WHAT MAKES A GOOD MR FLUID?

Transcription

1 Presented at the 8 th International Conference on Electrorheological (ER) Fluids and Magneto-rheological (MR) Suspensions, Nice, July 9-13, 2001 WHAT MAKES A GOOD MR FLUID? J. DAVID CARLSON Lord Corporation, 110 Lord Drive, Cary, North Carolina 27511, USA jdcarlson@lord.com Experience in manufacturing MR fluids for commercial application has shown that some of the greatest barriers to commercial success are not factors or conditions normally considered in the laboratory. The present paper looks at conditions found in MR fluid devices operating in real-world applications where shear rates may exceed 10 5 sec -1 and devices are called upon to operate for very long periods of time. The problem of In-Use-Thickening wherein a MR fluid subjected to long-term use progressively thickens until it eventually becomes an unworkable paste is presented. The search for a solution to this heretofore unrecognized problem delayed commercial introduction of the Lord truck seat damper system for several years. Today, good fluids are able to operate for long periods with minimum in-usethickening. 1 Introduction The most common response to the question of what makes a good MR fluid is likely to be "high yield strength" or "non-settling". However, those particular features are perhaps not the most critical when it comes to ultimate success of a magnetorheological fluid. The most challenging barriers to the successful commercialization of MR fluids and devices have actually been less academic concerns. As anyone who has made MR fluids knows, it is not hard to make a strong MR fluid. Over fifty years ago both Rabinow and Winslow described basic MR fluid formulations that were every bit as strong as fluids today. A typical MR fluid used by Rabinow consisted of 9 parts by weight of carbonyl iron to one part of silicone oil, petroleum oil or kerosene. 1 To this suspension he would optionally add grease or other thixotropic additive to improve settling stability. The strength of Rabinow s MR fluid can be estimated from the result of a simple demonstration that he performed. As shown in Figure 1, Rabinow was able to suspend the weight of a young woman from a simple direct shear MR fluid device. He described the device as having a total shear area of 8 square inches and the weight of the woman as 117 pounds. For this demonstration to be successful it was thus necessary for the MR fluid to have a yield strength of at least 100 kpa. Figure 1 - Rabinow's magnetic fluid demo. 1

2 Presented at ERMR01, Nice (2001) 2/7 MR fluids made by Winslow were likely to have been equally as strong. A typical fluid described by Winslow consisted of 10 parts by weight of carbonyl iron suspended in mineral oil. 2 To this suspension Winslow would add ferrous naphthenate or ferrous oleate as a dispersant and a metal soap such as lithium stearate or sodium stearate as a thixotropic additive. The formulations described by Rabinow and Winslow are relatively easy to make. The yield strength of the resulting MR fluids is entirely adequate for most applications. Additionally, the stability of these suspensions is remarkably good. It is certainly adequate for most common types of MR fluid application. As early as 1950 Rabinow pointed out that complete suspension stability, i.e. no supernatant clear layer formation, was not necessary for most MR fluid devices. MR fluid dampers and rotary brakes are in general highly efficient mixing devices. When the piston in a MR fluid damper moves, the MR fluid jets through the orifices quite rapidly causing it to swirl and eddy vigorously even for low piston speed. Similarly, the shear motion that occurs in a MR brake causes vigorous fluid motion. As long as the MR fluid does not settle into a hard sediment, normal motion of the device is generally sufficient to cause sufficient flow to quickly remix any stratified MR fluid back to a homogeneous state. For a small MR fluid damper such as the Lord Motion Master RD-1005, two or three strokes of a damper that has sat motionless for several months are sufficient to return it to a completely remixed condition. Except for very special cases such as seismic dampers, lack of complete suspension stability is not a necessity. It is sufficient for most applications to have a MR fluid that soft settles upon standing a clear layer may form at the top of the fluid but the sediment remains soft and easily remixed. Attempting to make these MR fluids absolutely stable may actually compromise their performance in a device. One of the areas where MR fluids find their greatest application is in linear dampers that effect semi-active control. These include small MR fluid dampers for controlling the motion of suspended seats in heavyduty trucks, larger MR fluid dampers for use as primary suspension shock absorbers and struts in passenger automobiles and special purpose MR fluid dampers for use in prosthetic devices. In all of these devices one of the most important fluid properties is a low-off state viscosity. While in all of these examples having a MR fluid with a high yield strength in the on-state is important, it is equally important that the fluid also have a very low offstate. The very ability of an MR fluid device to be effective at enabling a semi-active control strategy such as sky -hook damping depends on being able to achieve a sufficiently low off-state. Care must be taken in choosing fluid stabilizing additives so that they do not adversely affect the off-state viscosity. Earthquake dampers and other some other special applications in which the device will sit quiescent for very long periods of time represent special cases where fluid stability issues may have overriding importance. Because of the transient nature of seismic events these dampers never see regular motion, which can be relied on to keep the fluid mixed. This lack of motion also has it benefit. Unlike dampers used in highly dynamic environments, seismic dampers do not need to sustain millions of cycles. The fact that durability and wear are not issues gives the fluid designer grater latitude to formulate a highly stable fluid. MR fluids for these applications are typically formulated as shearingthinning thixotropic gels.

3 Presented at ERMR01, Nice (2001) 3/7 2 The MR Fluid Environment Outside the Laboratory In exploring the question of what make s a good MR fluid, it is instructive to consider the factors and conditions seen by an MR fluid, not in the laboratory, but in devices in actual service. Even under normal device operation these conditions can be quite extreme. The specific environment to which a MR fluid is exposed inside a MR device is quite different from the regime that is normally measured in the laboratory. Consider the conditions found inside a MotionMaster RD damper manufactured by Lord Corporation for use in heavy-duty truck and bus seating. 3,4,5 This damper is designed to generate a nominal resistive force of about 1200 N at an input current of 1 amp. The shear rate inside the MR fluid valve under normal operating conditions when the speed across the damper is in the range of m/s is 1-4 x 10 4 sec -1. Under extreme conditions this damper may experience speeds in excess of 1 m/s or shear rates greater than 2x10 5 sec -1. An automotive primary suspension MR fluid shock absorber may experience shear rates that are even larger, perhaps as much as 106 sec -1. The peak mechanical power dissipation in the RD damper is normally in the range of watts. Since the damper contains about 70 cm 3 of fluid this corresponds to about 0.9 to 3.4 watts/cm 3 of fluid in the damper. Under extreme conditions the peak power dissipation may exceed 1200 watts or 17 watts/cm 3. Conditions in a rotary brake are, in general, not as extreme until one gets to the power per unit volume level. The Rheonetic MRB-2107 brake is designed to produce a maximum dissipative torque of about 6 Nm at an input of 1 amp. 6 Under normal operating speeds of RPM this corresponds to shear rates of sec -1. Power dissipation ranges from watts. Since the fluid volume in the brake is only 5 cm 3, this corresponds to watts/cm 3. Before a MR fluid is placed into a device is must be made. The manufacturing process for commercial MR fluids is quite different from the mixing of MR fluids in a laboratory. A MR fluid that is to be commercially useful outside of the laboratory must be amenable to volume production. As such the process for making the fluid must be scalable well beyond a liter-sized batch by many orders of magnitude. The process must also yield a result that is repeatable and reproducible. Today s highly competitive, global market, particularly any market related to the automotive industry, demands that MR fluids be made with a process that is QS-9000 compliant and certified. 5,6 The scale of fluid production necessary for even a modest automotive application is large. A single MR fluid device such as a damper on a single automotive platform model can easily require a total fluid volume production on the order of 10 5 liters/year. Depending on the iron particle loading, that amount corresponds to 2-4 x 10 5 kg/year or 1-2 tons of MR fluid per day. A 50 gallon (185 liter) barrel of MR fluid weighs between 1/2 ton and 1 ton. Thus, one must be able to reliably and reproducibly make several barrels of MR fluid per day on a continuous basis to support even a single, modest automotive application. Being able to cope with the logistics of handling and mixing large quantities of heavy, pyrophoric materials is an absolute necessity. 3 The Problem of In-Use-Thickening During the development and commercialization phases of several recent MR fluid devices, problems with the MR fluid were discovered that were not apparent in the early research

4 Presented at ERMR01, Nice (2001) 4/7 phases of these projects. An example of one such problem is a phenomenon called "In - Use-Thickening" or IUT. If an ordinary MR fluid is subjected high stress and high shear rate over a long period of time, the fluid will thicken. Superficially, this process appears much like the process of churning cream to make butter. An originally low-viscosity, i.e. low off-state, MR fluid progressively becomes thicker and thicker until it eventually becomes an unmanageable paste having the consistency of shoe polish. An example of the in-use-thickening or IUT problem is illustrated in Figure 2. In this example, the off-state force of an early version of the MotionMaster RD-1005 truck seat damper is shown over the course of a life -cycle test. The damper was operated at 1 Hz with a 25.4 mm stroke in the on-state with 1 amp applied continuously such that the damper force was approximately 1200 N. Under these conditions, the RMS mechanical power dissipation was approximately 68 watts. Aluminum cooling fins were fastened to the damper body to augment convection cooling and keep the damper temperature from exceeding 100C. Periodically, the current applied to the damper was removed and the offstate force was measured. These periodic off-state forces over the course of about 600,000 on-state cycles or about 167 hours of operation are plotted in Figure 2. Figure 2 - Example of IUT problem in an early MR fluid formulation. The IUT phenomenon appears as a progressive increase in the off-state force. By the time the damper had experienced 600,000 on-state cycles, the off-state force has increased from 200 N to 500 N. This factor of 2.5 increase in off-state force is sufficient to render the damper unsuitable for implementing effective semi-active vibration control. For such, a low off-state force is just as important as a high on-state force. While this fluid provided very good performance when the damper was first tested, the fluid thickening that developed eventually rendered the damper unacceptable. The off-state force increase that is the hallmark of the IUT problem is due to an increase if the off-state viscosity of MR fluid subjected to long-term stress. The cause of this viscosity increase is believed to be due to spalling of the friable surface layer from the

5 Presented at ERMR01, Nice (2001) 5/7 surface of the carbonyl iron particles that typically comprise the particle component of MR fluid. This surface layer composed of iron oxides, carbides and nitrides is rather brittle. When subjected to high inter-part icle stresses this surface layer fractures and breaks into small pieces that separate from the primary particle. These very small nanometer-sized secondary particles have a very large surface area to weight ratio. As such even a very weight small amount of these secondary particles is capable of significantly affecting the rheology of the overall MR fluid. Viscosity (Pa-s) % 0.5 % K 400K 600K Cycles Figure 3 - Viscosity increase due to simulated surface spalling compared to observed IUT. The increase in MR fluid viscosity that results from the addition of a volume of nano-sized particulates that may have spalled from the surface of the carbonyl iron particles is illustrated in Figure 3. In this graph the observed progressive off-state viscosity increase due to IUT is compared to the viscosity increase caused by replacing a small part of the carbonyl iron mass with nano-sized ferrite particles.. The 0.5% and 1% dashed lines in this graph correspond to the increased viscosity that results if that amount of the carbonyl iron is replaced by nano-sized ferrite. Replacing 0.5% of the mass of carbonyl iron with the same mass of nano-sized ferrite particles results in almost a factor of two increase in MR fluid of-state viscosity. Replacing 1% of the carbonyl iron mass with ferrite results in approximately a factor of three increase in viscosity. Thus, even if only a small portion of the carbonyl iron particles is shed secondary nano-sized debris, a very large increase in off-state viscosity can occur. IUT was not identified as a serious problem until long-term life testing to qualify commercial MR fluid truck seat dampers was well underway. Finding and implementing a viable solution to the IUT problem required the focused efforts of a six-person team working over a period of about two years. The IUT problem delayed commercial introduction of the MotionMaster truck seat damper for over two years. Figure 4 shows the progressive progress that was made in dealing with the IUT problem Starting with the original MR fluid formu lation, MR fluid formulations that showed progressively less IUT were developed over the course of several years intensive development effort. Eventually, a fluid formulation that showed no significant IUT after two million cycles was found. Over two years of intensive fluid development were required to progress from Curve 1 to Curve 4, a fluid that could survive the minimum 2 million cycles required for a commercial MR fluid seat damper. Today the "in-use-thickening" problem has not only been

6 Presented at ERMR01, Nice (2001) 6/7 identified but has been solved. Today, good MR fluids show no measurable in-usethickening after more than 10 million cycles in the MotionMaster RD damper. Figure 4 - Progressive development of IUT resistant MR fluid formulations. 4 MR Fluid Life Depending on the conditions of the specific application, all MR fluids will eventually show some degree of deterioration. Such deterioration is usually manifested as a thickening of the fluid as described above although problems may occur as well. Silicone oil based fluids, for instance, are prone to cross-linking if exposed to high temperatures for extended periods or to ionizing radiation. The amount of deterioration generally depends on the shear rate, temperature and duration. A measure that is useful in predicting the expected life of a MR fluid is the lifetime dissipated energy or LDE defined in Equation 1. 1 LDE = V life P dt 0 where P is the instantaneous mechanical power being converted to heat in the MR device. Thus, LDE is simply the total mechanical energy dissipated per unit volume of MR fluid over the life of a device. In the case of MR Fluid #4 in Figure 3 the damper was operated at an amp litude of +/ mm at 1 Hz with 1 amp applied to generate 1200 N for 2.5 million cycles. The RMS power dissipation was 68 watts and the damper contained 70 cm3 of fluid. Thus, over the course of approximately 29 days of continuous operation the fluid in this damper accumulated a LDE of 2.4x10 6 J/cm 3. It is our experience that the best MR fluids today can (1)

7 Presented at ERMR01, Nice (2001) 7/7 sustain approximately 10 7 J/cm 3 before they have thickened to the point where they are no longer useful in a controllable MR fluid device. 5 Conclusion The answer to the question What makes a good MR fluid? is It depends. It depends on the type of device in which the MR fluid is used, the conditions to which the fluid is exposed and the duration of that exposure. MR fluids that are considered good in one application may fail miserably in another type of device. MR fluid development is of course a balancing act that is highly coupled with MR device design. In evaluating the quality of an MR fluid it is important to consider the actual conditions to which it will be exposed and not just the rheological behavior measured under normal laboratory conditions. MR fluid durability and life have been found to be more significant barriers to commercial success than yield strength or stability. Amenability of a particular MR fluid formulation to being scaled to volume production must also be considered. Challenges for future MR fluid development are fluids that operate in the high shear regime of 10 4 to 10 6 sec -1 and fluids able to sustain high LDE greater than 10 7 J/cm 3. References 1. Magnetic Fluid Clutch, Technical News Bulletin, National Bureau of Standards, 32/4 (1948) W. Winslow, Field Responsive Fluid Couplings, US Patent No. 2,886,151, (1959). 3. Motion Master Ride Management System, Pub No. PB8008a, Lord Corp., Cary (1998). 4. J.D.Carlson, and M.J.Chrzan, Magnetorheological Fluid Dampers, US Patent No. 5,277,281 (1994). 5. J.D.Carlson, K.A.St.Clair, M.J.Chrzan and D.R.Prindle, Controllable Vibration Apparatus, US Patent No. 5,878,851 (1999). 6. J.D.Carlson, D.M.Catanzarite and K.A.St.Clair, Commercial Magneto-Rheological Fluid Devices, Proc. 5th Int. Conf. on ER Fluids, MR Fluids and Assoc. Tech., July 1995, W.A.Bullough, Ed., World Scientific, Singapore (1996) International Automotive Sector Group, (1999). 8. Quality Systems Requirements QS -9000, AIAG-Automotive Industry Action Group, Southfield, MI (1998).