2009:011
M A S T E R ' S T H E S I S
Adsorption of ZDDP on Alpha-Phase Iron Oxide
Yang Zhendong
Luleå University of Technology Master Thesis, Continuation Courses Chemical and Biochemical Engineering Department of Chemical Engineering and Geosciences
Division of Chemistry
1. Introduction Background
Zinc dibutyldithiophosphates (ZDDP) has been synthesized since the late1930s, and now it is widely used as multifunctional lubricating oil additive [1].A simple representation of the structural formula is shown below:
Zn S S
S
S
P P
OR OR OR
OR
ZDDP has been used to protect lubricants against oxidation but its primary use is as an antiwear agent[2-5]. During recent years, some scientists mixed ZDDP with other additives to get better results [6].
The additives adsorb on metal surfaces and form protective films to reduce friction and wear. Because iron and iron oxides are frequently occurring in tribological contacts, many articles study the adsorption of ZDDP on iron, iron oxides and steel. S.Plaza found that iron could adsorb much more ZDDP than iron oxides [7]. They mainly used X-ray fluorescence (XRF) measurements [8], although some of them also used reflection absorption infrared spectroscopy (RAIRS) [9] and photometric titration with p-di -methyl amino phenyl mercury acetate with Michler’s thio-ketone as indicator [10], 65Zn and 14C radioactive labeling [11] and combined FT-IR with XPS for their studies [12].
XRF collected the thermal and tribological film data and from detailed analysis of the spectra it was possible to get the element composition and type of chemical bond that the surface consisted of as well as the thickness of the film. ATR can also give chemical bond information as well as information about functional groups. The residual ZDDP concentration could be measured by infrared spectroscopy, radioactive labeling, and the titration method.
These reported experimental works were mainly concerned with the determination of the amounts adsorbed at different temperature, concentration effects, film thickness and roughness, and film composition. The Mechanisms of ZDDP adsorption on the surface were discussed.
Scope of the thesis
Two different ZDDP solutions were prepared, one with ZDDP in hexadecane, the other one with ZDDP in Diethylene glycol dibutyl ether (DGDE). The concentration of them was increased from 0.1 weight percent to 5 weight percent of ZDDP. Alpha-phase Iron(Ⅲ) Oxide is used for ZDDP adsorption. The scope of the exam work was to first measure the molar absorption coefficient for ZDDP in DGDE as well as in hexadecane by recording ATR spectra of ZDDP in the two solvents as a function of concentration. Subsequently, the absorption coefficients would be used in batch experiments where ZDDP was adsorbed on alpha-phase Iron(III) oxide in order to construct adsorption isotherms at various temperatures. In the latter part of the
experiments, the temperature and time stability of ZDDP should also be examined.
Finally, the aim was to study the adsorption of ZDDP on alpha-phase iron(III) oxide in-situ, using the ATR method. However, the final part of the experiments was not meaningful, since it eventually turned out that the ZDDP chemical, received from SKF, was not pure enough as confirmed at our laboratories by means of NMR analysis.
2. Experimental Section
2.1 Materials and apparatus
Table 1: Materials used in experiment
Material Purity Producer
Alpha-phase Iron(Ⅲ) Oxide 99% Alfa Aesar
ZDDP ?? SKF
n-Hexadecane 99% Acros Organics
Diethylene glycol dibutyl either
99% Acros Organics
Table 2: Apparatus used in experiment
Instrument Model Producer
Shaking water baths Sw22 JULABO Labortechnic
Specific surface area and pore size analysis
ASAP 2010 Chemical instrument AB
Centrifuge Advanti J-E Beckman
Infrared spectrum 2000 FT-IR Perkin Elmer
2.2 Methods
2.2.1 Surface area characterization
The Iron Oxide was heated at 150 for 2 hours, then using N2
adsorption-desorption method samples were characterized. The specific surface area was calculated by the BET method.
2.2.2 Batch adsorption experiment
A certain quantity of ZDDP solid was dissolved in hexadecane and Diethylene glycol dibutyl either, respectively in order to get 5 weight percent solutions.
Subsequently these solutions were diluted to get 2.5%, 2.0%, 1.5%, 1%, 0.4%, and 0.2% solutions.
After that, a number of flasks (at least 6) were filled with iron oxide and one of the ZDDP concentrations according to a certain quantity ratio. This was repeated for all the six ZDDP concentrations. These mixtures were used individually to determine the adsorption isotherms of ZDDP at constant temperatures; 30 , 45 , 60 , and 80 . The flasks were shaken for 3 hours to get equilibrium. Then they were cooled to room temperature and centrifuged at 3000 rpm for 10 minutes. The liquid above the solid
phase was used for infrared spectrum characterization in order to obtain the amount of adsorbed ZDDP from the calibration curve.
In order to study possible change of the samples with time, a certain amount of pure 5% solutions and 5% solution/iron oxide mixtures were stored in the oven at 80 oC and characterized by infrared spectra at different times ranging from 1day to 2 months.
Calibration curves were made by recording infrared spectra of ZDDP solutions using the P-S stretching mode. The amount of ZDDP adsorbed by iron oxide was determined by the difference between spectra of pure solutions and spectra from the liquid above the ZDDP-iron oxide mixtures after adsorption.
3. Results and discussion
3.1 Specific surface area and pore size analysis
According to the BET method, the surface area of the iron oxide particles was 21.06±0.05 m2/g.
We supposed that the iron oxide particles were spheres and according to the following equation, the specific surface area, S, could be used to calculate the average size of the particles:
S=6/(ρ*d) (1)
Where ρ is the density of the solid, which is 5.2 g /cm3 for iron oxide, and d, is the diameter of the particles. According to this calculation the average diameter of the iron oxide particles is about 55 nm.
3.2 The solubility of ZDDP
ZDDP could dissolve in hexadecane completely，but not in diethylene glycol dibutyl ether. No matter if a new or old batch of ZDDP was used, in DGDE it seemed to dissolve under ultrasonic treatment at the beginning, however, a solid phase appeared on the bottom of the bottle one day later. We thought that the ZDDP used might contain some water, so we dissolved the ZDDP in chloroform and recrystalized it. However, that procedure did not help. From the NMR spectra we eventually found out that the ZDDP also contained PO4, PSO3, (DTP)2 and other species.
Although the un-dissolved amount was not that much (only about 3.5 weight percent), the NMR spectra showed that only about 50 % of the solid was pure ZDDP. Two new batches of ZDDP from SKF were slightly more pure according to NMR spctra, but this did not affect the calibration curves.
3.3 Calibration curves
Figure1 shows the FTIR spectra of various ZDDP concentrations in hexadecane. The IR band at 1005 cm-1 could be used for the calibration curve.
Integration from 974 to 1096 cm-1 gave rise to the calibration curve shown in Figure
2.
0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6
940 990
1040 1090
1140 1190
1240
Wavenumber ( cm- 1)
Absorbance
0. 20%
0. 40%
1. 0%
1. 5%
2. 0%
2. 5%
5. 0%
Figure1. Spectra of different weight percent ZDDP in hexadecane
y = 3.594x R2 = 0.9995
0 2 4 6 8 10 12 14 16 18 20
0 1 2 3 4 5 6
ZDDP weight percent in hexadecane
Integrated area
Figure2. Calibration curve for ZDDP in hexadecane
Figure3 shows the infrared spectra of different concentrations of ZDDP in DGDE. The shapes of the absorption bands and the absorption intensity are similar with these ZDDP in hexadecane solutions. Integrating the absorption bands between 975 cm-1 and 1070 cm-1 resulted in the calibration curve shown in Figure 4. Both calibration curves were linear with a R2- value larger than 0.999..
0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6
940 960
980 1000
1020 1040
1060 1080
wavenumber ( cm- 1)
Absorbance 0. 20%
0. 40%
1. 00%
1. 50%
2. 00%
2. 50%
5. 00%
Figure3 Spectra of different weight percent ZDDP in DGDE
y = 3.9302x R2 = 0.9994
0 5 10 15 20 25
0 1 2 3 4 5 6
ZDDP weight percent in DGDE
Integrated area
Figure4 Calibration curve of ZDDP in DGDE
These calibration curves were done using new ZDDP. For ZDDP in DGDE, the calibration curve showed no difference compared to old ZDDP. For ZDDP in hexadecane, the slope of the calibration curve was a little larger for the new ZDDP than for the old one. Both old and new ZDDP could dissolve in DGDE completely but not in hexadecane. There was less solid left by the new ZDDP dissolved in hexadecane than by the old one. It means that the new ZDDP was purer than the old one.
3.4 The stability of samples
For batch experiments at different temperature, the samples were shaked at a certain temperature for 3 hours to reach equilibrium [13]. Solution colors changed after adsorption on iron oxide. For ZDDP in hexadecane solutions containing Fe2O3, the color became green at 5% and 2.5% solutions at 30 , as shown in Figure 5. However, even a 1% solution could become green at 80 . After several weeks, the color was not as clear as before. For ZDDP in DGDE solutions containing Fe2O3 particles, the colors became shining yellow. Without ZDDP there was no color change, so the solvent did not react with Fe2O3. The colored solutions were centrifuged at 3000 rpm for 10 minutes without any change in color. Then they were centrifuged at 21000 rpm for 2 hours without any change in color. This means that the color could not come from dispersed iron oxide particles. Furthermore, there were no differences in recorded spectra, so the color should be due to the reaction between iron oxide and ZDDP.
Figure 5. Different concentrations of ZDDP in hexadecane supernatant after adsorption at 30
℃
The color of these solutions changed much quicker at 80 oC than at room temperature. For ZDDP in hexadecane solutions with dispersed Fe2O3, the color of the supernatant became green after 1 hour and changed to red after 2 months (Fig. 6).
From Figure 7 we can see that the solutions are stable for about 1 week, but the infrared spectral intensity decreases with time.
Figure 6. ZDDP in hexadecane supernatant after adsorption at different times
0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6
980 1000
1020 1040
wavenumber ( cm- 1)
Absorbance
hexadecane cal i br at i on cur ve hexadecane and Fe2O3 1day hexadecane and Fe2O3 1week hexadecane and Fe2O3 2weeks hexadecane and Fe2O3 2mont hs hexadecane 1day
hexadecane 1week hexadecane 2weeks hexadecane 2mont hs
Figure 7. IR spectra of ZDDP recorded at different reaction times
(5% ZDDP in hexadecane).
The supernatants of ZDDP in DGDE with iron oxide dispersed are shown in
figure 8. No color change showed up after the first day, but the supernatant changed to green during the second day and to yellow green the third day. After two months the color had changed to red. The solvent without iron oxide changed to light yellow after several days, but changed to red after 2 months. Of course, not so dark as solutions mentioned above with iron oxide. Figure 9 shows the infrared spectra with and without iron oxide. It seems that the color did not change so much from 1day to 2 weeks for the solvent itself, however, for these solutions after adsorbed by Fe2O3, the peak position of the band shifted to higher wavenumber with time.
Figure 8. ZDDP in DGDE supernatant after adsorption for different times, increasing from left to right.
0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6
940 960
980 1000
1020 1040
1060 1080
wavenumber ( cm- 1)
Absorbance
DGDE cai l br at i on cur ve DGDE 1day
DGDE 1week DGDE 2weeks
DGDE and Fe2O3 1day DGDE and Fe2O3 1week DGDE and Fe2O3 2weeks DGDE and Fe2O3 2mont hs
Figure 9. Spectra recorded after different reaction times (5% ZDDP in DGDE).
One explanation to the observed color changes could be that ZDDP is reduced to S and PO43- [14]. Sulfur is light yellow and if iron(Ⅲ) is reduced to Fe2+ by S it would give a green color. With time, ferrous ions may be oxidized to Fe3+, whose color is yellow or red dependent on concentration. For ZDDP in hexadecane (Figure 7) we know that there is no difference in the infrared spectra after one week at 80 oC. So there is very little of ZDDP decomposed to other structures after one week. For ZDDP in DGDE, the spectrum looks much different even after 1 day at 80 OC and the peak position changed when the system was subjected to Fe2O3. This observation implies that a new complex was generated. So, ZDDP is more stable in hexadecane than in DGDE or other words, its stability depends on the solvent.
3.5 Batch experiments
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
0 5 10 15 20 25
solid/liquid mass ratio
Amount adsorbed
Fe2O3 adsorb ZDDP in DGDE solutions 30℃
Figure 10. Adsorbed amount (wt-%) of ZDDP versus different mass ratios of Fe2O3/ZDDP in DGDE solution.
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6
0 1 2 3 4 5 6
ZDDP weight percent in DGDE(%)
Amount adsorbed
solid/liquid mass ratio:15/25 30℃
Figure 11. Adsorbed amount (wt-%) of ZDDP at different ZDDP concentration in DGDE.
Figure 10 shows the amount of ZDDP adsorbed on iron oxide with increasing iron oxide/liquid ratio from 1:25 to 20:25 using 5% ZDDP in DGDE solution. It is clear
that the adsorbed amount of ZDDP increased with the mass ratio. We chose Fe2O3: ZDDP solution ratio 15:25 to do batch experiments at 30 oC, as shown in Figure 11.
However, it is clear that the system did not reach equilibrium within the ZDDP concentration range used.
Therefore we did batch experiments with the solid/liquid ratio equal to 1:25 as shown in Figure 12 and Figure 13 for DGDE and hexadecane, respectively. Both of them could reach equilibrium at high ZDDP concentrations. For ZDDP in DGDE the plateau value starts at about 2 wt-% and for ZDDP in Hexadecane it starts at about 1.5%. However, we can not get typical Langmuir adsorption isotherms. For experiments at 45 oC, 60 oC, and 80 oC only small amounts were adsorbed, so small that it sometimes appeared as negative values according to our calibration curve.
These plots are not shown. However, From infrared spctra it is anyhow evident that the adsorbed amount decreased with increasing temperature (Figure 14 and Figure 15).
We also tried solid/liquid mass ratios equal to 3/25 and 5/25 and performed batch experiments at different temperatures. However, typical Langmuir curves could not be obtained, which seems reasonable since the ZDDP changed with temperature and was not even pure when the experiments were commenced.
0 0. 05 0. 1 0. 15 0. 2 0. 25
0 1 2 3 4 5 6
di f f er ent ZDDP wei ght per cent i n DGDE
Amount adsorbed
sol i d l i qui d mass r at i o: 1/ 25 30℃
Figure 12. ZDDP adsorbed onto iron oxide from DGDE solutions at 30 oC.
0 0. 05 0. 1 0. 15 0. 2 0. 25 0. 3
0 1 2 3 4 5 6
di f f er ent ZDDP wei ght per cent i n hexadecane( %)
Amount adsorbed sol i d/ l i qui d r at i o: 1/ 25 30℃
Figure 13. ZDDP adsorb onto iron oxide from Hexadecane solutions at 30
oC.
- 0. 1 0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6
940 960
980 1000
1020 1040
1060 1080
wavenumber ( cm- 1)
Absorbance
DGDE cal i br at i on cur ve DGDE 30℃
DGDE 45℃
DGDE 60℃
DGDE 80℃
Figure 14. IR spectra of ZDDP adsorbed onto iron oxide from DGDE solutions at various temperatures.
0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6
940 990
1040 1090
1140
wavenumber ( cm- 1)
Absorbance
DGDE cal i br at i on cur ve DGDE and Fe2O3 30℃
DGDE and Fe2O3 45℃
DGDE and Fe2O3 80℃
Figure 15. IR spectra of ZDDP adsorbed onto iron oxide from hexadecane solutions at different temperatures.
4. Conclusions
The ZDDP we used could dissolve in DGDE completely and easily but not in hexadecane. Some solid was left after dissolving in hexadecane and it should not be like that. It was demonstrated that the sample we used is not pure enough, no matter if the ZDDP batches were old or new. Less solid left for new ZDDP after dissolving in hexadecane compared with the old one, which means that the new ZDDP is better than old one. This affected the calibration curve in such a way that the integrated infrared area was somewhat higher for new ZDDP. Accordingly, a pure ZDDP would have resulted in a steeper calibration curve implying a better sensitivity and more precise quantitative measurements. Fortunately, only a small amount of solid could not dissolve in hexadecane although the ZDDP was still impure. However, there was not much difference between the two calibration curves, which implies that all of the ZDDP in the impure batch was dissolved and that the solvent itself had only a minor effect on the molar absorption corfficient.
The stability of ZDDP solutions depends on the solvent. ZDDP could readily be dissolved in hexadecane for more than one week whilst less than one day in DGDE at 80 oC. These results clearly show that ZDDP (as obtained from SKF) is more stable in hexadecane than in DGDE. Of course the impurities can certainly contribute to this difference between hexadecane and DGDE. According to literature ZDDP should be stable at least up to 80 oC, but eventually decompose at temperatures above 150 oC.
However, the infrared spectra recorded in this work are assigned to the thiophosphate group in ZDDP, so changes in these spectra are due to changes in ZDDP. The tricky part here is to exclude effects from impurities on the spectral behavior of ZDDP.
We did batch experiments at different temperatures at the beginning when we still did not know that the ZDDP was impure, with the aim to test our adsorption data in a Langmuir type of plot. The batch experiments were repeated many times and with different Fe2O3: ZDDP mass ratios. These measurements indicated that a Langmuir type of plot could be obtained at 30 oC, but at higher temperatures the curves looked very strange and were also hard to reproduce. To our opinion this was due to the impurities in ZDDP, as detected by NMR spectra.
5 Future works
It has been shown by this work that the method works and that it should be possible to perform adsorption/reaction experiments where ZDDP is adsorbed onto iron oxides. From such measurements it would be possible to obtain adsorption isotherms and thermodynamic parameters for the adsorption of ZDDP on iron oxides as well as kinetic data for the adsorption. However, it is of course not possible to obtain trustable results if the chemical used is impure. Accordingly, for future work our experience now tells us that the purity of the ZDDP chemical should always be tested before experiments are commenced.
6 ACKNOWLEDGEMENTS
First of all I would like to greatly acknowledge Professor Allan Holmgren and Mattias Grahn for helpful guidance and providing ideas during this work. I will also thank you for the fruitful discussions we have had during the work.
I would also like to thank research Engineer Maine Ranheimer and Dr Mats Lindberg for the laboratory and computer aspect.
I want to thank my friends Xiaofang Yang, Ivan Carabante, Anuttam Patra for their help on the instrument operation.
I also want to thank Prof.Oleg N Antzutkin for his good suggestion.
There are many people at the department of Chemical Engineering and Gesciences I want to thank, thank you all.
Finally, I would like to take the opportunity to send my best gratitude to my supervisor Professor Zhongxi Sun, for giving me this opportunity to study here. And he always give me support and guidance during this work though E-mail.
The financial support from the division of Chemical Engineering and SKF are gratefully acknowledged.
References
[1]G.M.Bancroft, M.Kasrai, M.uller, et al., Tribology Letters3(1997), 47-51.
[2]Kaizhong Fan, Jing Li, Haibing Ma, et al., Tribology International 41(2008),1226.
[3]K.L. Windahl, T.J. Cardwell, Journal of Chromatography A, 765(1997), 181.
[4]O.L. Warren,J.F. Graham, P.R.Norton, Tribology Letters 4(1998), 189.
[5]Ksenija Topolovec-Miklozic, T.Reg Forbus, et al., Tribology Letters 26(2007)161.
[6]S. Plaza, Tribology Transacions, 30(1987),241.
[7] S. Plaza, Tribology Transacions, 30(1987),233.
[8]S.A. Francis and A.H. Ellison, J.Chem.Eng.Data.6(1991),83.
[9]F.G. Rounds, ASLE Trans. 7(1964),11.
[10]S. Plaza, Analyst, 109(1984), 1313.
[11]B. Dacre, C.H. Bovington, ASLE Trans. 36(191982),333.
[12]Federica M.Piras, Antonella Rossi, Nicholas D. Spencer, Tribology Letters,15(2003), 181.
[13]B.Dacre, C.H.Bovingon, ASLE Trans. 24(1982), 546.
[14]H. Spikes, Tribology Letters, 17(2004), 469.
