Embed Size (px)
Citation preview
1234 I-VDUSTRIAL A N D EAVGILVEERISG CHEMISTRY Vol. 19. No. 11
The loss of volatile preservative from wood will be, of course, a function of its vapor pressure. Because of the difficulties in measuring vapor pressures of mixtures of oils a t the outdoor temperature, it is more convenient to form opinions as to the vaporization losses of oils by comparisons of boiling points a t 760 mm. pressure. It is known that oils below naphthalene in boiling point are of little permanence. Naphthalene, b. p. 218" C., is fairly permanent in wood be- low ground and under water, but when exposed to the winds it largely disappears by vaporization in five or six years. On the other hand, creosote oils boiling above 270" C. are quite permanent under any ordinary conditions.
A certain commercial tar acid was found to distil under 760 mm. pressure as follows: initial point, 200" c.; 10 per cent below 207" C.; 80 per cent below 230" C.; and 95 per cent below 285" C. After chlorination to 35 per cent com- bined chlorine, a distillation under 15 mm. pressure was made and from the data thus obtained a distillation curve a t atmospheric pressure was constructed. The error in this curve vas probably not greater than 4 degrees a t any point. This curve showed that the chlorination had raised the boil- ing point an average of 30" C. A fractionation under 760 mm. pressure was not possible, since the chlorinated tar acids commence to decompose a t 220" C.
Mixtures of Chloro Tar Acid with Petroleum
The Western Union Company has treated southern yellow pine, on an experimental scale, with petroleum containing from 1 to 10 per cent chloro tar acid. No difficulty was encountered in the treatment, and it was noticeable that the penetrations were better than when petroleum alone was used. The petroleum was asphaltic base, 18' to 22" BB grade, in most cases, although both lighter and heavier oils were used. Field tests are now in progress on the wood so treated.
With most of the petroleum oils used, the chloro tar acid dissolved readily. This was especially the case with dis- tillate oils or when the petroleum had been washed with dilute acid or when the chloro tar acid had been incompletely freed from hydrogen chloride. In certain other cases, par- ticularly with residuum oils, a film of gummy material was deposited. Some of this was collected and treated with a dilute solution of sodium hydroxide. A yellowish oil of powerful odor separated and rose to the top. This was easily identified by its appearance and odor as pyridine or homo- logs of pyridine. When the water layer was acidified, chloro tar acid was recovered. Evidently, the gummy material consisted in part of the pyridine salt of the chloro tar acid.
Protection of Marine Piling against Borer Attack' Chemical Aspects
By W. D. Ramage and J . S. Burd
r N I V R R S I T Y OF CALIFORNIA, BERKRLBY, CALIF.
H E chemical treatment of wooden piling to preserve it against the attack of marine borers has been practiced for many years, but only within the last few years has
its full economic bearing been generally recognized. Recently extensive investigations2 of the borer problem
have been carried out in the hope of improving methods of protection. The San Francisco Bay Marine Piling Com- mittee was formed in 1920 for the purpose of studying the engineering, chemical, and biological aspects of borer attack on marine piling. The material here reported represents the results of the chemical investigations.
T
Note-This committee was a coiiperative organization, supported by voluntary contributions from the industries interested in water front struc- tures on San Francisco Bay. The complete report of the committee will be published shortly in book form. The chemical investigative work of the committee was concluded in March, 1924, before the completion or publication of the work of any other large agency studying the same or similar problems.
Free use has been made of the publications listed in the bibliography on marine borers compiled by A. L. Barrows (National Research Council Report, 1924). On account of the large number of overlapping publications in this field. no attempt has been made to give specific credit. It is be- lieved that those familiar with the field will readily recognize how many of our data are new and how many merely confirm the previous findings of others. Particular reference should be made to the various studies of E. Bateman, of the U. S. Forest Products Laboratory, and of L. F. Shackell, with the U. S. Bureau of Fisheries, especially on the subject of toxicity.3
Our work was confined almost entirely to problems con- 1 Received February 23, 1927. : "Marine Structures-Their Deterioration and Preservation," Na-
tional Research Council Report, 1924; "The Deterioration of Structures in Sea Water," Reports 1 to 6 of the Committee of the Institute of Chemical Engineers (British).
3 Proc. A n . Wood-Preservers' Assocn.. 11, 232 (1915).
nected with the preservation of wood from marine borers. Within the field indicated, only those preservatives or pre- servative methods were considered which involve pene- tration of the preservative into the wood. This naturally resulted in centering much of the work upon coal-tar creo- sote. The principal preservatives studied were coal-tar creosote and possible inorganic inhibitants. Furthermore, since experience has shown that properly creosoted piling is very resistant to borer attack, no treatments were studied which might be expected to cost much more than a good creo- sote treatment. Outside the field indicated, some work was done in studying the effects of chlorine on marine borers.
Creosote Study
OBSERVATION OF TEST PIECES TREATED WITH VARIOUS CREOSOTE FRACTIONS-The test timbers used in these ex- periments were 2 by 4 by 48 inches. Half of them were Douglas fir and the rest were redwood, as indicated in Table I. As far as possible they were sapwood and free from large knots or defects. Both species are more or less refractory to impregnation. All the test pieces were treated to refusal, by the standard vacuum-pressure method used for impreg- nating marine piling with creosote. In commercial treating practice varying impregnations often result from substan- tially uniform treating conditions, owing to differences in the penetrability of different wood species or specimens used, or to varying viscosities of different oils or fractions. This is largely a function of the time factor, however, as well as of temperature and pressure, and the difficulty was practically overcome in our experimental work, as it can be whenever minimum treating time is not a requisite. On account of the small size of the test pieces there was practically no core of untreated wood. The amount of creosote remaining in all the test pieces was about 15 pounds per cubic foot, except
November, 1927 INDUSTRIAL .4ND ENGINEERING CHEMISTRY 1235
T a b l e I-Cornearison of Oils Used in T i m b e r T r e a t m e n t s w i t h Oils Extrac ted a f t e r Two a n d a Half Years in Water
D A T E O F DATE O F KISD O F T~~,~~~~~.:~ TREATMEST I M M E R S I O S I T O O D a
1
2
3
4
3
6
7
b
'3
10
11
12
13
14
15
16
6-b-21
6-9-21
6-10-21
6-11-21
6-12-21
6-30-2 1
7-1-21
6-16-21
i-6-21
6-13-2 1
7-9-21
6-15-21
7-7-QI
7-5-21
i-13-21,
i-26-21
7-7-21
7-7-21
7-7-21
7-7-21
7-7-21
8-4-21
8-4-21
8-1-21
8-4-21
7-7-21
8-4-21
7-7-2 1
8-4-21
8-4-2 1
8-4-21
8-4-2 1
- -
D.F. Whole oil
D.F. Fraction A (210-235' C.)
D.F. Fraction B (235-315' C.)
D.F.
D.F.
R.
R.
D.F.
R.
D.F.
R.
D.F.
R.
R.
R.
R.
Fraction C (315-355' C . )
Fraction D (residue above
Fraction D (repeated)
Whole oil + fra.ction h
Whole oil -t fraction B
Whole oil + fraction C
Whole oil - fraction D
Whole oil - fraction A
Whole oil - fraction B
Whole oil - fraction C
Whole oil - fraction D
\Thole oil - tar acids
Oil-tar distillate
3550 C.)
AN.ALYSIS OF OIL I N PER CENT
Amount ~~ . ~ ~ ~ .
Below 210- 235- 270- 31:- Residue Specific Tar solids in 210' C. 235O C. 2703 C . 31Z0 C. 3:j0 C. gravity acids
temperature
0 . 1 4 . 2 24.6 1 8 . 0 20 .3 32.5 1.090 7 . 5 0 . 1 1 . 2 6 . 1 20.2 29.7 42.6 1.098 3 . 9
44.0 34.3 1 2 . 1 5 . 1 4 .17 , above 315' C. 1.006 1 4 . 3
S o t enough oil for analysis in sample taken, on account of large loss 0 . 0 2 2 . 6 40.9 2 4 . 1 11.17 above
315* C. 1.041 5 . S 0 . 1 2 . 4 44.4 3 1 . 1 22,753 above
315' C. 1.048 3.0 0 . 0 0 .0 0.0 16.3 56.2 27.0 1.108 6 . 0 0 . 0 0 .0 0 .8 12.3 54.0 32.9 1.112 5 , 4
Poor penetration; repeated in next run; oil too viscous
4 . 0 % below 355' C. 96.0 1.182 0 . 0 6 .9 27.7 16.7 20.6 28.7 1.076 0 . 0 2 . 3 15.2 1 9 . 3 19.0 43.8 1.088 0 . 0 3 . 0 34 .0 20 .2 17.4 24.8 1.080 0 . 0 0 . 8 1 2 . 1 21 .0 24.3 41.6 1.099 0.0 3 . 8 2 0 . 6 18 .3 26 .8 30.4 1.065 0 . 0 0 . 3 7 . 5 17 .5 30.2 4 4 . 4 1.103 0 . 0 0 . 1 10.0 16.2 20 2 53.4 1.121 0.0 0 . 1 5 . 4 8 . 3 22.9 63 .1 1.128 0.0 0.0 14.6 1 9 . 6 28.6 37.2 1.101 0 . 0 0 . 0 4 . 8 22.3 28.9 43.8 1.105 0.0 0.0 19 .0 26.6 20.9 33.2 1.110 0.0 0 . 5 6 . 2 10 .4 28.9 53.9 1.129 0 . 4 17.8 19.3 12.3 10.1 39.9 1.084 . . . 3 . 5 13.4 17.6 1 0 . 3 5 0 . 0 1.092 0.0 8 . 9 28.2 22.6 30.3 1 0 . 0 1.062 0.0 2 . 1 8 . 6 29 .5 36.8 22.7 1.082 0.0 0 . 0 21.7 1 9 . 1 24.6 34.6 1,087 0.0 1.1 3.8 22.2 2 8 . 1 44.7 1.099 0.0 2 . 7 23 .3 7 . 8 11 .4 54.3 1.117 0.0 3 . 3 8 . 9 12.5 15.2 59.7 1.118
4.1'3" below 355O C. 95.9 1.180 4 . 0 2.8 6 . 4 4 . 6 6 . 0 3 . 8 6 . 6 5 . 2 6 . 0 4 . 5 4.6 3 . 5 5 .1 4 . 4 5 . 9 3 . 2 6 . 6 3.1 0. 0 0 .0 0 . 0 0 . 0
%d;UQl
. . . . Large
.... Small
Solid . . . .
. . . .
. . . . Solid
Large
Medium
Solid
Solid
. . . .
. . . .
. . . .
. . . .
i i id ium . . I .
Solid
Aiddlum
a ie i ium . . . .
Lledium
Solid . . . .
D F. = Douglas fir, R. = Redwood.
in the case of impregnation with fraction D (Table I). Here some difficulty was experienced, but a reasonaklly satisfactory treatment was finally obtained, which was approximately 10 pounds per cubic foot.
After treatment, the pieces were made up into gates, each holding eight pieces corresponding to eight different treat- ments. An untreated bait piece was then attached to each treated piece. As there were sixteen treatments, two such gates constitute a complete series. One complete series of treatments was immersed a t each of four stations on the bay: Southern Pacific Oakland Pier, July-August, 1921; San Franckco Pier 7 , August, 1921 ; Crockett, July-August, 1921; Mare Island, September, 1921.
Five, and in some cases six, pieces were subjected to the same treatment. This gave one piece with each treatment for each of the four stations, and one or two for laboratory use. The sixteen treatments were organized as shown in Table I.
These oils were synthesized so that the several fractions were added or subtracted in the proportions in which they occurred in the whole oil, which were as follows:
Fraction 210-235' C. Fraction 235-315' C. Fraction 315-355' C. Residue above 355' C.
Pev cent 10 40 23 27
In view of the difficulties incident to obtaining clear-cut separation by fractional distillation, and consequently to obtaining clearly marked differences in inhibitive or destruc- tive effects upon living organisms, it was thought that the doubling of effect gained by both adding and subtracting the required fraction might increase the decisiveness of the results.
An analysis was made of the oil from each run, as well as of that extracted from one of the test pieces after treatment. The analyses of oils soon after treatment are of no particular signifkance and are not given.
After planting, the gates were inspected from time to time for signs of borer attack. The bait pieces were in all cases heavily attacked, and in some cases were completely de-
stroyed and had to be replaced. However, no direct attack was found on the treated wood until January, 1924, after an exposure of two and one-half years, when slight Limnoria attack was discovered on two of the pieces in the gates a t Oakland Mole. These two were the pieces treated, re- spectively, with fractions B and D. The attack was not yet so severe that these pieces could be said to have completely failed. Furthermore, the same treatments a t the other stations were not attacked. It is possible that these pieces were not so well treated as some of the others.
At the time of this inspection 10-inch lengths were cut from the test pieces and the oils were extracted and analyzed. The results are given in Table I. They show nothing un- expected, except an apparent relation between the physical character of the extracted oil and the percentage of tar acids remaining in the oil. Where the extracted oil contains a large proportion of solids a greater percentage of the original tar acids remains in the oil. This is probably due to two causes. First, an oil containing a large amount of solid matter offers a smaller effective leaching surface to the water, since it prevents, to some extent, penetration of water. Second, the oils containing the larger amounts of solids are usually the higher boiling fractions, and naturally contain higher boiling tar acids, which are less subject to leaching. Table I1 shows the relation between the amount of solids in the oil and the loss of tar acids.
Table 11-Relations be tween Loss of T a r Acids and Physical Charac te r of Creosote Oil
AMOUNT OF SOLIDS IN OIL .AT ROOM TEJIPERATVRE T A R -4CIDS
Small/ Medium 1 Large I Solid
Original
Percentage loss of orig- inal tar acids 1 48 1 48 37 24 -16 53 1 28 110 30 21 25 14 -
1236
TREAT- M B ~ ~ T ~
INDVSTRIAL AND ENGINEERING CHEMISTRY
DEPTH OF BURROWS IN TREATED WOOD PER CENT
Surface &fore 'ELoW nicks '/*.to to than ' I4 INCH PIECEsh only inch 1 inch 1 inch
Vol. 19, No. 11
a 4 5
i 9
10 11 12 13 14 15 16
6
It is evident from Table I that there is a very considerable loss of tar acids in all cases. This is not surprising when we find that there remains in the wood practically nothing dis- tilling below 235' c , and that a considerable portion of the tar acids ordinarily present distils below that point, Since these very low boiling tar acids have no permanent value, it seems evident that the ordinary specification for tar acids can be considerably reduced, if it is limited to those distilling above 235" C. In fact, it is believed that the percentage of all constituents distilling below 235" C. should be as small as the requirements for satisfactory penetration will allow.
The loss of low-boiling fractions, as shown in Table I, is, of course, accompanied by a corresponding percentage of increase in the higher boiling fractions remaining. Further- more, calculation of the amount of oil originally and finally in the test pieces indicates that the loss falls almost entirely on these low fractions-that is, that there is no washing-out effect on the whole oil.
The last inspections made were in October, 1924, for the pieces a t Oakland Mole; July, 1924, for those a t Pier 7, San Francisco; November, 1925, for those a t Crockett; January, 1924, for the pieces a t Mare Island. No attack was noted on any of the pieces except those treated with fractions B, C, and D a t Oakland Mole, which had been slightly eroded by Limnoria.
No specimen treated with whole creosote was directly attacked. Furthermore, two of the oils which stood up -the whole creosote minus tar acids, and a sample of oil-tar distillate-contained no low-boiling tar acids. There are undoubtedly some tar bases, sulfur compounds, and other possibly toxic constituents in these oils; but since by far the larger proportion of coal-tar creosote consists of aromatic hydrocarbons, it seems probable that these bodies must themselves be responsible for much of the preservative value of the oils.
The oil-tar distillate apparently has a somewhat greater protective value than that with which it has ordinarily been credited. With sufficiently rigid specifications it might be used to some extent as a substitute for coal-tar creosote in localities where conditions are not particularly favorable to the borers. This result, however, is not conclusive with- out further study.
Although no direct attack of any consequence was found on the creosoted pieces, attack was heavier where the borers went from the bait pieces into the treated pieces. On April 12, 1923, the creosoted gates a t Pier 7 , San Francisco, were inspected. None of the treated pieces showed any signs of direct attack, even though the bait pieces attached to them had been heavily attacked by Bankia and, in all but one case, almost eaten away by Limnoria. In every piece but one, however, the Bankia had gone through from the untreated bait pieces into the treated pieces. I n some cases the pene- tration of the borers into the treated pieces was slight, but in others they had penetrated several inches. In several pieces where a large number of borers had nicked the treated wood, none had gone in far. That the borers showed some reluctance to entering the creosote is indicated by the de- flection of the burrows in the bait pieces as they approached the treated wood. In some cases they ceased boring and sealed off their burrows rather than enter the creosoted wood. The amount of crowding in the bait pieces apparently has a considerable effect on the number of borers trying to enter, since the only treated piece which showed no attack was the one with the least crowded bait piece. Table I11 gives the complete description of these pieces a t the time of inspection. There was no Limnoria attack on any of the treated pieces. The data are not extensive enough to warrant any final con- clusions, but it seems reasonable to attribute a high protective power to a treatment into which a large number of borers
. . . H.A.0 10 100 C.D. 10 2 100 C.D. 35 100 C.D.
140 i 100 C.D. 25 5 4 88 C.D. 12 3 80 C.D. 25 1 100 H.A. 45 4 5 83 C.D. 30 9 6 67 C.D. 13 5 1 4 78 C.D. 75 7 4 87 C.D. 25 s 1 97 C.D.
1 48 c n 40 2
have started to penetrate from the bait piece, if none of these borers penetrate to any depth. On this basis every fraction of the creosote shows a considerable protective value.
Table 111-Description of Creosoted Test Pieces Attacked through Untreated Bait Pieces -
, ! I 1:; I E: 1 9 3 .
2 1 1 5 1
I I I
a Treatment numbers described in Table I. b H.A. = heavily attacked by Bankza and Limnoria but still holding
together; C.D. = completely destroyed, only fragments hanging to bolts. c The crowding of the borers in this bait piece was not as great as in
the others.
..
The foregoing emphasizes the importance of avoiding un- treated braces or other untreated timbers in contact with treated structures. This point has been previously noted by Shackell and others.
LOSSES FROM CREOSOTE UNDER VARIOUS COXDITIONS- Small-scale tests were carried out to find the amount and character of the loss of creosote constituents from treated wood exposed in air and in sea water. They are not to be regarded as service tests, as they do not quantitatively rep- resent the losses from marine piling. They are, however, satisfactory as indicating the qualitative changes in creosote under various conditions, and to that extent they can prob- ably be safely regarded as somewhat accelerated service tests.
Ten test pieces, 6 inches in length by 5 inches in diameter, were cut from a Douglas fir sapling and treated to refusal with a medium-weight creosote. The pieces were weighed before and after treatment to determine the weight of oil absorbed. This was very high, from 25 to 30 pounds per cubic foot.
Two of the pieces were exposed in the air and the others were placed in the bay. At tjhe same time portions of the original oil were exposed in the laboratory in open dishes. The dishes and the air pieces were weighed from time to time to determine the amount of loss. Pieces were also removed from the bay a t intervals, and the oil extracted and weighed. From the weight of oil recovered and the weight initially present, the percentage of loss on these blocks can also be calculated.
Figure 1 shows the rate of loss of creosote constituents from treated wood in water, treated wood in air, and from the open dishes. All losses are corrected to the same mass of oil and same evaporation surface. Table IV shows the compara- tive losses in the three cases after various periods of time.
Table IV-Loss' of Creosote by Leaching and Evaporation Time, in days 10 30 90 222 475 512 787 Percentage loss from open
dish in air 5 .0 1 0 . 1 2 0 . 0 2 8 . 9 3 6 . 7 3 7 . 4 3 9 . 6 Percentage loss from
treated wood in air 1 . 2 3 . 3 7 . 2 1 6 . 6 2 4 . 6 2 6 . 3 2 8 . 0 Percentage loss from
treated wood in sea water . . , , . . . . . 1 1 . 6 17.55 18 . lb 21 .9)
a All losses corrected to same mass of oil and same exposed surface. b Some mechanical loss (possibly as much as 1 per cent) in sampling.
.
The rate of loss from treated wood in air is not nearly so great, especially a t first, as would be expected on the basis of the loss from the open dish. In fact, the form of the two loss curves is different throughout their whole range and the rate
Kovember, 1927 INDUSTRIAL A V D ESGI.VEERING CHE.IIISTRY 1237
of loss from treated wood reaches a fairly constant value a t a much smaller percentage loss than in the case of the open dish. This indicates that the surface of the treated wood does not act as a free oil surface. The rate of loss from the treated surface is dependent on the rate of feeding of low- boiling constituents to the surface, which would probably be considerably lower than in the free oil, merely on account of the cutting down of circulation in the oil. This is sub- stantiated by the slightly increased amounts of lowboiling constituents found as we go inward through the treated wood. If the rate of feeding to the surface were rapid, the composition of the oil should remain constant throughout the treated wood. The same slight variations in composition in different layers are also shown by a thirty-year-old pile from the Oakland Long Wharf. There is also a possibility that the rate of feeding to the surface is further limited by an actual absorp- tion of the oil by the wood, cutting down the effective vapor pressure of the oil. In either case, and particularly in the latter, the nature of the wood used would be likely to have a considerable effect upon the rate of loss. On this basis it appears that no evaporation study of free oil is of value, except, in a very general way, in determining the probable rate of loss from treated timbers.
The smaller rate of loss from treated wood in water as compared with treated wood in air indicates the adrisability
' of storing creosoted timbers in water when possible. The reason for the smaller loss in water can probably be ascribed roughly to the greater volatility (in air) than solubility (in water) of creosote constituents in general. Also the rate of feeding to the surface is probably cut down by the swelling of the wood in water, which is not prevented by the presence of oil, because water can penetrate the cell walls of wood while creosote does not do so.
40
35
(r3 30 325 5 20 5 IS
a IO
v)
c V
5
100 200 300 400 500 600 700 800 900 ?TME IN DAYS
Figure 1-Rate of Loss of Creosote under Various Conditions
In addition to knowing the absolute loss of creosote from treated wood, it is important that we know on what constitu- ents this loss falls. To do this, we must be able to effect complete recovery of the oil remaining in the wood, without altering its character. Experiments on the use of benzene as an extracting solvent indicate that with it the oil can be completely recovered in an unchanged condition. The com- mittee work had the benefit of detailed knowledge of the oils originally used in all such tests, including carefully preserved identical samples, the lack of which has invalidated, or a t least cast doubt upon, most previous attempts in this di- rection. The analysis of the original creosote and that recovered by extraction from treated wood gave practically identical results. The effect of benzene on the oil was further tested by boiling a 50-50 mixture of benzene and oil under a reflux condenser for 3 hours. Benzene was then added until it was 90 per cent of the whole and the digestion continued for 10 hours more. The recovered oil was the same
as the original oil, as shown in Table V, although the treat- ment was far more severe than in any ordinary extraction.
Table V-Effect of Boiling Benzene o n Creosote Oil
DISTILLATION ORIGINAL OIL DIGESTION OIL AFTER BENZENE
c. Per cent Per cenf u p t o 210
210-230 230-250 250-270 270-290 290-315 315-355
0 . 2 1 . 5 7 . 2
1 2 . 8 1 0 . 7 1 0 . 4 1 6 . 1
0.05 0 . 7 " 7 . 0
1 2 . 5 10 .8 1 0 . 5 16 8
Residue 4 1 . 1 4 1 . 5 Small amount of loss in removing benzene.
Table VI-Analyses of Oils after Exposure i n Treated Wood and Open Dishes i n the Air
ORIGINAL TEST PIECE OPEN DISH OPEX DISH OIL No. 1E X O . 1 h'o. 2
Time exposed, days . . . 263 530 1005 Per cent loss 1 9 . 4 3 7 . 6 4 1 . 3 Specific gravity i : 059 1 063 1 10 1.103 Per cent tar acids 1 0 . 7 8 . 0 5 . 4 3 . 7
DISTILLATION c. Per cent Per cent Per cent Per cent u p to 210 1 . 5 0 . 0 0 . 0 0 . 0
230-250 1 6 . 5 3 . 3 0 . 1 0 .1 250-270 1 2 . 6 1 4 . 7 0 . 9 0 . 6 270-290 8 . 1 1 4 . 4 5 . 7 1 . 9 290-315 8 . 2 1 4 . 0 2 1 . 1 1 9 . 1 315-355 1 6 . 4 1 8 . 5 2 9 . 4 3 2 . 8 Residue 2 7 . 2 3 2 . 1 4 2 . 8 4 5 . 1
210-230 9 . 5 G , ? 0 . 0 0 . 0
Table VI gives the composition of the recovered oils after exposure for some time, in treated wood and in open dishes, in the air. Table VI1 gives the results for the oil from treated wood exposed in sea water. Table VII-Comparison of Original and Extracted Oilso from
Small Test Pieces after Exposure in Sea Water TEST TEST TEST TEST
OIL 1-L 2-L 3-L 4-L 222 475 512 7117
ORIGINAL PIECE PIECE PIECE PIECE
Time exDOsed. davs ... Per cent'loss 1 1 . 6 1 7 . 5 1 8 . 1 21:9
Per cent tar acids 1 0 . 7 5 . 7 3 . 8 3 . 6 3 . 6 Specific gravity i :059 1 .063 1.067 1 .066 1.067
DISTILLATION c. Per cent Per cent Per cent Per cent Per cenf c'p to 210 210-230 230-250 250-270 270-290 290-315 315-355 Residue
1 . 5 9 . 5
1 6 . 5 1 2 . 6
8 . 1 8 . 2
1 6 . 4 2 7 . 2
0 . 0 2 . 6
1 7 . 8
1 5 . 9 1 4 . 5 1 7 . 8 3 1 . 2
0 . 0 1 . 6
1 2 . 0 1 0 . 6 17 .6 3 3 . 5
{;E 0 . 0 1 . 7
1 1 . 0 1 3 . 9 1 1 . 5 1 2 . 0 1 6 . 3 3 3 . 1
0 . 0 1 . 2 7 . 8
14 .1 1 0 . 7 1 2 . 2 1 9 . 6 3 4 . 2
4 The oils from the different layers of these test pieces were analyzed separately, but only the average analysis for each block is given, since the differences in composition between layers were small. There was, however a slightly smaller amount of low-boiling constituents and tar acids in the' outer layers.
The composition changes found in the oil in this series of tests are practically the same as those found in the earlier experiments. The loss falls on the low-boiling fractions, and a considerable portion of the tar acids is lost. The chief difference between the oils extracted from the water-exposed and air-exposed pieces lies in the percentage of tar acids. The rate of loss of tar acids from treated wood in water is greater than from treated wood in air, in spite of the fact that the rate of loss of the total oil is greater in the air. Analyses of oils extracted from piling show similar differences in air and water sections. A typical analysis is given in Table VI11 for a pile pulled from the Alameda Wharf of the Associated Oil Company after four years' service. Table VIII-Comparison of Oils Extracted from Air, Water, and
Mud Sections of a Pile after Four Years' Service ORIGINAL AIR WATER MUD
0 1 1 . ~ SECTION SECTION SECTION Per cent oil in treated wood 4 9 . 9 5 4 . 1 5 5 . 0 Specific gravity i : 0 6 + 1 .078 1 . 0 7 4 1 .072 Per cent tar acids . . . 3 6 3 . 0 3 . 3
c. Per cent Per cent Per cent Per cent DISTILLATION
u p to 210 210-235 235-270 270-315 315-355 Residue a Manufacturer's analysis.
0 . 0 0 . 0 0 . 0 0 . 0 5 . 5 1 . 1 1 . 8 1 . 8
2 3 . 5 1 6 . 4 2 1 . 5 2 1 . 9 2 3 . 1 2 6 . 5 2 5 . 7 2 5 . 9 1 9 . 3 2 4 . 3 2 1 . 8 2 1 . 3 2 8 . 0 3 1 . 4 2 8 . 8 2 8 . 8
1238 ISDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 19, No. 11
The foregoing results further emphasize the validity of our former conclusions that the low-boiling fractions should be limited to the amount necessary for good penetration and that the tar-acid specification should refer to the fractions above 235 O C.
As already noted, there are slight differences in composition in the oil recovered from different layers of the treated wood. This condition, however, is apparently one which arises after treatment when an equilibrium rate of loss has been established, since studies of successive layers in recently treated piles indicate that creosote penetrates the wood in an unaltered condition. That is, with straight or whole-run creosotes, there is no tendency toward filtration of the heavier constituents in the outer layers of wood.
suggested that the low-boiling fractions of creosote are not entirely lost, but that part of them polymerize or condense to higher boiling constituents, thus becoming more or less permanently fixed in the wood. Some of our earlier results suggested that this was the case, but very careful further study of this point gave only negative results. Although changes in composition were found in the direction to indicate polymerization, they were too small to be of any significance, and were probably due to evaporation losses.
The idea has also been put forward that certain creosote constituents are fixed in the wood by an actual combination with the wood. Any such combination must be very loose, since we have shown that it is possible to recover creosote completely by a benzene extraction. Furthermore, careful analyses of the wood itself, before treatment and after removal of the creosote, show no significant differences. Table I X gives the results of these analyses.
FIXATION O F CREOSOTE CONSTITUEKTS-It has been
Table IX-Analysis of Wood Substance of Douglas Fir Sapling before and after Creosote Treatment
BENZENE- ALCOHOL- SOLUBLE SOLUBLE
EXTRACT^ EXTRACT CELLULOSE LIGNINS 3 f A N N A N
Per cent Per cent Per cent Per cent Per cent Before treatment 0 20 3 28 56 66 28 25 7 39 After extraction 0 21 3 77 65 22 28 48 7 61
a All results calculated to 4 0 per cent moisture content.
Inorganic Treatments
For many years various inorganic salts have been used 111 the preservation of wood. The only ones which have attained general use, however, are zinc chloride, copper sulfate, and mercuric chloride. Even these have not been used to any great extent under high-moisture conditions, on account of their large solubility and consequently rapid leaching from the wood.
The value of any treatemnt depend3 on two factors, the immediate effectiveness of the protection afforded and the permanence of the protection. Obviously, even though such substances as those mentioned might afford protection for a short time, they have no value for the permanent preservation of marine piling, since they would soon be lost from the wood
It is readily seen that the high-moisture conditions to which marine piling is subjected make the preservative use of in- organic salts very difficult. Either they must be left in the wood in some very slightly soluble form, in which case their preservative value is presumably cut down, or else leaching must by some means be prevented.
The most readily apparent method for leaving inorganic salts in the wood in a soluble form and still having permanent protection is by the use of oil emulsions of aqueous solutions. Solutions or dispersions of metallic compounds in oils have also been tried and patented, but have not been commercially successful on marine piling. All available petroleum oils are lighter than water. The formation in these oils of stable emulsions of solutions much heavier than water is a matter of
some difficulty. The only high-gravity oils which are appre- ciably cheaper than creosote are certain oil-tar distillates. Emulsions in these oils of such solutions as zinc chloride, copper sulfate, and mercuric chloride would probably possess a considerable protective value, particularly since the oil-tar distillate has been found to possess some protective power in itself.
The methods which we studied chiefly were those involving precipitation in the wood of a so-called insoluble (i. e., slightly soluble), inorganic compound. One is the method of double treatment and the other involves decomposition of the treat- ing solution by some means after treatment.
To get satisfactory penetration and distribution of the second solution in the double treatment method, it is necessary to dry the wood fairly well after the first treatment. In practice this was found to be rather difficult, and it is probably not economically feasible on a large scale. The double treatment idea was t,herefore abandoned, except for prelimi- nary experiments.
A preliminary study was made by treating small Douglas fir blocks with various inorganic substances, including copper sulfide, arsenic sulfide, antimony sulfide, lead chloride, mercu- rous nitrate, mercuric sulfide, and metallic selenium. The pre- cipitation of some of these substances in the wood involved a double treatment. As already noted, this was not thought to be feasible on a large scale. #Hence, a second series of blocks was treated, using single treatment methods believed to be applicable in commercial practice. The treatments were as follows, the dry wood being impregnated with the aqueous solutions a t ordinary temperatures and 90 pounds pressure:
(1 ) Chandler’s Solution. Solution contained 8.5 per cent Na2C03, 1.5 per cent NaHC03, less than 1 per cent CuSOd: 5H30. This solution must be used soon after preparation, since it gradually decomposes until the copper content is less than 0.5 per cent CuS04.5H10.
(e) Impregnation in cold. No after-treatment. (Chandler’s method.)
( b ) Impregnation in cold. Steamed after-treatment to
.
precipitate CUZO in the wood. (2) NurSe03 + NaOH. Solution contained 5 per cent sele-
nium dioxide (SeOa) and 0.5 per cent free NaOH. Steamed after-treatment to precipitate metallic selenium in the wood.
(3) CuSOl + NaaSzOa. Solution contained 5 per cent CuS04.5HzO and 12 per cent Na~Sh08.5H20. Steamed after- treatment to precipitate CuS and sulfur in the wood. (4) HgNO3. Solution contained 1 per cent HgNOa.2HzO.
No after-treatment, Sea water precipitates calomel in the wood. (5 ) HgClz + Na2S203. Solution contained 1 per cent HgClz,
4 per cent Na&03.5H20, and 2 per cent NanCOs. Steamed after-treatment to precipitate HgS. Hg20, and other decompo- sition products of this solution.
These treatments utilize two ideas: (1) the use of solutions of metallic salts, which precipitate relatively insoluble metallic compounds on heating; and (2) the utilization of the reducing power of the wood in precipitating relatively insoluble metallic compounds in the wood. The second method has not to our knowledge been used prior to the work of this committee.
Sodium thiosulfate is useful in making up both types of solutions. An excess of sodium thiosulfate added to a so- lution of a heavy metal salt, such as copper sulfate, or mer- curic chloride, gives a solution which is stable for a long time a t room temperature but which decomposes on heating, giving a precipitate of the metallic sulfide and some free sulfur.
Note-In the case of the mercuric chloride the stability of the solution is somewhat increased by the addition of a little NaaCOs or NaOH.
Solutions of this type can also be made alkaline without the formation of any precipitate. They fall then in the second class.
The reducing power of the wood is due to the carbonyl groups in the cellulose molecule, and is only active in alka- line solutions.
November, 1927
TgiE C H L O R I S E CONCEXTR.4TIOX
P O S U R E EX- No. P O S U R E ~ Initial Final .4veraee
IYDUSTRIAL AND ENGINEERING CHEMISTRY
REACTIOS O F PERMANENT BORERS EFFECT
1239
H o u r s Block No. 1
1 24
.Vote-Determination of the reducing power of Douglas fir sapwood on Fehling's solution gave a value of 0 14 pound of CuSOc rt.duced to Cur0 per pound of wood. The determination was made on sawdust which had been extracted with benzene and alcohol to remove all except the structural material of the mood.
Since most metals precipitate in ordinary alkaline solutions, special solutions must be used, such as the thiosulfate so- lution abore.
The timber is impregnated with the alkaline solution a i d subsequently heated, whereupon the reducing action of the wood causes precipitation of a metallic oxide or the free metal. The heating merely speeds up the reaction.
Treatments (14) and (2) depend entirely on the reducing power of the wood. In (3) the decomposition depends on heat alone. Treatment (5) gives a decomposition which is a mixture of the two types.
Four sets of test pieces with the above treatments were placed a t various stations in the bay. All showed some retardant action against borers, but within one season all were attacked to some extent. Some of the treatments may have value for other purposes, but apparently none of the inorganic treatments which we have studied have any permanent value in the protection of marine piling against borer attack. This conclusion is strengthened by the rapid failure in marine use of proprietary metallic salt preparations, such as colloidal metallic selenium and tellurium in the protection tests of this committee, and of mercuric chloride and oleate as tested in the East.
Effects of Chlorine on Teredos
The following experments were carried out a t the request of the Kational Research Council for the purpose of deter- mining the effects of chlorine liberated in sea water upon adult teredos in their burrows.
Three pieces of 2 by 4-inch timber infested by Teredo navalis were obtained from Dumbarton through the com- mittee biologist. One, used as a control, was kept in fresh bay water through the test. The other two were exposed to various concentrations of chlorine in bay water.
The experiments extended over several weeks, during which time the teredos were alternately exposed to chlorine and fresh bay water. The time of exposure to chlorine varied from 24 to 72 hours. After being exposed to chlorine, the borers were usually left for about 24 hours in fresh bay water, but in all cases observed the activity of their siphons indicated complete recovery in a much shorter time.
The concentrations of chlorine possible when the gas is liberated in sea water under no pressure other than atmos- pheric are necessarily very small. The initial concentrations of chlorine used in our experiments were from 30 to 120 parts per million parts of sea water. These concentrations dropped off rapidly since the containers were not gas-tight. In prac- tically all cases, however, there was sufficient chlorine to give a distinct odor during most of the time of exposure.
In small concentrations, below 10 p. p. m. or practically the point where the odor of chlorine over the solution becomes no- ticeable, the teredos continued to feed, seemingly unharmed. It must be remembered that in maintaining even these small concentrations of chlorine considerable amounts will be given off into the surrounding atmosphere. At higher concentrations the siphons were always immediately withdrawn. However, after the highest concentration used (120 p. p. m.), applied for the longest time tried (72 hours), on the stick being placed in fresh sea water (chlorine-free) the borers again extended their siphons and were apparently none the worse for their experience. This concentration of 120 p. p. m. was sufficient to give off into the air enough of the gas to make human breathing very uncomfortable.
P. P. m. P. p. m. P. p . m. . . . Siphons all
retracted 40 <20 30
Table X-Effects of Chlor ine on Teredos i n T h e i r Burrows
40 _ _ 30
< 10 30 xn
173
Siphons all . . retracted
. . . . . . . . . . . . None
4 5 72
Block No._2
1 24
2 24
5 48 6 72
for 24 hours. After each
50 20 50 20 110 20 30 <10
60 20
80 30 40 <"O 30 1 1 0 50 10
120 40
chlorine exposure
35 35 65 20
, . . . . . . . .
Sone
Table X gives a summary of the experimental data. At the end of the series the teredos in all aquaria were apparently i n good health, and there were as many active siphons as a t the beginning. Some of the borers lived for several months, even under the unfavorable conditions of the aquaria. The borers in one of the pieces subjected to the chlorine treatment lived longer than those in the control piece, which had had no chlorine treatment.
Undoubtedly, the chlorine was offensive to the borers, When it was added slowly, the siphons were retracted pro- gressively along the wood in the direction in which the chlorine 3pread through the water. Furthermore, when the infested pieces were taken out of fresh bay water, exposed to the air for a moment and put back into the water again, the teredos extended their siphons and began to feed again after a very few minutes. On the other hand, if they were put into a solution giving a noticeable odor of chlorine, they did not extend their siphons again. They were never found feeding In Concentrations as great as 10 p. p. ni. The concentration of chlorine used in several of the experiments might, therefore, have been toxic to the teredos, had they not been able to close their burrows against the gas. This is one of the major premises overlooked by the promoters of chlorine extermina- tion of marine borers
There remains a possibility that a practically continuous maintenance of small concentrations of chlorine might prevent settling of the larvae on the piling, or that the young teredo larvae might be more sensitive to the action of the chlorine than are the adults. The results of our experiments indicate, however, the improbability that chlorine (entirely aside from its many objectionable features) could economically be main- tained around piles in sufficient concentration for a sufficient length of time to kill the borers when once established, or to be of any practical value in the protection of marine piling.
Conclusions
(1) Straight-run or whole creosotes penetrate wood in an un- altered condition; that is, there is no tendency toward filtra- tion of the heavier constituents in the outer layers of the wood.
( 2 ) There is no fixation of creosote constituents in the wood, and low-boiling substances do not polymerize to any appreciable extent to form high-boiling substances.
(3) The loss of creosote from treated wood is due to the combined volatility and solubility of the constituents in the low-boiling fractions, and the loss falls almost entirely on these low-boiling constituents.
(4) Creosote constituents boiling below 235" C. are largely lost from treated wood within two or three years; therefore, the percentage of these low-boiling constituents in the oil should be limited to the amount necessary for good pene-
1240 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 19, No. 11
tration. Also the tar-acid specification should refer to the (6) Inorganic substances, by themselves, have little pros- fractions above 235" C. pect of value in the preservation of marine piling against
( 5 ) Certain oil-tar distillates have a greater protective borer attack. value against marine borers than that with which they have (7) Chlorine treatment has no practical value in the pro- ordinarily been credited. tection of piling against marine borers.
Deterioration of Mineral Oils 11-Mechanism of Oxidation and Action of Negative Catalysts as Determined
by Static Methods'sz By Brian Mead and Collaborators
DEPARTMENT OF CHEMICAL ENGINEERING, MASSACHUSETTS INSTITUTE OR TECHNOLOGY, CAMBRIDGE, MASS.
N A previous paper on this subject3 results were shoxn of experiments, made by a dynamic method, to determine the mechanism of oxidation of mineral oils. This method
of testing by bubbling oxygen through a heated sample of oil, while generally satisfactory when the oil sample oxidizes a t a relatively high rate, was used with difficulty when the experiment was of long duration or when a large number of samples were to be studied, because small differences in oxy- gen supply made large differences in the deterioration. In one set of experiments oxygen was bubbled through several samples of oil by way of pieces of capillary tubing as nearly as possible of the same diameter. No concordance between the results was possible, however, on account of slight differences in internal diameter of the tubes. Another dis- advantage of this method is that by the passage of the stream of oxygen through the oil some of the volatile products of the oxidation are removed from the system, and since these may have an appreciable influence on the progress of the reaction, their removal may c?use erroneous results.
The mechanism of oxidation can also be determined by a static method, in either an open or a closed system. For the former, the method.consists in heating samples of oil in test tubes, the mouths of which are open to the air. By plac- ing a batch of such tubes in an air oven maintained a t con- stant temperature it is possible to obtain fairly satisfactory and concordant results. However, in this case, too, only the effects produced by the oxidation can be measured, and not the amount of oxygen producing them. This method, or a modification of it, is the principle of the French and the Sligh oxidation tests. In the Sligh test oxygen is substituted for air and the flasks in which the oxidation takes place are stoppered.
A more satisfactory method seems to be to have the oil and oxygen in an enclosed space. As oxygen is absorbed by the oil, the amount so used can be determined directly by actual measurement. To determine the resulting change in the physical and chemical properties of the oil, it is neces- sary to discontinue the oxidation and analyze the oil. This method has one disadvantage-that it requires a separate apparatus for each determination. The apparatus can, hov- ever, be so constructed that it is necessary to disconnect only the part in which the oxidation takes place, using the remainder for the next run in the series. By having a num-
I
1 Combination of papers presented under the titles "Mechanism of Static Oxidation," "Prevention by the Use of Antioxidants," and "Mech- anism of Static Deterioration of Transformer Oils," before the Di- vision of Petroleum Chemistry at the 72nd Meeting of the American Chemical Society, Philadelphia, Pa., September 5 to 11, 1926.
2 The experimental work reported in this paper was carried out during 1924.
8 Haslam and Frolich, THIS JOURNAL, 19, 292 (1927).
ber of reaction vessels made up ahead of time, it is a com- paratively simple matter to arrange the experiments so that they are practically continuous. Thus, for any given oil, if it is desired to determine oxygen absorption, acidity de- velopment, and sludge formation over a period of 20 days, it will be necessary to make twenty separate runs, the first lasting for one day, the second for two, and so on. It is possible by this method to get accurate oxygen absorption data for the whole period, and so to trace any unusual or unexpected trend in the reaction.
The results reported in this paper were obtained by static methods, using both open and closed static systems. Part A includes the results from the closed static method, study- ing the effects of temperature, surface exposure, degree of unsaturation of the oil, acidity development, and sludge for- mation. Part B consists of a series of accelerated tests on S u j o l and a commercial transformer oil (No. 10 Transil Oil) by the open static method. Part C describes a study of the effect of various antioxidants on the sludge formation in transformer oils as determined by the open static method.
PART A-STUDY OF OXIDATION BY CLOSED STATIC METHOD By Brian Mead, A. C. Phelps, C. R. Washburn, and J . P. Warner
Experimental
One of the most difficult factors to decide upon and to con- trol was the temperature of the reaction. Oils are rarely ex- posed t o temperatures greater than 90" to 100" C. in trans- formers. At this temperature, however, the oxidation is relatively slight, so that in order to get measurable amounts of the products of oxidation, or of oxygen absorption, it would be necessary to use a very large sample of oil. Many of the accelerated tests have been run a t temperatures as high as 150°, or even 200" C. At 200" C. there is not only conaider- able fire hazard from the sample being tested and from the heating bath, but difficulty is experienced in maintaining a heating bath. Even a t 150" C. both these factors are of importance. Accordingly, it was decided to operate a t a temperature of 130" C.
The oil used for the heating bath was a cylinder stock of high flash and low free-carbon content. Bulb heaters were employed, since any type of open coil heatei gave such quan- tities of sludge that the bath could be operated for only very short periods. Efficient stirrers were installed, in order to maintain the temperature throughout the bath constant within 0.5" C. The heaters were so adjusted in value that the intermittent one, operated through a relay by a mercury thermostatic regulator, was in operation for the minimum amount of time.
The apparatus is shown 'diagrammatically in Figure 1.
