Physicochemical Characterization of Sludge Originating from Vegetable Oil–Based Cutting Fluids

A fresh (same day sampled) homogenized sample of sludge was used for analysis for all of these tests. A sample was placed on a microscope slide and carefully compressed using another slide. Magnifications between $10\times$ and $200\times$ were used to observe the sample, using dark field, light field, and phase contrast microscopy. Alongside this a dip slide was used to calculate the number of living bacteria within the sample. Both sides of the dip slide were wiped across the surface of the homogenized sludge and then incubated at 37°C for 48 h.

Vacuum-dehydrated sludge was used for carbon content analysis. Total carbon (TC) and total inorganic carbon (TIC) analysis was conducted using a Shimadzu SSM-5000 A solid sample module and Shimadzu TNM-1 detector; total organic carbon (TOC) was calculated by the difference between the two. TC in triplicate was measured using 0.5 g of dehydrated sludge. A calibration check was run and showed accuracy to be $\pm 5\%$ against a glycerol standard for TC and against calcium carbonate standard for TIC prior to analysis of the sludge. Three repetitions of each analysis were undertaken, with the results showing the average of the three, giving Sludge 1 a standard deviation of $\pm 0.6\%$ and Sludge $2\pm 1.5\%$.

A Varian 3100 FTIR was used along with a Varian 600 UMA FTIR microscope to analyze the unadulterated sludge, which was placed directly on the crystal. Scans at a resolution of 4 cm, wave number of $\mathrm{4,000}–650{\mathrm{cm}}^{-1}$, and 100 cycles were conducted.

Melting point tests have been used in analyses of FOG samples (e.g., Williams et al. 2012; Nitayapat and Chitprasert 2014). The melting point test in this case was conducted to aid in differentiation between metal-bound fatty acids and FFAs. Using a syringe, 10 g of sludge was injected into the bottom of a test tube and placed into a water bath, which was heated at a rate of $5°\mathrm{C}/\min$ from ambient. Once the sludge started to melt the bath was allowed to cool by 10°C and the test tube replaced with fresh sludge. The heating rate was altered to $1°\mathrm{C}/\min$ increase, where the initial and full melting points were measured to determine the melting range. This was repeated three times for each sample.

Separation and removal of the water is required for organics analyses by GC-MS and has been achieved using various methods in the literature applied to FOG samples and oily wastes and soils: (1) melting the lipid followed by and water-fat separation by centrifugation or use of anhydrous sodium sulphate to absorb water, or by drying at 103°C (Hamilton and Rossell 1986) or 105°C (Nitayapat and Chitprasert 2014), (2) entrainment distillation using the Dean and Stark method (Hamilton and Rossell 1986), (3) direct oven drying at temperatures between 101 and 105°C (e.g., Hamilton and Rossell 1986; Williams et al. 2012; He et al. 2011; Keener et al. 2008), (4) freeze drying, (5) microwave drying [ASTM D4643-08 (ASTM 2008)], and (6) vacuum dehydration using a vacuum desiccator with silica gel crystals. Water determination in oil sludge, but not water separation can be achieved using nuclear magnetic resonance (Zheng et al. 2013) or Karl-Fischer titration.

The properties of fats can be affected by excessive heat, chemical addition, and oxidation. In particular, heating can cause oxidation of fatty acids. Although drying is the most popular method, in this separation the aqueous inorganic elements remain within the dried sludge. Thus in this study liquid-liquid separation was the preferred method to allow determination of the deportment of inorganic elements between organic and water portions of the sludge.

Water separation was conducted for GC-MS and ICP-OES analysis using the following technique. Into a separating funnel 25 g of homogenized sludge and 50 mL of dichloromethane (DCM) were added and shaken vigorously, then left to settle for 15 min to develop a boundary layer. The lower layer with DCM and dissolved organics was drained. This was repeated five times, at which point the DCM was clear, indicating that minimal solvent-soluble organics remained within the separating funnel. The solvent was evaporated from the solution, leaving the resultant oily residue to be weighed and subject to further analyses; this residue is referred to as solvent-separated sludge (SSS).Three repetitions were undertaken, with the results showing the average of the three.

Water-insoluble organics are typically extracted from environmental solids, such as soils, clays, sediments, and sludge using Soxhlet extraction [e.g., Furniss et al. 1989; U.S. EPA method 3545 (U.S. EPA 2007)]. In the present study the extractor used was a Dionex accelerated solvent extractor (ASE100). A polar (hexane) and nonpolar (acetone) solvent mix, of $1:1$ ratio was used as the extractant. The following ASE operating conditions were used: 175°C oven temperature, 10,300–13,800 kPa (1,500–2,000 psi) pressure, 5–10 min static time, 60–75% of cell volume for flush, 60 s nitrogen purge at 1,030 kPa (150 psi), and one cycle for extraction. (U.S. EPA 2007), repeated once to ensure complete extraction. Non-solvent-extractable (NSE) solids and solids $>0.2\mathrm{\mu m}$ were filtered out during the extraction, leaving the ASE sludge for further processing. After extraction the solvent extract was cooled to room temperature, where solids were seen to have formed and were filtered using a 0.2-μm filter. Solids filtered out at this point are hereafter referred to as solvent extract filter residue (SEFR), while the liquid to be analyzed by GC-MS is known as the solvent extract filtrate.

The solvent extract filtrate was qualitatively and quantitatively analyzed using a Perkin Elmer Clarus 500 GC-MS using an Elite 5MS column of 30 m length, 0.25 mm inner diameter, and film thickness of 0.25 μm, with the following settings: system pressure 72 kPa ($\sim 10.5\mathrm{psi}$), $1\mathrm{mL}/\min$ helium carrier, injection temperature of 250°C, and temperature program comprising 100°C initial temperature, hold for 2 min, increasing at $4°\mathrm{C}/\min$ until 310°C, hold for 4 min. The column was cleaned between analysis runs using blank DCM samples. Qualitative analysis was performed using the NIST MS Search 2.0 database. Qualitative analyses indicated that palmitic and oleic acid were the dominant chemicals, with quantitative analyses being performed using a three-point calibration using standards made up from analytical grade palmitic acid and oleic acid in DCM. Triglycerides with more than 58 carbon bonds in their structure were suspected to be in the sludge, however due to their high molecular mass and boiling point (e.g., 851°C for trimethylolpropane trioleate) they would not elute from the columns. Samples of 99.5% purity of palmitic and oleic acid were run using the same operating conditions to confirm that the NIST database was correctly identifying the compounds.

Due to the organic content of the sludge (approximately 99% excluding water), digestion was required. Typical digestions in the literature include the following examples used for FOG-type samples: (1) refluxing with 50% ${\mathrm{HNO}}_3$ at 95°C (Williams et al. 2012), (2) digestion using concentrated nitric acid at 95°C followed by addition of hydrogen peroxide (Keener et al. 2008; U.S. EPA 1994), (3) microwave-assisted acid digestion using ${\mathrm{HNO}}_3$ and of HF (or other appropriate acid), ${\mathrm{KMnO}}_4$ or ${\mathrm{H}}_2{\mathrm{O}}_2$ used as oxidants (U.S.EPA 1996), and (4) ashing at 500°C followed by acid ($\mathrm{HCl}/{\mathrm{HNO}}_3$) digestion (He et al. 2011).

Ashing was trialed but found to be problematic, therefore microwave digestion was used: To 0.5 g of sludge 9 mL of ${\mathrm{HNO}}_3$, 3 mL of HCl, and 2 mL of ${\mathrm{H}}_2{\mathrm{O}}_2$ were added. The sample was heated to 180°C in 5.5 min and maintained at that temperature until the total heating time reached 10 min. The digested sample was diluted to 50 mL using deionized water and subjected to ICP-OES analyses (Perkin Elmer Optima 2100DV) using a multielement certified standard.

Ten repetitions of the full sludge extraction, microwave digestion, and ICP-OES analysis were conducted.

Microscopic examination of wet sludge spread onto a glass slide revealed that the sludge samples predominantly consisted of an emulsion of spherical liquid droplets, typical in appearance to an oil emulsion. A very small amount of microbial biomass, namely, some fungal fragments, were also present. Dip-slide testing (data not included) of the sludge revealed only moderate microbial growth. Fig. 3 shows spherical droplets of a liquid with another liquid. This is likely to be a water-in-oil emulsion (inverted emulsion) due to the low water content of the sludge. Also visible are minor fragments of swarf from the machining processes.

The TC, TIC, and TOC analyses sludge showed that none of the TC present was present as TIC, ruling out the presence of dissolved or solid phase carbonates in the sludge and that all the $\mathrm{TC}=\mathrm{TOC}$ for the sludge. The values of TC for the vacuum dehydrated sludge were 79.84% ($\pm 0.6\%$) and 75.47% ($\pm 1.5\%$) for Sludges 1 and 2, respectively. TC corrected to the wet sludge was 55.60 and 48.64% for Sludges 1 and 2, respectively.

A number of organic compounds were detected (Table 1) in the solvent-separated sludge using the NIST library. Of these compounds palmitic acid and oleic acid had by far the largest peak areas, therefore these were subsequently analyzed quantitatively. No vegetable triglycerides were detected, although their presence is suspected due to their existence in the virgin VOCF, but the GC-MS settings did not allow elution of triglycerides.

The quantitative analysis of the sludge revealed that palmitic acid formed 23.95 and 11.78% of total sludge mass and oleic acid content was 17.18 and 23.13% of Sludge 1 and Sludge 2, respectively. These are within the range of the 10–70% oleic acid content found in FOG (Nitayapat and Chitprasert 2014). Oleic acid and palmitic acid were found in many FOG samples (Nitayapat and Chitprasert 2014; Keener et al. 2008; Williams et al. 2012) in varying concentrations and ratios.

Non-solvent-extractable solids account for 2.29 and 1.50% of Sludge 1 and Sludge 2. The solvent extract filter residue comprised 0.49% of Sludge 1 and 0.71% of Sludge 2 by mass. These solids are suspected to be calcium soaps, i.e., calcium palmitate, oleate, and stearate, which while not soluble in organic solvents at room temperature and pressure (Harrison 1924) may have been solubilized under the elevated pressure and temperature of the ASE extraction to later precipitate as the solvent cooled.

The results of the 28-element ICP-OES analysis are detailed in Table 2. The relative concentrations are $\mathrm{Ca}>\mathrm{Na}>\mathrm{Al}>\mathrm{Fe}>\mathrm{Mg}>\mathrm{Li}>\mathrm{Zn}$, with these seven elements comprising 94% of the elements found within the sludge. Relatively high concentrations of the main elements present remain within the water after solvent separation, showing that significant amounts of the calcium, sodium, and lithium are not bound to the solvent soluble organic fraction. This is important because studies on FOG by authors such as Nitayapat and Chitprasert (2014), He et al. (2011), Williams et al. (2012), and Keener et al. (2008) have not separated the water from the organics in this way and have assumed that all of the inorganic elements detected are inherent within the organic phase, which is not the case for this sludge.

Calcium concentrations within the organic portion of the sludge were 135.5 and $\mathrm{1,303}\mathrm{mg}/\mathrm{kg}$ for Sludge 1 and Sludge 2, respectively. Total concentrations in the wet sludge are $932.3\mathrm{mg}/\mathrm{kg}$ ($\pm 3.6\%$) for Sludge 1 and $\mathrm{1,973}\mathrm{mg}/\mathrm{kg}$ ($\pm 3.6\%$) for Sludge 2. These concentrations are only 10–20% of the total concentrations found by Williams et al. (2012), although those results are based on dried sludge, so could be up to 20 times lower if water content of up to 95% was included in the concentration calculation. FOG studied by Keener et al. (2008) had a calcium content of two to four times greater than that of the sludge under study and a sodium concentration within the range found in Sludge 1 and Sludge 2.

The FTIR spectra are almost identical for Sludge 1 (Fig. 4) and Sludge 2 (Fig. 5), which show intense, sharp peaks at $\mathrm{2,800}–\mathrm{3,000}{\mathrm{cm}}^{-1}$, which correspond to the $\mathrm{C}─\mathrm{H}$ bond. Smaller but still sharp peaks are seen at $\mathrm{1,600}–\mathrm{1,800}{\mathrm{cm}}^{-1}$, representing $\mathrm{C}═\mathrm{O}$ bonds; more precisely, the peak is between 1,650 and $\mathrm{1,750}{\mathrm{cm}}^{-1}$, which relates to acids and esters. Another $\mathrm{C}─\mathrm{O}$ bond is detected at $\mathrm{1,200}–\mathrm{1,300}{\mathrm{cm}}^{-1}$. Overall there are $\mathrm{C}─\mathrm{O}$ bonds, $\mathrm{C}─\mathrm{H}$ bonds, and $\mathrm{C}═\mathrm{O}$ bonds, along with acids and ester bonds. All of these bonds are present within palmitic and oleic acid, however the most common bond in these chemicals was the $\mathrm{C}─\mathrm{H}$ bond, which relates to the intense peak seen at $\mathrm{2,800}–\mathrm{3,000}{\mathrm{cm}}^{-1}$. The detection of esters confirms that not all of the oil had been degraded into free fatty acids.

Direct comparisons with calcium palmitate, palmitic acid, oleic acid, calcium stearate, stearic acid, and the mineral oil used in the metal machining factory showed that no one chemical alone matched the sludge spectra, but each had their similarities on different peaks, except the mineral oil. Therefore it is likely that the sludge is formed of a mix of these chemicals, along with others chemicals that were not analyzed.

The results from water separation, GC-MS, and ICP-OES analyses were combined to form an overall composition for the sludge, shown in Fig. 6. The sludge has been shown to be an organically based inverted emulsion, composed mostly of water, oleic acid, palmitic acid, and unidentified organics, with little or no mineral oil present. These unidentified organics are suspected to be triglycerides (rapeseed oil or trimethylolpropane trioleate) found in the virgin VOCF and other organic compounds that have been qualitatively identified (Table 1).

The composition of the sludge can be described as similar to that of FOG because water, palmitic acid, oleic acid, and calcium are present in similar concentrations. Water (moisture) content of the sludge is on average 32.6%, compared with 54.7% (Williams et al. 2012) and 53.9% (Keener et al. 2008), although the range of moisture content in the studies by Williams et al. (2012), Keener et al. (2008), and Nitayapat and Chitprasert (2014) have a wide variation of 0–95%.

Soap is defined as a salt of a fatty acid, saturated or unsaturated, containing at least eight carbon atoms or a mixture of such salts (IUPAC 2014). Soaps have been identified as a principal component of FOG (e.g., Keener et al. 2008), in addition saponification has been identified as a suspected causal mechanism in the formation of FOG deposits (He et al. 2011; Williams et al. 2014). Thus it is important to establish whether soaps, particularly calcium palmitate and calcium oleate, contribute significantly to the composition of the VOCF sludge or whether the FFAs dominate.

The mass requirements of calcium bonds to two fatty acids to form soap (Ca:FFA) from stoichiometry are 1:12.8, 1:14.1, and 1:14.2 for palmitic acid, oleic acid, and stearic acid respectively. The mass requirement of sodium bonds to one fatty acid to form soap (Na:FFA) from stoichiometry are 1:11.1, 1:12.3, and 1:12.3 for palmitic acid, oleic acid, and stearic acid respectively. Assuming that all of the calcium, sodium, and lithium are bonded to oleic acid and palmitic acid in a $1:1$ ratio, the soap content of Sludge 1 would be 0.35%, and it would be 1.85% for sludge 2. Soap is a minor contributor to the chemical makeup of the sludge compared to water and the FFAs, oleic acid, and palmitic acid. While it is feasible that the soap may act as a surfactant in the inverse water in oil emulsion, it is unlikely that soap is important in the formation and accumulation of the VOCF ETP sludge.

Furthermore, the level of Ca present as calcium palmitate is lower than this because calcium palmitate and calcium oleate are insoluble in water (Harrison 1924), and the water contains 86% of the Ca in the sludge (Table 2). This highlights the benefit of using solvent extraction to separate water from the sludge compared to drying, which causes aqueous constituents to remain within the dried sludge and appear as part of an undifferentiated total sludge assay.

Ca is present in high concentrations, which is also observed by Williams et al. (2014) in FOG deposits. They suggested that biocalcification may be an important mechanism in this Ca accumulation (and in FOG formation), however in the case of the VOCF sludge the TC, TIC, and TOC analyses show that inorganic carbonate such as calcite is not present in significant concentrations within the sludge.

The presence of FFA rather than soaps is further supported by the melting point analysis (see comparative data in Tables 1 and 3). Calcium palmitate, calcium oleate, and calcium stearate have no definite melting point, rather they are observed to conglutinate at 110°C and 115–120°C, respectively, and decompose at temperatures 140–160°C (Harrison 1924). The melting point of palmitic acid and oleic acid, principal components of the sludge, are reported to be 63 and 14°C, respectively, which ties in with the melting point analysis for palmitic acid and also explains the anecdotally operator-reported thickening of the sludge in winter months, which on consideration of the compositional data was probably due to solidification of the oleic acid fraction of the sludge.

Ca is known to accumulate in cutting fluid circuits because it is a closed loop where water is topped up to replace water lost to evaporation, which naturally leads to concentration of Ca and other ions. Ruling out Ca soaps as a major component of the VOCF sludge precludes control of Ca concentrations in the effluent, which was one of the early proposed strategies considered in preventing the VOCF sludge formation.

Understanding the routes of formation is an important step in being able to halt the formation of the floating cutting fluid–based sludge. The two main oils used at the metal-machining factory are trimethylolpropane trioleate and rapeseed oil. Trimethylolpropane trioleate is a triglyceride containing three oleic acids and trimethylolpropane as the base. The fatty acids in rapeseed oil vary, but on average are composed of 60% oleic acid, 4% palmitic acid, 20% linoleic acid, and 10% alpha-linoleic acid (Gunstone 2009). It is suspected that the trimethylolpropane trioleate and rapeseed oil are being microbiologically degraded to free-floating oleic acid and palmitic acid. This route is mostly biologically achieved, however low and high pH can also lead to the formation of free fatty acids. Conditions within the factory VOCF coolant circuits are constantly monitored, with the pH being kept around 9 and the emulsion dosed with biocide to slow biodegradation of the cutting fluids. However, during prolonged storage in the ETP storage tanks the biocide itself will degrade or deplete, thus allowing the microorganisms to become active and biodegrade the oils.

The first stage in the degradation is bacterial lipase-promoted hydrolysis of the triglyceride (microbial rancidity) (VIU 2010). Evidence for this mechanism occurring is the qualitative detection of palmitic acid esters and decyl oleate within Sludge 2, which were not detected in the virgin VOCFs. During the triglyceride degradation mechanism other biodegradation processes can take place, including that of the degradation of the surfactant stabilizing the emulsion, triethanolamine (West and Gonsior 1996; Speranza et al. 2006). This degradation causes a decrease in the concentration of surfactant in the solution and decrease in pH. Degradation of surfactant and lowering of pH reduces the stability of micelles, thus allowing the oil and its components to separate from the solution. Because the free fatty acids densities are less than that of water, they float where they agglomerate into the mass of sludge seen in the ETP storage tanks, a phenomenon known as creaming. The glycerol or trimethylolpropane is miscible in water and remains dissolved and is processed through the ETP, which is why it is not detected within the holding tank sludge. The final stage of the sludge formation is via the generation of palmitic acid through breakdown of the oleic, linoleic, and alpha-linoleic acids. It was determined that the concentration of palmitic acid found within the sludge could not solely come from the fatty acid separation of the rapeseed ester, which contains on average only 4% palmitic acid (Gunstone 2009), therefore it must come from the degradation of the unsaturated fatty acids. The first stage in the process of fatty acid degradation starts by the breaking of double bonds, which may be why there is no linoleic or alpha-linoleic acid found within the sludge (although alpha-linoleic, linoleic, and oleic acid have almost identical retention times on the GC-MS columns). Once all of the double bonds have been (bio)hydrogenated, then no further biohydrogenation can occur, which leaves stearic acid. Stearic acid, also known as octadecanoic acid, is then broken down to palmitic acid. Myristic acid, which is the next breakdown product of palmitic acid, was detected during the qualitative analyses of the sludge.

The buildup of the floating sludge on the surface of the water and the subsequent exclusion of water from its physical structure results in stabilization of the sludge because the microbial lipases can no longer interact with the fatty acids and degrade them further (Gurr et al. 2002); this has important implications because in the sludge state rapid biodegradation dramatically slows, which is counter to the intended full biodegradability of the VOCF products. Limitation of biodegradation due to sludge flotation was also noted by Hwu et al. (1998); other factors that may be important in preventing rapid breakdown of the sludge include lack of nutrients and high oleic acid concentrations (Pereira et al. 2002).