Ind. Eng. Chem. Prod, Res. Dev. 1984, 23, 21-27
for transplantation. Finally, the use of agarose to encapsulate activated charcoal and ion-exchange resins in agarose beads for hemowrfusion to detoxifv individuals in the case Of drug overdose looks promising (Holloway et 1979). ~. New biotechnology applications are continually being discovered for agarose and its derivatives. can be seen from the representative examples cited, agar and agarose, polysaccharides from seaweeds, or marine algae, are inOne Of partners in advancing the most exciting frontiers of science. Registry No. Agar, 9002-18-0; agarose, 9012-36-6.
21
Coulson, S. E.; Cook, R. 8. “Electrophoresls ‘82”, Stathakos, D., Ed.; W. deGruyter and Co.: Berlln, 1982. D u c k w h , M.; Yaphe, w. cerbohyd. Res. 1971, 76, Guiselev, K. B. cerb.ohvd. R ~ S 1970. . 73, 247-256. Hotstw, B. H. J. Bloctiem. Res. co”.1973, 50, 751-757. Holloway, C. J.; Harstlck, K.; Brunner, 0. Int. J. A M . Org8ns 1979, 2 ,
n i-RR ””. “ I
Howell, S.L.; Ishaq,S.;Tyhwst, M. J . physw. 1982, 324, 20. Nochumson, S.; Cook, R. B.; Wllllams, K. W. Presentation at Ekctrophoresls Forum ‘80; Technlcal Unhrerslty, Munich, West Germany, 1980. Nochumson, s.; Qlbson, s. Bbrechniques 1983, 7 , 18-20. Schell. H. D.; ohetie, V. Rev. Roum. Blochim. 1972, 9 , 165-177. Serwer, P. Anal. Blochem. 1981. 172, 351-356. Shllllto, R. D., Frledrlch Mlescher-Institut, Basel, Switzerland, personal c0t-t-tmunlcatlon, 1982. Urlel, J.; Berges. J. C. R. Aced. Scl. farls 1986, 262, 164-167. Womer, M. C., Ed., ”The Agarose Monograph”, Marlne Collolds Dlvlsion, FMC Corporation, Rockland. ME, 1982.
Literature Cited
Received for review April 20, 1983 Revised manuscript received August 30, 1983 Accepted October 3, 1983
Buds, 2.; Chrambach, A. Electrophoresis 1982, 3, 130-134. CMn. C. I.; Banquerlgo, M. L., Johns Hopklns Unhrerslty. Baltimore, MD, personal communication, 1983.
I I I. Symposium on Synthetic and Petroleum Based Lubricants B. L. Cupples, Chairman 183rd National Meeting of the American Chemical Society Las Vegas, Nevada, March 1982 (Continued from June 1983 Issue)
Some Synergistic Antioxidants for Synthetic Lubricants Tal S. Chao,’ David A. Hutchison, and Manley Kjonaas ARC0 Petroleum Products Company, Dlviskm of Atlantlc Richflekl Company, Harvey, Illlnols 60426
Oxidation stability is a major requirement for synthetic lubricants, especially for those used in aircraft gas turbine engines, owing to the high-temperature oxidative environment. This paper describes some antioxidant systems which we found to be highly effective in improving the oxidation stability of synthetic lubricants. These systems contain a primary antioxidant, usually an aromatic monoamine, and a synergistic antioxidant. The latter is a compound which, when added in small concentrationsto the primary antioxidant, greatly improves its effectiveness. Synergistic antioxidants discussed will include alkali metal compounds, heterocyclic amines, aromatic diamines, hydroxybenzophenones, and certain sulfur and phosphorus compounds. Discussions will also be made of a bench test used for screening antioxidants and possible mechanisms of action of these antioxidants.
Introduction Oxidation stability is a major requirement for all lubricants. It is even more important for synthetic lubricants used in aircraft gas turbine engines, owing to the high oil temperature and air exposure. Two basic means are available to improve oxidation stability of synthetic lubricants: choice of base fluid and use of antioxidants. It is well-known that fluids such as polyphenyl ethers are much more oxidatively stable than fluids such as dibasic acid esters. Discussions of the effect of the nature of the C-H bond on oxidation stability have been made in our previous papers (Chao et al., 1979, 1983) and by other authors such as Ingold (1961a,b) and Sniegoski (1963). However, the choice of base stock is severely limited by cost and other considerations such as volatility, flow characteristics, etc. When the choice of base fluid has been made, use of a better antioxidant system is the only avenue available to improve the oxidation stability of the lubricant. This paper describes some of the antioxidant systems which were found to be especially effective for ester type 0196-432118411223-0021$01.50/0
synthetic lubricants. The esters used included those based on dibasic acids, neopentylpolyols, and complex esters. The antioxidant system discussed includes one or two primary antioxidants and a synergistic antioxidant. The primary antioxidants used included N-phenyl-anaphthylamine (PANA), p,p’-dioctyldiphenylamine (DODPA), and other aromatic amines. Synergistic antioxidants which were found to be very proficient in improving the effectiveness of the primary antioxidants include alkali metal compounds, aromatic diamines, heterocyclic amines,hydroxybenzophenones, and certain sulfurand phosphorus-containing compounds. Experimental Section Preparation of Alkali Metal Compounds. Alkali metal salts of partial esters of ethylenediaminetetracetic acid (EDTA) were prepared either by partial esterification of EDTA with alcohol, followed by neutralization with Na2C03or K2C03,or by the partial saponification of the tetraester, with NaOH or KOH. The latter method gave better control and provided relatively pure products. In 0 1984 American Chemical Society
22
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23,No. 1, 1984
most cases the product used contained a mixture of the tetraester with the Na or K salt of the triester, and possibly small amounts of salts of the diester. These are designated by notations such as Na Bu3(EDTA) representing the Na salt of the tri-n-butyl ester of EDTA. Alkali metal salts of perfluorobutyric acid and other acidic compounds were prepared by neutralization with alkali metal hydroxides with ethanol as solvent. The neutralized product was filtered and crystallized to yield the alkali metal salt. Sodium acetylacetonate and sodium dinonylnaphthalene sulfonate were purchased. Sources of Other Materials. Antioxidants such as PANA, DODPA, naphthalene diamines, hydroxybenzophenones, thioureas, and starting meterials were purchased from commercial sources. Benzilidine 2,2'-di(4-picolyl)amine was prepared by treating 2 mol of 4-picolylamine with 1 mol of benzaldehyde. Base fluids used in testing the antioxidants were mostly commercial products except a few specifically mentioned. Ester A is a pentaerythritol ester of a mixture of C6-Cl0 fatty acids, and ester F is a dipentaerythritol ester of similar acids. Diisooctyl azelate was purchased. Complex ester N was prepared by esterifying 1 mol of neopentyl glycol with 2 mol of azelaic acid and 2 mol of isooctyl alcohol. The trimethylolpropane ester of pelargonic acid was prepared by esterifying trimethylolpropane with a slight excess of pelargonic acid. Oxygen Absorption Test. This test is a modified Dornte O2absorption test and involves the bubbling of O2 at 28.3 L/h through 75 g of test oil held at 232 "C (Dornte, 1936). The volume of oxygen absorbed is monitored with a transducer and recorded automatically. The test is run until 2500 mL of O2 (measured at 25 "C and 760 mmHg) is absorbed. The test can be run either in the presence or absence of 12 ppm of Fe as Fe octoate. The results of the test are reported as Ti,the induction period, and T,, the time required to absorb 2500 mL of O2 The induction period is the time elapsed prior to rapid change of the rate of O2 absorption. Details of the test apparatus and procedure have been reported (Chao et al., 1970). Oxidation-Corrosion Tests. Oxidation-corrosion tests are standard tests specified in Federal Test Method Standard 791b, Method 5308 and Pratt and Whitney PWA521B. They generally involve the bubbling of air at 5 L/h through 100 mL of test oil in the presence of five or six metal coupons and determining the changes in metal weight and the viscosity and acid number of the oil. Temperature and test duration vary from one version to another. Bearing Head Test. The bearing head test was conducted according to the procedures established by Wright Air Development Division (WADD) and Pratt and Whitney (P&W) and summarized in Section 3.3.8 of MIL-L9236(USAF). The test employs an Erdco Universal Tester equipped with a 100-mm roller bearing which rotates at a speed of 10000 rpm under a radial load of 227 kg; 7.6 L of the test lubricant is circulated through the bearing head at a rate of 600 mL/min while moist air is supplied at a rate of 9900 mL/min. Under the type of test conditions used in this paper the temperatures of the bearing and lubricant are as follows: lubricant inlet, 204 "C; lubricant sump, 227 "C; bearing, 260 "C. The test is run with 15-h intervals each followed by a 9-h cooling-off time until a total running time of 100 h is accumulated. Properties of the oil including kinematic viscosity at 98.8 and 37.7 "C, total acid number (TAN), pentane and benzene insolubles, and ppm Fe are monitored during the test. A t the end of the test the hearing bead is examined and rated by both
the WADD and P&W system. The filter deposit is weighed.
Results and Discussion Definition of Synergism and Synergist. Dugan (1963) stated that the term synergism refers to the cooperative action of two or more agents in such a way that the total effect is greater than the sum of the individual effects taken independently. Emanuel et al. (1967) expressed this idea mathematically. Synergism is obtained when Tl,2> T1+ T2,where T1and T2are induction periods for each is the induction period under the antioxidant and T1,2 action of both. A similar mathematical definition was given by Scott (1965). The classes of compounds described below are believed to be synergists because they are poor antioxidants by themselves and when added in small concentrations to the primary antioxidant greatly increased the induction period in the oxygen absorption test, significantly reduced both the viscosity and acid number increases, and the metal losses in the oxidation-corrosion test; many of them markedly improved the performance in the bearing head test. Data to satisfy the mathematic definition of synergism will be shown for all except a few cases. In these few cases no data were obtained to provide direct comparisons, but the fact that these synergists were poor antioxidants when used alone was well established by related data. Alkali Metal Compounds. Davis and Thompson (1965, 1966) reported the synergistic effect of alkali metal salts of carboxylic acids with aromatic amines in improving oxidation stability of synthetic lubricants. By using the 218 "C oxidation-corrosion test, they showed that Na and K stearates, when used with PANA and 5-ethyl-10,lOdiphenylphenazasiline, were outstanding among other metal stearates in reducing increases in viscosity and acid number. They used the potassium salt of the partial amide of EDTA with Primene 81R (a tertiary alkylamine of 11-14 C atoms). 0
0
II RNHCCH2
CH2CNHR
\
II
/
NCH~CH~N
Based on this synergist, they formulated a MILL-9236B candidate lubricant which showed good results in the oxidation-corrosion tests and a 274 "C bearing head test. However, it failed a 100-h turbine engine test required for MIL-L-9236B lubricants, owing to excessive viscosity increase. Our discovery of the synergism between alkali metal compounds and aromatic amines actually dated as far back as 1962. The discovery was incidental, since our objective was to develop metal deactivators based on the tetraesters of EDTA. We found that an impure product of the tetra-n-butyl ester of EDTA containing a small quantity of Na was a strong synergistic antioxidant with PANA, while the purified ester containing no Na was not. Following this lead we prepared Na and K salts of partial esters of EDTA and other polyaminopolycarboxylic acids and a number of other alkali metal compounds and tested them as synergistic antioxidants for aromatic amines. Table I shows the type of alkali metal compounds which we found to be effective. For the Na and K salts of partial esters of EDTA we generally used products containing the tetraester and the
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 1, 1984 23
Table 11. Synergism between Alkali Metal Compounds and PANA (Base Fluid: Ester A, N o Fe Catalyst)
Table I. Types of Alkali Metal Compounds Effective as Synergistic Antioxidants 1. alkali metal salts of partial esters of polyaminopolycarboxylic acids
e.g. BuOZCCHZ
o,
absorption data
CH2CCZNa
‘NCH~CH~N
mi n
9
24 2 244
649
291 296 69 1
none 1081
73 1115
none 440
84 487
138
181
878
92 0
none 1117
81 1147
none 896
90 930
C s H ~ N ~ C I i z C H z N = ” L ~ ~ ]
PANA
6. alkali metal derivatives of enols e.g. c H 3- c T c H +-c H~
li
!i
CgHsN
I
I
I
I
I
1
I
B a s Fluld . Ester A Prlmary Antloxldant . 1% PANA Synecglstic Antioxldanl . K or Na Bu, (EDTA) in Bu, (EDTA)
P
4 ““a], CHzCHzN=CH
Na acetylacetonate PANA Na acetylacetonate Na Bu,(EDTA) PANA Na Bu,(EDTA)
u~Nu’“ I
1.0 0.1
Tested in presence of 12 ppm 01 Fe
a
1.1 1.0 0.1 l.la 1.0 O.la
As a 10%solution in Bu,(EI)TA).
Table 111. Effect of Fe Catalyst on Antioxidant Synergism of Alkali Metals. Alkali Metal Salt of Partial Ester 0,
I
0
20
I
40
I
60
I
I
I
I
I
80
100
120
140
160
absorption data 180
Concentration of Na or K (ppm)
Figure I. Effect of concentration of alkali metal on oxygen absorption.
Na or K salts of partial esters. The effectiveness of these products is directly proportional to the concentration of Na or K in the product. This is shown in Figure 1,where T,, the time required to consume 2500 mL of 02,was plotted against the concentration of Na or K. It can also be seen that the K salts are much more effective than the Na salts. The fact that these compounds are synergistic antioxidants is indicated in Table 11, where O2absorption test data in the absence of Fe catalyst were shown. Thus, 1.1% PANA provided a Tt of 291 min, 1.1% Na Bu3(EDTA) provided a Ttof 99 min, but the combination of 1% PANA and 0.1 % Na Bu3(EDTA) gave a Ttof 930 min. Similar conclusions of synergism can be reached by comparing Ti.Data showing the effect of Fe catalyst are shown in Table 111. By far the best alkali metal synergists we found are those derived from halogen-containing acids, especirtHy perfluorobutyric acid. They are strong organic acids and their alkdi metal salts are much more soluble in ester fluids than those from unsubstituted carboxylic acids. For example, a 50% solution of C3F,COOLi remained as a clear liquid in ester A for over one year before some cryatalliiation took place. Among the alkali metal salts, Li ones are most soluble, followed by those of Na, K, and Rb. Data indicating synergism between C3F7COONa and PANA are shown in Table 11. The combination of 0.1% C3F7COONaand 1%PANA gave a Ti of 1081 min, while each component alone gave Tiof 0 min and 242 min, re-
base fluid
ester A ester A ester A TMP-C,‘ TMP-C.‘
additives PANA PANA Fe a PANA
1.0 1.0 0.0012 1.0
Fe 0.0012 K B u , ( E D T A ) ~ 1.0 PANA 1.0
PANA Fe a
TMP-C,‘
wt%
1.0
0.001 2 1.0 Fe a 0.0012 K B U , ( E D T A ) ~ 1.0
PANA
Tt,
Ti, min
min
187 66
240 117
650
733
138 44
187 92
289
330
a As octoate. Mixture of Bu,(EDTA) and K Bu,(EDTA), containing 0.297% K. Trimethylolpropane pelargonate.
spectively. Effectiveness of alkali metal salts of other halogen-containing acids are shown in Table IV. Among the salts of C3F7COOH,the order of effectiveness is Rb > K > Na > Li. At a concentration of 1 mmol/kg of finished oil, containing esters A and F (75 and 25% each) as base fluid and 1% each of PANA and DODPA as primary antioxidants, the Li, Na, and K salts gave Ttof 696,789, and 1379 min, respectively. The Rb salt was soluble at only 0.1 mmol/kg and gave a T,of 1471 min a t this concentration. These alkali metal salts were not only synergistic to PANA and DODPA, but they were also synergisticto other aromatic amines such as N,N’-diphenyl-p-phenylenediamine. A mixture of 1%of this amine and 0.1% C3F7COONa gave a T,of 1245 min in a base fluid consisting of
.
24
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23,No. 1, 1984
Table IV. Effectiveness of Alkali Metal Salts of Halogen-Containing Acidsa 0,
absorption data salt of halogen-contg acid none CCl3CO Na C SF,CO,Na C,F,,CO,Li H(CF,),CO,Na H(CF,),CO,Li r Fj
i
z
metal, PPm
Ti, min
T: mn
65 18 11.5 19 20 64
224 748 626 442 670 528 835
266 768 661 472 700 550 882
37
660
698
_-__
3
Base formulation: ester A, 75%; ester F, 25%; PANA, Incompletely soluble, filtered before test. a
+ 1%. Tested in absence of Fe catalyst.
Table V. Comparison of Different Types of Na Compounds at Comparable Na Levela 0, absorption data
Na, ppm
Ti min
T, mi n
none Na acetylacetonate Na Bu,(EDTA) C ,F$0, Na
-__11 12 10
2 24 594 598 5 57
26 6 638 628 595
Na lauryl phosphates
10
148
180
Na compd
1
3
a Base formulation: ester A, 75%; ester F, 25%; PANA, +1%. Tested in absence of Fe catalyst.
a 75-25 mixture of esters A and F in the absence of Fe catalyst. More data are shown in U S . patent publication
3 436 348 (Chao et al., 1969). This synergism appears to exist in all ester fluids, including diabasic acid esters, neopentyl poly01 esters, and complex esters prepared from neopentylglycol, a dibasic acid, and an alcohol. Data to support this point can also be found in this patent and in US. patent 3329611 (Chao 1967). The effectiveness of different alkali metal compounds is different. Table V shows a comparison of six different types of Na compounds, all of which are used at an Na concentration of 10-12 ppm in the presence of 1% PANA. It can be seen that sodium acetylacetonate, Na Bu3(EDTA), and C3F7COONaall gave about the same Ti and Tv On the other hand, sodium dinonylnaphthalene sulfonate was much less effective and sodium lauryl phosphate, a mixture of Na salts of mono- and dilauryl acid phosphates, is antagonistic toward PANA. The effectiveness of C3F7COONaas a synergistic antioxidant to PANA and DODPA was also proven in the oxidation-corrosion test and the bearing head test. Data are shown in Tables VI and VII. In Table VI we can see that the use of 0.02-0.06% of C3F,COONa substantially reduced both metal corrosion and increases in viscosity and acid number. In Table VI1 it can be seen that the use of 0.06 5% C3F7COONagreatly reduced increases in viscosity, acid number, and filter deposit, while WADD and P & W deposit ratings were very good. Table VI1 also includes bearing head test data which show the effect of K Ru3(EDTA) in improving all the performance characteristics. Additional bench and engine test data on a type I1 synthetic lubricant containing 0.06% C3F7COONaas the synergist can be found in US. patent 3329611 (Chao 1967). The effectiveness of these alkali metal compounds as synergistic antioxidants with aromatic amines has also been proven in a finished type I1 synthetic lubricant which was commercialized after accumulating 2030 h of successful testing in eight JT8D engines, more than 1830 h in JT3D-3B, JT3C-12, JT4, JT-12, and TF33P7 engines, and approximately 100OOO h of flight evaluation. It was fully approved by both Pratt and Whitney and General Electric. Aromatic and Heterocyclic Polyamines. Many aromatic diamines, including those having primary and secondary amine groups, were found to be synergistic with primary antioxidants such as PANA. Table VI11 shows the synergism between PANA with 1,5- and 1,8-diaminonaphthalenes and their N-alkyl derivatives. Other diamines which we found to behave similarly include o-, mand p-phenylenediamines,4,4'-diaminodiphenyl ether, and 4,4'-diaminodiphenylmethane,etc.
Table VI. Oxidation-Corrosion Test Data. Effect of C ,F,CO,Na base formulation ester A ester F PANA DODPA Bu,(EDTA) saturated dimer acid monoamide DC Fluid (200 cSt; 60 000 M,) test conditions 204 "C, 72 h C,F,CO,Na, wt % wt change, mg/cm2 CU Ag
steel Al alloy Mg alloy Ti viscosity increase, 37.8 'C, % acid no. increase
0
-1.42
+ 0.06 -0.10
-0.04 -0.03
0.02 -0.61 0.00 0.00
+ 0.03 1.0.05
_._
___
30.1 1.82
23.0 0.57
72.22 24.08 1.00 1.00 1.50 0.20 0.001 218 "C, 48 h 0
-0.93 +0.03 +0.03 -0.02 -1.72 -0.02 30.3 2.21
0.02 -0.28
+ 0.06 -0.02
+ 0.02 + 0.01 + 0.01 23.4 1.06
232 "C, 48 h 0
__ __ __ ___ ___
__ __ __ __-
___
0.06 -0.30 +0.02 +0.05 +0.01
-0.77
+ 0.04
22.2 1.27
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 1, 1984 25
Table VII. Type I1 Bearing Rig Test Data. Effect of C,F,C02Na and K Bu,EDTA formula composition, wt % ester A ester F PANA DODPA Bu,(EDTA) K Bu,(EDTA) (0.297% K in Bu,EDTA) C ,F,CO,Na saturated dimer acid monoamide DC Fluid (200 cSt; 60 000 M,)
A
B
--_
100
_100 __
___ ___ ___
_-_
_ +_ _1.0
t 1.0
_-_ _..
results viscosity increase, 37.8 ‘C,% final acid number 100 mesh filter deposit, mg WADD demerit rating P&W rating
_- +-_ __1.0 ___ ___
C
D
72.09 24.21 1.00 1.00 1.50
71.99 24.25 1.00 1.00 1.50
---
___
0.20 0.001
___
0.06 0.20 0.001
2190 29 1 271 50.0 9.98 5.65 5.69 0.93 5.60 0.87 0.29 1.24 128 38 53 58 2,372 1,1,1 1,131 1,191 Table IX. Heterocyclic Amines as Synergistic Antioxidants. (Base Fluid: 80% Complex Ester N, 20% Dioctyl Azelate, 12 ppm Fe)
Table VIII. Naphthalene Diamines as Synergistic Antioxidants. (Base Fluid: Ester A, + 1 2 m m Fe) 0 2
absorption data antioxidant
wt %
PANA 1.5 l&diaminonaphthalene 1.5 PANA 1.0 l,&diaminonaphthalene 0.5 1,5-diaminonaphthalene 1.5 PANA 1.0 1,5-diaminonaphthalene 0.5 N-octyl-l,5-diaminonaphthalene 1.5 PANA 1.0 N-octyl-l,5-diaminonaphthalene 0.5 N-octadecyl-1,8-diaminonaphthalene2.0 PANA 2.0 PANA 1.0 N-octadecyl-l,8-diaminonaphthalene 1.0
Ti, min
0 2
absorption data
T,, min
62 109 131 181 265 308 150 1 9 3 257 317 112 157 258 297 100 142 74 126 295 335
The effectiveness of heterocyclic amines as synergistic antioxidants was first discovered by Eickemeyer and Ovist (Eickemeyer and Ovist, 1962; Eickemeyer 1964a,b, 1965). They found that compounds such as 2-aminopyridine, 3-aminoquinolineand 2,2‘-dipyridylamine were synergistic with antioxidants such as PANA, phenothiazine, and 5ethyl-10,lO-diphenylphenazasiline.This work was continued and several derivatives which were somewhat more effective and more compatiblewith Viton rubber seals were discovered. Table M shows the synergistic effect of these compounds with PANA in a type I base fluid containing 20% dioctyl azelate and 80% of complex ester N. Polyhydroxyphenols. Synergism between aromatic amines and phenols has been reported by Meskia et al. (1972). Our search for synergistic antioxidants containing one or more phenolic hydroxy groups for use in synthetic lubricants was mostly unsuccessful. The only exception was hydroxybenzophenonescontaining three and four OH groups. Table X shows that while these compounds were quite effective, other hindered phenolic compounds containing two OH groups were ineffective. A type I1 synthetic lubricant containing 0.30% 2,2‘,4,4‘-tetrahydroxybenzophenone gave a viscosity increase of 20.0% and a final acid number of 1.37 in a type I1 bearing head test. Details can be found in U.S.patent 3493510 (Chao 1970). Sulfur and Phosphorus Compounds. Synergism between aromatic amines and sulfur compounds is wellknown. During our screening program for synergistic an-
antioxidant PANA 2-anilinoquinoline 2-anilinoquinoline PANA benzilidene 2,2’-di-(4-picolyl)amine benzilidene 2,2’-di-(4-picolyl)aine PANA 2,6-di(pyridylamino )-4-tert-octylp hen01 PANA
0.5 76 1.0 1 4 1.0 1 6 5 0.5 0.2 1 2 0.2 168 0.5 1.O 2 11 0.5
163 109 210 81 218 28 1
tioxidants, we found that almost every organic sulfur compound tested acted as a synergist for aromatic amines. The problem is that most sulfur compounds contribute heavily to deposit formation, rendering them unsuitable for use in synthetic lubricants. The solution is, therefore, to find a sulfur compound which contributes least to deposits or one which is so effective that only a very small concentration is required. Table XI shows a number of such compounds and their effectiveness in improving the induction period provided by PANA. It also includes two compounds containing both sulfur and phosphorus. Data on a few other sulfur compounds are shown in our patent publications (Eickemeyer and Chao 1966a,b; Chao and Eickemeyer 1969). Since synergism between aromatic amines and sulfur compounds is known, very little effort was made to develop data to prove synergism. However, we did obtain data showing that 1% of 0,O-dicresylphenylthiophosphonategave a Ti of 10 min and Tt of 96 min in the test. Since 0.5 wt % PANA gave a Tiof 70 min and the combination a Tiof 243 min, their synergistic effect is well established. Possible Mechanism of Action. Mechanism of synergistic action of antioxidants has been reviewed and discussed by Scott (1965), Emanuel et al. (1967), Ingold (1961a), Meskia et al. (1972), and Kikkawa et al. (1972). In general, the antioxidants in a synergistic mixture operate by different mechanisms of oxidation inhibition, and in doing so, reenforce each other to such an extent that their combined effect becomes greater than the sum of their individual effecta. There are three major mechanisms of oxidation inhibition: free radical trapping, peroxide decomposition, and metal deactivation.
26
Ind. Eng. Chem. Prod. Res. Dev., Vol. 23, No. 1, 1984
Table X. Hydroxybenzophenones as Synergistic Antioxidants 0 2
absorption data ~
base fluid blend + 1 2 ppm Fe ester ester ester ester ester ester
A F A F A F
antioxidant
75% 25% 75% 25% 75% 25%
ester A 72.22% ester F 24.08% Bu,(EDTA) 1.5% saturated dimer acid ester A 72.22% ester F 24.08% Bu,(EDTA) 1.5% saturated dimer acid ester A 72.22% ester F 24.08% Bu,(EDTA) 1.5% saturated dimer acid ester A 72.22% ester F 24.08% Bu,(EDTA) 1.5% saturated dimer acid
wt %
Ti, min
T,, min
PANA DODPA 2,2',4,4'-tetrahydroxybenzophenone
1.o 1.0 0.5
46
93
0
69
2,2',4,4'-tetrahydroxybenzophenone PANA DODPA PANA DODPA
0.5 1.o 1.0
266
3 04
1.0
23
87
PANA DODPA 2,4,4'-trihydroxybenzophenone
1.0 1.0 0.5
238
272
PANA DODPA 3,3',5,5'-tetra-tert-Bu-4,4 '-dihydroxybiphenyl
1.0 1.0
43
86
PANA DODPA bis-(3,5-di-tert-Bu-4-hydroxy)methane
1.0 1.0 1.0
61
101
-
1.0
monoamide 0.2%
monoamide 0.2% 1.0
monoamide 0.2%
monoamide 0.2%
Table XI. Effectiveness of Sulfur and Phosphorus Compounds in Improving Effectiveness of PANA. (Base Fluid: 80% Complex Ester N, 20% Dioctyl Azelate, No Fe, Primary Antioxidant: 0.5%W PANA) 0 2
absorption data synergistic antioxidant
wt %
Ti, min
T,, min
--none 76 163 0.1 152 202 s-diphenylthiourea 1.0 >492 492 s-dipheny lthiourea 0.2 380 406 N-phenyl-"- 2-p yridylthiourea s-ditolylthiourea 0.5 >399 3 99 4,4'-bis(dimethy lamino )0.1 241 275 thiobenzophenone 2-thiobenzoylaminopyridine 0.5 285 349 N,N'-dilaurylthioxamide 0.5 263 3 27 2,2'-dithiobenzothiazole 0.2 160 235 O,O, 0-tributylphosphorothioate 1.0 143 195 0,O-dicresylphenylthiophosphonate 1.0 243 338
Other mechanisms which have been proposed for synergism include regeneration of the primary antioxidant through reduction (Kikkawa et al., 1972), formation of complexes (Privett and Quackenbush 1954; Kikkawa et al., 1972), and inhibition of the initiation step. For the synergistic antioxidants covered in this paper, we believe the sulfur compounds act as peroxide decomposers and the synergism is caused by the presence of both the peroxide decomposer and the aromatic amine free radical traps. The tri- and tetrahydroxybenzophenones are UV absorbers. Scott (1965) indicated that UV absorbers can suppress the decomposition of hydroperoxides into chain-initiating radicals. It is also possible that they act as metal deactivators involving the complexing of 2 mol of the tri- or tetrahydroxy compounds with each atom of Fe. Synergism involving antioxidants of similar structure has been described by Scott (1965) as homo-synergism and is illustrated by work of Knowton and Stubbs (1958). They believe that amino compounds act similarly. The mech-
anism of homo-synergism has been attributed to regeneration of the primary antioxidant AH2 (Spetsig 1959). A A'H2 AH2 A' (1) Thus, although the synergistic antioxidant A'H2 is a weak antioxidant, the regeneration of the stronger antioxidant greatly increased their combined effect. Another regeneration mechanism was given by Mahoney and DaRooge (1967) for the synergism between 2,6-di-tertbutylphenol and 4-methoxyphenol. They showed that the rate of regeneration of the nonhindered phenol is the most important step, followed by the rapid cross-termination reaction of the less reactive hindered phenoxy radical with the nonhindered phenoxy radical or the peroxy radical to prevent any undesired chain restarting reactions. It is believed that the aromatic diamines and heterocyclic amines discussed in this paper act in a similar manner. The regeneration mechanism also explains why a diamine is better than a monoamine. We also believe that different amines have different reaction rates with free radicals. Some will react faster and be consumed earlier, while others react slowly but last longer. Thus, their combined use becomes more effective in trapping the free radicals. It has been pointed out by Ingold (1961a) that aromatic amines also act as peroxide decomposers. Synergism can take place if one of the amine antioxidants acta as a peroxide decomposer while the other acts as a free radical trap. The least amount of literature is available to explain the synergistic effect of alkali metal compounds. Davis and Thompson (1965,1966) suggested three possible mechanisms. First, they propose that alkali metal salts direct the base oil to form oxidation products other than acid or sludge. This may be true, but it implies that they do not change the rate of oxidation. Our data on the O2 absorption test clearly indicated that the rate of oxidation was greatly reduced. Secondly, they suggest the formation of complexes which are more potent antioxidants than the original material. We believe this is a definite possibility. Thirdly, they believe the salts may form free radicals or catalyze the antioxidants to form free radicals and interfere
+
-
+
Ind. Eng. Chem. Prod. Res. Dev., Vol.
with autoxidation, with which we do not agree. We believe that the mechanism of action of alkali metal salts probably involves an alkali metal cation species. This is based on the fact that their effectiveness is dependent on the concentration of alkali metals (Figure 1) and that their effectiveness increases in the order of Li, Na, K, and Rb. The fact that although the associated anions may have some influence, the alkali metals synergistic effects are not limited to just one or two types of anions, would indicate a mechanism largely independent of effects such as anion basicity. This order of synergistic effectiveness would tend to suggest that the active species is indeed derived from the alkali metal cation alone and not a charge transfer complex or neutral species derived from the original salt. There may be several ways for the alkali metal ions to improve the effectiveness of aromatic amine antioxidants, but the data would tend to suggest a coordination effect. The antioxidant effectiveness of phenols increases when the OH group is shielded by bulky groups. Coordination of alkali metal ions with the N atoms of the aromatic amine may exert a similar effect. It has also been suggested that the alkali metal compounds can probably in some way regenerate the aromatic amine or reduce its consumption and that it may inhibit the initiation step. A great deal of work is required to test these hypotheses and elucidate a viable mechanism. Finally, it should be pointed out that the alkali metal compounds are more effective than metal deactivators. By definition, metal deactivators only cancel out, up to loo%, the catalytic effect of metals. Data in Table I11 show that 12 ppm Fe reduced Ti of 1% PANA in ester A from 187 min to 66 min. With the addition of K Bu3(EDTA) a mixture of Bu,(EDTA) and K Bu&EDTA) containing 0.297% K, Ti was raised to 650 min, much longer than the original 187 min. Thus, with the data shown in Table I1 where substantial increases were seen in Ti in the absence of Fe, we should consider these compounds as synergistic antioxidants rather than metal deactivators. Acknowledgment The authors wish to express their gratitude to Atlantic Richfield Company for permission to publish this paper; to Dr. B. W. Turnquest, Dr. E. B. Ovist, and Messrs. W. J. Wostl and K. L. Boekhaus for encouragement and guidance; and to Dr. D. B. Eickemeyer for his contributions in designing the O2 absorption test and discovery of the synergistic effects of the heterocyclic amines. Registry No. Diisooctyl azelate, 26544-17-2; trimethylolpropane tripelargonate,88426-26-0;N-phenyl-a-naphthylamine, 90-30-2; p,p ’-dioctyldiphenylamine, 101-67-7; sodium p-octylphenylphenate, 78899-79-3; sodium perfluorobutylate,2218-54-4; sodium 4-thiadode~anoate~ 87441-93-8; sodium acetylacetonate, 15435-71-9;EDTA tributyl ester sodium salt, 18459-40-0;EDTA tetrabutyl ester, 14531-12-5; iron octanoate, 6535-20-2; EDTA
23, No. 1, 1984 27
tributyl ester potassium salt, 18292-51-8;sodium trichloroacetate, 650-51-1; lithium perfluorooctanoate, 17125-58-5; sodium 7hydroperfluoroheptanoate, 2264-25-7;lithium 7-hydroperfluoroheptanoate, 87441-95-0;sodium 3,5-bis(trifluoromethyl)benzoate, 87441-96-1; sodium perfluorobenzoate, 602-94-8; lithium perfluorobenzoate,87442-01-1;sodium dinonylnaphthalenesulfonate, 26834-28-6;sodium lauryl hydrogen phosphate, 24613-62-5;disodium lauryl phosphate, 7423-32-7; sodium dilauryl phosphate, 17026-45-8; l&diaminonaphthalene, 479-27-6; 1,5-diaminonaphthalene, 2243-62-1; N-octyl-1,5-diaminonaphthalene, 87441-97-2; N-octadecyl-l,8-diaminonaphthalene,88426-27-1; 2-anilinoquinoline,546&85-9;benzilidene 2,2’-di(4-picolyl)amine, 88430-54-0;2,6-bis(pyridylamine)-4-tert-octylphenol,88430-55-1; 2,2’,4,4’-tetrahydroxybenzophenone, 131-55-5;2,4,4’-trihydroxybenzophenone, 1470-79-7; 3,3’,5,5’-tetra-tert-butyl-4,4’-dihydroxybiphenyl, 128-38-1; bis(3,5-di-tert-butyl-4-hydroxy)methane, 118-82-1;s-diphenylthiourea, 102-08-9;N-phenyl-N’(2-pyridyl)thiourea, 886-60-2; s-ditolylthiourea, 50322-92-4; 4,4r-bis(dimethylamino)thiobenzophenone,1226-46-6; 2-thiobenzoylaminopyridine, 27060-34-0; N,N‘-dilaurylthioxamide, 120-88-7; 2,2’-dithiobenzothiazole,120-78-5; O,O,O-tributyl phosphorothioate,78-47-7; 0,O-dicresylphenylthiophosphonate,
52761-19-0. 25619-52-7;C&~(CH2CH2N=CHC6H4-o-OH),.2Na,
Literature Cited Chao, T. S. US. Patent 3329611, 1967. Chao, T. S.; Matson, H. J.; Kjonaas, M. U.S. Patent 3436348, 1969. Chao, T. S. U S . Patent 3493410, 1970. Chao, T. S.; Kjonaas, M.; Vltchus, E. C. SAE Paper 700890, SAE National Combined Fuels 8 Lubricants and Transportation Meeting, Phlaideiphla. PA. . . ., Nov . .- . 4-6. . ., 1970. . .. .. Chao, T. S.; Kjonaas, M.; DeJovine, J. M. Repr. Div. Petrol. Chem. Am. Chem. Soc. 1979, 24(3), 836. Chao, T. S.; Kjonaas, M.; DeJovlne, J. M. Ind. Eng. Chem. Rod. Res. Dev. 1963. 22. 357. Davis, T: G.;’Thompson, J. W. Ind. Eng. Chem. Prod. Res. Dev. 1966, 5 , 76. Davis, T. 0.;Thompson, J. W. Repr. Div. Petrol. Chem., Am. Chem. SOC. 1965, 10(2), 127. Dornte, R. W. Ind. Eng. Chem. 1936, 28, 26. Dugan, L. R., Jr. I n Kirk-Othmer “Encyclopedia of Chemical Technology”, 2nd ed.;Wiiey: New York, 1963; Voi 2, p 588. Eickemeyer, D. E.; Ovlst, E. E. US. Patent 3038859, 1962. Elckemeyer, D. E. U S . Patent 3 121691, 1964a. Eickemeyer, D. E. U S . Patent 3 150093, 1964b. Elckemeyer, D. E. U S . Patent 3226324, 1965. Eickemeyer, D. E.; Chao, T. S. U S . Patent 3288713, 1966a. Eickemeyer, D. E.; Chao, T. S. U.S. Patent 3296 136, 1966b. Emanuei, N. M.; Denison, E. T.; Maizus, 2 . K. I n “Liquid Phase Oxidation of Hydrocarbons”; Plenum: New York, 1967; p 273. Ingoid, K. U. J. I n s t . Pet. 1961a, 47(456). 375. Ingoid, K. U. Chem. Rev. 196lb, 87, 561. Kikkawa, S.; Hayashi, T.; Fujlta. K. Nlppon Kagaku Kaishi. 1972 (2), 402. Knowlton, R. E.; Stubbs, H. W. D.; British Patent 796603, 1958. Mahoney, L. R.; DaRooge, M. A. J. Am. Chem. SOC. 1967, 89, 5619. Meskla, M. Ya.; Karpukhina, 0. V.; Maizus, 2 . K. Neffekhimiye 1972, 12(5), 731. Privett, 0. S.; Quackenbush, F. W. J. Am. Oil Chem. SOC. 1954, 31 321. Scott, G. Chem. Ind. 1863, 271. Scott, G. I n “Atmospheric Oxidation and Antioxidants”; Elsevier: New York, 1965; p 203. Sniegoski, P. J. ASLE Trans. 1963, 6(4), 316. Spetsig, L. 0. Arkiv Kemi 1959, 15, 5.
Received for review September 2, 1982 Revised manuscript received August 29, 1983 Accepted September 19, 1983
