The flavor contribution of Turkish tobacco in blended cigarettes

Historic, archived document

Do not assume content reflects current
scientific knowledge, policies, or practices.

33/ |
3/)A ARS 73-43

}
: oe fi } i Parchas. |
: Agricultura} Researc?; Service

M.S. Department oi Agriculture

The For Official} ee
Flavor "

Contribution
of

Turkish
Tobacco

Blended
Cigarettes

U.S. DEPT. 9
NATIONAL nepie Vr AGRICULTURE

& “" VIBRA Dy:
M AY hana win
1964 cueiiter 4

Agricultural Research Service |
U. S. DEPARTMENT OF AGRICULTURE

THE FLAVOR CONTRIBUTION OF TURKISH TOBACCO
Iz]

BLENDED CIGARETTES

:

R.. Lb. Stedman and Gb. ssorllls

ABSTRACT

A study was made of the effect of added isovaleric and
B-methylvaleric acids on the flavor of smoke from blended
cigarettes that did not contain Turkish tobacco. The smoke
from blended cigarettes containing flue-cured, burley, and

Maryland tobaccos with added acids had organoleptic prop-

erties that were identical to the smoke from a blend con-

taining flue-cured, burley, Maryland, and Turkish tobaccos.

Thus, isovaleric and B-methylvaleric acids are important

constituents in the smoke flavor contributed by Turkish to-

bacco in blended cigarettes containing the four major ciga-_
rette tobaccos.

Most cigarettes made in the United States for domestic
consumption contain blends of different tobacco types. The
important types used in these blends are flue-cured, burley,
Maryland, and Turkish tobaccos.* Each type is believed to
contribute specific physical or organoleptic properties to the
finished product resulting in the overall aroma, flavor, and
burn characteristics typical of domestic cigarettes (3). **

Flue-cured tobacco is believed to contribute most of the
aroma of the cigarette and a large amount of the smoke flavor;
this type is the major component in domestic cigarette tobacco
blends and is used in amounts ranging from approximately 45
to 75 percent of the blend. Burley tobacco is added to the
blend to improve the burn rate, to contribute bulk to the ciga-
rette, and to give strength to the smoke; the levels of burley

* "Bright" and ''Virginia'' are frequently used synonyms for
flue-cured tobacco, and "Oriental'' and ''Aromatic'' are simi-
larly used for Turkish tobacco.

**Numbers in parentheses refer to literature cited at end of
this report.

in blends are usually 15 to 45 percent. Maryland tobacco con-
tributes filling power to the cigarette and is believed to im-
prove generally the smoke flavor; this type is added in small
amounts, 1 to 5 percent, to domestic blends. Turkish tobacco
imparts aroma, mildness, and a distinctive, pleasant taste to
the smoke; this tobaccois used in amounts ranging from about
5 to 15 percent in blends.

Although these generalized contributions of the tobacco
types are known, little information is available on the specific
‘chemical components that are responsible for the flavor im-
parted to smoke by the types. In fact, fundamental chemical
differences in the smoke of different tobacco types have been
investigated onlyrecently. Although of significant value, such
chemical differences may not be relatedto observed differences
in smoke flavor since the chemicalcomponents in question may
play no role in the overall smoke flavor. The present report
describes an attempt torelate previous findings on Turkishto-
bacco smoke composition to the organoleptic properties of
smoke from blended cigarettes containing such tobacco.

RESULTS AND DISCUSSION

Technical background

As a result of basic studies on the composition of tobacco
leaf and smoke conducted recently at this Division, certain
fundamental differences were found in the chemical composi-
tion of the various tobacco types and the smoke produced from
such types. Included in this work were comparative studies of
the steam-volatile neutrals of leaf (1) and cigarette smoke (2)
and steam-volatile acids of leaf (4, 5) and smoke (7). Among
the significant findings was the observation that Turkish to-
bacco leaf and cigarette smoke were strikingly different from
other cigarette tobacco types in the amounts and proportions
of B-methylvaleric and isovaleric acids (4, 5, 6, 7). Table l
presents pertinent data on this point in the case of cigarette
smoke. Also, it was shown that the addition of Turkish tobac-
co to blends of bright, burley, and Maryland tobaccos resulted
in an increase in the amounts of the steam-volatile acids found
in the smoke compared to the levels expected from an additive
contribution of each tobacco (7). For isovaleric and B-methyl-
valeric acids, increases of 48 to 62 percent greater than ex-
pected were obtained.

These unusual properties of Turkish tobacco prompted fur-
ther investigations of the influence of isovaleric and 8B-methyl-
valeric acids on the flavor of cigarette smoke which are de-
scribed below.

Table 1 -- Quantitative differences in certain volatile acids of smoke
condensates from various blended or unblended cigarettes.

Milligrams per

105 cigarettes Ratios of amounts**

Acid (flue-cured)* F B M ie FBM FBMT
Isobutyric 2.4 LO) OS O anh aloe 0.74 es
n-Butyric 4.6 Lew ou BOS oe . 46 nod
Isovaleric Dad 10) . 84 veo Or ed E63
Unidentified A 0.6 Eo) SOG, mere Gis 3) ) 04
n-Valeric T0 TO . oo . 46 . 90 . 61 . 68
Unidentified B On Lad re 215 7) . 62 . 89 init
beta-Methylvaleric 2.6 10 Bef 0) Gi Cokin . Lo 3.93
Isocaproic se 7 170 . 64 aU = . 80 -
n-Caproic 0.6 1.0 poo ge AG tans “00 1. 23
Unidentified C Ore 1; '0 . 48 Ft aL aoo be . 96
Unidentified D Vet, 0 Ril aie) ra . 68 al
Unidentified E 3.0 1.0 . 390 , Oo .95 . 62 , 90
n-Heptylic 5). 0 1; 0 . 46 . 68 . 60 wad 1 |
Unidentified F 2 70! Wao eadegwe Rig TERRE

* Total weights (milligrams per 105 cigarettes) for indicated acids were as
follows: Turkish 48.3;flue-cured, 30.6; Maryland, 23.7; burley, 16.5;
FBM blend, 20.5; FBMT blend, 40.5. See footnote below for blend com-
positions. Unidentified acids were calculated as n-caproic acid.

**Relative amounts based on the levels in flue-cured cigarettes. Weights
were based on smoking 55 percent of cigarette length. F = flue-cured,
B = burley, M = Maryland, T = Turkish, FBM = blend (60 percent flue-
cured, 35 percent burley, 5 percent Maryland), FBMT = blend (40 per-
cent flue-cured, 35 percent burley, 5 percent Maryland, 20 percent
Turkish).

Organoleptic studies

Essentially, the studies were designed to compare by panel
test the flavor of cigarettes containing Turkish tobacco tociga-
rettes containing isovaleric and B-methylvaleric acids in place
of Turkish tobacco. Initially, candidates for the test panel
were screened repeatedly by the Triangle Method, using one
cigarette of a blend containing flue-cured, burley, Maryland,
and Turkish tobaccos and two cigarettes containing the same
blend except that the Turkish was replaced with an equivalent
amount of flue-cured. Eight out of twenty-five prospective
panelists were ultimately accepted in this screening and used
in the tests described below.

In subsequent tests, the Paired Comparison Technique was
employed, using two cigarettes. The first cigarette contained
flue-cured, burley, Maryland, and Turkish tobaccos. ‘The
second cigarette had the same composition as the first ciga-
rette except that the Turkish tobacco was replaced by a com-
parable amount of flue-cured tobacco and a combination of iso-
valeric and B-methylvaleric acid. Panelists were asked to
smoke the twocigarettes and indicate whether or not the ciga-
rettes were identical inorganoleptic properties. If differences
were found, the panelists were requested to describe the dif-
ference in detail. Various control tests of acuity were pre-
sented periodically tothe panelists as a check onthe well-known
daily fluctuations in human sensitivities. When a panelist was
found to be relatively insensitive, his results were discarded.
The inadequacies of ‘panel testing are well known and should
be realized in evaluating such data.

A wide range of concentration of constituent tobaccos and
of added acids was investigated. Table 2 contains dataon some
combinations of tobaccos with or without added acids which
were tested comparatively for similarities in smoke flavor.
The cigarettes in test No. 3 were commercially manufactured;
all other cigarettes were hand-rolled in the laboratory. The
low grade flue-cured tobacco was a sample previously judged
to be flavor-deficient in paneltests by adomestic manufacturer.
The cigarettes contained no filters, humectants, or flavoring
agents other than the added acids, and the moisture contents
were adjusted to 13.0 percent before panel testing. The acids
were addedto the cigarettes by injection of an acetone solution
containing 0. 026 percent isovaleric and 0. 094 percent B-methyl-
valeric acids, respectively, using a 50=microliter hypodermic
syringe. The syringe was provided with a 19-gage needle hav-
ing 4 holes of 1/32-inch diameter uniformly distributed along
the needle with the original aperture sealed. The solvent was
removed from each cigarette by a gentle stream of warm air
or nitrogen.

)

Table 2 -- Blend compositions used in panel tests.

Control
Test

Control
Test

Control
Test

Control
Test

* Blends composed of high grade flue-cured (HGF), burley (B),
Maryland (M), Turkish (T) and/or low grade flue-cured (LGF)
tobaccos. Cigarettes in test No. 3 were 85 millimeters in
length and all others were 70 millimeters.

**Micrograms of isovaleric (IV) or beta-methylvaleric (BMV)
acids added per cigarette.

The panel tests on all cigarettes in table 2 showed that no
organoleptic differences could be detected between the test -
and control cigarettes in each combination. These findings
indicate that isovaleric and B-methylvaleric acids are impor-
tant constituents in the smoke flavor contributed bythe Turkish
tobacco in domestic blends of cigarettes. In addition, these
constituents may contribute greatly to the distinctive aroma of
the sidestream smoke of Turkish cigarettes and blended ciga-
rettes containing Turkish tobacco.

Stability

Although isovaleric and B-methylvaleric acids are not highly
volatile substances, it was of interest todetermine the stability
of the acids under certain conditions.

A series of stability tests was conducted on cigarettes con-
taining a blend of tobaccos (60 percent flue-cured, 35 percent
burley, and5 percent Maryland) and the added acids (6.5 micro-
grams isovaleric acid and 11.0 micrograms B-methylvaleric
acid per cigarette). The cigarettes were sealed in standard
cigarette packages in the usual manner and then tested peri-
odically by the panel for changes in organoleptic properties.

6

After a storage of 2 months, no changes in flavor were ob-
served. These studies are continuing.

In another investigation, a series of cigarettes was pre-
pared by spraying the acids on the burley tobacco to be used
in the blend. Theburley and other tobaccos used in the blend
were then blended for 5 hours in a vigorous tumbling operation.
Subsequently, thetobacco blend was made into cigarettes and
tested by the panel. No difference in organoleptic properties
was observed between these cigarettes and control cigarettes
in which the additives were injected from an acetone solution
in the usual manner.

Other considerations

Recently, there has been some interest in the addition of
chemical substances to the filters of filter-tipped cigarettes
to alter the flavor or other properties of the smoke. Also,
there has been some effort in developing nonvolatile tobacco
additives that release volatile flavoring agents on burning. In
the present study, the effect of adding the calcium salt of iso-
valeric acid to the filter of filter-tipped cigarettes was inves-
tigated superficially. Althoughthe calcium salts are known to
decompose pyrolytically to ketones, it was felt that the moist,
acidic condition of cigarette smoke might effect some hydroly-
sis of the salt and release free acid.

The calcium salt of isovaleric acid was prepared by the
following method: 0.1 moleof the acid and 0.05 mole of calcium
carbonate were added to 50 ml. of water and the mixture was
refluxed for 5 hours. Thirty ml. of benzene were added tothe
reaction flask after cooling,and a benzene-water azeotrope was
continuously distilled intoa trapfrom whichthe water was sep-
arated. The salt was collected, washed several times with
ether, and dried in a vacuum desiccator for 3 days. Although
slightly low, the analytical value for the calcium content of the
product was considered satisfactory: calculated, 16.5 percent;
observed, 15.4 percent. The addition of approximately 0.5
milligram of this calcium salt to the filter of a filter-tipped
cigarette produced some change in the flavor of a cigarette
containing only flue-cured, burley, and Maryland tobaccos.
The test panel reported the taste to be ''different'' than a con-
trol cigarette without added calcium salt and ''somewhat simi-
lar to but not identical'' with a control cigarette containing all
four tobacco types.

some difficulty was encountered in the synthesis of calcium
B-methylvalerate by the above method and work on this problem
is continuing.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

LITERATURE CITED

Burdick, D., Schmeltz, I., Miller, R. L., and Stedman,

1 ee

1963. Composition studies on tobacco. XIV. Steam-vol-
atile, neutral substances in various types and grades.
Tebacco,sei,.. 73; 945100.

and Stedman, R. L.
1963. Composition studies on tobacco. XV. Steam vola-
tile, neutral substances in smoke from blended and
unblended cigarettes. “Tobacco, Sey fale

Darkis, »-E’..-Ri.j-and, Hackney); Ea+J:

1952. Cigarette tobaccos. Chemical changes that occur
during processing. Indus. and Engin. Chem. 44:
284-291.

schmeltz,..t..,.Miller,..R., Ls; and Stedman, -h ale

1963. Gas chromatographic study of the steam-volatile
fatty acids of various tobaccos., Jour. Gas Chrom-
atog. 1,(8): 27-28.

Stedman, R. L., and Miller, R. L.
1963. Composition studies on tobacco. XVI. Steam-vol-
atile acids. Assoc. Off. Agr. Chem. Jour. © 46:
779-784.

Stedman, R. L.

1964, Aroma, flavor and chemical composition of ciga-
rette tobacco and smoke, Bull. d'Inform. CORESTA,
No, 1963-4; ,1L-23)

Burdick, D., and Schmeltz, I.
1963. Composition studies on tobacco. XVII. Steam-vol-
atile acidic fraction of cigarette smoke. Tobacco
eci.. 7; 166-169,

A report of work done at the

EASTERN UTILIZATION
RESEARCH AND DEVELOPMENT DIVISION

600 E. Mermaid Lane
Philadelphia, Pa. 19118

-
+
‘ |
'
J
i
b
:
4