Shale Cuttings Addition to Wellbore Cement and Their Effect on Unconfined Compressive Strength

Abstract

Mitigation of greenhouse gas emissions is becoming a significant factor in all industries. Cement manufacturing is one of the industries responsible for greenhouse gas emissions, specifically carbon dioxide emissions. Pozzolanic materials have long been used as cement additives due to the pozzolanic reaction that occurs when hydrated and the formation a cementitious material similar to that of cement. In this study, shale, which is a common component found in wellbore drill cuttings, was used in various sizes and quantities to determine the effect it had on the mechanical properties of wellbore cement. The unconfined compressive strength of the cement containing shale was compared to the cement without shale to observe the effect that both the quantity and particle size had on this property. SEM–EDS microscopy was also performed to understand any notable variations in the cement microstructure or composition. The samples containing micron shale appeared to have the best results of all the samples containing shale, and some of the samples had a higher UCS than one or more of the base case samples. Utilization of cuttings as a cement additive is not just beneficial in that it minimizes the need for cuttings removal and recycling, but also in that it reduces the amount of greenhouse gas emissions associated with cement manufacturing.

1. Introduction

Wellbore cement is the most used barrier material in hydrocarbon, carbon capture and storage, geothermal, and hydrocarbon wellbores, primarily due to the ability to manipulate the cement properties to adhere to the conditions at hand. For cement to mitigate hydrocarbon leakage, provide zonal isolation, and support the casing, the cement sheath must be able to withstand the overburden stress, remain intact during any subsequent wellbore operations, have the ability to securely bond to both the casing and formation, and be resistant to hydrocarbon migration. The cement must also be designed so that when it is in the liquid phase and being pumped downhole, it will be properly placed, remove any residual drilling mud, exhibit optimal rheological properties, and set in the appropriate period of time [1]. Cementitious materials are inflexible and susceptible to failures such as cracking and shrinking [2]. To enhance the mechanical properties of cement and prevent short- and long-term well integrity issues, additives are mixed with the cement slurry to procure the desired properties.

Additives can be divided into seven primary categories pertaining to the effect that they have on cement: densifiers, accelerators, retarders, viscosifiers, density reducers, friction reducers, and fluid loss prevention. Oftentimes, an additive will fall into more than one of these categories which makes finding a combination of additives to achieve each desired property difficult. Additive selection must take into consideration the pressure, temperature, chemical composition of the formation, mud type, and the presence of formation water, oil, and gas as well as the compatibility with other additives [3]. While the list of additive material that has been used to modify cement mechanical properties is ever-growing, some common additives are nanoparticles, cellulose material, polymers, and pozzolanic material.

Nanoparticles are often used to enhance the mechanical properties of cement because they have large surface areas and are thus able to be more reactive with surrounding cementitious material [4,5]. Nanoparticles used in wellbore cements include, but are not limited to, silica nanoparticles, magnesium oxide nanoparticles, alumina nanoparticles, iron nanoparticles, carbon nanotubes, and magnetic nanoparticles [6]. While nanoparticles have significant promise for enhancing cement mechanical properties, they are often expensive and significantly increase the cost of cementing [7].

Cellulose materials are often used in cement slurries to prevent fluid loss but have proved to be unreliable in both high- and low-temperatures as well as in areas with a high salt concentration [8]. Examples of cellulose materials commonly used in wellbore cement include hydroxyethyl cellulose (HEC), methyl hydroxyethyl cellulose (MHEC), and carboxymethyl hydroxyethyl cellulose (CMHEC), which when hydrated, immediately increase the slurry viscosity [9]. Cellulose materials can prove problematic in that this increased viscosity can lead to issues with pumping the slurry.

Polymers are often added to cement to enhance the elastic properties [10]. Encapsulation of polymers has also been investigated and it was found that such additives beneficially impact the cement mechanical properties and bonding, but due to the complex technique required to formulate such additives may prove costly [11]. The use of shape memory polymers (SMPs), which is a type of polymer that is composed of low-molecular weight pre-polymers and crosslinking agents, can alter its shape when stimulated [12,13,14]. The addition of SMPs would allow for a less brittle and more ductile cement which could prove beneficial for maintaining wellbore integrity for various aspects within the life of a well [15]. One of the drawbacks of using SMPs as sealing material is that they can be impacted by temperature fluctuations [16]. Another downside to this additive material is that little testing has been undertaken to determine how SMPs behave under downhole conditions [17]. Polymers, while improving elasticity in the short-term, could negatively impact the long-term well integrity in that the bond between the cement and polymers is often weak and the polymers could degrade under elevated temperatures and pressures [10].

There are a wide variety of pozzalanic materials that have been used as wellbore cement additives. Pozzolanic material works when the amorphous silica within the pozzolans react with calcium hydroxide that is formed during cement hydration [18]. This reaction leads to an increase in the compressive strength and durability by filling the effective pore space, thus, reducing the permeability, enhancing the calcium-silicate-hydrate (C-S-H) phases associated with the pozzolanic reaction, resulting in higher amounts of inert pozzolanic minerals, and increased nucleation sites for C-S-H formation [19,20]. Common pozzolanic materials used as wellbore cement additives are fly ash, metakaolin, blast furnace slag, and glass microspheres. Fly ash, metakaolin, and blast furnace slag are biproducts of industrial manufacturing processes and are readily available. The addition of fly ash created from the combustion of coal to cement has been shown to increase the compressive strength [21]. Metakaolin is made from calcined kaolin clay and when added to cement, decreases the porosity and permeability, minimizes shrinkage, and minimizes chemical degradation [22]. Blast furnace slag formed during the manufacturing of iron has been shown to aid in preventing gas migration, setting time, and bonding [23,24]. Glass microspheres are a type of pozzalanic material that is used to lower the density of cement but have many drawbacks such as their cost, separation tendencies, and the tendency to be crushed under pressure [25,26,27].

While the use of pozzolanic materials has proven beneficial for a variety of reasons, one of the key aspects that make them so desirable is the potential to minimize the amount of cement needed. Cement manufacturing is one of the largest producers of greenhouse gases, specifically, carbon dioxide (CO₂). Using pozzolanic material as a cement additive would reduce the amount of cement used in the petroleum industry which could lead to a significant decrease in CO₂ emissions.

While the aforementioned pozzolanic materials have been and continue to be investigated as cement additives, there is one type of pozzolanic material that is readily available in the petroleum industry yet has not been investigated: drill cuttings. Drill cuttings are fragments of formation coated in drilling mud that are brought to the surface during wellbore drilling operations. Cuttings disposal is dictated by the type of drilling fluid used and the disposal methods vary by geographic location.

The purpose of this research was to investigate the feasibility of using cuttings as a cement additive and to understand the effect on the cement’s mechanical properties. Various particle sizes and quantities of shale were added to class H cement to understand how compressive strength was altered with the novel geologic additive. Woodford shale has a highly similar chemical composition to other pozzolanic additives, specifically class F fly ash and metakaolin, and its similarities can be seen in Figure 1 [28,29]. Class F fly ash and metakaolin have both been added to slurries that are pumped in the field, indicating the promise of shale additive potential.

While the purpose of this study was solely for feasibility purposes, it could prove to have significant beneficial effects if the work is furthered and various other properties of such slurries are tested and better understood. Annually, roughly 4.1 billion tons of cement is produced globally and is accountable for approximately 7.1% of global CO₂ emissions [30,31,32]. The use of cuttings within cement could lead to a reduction in the amount of cement used in wellbores, thus, reducing the amount of CO₂ emissions as well as decreasing the cost needed to recycle drill cuttings and any associated emissions.

2. Materials and Methods

To understand the feasibility of using cuttings as a cement additive, various sizes and quantities of shale were added to the Central Plains Cement Company class H wellbore cement. The class H cement was from the Central Plains Cement Company, and from the mill sheet the presence of calcium, oxygen, and silica were expected to be the primary constituents of the matrix with smaller quantities of aluminum, iron, magnesium, and sulfur as well as various elements attributed to the insoluble residue as shown in

Table 1. Central Plains class H cement chemical compositions [33].

Component	Amount in Cement (%)
SiO2	21.40
Al2O3	3.20
Fe2O3	5.30
CaO	63.40
MgO	1.70
SO3	2.40
Loss on Ignition	0.90
Na Eq.	0.40
Insoluble Residue	0.28
Free Lime	0.60

[33].

Woodford shale obtained from an outcrop in Murray County, Oklahoma was collected and ground so that there were equal amounts of millimeter (2–4.7 mm), micron (74–210 μm), and submicron (<1 μm) particles. While the millimeter and micron size shale were able to be obtained using conventional grinding techniques with a ring and puck grinder and sieved using a 10-mesh sieve, a 200-mesh sieve, and a Retsch sieve shaker, the submicron samples were obtained using a Retsch EMAX high energy ball mill and the particle size was verified using SEM techniques.

To make the samples, shale was added in 5%, 10%, 15%, 25%, 50%, and 75% quantities to a slurry composed of Central Plains class H cement, a defoamer, and deionized (DI) water. To ensure mixability, the density of each slurry was calculated to be 16.49 ppg and was obtained by changing the water content depending upon the amount of shale added to the system. To mix the slurry, a Chandler 3260 Constant Speed Mixer was used and mixing followed the American Petroleum Institute (API) 10 RB specifications [34]. The slurry was then poured into greased 1.5″ by 4.0″ stainless steel molds and allowed to set for 24 h. After this time, the samples were demolded and submerged in a sodium hydroxide (NaOH)-DI brine in a 150 °F oven for 28 days. After 28 days, the samples were removed from the oven, weighed wet, and placed in a 150 °F vacuum oven for 24 h, then removed, weighted dry and cut/ground to ensure both ends of the sample were smooth and free from any abnormalities, as shown in

Table 2. Specimen density measurements.

Particle Size	% Shale	Avg. Dry Density (10−3 kg/m3)	S. Dev. (10−3 kg/m3)
Base	0	1.82	0.04
Millimeter	5	1.70	0.04
	10	1.78	0.01
	15	1.72	0.01
	25	1.69	0.05
	50	1.40	0.04
	75	1.63	0.05
Micron	5	1.78	0.03
	10	1.62	0.02
	15	1.57	0.02
	25	1.59	0.02
	50	1.43	0.04
	75	1.37	0.02
Submicron	5	1.69	0.02
	10	1.65	0.04
	15	1.64	0.03
	25	1.55	0.04

. Uniformity was confirmed using a caliper and rotating in at three locations with 90° spacing and the values were all within one-hundredth of a millimeter.

UCS testing was performed using a New England Research Autolab-500 uniaxial load frame with a strain rate of 0.04 mm/min. Once testing was initiated, samples were tested until axial force rapidly decreased to the initial axial force value, indicating that the sample had failed and UCS was reached indicating that the test was concluded (Figure 2).

After the samples were tested, scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) was performed on the millimeter and submicron sample sets that had the highest and lowest UCS values using a FEI Quanta 250 and Bruker XFlash 6130. Small pieces of the samples were sputter coated with gold and palladium using a Denton Vacuum Desk V Thin Film Deposition Solution. Once coated, SEM and EDS were performed on various locations at multiple magnifications on the sample surface.

3. Results

The mean UCS and standard deviation for all tests are given in

Table 3. UCS test results for the samples cured over 28 days.

Shale Size	% Shale	Sample Number	UCS (MPa)	Avg. UCS (MPa)	Std. Dev. (MPa)
Base	0%	1	10.52	10.97	0.38
		2	11.44
		3	10.94
Millimeter	5%	4	4.29	6.06	1.29
		5	7.33
		6	6.57
	10%	7	4.90	5.54	1.04
		8	4.71
		9	7.01
	15%	10	8.46	7.71	2.55
		11	4.28
		12	10.39
	25%	13	6.83	5.21	1.14
		14	4.34
		15	4.47
	50%	16	3.71	4.87	0.86
		17	5.13
		18	5.77
	75%	19	8.56	6.04	2.45
		20	2.72
		21	6.85
Micron	5%	22	12.06	9.79	2.02
		23	10.17
		24	7.15
	10%	25	10.83	8.85	2.61
		26	5.16
		27	10.57
	15%	28	11.60	9.26	2.55
		29	10.48
		30	5.71
	25%	31	10.93	7.96	2.25
		32	5.48
		33	7.46
	50%	34	5.63	5.75	0.64
		35	6.58
		36	5.03
	75%	37	7.29	6.07	1.11
		38	6.30
		39	4.61
	5%	40	10.16	9.58	0.63
		41	9.87
Submicron		42	8.71
	10%	43	5.14	7.00	2.03
		44	6.03
		45	9.82
	15%	46	6.00	6.08	0.90
		47	7.21
		48	5.02
	25%	49	6.76	4.88	1.34
		50	4.11
		51	3.77

and shown in Figure 3, Figure 4 and Figure 5. For the millimeter and micron sample sets, all samples were able to be tested regardless of the amount of shale. For the submicron sample set, the 50% and 75% samples broke during the curing process and could not be tested.

Single-factor ANOVA tests were performed on the results in Table 3 to evaluate the statistical significance of the effect of shale size and quantity on UCS (

Table A1. Single-factor ANOVA results for the class H cement samples containing various sizes and quantities of shale cured for 28 days.

Shale Size	Assumption	Groups	Count	Sum	Average	Variance	p-Value
Millimeter	Addition of millimeter shale has no effect on UCS.	0	3	32.8	10.9	0.2	0.02
		5	3	18.2	6.1	2.5
		10	3	16.6	5.5	1.6
		15	3	23.2	7.7	9.7
		25	3	15.6	5.2	1.9
		50	3	14.6	4.9	1.1
		75	3	18.2	6.1	9.2
	Concentration does not matter for millimeter samples.	5	3	18.2	6.1	2.5	0.6
		10	3	16.6	5.5	1.6
		15	3	23.2	7.7	9.7
		25	3	15.6	5.2	1.9
		50	3	14.6	4.9	1.1
		75	3	18.2	6.1	9.2
Micron	Addition of micron shale has no effect on UCS.	0	3	32.8	10.9	0.2	0.1
		5	3	29.5	9.8	6.1
		10	3	26.6	8.9	10.1
		15	3	27.8	9.3	9.8
		25	3	23.9	8.0	7.5
		50	3	17.2	5.7	0.7
		75	3	18.2	6.1	1.9
	Concentration does not matter for micron samples.	5	3	29.5	9.8	6.1	0.3
		10	3	26.6	8.9	10.1
		15	3	27.8	9.3	9.8
		25	3	23.9	8.0	7.5
		50	3	17.2	5.7	0.7
		75	3	18.2	6.1	1.9
Submicron	Addition of sub-micron shale has no effect on UCS.	0	3	32.8	10.9	0.2	0.008
		5	3	28.8	9.6	0.6
		10	3	20.9	7.0	6.2
		15	3	17.6	5.9	5.4
		25	3	14.7	4.9	2.7
	Concentration does not matter for submicron samples.	5	3	28.8	9.6	0.6	0.08
		10	3	20.9	7.0	6.2
		15	3	17.6	5.9	5.4
		25	3	14.7	4.9	2.7

).

The UCS test results fluctuated. From the UCS results, the base case achieved the highest average UCS of 10.97 MPa and the lowest standard deviation (Table 3). Adding the millimeter-sized cuttings lowered the UCS (Figure 3). For the millimeter sample sets, the 15% samples yielded the highest UCS while the 50%-millimeter shale samples had the lowest average UCS. From the 15%-millimeter sample set, sample 12 had the greatest UCS of 10.39 MPa, which was comparable to the base case results (Table 3). The standard deviation for the 15%-millimeter samples was the highest of all the millimeter sample sets with the difference between the highest and lowest samples being 6.11 MPa. Sample 11 achieved a UCS that was less than half of sample 12. The 50%-millimeter sample set had the lowest average UCS values but had the lowest standard deviation, indicating that the results for this sample set were fairly consistent. While the 75%-millimeter sample set had an average a UCS of 6.04 MPa, it also contained the sample with the lowest UCS of all the millimeter samples, sample 20, which had a UCS of 2.72 MPa. As shown in Figure 3, there was no trend in the reduction in strength with increasing cuttings volume.

Figure 4 shows that the micron-sized shale exhibited the most promising compressive strength properties. The samples containing the micron-sized shale appeared to achieve the highest strengths of all the samples containing shale. Of the micron sample set, the 5%, 10%, and 15% samples achieved the highest average UCS but also had high standard deviations. Sample 31, which contained 25%-micron shale, had a high UCS that was comparable to the samples containing lesser amounts of micron shale. The highest and lowest average UCS sample sets with micron shale addition were 5% and 50%, respectively. The 5%-micron sample set had an average UCS of 9.79 MPa with a standard deviation of 2.02 MPa. Sample 22 had the highest UCS of all the micron samples, 12.06 MPa, which was actually higher than any of the base case UCS values. Sample 28, which contained 15%-micron shale, had a UCS of 11.60 MPa, which was also higher than any of the base cases. The 50%-micron samples had both the lowest UCS and standard deviations, indicating that the samples all exhibited similar behavior and the UCS values showed little variation.

Figure 5 shows the submicron sample additives. The 5% and 10% submicron sample sets proved to have higher UCS values than those containing the same amounts of millimeter shale, yet the opposite was seen in the 15% and 25% sample sets with the same particle sizes. Of the submicron samples, the 5% samples had the highest average UCS of 9.58 MPa with a standard deviation of 0.63 MPa, while the 25% samples had the lowest average UCS of 4.88 MPa with a standard deviation of 1.34 MPa. The 10%-submicron samples had a high standard deviation primarily due to sample 45 having a significantly higher UCS than samples 43 and 44.

Figure 6 shows a comparison of the average UCS values for the base case, millimeter, micron, and submicron sample sets to the previously reported UCS values for class G and H cement [35].

SEM and EDS was performed on all the samples at low and high resolutions to understand the variations in both the chemical composition within the samples as well as the surface microstructure. Secondary electron and backscatter electron (BSE) images were created to analyze the sample topography and variations in the chemical composition, respectively. These dark areas were not wholly representative of the pores or voids within the sample. EDS was performed on a designated spot or a specified area on the surface of the sample to provide information about the sample at both a small scale and a macroscale. EDS was performed at low magnification at multiple locations to understand how the quantity of the shale could impact the overall chemical composition of the cement. Locations that appeared to contain shale particles were also selected for EDS analysis, to understand the compositional variations between the geologic material and cement. These locations were also observed as to whether any atypical reactions may have occurred at the cement–shale interface. SEM and EDS was performed on all of the 5%- and 10%-submicron samples as well as the 15%- and 50%-millimeter samples to investigate whether there were any differences in the microstructure, hydration, porosity, or chemical composition amongst the samples containing the same amounts of shale as well as the with the same particle sizes.

Areas and points were used in the EDS analyses; points were specifically selected when it appeared that the minerology differed. EDS analysis for sample 12 containing 15%-millimeter shale is shown in Figure 7.

SEM analysis taken at various surface locations on the 15%-millimeter sample 12 is shown in Figure 8.

EDS analysis to identify the elemental composition on the surface of the 50%-millimeter sample 18 is shown in Figure 9.

SEM analysis of the 50%-millimeter sample 18 is shown in Figure 10.

EDS analysis to determine the compositional variations on the surface of the 5%-submicron sample 40 is shown in Figure 11.

SEM analysis of samples 40 containing 5% submicron shale is shown in Figure 12.

EDS analysis for the 10% submicron sample 45 is shown in Figure 13. EDS analysis was performed for the points taken at 28× magnification and 380× magnification.

SEM analysis of samples 45 containing 10% submicron shale is shown in Figure 14.

4. Discussion

After the samples were crushed, a visual check was performed to determine whether the shale particles appeared to be evenly distributed within the samples. Visual inspection was not able to be carried out for the micron and submicron samples due to the particles being too small to see with the naked eye. For all the millimeter samples, particle distribution appeared homogenous; the shale did not settle to the bottom or float to the top and did not agglomerate within the sample and were evenly distributed throughout.

For all the millimeter sample sets, two samples were similar and one had a significantly higher or lower compressive strength. This indicated that the shale distribution varied between the samples yet upon visual inspection, this did not appear to be the case. From the results, the millimeter sample sets all achieved lower UCS values than the base case and the previously reported UCS values for class G and H cement. First, the addition of shale reduced the dry density of the samples (Table 2). These particles were much larger than the cement particle size and hindered the C-S-H nucleation and growth. This inhibition on hydration leads to weaker points within the cement where the cement is likely to experience failure. It also increases the porosity and permeability of the samples in that there is more free space where hydration products should be forming. In the SEM images of the 15%-shale sample 12, a large void and a number of large shale particles were apparent at low magnification. At higher magnification, the predominant microstructures present included ettringite needles and CH. Secondary electron images show that there were regions of the surface that appear to be slightly elevated. BSE images showed that there did appear to be some charging that occurred in sample 12, but from Figure 8d, it can be seen that the partially hydrated cement appeared darker in color than the surrounding structures. This was also true of the linear rod-like structures, which were CH. Sample 12 had the highest UCS of the three 15%-millimeter samples and showed that the shale was able to bond with the cement and the addition of shale to cement positively impacted compressive strength. The ettringite crystals seen in Figure 8f were similar in structure to those of a Portland cement [36]. EDS analysis for the 15%-millimeter sample locations all appeared highly similar. Figure 7 shows the seven locations where EDS was performed on sample 12. Points 2, 4, and 5 were darker than the other locations and appeared to be shale, but from the elemental analysis only points 6 and 7 show the presence of carbon, indicating that shale was present within the location areas. All sampled locations showed that oxygen was the most prevalent element followed by calcium, silicon, sodium, aluminum, and potassium.

From the SEM images of sample 18, which contained 50%-millimeter shale, an understanding of the interaction between the cement and shale was shown. Shale, which is the smooth, bedded structure, had begun to bond with the cement as shown in the high magnification secondary electron images. Ettringite and C-S-H appeared to be nucleating on the shale, thus, indicating that a reaction had occurred. Ettringite also surrounded a majority of the shale particles and at 2554× magnification, it appeared that the cement particles may actually have bonded to a portion of the shale. It was apparent that the shale was nearly the same brightness/color contrast as the surrounding matrix, meaning that the shale was of similar composition to the cement. Of the 50%-millimeter samples, sample 16 had the lowest UCS due to the occurrence of microcracking or tobermorite formation rather than other hydration products such as C-S-H and CH. While samples 17 and 18 had a similar UCS, sample 18 achieved a slightly higher UCS; this was indicative of the success of the shale bonding to the cement, as evidenced by the BSE and secondary electron images. Figure 10f had a similar microstructure to class G cement containing silica fume, which is a pozzolanic material [37]. The 50%-millimeter EDS results appeared to be consistent with one another and the presence of shale was more prevalent in these three samples. Sample 18 had the presence of carbon in five of the nine EDS locations but the carbon was present in small amounts ranging from 0.5% to 2.3%. Oxygen content appeared higher in the locations without carbon, which was due to the amount of oxygen within the class H cement. All the points also had high amounts of calcium except for point two, which had a higher silicon and aluminum content than the other sample locations. This could also be indicative of the presence of shale.

The micron sample sets had tests that either achieved a higher or comparable UCS than the base case but as a whole, achieved a lower UCS than the previously reported data for class G and H cement. Sample 22 (5%-micron shale) achieved the highest UCS of all the samples tested. Sample 28, which contained 15%-micron shale, had a higher UCS than any of the base case samples. This implied that at lower quantities of shale addition, there was a potential for increased compressive strength due to the pozzolanic reaction occurring between the pozzolanic material, the shale, and the cement. At lower quantities, the micron particles appeared to have provided enough surface area to allow for the hydration of the cementitious materials and allow for ample nucleation points for C-S-H to form and grow. While the 5%-micron shale had the highest average UCS, it also had a high standard deviation. This could have been due to one or more of the samples containing a higher amount of shale, problems during the debonding process, or increased porosity and permeability. The 50% and 75% sample sets all had UCS values that were considerably lower than the other micron sample sets. This indicated that higher amounts of shale hindered hydration and minimized C-S-H formation.

It appeared that as the amount of submicron shale was added to the cement, the UCS decreased linearly, indicating that additives or alterations to the water–cement (w/c) ratio were needed to increase the compressive strength. The addition of 5%-submicron shale particles yielded the highest compressive strength and lowest standard deviation of all the submicron sample sets. This showed that the shale was evenly distributed within the 5%-submicron samples, meaning that these particles aided the hydration and the formation of C-S-H.

The 10%-submicron sample 45 had a UCS similar to the 5%-submicron samples but the other two 10%-submicron samples had a much lower UCS. This indicated the potential for discrepancies in the composition, issues during debonding from the mold, or higher sample permeability. The 50%- and 75%-submicron samples were able to set after 24 h and were removed intact and heated in a brine bath. During the early hydration period in which the samples were being cured in the brine bath, the 50%- and 75%-submicron samples had all broken into smaller pieces and were unable to be tested. This occurrence could have been due to thermal degradation or poor bonding between the submicron particles and the cement. There are very few studies that have investigated the effect of such high concentrations of pozzolanic material being added to cement, specifically past the early stages of hydration. The 5%-submicron shale SEM images appeared to have more spherical particles than either the 15%- or 50%-millimeter shale samples. Sample 40 shows a submicron shale particle that had a thin, smooth, platy structure that was elevated within the cement. There was little to no presence of ettringite, implying that the shale did not reliably bond to the cement. A number of unhydrated and partially hydrated cement particles were also seen in the same area of the shale particle. This implied that there was a lack of water to appropriately hydrate the particles. From the secondary electron images, it was seen that the shale particle was raised within the cement with areas of the shale being higher than others. Figure 12e showed areas of dense spherical C-S-H on the edge of the smooth shale. From the UCS testing for the 5%-submicron shale addition to the cement, sample 40 had the highest UCS, and from the SEM it was seen that there appeared to be large amounts of CH and C-S-H, two of the primary cement hydration products within the sample, even around the shale particles. Secondary electron images showed that there were some shale particles that appeared to be raised above the matrix while others appeared to be within the cement. Sample 42 had the lowest UCS of the sample set and from the SEM images it appeared that there was less CH and C-S-H present but more C-S-H gel. For the three 5%-submicron samples, there did not appear to be a significant amount of ettringite. For the 10%-submicron samples, similar structures to those seen in both the millimeter and 5%-submicron samples were observed. The sample 45 secondary electron images showed that the topography of the surface was smooth; at high magnifications, there appeared to be a slightly raised area where the cement was bonding to the partially hydrated grain, but this elevation was minimal (Figure 14). The edges of the partially hydrated cement particle appeared to be more angular and there was an evident distinction between the smooth, unhydrated portion and the partially hydrated portion. C-S-H and ettringite were also present at the interface. At low magnification, BSE images showed a number of partially hydrated cement particles and a number of pores. The center of the partially hydrated cement was the cement that had not been hydrated and appeared much lighter than the surrounding material; while the outer area of these particles was somewhat darker than the unhydrated cement, it was not homogeneous to the matrix. The honeycomb-like structures seen around the unhydrated cement was C-S-H, which is a common crystalline structure in Portland cement [36]. For the 5%-submicron sample, shale was detected in two of the three samples as evidenced by the presence of carbon. The EDS weight percent results for the 5%-submicron shale sample locations were in good agreement with one another; calcium and oxygen were the two most prevalent elements with trace amounts of potassium, silicon, sulfur, iron, and sodium. Sample 40 showed that shale was present within the selected areas, and often portions of the sample locations appeared to be darker in areas. The cement material, which appeared lighter, was primarily composed of oxygen and calcium, which is in agreement with the Central Plains Cement Mill report. For the 10%-submicron sample, oxygen and calcium appeared to be the most present elements at all the sample locations. This indicated that this piece of the sample had high amounts of cement and lower amounts of shale. For sample 45, only two of the nine locations where EDS was performed appeared to have any trace of shale, being locations five and eight. Along with having the presence of carbon, these two locations also had higher silicon content than the other locations. Silicon is a known constituent of Woodford shale but is also present in the cement. The two locations containing shale appeared to be darker than the surrounding areas on the sample surface.

From the UCS test results, it can be seen that the samples containing micron-sized shale exhibited the most promising compressive strength properties. The 5%- and 10%- submicron sample sets proved to have higher UCS values than those containing the same amounts of millimeter shale, yet the opposite was seen in the 15% and 25% sample sets with the same particle sizes. This indicated that the particle size played a significant role in the cement hydration and that there is an optimal particle size that should be used for future testing to ensure C-S-H formation.

The UCS results from a previous study by [38], which described the effect of 5%-, 10%-, and 15%-millimeter, micron, and submicron shale cuttings addition on the mechanical properties of cement after 7-days curing, were compared to the results found in this study due to the sizes and type of shale used being identical. For both short-term and long-term curing times, the average base case results achieved higher strengths than those with the shale. In both cases, it was seen that the millimeter samples had inferior strengths in comparison to the micron samples. For the long-term study, the average UCS for the 5%-submicron samples was closest to that of the base case. Altering the w/c ratio or using various additives would help to strengthen the systems and aid in achieving higher compressive strengths.

The SEM results were also indicative of the porosity and permeability of the cementitious material. From Figure 8, there were no large pores visible on the 15%-millimeter sample surface at high magnifications and at low magnifications, the pores were less than 10 microns in diameter, indicating that there was some porosity in the sample. Microcracks were not evident at low magnifications, but in Figure 8c,d, thin microcracks were seen on the material’s surface. This was due to a combination of permeability and fracturing from the destructive UCS testing. Figure 10 showed that for sample 18 containing the 50%-millimeter shale, the presence of a long, thin microcrack could be seen at high magnification. At higher magnifications, there were microcracks in the matrix surrounding the dark shale particle indicating an increase in permeability near the shale. Porosity was also higher than the 15%-millimeter sample as the number of pores was greater throughout the matrix. The samples that contained the millimeter shale exhibited some permeability on the fractured sample surface, and this could have an impact on gas migration and fingering in a downhole environment. The increased porosity of the 50%-millimeter sample explained why the compressive strength decreased and indicated that the addition of millimeter shale increased the void space in the cementitious materials. For sample 40 shown in Figure 12, microcracks can be seen at every magnification. While the longer and wider microcracks seen in Figure 12a,b were due to the destructive testing that the sample underwent, the microcracks seen at higher magnifications were indicative of the sample permeability. It can be seen in Figure 12b that the matrix surrounding the shale particle had short, thin microcracks and few pores that were below 20 microns in diameter. Figure 14 shows that there were more pores in sample 45 but they were homogenous in size. The microcracks that were indicative of permeability within the sample are seen in Figure 14d–f; the cracks had penetrated the partially hydrated cement grains. This indicated that fluid flow within the matrix was possible and could lead to the formation of microannuli within the downhole environment. This phenomenon is responsible for oil and gas migration within the wellbore and possibly to the water table or surface. The high permeability needs to be remediated in future testing.

Overall, the micron sample sets appeared to have the best results of all the samples containing shale. This could be due to the fact that the micron shale particles were the most similar in size to that of the class H tested. Having additives with larger or smaller particle size in comparison to the cement can have an effect on the hydration and permeability of the system. Ideally, cement particles should be between 7–200 microns [39].

The results showed that cuttings as a cement additive reduced the strength of the cement, but the cement retained most of its strength (Figure 6) and showed promise as a method for cuttings upcycling. Additional work is needed to further understand the effect of shale on cement’s mechanical properties. Determination of the Blaine fineness of the millimeter, micron, and submicron shale should be performed to understand the particle size of these additives in comparison to the cement; research on the rheology of cement with calcined clay showed that smaller particle size may lead to an increased yield point [40]. Another aspect that needs to be considered is the rheology, consistency, and thickening time of cement containing shale and how various sizes could impact pumpability [41]. An understanding of the ideal w/c ratio is needed to properly hydrate the slurries with this novel pozzolanic additive. It has been noted that pozzolanic materials such as fly ash require an increase in the amount of water needed for hydration, so using the conventional API water–cement ratio may not be ideal for shale or cuttings cements [42].

5. Conclusions

This paper addressed the effect of the use of an environmentally friendly, readily available, upcycled material on the mechanical properties of wellbore cement, to evaluate the feasibility of using the cuttings as a cement additive or replacement material. It is beneficial to reduce the need for cuttings removal and recycling reduces the amount of greenhouse gas emissions, specifically CO₂, associated with cement manufacturing. While this research is the first step in using cuttings as a wellbore cement additive, several insights were gained that may be beneficial for future applications.

• While the base case samples, or samples containing no shale, achieved the highest average UCS, micron samples with 5%, 10%, 15% and 25% shale addition had at least one sample that had a comparable UCS.
• The micron samples proved to have the highest average UCS for all sample sets, which also coincide with the particle sizes of the cement.
• While the submicron samples containing 50% and 75% shale split during the curing process, the samples had a decreased UCS with an increased shale content.
• For the millimeter and micron sample sets, the 75% average UCS values were similar and in both cases, were greater than the samples containing 50% shale, but the 75% sample sets had larger standard deviations. The 50%-shale sample sets for both the millimeter and micron shale had the lowest average UCS of all the sample sets indicating that this amount of shale had a negative impact on hydration and the formation of critical microstructure. The 75% sample sets had a higher UCS than the 50% millimeter and micron, but had a higher standard deviation due to variations in the shale distribution, debonding issues, or increased porosity and permeability. It was apparent from the decrease in density that the addition of shale increased the porosity of the samples with shale.
• This paper shows promise for the upcycling of cuttings as a cement additive. Additional work is needed to further understand the effect of shale on cement’s mechanical properties rheology, consistency, and thickening time. An understanding of the ideal w/c ratio is needed to properly hydrate the slurries with this novel pozzolanic additive.