation in the lignin raw materials (molecular weights vary from 20 000 to 50 000 g/mol). Lignosulfonates have only a weak effect on the cement setting processes, but, because of their substantial residual sugar content, they can cause pronounced setting retardation, even at low addition levels. This means that the setting retardation is due essentially to the residual sugar content of the lignosulfonates. As macromolecules (three-dimensional network of phenyl propanoids) the lignosulfonates act predominantly as dispersants or plasticizers for cement suspensions.

The low-molecular mono-/di-saccharides, sugar alcohols and oligosaccharides have proved to be particularly effective as set retarders. In extensive test series on about 20 different, chemically pure sugars and sugar alcohols, Akstinat was able to demonstrate the dual character of the individual low-molecular saccharides [10]. Contrary to general opinion, this showed that sugars act both as set retarders and as set accelerators (Figs. 1 and 2).

The following sugars and sugar alcohols were investigated:

Group I: trehalose, meso-erythrite, rhamnose, ribose, sorbose, meso-inosite.

Group II: cellobiose, arabinose, xylose, fructose, maltose, lactose, glucose, galactose, cane sucrose, mannose, raffinose, mannite, sorbite.

The overwhelming number of sugars in Group I and II exhibited marked set acceleration in the concentration range of < 0.01-0.1 % BWOC. Sugars of Group I may still act as set accelerators up to 0.4 % BWOC (and in some exceptional cases even above this value), but the sugars of Group I show a strong  set retarding action at concentrations > 0.3-0.4 % BWOC already. From this it follows that - due to their moöecular structures - saccharides do not have a consistent effect on the early setting processes of cements and develop very specific characteristics in Portland cement suspensions. Great accuracy of addition of these inexpensive basic compounds or of their mixtures with other set accelerators/set retarders is therefore essential for saccharides.

2.2 Amines, polyamino and polyamido-compounds

This class of compounds not only contains well known natural substances, such as:

Proteins and substances containing proteins, such as:

– Casein and modified derivatives,

– Whey and whey derivatives,

– Oil press cake (with restrictions), as well as

– Types of flour (with restrictions),

but also fully synthetic products, such as:

– Polyacrylamides and their derivatives/copolymers,

– Polyethylene imines (PEI),

– Polyaspartic acid (PAS) and its derivatives,

– Polyalkylene-polyamine condensates (EDA),

– Copolymers of epichlorohydrin and amines (DMA),

– Primary, secondary and tertiary amines.

2.3 Inorganic salts

The most important inorganic salts that can interact with the Ca²⁺-ions or the ionic equilibria during the cement setting process are, in particular:

Boron compounds, such as boric acid, borates, tetraborates, pentaborates, etc.,

Various phosphates, such as sodium hexa-metaphosphate, hydrogen phosphates and polyphosphates,

Aluminium salts,

Silicofluorides.

Some authors have mentioned raising the Ca²⁺-content by addition of CaBr₂, Ca(NO₃)₂ or Ca(OCl)₂ in order to achieve set retardation [3, 6]. However, it has not been possible to verify this information.

High levels of NaCl >> 200 g/l in the mixing water can also significantly extend the setting times [18].

2.4 Complexing agents

As early as the 1970s Akstinat [11] had investigated numerous complexing agents (EDTA, NTA, TETA, DTPA, phosphonates, gluconates, glucoheptonates, etc.)¹ in regard to the elimination of mineral scales and deposits in oilfield pipelines, with the assumption, that such products also affect the early crystallization processes in cement slurries for cementing. Simple phosphonobutane tricarboxylic acids were in fact known as flow improvers [12] and free phosphonic acids as soil stabilizers [13], but not as set retarders for cements. From the large number of phosphonates available on the market Akstinat was able, rather surprisingly, to identify amino-/hydroxyl-phosphonic acids and their derivates (DEQUEST^® Group [14]) as excellent set retarders for cements (Figs. 3 and 4).

Although these research results had already been published around 1980 [15-17], numerous U.S. companies (including Halliburton, Dow, ­Sandox) [20] attempted later to patent various complexing agents of this type as set retarders for oil well cementation.

Still more extensive optimization trials with three-dimensional (3D) molecular-modelling and sythesis of  novel phosphonic acid products were carried out in the following years jointly between Monsanto Europe and Akstinat.

Ultimately, the low-molecular products (DEQUEST 2041^®: Ethylenediamino-tetramethylene phosphonic acid and Dequest 2060^®: Diethylenetriamino-pentamethylene phosphonic acid) proved to be particularly effective set retarders for oil well cements. Research products of BASF with quarternized polyethylene iminoethylene phosphonates (PEIP) and amine oxides of PEIP also gave comparable results [21].

Specific amino- and hydroxy-phosphonic acids and their derivatives have a strong and persistent effect even in concentrations << 0.03 % BWOC at 50°C or < 0.3 % BWOC at > 130°C.

None of the familiar, commercially available set retarders achieve such favourable characteristic values. Modified amino-/hydroxy-phosphonates and their derivatives are therefore often also designated “super set retarders” or “long-term retarders (LTR)”. This means that the phosphonic acid derivates first proposed by Akstinat have a leading role as set retarders for oil well cements based on Portland cement.

3 Working mechanisms of set retarders

The various groups of set retarders do not exhibit a consistent mode of action. The technical literature discusses topics, such as:

Calcium complex formation,

Surface adsorption,

Blocking of crystal nucleus formation (nucleation poisoning), and

Surface reactions through formation of insoluble barrier layers.

However, several processes are usually superimposed, so that it is not unusual to find controversial information in the literature.

Compounds with several OH-, COOH-, NH₂- or amide-functions tend essentially towards surface adsorption on sulfate activators, ettringite and C₃A, so that an interference of the crystal growth is also involved. Especially with sugars, Akstinat [10] was able to use molecular models to document the spontaneous adsorption processes of saccharides (Fig. 5).

The assumption that insoluble calcium saccharate layers are formed must be contradicted, because no saccharic acid appears or is newly formed during the spontaneous interactions during the setting process in the presence of saccharides. In fact, a dissociation of the various saccharides/polysaccharides often occurs in alkaline media by ring opening in line with a Lobry-de-Bruyn-Alberda-van-Ekenstein rearrangement (special case of Keto-Enol-Tautomerism) [19]. These rearrangements can - depending on the molecular structure - be accompanied by ionization (equilibrium of neutral and anionic Monosaccharide molecules) and mutarotation.

Free electron pairs lead to a stereo-chemical
molecular orientation or deformation not only at the hydrating, positively charged surfaces of the cement particles but also at the surface of the sulfate activators. Mineral fracture-edges, -tips and -corners lead to high charge densities at which molecules with free electron pairs are taken up preferentially (unless already saturated by cement grinding aids).

As is also known from other professional disciplines [23], partial areas of varying extent of a mineral surface can, depending on the type of adsorption and size of molecule, become “shielded-off”, which could also account for the differing, non-stoichiometric application concentrations. Mobile adsorption types (i.e. funnel- or umbrella-adsorption, etc.) as well as stable adsorption types (i.e. comb-, bridge-, horizontal-chain- or cloud-adsorption) are known.

In most cases the action of mobile types of adsorption with only one adsorption centre fades significantly at higher temperatures, while increased temperatures will have much less impact on the stable types of adsorption with several adsorption centres (e.g. with saccharides).

Small molecules lead to tighter packing densities on mineral surfaces, with more favourable energy characteristics, while macromolecules usually still have large gaps in their coverage and can therefore only achieve moderate to low set retardation. Network structures (such as lignosulfonates) therefore only provide slight set retardation, while low-molecular saccharides achieve a significantly greater inhibition of the setting processes. From other specialist fields it is well known that surface adsorption with simultaneous complexing [22, 24, 28] can disrupt or completely prevent the diffusion processes and crystal formation, and is known as inhibition.

With inorganic salts a set retardation can usually be explained by the deposition of sparingly soluble compounds on the mineral surfaces. In some cases the salts that have been added reduce the Al³⁺- and Ca²⁺-concentrations to such an extent by precipitation and complexing that the formation of ettringite and syngenite is virtually suppressed. Sparingly soluble borates, phosphates and silicofluorides of calcium and aluminium are well known and are basically responsible for the set retardation by forming barriers.

The action mechanism of complexing agents has to be explained differently. At first it was mistakenly assumed that a strong surface adsorption occurs with modified phosphonates. The application of 3D-modelling computer programs on the adsorption process by Akstinat and Monsanto [11] were not able to explain either the non-stoichiometric concentration phenomena or the small effect of specific structure modifications.

These investigations led to the conclusion, that specific phosphonates do in fact also cause complexing, but that the set retardation is based on a chemical surface reaction, which forms sparingly soluble calcium phosphonate complexes and in this way effectively inhibits the cement hydration [25, 26]. These findings were also confirmed by later work in 2003 by Bishop et al. [25] and in 2004 by Rickert and Thielen [26, 27]. Only the product-specific stability constants of these complexes are responsible for the solubility (redissolving, dissociation and temperature resistance) of the barrier layers.

Some set retarders, especially Sugars and certain modified Phosphonic Acid derivatives, have a tendency in certain concentration and temperature ranges to “switch”, i.e. a set retarder becomes a set accelerator and vice versa. This process can occur unexpectedly during oil well cementing due to underground dilution effects (mixing, inflows) and therefore represents a serious risk, if an oil well cement job is not carried out properly.

Excess addition of super set retarders (LTR) can also completely prevent the cement from setting, so the addition of these products must be carried out very carefully.

4 Areas of use of set retarders

Set retarders are used in drilling technology wherever:

Long pumping and workability times (deep and ultra-deep wells) can be expected,

High bore hole temperature and pressure conditions occur, and

Special binder systems and compositions are to be used.

The applications of set retarders are in no way limited just to the API oil well cements / Classes A to J (PC), but in most cases also cover the range of special cements (e.g. pozzolanic systems, gypsum-based and Sorel cements, lime binders, etc.).

Set retarders are predominantly encoded with codes or brand names and are used directly and expensively by the cementing service companies, who attempt to keep their product know-how and the chemical composition of their admixtures strictly secret.

4.1 Comparison of commercial set retarders
with industrial raw materials

Akstinat and Arens [16] compared commercially available set retarders from Halliburton (Duncan, OK/USA), the leading service company for oil well cementation,

HR-4^® and HR-7^® recommended for ≤ 125 °C static BHT,

HR-12^® recommended for ≤ 190 °C static BHT,

HR-20^® recommended for > 150 °C static BHT,

with industrial raw materials.

Determination of the chemical compositions of these standard admixtures from Halliburton gave:

HR-4^® and HR-7^® mixtures of lignosulfonates,

HR-12^® mixture of lignosulfonate and gluconate,

HR-20^® mixture of HR-12^® and potassium
pentaborate.

For comparability the setting time was selected as a function of the concentration f(c_admixture) at three temperatures (50°C, 90°C and 132°C).

The quantities of the traditional set retarder added are 0.1-1.5 % BWOC (for ~ 50°C); addition levels of ≥ 2.5 % BWOC are recommended for higher temperatures (≥ 132°C).

4.1.1 Gluconates/glucoheptonate

Three products were examined from the group of gluconates/glucoheptonates because the gluconates in particular are said to have a good performance at high temperatures:

SEQLENE^® 540 (Pfanstiehl Lab., Waukegan, IL/USA) = Na-salt of D-glucoheptonate C₇H₁₃O₈Na·2H₂O (chemically pure, as dry powder)

SEQLENE^® ES-40 = identical industrial product to SEQLENE^® 540, but as 40 % aqueous solution

Calcium gluconate = intermediate product from the oxidation of aldonic acid to 2-keto-aldonic acid Ca(C₆H₁₁O₇)₂, as dry powder.

Even at addition levels of 0.05 % BWOC and low temperatures (50°C) the gluconates and glucoheptonates prove to be highly effective as set retarders. As expected, the effectiveness of all three trial products significantly exceeded that of the commercially available HR-7®/HR-12® set retarders both at 50°C and at 132°C.

4.1.2 Calcium glycerophosphate

Another trial product, the organophosphate Ca(PC₃H₈O₆)₂ in 50 % aqueous solution, was also checked for possible use as a set retarder. The effectiveness of glycerophosphate is comparable with that of the commercial HR-7^®/HR-12^® set retarder.

4.1.3 Lignin-based set retarders

KELIG^® 32 (American Can Comp., Greenwich, CT/USA) is an industrial sodium lignosulfonate, that contains the Na-salts of hydroxy acids and other substances contained in wood. Its effectiveness at both 50°C and 132°C is also significantly superior to that of the HR-7^®/HR-12^® commercial products provided by the service company Halliburton.

4.1.4 Modified amino and hydroxy phosphonic acids

Six different phosphonic acid derivatives of the DEQUEST^® product range were chosen from this group and compared with the high temperature set retarder HR-20^®. At a temperature of 132°C and an addition level ≤ 1 % BWOC all the phosphonic acid derivatives exhibited in various ways comparable or better set retardation than HR-20^®. However, at target retardation times > 1.5 hours two of the chosen phosphonates (DEQUEST^® 2060 and DEQUEST^® 2041) were far superior to the commercial set retarder HR-20^®. In comparison with standard products from service companies none of the industrial commercial products examined exhibited any losses in compressive strength or elastic modulus.

4.2 Action of set retarders in suspensions
of non-standard binders

In view of the importance of set-controlling admixtures in drilling technology, Arens and Akstinat [15] also compared the behaviour of other non-standard binders and of oil well cement G with different classes of set retarders. Among others, the following material combinations were examined:

Binders:

Mortar binders (ROCADUR^®)/MB²,

Normal high-alumina cement (TSZ),

Oil-shale cement (ÖZ),

Oil well cement, class G (TBZ G),

Set retarders:

HR-12^® and/or HR-20^® (borates, etc.),

KELIG^® 32 (lignin-based),

SEQLENE^® 540 (glucoheptonate),

DEQUEST^® 2041 (modified phosphonic acid derivative).

All tests were carried out at a temperature of 132°C (API schedule 9 g/6) [42] as the aim was, that the results achieved should be largely representative and applicable to drilling technology. The binder systems are characterized here in greater detail (see Table 1 for average analysis values).

4.2.1 Oil-shale cement (ÖZ)

This cement cannot be retarded as well as the standard oil well cement (TBZ G) tested previously. In most cases it was only possible to carry out measurements in the higher concentration range of set retarders, for which the values shown in Table 2 were obtained for a setting time of 200 min.

4.2.2 Normal high-alumina cement (TSZ)

The very short setting time of high-alumina cement (without set retarder) of only about 10 min must be stressed. This cement does not respond, or only responds very poorly, to the set retarders used here. The reference setting time of 200 min (compared with TBZ G) can only be achieved with TSZ by addition of about 1.8 % BWOC of HR-20^® (because of the presence of pentaborates). The target setting time of 200 min cannot be approached with any of the other organic set retarders used, even with addition levels > 3 % BWOC.

4.2.3 Mortar binder (ROCADUR^®)/MB [lime cement]

The characteristic concentration-setting curves of all the set retarders used for the mortar binder are virtually linear with relatively steep gradients, so that any cost-effective application must be considered as restricted to a maximum addition level of only < 1.5 % BWOC. For this reason the reference setting time of 200 min was abandoned and a time of 100 min was chosen (Table 3). This means that the set retarders exhibit a significantly lower response with lime cement (MB) than with TBZ G.

5 Conclusion

With these trials by Arens and Akstinat it was then possible to show that - apart of API well cements:

the setting behaviour of different other binder systems, that are rarely used or have not yet been used for securing bore holes, can be controlled with both traditional and novel admixtures,

the industrial products described here can also be used as effective set retarders for non-standard binders up to at least 132°C, and

high-alumina cement exhibits a completely different setting behaviour from Portland cement, oil-shale cement and mortar binder and cannot be controlled with the known organic set retarders based on lignins, gluconates or phosphonic acids. Only inorganic salts (e.g. borates) can achieve reliable control of the hydration of high-alumina cements.

Basically, it must be assumed that preliminary testing with different set retarders should always be carried out before non-standard binder systems are used.