Tribology Letters 5 (1998) 283–286 283 The relationship between PFPE molecular rheology and tribology T.E. Karis a and M.S. Jhon b a IBM Research Division, Almaden Research Center, 650 Harry Road, San Jose, CA 95120-6099, USA b Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213-3890, USA Received 22 February 1998; accepted 21 June 1998 The tribology of several perfluoropolyether (PFPE) lubricants was studied using a pin on disk (POD) test. During the POD test, PFPE is incrementally removed from the track with each sliding cycle. The number cycles to failure, N F , is detected as a sudden increase in the friction coefficient. Molecular theory for polymer melt rheology was employed to develop a universal scaling rule. The PFPE removal rate coefficient is proportional to a parameter containing the bulk viscosity, degree of polymerization, and temperature and structure scaling coefficients. The parameter is a measure of the frictional resistance to segmental sliding along the surface in the contact zone. The temperature scaling coefficient corrects for the absence of free volume in the molecularly-thin lubricant film. The structure scaling coefficient accounts for differences in the energy barriers to internal rotation. This is the first description of a relationship for the tribological properties of PFPEs that takes into account their viscosity, molecular structure, degree of polymerization, and temperature. Keywords: magnetic recording, perfluorinated lubricant, perfluoropolyether, perfluoropolyalkylether, pin on disk tribology, molecular rheology, thin film viscosity 1. Introduction Perfluoropolyether (PFPE) lubricants are used on mag- netic recording media [1] and are being studied for use in advanced turbine engines [2]. PFPE are selected as lubri- cants in these applications because of their low vapor pres- sure and surface energy as well as their good thermal sta- bility. PFPE tribology was studied in sliding wear tests on magnetic recording media [3] and on stainless steel [4]. The mechanism of tribological tests on PFPE has been inves- tigated by sensitive microanalytical techniques [2–7]. The physicochemical properties of PFPE were measured [8]. Despite considerable research on this subject, the relation- ship between PFPE molecular rheology and tribology is poorly understood. We explore a novel approach to under- standing this relationship. A comparison is made between the tribological perfor- mance of three different PFPE molecular structures and sev- eral molecular weights on a model surface using the pin on disk (POD) sliding test. The PFPE considered in this study are Z, Y, and K. The K and Y have a –CF 3 side group, and Z is a linear chain. The molecular structures of the PFPE are shown in figure 1. The results are discussed in terms of polymer theory. A universal scaling rule is proposed. 2. Apparatus and materials The POD apparatus consists of a disk test stand, a pin mounting device, and equipment for recording the friction force on the pin [9]. The disks were rotated at 18 rad/s. The pin was hemispherical with a 10 cm radius of curva- ture, mirror polished, and made of the same TiC–Al 2 O 3 ceramic used in magnetic recording sliders. The pin was maintained in contact with the disk by a 150 mN load. The friction force is monitored during the test to determine the number of cycles to failure, N F . The endpoint is in- dicated by a sudden increase in the friction force, and at the endpoint there is no detectable physical damage to the media surface [6]. The test surfaces were particulate mag- netic recording media (a 0.8 μm thick film of magnetic oxide particles in a porous epoxy phenolic binder on alu- minum magnesium alloy disk). A solution of the PFPE in a volatile solvent is sprayed onto the disk, and the excess PFPE is wiped off. Wiping forces the PFPE into pores within the media [10]. The PFPE molecular weight, M , was measured by gel permeation chromatography, and the viscosity, η, was measured using a Ubbelohde viscome- ter [11]. The degree of polymerization, D p , was calculated from M and the molecular structures in figure 1. M , η, and D p are listed in table 1. 3. Results The N F depends on both the initial areal density and the type of PFPE. A model was adopted to separate the PFPE- dependence of N F from the dependence on the initial PFPE areal density. Each time the pin passes over a point on the track, an incremental amount of lubricant is removed [3], giving rise to an overall rate of change in the areal density of PFPE in the track, dm/dN , where m is the lubricant areal density, and N is the number of sliding cycles. Since details of the tribochemistry at the interface are unknown, an nth-order rate equation was employed for the purpose of data analysis: dm dN = −K n m n , (1)  J.C. Baltzer AG, Science Publishers