This is an author version of the contribution published on:

Questa è la versione dell’autore dell’opera:

Industrial & Engineering Chemistry Research 2014

Volume 53, fascicolo 21, pp 8675–8684, DOI: 10.1021/ie4032814

The definitive version is available at:

La versione definitiva è disponibile alla URL:
http://pubs.acs.org/doi/abs/10.1021/ie4032814

1
 Stability and reactivity of sulfur compounds against

copper in insulating mineral oil: definition of a

corrosiveness ranking

Maria Concetta Bruzzoniti*†*, Rosa Maria De Carlo†, Corrado Sarzanini†, Riccardo Maina‡,

Vander Tumiatti‡

†*Department of Chemistry, University of Torino, Via P. Giuria 5, 10125 Torino (Italy)

‡ Sea Marconi Technologies, Via Ungheria 20, 10090 Collegno (Torino)

AUTHOR INFORMATION

*Corresponding Author

Maria Concetta Bruzzoniti

Department of Chemistry, University of Torino, Via P. Giuria 5, 10125 Torino Italy

Ph: +39 011 6705277

Fax: +39 011 6705242

E-mail: [email protected]

2
ABSTRACT

Corrosive sulfur is a well-known critical issue for mineral insulating oils used in power

transformers. The objective of this work is to study the reactivity of different groups of sulfur

containing compounds (aliphatic and aromatic mercaptans, sulfides, disulfides, thiophenes) that

may be present in insulating oils. The stability of nine compounds, pertaining to the

aforementioned groups, was investigated with regards to temperature and aging time, under

oxidative and inert atmosphere, in order to simulate the natural aging of both hermetically sealed

or free breathing type transformers. Their reactivity against copper was also investigated. A gas

chromatography atomic emission detection method was optimized to quantify the consumption

of each compound. The tendency of 22 sulfur compounds to react with copper and form copper

sulfide was quantitatively assessed at different temperatures. The resulting corrosiveness ranking

for sulfur compounds families differs from what is usually reported in literature, and was proven

to be temperature dependent.

KEYWORDS: insulating oil, corrosive compounds, corrosive sulfur, corrosiveness ranking

3
 INTRODUCTION

An increasing occurrence of copper sulfide formation in paper-insulated windings of power

transformers and shunt reactors has been observed in the last decades. Some failures that

occurred in large power transformers have been ascribed to this phenomenon, as a result of

chemical reactions occurring between copper conductors and insulating materials (Kraft paper

and oil).1-3 It is worldwide accepted that corrosion is mainly promoted, at the standard operating

temperature of the transformer, by the reaction of some forms of sulfur, defined as "corrosive

sulfur", and the naked copper surfaces, resulting in the formation of copper sulfide.4

Sources of sulfur compounds are numerous. Crude oil, from which insulating oils are obtained,

has a relatively high sulfur content, a great deal of which is reactive towards copper components

of the transformer.5 Thanks to the improvements in the petroleum refining processes (hydro-

treating) and the increasing demand of petroleum derivates with a low sulfur content, in the last

25 years there was a marked reduction of the concentration of total sulfur in insulating oils.6-8

Nowadays, the amount of sulfur contained in a typical transformer oil may range from 0.001 to

0.5 %.9 As a side effect, the same refining processes, that reduce the total amount of sulfur, lead

to a decrease in the concentration of aromatic and hetero-atom compounds that act as natural

antioxidants.10,11 This sometimes requires the restoring of part of the original sulfur compounds

by the so-called “back-blending”, a common practice that entails the mixing of heavier refining

streams (rich in sulfur and aromatic compounds) with the highly refined base oil. Among the

crude oil -derived sulfur compounds some are beneficial to the oils, as they act as retardants or

passivators in oxidation reactions, but others may cause corrosion.4 As a consequence, the

reduction in total sulfur amount apparently promotes an increase in the occurrence of corrosive

sulfur.

4
 Another source of corrosive sulfur in insulating oil can be the chemicals usually added to

improve the antioxidant properties.12

As a result of the above described causes, sulfur can be present either in the elemental form or

contained in organic molecules (e.g.: thiophenes, disulfides and polysulfides, thio-ethers and

mercaptans). Not all these substances are corrosive, or have the same reaction rate against

copper: the degree of corrosiveness of a mineral oil is usually a function of the type and

concentration of compounds present, the working temperature and the presence of oxygen.4

A generally accepted corrosiveness ranking is: S x > R-SH > R-S-R >> R-S-S-R >>

thiophenes.13 Elemental sulfur (usually in the S 4 or S 8 form) is considered highly reactive against

copper, while disulfides are relatively stable and thiophenes are not reactive. Nevertheless, very

few data about their reactivity and degradation trend are available in literature. Jian et al.14

investigated the effect of corrosive sulfur at different temperatures, and Lukic et al.15 studied

different ways to remove sulfur compounds from insulating oils. Unfortunately neither of them

detailed the types of sulfur species present. Maina et al.16 started in 2009 to study the degree of

corrosiveness of dibenzyldisulfide. Recently, only Kamishima17 investigated how changes in the

molecular structure of sulfur compounds in mineral insulating oils affect their corrosiveness.

The fact that the majority of the studies on corrosiveness in mineral insulating oils were

focused on dibenzyldisulfide (DBDS)5, 18-20

, a sulfur compound considered relatively

stable13,reveals that the traditional reactivity scale may not reflect the reality. Previous studies

highlighted that DBDS reduction to benzyl mercaptan can result in more corrosive properties16,10

and that by-products such as bis-benzyl and dibenzyl sulfide can be present in in-service mineral

oils.16,21 Finally, although the corrosiveness of almost 90 % of corrosive oils in transformers is

determined by DBDS,16 a significant amount of insulating oils were found to be corrosive, and

5
namely capable to transfer copper sulfide onto the Kraft paper tapes, independently of the

presence of DBDS. This shows that most of the sulfur compounds responsible for the corrosion

process remain unidentified.

For the above mentioned reasons, the aim of this work was to investigate the stability and the

potential corrosive behavior of different types of organic sulfur compounds towards copper and

paper insulation. Therefore, different reaction conditions were explored: the behavior of the

tested sulfur compounds was investigated at temperatures ranging from 80 to 200 °C, in the

presence and absence of oxygen and bare copper. Experiments were conducted at different

contact times (72 – 240 hours) to elucidate the rate of depletion of each compound. A

corrosiveness ranking was assessed by determining the total corrosive sulfur (TCS) for each

compound at different temperatures. The corrosiveness of each compound, obtained by TCS test,

was compared with the results of the official method for corrosive sulfur detection (ASTM

D1275 B).

EXPERIMENTAL SECTION

Materials

Isooctane was from Panreac (Barcelona, Spain), Dibenzothiophene (≥ 99 %, DBTf), Dodecyl

methyl sulfoxide (98 %, DcMSOx), Phenyl vinyl sulfide (97 %, PVS), sec-Butyl sulfide (98 %,

sBS), Dibenzyl disulfide (98 %, DBDS), Diphenyl disulfide (99 %, DPDS), Isopropyl disulfide

(96 %, IPDS), Benzyl mercaptan (99 %, BMt), Benzene-1,3-dithiol (99 %, BDT),

Butylmercaptan (99 %, BuMt) and Heptylmercaptan (98 %, HeMt) were from Sigma-Aldrich

(St. Louis, MO, USA). Dibenzyl sulfoxide (DBSOx), Dibutyl sulfoxide (DBuSOx), Dibenzyl

sulfide (DBS), 2-Methyl-1,3-dithiane (MeDT), Dodecyl Mercaptan (DCMt), Hexadecyl

6
Mercaptan (≥ 95.0 %, HDMt), 1,2-Ethanedithiol (EtDt) and Cyclohexanethiol (97 %, CET) were

from Acros Organics (Thermo Fisher Scientific, Geel, Belgium). Finally, Pentamethylene

Sulfide (PMeS) and Butyl Disulfide (BuDS) were from Wako Pure Chemical Industries (Osaka,

Japan), and Thioanisole (TAS) was from VWR International (Milano, Italy).

Chemical structures and abbreviations of each compound are collected in Table S1 of the

Supporting Information.

The samples were prepared in a sulfur-free light mineral oil (Sigma Aldrich, Milan, Italy—

d(25 °C) = 838 to 854 kg/m3; kinematic viscosity at 40 °C = 14.2 to 17.0 cSt) at a concentration

of 40 mg/kg expressed as elemental sulfur for each compound. Samples were diluted 1:10 v:v in

isooctane before injection.

For the reactivity study and for the ASTM test, copper foils of 4 mm x 10 mm and 6 mm x 25

mm, respectively were used.

Vials of 20 mL (22 mm diameter) with PTFE/butyl rubber septa and aluminum crimp caps

(Agilent) were used for total corrosive sulfur (TCS) determination.

Methods

The amount of each compound during the corrosiveness study was determined using a 6890

Model gas chromatograph (Agilent Technologies, CA, United States) equipped with a MPS2

autosampler (Gerstel GmbH, Germany) and coupled with an Atomic Emission Detector (AED)

from Joint Analytical Systems GmbH, Germany. A HP-1 Methyl Siloxane capillary column

(60m×250μm×0.25μm) from Agilent was used.

7
 Corrosiveness studies

Reactivity study in white oil (light mineral oil). With the objective of simulating the stress

conditions to which sulfur compounds may be subject in power transformers during service, the

influence of different parameters was investigated: temperature (to simulate the thermal stress),

presence of oxygen (to understand the influence of the conservator segregation mode—sealed or

free breathing) and contact with copper conductor.

The Design Of Experiments (DOX) can be detailed as follows: 1.5 ml of white oil containing

nine selected compounds (dibenzothiophene, dibenzyl sulfide, phenyl vinyl sulfide, dibenzyl

disulfide, diphenyl disulfide, butyl disulfide, benzyl mercaptan, dodecyl mercaptan, hexadecyl

mercaptan) at 40 mg/kg each (expressed as elemental sulfur) were put in 2 ml glass vials for gas

chromatography, sealed with rubber septum and artificially aged at 80°C, 100°C, 120°C, 150°C,

200°C. At each temperature, aging times of 36, 72, 144 and 288 hours were applied.

The behavior of the nine compounds was studied, according to the procedure detailed above,

under the following conditions:

-thermal aging (Heating): 1.5 ml of the solutions aged in sealed vials, without contact with air

or copper;

-thermal-oxidative aging (Heating + O 2 ): 1.5 ml of the solutions aged in sealed vials, intake of

air through a stainless steel needle (i.d. 0.8 mm) breaking through the septum without contact

with copper;

-non-oxidative corrosion (Heating + Cu): 1.5 ml of the solutions aged in sealed vials,

containing a piece of naked copper strip (72 mm2 total surface) , without contact with air;

8
 -oxidative corrosion (Heating + O 2 + Cu): 1.5 ml of the solutions aged in sealed vials, intake

of air through a stainless steel needle breaking through the septum (i.d. 0.8 mm), and containing

a piece of naked copper strip (72 mm2 total surface).

Copper strips were polished before use with sand-paper and acetone; the procedure was

derived by method ASTM D1275.

The residual amount of each compound was determined by an optimized GC-AED method

(optimized experimental conditions are reported in Table S2 of the Supporting Information) on

diluted samples (1:10 with isooctane) comparing chromatographic peak areas before and after

aging.

Corrosiveness ranking by total corrosive sulfur (TCS) determination. To estimate the tendency

of different families of sulfur compounds to form copper sulfide as a consequence of the contact

with bare copper surfaces, we measured the amount of Cu 2 S formed by 22 different compounds

(listed in Table S1, Supporting Information) under different temperature conditions. This test

was applied as a modification of the Total Corrosive Sulfur (TCS) Test previously described22.

Individual solutions, containing each sulfur compound at 40 mg/Kg (expressed as elemental

sulfur), were prepared in light mineral oil; aliquots of 10 g of oil were put in contact with 3 g of

copper powder and heated and stirred for 12 hours, under Argon atmosphere. Finally, the oil was

pipetted out and the powder rinsed twice with isooctane and pentane. Copper sulfide was formed

and then oxidized to sulfate by KNO 3 at 370°C for 2 hours. Sulfate ions were extracted in

deionized water and measured by turbidimetry.

The resulting TCS is calculated according to the following equation:

TCS = [(C Sulfate / 96.06) x V sol x 32.07] / kg oil

9
 where C Sulfate is the sulfate concentration (mg/L), V sol the volume of the solution in which

sulfates are extracted (L) and kg oil is the weight of oil sample. The corrosiveness has been

calculated as corrosion yield, using the formula:

R c = 100 x [TCS]/[TCS] theoretical

where [TCS] theoretical is the total corrosive sulfur value for a quantitative conversion of the

sulfur compound in Cu 2 S.

TCS has been applied as an internal method, and its reproducibility has not yet been evaluated

at the time of preparation of this paper. All the results reported in this paper were obtained as

average of duplicate tests (spread between the two replicates ranged from 3 to 25 %). TCS’

reproducibility should be expected to be quite large, due to the number of steps involved inside

the test procedure. Therefore, results obtained by TCS test shall be considered as comparisons of

the order of magnitude of the corrosiveness yield of each individually tested compound.

TCS tests were performed at the following temperatures: 80 °C, 100 °C, 120 °C, 150 °C and

180°C.

Corrosiveness estimated by standardized method. Each solution of the 22 compounds studied,

containing 40 mg of elemental sulfur per kg of light mineral oil was tested by an official method

(ASTM D1275 B23) to evaluate the corrosiveness. Accordingly with ASTM D1275 B procedure,

a naked copper strip is put in contact with 250 ml of oil at 150 °C for 48 hours. The oil is

classified corrosive when the copper sulfide film formed upon the conductor surface is black,

dark grey or dark violet colors, according to the color scale given in the ASTM D130/IP 154.

In order to verify the absence of corrosive effect from the technical oil, a blank sample

(technical new oil without sulfur compounds) was processed.

10
 In order to compare the results obtained in the reactivity study (see Results and Discussion

section) and in the corrosiveness ranking study (see Results and Discussion section), ASTM

D1275 B method was performed at different temperatures: 80°C, 100°C, 120°C, 150°C and

180°C.

RESULTS AND DISCUSSION

Selection of sulfur compounds

The behavior of 22 substances belonging to the following families of sulfur compounds were

investigated (table S1): thiophenes (1), sulfoxides (3), sulfides (6), disulfides (4) and mercaptans

(8). Compounds were chosen in order to cover the corrosiveness scale and to include molecules

actually found in in-service transformer oils (DBDS, DBS, BMt and DBTf).

In order to correlate the structure with the corrosiveness, within each class, where possible,

both aliphatic and aromatic compounds were selected.

Reactivity study

Among possible faulty conditions that can occur in transformers, overheating is one of the

most important.24 Copper sulfide formation on conductors and Kraft paper tapes has been often

related to increased temperatures, determined by either faulty conditions or severe thermal

profile.25 Consequently, it is important to assess the stability of different sulfur compounds under

different thermal stress conditions.

The reactivity study was performed on nine selected compounds (BMt, PVS, BuDS, DCMt,

DBTf, DBS, DPDS, HDMt, and DBDS) representative of the above mentioned classes. A GC-

AED method was optimized with the aim of quantitatively estimating the consumption of each

11
compound after the aging step. Analytical conditions are summarized in Table S2 (Supporting

Information). This method allowed the separation and detection of the nine compounds, in 46

minutes (Figure 1).

300

BMT
BuDS DCMT

DPDS

DBS HDMT
Counts PVS
DBDS
DBTF

0 10 20 30 40 50

Time (min)

Figure 1: Gas chromatographic separation of the nine selected sulfur compounds.

Chromatographic conditions are reported in Table S2 (Supporting Information).

The reactivity of BMt, PVS, BuDS, DCMt, DBTf, DBS, DPDS, HDMt, and DBDS was

studied at five different temperature levels: 80°C, 100°C, 120°C, 150°C and 200°C. These levels

were chosen to include the operating temperature of a transformer under normal conditions (<

105°C), and simulate a slight overheating (105°C-200°C), typical of many electrical in-service

equipment. The role of oxygen and copper in degradation was also studied.

As a general consideration, compounds degrade faster with the increase of reaction conditions

harshness (augmented temperature and time of exposure). It has been generally observed that

harder reaction conditions promote a faster compound degradation, especially if oxygen and

copper are present. This behavior is shown for DBDS (figure 2), as a general example.

12
 120 120
residual % after reaction

residual % after reaction

a) b)
100 100

80 80
80°C without Cu and O2
60 100°C 60

120°C Cu
40 40 O2
20 20
200°C 150°C
Cu + O2
0 0
0 50 100 150 200 250 300 0 50 100 150 200 250 300

time (h) time (h)

Figure 2: Degradation of DBDS. Conditions: (a): thermal aging in presence of copper; (b):

oxidative degradation in presence and without copper at T= 150 °C.

The behavior of the various classes of substances, at different temperatures, is hereafter

described (Figure 3a-3e).

Reactivity and depletion rate at 80 °C (figure 3a): mercaptans (BMt, DCMt, HDMt) are stable

with temperature and moderately influenced by the presence of oxygen. Butyl mercaptan and

phenyl vinyl sulfide are the most and the least stable compounds, respectively. Mercaptans and,

to a minor extent, disulfides exhibit reactivity towards copper; in presence of copper, reactivity

of disulfides is fairly influenced by oxygen. Conversely, dibenzylsulfide and mercaptans

degradation is enhanced by oxygen in the presence of copper.

13
 BMT DCMT HDMT BMT DCMT HDMT
80°C PVS DBS DPDS
80°C + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h
BMT DCMT HDMT BMT DCMT HDMT
80°C + Cu PVS DBS DPDS
80°C + Cu + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h

Figure 3a. Thermal and oxidative degradation in presence and without copper at 80 °C

Reactivity and depletion rate 100 °C (figure 3b): mercaptans and disulfides reacts with metal

at exposure times lower than what observed at 80°C. The reactivity of sulfides towards copper is

observed only in presence of oxygen. As a general consideration, aromatic compounds seem to

be more stable than aliphatic ones (e.g: compare the behavior of dibenzyl disulfide with butyl

disulfide).

14
 BMT DCMT HDMT BMT DCMT HDMT
100°C PVS DBS DPDS
100°C + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120

residual %
residual %

100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h
BMT DCMT HDMT BMT DCMT HDMT
100°C + Cu PVS DBS DPDS
100°C + Cu + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h

Figure 3b. Thermal and oxidative degradation in presence and without copper at 100 °C

Reactivity and depletion rate 120 °C (figure 3c): heating under inert atmosphere does not

induce appreciable depletion of the sulfur compounds. If oxygen is present, some sulfides and

mercaptans are significantly consumed. All the mercaptans and disulfides have a strong

reactivity towards copper.

15
 BMT DCMT HDMT BMT DCMT HDMT
120°C PVS DBS DPDS
120°C + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h
BMT DCMT HDMT BMT DCMT HDMT
120°C + Cu PVS DBS DPDS
120°C + Cu + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h

Figure 3c. Thermal and oxidative degradation in presence and without copper at 120 °C

Reactivity and depletion rate 150 °C (figure 3d): at this temperature, thermal degradation

occurs. Oxidative degradation as well as reactivity towards copper is enhanced; in particular,

mercaptans are totally degraded within 36 hours of reaction, showing a higher corrosiveness

effect compared with sulfides and disulfides. It is worth mentioning that dibenzothiophene is the

only compound that does not show reactivity towards copper.

16
 BMT DCMT HDMT BMT DCMT HDMT
150°C PVS DBS DPDS
150°C + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h
BMT DCMT HDMT BMT DCMT HDMT
150°C + Cu PVS DBS DPDS
150°C + Cu + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h

Figure 3d. Thermal and oxidative degradation in presence and without copper at 150 °C

Reactivity and depletion rate 200 °C (figure 3e): strong thermal degradation is observed for all

compounds, except for dibenzothiophene. In the presence of oxygen 80 % of DBTF is still

unreacted (50% residual after 288 hours also in oxidative conditions). These data suggest a non-

corrosive behavior of dibenzothiophene, even in the event of harsh temperature conditions of

open-air transformers. This was also recently suggested by Kamishima et al.,17 who have

evaluated the absence of corrosiveness of a dibenzothiophene fortified oil through ASTM D1275

B, and verified the corresponding negligible formation of sulfur on the surface of a copper strip

by wavelength dispersive X-ray fluorescence.

17
 BMT DCMT HDMT BMT DCMT HDMT
200°C PVS DBS DPDS
200°C + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120

residual %
residual %

100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h
BMT DCMT HDMT BMT DCMT HDMT
200°C + Cu PVS DBS DPDS
200°C + Cu + O2 PVS DBS DPDS
DBDS BuDS DBTF DBDS BuDS DBTF
120 120
residual %

residual %
100 100
80 80
60 60
40 40
20 20
0 0
0 100 200 300 0 100 200 300
h h

Figure 3e. Thermal and oxidative degradation in presence and without copper at 200 °C

To summarize, it can be stated that all nine sulfur compounds are quite resistant to thermal and

oxidative degradation, at the temperatures close to the normal working condition of transformers

(80°-120°C). As a general rule, air (oxygen) accelerates the decomposition at high temperatures.

For a temperature higher than 120°C, other than DBTF, only the compounds belonging to the

disulfide family remain stable (especially the aromatic ones), while substances in the mercaptans

and sulfides families are quickly depleted, in particular in the presence of oxygen.

Mercaptans exhibit a high reactivity to copper even at low temperature, whereas sulfides and

disulfides require a higher temperature (above 120°C). This result is in accordance with the

theory of a peculiar reaction path of DBDS, postulated by some researchers13, that ascribes its

corrosiveness towards copper to DBDS degradation to benzyl mercaptan (BMt).

18
 The copper pieces collected after the aging experiments gave a qualitative evaluation of the

corrosion phenomena (figure 4).

Figure 4. Images of copper conductors after the aging experiments according to the DOX.

Copper pieces become progressively darker with increasing temperature and aging time,

mainly due to the formation of Cu x O, Cu x S y and other insoluble compounds. It is noteworthy

that, at 80°C and 100°C, the presence of oxygen considerably enhances copper corrosion, in

19
particular after 144 and 288 hours. At higher temperature no particular differences are observed

between the pieces in contact with air and those aged without direct exposure to air.

Finally, it can be hypothesized that oxidation and corrosiveness are two competitive reactions:

compounds with a lower thermal stability and/or more prone to oxidation should be less

corrosive at high temperatures, where the thermal/oxidative degradation competes with the

corrosive reaction towards copper.

According to the same hypothesis, compounds with higher thermal stability, but having a good

affinity to copper, should show enhanced corrosiveness at higher temperatures.

Corrosiveness ranking

The reactivity study described in the previous paragraph provides a rough corrosiveness

ranking for the families of tested compounds. In particular, for temperatures comprised between

80 and 120 °C, mercaptans appear to be the most corrosive compounds, whereas between 150

and 180 °C, the highest corrosiveness properties are exhibited by disulfides. To deeply

investigate the dependence of corrosiveness on temperature, and correlate it with the reactivity

data previously obtained, the TCS method was performed at the same temperatures used for the

reactivity study. In this way, the corrosiveness of each compound towards copper was calculated

(and expressed as yield of copper sulfide formed against the stoichiometric theoretical amount of

available sulfur), thus allowing to define a corrosiveness ranking of the investigated compounds

at the considered temperatures.

As reported in table 1, the results are in agreement with what observed during the reactivity

study. In detail, at 80 °C (temperature typical of transformers operating at normal conditions),

20
only mercaptans show corrosion towards copper, with a yield proportional to the number of

sulfur atoms.

Table 1. Corrosion yield for all the 22 sulfur compounds calculated according to TCS tests.

Class Compound Corrosion yield (%)

80°C 100°C 120°C 150°C 180°C
Thiophenes Dibenzothiophene N.D. N.D. N.D. N.D. N.D.
1,2-Ethanedithiol 64.2 49 51.5 35.7 51.9
Benzene-1,3-dithiol 60.6 29.8 78.4 77.5 79.5
Dodecyl mercaptan 33.6 66.2 32.9 50 75.4
Butyl Mercaptan 30.4 28 24.5 45.7 58.9
Mercaptans
Cyclohexanethiol 26.5 30.8 41.8 80.5 91.8
Benzyl mercaptan 24.8 0 50.8 67.5 60.3
Hexadecyl mercaptan 23.9 7.3 23 35 76.8
Heptyl mercaptan 8 0 34.8 73.8 63.5
Dibenzyl sulfoxide N.D. N.D. N.D. 55 100.2
Sulfoxides Dibutyl sulfoxide N.D. N.D. N.D. 45.4 81.9
Dodecyl methyl
N.D. N.D. N.D.
sulfoxide 45.2 69
2-Methyl-1,3-Dithiane N.D. N.D. 11.3 19.2 66.7
Phenyl vinyl sulfide N.D. 22.6 34.7 15 55.3
Dibenzyl sulfide N.D. N.D. N.D. N.D. 77.8
Sulfides
sec-Butyl sulfide N.D. N.D. N.D. N.D. N.D.
Pentamethylene sulfide N.D. N.D. N.D. N.D. N.D.
Thioanisole N.D. N.D. N.D. N.D. N.D.
Dibenzyl disulfide N.D. 47.4 73.8 90 99.7
Isopropyl disulfide N.D. N.D. 23.7 71.8 95.6
Disulfides
Butyl disulfide N.D. N.D. 16.4 65 75.2
Diphenyl disulfide N.D. 23.8 27.4 45 54.6

N.D. = Not Detectable. Detection Limit for TCS was fixed at 2 mg/kg of elemental sulfur, thus

corresponding to 5 % under the experimental condition applied.

21
 At 100 °C, sulfides (PVS) and disulfides (DBDS and DPDS) begin to exhibit reactivity against

copper, even if mercaptans remain on average the most corrosive compounds. Only at

moderately high temperature (120 °C) all disulfides are corrosive, and their corrosiveness yield,

in particular in the case of an aromatic molecule like DBDS, becomes comparable to

mercaptans.

At 150 °C an inversion in the corrosion scale is observed since disulfides become more

reactive than mercaptans; the same behavior can be observed for sulfoxides that show, in

particular the aromatic ones, to be aggressive towards copper.

Finally, at 180 °C (a temperature that simulates overheating conditions, as thermal fault type

T1 as defined in IEC 60599) the corrosion effect further increases for all the families of

compounds, and the conversion of some disulfides (DBDS, isopropyl disulfide) and sulfoxides

(dibenzyl sulfoxide) to Cu 2 S becomes quantitative (corrosion yield 100%).

According to the results obtained, the following corrosiveness ranking can be proposed (figure

5), where capital letters were used to emphasize a more pronounced corrosiveness:

80 °C: MERCAPTANS >>> sulfides ≈ disulfides ≈ sulfoxides ≈ thiophenes;

100 °C MERCAPTANS >> DISULFIDES > SULFIDES > sulfoxides ≈ tiophenes;

120 °C MERCAPTANS ≈ DISULFIDES >> SULFIDES > sulfoxides ≈ thiophenes;

150 °C DISULFIDES > MERCAPTANS > SULFOXIDES > SULFIDES >> thiophenes;

180 °C SULFOXIDES ≈ DISULFIDES > MERCAPTANS > SULFIDES >> thiophenes.

22
Figure 5. Corrosiveness of family of compounds (averaged on all the investigated compounds)

at different temperatures.

The results indicate that up to 120 °C, the -SH group is responsible for the major corrosive

effect. Other species (i.e. disulfides or sulfoxides) begin to exhibit relevant corrosiveness only

from an aging temperature higher than 150°C. This evidence (deduced from the results obtained

under the applied experimental conditions) is in opposition to the classical scale of reactivity

based only on the thermodynamic affinity between sulfur compounds and copper.

To investigate the corrosiveness also in the presence of oxygen, some of the studied

compounds were selected on the basis of their behavior in the reactivity study and higher values

of corrosion provided by the TCS method. For this purpose, the original TCS test was repeated

using a septum drilled with a stainless steel needle (i.d. 0.8 mm), allowing air intake.

23
 It can be observed from these data (see Table 2) that corrosiveness increases proportionally

with temperature for all the investigated compounds. In agreement with what highlighted by the

reactivity study, the presence of oxygen hampers corrosion to a certain extent.

Table 2. Corrosiveness yield for selected sulfur compounds calculated according to TCS tests,

performed in presence of oxygen.

Class Compound Corrosion yield (%)

80°C 100°C 120°C 150°C
Dodecyl mercaptan N.D. 32.7 49.6 58.6
Mercaptans
Cyclohexanethiol 7.3 36.3 52.1 74.7
Benzyl mercaptan 24.9 52.2 67.2 61.9
Sulfoxides Dibenzyl sulfoxide N.D. N.D. N.D. 23.2
Sulfides Phenylvinylsulfide 7.3 27.6 22.9 43.9
Dibenzylsulfide N.D. N.D. 5.3 5.3
Disulfides Dibenzyldisulfide N.D. 48.7 68.3 75.9
Diphenyldisulfide N.D. 42.2 51.2 51.6
N.D. = Not Detectable.

As it can be perceived from data in Table 2, sulfur compounds like aliphatic mercaptans, that

showed to be the most easily oxidized (see figures 3a to 3e) even at lower temperatures, have a

lower corrosion yield in the presence of oxygen. On the other hand, compounds like DBDS, that

are less prone to oxidation, show similar corrosion yield regardless the presence of oxygen. This

result confirms the competitive effect of an oxidative atmosphere, and confirms other results

obtained by some researchers, that showed how corrosive oils have a peak of reactivity at

relatively low oxygen content.2

24
 Comparison with official measurements

Methods available to assess the degree of corrosiveness of an oil upon copper usually imply a

visual evaluation after the reaction between the oil and a copper surface. Generally, they can be

classified into two families: (i) methods using naked copper strips (e.g. method ASTM D1275),

and (ii) methods using paper-wrapped copper conductors (method IEC 6253526). Methods

involving paper are aimed at evaluating the capability of the oil to form copper sulfide on the

paper surface, whereas methods without paper specifically determine reactivity of corrosive

species towards copper surface. These tests are often considered to be complementary. In this

work, method ADSTM D1275 B was applied in order to correlate standard methods with the

reactivity study and corrosiveness ranking studies previously discussed. IEC 62535 method was

not used to compare or confirm the results of this study because it implies the use of cellulose

(Kraft paper), that can interfere in the corrosion reactions. ASTM D1275 B method was

performed at 80°C, 100°C, 120°C, 150°C and 180°C.

Pictures reported in Figures S1-S2 (Supporting Information) clearly demonstrate that benzyl

mercaptan is, among the studied compounds, the one more prone to form copper sulfide onto

bare copper surfaces, even at low temperatures.

This result is in agreement both with the reactivity study and corrosiveness ranking previously

discussed where, among mercaptans, benzyl mercaptan resulted to be the compound with the

highest corrosion yield.

Also DBDS brings about a significant copper sulfide deposition on conductor, but this effect is

evident only for temperatures higher than 150 °C, at least under the experimental conditions

evaluated. This result supports, once again, that BMt attacks copper more strongly, and that, as

mentioned above, corrosive effect of DBDS can be due to its conversion into BMt. For the other

25
mercaptans a clear corrosiveness can be established only at very high temperature (180 °C,

except EtDt from 150 °C). The other important outcome produced by this investigation is the

high corrosiveness showed by the three sulfoxides: beyond 150 °C, they cause an increasing

black deposit, thus confirming their corrosive role highlighted by the TCS method. For the other

investigated compounds, the assessment of a corrosiveness is tentative, because the observed

colors can be subjectively interpreted (e.g.: DPDS and PVS).

CONCLUSIONS

This study allowed us to better understand the reactivity of several sulfur compounds, both in

terms of stability under conditions of thermal and oxidative stress, and tendency to form copper

sulfide in the presence of bare copper surfaces. Sulfur compounds belonging to different families

(mercaptans, sulfides, disulfides, sulfoxides, thiophenes) showed different affinity against

copper, and their reactivity scale is temperature-dependant. In fact, at relatively low temperatures

(80 °C) mercaptans seem to be largely more corrosive then all other families of sulfur

compounds. But, when temperature increases (up to 180 °C) we have observed an apparent

inversion of reactivity scale, having disulfides and sulfoxides as the most corrosive families. This

behavior, that cannot be explained only by considering the affinity of each compound against

copper, or by the thermodynamic stability of the reaction by-products, may be caused by the

concomitant presence of at least two different reactions: oxidative degradation and corrosive

attack on copper. Both reactions involve all sulfur compounds, but the less thermally stable (i.e.

mercaptans, as depicted by the reactivity study in this work), can be degraded at a rate faster than

the corrosive reaction rate. This may explain the reason why method ASTM D1275 B is usually

more sensitive to disulfides than to mercaptans: the applied temperature (150 °C) is high enough

26
to rapidly induce the oxidation of mercaptans before they can react with copper to an appreciable

extent. On the other hand, disulfides exhibit a better resistance to oxidation at the test

temperature, and the reaction of copper sulfide formation is dominant compared to the oxidative

degradation. Other empirical tests (e.g. DIN 5135327) are known to be more sensitive to

mercaptans, and are also recommended by international standards12,28 to assess the corrosiveness

of mineral oils. Their better response to mercaptans may be explained not only by the use of

silver instead of copper as a reactive metal, but also by the lower applied temperature (100 °C),

at which mercaptans are relatively thermally stable and can exhibit their corrosive tendency in an

appreciable way, thus giving a visible color change in the test strip.

Our studies demonstrated that in case of thermally stable compounds, such as DBDS, the

formation of copper sulfide increases with temperature.

Our research underlines the need for monitoring the presence of sulfoxides in mineral oil,

especially in the case of free breathing units. As a matter of fact the oxidative deterioration of

the oil may convert sulfides into corrosive sulfoxides, thus inducing an increased corrosiveness

of the oil by effect of oil oxidation. It is worth mentioning the totally non-corrosive effect of

dibenzothiophene.

ACKNOWLEDGMENT

The Authors would like to thank Dr. Francesco Armigliato for his valuable help and Dr. Carlo

Maria Roggero for language revision. Financial support from MIUR (Ministero dell’Istruzione,

dell’Università e della Ricerca, Italy) is gratefully acknowledged. Financial support from

Compagnia di San Paolo and University of Torino (Progetto d’Ateneo 2012) is also

acknowledged.

27
 ASSOCIATED CONTENT

Supporting Information. Chemical structures of the sulfur compounds investigated. GC-AED

conditions optimized for the quantification of sulfur compounds. Photographs of copper

conductors after contact with each sulfur compound investigated according to the official method

ASTM D1275B tested at different temperatures. This material is available free of charge via the

Internet at http://pubs.acs.org.

28
 REFERENCES

(1) Maina R., Corrosion Phenomena. Proc. MyTransfo Meeting, Turin, Italy, 2008.

(2) Bengtsson, C.; Dahlund, M.; Hajek, J.; Pettersson, L.F.; Gustafsson, K.; Leandersson, R.;

Hjortsberg, A. Oil corrosion and conducting Cu 2 S deposition in power transformer windings.

41st CIGRE Session, Paris, France, 2006; Paper A2-111.

(3) Maina, R. New diagnosis on insulating liquids during the operational life of electrical

equipment. Procs. Of MyTransfo Brasil, Rio de Janeiro, 2007.

(4) Scatiggio, F.; Tumiatti, V.; Maina, R.; Tumiatti, M.; Pompili, M.; Bartnikas, R. Corrosive

Sulfur in Insulating Oils: its Detection and Correlated Power Apparatus Failures. IEEE T. Power

Deliver. 2008, 23, 508.

(5) Grant, D.H. Quantification of reactive sulfur in transformer oil. IEEE Electrical

Insulation Conference, Montreal, Canada, 2009.

(6) Kaplan, I.R.; Rasco, J.; Lu, S.T. Chemical Characterization of Transformer Mineral-

Insulating Oils. Environ. Forensics 2010, 11, 117.

(7) Tumiatti, V. Condition monitoring by oil chemical analysis”, CIGRE General Meeting,

Paris, 2000.

(8) Varga, Z.; Hancsok, J. Deep Hydrodesulfurization of Gas Oils, Petroleum and Coal

2003, 45, 147.

29
 (9) Duval, M.; Rouse, T.O. CHAPTER 4—Physical and Chemical Properties of Mineral

Insulating Oils—in Electrical Insulating Liquids, Engineering Dielectrics, Vol. III, Bartnikas R.

Ed. 1994, ASTM, USA, 310.

(10) Maina, R.; Scatiggio, F.; Kapila, S.; Tumiatti, V.; Tumiatti, M.; Pompili, M. Dibenzyl

disulfide (DBDS) as corrosive sulfur contaminant in used and unused mineral insulating oils.

CIGRE A2, Paris, France, 2006.

(11) Sulfur Speciation in Insulating Oil. CIGRE TF A2-32-03, Final Report, June 2010

(12) Fluids for electrotechnical applications – Unused mineral insulating oils for

transformers and switchgear. International Standard International Electrotechnical Commission

IEC60296 2003-11, Geneva, Switzerland.

(13) Lewand, R.L. Dibenzyl Disulfide in Transformer Oil. Doble Engineering Company-Neta

World, USA, Fall 2008.

(14) Li, J.; He, Z.; Bao, L., Yang L. Influences of Corrosive Sulfur on Copper Wires and Oil-

Paper Insulation in Transformers. Energies 2011, 4, 1563.

(15) Lukic, J.M.; Nikolic, D.; Mandic, V.; Glisic, S.B.; Antonovic, D.; Orlovic A.M. Removal

of Sulfur Compounds from Mineral Insulating Oils by Extractive Refining with N-Methyl-2-

pyrrolidone, Ind. Eng. Chem. Res. 2012, 51, 4472.

(16) Maina, R.; Tumiatti, V.; Pompili, M.; Bartnikas, R. Corrosive Sulfur Effects in

Transformer Oils and Remedial Procedures. IEEE T. Dielect. El. In. 2009, 16, 1655.

(17) Kamishima, S.; Ito, T.; Morishima, Y. Change in Corrosivity of Insulating Oil Caused by

Oxidative Deterioration of the Oil. IEEE T. Dielect. El. In. 2012, 19, 505.

30
 (18) Oweimreen, G.A.; Jaber, A.M.Y.; Abulkibash, A.M.; Mehanna, N.A. The Depletion of

Dibenzyl disulfide from a Mineral Transformer Insulating Oil. IEEE T. Dielect. El. In. 2012, 19,

1962.

(19) Kato, F.; Amimoto, T.; Nagao, E.; Hosokawa, N.; Toyama, S.; Tanimura, J. Effect of

DBDS Concentration and Heating Duration on Copper Sulfide Formation in Oil-immersed

Transformer Insulation. IEEE T. Dielect. El. In. 2011, 18, 1869.

(20) Ren, Shuangzan; Zhong, Lisheng; Yu, Qinxue; Cao, X.; Li, S. Influence of the

Atmosphere on the Reaction of Dibenzyl Disulfide with Copper in Mineral Insulation Oil. IEEE

T. Dielect. El. In. 2012, 19, 849.

(21) Toyama, S.; Tanimura, J.; Yamada, N.; Nagao, E.; Amimoto, T. Highly sensitive

detection method of dibenzyl disulfide and the elucidation of the mechanism of copper sulfide

generation in insulating oil. IEEE T. Dielect. El. In. 2009, 16, 509.

(22) Tumiatti, V.; Kapila, S.; Roggero, C.M.; Di Stefano, C.; Tumiatti, M.; Maina, R.;

Anderson, K.R. Sea Marconi-Total Corrosive Sulfur (SM-TCS). Patent WO/2010/070590 A1.

(23) Standard Test Method for Corrosive Sulfur in Electrical Insulating Oils. Designation: D

1275 – 06, 315/98, ASTM International, March 2006

(24) Mineral oil-impregnated electrical equipment in

service: Guide to the interpretation of dissolved and free gases

analysis. International Electrochemical Commission, IEC60599 (1999–2003), March 1999

(25) Copper Sulphide in Transformer Insulation. CIGRÉ SC A2, Brochure TB-378, 2009

31
 (26) Insulating liquids - Test method for detection of potentially corrosive sulfur in used and

unused insulating oil . IEC 62535 – International Standard, Edition 1.0, 2008-10

(27) Testing of insulating oils; detection of corrosive sulfur; silber strip test. Deutsches Institut

Fur Normung DIN 51353. 1985

(28) Mineral insulating oils in electrical equipment - Supervision and maintenance guidance.

International Electrochemical Commission, IEC60422 Second Edition 1999-03

32
Supporting Information for

Stability and reactivity of sulfur compounds against

copper in insulating mineral oil: definition of a

corrosiveness ranking

Maria Concetta Bruzzoniti*†*, Rosa Maria De Carlo†, Corrado Sarzanini†, Riccardo Maina‡, Vander

Tumiatti‡

†*Department of Chemistry, University of Torino, Via P. Giuria 5, 10125 Torino (Italy)

‡ Sea Marconi Technologies, Via Ungheria 20, 10090 Collegno (Torino)

AUTHOR INFORMATION

*Corresponding Author

Maria Concetta Bruzzoniti

Department of Chemistry, University of Torino, Via P. Giuria 5, 10125 Torino Italy

Ph: +39 011 6705277

Fax: +39 011 6705242

E-mail: [email protected]

S1
Table S1: Sulfur compounds investigated. In italic: compounds found in transformer oils. In bold: species considered for the reactivity

study as representative of their class.

Class Structure Name Abbreviation Class Structure Name Abbreviation

Dibenzothio- Dibenzyl
Thiophenes DBTf DBDS
phene S S Disulfide
S

O Dibenzyl Diphenyl
DBSOx DPDS
S Sulfoxide Disulfide
Disulfides S S
O Dibutyl Butyl
S
Sulfoxides DBuSOx BuDS
S Sulfoxide S Disulfide

Dodecyl methyl S Isopropyl

DcMeSOx IPDS
S sulfoxide S Disulfide
O

Dibenzyl Benzyl
DBS BMt
S Sulfide SH Mercaptan

Phenyl Vinyl Dodecyl

PVS SH DCMt
Sulfide Mercaptan
S

Pentamethylene Hexadecyl
Sulfides PMeS Mercaptans HDMt
Sulfide Mercaptan
SH
S

S
sec-Butyl Benzene-1,3-
sBS BDT
Sulfide dithiol
HS SH
S
Butyl
Thioanisole TAS SH BuMt
Mercaptan

S2
S
2-Methyl-1,3- Heptyl
MeDT SH HeMt
S Dithiane Mercaptan

1,2-
HS
EtDt
SH Ethanedithiol

Cyclohexanet
CET
SH hiol

S3
Table S2: GC-AED conditions optimized for the quantification of sulfur compounds.

Split type
Injector
300°C; 1 µL; splitting ratio 1:10

HP-1 Methyl Siloxane Capillary (100% PMSO)

Column
60m×250μm×0,25μm

Carrier gas He 6.0, 1 mL/min

0-2 min: 100°C

Program
temperature 3°C/min until 300°C

300°C for 12 min

Gas reagents: O 2 , H 2
Temperature of transfer line: 310 °C
Detector conditions
Cavity temperature: 310 °C
λ = 181 nm (carbon spectral line)

S4
Figures S1-S2. Photographs of copper conductors after contact with each sulphur compound
investigated according to the official method ASTM D1275B tested at different temperatures

S5
Figure S2.

S6