Oil characterization
Prior to usage the oils were characterized using FT-IR and NMR (1H and 13C) techniques. The FT-IR (Fig. 1) display a similar overall profile across all the oil samples. All the samples presented bands νC-H sp2 at 3002 cm− 1, ν C-H sp3 between 2800 and 3000 cm− 1 from fatty unsaturated chains. The νC = O band at 1743 cm− 1 is assigned to ester carbonyl groups with different degrees of symmetry, higher on SO and lower for RO. Higher symmetry for νC = O in oil samples indicates lesser carbonyl signal overlay, and therefore, lesser acidity or lesser contamination with different esters. Below 1500 cm− 1 is the fingerprint region with overlap of different signals, slightly different for every sample. The RO FT-IR spectra displays a different band at 3410 cm− 1 attributed to νO-H, while the carbonyl signal was less symmetrical, thus indicating the expected presence of ricinoleic fatty chains, and also the presence of free fatty acids.
To corroborate the structural assumptions made using FT-IR data, the 1H and 13C NMR spectra were acquired (S-Fig.s 1–8). The NMR spectra display a similar overall profile across all the oil samples. Comparing 1H NMR spectrums, all samples presented unsaturation signals at δ 5.3 ppm, internal glycerol hydrogens at δ 5.2 ppm, external glycerol hydrogens between δ 4.0 and 4.4 ppm, followed by a typical fatty compound profile. All spectrums were normalized using the external glycerol peaks which allowed the comparison of unsaturation hydrogen integrals. An increasing order of unsaturation hydrogen amount was verified, with SO < WCO < RO < SPO, revealing that our refined SO contained less unsaturations than the collected waste cooking oil. The 13C NMR spectrums allowed to identify greater unsaturation complexity between 135 − 125 ppm with the same unsaturation hydrogen amount order, as expected based on known fatty compositions, accompanying a typical fatty compound profile.
The extracted SPO, is an unsaturated rich fatty substrate as seen on FT-IR and 1H and 13C NMR spectra. SPO is comprised mainly by linolenic acid chains (C18:3) with 38.18%, followed by oleic acid chains (C18:1) and palmitic acid chains (C16:0). [28] This fatty composition of SPO, polyunsaturated, is very distinct from fatty subtracts used previously, such as Caryocar brasiliense (pequi), Magonia pubescens (tingui) and sewage sludge, [51–53] which were used for fuel. Based on the results, the SPO, given the oil chemistry possibilities, was functionalized into polyol, and used on the composition of higher value eco-friendly greases looking for better possibility applications according to circular economy concepts.
The extracted SPO was epoxidized using a simple homogeneous peracid, resulting on the formation of oxiranic rings and an epoxidized material (Silkwormm Pupae Epoxide - SPE). The generated SPE was simultaneously hydrolyzed by a similarly simple acid hydrolysis methodology followed by washing. The resulting solid material, was hydroxyl rich and called Silkworm Pupae Polyol (SPP), as represented in Fig. 2.
The comparison between SPO and SPP FT-IR spectra (Fig. 3) allow to visualize the unsaturation consumption by the disappearance of band νC-H sp2 at 3002 cm− 1, the polyol formation, by the increase in νO-H band at 3410 cm− 1 and the symmetry loss in νC = O band due to superposition between ester and carboxylic signals. The 1H NMR spectra superposition of SPO and SPP (Fig. 4) shows the consumption of the SPO’s unsaturation peak at δ 5.3 ppm, the decrease in intensity of internal (δ 5.2 ppm) and external (δ 4.0–4.4 ppm) glycerol hydrogens due to hydrolysis and washing and also the disappearance of unsaturation vicinal hydrogens at δ 2.7 and δ 1.9 ppm. On the SPP it is also possible to identify superposition of labile hydroxyl groups at δ 4.91 ppm and hydroxyl vicinal hydrogens uncoupled at δ 3.29 ppm.
The 13C NMR spectrums of SPO and SPP (S-Figs. 8 and 10) support the mentioned hydrolysis by the presence of carboxyl acid peaks in δ 177 and 178 ppm the in SPP spectra, the unsaturation consumption by the disappearance of signals between δ 135 − 125 ppm and the formation of signals between δ 60–75 ppm attributed mostly to carbons linked to hydroxyl groups.
Grease rheological and thixotropic characterization
The material rheological characterization for greases is mainly comprised of two complementary experiments, the amplitude and the frequency sweeps. On the amplitude sweep experiment, samples sustain shear strain sweeps and have its G’ and G” recorded, allowing the identification of elastic (G’ > G”) and viscous (G” > G’) behaviors, yield points (τr), flow points (τf) and the identification of viscoelastic limit region (LVE).
The τr value represent the point after which the grease begins to flow as a lubricant oil. [69] The τf is the intersect point of G’ and G” curves, representing the point in which the grease the viscous component is equal to the elastic component and therefore, the point beyond which the grease is more viscous than elastic. [70] The obtained data for τr and τf were compiled into Table 3. The LVE value of 0.01% for all the prepared greases were obtained graphically (Fig. 6) from amplitude sweeps and considering a deviation of 3% from starting values.
The obtained amplitude sweep curves (Fig. 6) allow the identification of a typical commercial grease viscoelastic behavior, whereas G’ > G” under lower shear strains and G” > G’ under higher shear strains. [36] Comparing SPP/ROG and WCOG starting values of G’, directly related to how solid the material is under resting conditions, demonstrates that overall Mb results in stiffer greases. Among SSP/ROG and WCOG, Mb greases resulted in higher values of τr and τf, The reason is both the higher hydroxide content and the stronger mechanical stirring, resulting in a steady emulsion.
The ion substitution from Li to Na, using Ma, in SPP/ROG results in decreased τr (from 6.225 to 3.41 Pa) and τf (from 457.35 to 70.31 Pa) values, while the Li to Na substitution on greases prepared with Mb presented almost constant τr value (from 27.85 to 27.95 Pa) with decreased τf (from 810.95 to 565.7 Pa). Comparison between SPP/ROG(Li)-Mb and ROG(Li)-Mb allow to investigate the polyol addition effect. The polyol addition resulted in a constant τr value (from 27.35 to 27.85 Pa), while greatly increasing τf (from 259.2 to 810.95 Pa).
The ROG relation between Mb and Mb’ allows to investigate the hydroxide proportion effect on commercial ricin oil. Comparing ROG(Li)-Mb with ROG(Li)-Mb’, the decrease in hydroxide proportion also decreased the τr (from 27.35 to 0.75 Pa) and τf (from 259.2 to 27.35 Pa) values. The ROG Mb’ substitution of Li for Na resulted in increase on τr.(from 0.75 to 49 Pa) and τf (from 27.35 to 2489 Pa), an indicative of possible ion size effect on the structure of ROG.
The ion substitution from Li to Na, using Ma, in WCOG results in slight decrease in τr (from 2.95 to 2.02 Pa), with three times increase in τf (from 32.06 to 96.90 Pa). The Li to Na ion substitution on WCOG, greases prepared with Mb, showed a strong increase in τr (from 15.7 to 56.9 Pa) with almost double increase in τf (from 395.15 to 645.4 Pa). Comparison between SOG(Na)-Ma and WCOG(Na)-Ma allow to investigate the usage effect (acidity increase and saponification value decrease). The usage effect resulted in slight increase in τr value (from 1.19 to 2.02 Pa), with three times increase in τf (from 38.44 to 96.90 Pa).
The SOG(Na)Ma and SOG(Na)-Ma’ comparison allow to investigate the hydroxide proportion effect on refined soy oil. The increase in hydroxide proportion, SOG(Na)-Ma to SOG(Na)-Ma’, results in slightly decreased τr (from 1.19 to 0.94 Pa) with three times increase in τf (from 38.44 to 106.25 Pa). The change of ion from Na to Li, SOG(Na)-Ma’ and SOG(Li)-Ma’ comparison, results in τr increase (from 0.94 to 53.4 Pa) and doubling in τf (from 106.25 to 233.05 Pa) values.
Saxena and coworkers described a preparation of eco-friendly nano-greases based on soybean oil and organoclay as thickener and CaCO3 nano-powder as additive. The obtained values for τr ranged between 497–992 Pa, while τf ranged between 118.2-292.7 Pa, at 25°C, with a commercial all-purpose presented τr and τf of 1100 Pa and 289.1 Pa, respectively. [36] Padgurskas et al. described biogreases based on rapeseed oil and beeswax as thickener with calculated yield points ranging between 47–301 Pa at 20°C. [71] Roman et al. described ricin oil based and cellulose nanofiber thickened greases, with τr of 300 Pa at 25°C. [72] While classical aluminum, lithium, calcium and polyuria thickened greases displays τr values between 500 and 1000 Pa. [73] Our prepared materials is therefore comparable with other green eco-friendly greases examples, but with lower yield point values if compared with classical greases. But, they show a lot of benefits to environmental benefits, as waste disposal reduced, renewable materials, circular economy and green chemistry concepts.
On the frequency sweep experiment (Fig. 7), samples sustain angular frequency sweeps within the LVE region (constant 0.01% shear strain), while it’s G’ and G” are recorded, allowing to comprehend the grease’s time dependency behavior. Typical commercial greases display G’ higher than G” which run parallel at lower frequencies and slight increase in G’ and G” at higher frequencies with increase rate slightly higher for G”. [36, 74] The higher rate approach between G’ and G” is an indicative of structural breakdown. [69]
Among SSP/ROG samples, Mb greases showed higher values for G’ and G” within the 10− 2-102 rad s− 1 range. The samples SSP/ROG(Li)-Mb and SSP/ROG(Na)-Ma displayed commercial like behavior, while SSP/ROG(Li)-Ma and SSP/ROG(Na)-Mb were deviant. The SSP/ROG(Li)-Ma and SSP/ROG(Na)-Mb G’ and G” increased at higher and lower frequencies, creating a concavity like profile for G’ and G” values. The loss tangent graph shows greater stability for SSP/ROG(Na)-Ma, with considerable linearity for SSP/ROG(Li)-Mb and SSP/ROG(Na)-Ma. The SSP/ROG(Li)-Ma displayed greater loss tangent, with considerable instability at lower frequencies.
The ROG greases displayed commercial like behavior. The comparison between SSP/ROG(Li)-Mb and ROG(Li)-Mb G’ and G” curves shows very similar profiles. The decrease in hydroxide content (Mb to Mb’) results in lower G’ and G” curves and instability below 10 rad s− 1 as displayed by the ROG(Li)-Mb’ loss tangent. The ion change in ROG, from Li to Na, results in greater stability as shown by the ROG(Na)-Mb’.
Among WCOG samples, Mb greases (WCOG(Li)-Mb and WCOG(Na)-Mb) showed higher values for G’ and G”, within the 10− 2-102 rad s− 1 range, and also greater stability as shown by the loss tangent. The Ma greases (WCOG(Li)-Ma and WCOG(Na)-Ma) displayed lower stability at lower frequencies. The SOG greases displayed commercial like behavior. The comparison between WCOG(Na)-Ma and SOG(Na)-Ma G’ and G” curves shows very similar profiles, with slightly higher values for the WCOG, explainable by it’s higher acidity and therefore more effective saponification. The increase in hydroxide content (Ma to Ma’) results in similar G’ and G” curve profiles. The ion change in SOG, from Na to Li, results in greater stability as shown by the SOG(Li)-Ma’ loss tangent curve despite instabilities in middle range frequencies.
The thixotropic experiments for all the samples displayed profiles consistent with expected commercial greases (Fig. 8), which were treated for thixotropic hysteresis area (THA), also compiled in Table 3.
The thixotropic behavior represents the structural recovery within increase and decrease of shear rates and time. [75] To identify thixotropic behavior, a simple hysteresis loop technique can be used resulting in the quantifiable value of thixotropic hysteresis area (THA). [76] High THA values represents greater structural time dependency, while low THA values represent how easily a material recovers its structure post-shear. [77] The THA values are, therefore, the kinetic difference between network breakup (faster) and network buildup (slower). [75]
The Li based SSP/ROG samples displayed higher THA values indicating greater time dependency. The comparison between ROG(Li)-Mb and SSP/ROG(Li)-Mb reveal increased time dependency on occasion of polyol addition (from 2.70 104 to 9.16 104 Pa s− 1). The ROG(Li)-Mb and ROG(Li)-Mb’ correlation shows the hydroxide content effect on THA, diminishing hydroxide content results in lesser time dependency (from 2.70 104 to -1.05 104 Pa s− 1). The negative signal on THA for ROG(LI)-Mb indicates that not only the material displayed fast structural recovery, but also achieved greater organization post shear. The ion change from Li to Na in ROG results in great time dependency (from − 1.05 104 to 1.29 105 Pa s− 1), another indicative of possible ion size effect on the structure of ROG.
Table 3
Rheological and thixotropic relevant values obtained for the studied greases.
Grease
|
Yield Point, τr (Pa)*
|
Thixotropic Hysteresis Area (Pa s− 1)
|
Flow Point, τf@G’=G” (Pa)
|
SPP/ROG(Li)-Ma
|
6.225
|
3.83 104
|
457.35
|
SPP/ROG(Li)-Mb
|
27.85
|
9.16 104
|
810.95
|
SPP/ROG(Na)-Ma
|
3.41
|
2.32 104
|
70.31
|
SPP/ROG(Na)-Mb
|
27.95
|
1.85 104
|
565.7
|
WCOG(Li)-Ma
|
2.95
|
-1.89 104
|
32.06
|
WCOG(Li)-Mb
|
15.7
|
4.61 104
|
395.15
|
WCOG(Na)-Ma
|
2.02
|
2.18 103
|
96.90
|
WCOG(Na)-Mb
|
56.9
|
1.59 104
|
645.4
|
SOG(Na)-Ma
|
1.19
|
1.55 103
|
38.44
|
SOG(Na)-Ma’
|
0.94
|
-9.44 101
|
106.25
|
SOG(Li)-Ma’
|
53.4
|
3.55 104
|
233.05
|
ROG(Li)-Mb
|
27.35
|
2.70 104
|
259.2
|
ROG(Li)-Mb’
|
0.75
|
-1.05 103
|
27.35
|
ROG(Na)-Mb’
|
49
|
1.29 105
|
2489
|
* Parameters directly provided by Anton Paar MCR 72 software.
The WCOG samples prepared on Mb displayed greater THA. On WCOG the change from Li to Na, using Ma, resulted in great time dependency (from − 1.89 104 to 2.18 103 Pa s− 1), while using Mb resulting in decreased time dependency (from 4.61 104 to 1.59 104 Pa s− 1). The comparison between SOG(Na)-Ma and WCOG(Na)-Ma reveal a slight increased time dependency with usage effect (from 1.55 103 to 2.18 103 Pa s− 1). The SOG(Na)-Ma and SOG(Na)-Ma’ relation shows the relation between hydroxide content and THA, increasing hydroxide content results in lower time dependency (from 1.55 103 to -9.44 101 Pa s− 1). The ion change from Na to Li for SOG results in greater THA (from − 9.44 101 to 3.55 104 Pa s− 1).
Saxena et al. described THA values ranging between 9.91-24 103 Pa s− 1 for soybean organoclay greases [36] Ren et al. described the preparation of synthetic oil (polyalphaolefin - PAO 40) greases with lithium complex thickener with THA values ranging between 2–6 104 Pa s− 1. [78] Our prepared greases are, therefore, within range with both eco-friendly greases and classical greases.
A direct application for our residual oil prepared greases depends on desired lubrication properties which are intrinsic to each system. For instance, on SSP/ROG, greases displayed low and high τr and τf values, some presented commercial like frequency sweep profiles, but only slow post-shear recovery (high values of THA). Based on the high values of τr and τf Mb greases would be better for more intensive applications, to be confirmed by tribological tests. Based on flow point data, SSP/ROG(Na)-Ma seems to be in range with microcristanile and paraffin waxes. [79] On the other hand, WCOG also displayed low and high τr and τf values, presented almost commercial like frequency sweep profiles and also varying THA values, with applications well discussed. [24]
In this study WCOG despite been well known was prepared and analyzed similarly to SPP/ROG aiming to access possible compatibilities between WCO and SPP. Our focus now on it to study the polyol effect on WCO compositions, linking two different residues into a common biorefinery application on an industrial scale, effectively aligning with UN SDG’s, while advancing the TRL (Technology Readiness Level), as summarized in Fig. 9.
Nevertheless, despite using each residual oil separately, the residual oils biorefinery proposed herein enhances the overall value of the applied residual oils. The grease compositions were prepared by partial saponification, removing unitary operation and simplifying the preparation process (relative to classical grease preparation). Our prepared greases uses no further modified polyol bases, is considerable eco-friendly based on the main renewable components, while also describing reusage/ion effect on grease composition. The prepared grease properties are also on par with other eco-friendly greases and therefore present feasible commercial application.