Journal of Biomaterials and Nanobiotechnology, 2011, 2, 329-334
doi:10.4236/jbnb.2011.23040 Published Online July 2011 (http://www.SciRP.org/journal/jbnb)
329
Microleakage of Nanofilled Composite Resin
Restorative Material
Ibrahim M. Hamouda1*, Haga g Abd Elkader 1, Manal F. Badawi1
Dental Biomaterials Department, Faculty of Dentistry, Mansoura University, Mansoura, Egypt.
Received November 11th, 2010; revised March 30th 2011; accepted May 28th, 2011.
ABSTRACT
The role of nanofillers in reducing the micro leaka ge of dental compo site resins has not been previously in vestigated . So
this study was designed to evaluate microleakage of nanofilled composite resin in comparison to the conventional hy-
brid composite. Twenty extracted so und molars were selected. Class II cavities were prep ared. All cavities were etched
(enamel and dentin) with 37% phosphoric acid. Dentin bonding agents were applied to etched tooth surfaces and re-
stored with nanofilled and hybrid composite restorative materials. The restored teeth were thermocycled. Specimens
were immersed in 2% methylene blue dye, sectioned along the mesio-distal direction; dye penetration of occlusal and
gingival margin s of each section wa s evaluated using a stereo-micro scope. No significant difference was fou nd between
the microleakage of nanofilled and hybrid composite restorations at occlusal/enamel and at gingival/dentin margins.
Also, there were no significant differenc es for nanofilled composite restorations at occlusa l/enamel margins and gingi-
val/dentin marg ins. On the other hand, there were a significan t differences for hybrid composite restorations a t occlu-
sal/enamel margins and gingival/dentin margins.
Keywords: Microleakage, Nanofilled Composite Resin, Hybrid Composite Resin
1. Introduction
Resin composites are increasingly used for restorative
purposes because of good esthetic and the capability of
establishing a bond to enamel and dentin [1]. However,
like all dental materials, composites have their own limi-
tations, such as the gap formation caused by polymeriza-
tion contraction during setting, leading to marginal dis-
coloration and leakage [2]. Improvements of mechanical
properties of the composite have permitted its use in
posterior teeth with greater reliability than was the case
some years ago. This improvement included; development
of smaller particle sizes of filler, better bonding systems,
curing refinements and sealing systems [3].
Composite resin materials have progressed from mac-
rofills to microfills and from hybrids to microhybrids,
and new materials such as packable and nanofilled com-
posites have been introduced to the dental market [4,5].
Each type of composite resin has certain advantages and
limitations. The universal hybrid composites provide the
best general blend of good material properties and clini-
cal performance for routine anterior and posterior resto-
rations [6]. A new brand of composite resins called nan-
ofilled composites has been introduced to the dental mar-
ket, which has been produced with nanofiller technology
and formulated with nanomer and nanocluster filler par-
ticles. Nanomers are discrete nanoagglomerated particles
of 20 - 75 nm in size, and nanoclusters are loosely bound
agglomerates of nano-sized particles. The combination of
nanomer-sized particles and nanocluster formulations
reduces the interstitial spacing of the filler particles and,
therefore, provides increased filler loading, better physi-
cal properties, and improved polish retention [3].
This investigation was designed to evaluate enamel
and dentin microleakage of a nanofilled composite resin
in comparison with conventional hybrid composite re-
storative materials.
2. Materials and Methods
The materials used in this study are presented in Table 1.
A total of 20 specimens were prepared from both nano-
filled and conventional hybrid composite resins. Speci-
mens were cured with a light curing device (Chromalux
E, Germany) according to the manufacturer’s instruc-
tions.
Twenty freshly extracted sound (non-carious and non-
restored) mandibular human molars were selected and

Microleakage of Nanofilled Composite Resin Restorative Material 330
Table 1. Materials used in this study.
Materials Type Composition Batch/lot No. Manufacturer
FiltekTM Supreme
Light curing
nanofilled
Composite Resin
Monomer matrix contains Bis-GMA, urethane
dimethacrylate, triethylene glycol dimethacrylate and
bis-EMA resin. Inorganic filler particles are a
combination of aggregated zirconia/silica cluster and
a non-agglomerated/non- aggregated silica filler.
3910A3.5B 3M ESPE Dental Products
St. Paul, MN 55144
Prime-Dent®
Visible light
cured hybrid
composite resin
It is based on BIS-GMA resin and inorganic filler
particles with average diameter of 1.40 microns MB010
Prime Dental Manufacturing
INC. 3735 W. Belmont Ave.
Chicago, IL. 60618
AperTM Single
Bond2 Adhesive
Light curing
Bonding agent Adhesive containing 10%, 5 nm colloidal filler 4BK51202
3M ESPE Dental Products
St. Paul,
MN 55144
SDI Super Etch Etchant Gel 37% wt phosphoric acid 00601 SDI
cleaned, polished using scalars and pumice, and were
stored in distilled water until being used. Class II cavities
were prepared with a number 836 cylindrical diamond
bur (Diatech diamond AG, Swiss). The cavities were
prepared following a standardized pattern in which class
II cavity had a length of 3.0 mm, width of 2.0 mm, and
depth of 2.0 mm occlusally. The proximal box had an
axial depth of 1.5 mm and buccolingual width of 4.0 mm.
The cervical margin of the proximal box was located 1.0
mm below the CEJ [7]. The specimens were then ran-
domly divided into two experimental groups, with 10
teeth each. All cavities were etched (enamel and dentin)
with 37% phosphoric acid for 15 seconds according to
the manufacturer’s instructions, rinsed with air/water
spray for 20 seconds followed by gentle drying for 5 se-
conds. Dentin bonding agent was applied to the etched
tooth surface. AdperTM Single Bond 2 Adhesive was ap-
plied to group a1 and Prime- Dent® was applied group a2.
The teeth were then restored with Filtek Supreme (group
a1) and Prime-Dent® (group a2). The resin-based com-
posites were placed incrementally in three layers, and
each layer was cured for 40 seconds from the occlusal
direction according to the manufacturer’s instructions for
cavity class II compound. After the last increment was
placed, the matrix band was removed. The restoration
was also light-cured for 40 seconds from both the buccal
and lingual walls according to the manufacturer’s in-
structions [7].
The restored teeth were stored for 24 hours in distilled
water, and thermocycled for 2500 cycles between 5˚C
and 55˚C with a dwell time of 30 seconds in each bath
[8]. The apices of the specimens were sealed with sticky
wax, and all tooth surfaces were covered with two coats
of clear nail polish with exception of 1.0 mm around the
tooth-restoration margins and allowed to air dry. Speci-
mens were then immersed in 2% methylene blue dye.
The teeth were sectioned along the mesio-distal direction,
coincident with the center of the restoration, with a sec-
tioning diamond disc under water spray from chip sy-
ringe. The dye penetration of the occlusal and gingival
margins of each section was evaluated independently by
the observers using a stereo-microscope (Olymbus SZ 60,
Japan) at a magnification of X 10 and scored as follow
[7]: 0-No dye penetration; 1-Dye penetrations up to but
not beyond 1/2 the occlusal or gingival wall; 2-Dye pe-
netration up to but not contacting the axial wall.
3. Statistical Analysis
The data obtained were tabulated for statistical analysis
which was conducted using SPSS (Statistical Package for
Social Science) version 10. Chi-Square test was used to
detect the significant differences among the variables
tested.
4. Results
The results of microleakage are presented in Tables 2-5.
Chi-square test demonstrated no significant difference
between the microleakage of nanofilled composite (Filtek
Supreme) and hybrid composite (Prime-Dent) restora-
tions at occlusal/enamel margins (P > 0.05) (Table 2). At
the same time, there were no significant differences at
gingival/dentin margins (P > 0.05) (Table 3).
There were no significant differences for nanofilled
composite restorations at occlusal/enamel margins and
gingival/dentin margins (P > 0.05) (Table 4). On the
other hand, there were a significant differences for hybrid
composite restorations at occlusal/enamel margins and
gingival/dentin margins (P ≤ 0.01) (Table 5).
5. Discussion
Microleakage is defined as dynamic clinically undetect-
able passage of bacteria, fluids, chemical substances,
molecules and ions between the cavity walls and the re-
storative material applied. Microleakage is used as a

Microleakage of Nanofilled Composite Resin Restorative Material 331
Table 2. Chi-square (x2) and leakage score values of test
materials in occlusal/enamel margins.
Materials Leakage Scores Chi-Square (x2)P-value
Scores 0 1 23
Nanofilled composite
(n = 10) 5 2 1 2
Hybrid composite
(n = 10) 3 2 2 3
1.62 0.65
Table 3. Chi-square (x2) and leakage score values of test
materials in gingival/dentin margins.
Materials Leakage Scores Chi-Square (x2)P-value
Scores 0 1 23
Nanofilled composite
(n = 10) 1 3 1 5
Hybrid composite
(n = 10) 0 1 3 6
6.47 0.09
Table 4. Chi-square (x2) and leakage score values of nano-
filled composite restorative materials in occlusal/enamel
and gingival/dentin margins.
Materials Leakage Scores Chi-Square (x2)P-value
Scores 0 1 23
Nanofilled composite
(n = 10) 1 3 1 5
Hybrid composite
(n = 10) 0 1 3 6
6.47 0.09
Table 5. Chi-square (x2) and leakage score values of hybrid
composite restorative materials in occlusal/enamel and gin-
gival/dentin margins.
Materials Leakage Scores Chi-Square (x2)P-value
Scores 0 1 23
Nanofilled composite
(n = 10) 3 2 2 3
Hybrid composite
(n = 10) 0 1 3 6
6.47 0.09
measure by which clinicians and researchers can predict
the performance of restorative materials in the oral envi-
ronment [9,10].
Microleakage is caused by polymerization shrinkage
of composite restorative materials. High bond strength
between the restoration and the dentin surface may resist
the polymerization shrinkage of the restoration and sub-
sequent microgap formation at the tooth-restoration in-
terface [11]. Poor adaptation of the restorative materials
to cavity walls and margins and the method by which the
restorative material is inserted may affect the sealing
properties of the restorative material [12,13]. Difference
in coefficient of thermal expansion and contraction be-
tween tooth structure and the applied restorative material
had been implicated in microleakage through marginal
percolation or through disruption of the marginal enamel
etch bond, allowing microleakage in the space resulting
from thermal contraction [14]. It has also been found that
the type of occlusion and masticatory forces have a
marked effect on the development of marginal leakage in
composite restorations. The frequency of marginal leak-
age was significantly greater in teeth that were in func-
tional occlusion than in similar teeth without antagonist
[14,15]. The oral environment is also of importance in
determining the extent of marginal leakage, where both
restoration and surrounding tooth substance are subjected
to mechanical loading and temperature variations when
become in contact with food, saliva and microorganisms
[15,16].
Modulus of elasticity of the restorative material can also
be considered one of the causes of marginal microleakage
[17]. Therefore, the importance of applying an intermedi-
ary layer with a low elasticity module or a stress breaker
layer. This layer would then provide enough flexibility to
compensate the tension generated by polymerization
shrinkage [18].
In clinical practice, three commonly encountered prob-
lems may be associated with microleakage in dental res-
torations. These problems are postoperative sensitivity
[19], marginal percolation [14], secondary marginal car-
ies [19]. The incremental placement technique in the re-
storation of Class II cavity preparations seems to im-
prove the marginal seal of the proximal walls of finished
restorations [20].
In this study dye penetration method was used to
evaluate the microleakage because methylene blue dye
penetration method provides the evaluators with a perfect
and easy visualization of the prepared cavity in the digi-
tal images which provide the evaluators with a clear ref-
erence point from which to score. The dye also provides
an excellent contrast with the surrounding environment
[21].
All tested groups showed dye penetration at the
tooth-restoration interface. This could be attributed to the
dimensional changes of the resin material which often
result from polymerization shrinkage of the restorative
resin, and differences in coefficient of thermal expansion
and contraction between the tooth and the restorative
material. These changes in the material produce internal
forces that results in gap formation at the tooth-restora-
tion interface, which in turn causes microleakage [22].

Microleakage of Nanofilled Composite Resin Restorative Material 332
The results of the present study showed that, the score
of microleakage values was lower for nanofilled com-
posite than hybrid composite in all groups, but the dif-
ference was not significant. The score of microleakage
values was lower for two types of composite restorations
at the occlusal margins than at the gingival margins
placed below the cemento-enamel junction (CEJ) and the
difference was significant for hybrid composite restora-
tions, but was not significant for nanofilled composite
restorations. These results could be attributed to the in-
adequate adaptation of resin composite restoration in
posterior teeth caused by the great dentin/enamel propor-
tion in the cervical area and the critical difference of the
thermal expansion coefficient between the tooth and the
restorative material which allowed tracing agent penetra-
tion [23-26].
In this study, gingival/dentin margins showed signifi-
cantly higher leakage than occlusal/enamel margins in
hybrid composite restorative materials. This was ex-
pected as the bond strength to enamel is usually higher
than bond strength to dentin, as dentin is a less favorable
bonding substrate and the heterogeneous structure of
dentin also affects the quality of bonding of the current
dentine bonding systems [27-31]. Also, the orientation of
dentinal tubules can affect the formation of the hybrid
layer. In areas with perpendicular tubule orientation, the
hybrid layer was significantly thicker than areas with
parallel tubule orientation. Therefore, the dentine surface
on the gingival floor of class II preparations may be a
surface on which good hybrid layer formation is difficult
the fact that contributed to the results of the present study
in which substantial leakage occurred [32]. In the present
study, the absence of statistically significant differences
between occlusal and gingival margins could attributed
to the high and reliable dentin bond strength of the used
adhesives. The gaps between restorative material and
cavity walls generally occur when the bonding capacity
of the adhesive systems is insufficient to resist the forces
of polymerization shrinkage of the composites [33].
Clinically, there were many attempts to reduce micro-
leakage around restorations. During cavity preparation,
all surfaces should be smoothed, cleaned dried to achieve
maximum adaptation of the restoration to the cavity
walls. It has also been found that beveling of cavosurface
margins would provide for an increase in marginal sur-
faces which in turn, would compensate for polymeriza-
tion shrinkage to some extent [34-36]. Removal of smear
layer by acid etching increase the permeability of dentin
by exposing the orifices of dentinal tubules [37,38]. In-
creasing surface area for bonding by means of acid etch-
ing technique followed by application of direct adhesive
was found to significantly reduce microleakage level
[39-41]. Good adhesion of the restorative materials to
tooth structure [9,35,37].
6. Conclusions
There was no significant difference among the tested
materials regarding the microleakage.
REFERENCES
[1] J. Mahart and R. Hickel, “Esthetic Compomer Restora-
tions in Posterior Teeth Using a New All-in-one Adhe-
sives: Case Presentation,” Journal of Esthetic and Re-
storative Dentistry, Vol. 11, No. 5, 1999, pp. 250-258.
doi:10.1111/j.1708-8240.1999.tb00406.x
[2] J. F. Roulet, “The Problems Associated with Substituting
Composite Resins for Amalgam: A Status Report on Pos-
terior Composites,” Journal of Dentistry, Vol. 16, No. 3,
1988, pp. 101-103. doi:10.1016/0300-5712(88)90001-2
[3] S. C. Bayne, H. O. Heymann and E. J. Swift, “Update on
Dental Composite Restoration,” The Journal of the Ame-
rican Dental Association, Vol. 125, No. 6, 1994, pp. 687-
701.
[4] N. Attar and M. D. Turgut, “Fluoride Release and Uptake
Capacities of Fluoride Releasing Restorative Materials,”
Operative Dentistry, Vol. 28, No. 4, 2003, pp. 395-402.
[5] J. Mahart, H. Y. Chen and R. Hickel, “The Suitability of
Packable Resin-Based Composites for Posterior Restora-
tions,” The Journal of the American Dental Association,
Vol. 132, No. 5, 2001, pp. 639-645.
[6] K. F. Leinfelder, “Posterior Composite Resins: The Ma-
terials and Their Clinical Performance,” The Journal of
the American Dental Association, Vol. 126, No. 5, 1995,
pp. 663-672.
[7] O. Bala, M. B. Üctasli and I. Ünlu, “The Leakage of
Class II Cavities Restored with Packable Resin-Based
Composites,” Journal of Contemporary Dental Practice,
Vol. 4, No. 4, 2003, pp. 1-11.
[8] D. Simone, N. B. David, P. Aikaterini and P. Ronald,
“Microleakage of Resin-Based Liner Materials and Con-
Densable Composites Using Filled and Unfilled Adhe-
sives,” American Journal of Dentistry, Vol. 16, No. 5,
2003, pp. 351-355.
[9] J. G. Bauer and J. L. Henson, “Microleakage. A Measure
of the Performance of Direct Filling Materials,” Opera-
tive Dentistry, Vol. 9, No. 1, 1984, pp. 2-9.
[10] B. Torsten, M. Brannstrom and B. Maltson, “A New
Method for Sealing Composite Resin Contraction Gaps in
Lined Cavities,” Journal of Dental Research, Vol. 64, No.
3, 1985, pp. 450-453.
doi:10.1177/00220345850640031201
[11] J. D. Eick and F. H. Welch, “Polymerization of Posterior
Composite Resins and Its Possible Influences on Postop-
erative Sensitivity,” Quintessence International, Vol. 17,
No. 2, 1986, pp. 103-111.
[12] D. Fortin, E. J. Swift, G. E. Denehy and J. W. Reinhardt,
“Bond Strength and Microleakage of Current Dentin Ad-
hesives,” Dental Materials, Vol. 10, No. 4, 1994, pp. 253-

Microleakage of Nanofilled Composite Resin Restorative Material 333
258. doi:10.1016/0109-5641(94)90070-1
[13] J. R. Holtan, G. P. Nystrom, S. E. Rensen, R. A. Phelps
and W. H. Douglas, “Microleakage of Five Dental Adhe-
sives,” Operative Dentistry, Vol. 19, No. 2, 1993, pp.
189-193.
[14] V. Quist, “The Effect of Mastication on Marginal Adap-
tation of Composite Restorations in vivo,” Journal of
Dental Research, Vol. 72, No. 1, 1993, pp. 490-494.
[15] D. B. Mohler and L. W. Nelson, “Factors Affecting on
the Marginal Leakage of Amalgam,” The Journal of the
American Dental Association, Vol. 108, No. 1, 1984, pp.
51-54.
[16] M. Brannstrom, “Communication between the Oral Cav-
ity and the Dental Pulp Associated with Restorative
Treat- ment,” Operative Dentistry, Vol. 9, No. 2, 1984,
pp. 57-66.
[17] D. H. Retief, “Do Adhesive Prevent Microleakage?” In-
ternational Dental Journal, Vol. 44, No. 1, 1994, pp.
19-26.
[18] T. C. Steet, J. Perdigão and E. J. Swift, “Marginal Adap-
tation of Composite Restorations with and without Flow-
able Liner,” Journal of Dental Research, Vol. 79, No. 2,
2001, pp. 269-275.
[19] M. Brannstrom and K. J. Nordenvall, “Bacterial Penetra-
tion, Pulp Reaction and Inner Surface of Concise Enamel
Bond Composite Fillings in Etched and un Etched Cavi-
ties,” Journal of Dental Research, Vol. 57, No. 1, 1978,
pp. 3-10. doi:10.1177/00220345780570011301
[20] K. W. Fouad and J. S. H. Firas, “Evaluation of the Mi-
croLeakage at the Proximal Walls of Class II Cavities
Restored Using Resin Composite and Procured Compos-
ite Inserts,” Quintessence International, Vol. 34, 2003, pp.
600-606.
[21] J. B. Almeida, J. A. Platt, Y. Oshida, B. K. Moore, M. A.
Cochran and G. J. Eckert, “Three Different Methods to
Evaluate Microleakage of Packable Composites in Class
II Restorations,” Operative Dentistry, Vol. 28, No. 4 ,
2003, pp. 453-460.
[22] E. Elias and G. Sajjan, “Effect of Bleaching on Micro-
Leakage of Resin Composite Restorations in Non-Vital
Teeth. An in-vitro Study,” Journal of Endodontics, Vol.
14, 2002, pp. 9-13.
[23] W. W. Barkmeier and A. L. Cooley, “Laboratory Evalua-
tion of Adhesive System,” Operative Dentistry, Vol. 17,
Suppl 5, 1992, pp. 50-61.
[24] J. Kanca, “Posterior Composite Microleakage Below the
Cementoenamel Junction,” Operative Dentistry, Vol. 18,
No. 5, 1987, pp. 347-349.
[25] N. Nakabayashi, K. Kojima and E. Masuhara, “The Pro-
motion of Adhesion by Infiltration of Monomers into
Tooth Substrates,” Journal of Biomedical Materials Re-
search, Vol. 16, No. 3, 1982, pp. 265-273.
doi:10.1002/jbm.820160307
[26] R. R. Russel and R. B. Mazer, “Should Flowable Com-
posites be Used as Liners for Class II Restorations?”
Journal of Dental Research, Vol. 78, No. 3, 1999, pp.
389-392.
[27] W. S. Eakle and R. K. Ito, “Effect of Insertion Technique
on Microleakage in Mesio-Occlusodistal Composite Re-
sin Restorations,” Quintessence International, Vol. 21,
No. 5, 1990, pp. 369-374.
[28] T. J. Hilton, R. S. Schwartzs and J. L. Ferracane, “Micro-
leakage of Four Class II Resin Composite Insertion Tech-
niques at Intraoral Temperature,” Quintessence Interna-
tional, Vol. 28, No. 2, 1997, pp. 135-144.
[29] C. Leevailoj, M. A. Cochran, B. A. Martis, B. K. Moore
and J. Platt, “Microleakage of Posterior Packable Resin
Composites with and without Flowable Liners,” Opera-
tive Dentistry, Vol. 26, No. 3, 2001, pp. 302-307.
[30] L. A. Linden, O. Kallskog and M. Wolgast, “Human Den-
tin as a Hydrogel,” Archives of Oral Biology, Vol. 40, No.
11, 1996, pp. 991-1004.
doi:10.1016/0003-9969(95)00078-4
[31] A. L. Neme, B. B. Maxson and F. E. Pink, “Microleakage
of Class II Packable Resin Composites Lined with Flow-
ables: An in vitro Study,” Operative Dentistry, Vol. 27,
No. 6, 2002, pp. 600-605.
[32] M. Ogata, M. Okuda and M. Nakajima, “Influence of the
Direction of Tubules on Bond Strength to Dentin,” Op-
erative Dentistry, Vol. 26, 2001, pp. 27-35.
[33] C. Parti, L. Tao, M. Simpson and D. H. Pashley, “Per-
meability and Microleakage of Class II Resin Composite
Restorations,” Journal of Dentistry, Vol. 22, No. 1, 1994,
pp. 49-56.
[34] A. Ben-Amar and H. S. Cardash, “The Fluid-Filled Gap
under Amalgam and Resin Composite Restorations,”
American Journal of Dentistry, Vol. 4, No. 5, 1991, pp.
226-230.
[35] F. Lutz, T. Imfeld, F. Barbakow and W. Iselin, “Optimiz-
ing the Marginal Adaptation of MOD Composite Resto-
rations. In: G. Vanherle and D. C. Smith, Eds., Posterior
Composite Resin Dental Restorative Materials, Peter
Szulc Publishing Co., Netherlands, 1985, pp. 405-419.
[36] M. Staninec, A. Mochizuki and K. Fuckuda, “Interfacial
space, Marginal Leakage and Enamel Cracks around
Composite Resins,” Operative Dentistry, Vol. 11, No. 1,
1986, pp. 14-24.
[37] R. L. Bowen, K. R. Nemoto and J. E. Rapson, “Adhesive
Bonding of Various Material to Hard Tooth Tissue Force
Developed in Composite Materials during Hardening,”
The Journal of the American Dental Association, Vol.
106, No. 4, 1983, pp. 475-477.
[38] M. G. Buonocore, “A Simple Method of Increasing the
Adhesion of Acrylic Filling Materials to Enamel Sur-
face,” Journal of Dental Research, Vol. 34, No. 6, 1955,
pp. 849-853. doi:10.1177/00220345550340060801
[39] M. Brannstrom and H. Nyborg, “Cavity Treatment with a
Microbiocidal Fluoride Solution: Growth of Bacteria and
Effect on Pulp,” The Journal of Prosthetic Dentistry, Vol.
30, No. 3, 1973, pp. 303-310.
doi:10.1016/0022-3913(73)90187-X
[40] S. P. Gray and B. D. Chewing, “Reducing Marginal Lea-

Microleakage of Nanofilled Composite Resin Restorative Material 334
kage in Posterior Composite Resin Restorations. A Re-
view of Clinical Technique,” The Journal of Prosthetic
Dentistry, Vol. 63, No. 3, 1990, pp. 286-288.
[41] D. H. Pashley, V. Michelich and T. Kehl, “Dentin Per-
Meability Effect of Smear Layer Removal,” The Journal
of Prosthetic Dentistry, Vol. 46, No. 5, 1981, pp. 531-537.
doi:10.1016/0022-3913(81)90243-2