Feasibility of high pressure tribochemistry experiments from Raman study of lubricant additives at ambient conditions

Full text

(1)2009:147 CIV. MASTER'S THESIS. Feasibility of high pressure tribochemistry experiments from Raman study of lubricant additives at ambient conditions. Joel Andersson. Luleå University of Technology MSc Programmes in Engineering Engineering Physics Department of Applied Physics and Mechanical Engineering Division of Physics 2009:147 CIV - ISSN: 1402-1617 - ISRN: LTU-EX--09/147--SE.

(2) Acknowledgements Many people have taken part in the making of this master thesis. My special thanks go to my associate and good friend Shuai Wei, who worked alongside me on this project during the autumn of 2008. Without his hard work and ideas this thesis would not have been what it is today. A lot of help and teaching from master students Mattias Mases and Andreas Müller was crucial for learning about Raman spectroscopy and the equipment in the lab. My warm thanks go to them and I wish them all the luck in the future. I want also to thank all the people in the high pressure physics group, whose joyous spirit and nice company I’ve had the pleasure to experience. Great thanks go to my supervisor Alexander Soldatov who made this thesis possible and whose ideas and engagement were and are crucial and an inspiration for the whole group. I want to thank Anna-Carin Larsson at the chemistry department of LTU for her valuable input and interesting discussions which were a shot of energy at all times. My warmth and love goes to my family and my friends who have supported me at all times..

(3) Abstract In this work a possibillity of studying ZDDP lubricant additive in-situ under high pressure in a diamond anvil cell (DAC) using Raman spectroscopy is examined. Pure ZDDP does not exhibit anti-wear properties whereas its presence in the lubricant results in forming a protective layer (tribofilm), the process which mechanism is still unknown. Therefore it is desirable to study chemical transformation of ZDDP in a solution at real tribological conditions, i.e. high pressure and temperature. Recent molecular dynamics simulations show cross-linking behaviour of the decomposition products of ZDDP under high pressure which may explain the mechanism of tribofilm formation thus making high pressure experimental verification of the model necessary. We use three different lasers and two different solvents to examine the possibillities of studying ZDDP solutions by Raman spectroscopy. The Raman spectrum of ZDDP in solid form was recorded and served as a reference for further experiments. The hydrocarbons − hexadecane and diethyleneglycoldibutylether (molecules with higher polarity) − were used as ZDDP solvents. Raman spectra of the solutions with different ZDDP concentrations were recorded and compared to the spectra of pure solvents and solid ZDDP. The signal from hydrocarbons is overlapping with that from the ZDDP in the solution which makes study of the latter in a DAC problematic. Nevertheless we demonstrate in this work feasibility of the tribochemical experiments in a DAC by showing that the unwanted contribution of the solvents to the Raman spectrum of ZDDP can be eliminated. Our experiments show that even more clear spectral separation can be made using an ethanol/methanol mixture as a solvent although in this system a chemical alteration of the ZDDP molecules may occur. The Raman peak with highest intensity in ZDDP spectrum is assigned to the bonds in the core of the ZDDP molecule. Following evolution of this peak at high pressure/temperature would provide information on the molecular structure changes, a first step to the tribofilm formation.. iii.

(4) iv.

(5) Contents 1 Introduction 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Theses outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Theory 2.1 Lubrication and motor oil additives . . . . 2.2 The chemistry of ZDDP . . . . . . . . . . 2.2.1 ZDDP in solution . . . . . . . . . . 2.2.2 Ligand exchange . . . . . . . . . . 2.2.3 Oxidation inhibition . . . . . . . . 2.2.4 Decomposition of ZDDP . . . . . . 2.3 ZDDP film formation . . . . . . . . . . . . 2.3.1 Thermal film . . . . . . . . . . . . 2.3.2 Tribofilm . . . . . . . . . . . . . . 2.3.3 Step by step ZDDP film formation 2.4 Computer simulation . . . . . . . . . . . . 2.5 Cross-linkage in experiments . . . . . . . . 2.6 Unsolved problems . . . . . . . . . . . . . 2.7 Raman spectroscopy . . . . . . . . . . . .. 1 1 1 2. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. . . . . . . . . . . . . . .. 3 3 3 3 5 5 6 6 7 7 7 9 11 12 12. 3 Methods and equipment 3.1 Raman Spectroscopy . . . . . . . . . . . . . . 3.1.1 Exitation Sources . . . . . . . . . . . . 3.1.2 Confocal Raman Microscope . . . . . 3.2 High pressure . . . . . . . . . . . . . . . . . . 3.2.1 Membrane Diamond Anvil Cell(DAC) 3.2.2 Diamond anvil and gasket . . . . . . . 3.2.3 Pressure measurement in the DAC . . 3.3 High Temperature . . . . . . . . . . . . . . . 3.4 Experimental details . . . . . . . . . . . . . . 3.4.1 The Samples . . . . . . . . . . . . . . 3.4.2 excitation laser details . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . .. 15 15 15 15 18 18 18 20 20 21 21 22. . . . . . . . . . . . . . .. 4 Results and discussion 4.1 Pure ZDDP substances . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Study of suitable solvents for high pressure experiments on ZDDP 4.2.1 Hexadecane/ZDDP solution . . . . . . . . . . . . . . . . . 4.2.2 Diethyleneglycoldibutylether/ZDDP solution . . . . . . . 4.2.3 Methanol/ethanol/ZDDP solution . . . . . . . . . . . . . 4.3 ZDDP in hexadecane . . . . . . . . . . . . . . . . . . . . . . . . . 4.3.1 Excitation source determination . . . . . . . . . . . . . . v. 25 25 28 28 30 30 31 32.

(6) vi. CONTENTS. 4.4 4.5 4.6. 4.3.2 Peak assignment . . . . . . . Raman study of the tribofilm . . . . High pressure and high temperature 4.5.1 Pre-heated sample analysis . High pressure measurements plan . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. . . . . .. 33 35 37 37 39. 5 Conclusions and Future work. 43. 830 nm Laser Data. 49. Peakfit files, ZDDP and hexadecane. 51. Nuclear magnetic resonance data of ZDDP. 57. Details for high pressure and temperature spectroscopy investigations 59.

(7) Chapter 1. Introduction 1.1. Background. Triboligy originates from the Greek language, tribo meaning rub and logos meaning principle or logic. Tribology is the study of interacting surfaces and it includes the fields of lubrication, wear and friction. The first recorded tribology investigations were conducted in the field of friction by Leonardo da Vinci in the 15th century. This thesis is a work on additives in lubricants. Lubricants were used first by the Romans while the research of lubricants began for real in the 19th century during the industrial revolution. The lubricants have grown more sophisticated and today it is common to use reactive lubricants. Reactive lubricants are divided into mineral oils with additives and synthetic oils. This report treats a lubricant additive, namely Zinc Dialkyl DithioPhosphate, henceforth referred to as ZDDP. It is both one of the cheapest and one of the most efficient friction and wear reducing additives used today. It has been known for some time that ZDDP in solution between interacting surfaces creates a tribofilm. Although extensively used the chemical processes involved in the formation of this film is still not fully understood. Incompatibility with aluminium engines and environmental issues makes finding replacements for ZDDP urgent. Except for curiosity, this also stimulates a need for understanding. Several recent experiments have been conducted to understand the functionality of ZDDP and many methods have been used to analyse the chemistry of ZDDP as well as the tribofilm formation ( [1], [2]) . Furthermore many studies have focused on thermal decomposition of ZDDP, a process seemingly important to tribofilm formation ( [3], [4]). Recently influence of not only of temperature but also of pressure on film formation has come into focus.. 1.2. Motivation. As ambitious strife is focused on ZDDP functionality, understanding of the processes begin to deepen. In a computer simulation, high pressure was found to induce cross linkage of Zn-phosphates. This lead to an increased interest of the pressure effect on ZDDP, as the anti-wear property could be derived from this cross-linkage. IR-studies of ZDDP under high pressure has been conducted [5] but Raman studies have not yet been performed. In this study we will use 1.

(8) 2. Introduction. hexadecane as a model system for oil and study the behaviour of ZDDP in solution. All as a preparation for using a diamond anvil cell combined with Raman spectroscopy. The project is a pilot project for the High Pressure Spectroscopy lab under supervision of Alexander Soldatov. It will be possible to study ZDDP solution in a diamond anvil cell using Raman spectroscopy.. 1.3. Theses outline. The thesis starts with a short introduction followed by the theory of ZDDP as well as Raman spectroscopy. Following this is a description of the experimental methods and equipment..

(9) Chapter 2. Theory 2.1. Lubrication and motor oil additives. Lubrication is a broad field, wherein dry lubrication, lubrication with mineral oil and lubrication with synthetic oils are all included. Different kinds of lubrication are applied depending on the situation or application. The most common lubricant is oil. In turn, the most common way to enhance and specialize oil applicability is by using an additive. For parts with iron, the most common anti-wear additive is ZDDP.. 2.2. The chemistry of ZDDP. Although initially used as an oxidization inhibitor in sub-marines in the 30’s, its anti-wear abilities were soon discovered to be superior. The cheap and effective ZDDP is the most successful motor-oil additive of the 20th century. The tendency of phosphorus to form a harmful surface on the catalytic converter of vehicle engines, along with the incapacity of ZDDP films to form on aluminium, has accelerated the need to understand the chemical processes of ZDDP, in order to produce new additives. The current understanding of ZDDP includes many chemical reactions. This theoretical part will first discuss the ZDDP molecule and some of the reactions that ZDDP undergo in solution.. 2.2.1. ZDDP in solution. The ZDDP molecule consists of a Zn atom in the centre, connected to 4 sulphur atoms, which in turn are connected to two phosphorus atoms1 . The phosphorus connects to an oxygen atom which in turn connects with an alkyl, aryl or alkaryl group. See figure 2.1 for an image of the ZDDP molecule. In solution ZDDP exist in equilibrium between monomeric and dimeric forms. In solvents of low polarity ZDDP may also exist in tetrameric or even higher forms [7]. figure 2.2 shows the dimeric and monomeric forms of ZDDP. In a paper by E.S. Yamaguchi [8] dynamic light scattering has even indicated that a tetrameric form of ZDDP dominates in a solution of low polarity. Except for the neutral ZDDP seen in figure 2.1 there is also a basic form of ZDDP, see figure 2.3. The basic 1 Best. described with a resonant sesquibond description [6]. 3.

(10) 4. Theory. Figure 2.1: The ZDDP molecule. The white "atoms" correspond to the side chain, often denoted by R. Image source: [7] form of ZDDP has been observed in solution. This form can convert to the neutral form of ZDDP if heated. The literature is divided in terms of the efficiency of the different forms. For example [8] claims that there is a difference between the forms of ZDDP as far as wear properties are concerned, [7] concludes that they behave similarly in general.. Figure 2.2: The dimeric and monomeric form of ZDDP. Image source: [7]. Figure 2.3: The "Basic" form of ZDDP that can be observed in solution. Image source: [7] In solution and application ZDDP reacts chemically. First of all there occur reactions directly in solution. Secondly, if a surface is present, adsorption of ZDDP is possible. If a surface is present and there is adsorption film formation is possible, resulting in a tribofilm or a thermal film. The possible reactions immediately in solution are ligand exchange, peroxide decomposition and thermal.

(11) 2.2 The chemistry of ZDDP. 5. degradation of ZDDP. figure 2.4 shows the different possible chemical reactions of ZDDP in solution.. Figure 2.4: Possible reactions of ZDDP in solution. Image source: [7]. 2.2.2. Ligand exchange. Ligand exchange is a process where the Zn atom in ZDDP is exchanged with another metal ion, such as Cu or Fe to form another MDDP molecule. The chemical reaction can be described by ((RO)2 PS2 )2 Zn + M2+ A ((RO)2 PS2 )2 M + Zn2+ .. (2.1). The ligand exchange has been investigated thoroughly in [9] where the relative order of extraction was found, Pd+2 > Au+3 > Cu+ > Hg+2 ≥ Ag+ > Cu+2 > Sb+3 > Bi+3 > Pb+2 > Cd > Ni+2 > Zn+2 .. (2.2). This should be interpreted as ions to the left will replace ions to the right of them in MDDP.. 2.2.3. Oxidation inhibition. ZDDP has the ability both to decompose hydro peroxides and peroxy radicals. This results in the good oxidation inhibition properties of the additive. Once the ZDDP has acted as oxidization inhibitor the rest products are no longer able to act as an anti-wear agent [10] [11]. This means care must be taken so that our samples are not extensively exposed to oxygen in which case processes that lead to film-formation and anti-wear properties of ZDDP may be disturbed. The anti-oxidizing properties of ZDDP are further enhanced by the thermal film described in 2.3.1..

(12) 6. 2.2.4. Theory. Decomposition of ZDDP. Under high temperature ZDDP acts in a different manner than under room temperature with oxygen. ZDDP will decompose to several different molecules in the temperature range 130-230℃, see for example [4] and [3]. The temperatures vary with the alkyl groups present as well as on the metal ion involved in the MDDP. Heating ZDDP will lead to Zn phosphate solid deposit, alkyl sulphides, mercaptans (thiols), hydrogen sulphide and olefins (alkenes) [7]. figure 2.5 shows some different decomposition products of ZDDP. A first step in this thermal degradation process is that sulphur and oxygen exchange connection with the R-group that was initially connected with the oxygen. That is, the R-group which is bounded with the oxygen is transferred to the sulphur atom. This explains the large amount of decomposition products with sulphur bounded to the R-group. This transfer occurs when ZDDP is heated to relatively. Figure 2.5: Some typical decomposition products of ZDDP, when it has been heated in solution. mild temperatures, or even during room temperature in the presence of certain Lewis acids [12]. Hihgetag and Teichmann also point out that the whole set of reactions of the mono and dithiophosphates are effectively aimed at replacing P-S, P=S and C-O bonds by C-S bonds (leading to mercaptans, sulphides and disulphides as products), P-O and P=O bonds (leading to polyphosphates as products) [12].. 2.3. ZDDP film formation. The first step in ZDDP film formation is most likely the adsorption of ZDDP on a surface. This means ZDDP will react with iron on the surface. [13] proposed that the connection with iron is with the sulphur atom of ZDDP. [14] discusses the adsorption of ZDDP on Iron powder and conclude that ZDDP adsorbs without chemical changes up to 40℃. At temperatures over 60℃ Zn will leave the surface and once again blend in the solution. ZDDP can in other words adsorb on iron or steel at room temperature or higher with a result that vary depending on the.

(13) 2.3 ZDDP film formation. 7. temperature. K. Homan et al. [15] suggest the following reactions of ZDDP in solution with oxidized iron or nickel surfaces, 3((RO)2 PS2 )2 Zn + Fe2 O3 A 3((RO)2 PS2 )2 + 2Fe + 3ZnO.. (2.3). It seems like ZDDP interacts with an oxidized surface by “cleaning” it from Oxide. It is not completely clear today how ZDDP films take shape, although it is quite likely that the S/O exchange mechanism is important. Further analysis is still needed to identify all the reaction intermediates of the film formation.. 2.3.1. Thermal film. The transparent solid so called thermal film form on a steel surface when the ZDDP solution in contact with a surface is heated above 100℃. The film consists of a thin (about 10 nm) outer layer of polyphosphate grading to pyro or orthophosphate in the bulk. The main cation in ZDDP thermal films are Zn, unlike tribofilms whose main cation is Fe [16]. The thermal film has been extensively studied. The most recent paper on the thermal film that the author knows of is [17].. 2.3.2. Tribofilm. A film often referred to as tribofilm can form at much lower temperatures than the thermal film, even as low as room temperature. The rate of formation depends on the temperature [18]. The tribofilm forms in rubbing tracks and if there is a sliding contact rather than a rolling contact or if the lubricant layer thickness is much greater than the surface roughness [19]. The structure of the tribofilm has been found to be Pad-like. This can be seen in figure 2.6. As can be observed, the cross section of the film is separated by valleys [20]. The pads consist mainly of glassy phosphate with a thin, outer layer of Zn polyphosphate and with pyro- or ortho-phosphate in the bulk [7]. see figure 2.7 for an image of the individual pads. It has been found that the pads are smart in the sense that they become harder during nano indentation [21]. This could be explained by cross-linkage of Zn phosphates under pressure. A crosslinked network that takes form would distribute the load and spread it to a larger contact area and thus protect the surface from wear. A computer simulation published in 2005 was the first indication of cross-linkage in motor oil additives that the author know of [22]. The proposals in that study were experimentally tested in [23]. They conclude that there are still a lot of experiments required to examine the cross-linkage theory.. 2.3.3. Step by step ZDDP film formation. Marina L. Suominen Fuller et al. proposed an early model for the ZDDP film formation. [24]. In the first step ZDDP is adsorbed on the rubbing surfaces as described in equation 2.4. Zn[(RO)2 PS2 ]2 (solution) A Zn[(RO)2 PS2 ]2 (adsorbed). (2.4). After some time, ZDDP (partially) converts to a LI-ZDDP in solution, Zn[(RO)2 PS2 ]2 (solution) A Zn[O2 P(SR)2 ]2 (LI-ZDDP in solution).. (2.5).

(14) 8. Theory. Figure 2.6: An SFM Image of the bulky structure of ZDDP tribofilm. The films have been formed in a Cameron-Plint wear rig for different times. (A) 10 min (B) 40 min (C) 60 min and (D) 120 min. Image source: [20]. Figure 2.7: The pad like structure of the tribofilm. Image source: [7].

(15) 2.4 Computer simulation. 9. The LI-ZDDP will later adsorb along with ZDDP on the surface, Zn[O2 P(SR)2 ]2 (LI-ZDDP in solution) A Zn[O2 P(SR)2 ]2 (adsorbed).. (2.6). After this, the adsorbed species of ZDDP and LI-ZDDP react thermooxidatively and a long chain polyphosphate results on the surface (equation 2.7). Zn(RO)4 P2 S4 + O2 A Zn(PO3 )2 + sulphur species. (2.7). As more rubbing occurs phosphates come in contact with water and form short chain polyphosphate (equation 2.8 and/or equation 2.9). 7Zn(PO3 )2 + 6H2 O A Zn7 (P5 O16 )2 + 4H3 PO4. (2.8). 2Zn(PO3 )2 + 3H2 O A Zn2 P2 O7 + 2H3 PO4. (2.9). The amount of LI-ZDDP and ZDDP varry between tribofilms and Thermal films, but the final structure of the film is similar. Figure 2.8 shows scematically the micrometer of the film closest to the surface.. Figure 2.8: The ZDDP film constituents. Image source: Surface Chemistry in Tribology by Andrew J. Gellman and Nicholas D. Spencer. 2.4. Computer simulation. Computer simulations can be effectively used to investigate the molecular level details of physical processes. The simulations results can be directly compared with experimental results and help us understand what lies behind larger scale observations. The computer simulations are limited by efficiency of algorithms in combination with computer muscle limitations. For this reason most tribological simulations deal with simple hydrocarbon models instead of complex lubricant systems. Nicholas J. Mosey et al. reported an ab-initio quantum chemical simulation of tribochemical reaction of Zn phosphate2 in 2005 [22]. Their results suggest 2 Zn. phosphate is one decomposition product of ZDDP.

(16) 10. Theory. that the formation, functionality and friction properties of ZDDP anti wear film are directly depending on pressure induced changes in the atomic bonds of Zn. The initially viscoelastic system of Zn phosphate is cross-linked and shape a polymer like structure when put under pressure. The degree of connectivity within the system strongly depends on the maximum pressure to which the system has been exposed. The pressure and the temperature to which a system is exposed play a key role to determine the antiwear and antioxidant properties of ZDDPs in this tribological system. Previous attempts to simulate the tribological system have failed to incorporate these effects adequately. The study by J.Mosey [22] was conduced using a modified version of the software package Car-Parrinello ab initio molecular dynamics(AIMD) and was successful in incorporating the pressure conditions. In the simulations compression and decompression cycles were used. The pressure was increased at a rate of 2, 5 GPa/ps and 10, 0 GPa/ps from 0,25 to 2,5, 4,0, 7,0, and 32,5 GPa-. When this pressure was reached, the pressure was released at the same rates. It was reported that during the initial compression, the Zn atom fluctuates between di-coordinate, tricoordinate and tetra-coordinate bonding arrangements. At a pressure of about 6 GPa a seesaw coordiation geometry [22] is taken by the Zn sites. At this pressure, the bonding arrangement is irreversibly changed and a cross-linked structure could be seen even after pressure release. The density of the cross-linked products was depending on the maximum pressure value. When the pressure was increased to more than 17 GPa, a highly cross-linked structure took form and Zn went into an even higher coordinate state [22]. The highly cross linked state remains under pressure release down to 7GPa where Zn returns to its tetrahedral geometry. But the structure now consist of a fully connected network dissimilar from the original state, see figure 2.9. This cross-linked network is considered to. Figure 2.9: (A) Initial structure of the Zinc phosphate. (B) At a pressure of slighly above 6GPa the irreversible change of the networks form. (C) At a pressure exceeding 17 GPa a highly cross-linked configuration is observed. Image source: [22]. increase the bulk modulus and shear modulus of the films and this is possibly the main reason for the wear inhibition..

(17) 2.5 Cross-linkage in experiments. 2.5. 11. Cross-linkage in experiments. Despite great efforts, the cross-linkage of ZDDP has not been illustrated by experiment. Two great articles that have been published with a clear aim to investigate cross-linkage are [23] and [5]. As the cross-linkage is predicted to be caused by an increased connectivity of Zinc or other metal cations in MDDP, it is more than likely that this high pressure effect would induce changes in the vibrational modes of the ZDDP. Therefore [5] uses IR-spectroscopy as a tool to investigate the ZDDP in a diamond anvil cell. The ZDDP they use is industrially purchased and no further details are given in the article. The pressure conditions are non hydrostatic, as the solid form of ZDDP is used. Not unlike in the simulations described in section 2.4 the pressure is first increased, with Raman information collected at different pre-determined pressure steps and the same procedure is followed during pressure release. The pure pressure effect is first investigated and pressures up to 21, 2 GPa are reported. The conclusions in this study from these measurements state that there are no large structural changes of the ZDDP. For investigating the high pressure and temperature effect, the pressure is initially increased to 15 GPa before heating. After that no substantial change in the IR-spectrum is seen up to 17 GPa and 200℃. At 225℃there are changes in the spectrum that are also permanent during decompression. From this stability of ZDDP under cold compression as well as some exposure of high temperature is concluded and the changes occurring at 225℃are deemed unachievable during normal tribological conditions. As we can see, cross-linkage is not observed of the pure ZDDP. Verification of this result by Raman spectroscopy would be useful as a compliment. Another great piece of work clearly aimed at supporting the cross-linkage theory, this time of decomposition products of ZDDP, [23], investigates in a similar fashion the proposed decomposition product of ZDDP, α-Zn3 (PO4 )2 , 99,999% pure and hydrogenated Zinc phosphate Zn(H2 PO4 )2 . In the experimental part of the article, the different molecules are investigated in situ by Raman spectroscopy. The pressure is increased and decreased in the same way as in [5]. Lower maximum pressures are used and irreversible changes of the α-Zn3 (PO4 )2 system are observed at 6, 6 GPa. Similar changes for the Zn(H2 PO4 )2 are observed at even lower pressure. Changes occurring at lower pressure are explained by mobile hydrogen that causes large local stresses. The change is seen as a broadening of the peak at 960 cm−1 along with the loss of low frequency peaks. This result is interpreted as cross-linkage. That means that just as in the simulation discussed in section 2.4, there is cross-linkage of something similar to ZDDP. But why can’t we observe the cross-linkage in the ZDDP? [23] suggests that there is some run-in, that the ZDDP will only start to cross-link after being decomposed in an engine system. In other words, in fact it is the decomposition products of ZDDP that may cross-link. One of the points of [23] is that experiments under more realistic tribological experiments are needed in order to show that cross-linkage of ZDDP is really happening. Is it possible that the incredibly advantageous method of high pressure, insitu Raman spectroscopy can be used to observe cross-linkage of the decomposed ZDDP?.

(18) 12. 2.6. Theory. Unsolved problems. Observations of tribofilms forming under low temperature conditions rise question about the influence of pressure on ZDDP. There are at least two possibilities. Maybe thermal degradation (exchange of O/S) happens at low temperatures in bulk solution where the degradation energy comes from the sliding. Otherwise the degradation is driven by rubbing processes, molecular strain and exoelectron or free surface catalysis. If in-contact analysis and observation of chemical intermediates are possible this would be a good method to observe the behaviour of ZDDP. Raman spectroscopy, which has a high sensitivity for the carbon-sulphur bond3 , should be a good tool to use. There are still questions how ZDDP both demand high pressure and temperature conditions and protect the surfaces from its effect. It seems like ZDDP is versatile indeed. Furthermore the chemical mechanisms of formation of the tribofilm is not sufficiently examined. Understanding of the reacting species from ZDDP to the formation of tribofilm is also desirable. All kinetics are important in order to understand ZDDP so well as to be able to replace it with an equally effective additive.. 2.7. Raman spectroscopy. Raman spectroscopy utilizes the Raman effect. The Raman effect was first discovered by Sir Chandrasekhra Venkata Raman in 1928. He showed, using white light as excitation source, a telescope as collector and his eye as a detector that some reflected light on a sample was reflected with a different wavelength. The phenomenon was named Raman scattering. With improvement in instrumentation Raman spectroscopy has become an important means for scientific analysis. Today monochromatic laser light is used as excitations source and advanced monochromators are used as detectors. The Raman effect is the response of a molecule to a photon. When the energy that is charged in the molecular vibration is released, it will sometimes fall into a state which has a different energy from the original one. If the energy is higher than the light of the incoming ray it is called stokes scattering and if the energy is lower than that of the incoming light it is called anti-stokes scattering. If the energy is the same it is called elastic scattering. Figure 2.10 shows a schematic sketch of the Raman Effect. If the intensity of the reflected light is plotted against its wavelength the result will be a map of the Raman-active vibrations in the sample called a spectrum. This spectrum gives information not only about which molecules of the sample but can also show properties such as temperature, pressure, conductivity and more of a well characterized sample. Figure 2.11 shows a typical spectrum.. 3 See. [25] for more information about Raman spectroscopy in general and which bonds make a strong signal in Raman spectroscopy..

(19) 2.7 Raman spectroscopy. 13. Figure 2.10: This is a simple illustration of the Raman Effect. (1) An incoming photon hits a Raman active vibrational bond. (2) The bond is excited to a virtual state. Energy is released when the excited state is de-excited and the molecule starts to vibrate in one of its vibrational modes again. (3) Depending on if the new mode has a higher energy or a lower energy or is the same as before excitation, the energy of the emitted photon will vary.. Figure 2.11: A typical raman spectrum. We can see the peaks both of antistokes and stokes scattered light. Each peak corresponds to light emitted by a Raman-active bond. Image source: [26].

(20) 14. Theory.

(21) Chapter 3. Methods and equipment 3.1 3.1.1. Raman Spectroscopy Exitation Sources. It is common practice to use lasers as excitation sources for Raman spectroscopy. There are 2 main reasons for this. The first is that lasers have small rays that don’t need so much expensive optical equipment to analyse small samples. The second is that they operate at single wavelengths. In this study three different lasers were used as excitation sources to analyse the efficiency of the different wavelengths as far as studying this system is concerned. The first laser used was a green laser with an output wavelength of 532 nm corresponding to a photon energy of 2,33 eV. The second laser used was a He-Ne red laser operating at a wavelength of 632,8 nm. This corresponds to the energy 1,96 eV. The third laser was an infrared laser of wavelength 830 nm.. 3.1.2. Confocal Raman Microscope. In confocal microscopy the illuminating and the collected light takes the same path in the microscope near the sample and is separated by a beam splitter. After the beam splitter the collected light is filtered through a small pinhole that makes sure only the image focal plane will hit the detector. This will exclude disturbing light from lens flare or out of focus light. The effect is much higher contrast and if the pinhole size is correctly chosen a small increase in resolution. In Raman microscopy both are important. Confocal Raman microscopy is the most efficient means to study Raman scattering today. Figure 3.1 shows the confocal Raman microscope CRM 200 from Witec focus innovations that has been used with the 632,8 nm laser and 532 nm laser. Figure 3.2 shows the inside of the CRM 200 and the beam path both for the incoming Laser light as well as using the CRM 200 as an ordinary microscope. The green ray shows the laser path. What can be seen is that the excitation laser (12) enters the microscope by a polarization maintaining single mode fibre (14) that is coupled in both ends by fibre couplers (13 and 15). The Achromatic lens system (16) makes the laser beam enter the microscope in parallel. A beam holographic band pass filter (7) is used to reflect the ray down through a beam splitter(3) and the objective 15.

(22) 16. Methods and equipment. Figure 3.1: The Confocal Raman microscope. Image source: [26].

(23) 3.1 Raman Spectroscopy. 17. (4) focuses the laser on the sample (5) resting on a piezoelectric table(6). The Raman scattered (and fluorescence) light is then scattered a 180◦ angle to go the same path back all the way to the laser blocking filter (8). This filter is set for the wavelength of the laser and will protect the sensitive photon collecting equipment from the relatively strong elastic scattering of the laser. After this weakening of the laser line the signal will continue to the beam splitter (10) which in practice works like a mirror leading the light to the correct collector depending on application. A multimode fibre (19) connects the microscope to the Spectrometer. The Signal is finally collected at either the CCD camera (22) or the photon counting apd-system (23). The flip mirror (21) directs the light to the collector.. Figure 3.2: The beam path of the CRM200. Image source: [26] Several means for collecting spectra using CRM 200 are available. The ones.

(24) 18. Methods and equipment. used in this project are: • Single spectrum - Records the Raman spectrum at one single spot. • Fast image spectrum - The APD collects and counts photons of a specified wavelength at a larger sample area. With this method we can scan an area for characteristic peaks for example of ruby. • Image spectrum - A full spectrum is taken for each pixel at a great amount of points of the sample.. 3.2. High pressure. Normal high pressure studies will usually refer to studies at any pressure over around 100 MPa up to the measured limit. There are examples of studies at thousands of GPa of shock pressure. High pressure studies are interesting because of the great influence pressure has on all materials around us. Examples of pressure effects are phase transitions, polymeric phase changes, compression, changes in magnetic, optical, electrical, viscous, chemical properties and changes of the strength of most solids. There are two main branches of high pressure, namely static pressure and shock pressure. The later uses shockwaves produced by explosives or projectile impact to produce pressures up to 1 TPa. The highest static pressures achieved are in Diamond anvil cells, where pressures as high as 400 GPa are possible for extremely small samples. Information of high pressure physics can be found at [27].. 3.2.1. Membrane Diamond Anvil Cell(DAC). The diamond anvil cell operates by squeezing a gasket containing a sample between two diamonds. The extreme high pressures that can be achieved are due to the hardness of the diamonds. Failure of the diamonds will occur if the diamonds come into contact with one another while there is still a force pressing them towards each other. A basic Diamond anvil cell operates by screws that clamp the diamonds together. The downside of this is that pressure steps are difficult to control and pressure steps below 1GPa are hard to achieve. Therefore a membrane diamond anvil cell is a good alternative. This membrane works like a spring which expands when exposed to pressure. With small changes in pressure within the membrane, precise pressure control in steps as small as 0,1 GPa are achievable.. 3.2.2. Diamond anvil and gasket. Diamonds are used in the DAC because of their hardness, and because of their transparency to radiation at a wide range of frequencies. To achieve a safe pressure distribution the diamonds need to be aligned perfectly horizontal and opposite to one another. The diamond culets, the tip of the diamond in between which the sample is squeezed, are as small as 450 µm on our image. Alignment is done using screws. To assure that the alignment is perfect interference fringes that are seen are eliminated. figure 3.3 shows the diamonds before alignment. The design and quality of the diamond anvils vary. The diamond type for Raman.

(25) 3.2 High pressure. 19. spectroscopy are ultra low fluorescence type I diamonds. For high pressures the diamond culet is sometimes bevelled to improve the pressure distribution. The diamonds we use have been bevelled to make higher pressure possible reducing the culet face from 500 µm to 450 µm.. Figure 3.3: The diamond anvils during alignment. The tip of the diamonds facing each other here is only 500µm.. The sample chamber of the Diamond anvil cell is called the gasket. The gasket is commonly made of hard rolled austenitic stainless steel for measurements up to 10 GPa. Other possibilities are to use rhenium (for temperatures over 600℃ or over 30GPa) or inconel. The gasket has a hole in it called sample chamber which contains the sample that is to be studied. Its diameter is typically 1/3 ( [28], [29]) of the culet diameter, so in our case a 150 µm sample chamber diameter is appropriate. The initial thickness of the gasket is about 200 µm. Before experiment the gasket is squeezed between the diamonds during pre-indentation to about 60 µm. When this pre-indentation has been made the gasket has a crater-like sink in the middle where the diamonds have squeezed it. It is advantageous both because the centring of the sample chamber will be easier and because the metal forms a protective ring around the diamonds. In the exact middle of this the sample chamber is dug using the MHM20 Spark eroder from BETSA technologies. The equipment is very precise and can drill the 150 µm hole with a precision of a few µm. When the sample chamber has been drilled the sample is placed in the chamber. Depending on the required properties of the sample some pressure transmitting medium can be used. The properties of this medium should be a high solidification pressure to obtain hydrostatic pressure. Commonly used is a Methanol/Ethanol 4:1 solution or cryogenic gas. Hydrostaticity is maintained up to 10GPa for Methanol/Ethanol mixture and up to 90 kbar for Argon. Depending on the pressure transmitting media the gasket hole diameter may vary (some media expand while some contract under pressure)..

(26) 20. Methods and equipment. Figure 3.4: Some of the dimensions in the heart of the Diamond anvil cell.. 3.2.3. Pressure measurement in the DAC. The pressure in the DAC can be measured using the ruby fluorescence according to the method of Forman and his coworkers [30]. the Cr3+ -doped Al2 O3 emits a strong signal of light that peak at two different wavelengths. The two lines that take shape are called R1 and R2 . The wavelength depends on the pressure of the environment. Equation 3.1 shows the equation that derives the pressure from the measured wavelength of the fluorescence. " # B λ A −1 (3.1) P = B λ0 In equation 3.1 λ is the measured wavelength of the ruby R1 line, λ0 is the wavelength at ambient pressure and ambient temperature while A = 1904 and B = 5 are least-squares-fit parameters.. 3.3. High Temperature. To analyse the effect of temperature on hexadecane with a 5wt% concentration of ZDDP in it samples were prepared by heating for different time at well chosen temperatures according to the observed decomposition temperatures of [31], [4] and the summary of thermal effects presented in [7]. The presented decomposition temperatures were 130-230℃, so we choose temperatures within and around this temperature frame. A few drops of the Hexadecane/ZDDP solution were placed in a small glass tube filled with air. After this the glass tube with a lid on it was inserted into a pre-heated oven for one hour. Finally the sample was removed from heat and allowed to cool in room temperature. The temperatures and times are listed in table 3.1. In order to study the effect of temperature and pressure in combination we would use the same diamond anvil cell as for the pure high pressure measurements. Added to the setup is a small heater (See figure 3.5) that is connected to a programmable voltage source. The heater is placed inside of the diamond anvil cell. The cables are connected with the diamond anvil cell by heat resistant ceramic tubing. When heated to 1000℃, achievable with a voltage of 20 Volts.

(27) 3.4 Experimental details Temperature(◦ C) 100 150 200. 21 Time(minutes) 60 60 60. Table 3.1: The temperatures and time chosen for examining the effect of temperature on hexadecane/ZDDP solution the exterior of the diamond anvil cell reaches temperature up to 400℃. This is dangerous to the optical equipment of the Raman microscope, the lens operates only up to 40℃. To protect the sensitive equipment as well as the user a heat insulating box has been designed by Master of Science Magnus Grennvall, see [32]. The heat insulation box protects the surrounding environment as well as the operator from heat.. Figure 3.5: The small heater used in the diamond anvil cell.. 3.4 3.4.1. Experimental details The Samples. We have done Raman spectroscopy of hexadecane (99,5% pure), diethyleneglycoldibutylether(99% pure), ZDDP (50% pure and 99% pure1 for NMR data on our substances) as well as solutions of ZDDP and the solvents. We have also done spectroscopy of methanol/ethanol 4:1 mixture and of solution between methanol, ethanol and ZDDP, saturated with ZDDP. In hexadecane we used 3 different concentrations of ZDDP in solution. They were 2wt%, 5wt% as well as saturated solution. As a complement we also studied the precipitate of hexadecane/ZDDP solution. In diethyleneglycoldibutylether we used 2wt% 4wt% and 6wt% solution. The solution was mixed using a scale. All samples for ambient conditions were put on glass slides using glass or plastic pipettes. The saturated solutions were mixed and blended until precipitate was observed. Once the precipitate was resting on the bottom of container, a syringe was used to collect the solution. The containers we used for all our fluid samples were small glass tubes with air-tight lids that were filled to about half with hexadecane/ZDDP solution. The 1 See. appendix 5.

(28) 22. Methods and equipment. hexadecane solution was studied with the 532 nm laser, the 830 nm laser and the 632,8 nm laser. The diethyleneglycoldibutylether solution was investigated with the 632,8 nm laser. We also looked at one metal surface that had been used in a pin-on-disk test and one surface that had been used in a three-ball-test. The surface had been lubricated with 2wt% ZDDP in hexadecane/ZDDP solution.. 3.4.2. excitation laser details. An Olympus 20x objective focused the laser to a spherical spot with a diameter of 15µm for all measurements with the 532 nm laser and 632,8 nm laser. The power was measured on stage using a handheld laser meter with a 5% calibrated accuracy from 400-1060 nm in the power range 0, 5µW - 1 W. Corresponding power density to laser power can be seen in table 3.22 . The first measurements Power output(mW) 0,7 1,7 2,0 2,3 2,7 4,0 4,3 4,6 6,4 6,7 6,2 7,6 8,1. Power density(kW/cm2 ) 0,40 0,96 1,13 1,3 1,5 2,3 2,5 2,6 3,6 3,8 3,5 4,3 4,6. Sample ZDDP Hexadecane Solution Hexsol ZDDP Methanol/Ethanol ZDDP, Tribofilm Methanol/Ethanol Hexadecane Heated Hexsol Diethylene. . . Hexadecane and Hexsol Hexsol. Table 3.2: The power and corresponding power density for the measurements with the 532 nm laser (above the line) and 632,8 nm laser (below the line). were done with the 532 nm laser. On hexadecane a laser power of 1, 7mW was used. On Hexadecane/ZDDP solution a power of 2, 0 mW was used. On ZDDP a power of 0, 7 mW was used. For all measurements with the 830 nm laser a power of 10mW was used. Measurements with the 632,8 nm laser were done with many different laser powers. Eventually a standard around 7 mW was established for hexadecane/ZDDP solution. For hexadecane laser power of 6, 4 mW and 7, 6 mW was used. For measurements on ZDDP a laser power of 2, 5 − 2, 7 mW was used with the 600 grating and a power of 4, 3 mW with the 1800 grating. For diethyleneglycoldibutylether as well as for the diethyleneglycoldibutylether/ZDDP solution a power of 6, 2 mW was used. for the saturated hexadecane/ZDDP solution, the powers 6,4 mW and 8, 1 mW were used. For hexadecane/ZDDP solution of 5 wt% concentration 7, 6 mW power was used. For hexadecane/ZDDP solution of 2 wt% a power of 2, 3 mW was used. For methanol/ethanol/ZDDP solution 2 In. this table “Diethylene...” refers to diethyleneglycoldibutylether and “Hexsol” refers to hexadecane/ZDDP solution.

(29) 3.4 Experimental details. 23. and precipitate laser power from 4, 0 mW to 4, 6 mW was used. On the heated hexadecane/ZDDP solution samples, a power between 6, 2 mW and 6, 7 mW was used. For the steel3 sample with a tribofilm on it 4, 2 mW was used. After that results are displayed and discussed. The thesis finally ends with the author’s conclusions.. 3 The. steel on the steel samples was 100Cr6 bearing steel.

(30) 24. Methods and equipment.

(31) Chapter 4. Results and discussion The investigation that lay behind this master thesis is divided in to two main parts. The first part includes characterization of ZDDP under ambient conditions as well as ZDDP in three different solvents. 3 different lasers are used for this to see which laser is optimal for proceeding with to the second part. In the first part we also study the raman spectrum of a tribofilm synthesized by hexadecane/ZDDP solution. The second part is where we analyse the effect of temperature on ZDDP.. 4.1. Pure ZDDP substances. In our initial investigation, a ZDDP substance which was considered pure was used, this substance we refer to as active ZDDP. NMR analysis of our samples showed that this was not the case, in fact this sample had reacted and was in itself a complex system. It may be that this substance is still considered pure in the industry, just consisting of different forms of ZDDP and its reaction products. We will prefer to study a substance which is more simple and therefore, we have recieved a substance which we call chemically synthesized ZDDP. This substance consist of ZDDP only in its crystal form. This form is the form of dimeric ZDDP. The R-group is a tert-butyl group. Figure .10 of appendix 5 shows the nuclear magnetic resonanceresults from the chemically synthesized ZDDP. Figure .11 shows the nuclear magnetic resonance results from active ZDDP. The advantage of studying active ZDDP is that in running conditions ZDDP certainly evolves (thermal decomposition, pressure induced changes and so on) and this may cause a difference compared to the chemically synthesized ZDDP. In [23] there is a discussion about this concerning the results of [5]. However, for clear interpretation, chemically synthesized ZDDP is the first choice and active ZDDP the seconde concerning high pressure studies. The figures in 5 are marked with explanations of the different peaks, and the meaning is explained in the appendix. When we knew about the difference in the samples by NMR we have a good advantage to study 2 different samples and compare them. If this is done we can identify the “real” peaks of ZDDP in the active ZDDP. Several spectra were 25.

(32) 26. Results and discussion. Chemicaly pure ZDDP, Spot 1 Chemicaly pure ZDDP, Spot 2 Chemicaly pure ZDDP, Spot 3 Chemicaly pure ZDDP, Spot 4 Activated/Reacted ZDDP Spot 1 Activated/Reacted ZDDP Spot 2. Intensity. Activated/Reacted ZDDP Spot 3. 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 Raman shift (1/cm). Figure 4.1: Spectra of different spots of active ZDDP and chemically synthesized ZDDP samples.. taken of both substances and a great variation in intensity was found in each substance. That is even between two different spots in the same sample there is a great difference. This is normal concerning the samples may not be completely homogenous. What is more interesting though is the differences between the spectra of the chemically synthesized ZDDP and the active ZDDP. Comparing 3 different spots of the active ZDDP with 4 different spots of chemically synthesized ZDDP (all choosen for their relatively high intensity) some general differenses can be observed. Figure 4.1 displays the spectra. The first two things that we noticed was the peak1 at 667 cm−1 and the higher fluorescence background of the active ZDDP sample. When we look closer on the spectra there are several “small peaks” that exist only in the active ZDDP. This can be seen in figure 4.2 where the small peaks are clearly seen at all points of the active ZDDP sample while in neither of the points of the chemically synthesized ZDDP. The smallness of these peaks may lead the reader to suggest that the peaks are background noise, or some shakiness caused by humidity in the sample. This could be the case but such shakiness would occur randomly, while the peaks of this sample appear at the same wave numbers, which can be clearly seen in figure 4.2. This instead indicates that the peaks correspond to real vibrations of the decomposition products. The wave numbers of the extra peaks are listed in table 4.1. 1 The. author would assign this peak to a P=S (double) bond. The assignment of this peak was based on [33].

(33) 4.1 Pure ZDDP substances. 27. Chemicaly pure ZDDP Spot1 Chemicaly pure ZDDP Spot2 Chemicaly pure ZDDP Spot3 Chemicaly pure ZDDP Spot4 Activated/Reacted ZDDP Spot1 Activated/Reacted ZDDP Spot2. Intensity. Activated/Reacted ZDDP Spot3. 1220. 1240. 1260. 1280. 1300. 1320. 1340. 1360. 1380. Raman shift (1/cm). Figure 4.2: Spectra of different spots of active ZDDP and chemically synthesized ZDDP samples in the spectral region 1200 − 1400 cm−1 .. Raman shift (cm−1 ) 241 368 425 437 714 745 754 765 978 1000 1203 1225 1278 1312 1318 1328 1362 1384 1394 1408 1482 1492 1507 1516 1526 Table 4.1: The wave numbers of tiny Raman peaks that indicate the relative decomposition of active ZDDP compared to chemically synthesized ZDDP..

(34) 28. Results and discussion. ZDDP saturated solution 5 wt% solution 2 wt% solution. Intensity. Hexadecane. 700. 800. 900. 1000. 1100. 1200. -1. Raman shift [cm ]. Figure 4.3: Spectra of ZDDP, Hexadecane and hexadecane/ZDDP solution of different concentration.. 4.2. Study of suitable solvents for high pressure experiments on ZDDP. The main goal is to study ZDDP and its behaviour. In real application ZDDP is only used as an additive of small relative weight compared to the base oil. Therefore, to get similar effects that are observed in real systems we should have some solvent and study ZDDP in an environment resembling to a motoroil. Possible choices are synthetic oils, mineral oils, vegetable oils as well as animal oils. To observe ZDDP in the solution we want minimal interference from the base oil on the system. Simple oil with only one molecule would be a suitable candidate. We have chosen to use hexadecane (a type of paraffin) and diethyleneglycoldibutylether as our solvents in the characterisation phase. The two solvents differ in polarity, hexadecane is less polar. The study initially examines ZDDP and Solvent separately and thereafter proceeds with examination of the solution. This order of examination is continuously used in this study. The straightforward reason is that we are mostly interested in the behaviour of ZDDP in the solution and will therefore follow the difference between solution behaviour and pure base oil behaviour.. 4.2.1. Hexadecane/ZDDP solution. When we are familiar with the peaks of both hexadecane and ZDDP it is time to start studying the solution. We started this work with hexadecane as a solvent. We started out with 2wt% ZDDP in the solution because the same wt% was used in a previous study. To see the influence of ZDDP on the solution more clearly.

(35) 4.2 Study of suitable solvents for high pressure experiments on ZDDP. 29. Intensity. ZDDP Hexadecane subtracted from 5% solution Hexadecane subtracted from saturated solution. 700. 800. 900. 1000. 1100. 1200. Raman shift [cm ] -1 . Figure 4.4: Spectrum of ZDDP and the resulting spectra from subtracting hexadecane from hexadecane/ZDDP solution of 2 different concentrations. higher wt%ages were also used. Figure 4.3 shows the increasing concentration of ZDDP in the spectral region where ZDDP activity can be most clearly seen. In this image we can see how weak the very strongest ZDDP peak occurs in solution. But with concentration increase the intensity of the peak also increases. We can draw two conclusions from this; the first is that ZDDP is a weaker Raman scattered than hexadecane (at least for the 632,8 nm laser that we use). The second conclusion is that we can only follow a few of the strongest ZDDP peaks in the hexadecane/ZDDP solution. So the important question remaining is, are there peaks we can follow important to our analysis of ZDDP behaviour? As it seems the peak around 850cm−1 change is character by its left shoulder and the peak around 950cm−1 increase in intensity with increasing concentration. The exact wave numbers of the concentration depending peak has been decided to 832cm−1 and 962cm−1 using peak fit convolution. The corresponding vibrations could be a vibration of the Zn-S2 for the bond at 832cm−1 and vibrations from CH2 , CH3 , P-O, C-O or C-C for the bond at 962cm−1 according to our section 4.3.2. Too see the peaks more clearly a decision to try and subtract the spectrum of hexadecane from the spectrum of hexadecane/ZDDP solution point-by-point was made. The result we hoped for was a spectrum only ZDDP inside the hexadecane/ZDDP solution. This can be done under the assumption that there is no chemical reaction between the solvent and the solute, that they exist separately and that they don’t effect the vibrations of one another. To do a subtraction there are three steps. First we subtract the background from both solution and solvent. After that we will rescale the two spectra which have different absolute intensity after the highest peak available. The final step.

(36) 30. Results and discussion. Intensity. ZDDP 6% ZDDP in C12H26O3 4% ZDDP in C12H26O3 2% ZDDP in C12H26O3 C12H26O3. 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300 -1. Raman shift [cm ] . Figure 4.5: Spectra from diethyleneglycoldibutyletherand diethyleneglycoldibutylether/ZDDP solution of different concentrations. is to subtract the y values of the solvent from the y values of the solution at each x value. The result of the subtraction can be seen in figure 4.4. It is also desirable to achieve a stronger signal from the ZDDP inside of the hexadecane/ZDDP solution. In this study we proceeded by seeing if the peaks could be seen more clearly with another solvent, namely, diethyleneglycoldibutylether/ZDDP solution. We also tried to mix ZDDP with a methanol: ethanol 4:1 solution because of the superior high pressure properties of the methanol/ethanol solution.. 4.2.2. Diethyleneglycoldibutylether/ZDDP solution. The solution of diethyleneglycoldibutylether and ZDDP was examined. The results from measurements of different concentrations of diethyleneglycoldibutylether can be seen in figure 4.5. As we can see in this figure we have a similar (or worse) problem as with hexadecane/ZDDP solution except that diethyleneglycoldibutylether has replaced hexadecane. Therefore the diethyleneglycoldibutylether/ZDDP solution was decided to be equal to hexadecane/ZDDP solution as far as studying ZDDP is concerned. We decided hexadecane had an advantage in the previous study on hexadecane behaviour under high pressure [34].. 4.2.3. Methanol/ethanol/ZDDP solution. To go into high pressure a medium with a high solidification pressure (usually combined with a low melting point) is desirable. This is because structural changes occur during phase transformation, and they will both have effect on.

(37) 4.3 ZDDP in hexadecane. 31. Intensity. ZDDP Methanol/Ethanol/ZDDP solution Methanol/Ethanol Solution. 800. 1000. 1200. Raman shift [cm ] -1. Figure 4.6: Spectra of ZDDP, methanol/ethanol 4:1 solution and methanol/ethanol/ZDDP solution. The new peaks in the methanol/ethanol/ZDDP solution indicate a chemical reaction. the raman spectra and induce an un-even pressure distribution during high pressure runs. Therefore it is common when studying small particles that they are put in a pressure transmitting medium. Commonly used is a solution between methanol and ethanol, which is not a strong Raman scatterer and has a high solidification pressure. In this study we want to study ZDDP under high pressure and even in solution. Therefore it would be good if we could use a methanol/ethanol and immediately solve ZDDP in this. However alcohol and ZDDP have a strong chemical connection and there may be chemical reactions between them. We studied the system of methanol/ethanol/ZDDP solution to see if this was the case or if we could use the new system for our study. The measurements were done on the methanol: ethanol 4:1 solution, and the methanol/ethanol/ZDDP solution. The resulting spectra are shown in figure 4.6. The spectrum of pure ZDDP is also present for comparison. In the image we can se several new peaks in the solution, which are not present in neither of the two pure substances. This was interpreted as an indication of a chemical reaction and therefore methanol/ethanol/ZDDP solution as solvent is not recommended for high pressure studies.. 4.3. ZDDP in hexadecane. After investigating solvents, hexadecane/ZDDP solution was found the strongest candidate to study ZDDP in solution under high pressure. In this section we decide upon what excitation source is appropriate as well as analyse the Raman spectrum of each component..

(38) 32. 4.3.1. Results and discussion. Excitation source determination. Green laser measurements The logical choice for first investigation fall on the laser with the lower wavelength, in our case the 532 nm laser The intensity of the signal is governed by the ν 4 rule [25] and is supposedly higher with lower wavelength under otherwise equal conditions. Our first measurements were done with the 532 nm laser A huge fluorescence peak was discovered in the Stokes region near the laser line when the solution of hexadecane and ZDDP was investigated. See figure 4.7. This was interpreted as using the 532 nm laseras a bad idea and so we continued. Intensity. Hexadecane solution with 5wt% ZDDP (532nm). 400. 800. 1200. 1600. 2000. 2400. 2800. 3200. -1. Raman shift [cm ]. Figure 4.7: Spectrum of hexadecane/ZDDP solution containing 2wt% ZDDP illustrating the flourescence peak seen with the 532 nm laser by trying other lasers. Infrared laser measurement Investigations were also done with the 830 nm laser. The measurements got a similar result as those of the measurements with the 532 nm laser and are therefore presented here. Hopes that a higher excitation wavelength would give less fluorescence because this was the case with an intermediate wavelength were dashed. The raw spectra from all measurements are presented in appendix 5. As we can see both hexadecane and the solution spectra have a fluorescence peak similar to that seen in the measurements with the 532 nm laser. We can also see an exponential background in these spectra. With this exponential background subtracted the final spectra are displayed in figure 4.8. For this reason we do not recommend further investigation with the 830 nm laser on our system, certainly not under high pressure..

(39) 4.3 ZDDP in hexadecane. 33. Hexadecane Hexadecane solution with 5wt% ZDDP. 0. 500. 1000. 1500. 2000. Raman shift 1/cm. Figure 4.8: Spectra of hexadecane and hexadecane/ZDDP solution excited by the 830 nm laser. Red laser measurements Collecting Raman spectra of hexadecane/ZDDP solution with the 632,8 nm laser shows no sign of any strong fluorescence peak near the laser line in the Stokes region of the spectrum. Thus, the 632,8 nm laser is chosen as the best laser we have at our disposal to study the hexadecane/ZDDP solution system. Figure 4.9 shows the spectrum of Hexadecane excited with the red laser.. 4.3.2. Peak assignment. the peaks of ZDDP Before we can analyse ZDDP in solution we should analyse the pure substance of ZDDP to characterize the peaks that give the most useful information about the ZDDP molecule. Therefore, ZDDP spectra were collected with the 632,8 nm laser. Figure 4.10 shows the spectrum of ZDDP. By comparing the spectra of molecules containing the same bonds as those present in ZDDP with our own spectra, we can construct a table with suggested assignment of ZDDP peaks. This is a possible way to find vibrations according to [25]. There is also a theoretical prediction of the Vibrational frequencies of Potassium O,O’-Dibutyldithiophosphate and their assignments to be found in [35]. This molecule is similar to our own and has the same R-group. Table 4.2 shows our assignment of the peaks seen in the ZDDP spectrum. To find the exact positions of ZDDP peaks we used software for peakfit convolution. Peak fitted data of spectra from ambient conditions measurements using the 632,8 nm laser can be seen in appendix 5..

(40) 34. Results and discussion. Intensity. Hexadecane. 400. 800. 1200. 1600. 2000. 2400. 2800. -1. Raman shift [cm ]. Figure 4.9: Spectrum of the pure hexadecane excited by the 632,8 nm laser.. Intensity. ZDDP. 400. 800. 1200. 1600. 2000. 2400. 2800. -1. Raman shift [cm ]. Figure 4.10: Spectrum of ZDDP excited by the 632,8 nm laser..

(41) 4.4 Raman study of the tribofilm Wave number (cm−1 ) 316 486 555 587 670 790 824 964 1120-1300 2800-3000. 35. Assignment ν4 mode of PO4 S-S P-S, (P-S2 ), P-O or CH2 ν2 mode of PO4 or C-C P-S or P-O P-S or P-O ZnO2 CH2 , CH3 , P-O, C-O or C-C C-C, C-O, CH2 and CH3 CH, CH2 and CH3. Reference [36] [33] [33], [35] [37], [38] [35] [35] [33] [35] [34], [35] [34], [37]. Table 4.2: Suggested assignment of ZDDP peaks. Wave number (cm−1 ) 1061,9 1079,3 1130,1 1413 2800-3000. Bond C-C C-C C-C CH2 CH. Vibration Streching Streching Streching Bending Stretching. Reference [34] [34] [34] [34] [34], [37]. Table 4.3: Suggested assignment of hexadecane peaks. As can be seen in table 4.2 There are a lot of assignments for each peak. This ambiguity indicates a flaw of the use of Raman spectroscopy to study ZDDP. The peak at 824 cm−1 which in general is the strongest peak, however, has been assigned to the Zin-oxygen double bond. This bond is not present in the ZDDP molecule and the authors instead suggest that the peak originates from the Zinc-sulphur double bond. Concerning cross-linkage of the pure ZDDP substance this bond would be crucial and we would expect a clear change of this peak in the Raman spectrum if there would be cross-linkage. The peaks of hexadecane A similar investigation was done on hexadecane. Figure 4.9 shows the spectrum and table 4.3 shows our assignment of the peaks. As can be seen in this table, great help to assign the peaks of the spectrum to vibraitonal modes of molecular bonds come from [34]. This study is a high pressure study of hexadecane previously mentioned.. 4.4. Raman study of the tribofilm. Now when we have seen what laser is best to use as an excitation source and which solvent would be most appropriate to study with ZDDP, thoughts turn towards step 2. Before doing high pressure studies in the diamond anvil cell it is useful to identify the tribofilm spectrum. At our disposal we have a Steel ball which has been used in a three ball test. For this test hexadecane/ZDDP solution with a ZDDP wt% of 2 was used. The resulting metal surface was covered.

(42) 36. Results and discussion. Figure 4.11: Microscope image of a tribofilm on a steel ball. The lubricant to used for creating the tribofilm was hexadecane/ZDDP solution.. 4000. Tribofilm on Steel Ball. 3500. Intensity. 3000 2500 2000 1500 1000 500 0 600. 800. 1000. 1200. 1400. 1600. 1800. 2000. Raman shift (1/cm). Figure 4.12: Spectrum of a tribofilm steel ball. The lubricant used to induce the film was hexadecane/ZDDP solution..

(43) 4.5 High pressure and high temperature Wave number (cm−1 ) 641 682 813 976 1080 1212 1280 1442 1534. Bond P=S Fe3 O4 PO3 S or ZnO2 PO4 PO4 ,-PO3 or C-C PO3 or (Zn3 (PO3 )2 )n Zn4 P6 O14 -glass CH2 CH2. 37 Reference [33] [38] [38] [34], [38] [23] [23] [34] [34]. Table 4.4: Suggested molecular pedigree of peaks in the tribofilm spectrum. with a tribofilm. Figure 4.11 shows a microscope image of this “triboflilm”, on the steel sample. In this image it is hard to tell if there is a tribofilm, but Raman spectroscopy reveals several tribofilm products. Figure 4.12 shows the Raman spectrum of the steel ball. This Raman spectrum can be used as a reference if we manage to create a tribofilm within the diamond anvil cell. In a similar fashion as before, assignment of the peaks of the spectrum in figure 4.12 has been done. In this case, however, we are not identifying molecular bonds, but rather we are searching for the molecular components of the tribofilm and their most characteristic peaks. Therefore the assignment of peaks are also done to larger molecules. Table 4.4 shows the analysis.. 4.5. High pressure and high temperature. The reference spectra we have collected will be a great guideline for upcoming experiments which has as a goal to show how temperature and pressure is inflicting the ZDDP in solution. We will focus on experiments which give us higher quality data at certain more specific areas, both concerning spectral area and pressure temperature pathways. In this work we choose quality before amount. First we will show the effect of temperature on ZDDP in solution. The temperature-only effect on ZDDP crystal will not be studied in this work. High pressure measurements will be the final step in this master thesis but is suggested in future work.. 4.5.1. Pre-heated sample analysis. As described in chapter 3.3 samples of hexadecane/ZDDP solutionhas been prepared. Spectra from measurements on hexadecane/ZDDP solution heated to 100℃ and 150℃ along with spectra from hexadecane/ZDDP solution that was not heated are displayed in figure 4.13. We can see that the difference in the spectra between heating below and above 130℃ for one hour is small. this may indicate that in hexadecane the decomposition starts at lower temperatures or higher temperatures. It is possible to see some difference between the hexadecane/ZDDP solution that was not heated compared to the other two. This becomes extra clear if.

(44) 38. Results and discussion. ZDDP/Hexadecane Solution ZDDP/Hexadecane Solution heated to 100C. Intnensity. ZDDP/Hexadecane Solution heated to 150C. 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 Raman shift (1/cm). Figure 4.13: Spectra of hexadecane/ZDDP solution, heated to different temperatures.. ZDDP/Hexadecane Solution ZDDP/Hexadecane Solution heated to 100C. Intnensity. ZDDP/Hexadecane Solution heated to 150C. 800 825 850 875 900 925 950 975 1000 1025 1050 1075 1100 Raman shift (1/cm). Figure 4.14: Spectra of hexadecane/ZDDP solution, heated to different temperatures im the spectral region 800 − 1100 cm−1.

(45) 4.6 High pressure measurements plan. 39. ZDDP/Hexadecane Solution heated to 100C. Intensity. ZDDP/Hexadecane Solution heated to 200C. 600. 1200. 1800. Raman shift 1/cm. Figure 4.15: Spectra of hexadecane/ZDDP solution, heated to two different temperatures. we focus on the area shown in figure 4.14. We can see at least 2 clear peaks around 920cm− 1 and also some other different characteristics. We conclude that the change occurs already bellow 100℃. The new peaks could represent the movement of the R-group from oxygen to sulphur in the ZDDP molecule in that case they would originate from the Sulphur-Carbon bond. We also collected Raman data from a sample of hexadecane/ZDDP solution that had been heated to 200℃. The results from this measurement are displayed in figure 4.15. In this sample we can see huge fluorescence background and compared to the sample heated to 100℃ a generally rough characteristic. This may indicate severe decomposition of the sample, with so many molecules involved in the sample the signal from all of them become a mess. We can conclude that going above this temperature will make analysis of hexadecane/ZDDP solution very complicated.. 4.6. High pressure measurements plan. The reference spectra we have collected will be a great guideline for upcoming experiments but the future goal is to see how temperature and pressure is inflicting the ZDDP in solution. There are two ways to proceed. It is possible to focus on experiments which will give us a lot of data at a lot of different pressure-temperature pathways or to focus on experiments which give us higher quality data at certain more specific areas, both concerning spectral area and pressure temperature pathways. It is in other words a choice between quality and amount. We recommend doing highest quality measurements at a lowest possible amount of well motivated temperature/pressure points. See appendix 5 for some discussion about different ways to go about..

(46) 40. Results and discussion. Figure 4.16: The shift of hexadecane peaks under high pressure. We suggest to proceed starting with pure high pressure measurements of pure ZDDP, because there is a need to confirm the results of [5] and get good reference for in-solution spectra. One reason to begin with ZDDP and not hexadecane is the existing study of hexadecane, [34]. This study illustrates the movement of some important peaks under pressure. Figure 4.16 roughly shows the effect of pressure on spectral regions in the spectra of hexadecane and figure 4.17 gives an idea of what might happen to the spectrum of the solution. It is recommended to adapt and confirm the results of [34] after the system that will be used for high pressure hexadecane/ZDDP solutionstudies. More detailed suggestions about initial high pressure investigations are presented in Appendix 5.

(47) 4.6 High pressure measurements plan. 41. Figure 4.17: The peaks around 1000 cm-1 and the band around 1400 cm-1 will change as pressure goes up according to [34]..

(48) 42. Results and discussion.

(49) Chapter 5. Conclusions and Future work When trying to study ZDDP with Raman spectroscopy in a model system several problems have occurred. These problems have been addressed and we have solved them to a point where high pressure studies would be the next step. Two different kinds of ZDDP, namely active ZDDP and chemically synthesized ZDDP were studied. This study helped in identifying vibrations originating from the P=S bond. The study of the different ZDDPs was unintentional and the small differences between the samples in Raman spectroscopy compared to nuclear magnetic resonanceshows the importance of using different methods when conducting this type of study. Concerning fluorescence a 532 nm laser and a 830 nm laser where inferior to using a 632,8 nm laser which does not experience the same fluorescence problems. We have thus seen that a 632,8 nm laser would be the best one at our disposal for studying ZDDP in solution under high pressure. We have found some advantages in using hexadecane instead of diethyleneglycoldibutylether as a solvent for studying this system. The peaks of ZDDP overlap less with hexadecane and can be seen more clearly. To study hexadecane/ZDDP solution under high pressure the study of Wong et.al. , [34] can be useful. We have also managed to extract from the spectra of hexadecane/ZDDP solution peaks that we can follow under high pressure studies, they will tell us in situ about the ZDDP behaviour under high pressure. We conclude that point-by-point subtraction is one good way to observe the ZDDP inside of the hexadecane/ZDDP solution. A chemical reaction between methanol/ethanol/ZDDP solution and active ZDDP has been identified. This reaction indicates that active ZDDP is not appropriate to use for high pressure studies with methanol/ethanol. Chemically synthesized ZDDP remains to be studied in methanol/ethanol solution, too see if chemical reactions can be avoided. The obvious steps for future work are investigations of ZDDP under high pressure as well as ZDDP in solution under high pressure. The first would be to study chemically synthesized ZDDP substance under high pressure. For high pressure studies on ZDDP the spectral region from 300-1300 cm−1 is of great interest, because of the active bonds of ZDDP as well as the peaks that can be seen in the end product1 . 1 We. can see highly interesting peaks in the tribofilm spectrum up to 1280 cm−1. 43.

(50) 44. Conclusions and Future work. After this, hexadecane and hexadecane/ZDDP solution would be first candidates for high pressure studies. To study ZDDP in hydrostatic pressure methanol/ethanol/ZDDP solution could be studied under high pressure, with some risk of chemical reaction interference. In order to study ZDDP or ZDDP in solution under high temperature and high pressure further investigation of the high temperature effect on ZDDP is recommended. The best way to do this would be to study this effect in-situ to exactly determine the reversible and irreversible changes of the ZDDP Raman spectrum due to temperature..

(51) Bibliography [1] Y.-R. Li and G. Pereira, “Studies on zddp thermal film formation by xanes spectroscopy, atomic force microscopy, fib/sem and 31p nmr,” Tribology Letters, vol. 29, no. 1, pp. 11–22, 2008. [2] A.Morina and A. Neville, “Understanding the composition and low friction tribofilm formation/removal in boundary lubrication,” Tribology International, vol. 40, pp. 1696–1704, 2007. [3] J. J., D. Jr., and C. N. Rowe, “The thermal decomposition of metal o,odialkylphophorodithioates,” Journal of organic chemistry, vol. 32, no. 3, pp. 647–653, 1967. [4] W. W. Hanneman and R. S. Porte, “The thermal decomposition of dialkyl phosphates and 0,o-dialkyl dithiophosphates,” Journal of Organic Chemistry, vol. 29, pp. 2996–2998, 1964. [5] Y. S. John S. Tse and Z. Liu, “Effects of temperature and pressure on zddp,” Tribology Letters, vol. 28, pp. 45–49, 2007. [6] S. Jiang, S. Dasgupta, M. Blanco, R. Frazier, E. S. Yamaguchi, Y. Tang, and W. A. Goddard, “Structures, vibrations, and force fields of dithiophosphate wear inhibitors from ab initio quantum chemistry,” Journal of physical chemistry, vol. 100, no. 39, pp. 15760–15769, 1996. [7] H. Spikes, “The history and mechanisms of zddp,” Tribology Letters, vol. 17, no. 3, pp. 469–489, 2004. [8] E. Yamaguchi, “The importance of progressive techniques in tribology research,” Tribology International, vol. 36, pp. 727–732, 2003. [9] T.H.Handley and J. Dean, “O,o’-dialkyl phosphorodithioic acids as extractants for metals,” Analytical Chemestry, vol. 34, no. 10, pp. 1312–1315, 1962. [10] P. Willermet, L. Mahoney, and C. Haas, “Effects of antioxidant reactions on the wear behavior of a zinc dialkyldithiophosphate,” ASLE Trans., vol. 22, no. 4, pp. 301–306, 1979. [11] P. Willermet, L. Mahoney, and C. Bishop, “Lubricant degradation and wear iii. antioxidant reactions and wear behaviour of a zinc dialkyldithiophosphate in a fully formulated lubricant,” ASLE Transactions, vol. 23, no. 3, pp. 225–231, 1980. 45.

No results found