Degradation resistance of cordierite diesel particulate filters to diesel fuel ash deposits

Article abstract of DOI:10.1111/j.1551-2916.2011.04997.x

The thermochemical degradation resistance of a typical cordierite diesel particulate filter (DPF) material by synthetic ashes typical of those arising in practice has been investigated over the temperature range 950°C-1250°C. Differential thermal analyses and heat treatment of pressed pellets were used to characterize the melting/transformation behavior of ashes representative of typical diesel fuel (ash A) and those typical of when the catalysts ferrocene (ash FeA) and cerium carboxylate (ash CeA) were present in the fuel and to study the interaction chemistry between powdered cordierite and the ash compositions. Additional experiments involved the application of surface coverings of ash to DPF specimens. The results obtained showed that filter performance would not be compromised by ash liquefaction/sintering as long as temperatures did not exceed 900°C for ash A, 970°C for ash FeA and 1100°C for ash CeA. For long time periods, compared to the expected application durations, liquid phosphates dissolve cordierite leading to the formation of Zn and Fe aluminate spinels. Overall, the results clearly indicate that thermochemical degradation of cordierite by ashes under conditions representative of typical diesel engine systems is highly unlikely at temperatures of 1100°C and below.

Full text of DOI:10.1111/j.1551-2916.2011.04997.x

J. Am. Ceram. Soc., 95 [2] 746–753 (2012)
DOI: 10.1111/j.1551-2916.2011.04997.x
© 2011 The American Ceramic Society
ournal
J
Degradation Resistance of Cordierite Diesel Particulate
Filters to Diesel Fuel Ash Deposits
M. J. Pomeroy,‡,† D. O’Sullivan,§ S. Hampshire,‡,§ and M. J. Murtagh¶
‡Materials and Surface Science Institute, University of Limerick, Limerick, Ireland
§Materials Ireland Research Centre, University of Limerick, Limerick, Ireland
¶Corning Incorporated, Corning, New York 14831
The thermochemical degradation resistance of a typical cordie-
rite diesel particulate filter (DPF) material by synthetic ashes
typical of those arising in practice has been investigated over
the temperature range 950°C–1250°C. Differential thermal
analyses and heat treatment of pressed pellets were used to
characterize the melting/transformation behavior of ashes rep-
resentative of typical diesel fuel (ash A) and those typical of
when the catalysts ferrocene (ash FeA) and cerium carboxylate
(ash CeA) were present in the fuel and to study the interaction
chemistry between powdered cordierite and the ash composi-
tions. Additional experiments involved the application of sur-
face coverings of ash to DPF specimens. The results obtained
showed that filter performance would not be compromised by
ash liquefaction/sintering as long as temperatures did not
exceed 900°C for ash A, 970°C for ash FeA and 1100°C for
ash CeA. For long time periods, compared to the expected
application durations, liquid phosphates dissolve cordierite lead-
ing to the formation of Zn and Fe aluminate spinels. Overall,
the results clearly indicate that thermochemical degradation of
cordierite by ashes under conditions representative of typical
diesel engine systems is highly unlikely at temperatures of
1100°C and below.
Cordierite (Mg2Al4Si5O18) ceramic bodies exhibit excellent
thermal shock resistance, making cordierite a good material
choice for the diesel particulate filter application. The ortho-
rhombic cordierite crystal structure contains six-member
coplanar rings (four SiO4 groups and two AlO4 groups) with
the corner-sharing tetrahedra allowing for an open channel
parallel to {00l}.3 These rings are cross-linked by (Si,Al)O4
tetrahedra and MgO6 octahedra. The large expansion of the
Mg–O bonds and lower expansion of the Si–O and Al–O
bonds result in a crystal having a low, anisotropic thermal
expansion.4 Cordierite DPF systems have been commercially
available for nearly 2 decades.5 The principal function of the
filter is to capture solid carbon particles with high efficiency
(up to 99%).
The carbon particulates captured by wall-flow filters are
intimately mixed with unburnt hydrocarbons as well as con-
taminants, such as calcium, zinc, phosphorous, and sulfur
compounds which arise from engine lubrication. Iron-con-
taining particles arising from engine abrasion can also be
captured. This composite solid fluid admixture comprises the
particulate matter requiring filtration. Clearly as time pro-
gresses, then carbon builds up on filter walls and if not
removed contributes to significant increases in backpressure.
Regeneration of the filter occurs by carbon particle combus-
tion at temperatures of 600°C and above,2 with typical tem-
peratures up to 890°C.6 Ferrocene ((C5H5)2Fe) and cerium
carboxylate (Ce(COOH)4) are both useful fuel additives for
diesel because they reduce carbon buildup in DPFs by lower-
ing the light-off temperature of the carbon particulates. Par-
ticulate matter burn-off temperatures can be reduced by as
much as 250°C by the use of such catalytic fuel additives.2,7
The use of these additives results in the deposition of Fe2O3
or CeO2 in the diesel particulate filter.
Particulate matter combustion is, of course, a highly exo-
thermic reaction, and if uncontrolled, can cause local temper-
ature increases well above 1000°C.8 As a result of multiple
regenerations, a residual ash comprising the contaminant oxi-
des from the lubricating oil, iron from engine abrasion, and
oxides from catalytic fuel additives, will remain on and in
the wall of the filter. At elevated temperatures (>1100°C),
which may be generated during in situ incineration, this
residual ash may react with the ceramic structure. Such reac-
tions may cause premature mechanical or chemomechanical
failures through formation of low eutectic surface liquids as
well as crystalline and vitreous phases of incompatible ther-
mal expansion with the diesel particulate filter ceramic
matrix.
This thermochemical durability of cordierite and other
DPF materials has been studied to a degree by reacting oxi-
des with them. Thus, cordierite has been reacted with
sodium, lead, iron, calcium, zinc, and vanadium oxides9–11
and been shown to be degraded to slight extents by each of
the oxides. Sodium and lead oxides were observed to cause
I. Introduction
NVIRONMENTAL friendly light (car) and heavy (truck) duty
diesel engines are required to have diesel particulate fil-
E
ters (DPFs) fitted to the exhaust train to remove unburnt
carbon particulates from exhaust emissions. Typically, such
particulate filters are of the honeycomb wall-flow type com-
prising ceramic honeycomb structures, which have a checker-
board pattern of open and closed cells at the entry face and
at the exit face. Open cells at the entry face are blocked at
the exit face and the blocking of alternate cells forces the
exhaust gas to flow through the porous walls trapping the
particulate matter in the pores and on the surfaces of the
open channels.1 Foamed ceramics and fiber mats as well as
sintered ceramics are used as DPFs. Materials’ requirements
for honeycomb wall-flow filters are typically: low coefficient
of thermal expansion, high thermal shock resistance, and
thermal stability.2 Adler2 has also listed suitable materials
which possess these properties as cordierite, aluminum tita-
nate, silicon carbide, silicon nitride, and mullite.
C. Jantzen—contributing editor
Manuscript No. 30293. Received September 08, 2011; approved November 11,
2011.
†Author to whom correspondence should be addressed. e-mail: michael.pomeroy@
ul.ie
746

February 2012
Degradation Resistance of Cordierite Diesel Particulate Filters
747
greatest damage to cordierite.9 Del Pin et al.12 examined the
degradation of DPF candidate material zircon with respect
to sodium, calcium, lead, zinc, iron (III), vanadium, and cer-
ium oxides over the temperature range 700°C–1000°C.
Results showed again that only sodium and lead oxides
induced significant chemical interaction by the formation of
sodium and lead zirconates. Maeir et al.13 have investigated
the chemical interaction between single crystal cordierite and
a soda–silica mixture of composition Na4Si3O8 with a lowest
liquidus of 830°C. As might be expected, the cordierite
reacted rapidly with the liquid alkali–silicate at temperatures
of 900°C and 1000°C allowing cordierite dissolution. After a
certain period of time, however, a protective nepheline cor-
rosion deposit formed on the cordierite, slowing down the
degradation process.
While the above chemical durability experiments were of
benefit in assessing the durability of cordierite to potential
diesel ash impurities, lead and sodium impurity levels in die-
sel fuels are miniscule (typically 2 ppm for sodium,14
although sodium levels may increase as Biodiesel becomes
more commonly used).13 O’Sullivan et al.15 studied the inter-
action of silicon carbide with a calcium–zinc phosphate–sul-
fate mixture based on analyses obtained for ash removed
from a DPF which had been operating for the equivalent of
160 000 km. While this ash composition alone may sinter to
damage the filtration efficiency of the SiC filter, no such
damage or chemical degradation of the filter material was
observed when the ash chemistry was adjusted to take into
account iron or cerium additions arising from the catalytic
additives referred to above.
Table I details the ash compositions studied and in addi-
tion, relates compositions reflecting an ash loading on a 152-
mm long, 144-mm diameter DPF representative of
160 000 km (100 000 miles) operation. Ash compositions
were prepared by dry ball milling relevant weighed powder
mixtures of CaSO4·2H2O (Alfa/98% min), CaO (BDH/99%
min), ZnSO4·7H2O (Alfa/98% min), ZnO (BDH/99% min),
P2O5 (BDH/97% min), Fe2O3 (Alfa/99.99%), CeO2 (Alfa/
99.9%), and graphite (BDH/99% min) for 1 h using sialon
milling media. The 50 mg portions of uniaxially pressed com-
pacts of the three ash compositions were subjected to simul-
taneous Differential Thermal Analysis/Thermogravimetric
Analysis (DTA/TGA) (Stanton Redcroft STA1640; Polymer
Labs Rheometric Scientific, Epsom, U.K.). These analyses
were carried out using a heating rate of 10°C/min to 1400°C
in air + 200 ppm SO2 flowing through the system at 70 cm3/
min. This quantity of sulfur dioxide was added to simulate
the presence of this gas in the exhaust train of a diesel
engine. Pressed pellets, typically 8-mm diameter and 5-mm
high were subjected to heat treatment in the same gas envi-
ronment for 24 h at temperatures selected on the basis of the
thermal analysis studies (900°C, 970°C, 1100°C, and 1250°C).
A heating rate of 50°C/min and furnace cooling were used
for all heat treatments. The dimensions of the pressed and
heat-treated pellets were measured using Vernier callipers
and radial shrinkages calculated. X-ray diffraction (XRD) of
heat-treated pellets ground to powder was carried out with
CuKa radiation in a Philips X’Pert system.
Interaction chemistry between cordierite and ash mixtures
was evaluated by DTA/TGA and heat-treatment analysis of
intimately mixed and uniaxially pressed cordierite–ash cou-
ples. Accordingly, cordierite filters supplied by Corning Inc.
were crushed and ground to <90 lm powder and dry ball
mill mixed for 1 h with each of the three ashes. Cordierite–
ash compositions corresponded to the loadings given in
Table I. The 50 mg portions of these pellets were subjected
to DTA/TGA in the same way as the ash samples and
8 mm 9 5 mm pellets were heat-treated in the same manner
as the ash pellets.
To evaluate the surface interaction of cordierite with the
ashes and thus simulate more likely thermochemical condi-
tions in an operating filter, samples of cordierite filter were
coated with each of the ash mixtures. Thus, surface coatings
of each ash (A, FeA and CeA) were deposited on flat cou-
pons (6.25 cm2 area and two cells thick) sectioned from a
cordierite filter. The coatings were applied by spraying the
coupon surfaces, preheated to ~200°C, with a solution of
nitrates of the metallic (Ca, Zn, Fe, Ce) constituents of the
ash, as well as P2O5 and H2SO4. The target ash loading after
spraying and drying corresponded to an in-service lifetime of
80 000 km. The coated specimens were heat-treated at the
same temperatures used for the powder compacts (900°C,
1000°C, 1100°C, and 1250°C). The heat-treatment times were
chosen to reflect the total estimated time that a diesel partic-
ulate filter would experience at the particular temperature as
a result of repeated regenerations during its in-service life-
time. Thus, heat treatment at 900°C was carried out for 24 h
and at the other temperatures for 30 min. The heat-treated
sprayed surfaces were analyzed by XRD and scanning elec-
tron microscopy (SEM) examination using a JEOL JSM840
(Tokyo, Japan). Compositional variations were analyzed
using energy dispersive X-ray analysis (EDS). Selected speci-
With respect to cordierite, Maier et al.13 have investigated
degradation of single crystal material by an equimolar mix-
ture of Ca3(PO4)2, Mg3(PO4)2, and Zn3(PO4)2. They observed
that single crystal cordierite was severely corroded by the
equimolar phosphate at 1050°C. Earlier preliminary work15,16
(J. Kracklauer, Econalytic Systems Inc., private communica-
tion (1998)) has indicated that cordierite filter elements have
withstood corrosion in this temperature range by a calcium–
zinc phosphate–sulfate ash composition based on ash
recovered from a DPF used in engine tests for durations
equivalent to 160 000 km service. This article reports the
results of the effect of particulate matter ash constituents
on the chemical stability of cordierite filter materials as a
function of the DPF temperature environment.
II. Experimental Procedure
As indicated above and determined from data obtained
from engine tests for the authors16 (J. Kracklauer, Econalyt-
ic Systems Inc., private communication (1998)), lubricant
and fuel impurities being deposited on diesel particulate fil-
ters are typically ZnO, CaO, P2O5, and SO3 with relative
mass ratios of 1:1.75:2:2.25, respectively. In addition, it was
decided to recreate the conditions when the one filter regen-
eration cycle was missed, i.e. there was double the maxi-
mum expected carbon load on the filter prior to burn off.
Based on this rationale, various ash compositions were
derived for the baseline ash (A), and two other ashes
derived from the combustion of diesel containing either fer-
rocene (FeA) or cerium carboxylate (CeA) as combustion
catalysts. For these ashes, the catalysts decompose to their
base oxides, Fe2O3 or CeO2.
Table I. Ash Compositions and Loadings (% by Weight)
Ash Code
CaSO4·2H2O
CaO
ZnSO4
ZnO
P2O5
Fe2O3
CeO2
C
Ash loading (%)†
A
FeA
CeA
35.424
25.705
20.083
8.553
4.306
1.889
19.064
8.716
2.998
1.869
2.891
3.279
23.054
14.477
9.676
0.000
36.240
0.000
0.000
0.000
56.999
12.036
7.664
5.077
24.0
33.8
43.8
†As a percentage of filter + ash weights

748
Journal of the American Ceramic Society—Pomeroy et al.
Vol. 95, No. 2
mens were fractured and the resulting cross-sections were
also examined using SEM and EDS.
fills a dilution effect, thus reducing the amount of liquid
formed compared to the baseline ash. Such dilution effects
were even more prominent for the CeA ash where the pres-
ence of nearly 63 wt% CeO2 suppressed endothermic events
at temperatures less than 1000°C. Radial shrinkages of heat-
treated pellets were negligible until a temperature of 1250°C
was reached when 13% was observed. The phase assemblages
of the heat-treated ceria containing ashes, shown in Table II,
comprised those of the baseline ash (i.e. Ca and Zn
phosphates and sulfates) plus ceria.
III. Results
(1) Ash Characterization
For each of the ashes, large exotherms at ~800°C, associated
with carbon burnout were observed. Associated weight losses
corresponded well with those expected from complete com-
bustion of the carbon. For the baseline ash, endotherms
occurred at 954°C and 996°C, which were associated with
weight losses. These weight losses totaled some 7–8 wt%.
Radial shrinkage and XRD analyses of sintered pellets
showed that while little change in ash character occurred
after heat treatment at 900°C, significant shrinkage (16%)
occurred after heat treatment at 970°C and this was associ-
ated with the disappearance of CaSO4 from the phase assem-
blage. These results indicate two factors: (i) that CaSO4
decomposition gives rise to the weight losses observed during
DTA and (ii) as the measured weight losses only account for
about 25% of the total SO3 content, that CaSO4 is incorpo-
rated into a liquid phase which facilitates shrinkage through
liquid phase sintering. After heat treatment at 1100°C, com-
plete melting of pellets was observed as was the case at 1250°C.
The XRD analyses (Table II) showed that calcium phosphate
(Ca3(PO4)2) and calcium zinc phosphate (CaZn2(PO4)2)
occurred after heat treatment at 970°C and 1100°C. After
heat treatment at 1250°C, (Ca3(PO4)2) was the only crystal-
line phase present. In addition to the crystalline phases pres-
(2) Cordierite–Ash Interactions
Figure 1 shows DTA/TGA data for 50 mg samples of
pressed cordierite and cordierite–ash mixtures. It is seen that
no thermal events arise for the cordierite and no weight
changes occur [Fig. 1(a)]. For each of the cordierite–ash mix-
tures, carbon burnout exotherms can be observed at temper-
atures of the order of 800°C [Figs. 1(b)–(d)]. For the
cordierite–baseline ash system, weak endotherms arise at
979°C, 998°C, and 1310°C, whereas a stronger endotherm is
observed at 1186°C [Fig. 1(b)]. From Fig. 1(c), it can be seen
that the incorporation of Fe2O3 into the system induces
lower temperature endothermic events at 927°C and 968°C,
although these are again weak in character. Two stronger en-
dotherms arise at 1168°C and 1305°C, which appear to cor-
respond reasonably well with those observed for the
cordierite–baseline ash system. Figure 1(d) shows the DTA/
TGA data for the cordierite–CeA system. It is seen that the
only significant endothermic event due to melting occurs at
1291°C. At lower temperatures, low intensity endotherms
arise at 995°C and 1209°C, with the latter being more pro-
nounced. At temperatures in excess of 1300°C, a series of
exothermic peaks arise which are thought to be associated
with phase transitions.
ent in samples heat-treated at 1100°C and 1250°C,
a
significant amorphous halo, consistent with the presence of
much glass, was observed. As CaSO4 and the mixed Ca, Zn
phosphate disappeared from the phase assemblage as temper-
atures increased, it appears that the liquid phase comprises
Ca, Zn, P, S, and O.
The ash mixture containing Fe2O3 showed endothermic
activity at temperatures of 935°C and 996°C, suggesting that
iron oxide lowered the temperature of the lower endotherm.
However, negligible radial shrinkage occurred after heat
treatment of pellets at 970°C suggesting that the appreciable
liquid phase sintering of the baseline ash was suppressed by
the presence of the 42 wt% Fe2O3. After heat treatment at
1100°C, 9% radial shrinkage occurred and at 1250°C, com-
plete melting of the FeA ash was observed. The phase assem-
blages (see Table II) showed that the same phases were
present as for the baseline ash. However, ZnFe2O4 was pres-
ent at all temperatures as well as Fe2O3. The formation of
ZnFe2O4 at all temperatures would remove ZnO from the
system and suppress liquid formation. The presence of resid-
ual Fe2O3 in the phase assemblages also indicates that it ful-
Heat treatment of the pressed pellets of cordierite and cor-
dierite–ash mixtures at 900°C, 1000°C, and 1100°C for 24 h
in flowing air—200 ppm SO2, caused little change in the
physical character of the pellets. However, at 1250°C, the
cordierite–baseline ash and cordierite–FeA ash pellets showed
complete slumping, while the cordierite–CeA ash pellets
showed partial slumping. The XRD data for the heat-treated
pellets are shown in Table III. It is seen that no change in
the crystallographic character of the cordierite occurs. For
the cordierite–baseline ash system, phases arising from the
ash are CaSO4 and Ca3(PO4)2 at 900°C and 1000°C, with the
relative intensity of reflections becoming more intense with
increasing temperature. At 1100°C, the CaSO4 reflections dis-
appear and at 1250°C, the Ca3(PO4)2 reflections disappear
and appear to be replaced by an amorphous halo associated
Table II. Phase Assemblages for Ash Pellets Heat-Treated in Air—200 ppm SO2 Gas Mixture for 24 h at Different
Temperatures (Phases Listed in Order Of Decreasing X-Ray Intensities)
Heat-treatment temperature
900°C
970°C
1100°C
1250°C
Ash A
Ca3(PO4)2
CaZn2(PO4)2
CaSO4
Ca3(PO4)2
CaZn2(PO4)2
Ca3(PO4)2
CaZn2(PO4)2
Ca3(PO4)2
Glassy phase
Ash FeA
Fe2O3
Fe2O3
ZnFe2O4
Ca3(PO4)2
Fe2O3
ZnFe2O4
Ca3(PO4)2
Fe2O3
Ca3(PO4)2
ZnFe2O4
CaZn2(PO4)2
CaSO4
ZnFe2O4
Ca3(PO4)2
CaZn2(PO4)2
Ash CeA
CeO2
CeO2
Ca3(PO4)2
CaZn2(PO4)2
CeO2
Ca3(PO4)2
CeO2
Ca3(PO4)2
Ca3(PO4)2
CaZn2(PO4)2
CaSO4

February 2012
Degradation Resistance of Cordierite Diesel Particulate Filters
749
(a)
(b)
(c)
(d)
Fig. 1. Thermal analysis data for (a) cordierite, (b) cordierite + ash A, (c) cordierite + ash FeA and (d) cordierite + ash CeA.
Table III. Phase Assemblages Following Heat Treatment of Cordierite and Cordierite–Ash Pellets in Air—200 ppm SO2 Gas
Mixture for 24 h at Different Temperatures (Phases Listed in Order of Decreasing X-Ray Intensities)
Heat-treatment temperature
900°C
1000°C
1100°C
1250°C
Cordierite
Cordierite + ash A
Mg2Al4Si5O18
Mg2Al4Si5O18
CaSO4
Ca3(PO4)2
ZnAl2O4
Mg2Al4Si5O18
Fe2O3
Ca3(PO4)2
ZnAl2O4
CaSO4
Mg2Al4Si5O18
CeO2
Mg2Al4Si5O18
Mg2Al4Si5O18
Ca3(PO4)2
CaSO4
ZnAl2O4
Mg2Al4Si5O18
Fe2O3
Mg2Al4Si5O18
Mg2Al4Si5O18
Ca3(PO4)2
Mg2Al4Si5O18
Mg2Al4Si5O18
ZnAl2O4
ZnAl2O4
Glassy phase
Cordierite + ash FeA
Cordierite + ash CeA
Mg2Al4Si5O18
Fe2O3
Mg2Al4Si5O18
Fe2O3
FeAl2O4
Ca3(PO4)2
ZnAl2O4
Ca3(PO4)2
ZnAl2O4
FeAl2O4
Mg2Al4Si5O18
CeO2
Glassy phase
Mg2Al4Si5O18
CeO2
Mg2Al4Si5O18
CeO2
Ca3(PO4)2
ZnAl2O4
CaSO4
Ca3(PO4)2
ZnAl2O4
CaSO4
Ca3(PO4)2
ZnAl2O4
CaSO4
Ca3(PO4)2
ZnAl2O4
CePO4
CePO4
Glassy phase
with a glassy phase. For all temperatures, zinc in the ash
appears to remove aluminum from the cordierite to form the
ZnAl2O4 spinel and it appears to be a relatively abundant
phase at all temperatures examined. The consumption of zinc
in the formation of this spinel presumably accounts for the
absence of the mixed Zn, Ca phosphate present for the heat-
treated ash (see Table II).
cordierite–baseline ash system, except, of course, the pres-
ence of Fe2O3 in the phase assemblage after heat treatment
at 900°C. It should also be noted that the relative intensi-
ties of the CaSO4 and Ca3(PO4)2 phases are reversed. After
heat treatment at 1000°C, CaSO4 reflections disappeared,
while after heat treatment at 1100°C, the iron analogue of
the ZnAl2O4 spinel, FeAl2O4, had formed. After heat treat-
ment at 1250°C, the resulting phase assemblage was cordie-
rite plus Fe2O3 plus FeAl2O4 plus glass. When ceria is
The incorporation of Fe2O3 into the system (cordierite–
FeA) gives rise to a similar phase assemblage as for the

750
Journal of the American Ceramic Society—Pomeroy et al.
Vol. 95, No. 2
Table IV. XRD Phase Assemblages for Cordierite Test Pieces, Sprayed with Ash Constituents and Heat-Treated
in Air—200 ppm SO2 (Phases Listed in Order of Decreasing X-Ray Intensities)
Heat-treatment temperatures and times
900°C 24 h
1000°C 30 min
1100°C 30 min
1250°C 30 min
Cordierite
Cordierite coated with A
Mg2Al4Si5O18
Mg2Al4Si5O18
Ca3(PO4)2
ZnAl2O4
Mg2Al4Si5O18
Mg2Al4Si5O18
Ca3(PO4)2
ZnAl2O4
Mg2Al4Si5O18
Mg2Al4Si5O18
Ca3(PO4)2
Mg2Al4Si5O18
Mg2Al4Si5O18
ZnAl2O4
ZnAl2O4
CaSO4
CaSO4
Mg2Al4Si5O18
Fe2O3
Cordierite coated with FeA
Cordierite coated with CeA
Mg2Al4Si5O18
Fe2O3
CaSO4
Mg2Al4Si5O18
Fe2O3
Mg2Al4Si5O18
Mg2Al4Si5O18
CeO2
Mg2Al4Si5O18
CeO2
Mg2Al4Si5O18
CeO2
Mg2Al4Si5O18
CeO2
added to the cordierite–ash system (cordierite–CeA), the
phase assemblages after heat treatment at 900°C and 1000°
C are the same as for the cordierite–baseline ash system
excepting, of course, the presence of CeO2. After heat
treatment at 1100°C, CePO4 arises in the phase assem-
blage, and in contrast to the cordierite–baseline ash system,
CaSO4 remains present. After heat treatment at 1250°C,
the phases present include both Ca3(PO4)2 and CePO4, in
contrast to the other two systems where no phosphates are
present. In contrast to the two other ash types, the relative
intensity of the ZnAl2O4 spinel in the reaction products
is less.
reflections disappear. There remains a phase assemblage of
cordierite plus ZnAl2O4 spinel after heat treatment at the
highest temperature. However, in contrast to the data
given in Table III, Table IV does not indicate the presence
of a glass phase. In the other two cordierite–ash systems,
the significant dilution effects of Fe2O3 and CeO2 reduce
other phases to levels where they are undetectable, except
for the cordierite–FeA system at 900°C where CaSO4 is
observed.
The SEM observations of the surface of cordierite cou-
pons after heat treatment showed that the temperature of
heat treatment had little effect on surface morphology.
Figure 2(a) details the surface morphology of a coupon
heat-treated at 900°C for 24 h and shows the characteristic
prismatic cordierite grain structure surrounding deliberately
induced porosity. Figure 2(b) shows the development of an
apparently friable particulate layer developed by the repre-
sentative sprayed coating of baseline ash. The surface
applied FeA ash appears to have interacted more with the
cordierite surface giving rise to what appears to be a matrix
glaze in which prismatic cordierite grains are visible
[Fig. 2(c)]. For the CeA ash coating [Fig. 2(d)], the cordie-
(3) Surface Interactions of Cordierite with Ashes
Table IV shows XRD results for the phases present on the
surfaces of coupons of cordierite spray coated with each of
the ash compositions and heat-treated in air—200 ppm
SO2. It is seen that the phase assemblages for cordierite
coupons coated with baseline ash are similar to those
observed for heat-treated pellets (see Table III). Thus, as
heat-treatment temperatures increase, CaSO4 and Ca3(PO4)2
(b)
(a)
(d)
(c)
Fig. 2. Surface topography of cordierite samples after heat treatment in air—200 ppm SO2 for 24 h at 900°C (a) cordierite only, (b) cordierite–
baseline ash deposit, (c) cordierite–Fe2O3 containing ash deposit, (d) cordierite–ceria containing ash deposit (width of scale bar is 10 lm).

February 2012
Degradation Resistance of Cordierite Diesel Particulate Filters
751
(a)
(b)
(c)
(d)
Fig. 3. Surface topography of cordierite samples after heat treatment in air—200 ppm SO2 for 30 min at 1250°C (a) cordierite only, (b)
cordierite–baseline ash deposit, (c) cordierite–Fe2O3 containing ash deposit, (d) cordierite–ceria containing ash deposit (width of scale bar is
10 lm)
rite coupon surface appears little affected. Indeed, the pris-
matic cordierite structure of the uncoated cordierite is
clearly apparent indicating a lack of interaction with the ash
and thus poor adhesion. Figure 3 shows the appearance of
the corresponding surfaces after heat treatment at 1250°C
for 30 min. The cordierite coupon surface [Fig. 3(a)] was
very similar in appearance to the one heat-treated at 900°C
for 24 h. The surface topographies of coupons coated with
ashes A and FeA after heat treatment at 1250°C for 30 min
[Figs. 3(b) and (c)] took on a glazed appearance with the
surface glaze appearing to contain particulates which would
be expected on the basis of XRD to be zinc aluminate and
Fe2O3, respectively. While the surface morphology after heat
treatment at 1250°C did have certain features in common
with those for the other ashes, the extent of glazing
appeared less as indicated in Fig. 3(d). For the coupons
heat-treated at 1000°C for 30 min, the coupons coated with
A, FeA, and CeA ash all had similar appearances to their
900°C counterparts. For those coupons heat-treated at
1100°C, the coupon coated with baseline ash A, was similar
in appearance to the coupon heat-treated at 1250°C, while
the cordierite surfaces coated with the other two ashes
looked very similar to the images collected for the 900°C
equivalents.
the disappearance of CaSO4 reflections from the phase
assemblage of heat-treated pellets at temperatures of 1100°C
and 1250°C. The resulting CaO, formed from the decomposi-
tion of CaSO4, would then be capable of reacting with P2O5
according to the reaction scheme:
3CaSO4 ! 3SO3 þ 3CaO
3CaO þ P2O5 ! Ca3ðPO4Þ3
(1)
(2)
A similar reaction scheme could be envisaged for zinc sul-
fate decomposition and reaction:
3ZnSO4 ! 3SO3 þ 3ZnO
3ZnO þ P2O5 ! Zn3ðPO4Þ2
(3)
(4)
Such reactions would provide Ca3(PO4)3 and Zn3(PO4)3
which can react to form a mixed phosphate, which has a
lowest liquidus at zinc rich compositions of 1048°C.18 This
chemistry does not, however, explain the significant sintering
(16% radial shrinkage) which arises when pressed pellets of
the baseline ash are heat-treated at 1000°C. It is important
to note that TGA weight losses of only about 25% SO3
arise, which would tend to suggest that the residual SO3
could suppress the lowest liquidus temperature in the
Ca3(PO4)3 and Zn3(PO4)3 system and thus account for the
significant shrinkage noted at 1000°C.
The addition of iron (III) oxide to the ash system might
be expected to cause two effects. On one hand, it might be
expected to induce liquid formation through reaction with
P2O5 as a lowest liquidus temperature of 960°C arises in the
Fe–P–O system for compositions where P2O5 is not stable.19
On the other hand, if reasonably inert, the replacement of
nearly 40% by weight of the baseline ash composition by
Fe2O3 would reduce liquid volumes significantly as well as
the observed formation of ZnFe2O4 which would reduce the
amount of zinc available for liquid formation. Given the
IV. Discussion
(1) Ash Properties
From the DTA/TGA studies, it is evident that in addition to
weight losses occurring at temperatures of the order of 800°C
due to carbon burnout, further weight losses occur at tem-
peratures in excess of 950°C for the baseline ash and these
are associated with endothermic processes. The SO3 loss
from calcium sulfate is one process which would account for
such thermal behavior and investigation of the necessary
thermodynamic data17 shows that in an atmosphere of air
containing 200 ppm SO2, calcium sulfate should begin to
decompose at temperatures of the order of 920°C.
Accordingly, these decomposition effects should arise and
would explain the weight losses, endothermic behavior, and

752
Journal of the American Ceramic Society—Pomeroy et al.
Vol. 95, No. 2
retardation of sintering by the addition of Fe2O3, it appears
that a reduction in the amount of liquid phase is more likely.
In the case of ceria replacement of 60% by weight of the
baseline ash, an even greater reduction in liquid volume
would be expected and this would be in agreement with the
observed results. However, it appears that the action of ceria
as a diluent is not its only function as CePO4 is formed when
ceria is present in the ash. The gettering of phosphorus from
the ash system in such a way by cerium would also decrease
liquid volume and accordingly reduce the sinterability of the
CeA ash system compared to both A and FeA ash systems.
The sintering and liquefaction characteristics of the ash
are important with respect to the degradation of cordierite
filter elements from two standpoints. One effect that signifi-
cant sintering of an ash covering on the filter surface may
have is to increase the pressure drop across the filter walls,
with negative consequences for the efficient operation of the
engine to which the filter is attached. Based upon the infor-
mation collected in this work, it is quite clear that such
effects are unlikely unless temperatures rise above 900°C for
the baseline ash, 970°C for ash derived from diesel contain-
ing the ferrocene additive, and 1100°C for ash derived from
diesel containing cerium carboxylate. These temperatures can
thus be used as guideline temperatures in assessing undue
increases in pressure drop which are not due to the mecha-
nisms indicated by Adler.2 It can thus be concluded that the
representative ash compositions studied herein could induce
poorer filter performance due to liquefaction/sintering effects
if temperatures exceed 900°C for the baseline ash, 970°C for
the Fe2O3 containing ash, and 1100°C for the ceria contain-
ing ash. The other significant effect of ash sintering and liq-
uefaction is catastrophic thermochemical degradation of the
cordierite filter material which was not observed at any tem-
perature or with any of the ashes. Interactions between the
cordierite and the various ashes were, however, observed and
these are now discussed.
tures which are extreme for a DPF system. A similar reaction
mechanism to the formation of the zinc aluminate would be
expected and accordingly, the activity of iron in the liquid
would be important. Given that the studies of the ash alone
indicates the diluent effect Fe2O3 has on the amount of liquid
phase formed, then it is expected that while FeAl2O4 forms
at 1100°C and 1250°C, its formation does not really affect
the amount of cordierite degradation as much as that of zinc
aluminate.
(3) Surface Interactions of Cordierite with Ashes
The XRD phase assemblage and the surface topographies
reported above are consistent with the analyses of the results
discussed above in subsections A and B. For the baseline ash,
the same phase assemblages arise for both the reacted pellets
and the ash deposit–filter interaction. For the iron and cerium
containing ashes, the phase assemblages effectively comprise
cordierite and Fe2O3 or CeO2. This information together with
the SEM image information endorses the argument that the
cordierite DPF material does not significantly degrade in the
presence of ash containing the combustion catalysts. While
the same cannot be said for the situation when the catalysts
are absent (baseline ash), significant surface glazing/liquefac-
tion would only appear to arise at temperatures of 1250°C.
Based on the observations made for the surface degradation
tests, it seems that cordierite DPFs could be used without fear
of thermochemical degradation at temperatures of 1100°C or
less. Of course, thermochemical degradation is a kinetic pro-
cess and will depend on times at temperature. Data obtained
by Corning20 indicate that the total time a DPF spends at
temperatures in excess of 900°C is about 30 min over the war-
ranted DPF life, 48 times less time than that investigated
herein. Ten minutes would be the apparent maximum dura-
tion at 1000°C (three times the time investigated herein) and
excursions to higher temperatures have not been recorded in
the working systems examined. Based on these observations,
then it could be fairly safely concluded that interactions
between cordierite and the three representative ashes exam-
ined herein would not be expected to cause thermochemically
induced failure of cordierite DPFs.
(2) Cordierite–Ash Interactions
Examination of the X-ray phase assemblage data for heat-
treated cordierite–ash mixture pellets clearly shows that the
only reaction products arising that contain elements from the
cordierite are the zinc or iron aluminate spinels. For these
species to form, aluminum must be extracted from the cor-
dierite with the result that the cordierite–ash interface should
become richer in Mg and Si. Maier et al.13 have observed a
similar effect for the interaction of single crystal cordierite
with an equimolar mixture of Mg, Ca, and Zn phosphates.
They envisage the degradation process occurring via the solu-
tion of cordierite by a liquid phosphate with subsequent for-
mation of a zinc aluminate spinel and silica. This reaction
type could certainly be the cause of ZnAl2O4 formation at
the temperatures examined herein for the baseline ash and
presumably the iron oxide-containing ash, although there is
little evidence of major liquefaction from either shrinkage or
DTA analyses at 900°C to support the solution reaction
mechanism. However, the enrichment of Mg and Si in the
corrosion product may reduce liquidus temperatures to facili-
tate this as the process goes on. Given the findings reported
herein and those reported by Maier et al.,13 it is clear that
cordierite degradation to form a ZnAl2O4 spinel is due to
solution by phosphate-based liquids. The gettering of phos-
phorus from the ash system by ceria, as discussed above,
reduces the probability of liquid phase formation and this is
why the zinc aluminate spinel occurs in less significant
amounts in the X-ray phase assemblage.
V. Conclusions
1. The representative ash compositions studied herein do
not lead to liquefaction/sintering effects unless the tem-
peratures exceed 900°C for the Ca, Zn, S, P containing
baseline ash, 970°C for the Fe2O3 containing ash, and
1100°C for the ceria containing ash. Accordingly, ash
agglomeration effects would not be expected during
normal cordierite filter lifetimes.
2. The baseline ash and iron oxide containing ash cause
cordierite to degrade by extracting aluminum from it
via a solution mechanism involving molten phosphate-
based liquids. The presence of ceria in the ash results
in phosphorus being removed from the ash, causing
less liquid formation and accordingly, with less
ZnAl2O4 formation.
3. At temperatures of 1100°C and 1250°C, FeAl2O4 for-
mation occurs via a similar solution mechanism to the
formation of ZnAl2O4. In contrast, however, Fe2O3 acts
as a diluent in the iron containing ash and reduces the
amount of liquid available for the degradation process.
4. The interaction of DPF surfaces with the ashes studied
herein strongly indicates that thermochemical degrada-
tion of cordierite by ashes representative of typical die-
sel engine systems is unlikely.
5. As the interaction times for a specific temperature used
in the experiments conducted in this work were at least
three times longer than that which might arise in an
actual DPF, it can be fairly safely concluded that
In addition to the formation of the ZnAl2O4 spinel, a cor-
responding FeAl2O4 spinel is formed by the interaction of
cordierite with the Fe2O3 containing ash at temperatures of
1100°C and 1250°C. It would therefore appear that if this
were an additional degradation mechanism, it would only be
likely to augment cordierite degradation at these tempera-

February 2012
Degradation Resistance of Cordierite Diesel Particulate Filters
753
10L. Montanaro, A. Bachiorrini, and A. Negro, “Deterioration of Cordierite
Honeycomb Structure for Diesel Emissions Control,” J. Euro. Ceram. Soc., 13
[2] 129–34 (1994).
interactions between cordierite and the three represen-
tative ashes examined herein would not be expected to
cause thermochemically induced failure of cordierite
DPFs over a 160 000 km (100 000 mile) warranted life
span.
11L. Montanaro and A. Bachiorrini, “Influence of Some Pollutants on the
Durability of Cordierite Filters for Diesel Cars,” Ceram. Int., 20 [3] 169–74
(1994).
12G. Del Pin, S. Maschio, S. Bruckner, and A. Bachiorrini, “Thermal Inter-
action Between Some Oxides and Zircon as a Material for Diesel Engines
Filter,” Ceram. Int., 30 [2] 279–83 (2004).
13N. Maier, K. G. Nickel, C. Engel, and A. Mattern, “Mechanisms and
Orientation Dependence of the Corrosion of Single Crystal Cordierite
by Model Diesel Particulate Ashes,” J. Euro. Ceram. Soc., 30 [7] 1629–40
(2010).
References
1R. J. Outland, “Ceramic Filters for Diesel Exhaust Particulates”; US Patent
No. 4,276,071, 1981.
14A. Sappok and V. W. Wong, “Lubricant-Derived Ash Properties and
Their Effects on Diesel Particulate Filter Pressure-Drop Performance,” J. Eng.
Gas Turb. Power, 133 [3] Article Number 032805 (2011).
2J. Adler, “Ceramic Diesel Particulate Filters,” Int. J. Appl. Ceram. Tech., 2
[6] 429–39 (2005).
3G. V. Gibbs, “The Polymorphism of Cordierite I: The Crystal Structure of
Low Cordierite,” Am. Min., 51 [7] 1068–87 (1966).
15D. O’Sullivan, M. J. Pomeroy, S. Hampshire, and M. J. Murtagh, “Degra-
dation Resistance of Silicon Carbide Diesel Particulate Filters to Diesel Fuel
Ash Deposits,” J. Mater. Res., 19 [10] 2913–21 (2004).
4D. Taylor, “Thermal-Expansion. 13. Complex Oxides with Chain, Ring and
Layer Structures and the Apatites,” Brit. Ceram. Trans. J., 87 [3] 88–95 (1988).
5M. J. Murtagh, D. L. Sherwood, and L. S. Socha, “EX-80, A New Diesel
Particulate Filter Composition with Improved Durability and Performance,”
SAE Paper No. 940235, SAE International Congress and Exposition, Detroit,
MI, U.S.A.,1994.
16D. O. Sullivan, S. Hampshire, M. J. Pomeroy, and M. J. Murtagh, “Com-
parison of Corrosion Resistance of Cordierite and Silicon Carbide Diesel Par-
ticulate Filters to Combustion Products of Diesel Fuel Containing Fe and Ce
Additives,” Ceram. Eng. Sci. Proc., 25 [3] 415–20 (2004).
17O. Kubaschewski, E. L. Evans, and C. B. Alcock, Metallurgical Thermo-
chemistry. 5th edition. Pergamon, Oxford, 1979.
6K. Chen, K. S. Martirosyan, and D. Luss, “Temperature Excursions during
Soot Combustion in a Diesel Particulate Filter (DPF),” Ind. Eng. Chem. Res.,
49 [21] 10358–63 (2010).
18E. M. Levin, C. R. Robbins, and H. F. McMurdie, “Figure 2489“; p. 136
in Phase Diagrams for Ceramists, 1969 Supplement, Edited by M. K. Reser.
Pub. Am. Ceram. Soc., Columbus, OH, 1969.
7J. P. A. Neeft, “Catalytic Oxidation of Soot: Potential for the Reduction of
Diesel Particulate Emissions”; Ph.D. Thesis, Technische Universiteit Delft, the
Netherlands, 1995.
19E. M. Levin, C. R. Robbins, and H. F. McMurdie, “Figure 338“; p. 133
in Phase Diagrams for Ceramists, Edited by M. K. Reser. Pub. Am. Ceram.
Soc., Columbus, OH, 1964.
8M. C. Law, A. Clarke, and C. P. Garner, “The Effects of Soot Properties
on the Regeneration Behaviour of Wall-Flow Diesel Particulate Filters,” Proc.
Inst. Mech. Eng. Part D J. Automob. Eng., 218 [D12] 1513–24 (2004).
9L. Montanaro, S. Pagliolico, and A. Negro, “Durability of Cordierite Hon-
eycomb Structure for Automotive Emissions Control,” Thermochim. Acta.,
227, 27–33 (1993).
20R. S. Ingram-Ogunwumi, Q. Dong, T. A. Murrin, R. Y. Bhargava, J. L.
Warkins, and K. Heibel, “Performance Evaluations of Aluminum Titanate
Diesel Particulate Filters” SAE Paper no. 2007-01-0656, SAE World Congress
and Exhibition, Detroit, MI, U.S.A., 2007.
h