Tribology and Lubrication Technology December 2012 : Page 66
Peer-reviewed New Method of Measuring Permanent Viscosity Loss of Polymer-Containing Lubricants JENNIFER HOLTZINGER, 1 JONATHAN GREEN, 2 GORDON LAMB, 2 DAVID ATKINSON, 2 and HUGH SPIKES 1 1 Mechanical Engineering, Tribology Group Imperial College, London, UK 2 BP Castrol Ltd., Pangbourne, UK Manuscript received February 3, 2012 Manuscript accepted April 22, 2012 Review led by Paul Bessette Copyright© STLE KEY WORDS Viscosity modifier; Shear stability; High shear rate; Permanent shear stability ABSTRACT An ultrahigh shear rate viscometer (USV) was used to measure the viscosity of polymer so-lutions. It was found that some polymer solutions in base oil, including those used as engine oil viscosity modifiers, show permanent viscosity loss when subjected to very high shear rates above 10 6 s −1 . The USV was modified to automatically carry out a series of viscosity measurements on the same test lubricant sample. This enabled the accumulation of perma-nent viscosity loss to be measured over successive strain cycles. As expected, permanent viscosity loss increased with both strain rate and molecular weight. When carried out at 5 × 10 6 s −1 and 100°C, the test was more severe than the Kurt Orbahn test because samples of lubricants subjected to the latter underwent further shear thinning in the USV. The USV test appears to be a rapid and convenient way to quantify the permanent viscosity loss of polymer-containing lubricants for engine use, and a protocol to assess per-manent viscosity loss (PVL) and permanent shear stability index (PSSI) based on viscosity measurements at 10 6 s −1 before and after shear thinning is outlined. The study also shows that it is important to take into account possible permanent viscos-ity loss when measuring the viscosity of polymer solutions in very high shear rate viscom-eters such as the USV. This can be done by minimizing the amount of shear to which the lubricant is subjected or by taking successive measurements and subtracting the permanent viscosity loss taking place in each of the first few strain rate cycles. Editor’s Note: Measurement of viscosity is critical for proper determination of a number of properties of a lubricating fluid, however, rarely does the act of measuring impact the result directly. Such is the case when using an Ultrahigh Shear Rate Viscometer. This instrument is able to induce permanent shear loss in the measured sample. While some may view this as an inconvenience when attempting multiple measure-ments, the authors of this month’s Editor’s Choice paper chose to advance this concept into a working model for measuring permanent viscosity loss in polymer-containing lubricants. The concept proves to be much faster than the Kurt Orbahn Diesel Injector Test, which is certainly a boon for researchers. Evan Zabawski, CLS Editor INTRODUCTION Many liquid lubricants, including almost all crankcase engine oils and many trans-mission lubricants, contain dissolved polymer additives. These polymers, commonly known as viscosity modifiers (VMs) or viscosity index improvers (VIIs), enhance the vis-cosity index of their blends and thus enable the formulation of multigrade oils. It is well known that these polymer solutions can exhibit both temporary and per-manent viscosity loss under the very high shear rate conditions present in the lubri-cated contacts of machine components such as pistons and bearings. Once regarded as undesirable, temporary—that is, reversible—viscosity loss due to shear thinning is now recognized to make a valuable contribution to reducing hydrodynamic friction in engines. However, permanent viscosity loss is still generally considered undesirable because it can result in a lubricant falling below its designated viscosity grade during use. This constrains the molecular weight (Mw) of linear polymers used in engine 66 “Grease Compatibility,” STLE Webinar with Heinrich Braun, ExxonMobil Lubricants
Peer Reviewed Paper
New Method of Measuring Permanent Viscosity Loss of Polymer-Containing Lubricants<br /> <br /> KEY WORDS<br /> Viscosity modifier; Shear stability; High shear rate; Permanent shear stability<br /> <br /> ABSTRACT<br /> An ultrahigh shear rate viscometer (USV) was used to measure the viscosity of polymer solutions. It was found that some polymer solutions in base oil, including those used as engine oil viscosity modifiers, show permanent viscosity loss when subjected to very high shear rates above 106 s-1. The USV was modified to automatically carry out a series of viscosity measurements on the same test lubricant sample. This enabled the accumulation of permanent viscosity loss to be measured over successive strain cycles.<br /> <br /> As expected, permanent viscosity loss increased with both strain rate and molecular weight. When carried out at 5 × 106 s-1 and 100°C, the test was more severe than the Kurt Orbahn test because samples of lubricants subjected to the latter underwent further shear thinning in the USV.<br /> <br /> The USV test appears to be a rapid and convenient way to quantify the permanent viscosity loss of polymer-containing lubricants for engine use, and a protocol to assess permanent viscosity loss (PVL) and permanent shear stability index (PSSI) based on viscosity measurements at 106 s-1 before and after shear thinning is outlined.<br /> <br /> The study also shows that it is important to take into account possible permanent viscosity loss when measuring the viscosity of polymer solutions in very high shear rate viscometers such as the USV. This can be done by minimizing the amount of shear to which the lubricant is subjected or by taking successive measurements and subtracting the permanent viscosity loss taking place in each of the first few strain rate cycles.<br /> <br /> INTRODUCTION<br /> Many liquid lubricants, including almost all crankcase engine oils and many transmission lubricants, contain dissolved polymer additives. These polymers, commonly known as viscosity modifiers (VMs) or viscosity index improvers (VIIs), enhance the viscosity index of their blends and thus enable the formulation of multigrade oils.<br /> <br /> It is well known that these polymer solutions can exhibit both temporary and permanent viscosity loss under the very high shear rate conditions present in the lubricated contacts of machine components such as pistons and bearings. Once regarded as undesirable, temporary—that is, reversible—viscosity loss due to shear thinning is now recognized to make a valuable contribution to reducing hydrodynamic friction in engines. However, permanent viscosity loss is still generally considered undesirable because it can result in a lubricant falling below its designated viscosity grade during use. This constrains the molecular weight (Mw) of linear polymers used in engine oils to typically less than 300,000. Permanent viscosity loss depends on a number of design factors, including the type (Mortier1), architecture (Covitch2), and molecular weight (Bartz3) of the polymer and polarity of the base oil (Zakin and Hunston4) and other additives present (Devlin, et al.5). It is thus important to be able to measure and explore the extent of permanent shear thinning of polymer containing lubricants under high shear rate conditions.<br /> <br /> This article describes a new method of measuring the permanent shear thinning behavior of polymer-containing lubricants using the recently developed ultrahigh shear rate viscometer (USV; PCS Instruments6). One advantage of the method is that permanent shear thinning occurs under quite well-defined shear rate and shear stress conditions, so the impact of the latter on polymer degradation can be explored quantitatively. A second is that the method is rapid and employs a very small volume of test fluid, which makes it well suited to lubricant development. The findings of this study also have implications for the use of the USV to measure temporary shear thinning of engine oils.<br /> <br /> BACKGROUND<br /> Polymer additives were first employed to increase the viscosity index (VI) of engine and hydraulic oils in 1936 (Boehm7), although the concept of multigrade engine oils and the key role of VIIs in developing these only emerged in the early 1950s (Larson8). Initially, very high-molecular-weight polymers were favored as VIIs because these had the greatest impact on the viscosity index. However, it was soon recognized that such polymers suffered unacceptable levels of permanent viscosity loss during use, so the molecular weights were reduced.<br /> <br /> In practice, polymer solutions experience two types of viscosity loss in the high shear rate conditions present in lubricated contacts. At shear rates above about 104 s-1 for engine oil VIIs, temporary shear thinning (temporary viscosity loss) occurs. This is believed to originate from partial alignment of polymer molecules in solution from their normal, random coil configuration (Selby9). This has the effect of reducing the extent to which the polymer molecules impede the flow of oil as well as the level of entanglement of the polymer molecules. Both phenomena reduce the dynamic viscosity. This shear thinning is fully recovered when the strain is removed and the polymer molecules revert to a random configuration.<br /> <br /> At strain rates above approximately 106 s-1 for engine oil VIIs, permanent viscosity loss takes place, which is not recoverable when the strain is removed. This originates from scission of the polymer molecules into shorter fragments, due primarily to mechanical forces exerted on the polymer by the fluid. Considerable research has been conducted on the degradation of polymers under mechanical stress (Casale, et al.10; Casale and Porter11; Sohma12). Polymer degradation occurs in part from the stresses imposed by the solvent on the partially elongated molecular chains, so that scission tends to occur in the middle of the chains, reducing the molecular weight by approximately half. However, at high polymer concentrations, entanglement of the chains themselves can also impart stresses and lead to chain breakage (Beuche13). These mechanisms of chain scission and thus permanent viscosity loss imply, as is found in practice, that the length of the polymer backbone is a more important determinant of polymer degradation than the molecular weight. For this reason, comb VIIs and also multi-branched polymer architectures such as stars and hyperbranched dendrites, generally show lower permanent viscosity loss than linear polymers of the same molecular weight (Covitch14; Suzuki, et al.15; Wang, et al.16).<br /> <br /> Figure 1 schematically illustrates the processes of temporary and permanent viscosity loss. For a monodisperse polymer, permanent viscosity loss should occur at a higher shear rate than temporary viscosity loss. However, the polydisperse nature of most VIIs means that considerable overlap can occur.<br /> <br /> Figure 2 shows schematic flow curves (i.e., plots of measured viscosity versus the logarithm of shear rate) for a VIIcontaining lubricant when freshly blended and also after mechanical shear and consequent permanent viscosity loss. The fresh oil shows first and second Newtonian regions separated by a transition region of progressively increased temporary shear thinning. The used oil illustrates the impact of permanent viscosity loss. The viscosity is reduced over the whole shear rate and, in the case of high permanent viscosity loss, the transition region is displaced to a higher shear rate as the polymer molecular weight is reduced.<br /> <br /> Although the flow curves in Figure 2, as well as the main equations used to describe polymer shear thinning, are expressed in terms of viscosity versus shear rate, the main determinant of both temporary and permanent shear thinning in lubricated contacts is generally believed to be the shear stress—that is, the shear rate multiplied by the prevailing dynamic viscosity—because this will determine the main mechanical forces experienced by the polymer molecules (Ram and Kadim17). For this reason, VII permanent viscosity loss tends to be greater in higher viscosity base oils (Horowitz18) and much more severe in high-pressure elastohydrodynamic contacts, such as gears and rolling bearings, than in relatively low-pressure conformal hydrodynamic contacts such as plain bearings (Walker, et al.19).<br /> <br /> Permanent viscosity loss of polymer-containing oils is most commonly expressed in two ways. One is the permanent viscosity loss (PVL), which is the percentage of permanent reduction in dynamic viscosity; that is, where KVFresh is the kinematic viscosity of the unsheared blend and KVSheared is the kinematic viscosity of the sheared lubricant, both measured at low shear rate and at 100°C. Also widely used is the permanent shear stability index (PSSI), defined as the percentage of the thickening effect of the polymer that is permanently lost due to shear; that is, where KVSolvent is the kinematic viscosity at 100°C of the lubricant in the absence of polymer. One reason for preferring the use of PSSI to PVL is that it allows better comparison of the shear thinning behavior of a polymer in different viscosity oils.<br /> <br /> Several different methods have been developed to measure the permanent viscosity loss of polymer-containing lubricants (Mortier1; Alexander, et al.20). For engine oils, where the main concern is viscosity loss in low-pressure contacts such as plain bearings, the most widely used method is the Kurt Orbahn test (ASTM International21,22). In this, the test lubricant is circulated a set number of times (usually 30 or 90) through a diesel injector nozzle at a temperature of 30– 35°C. Kinematic viscosity at 100°C of the test fluid is measured at the beginning and end of the test. This test has been used to study the influence of polymer type and molecular weight as well as solvent properties on permanent viscosity loss (Devlin, et al.5; Bezot and Hese- Bezot23).<br /> <br /> One limitation of the Kurt Orbahn and most other existing methods for measuring permanent viscosity loss of lubricants is that the shearing conditions are complex and quite poorly defined. It has been estimated that the representative shear rate in the Kurt Orbahn test is about 106 s-1, but in practice the test fluid experiences a wide range of different shear rates in its passage through a nozzle. It has also been suggested that polymer viscosity loss in hydrodynamic shear is not the same as in cavitation (Crouse and Wilkins24). Both of these mechanical processes occur in the Kurt Orbahn.<br /> <br /> In fundamental research on the mechanical degradation of polymers, well-defined shear systems are often employed, such as capillary or rotating cylinder devices, and this enables the impact of shear stress on degradation to be analyzed scientifically (Porter, et al.25). This is possible because such studies have focused on very high-molecular-weight polymers that shear thin at relatively low shear rates, attainable in conventional viscometers. In contrast, polymer additives used in lubricants are designed to be shear stable and to not permanently degrade at easily attainable shear rates. Thus, permanent viscosity loss of lubricants is measured in devices such as nozzles, bearings, and gears where high shear stresses can be achieved but at the expense of a poorly defined stress field.<br /> <br /> Quite recently, a rotating cylinder viscometer was developed that is able to achieve higher shear rates than previously possible in large shear devices, up to 107 s-1. Although designed primarily for measuring the extent of temporary shear thinning of lubricants, this article will show that it can also be used to quantify the permanent viscosity loss of polymercontaining lubricants.<br /> <br /> TEST METHOD<br /> The equipment used in this study is a USV from PCS Instruments6. This is a rotating cylinder viscometer with a gap of approximately 1 µm, designed to measure dynamic viscosities of liquids over the shear rate range 106 to 107 s-1. At such high shear rates, the rate of energy dissipation within the sheared fluid is faster than the rate at which it can be conducted through the film, so the fluid temperature increases rapidly. This generally precludes useful viscosity measurement. To overcome this problem, the USV employs a very short duration of shear. An external rotor is spun up to the required speed and then coupled to the internal rotor of the viscometer using an electromagnetic clutch. The speed of this internal rotor stabilizes after a few milliseconds and then torque measurements, from which viscosity can be calculated, are captured over a period of less than 40 ms. They are thus obtained before a significant amount of temperature rise occurs due to shear. After the measurement, the internal rotor is disengaged and it returns to stationary. Normally the above procedure is repeated three times in rapid succession and the viscosity is calculated from the average torque.<br /> <br /> The USV was developed primarily to measure the temporary shear thinning behavior of polymer-containing lubricants. However, in a recent study by the authors it was noted that when two successive measurements were made on a single polymer blend sample at the same shear rate the second measurement was slightly lower than the first, indicating some permanent viscosity loss. The control software of the USV was therefore modified to enable a succession of measurements to be taken at a set shear rate, with a user-controlled rest time between each measurement.<br /> <br /> As will be indicated in the Results section, this enables progressive shear thinning to be monitored. In the current study, a series of 30 sets of viscosity measurements per test was selected for most of the work, with a 5-min pause between each measurement (30 cycles). Five minutes was chosen to ensure both full recovery of the polymer solution to a zero-shear molecular conformation and to make sure that, in the case of any temperature increase due to shear, the fluid returned to its set test temperature. It is likely that a shorter time would be satisfactory if a greater test turnaround time were required. In terms of the amount of shear applied, within one measurement cycle the rotor accelerated from rest to the set rotational speed within one rotation of the rotor and friction force was measured and averaged over eight rotations (corresponding to a shearing distance of 302 mm). Deceleration from 107 s-1 to 10% of the set speed occurred over a sliding distance of ca. 300 mm. This procedure was repeated three times in rapid succession.<br /> <br /> TEST MATERIALS<br /> Two commercial VM polymers were investigated. Both were hydrogenated polystyrene isoprene polymers, one linear with Mw 150,000 (HPIL) and one star with Mw 127,000 (HPIS). These were used in solution in an API Group III mineral oil as listed in Table 1.<br /> <br /> Two series of model polymers were studied, polystyrene (PS) and polyisoprene (PIP), with molecular weights listed in Table 4. These are commercially available polymers that are normally used for calibration purposes in gel permeation chromatography and have a narrow molecular weight distribution. The polyisoprene was nonhydrogenated and composed of 80% of the cis-1,4, 15% of the trans-1,4, and 5% of the 3,4 configuration.<br /> <br /> Polystyrene is not soluble in saturated hydrocarbon oils, so both sets of polymers were dissolved in the ester base fluid di-2- ethylhexyl phthalate (DOP) to produce the blends described in Tables 2 and 3. The polymer concentrations used were chosen to provide blends with similar dynamic viscosities of ca. 7 mPa.s at 1 × 107 s-1 and 100°C.<br /> <br /> In Tables 1–3, ?o is the low shear rate dynamic viscosity.<br /> <br /> RESULTS<br /> Commercial Polymer Solutions<br /> Figure 3 shows viscosity measurements over 30 cycles at 5 × 106 s-1 for the two 1 wt% commercial polymer solutions and the polymer-free Group III base oil. Each of the data points shown is an average from three series of measurements on different samples of the same test fluid and in most cases the span of the three values is smaller than the size of the data point shown. The base oil viscosity remains constant, but the viscosity of the two polymer solutions falls progressively, indicating permanent viscosity loss, before leveling out after about 15 cycles. Figures 4 and 5 show the impact of shear rate on this response for the two solutions. It can be seen that increased shear rate led to both a reduction in the initial measured viscosity and a greater subsequent viscosity loss.<br /> <br /> It should be noted when comparing these results at different shear rates that the USV was configured to subject the fluid to shear and measure average viscosity over a set shear distance rather than for a set duration. This means that during a test at 1 × 106 s-1 the fluid sample was actually sheared five times as long as during a test at 5 × 106 s-1, although the amount of shear was the same for the two tests. It is arguable which is more appropriate when comparing permanent viscosity loss, the shear distance or the shear time. The USV software could be changed to maintain a fixed shear time if this were preferred, though this was not done in the current study. If a set shear time had been used, then the rate of accumulation of viscosity loss would have been increased at high shear rate relative to low shear rate, but the value of viscosity at which the fluid eventually stabilized would have remained unchanged.<br /> <br /> These viscosity loss results were converted to nondimensional values analogous to the PVL and PSSI terms described in Equations  and . Figures 6 and 7 show plots for 1 wt% HPIL of PVL (HS) and PSSI (HS) versus the number of cycles, where these terms are defined as;<br /> <br /> The viscosities ?Fresh, ?Sheared, and ?Solvent are the dynamic viscosities measured at the prevailing test temperature and shear rate, so the bracket (HS) indicates that PVL and PSSI are based on high shear rate viscosity measurements. The use of the dynamic rather than the kinematic viscosity used in Equations.  and  should have little effect because the density will not change as a result of permanent viscosity loss and will not differ greatly between oil and polymer solution. As can be seen in Figure 6, PVL (HS) is lower at a higher shear rate than at a lower shear rate. This is because ?Fresh decreases with shear rate due to temporary shear thinning faster than (?Fresh - ?Sheared) increases due to permanent viscosity loss over the shear rate range measured. Similar behavior was seen with the star polymer HPIS. This anomaly is further discussed and addressed later in the Discussion section.<br /> <br /> PSSI (HS), as shown in Figure 7, increases with shear rate but provides considerable scatter at high shear rates. This is because the viscosity at a high shear rate approaches the base oil viscosity, so the denominator in Equation  becomes very small. The full measured flow curve for this polymer solution is shown in Figure 8 and it can be seen that the polymer blend viscosity approaches the base oil viscosity quite rapidly around 107 s-1. Figure 9 shows PSSI (HS) plots at 5 × 106 s-1 of all four commercial polymer solutions tested.<br /> <br /> Model Polymers<br /> Figures 10 and 11 show how PSSI (HS) varies with number of cycles at a shear rate of 5 × 106 s-1 for the two sets of model polymers. It should be noted that the two graphs have different scales. The DOP diluent alone showed no permanent viscosity loss.<br /> <br /> It is evident that the permanent viscosity loss increased with molecular weight, as expected. The dependence on the molecular weight of PSSI (HS) after 30 cycles is shown in Figure 12. Polyisoprene showed much greater permanent viscosity loss than polystyrene. This is probably due mainly to the longer chain lengths of the former. The isoprene monomer has a molecular weight of 68 and contributes four carbon atoms to the polymer backbone, whereas styrene, with a pendant benzene ring, has a molecular weight of 104 and contributes only two carbon atoms. Thus, for the same molecular weight, polyisoprene will have a carbon backbone three times as long as that of polystyrene. It is also possible that the C=C bonds of polyisoprene, or the single bonds adjacent to them, are more susceptible to mechanical degradation than single bonds in alkyl chains. A cis conformation about the double bond may be architecturally weaker in a shear field than the more linear trans conformation.<br /> <br /> DISCUSSION<br /> Comparison with Kurt Orbahn Measurements<br /> Tests were also carried out on the commercial polymer solutions using the standard Kurt Orbahn test (ASTM International21). The PVL and PSSI results are summarized in Table 5, where they are compared with the values PVL (HS) and PSSI (HS) measured using the USV after 30 cycles at 100°C.<br /> <br /> The Kurt Orbahn and USV values of PSSI clearly differed. This is not just because the conditions to which the lubricant was subjected were different in the two tests but also because the permanent viscosity loss values were calculated in quite different ways. This is illustrated in Figure 13, which shows schematically the origins of the four terms. In Table 5 the PSSI (HS) values are much greater than the PSSI values because the denominator of the former, (?Fresh - ?Solvent), is very small at high shear rates. The PVL and PVL (HS) values are quite similar for the concentrated polymer solutions.<br /> <br /> Comparison of the high shear rate viscosities of samples tested in the 30-cycle Kurt Orbahn and after 30 cycles in the USV showed that the end-of-test values for the USV were lower than those for the Kurt Orbahn, suggesting that the shearing conditions in the USV were more severe than in the Kurt Orbahn on a cycle-by-cycle basis. In some experiments, the Kurt Orbahn test was first carried out and then a sample of the sheared solution was tested using the USV procedure. Considerable further permanent viscosity loss was observed in the USV, as shown in Figure 14.<br /> <br /> Implications for Measuring Temporary Shear Thinning<br /> The observation that permanent viscosity loss of polymer solutions occurs at high shear rates in the USV has some implications for the use of the USV to measure and study temporary shear thinning. A typical current protocol in the USV is to measure the viscosity of a test sample at a series of shear rates between 106 and 107 s-1. In such a test, permanent viscosity loss may accumulate, to contribute significantly to the measured viscosity in the later measurements. This is illustrated in Figure 15, where the viscosities of the two commercial polymers at high concentration were measured in two ways. In one, a single test sample was measured over a series of 10 increasing shear rate stages. In the second, four test samples were measured individually, each at a different shear rate. It can be seen that at high shear rates, the multiple measurements on the single sample fell progressively further below the single measurements on multiple samples. This was due to accumulating permanent viscosity loss and suggests that single measurements are preferred to multiple measurements on the same sample when studying temporary shear thinning at high shear rates.<br /> <br /> It is not possible to take a large shear viscosity measurement on a solution of a high-molecular-weight polymer at very high shear rates that does not contain a permanent shear-thinning component. However, the short shear duration of the USV does ensure that this is relatively small. Permanent shear thinning during the first measurement can also be accounted for, to some extent, by extrapolation. This is exemplified in Figure 16, where five successive measurements were made on three high-molecular-weight polystyrene polymer solution samples at three shear rates. Each set of measurements can be extrapolated back to estimate zero shear cycle values. In the figure, linear extrapolation was employed but more refined fits are possible. Such an approach clearly assumes the absence of anomalously high viscosity loss in the first cycle.<br /> <br /> With the commercial polymers, some permanent viscosity loss was observed in the USV even at 106 s-1. This suggests that high temperature high shear (HTHS) measurements made using other viscometers at 106 s-1 may contain a permanent viscosity loss component, although this does not appear to have been reported in the literature.<br /> <br /> Advantages and Disadvantages of USV Test Method<br /> The USV method for studying permanent viscosity loss has several advantages over alternative approaches such as the Kurt Orbahn. A key advantage is that it subjects the test fluid to relatively well-defined shear conditions, which means that the impact of different levels of shear stress can be investigated scientifically. The permanent viscosity loss can also be related quite convincingly to conditions experienced in hydrodynamically lubricated machine components.<br /> <br /> The test also allows permanent viscosity loss to be obtained and measured under realistic operating conditions, especially of temperature. Temperature as well as mechanical shear influences the level of permanent viscosity loss and the USV approach enables this to be explored.<br /> <br /> The key limitation of the USV approach is that, because of the very small sample tested, it is not possible to extract the sheared lubricant sample for subsequent analysis or measurement of low shear rate viscosity. This is why it is not possible to measure low shear rate PVL and PSSI values corresponding to those obtained from the Kurt Orbahn test and directly test whether a lubricant has stayed in grade.<br /> <br /> PSSI and PVL Referenced to 106 s-1 One effect of calculating PVL (HS) and PSSI (HS) based on the measured viscosity at high shear rate as shown in Equations  and  is that, as shown in Figure 6, these values may be higher at lower shear rate if ?Fresh decreases with shear rate due to temporary shear thinning faster than (?Fresh - ?Sheared) increases due to permanent shear thinning. This can be resolved by basing PVL and PSSI on measured viscosity at 106 s-1 even when the test sample has been sheared at a higher shear rate, thus defining; where ?Fresh,106, ?Sheared,106, and ?Solvent are the dynamic viscosities measured at the prevailing test temperature and 106 s-1. This is analogous to the conventional PVL and PSSI definitions except that measurements are made at 106 s-1 rather than low shear rate. This approach should always yield values that increase at higher test shear rate. The test protocol is to take an initial viscosity measurement at 106 s-1 (?Fresh,106); shear the sample 30 times at a fixed, higher shear rate; and finally make five measurements at 106 s-1 (?Sheared,106). Figure 17 shows a typical set of viscosity measurements made according to this protocol on 1 wt% HPIS, with the initial and final five measurements made at 106 s-1 and the 30 intermediate ones, which are primarily responsible for the observed permanent shear thinning, made at 5 × 106 s-1. It can be seen that when the final measurements were made at 106 s-1, the viscosity increased slightly over the first few measurements. This was not seen for all conditions and may reflect a progressive loss of disentanglement. The average value of the five final measurements was employed in calculating the PVL (106) and PSSI (106) values listed in Table 6. The base oil viscosity was 3.34 mPa.s at 100°C and 1.61 mPa.s at 150°C. The linear polymer solution showed higher PSSI (106) under most conditions, although the star polymer showed a higher value at a combination of high shear rate and high temperature.<br /> <br /> CONCLUSIONS<br /> It has been found that polymer solutions in base oil can show permanent viscosity loss when subjected to very high shear rates in the USV. The USV was modified to automatically carry out a series of viscosity measurements on the same test lubricant sample. This enables the accumulation of permanent viscosity loss to be measured over successive strain cycles.<br /> <br /> As expected, permanent viscosity loss increased with both strain rate and molecular weight. When carried out at 5 × 106 s-1 and 100°C the test was more severe than the Kurt Orbahn test because samples of lubricants subjected to the latter underwent further shear thinning in the USV.<br /> <br /> The USV test appears to be a rapid and convenient way to quantify the permanent viscosity loss of polymer containing lubricants for engine use, and a protocol to assess PVL and PSSI based on viscosity measurements at 106 s-1 before and after shear thinning was outlined in this article.<br /> <br /> It is also important to account for possible permanent viscosity loss when measuring the viscosity of polymer solutions in very high shear rate viscometers such as the USV. This can be done by minimizing the amount of shear to which the lubricant is subjected or by taking successive measurements and subtracting the permanent viscosity loss taking place in each of the first few strain rate cycles.<br /> <br /> ACKNOWLEDGEMENTS<br /> The authors thank Castrol Ltd. for their kind support that allowed the research described in this article to be carried out and PCS Instruments for software modifications to the USV that facilitated the research.