KARAKTERISTIK KOMPONEN VOLATIL, KADAR N-TOTAL DAN KADAR LIPID DALAM

KOPI ROBUSTA OLAH KERING DAN BASAH

CHARACTERIZTIC VOLATILE COMPONENTS, NITROGEN TOTAL AND PERCENT LIPID

FROM ROBUSTA COFFEE DRY AND WET PROCESS

Ika Oktavianawati, S.Si., M.Sc1)*, Nikmatus Sholeha 2)*, Laiqotul Lutfiah 3)*,
Drs. Siswoyo M.Sc., Ph.D4)*, Drs. Busroni M.Si 5)*
1)
Universitas jember, email: bag_cha@yahoo.com.
2)
Universitas jember, email: nikmatus_sholeha@outlook.com
3)
Universitas jember, email: laiqlutfya@gmail.com
4)
Universitas jember, email: s.siswoyo@gmail.com
5)
Universitas jember, email: bushyogiya@yahoo.com

ABSTRACT

This research has investigated the characterization of robusta coffee in dry and wet
proses from local producer brand Kopi Ketakasi located in Silo, Jember. Characterization is
about chemical change in coffee resulting from dry or wet processing and roasting, it include
volatile components, nitrogen total and percent lipid in coffee. Extraction method for obtaining
essential oil to know what the volatile components is steam distillation, continued with GC-MS
characterization. The nitrogen total from coffee were carried out by Kjehdahl method. Lipid
extraction from coffee was carried out using Soxhlet method with petroleum ether as the
solvent. The result show that volatile components are not conducted with literature, it contains
high boiling point compound mostly like long chain hydrocarbon also fatty acids and their
derivatives. The total nitrogen from dry process coffee has the bigger percentage from wet
process, but not much different from the effect of roasting process. Percent lipid from dry
processing coffee has the higher result from the wet, roasting process makes the percent lipid
increase in both.

Keyword: Coffee, dry and wet process, volatile component, total Nitrogen, percent lipid

ABSTRAK

Penelitian ini dilakukan untuk mengkarakterisasi kopi robusta yang melalui dua jenis
teknik pengolahan yaitu pengolahan secara kering dan secara basah. Sampel kopi didapat dari
produsen lokal kopi di Desa Sidomulyo Kecamatan Silo Kabupaten Jember. Karakterisasi yang
dilakukan bertujuan untuk mengetahui perubahan kimia yang terjadi pada kopi olah kering dan
olah basah serta akibat dari proses penyangraian keduanya. Karakterisasi dilakukan untuk
mengetahui komponen volatile, kadar N-total dan kadar lipida pada kopi. Komponen volatil kopi
diekstrak menggunakan metode destilasi uap dan dikarakterisasi menggunakan GC-MS. Kadar
N-total pada kopi diketahui menggunakan metode Kjehdahl dan ekstraksi Soxhlet dengan
pelarut PE dilakukan untuk mengetahui kadar lipida pada kopi. Senyawa volatil kopi pada
penelitian ini hasilnya berbeda dengan literatur, sebagian besar senyawa yang teridentifikasi
merupakan senyawa hidrokarbon dengan titik didih tinggi serta asam lemak dan turunannya.
Kadar N-total dari kopi olah kering memiliki prosentase yang lebih besar dari kopi olah basah,
semntara efek penyangraian tidak menghasilkan perbedaan yang signifikan pada kadar N-total
kopi. Kada lipid pada kopi olah kering lebih besar dari pada kopi olah basah. Kadar lipida pada
kedua pengolahan mengalami peningkatan akibat dari proses penyangraian.

Kata Kunci: Kopi, olah kering dan basah, senyawa volatil, kadar N-total, kadar lipid

INTRIDUCTION

Coffee is one of plant that has long been cultivated, become a lead commodity and has
high economic value in Indonesia (Rahardjo, 2012). Indonesia is ranked fourth from 10
countries that has biggest coffee production. Jember city has a major contribution of coffee
production in Indonesia and coffee became one of the leading commodities in this district. In
2013, coffee production in Jember city was recorded at 2,516 tons in one year with 733 Kg / Ha
productivity and the number of coffee farmers 15,866 households (Directorate General of
Plantations 2014).
The process of coffee cherry to be coffee beans in general through two types of
processing techniques, namely dry and wet processing. Dry coffee processing through the
process of fruit solving, drying, huller and sorting seeds, while processing coffee in a wet
manner through a process of soaking with water (fermentation) before the process of fruit
solving. Wet processing coffee through the stage of sorting cherry at the beginning of the
processing, it only using red cherry (perfectly mature). Meanwhile the dry processing coffee
using coffee cherry with various maturity phases without the sorting process. Dry processing
produces coffee with a buttery and nutty flavor, while wet processing produces coffee with a
fruity, flora and caramel taste (Lee et al; 2015).
The processing of coffee beans which have the most influence on the formation of flavor
and aroma of coffee products is the roasting process (Franca et al., 2009). The volatile
components of coffee, affect the aroma that has the most important role in showing the quality
of coffee (Andueza et al; 2003). Roasted coffee contains more than 950 volatile compounds
(Farah, 2012). Volatile compounds in roasted coffee are generally furan and piran, pirazin,
pyrrole, ketones, phenols, hydrocarbons, alcohols, aldehydes, acids and anhydrides, esters,
lactones, tiophen, oxazole, thiazoles, pyridin, amines, and various sulfur and compound
contents nitrogen (Franca et al; 2009). Research on volatile coffee compounds using steam
distillation method was carried out by Vitzthum et al (1976), Kumuzawa and Masuda (2003) and
also carried out by Stoffelsma et al., (1968). GC-MS is a method of separating organic
compounds that combines two methods of compound analysis, namely gas chromatography
(GC) to analyze the number of compounds quantitatively and mass spectrometry (MS) to
analyze the molecular structure of compound analytes (Pavia, 2006). The advantage of GC-MS
analysis is its high sensitivity and ability to analyze various compounds (Fowlis, 1998).
The non-volatile components of roasted coffee are alkaloids (caffeine, trigonelline),
chlorogenic acid, carboxylic acids, carbohydrates and polysaccharides, lipids, proteins,
pigments, melanoidins, and minerals (Buffo and Cardelli, 2004). Components that contain
nitrogen are the main compounds which have an important influence in providing coffee flavor.
Three main sources of nitrogen-containing compounds in coffee are caffeine, trigonelline and
protein (Underwood and Deatherage, 1952). Types of amino acids in coffee include alanine,
arginine, aspartic acid, asparagine, glutamic acid, glycine, histidine, 3-methylhistidine,
isoleucine, leucine, lysine, proline, serine, tyrosine, valine and triptopane (Illy and Viani, 1995).
Trigonelline is an alkaloid compound which is a material of its derivative compound, nicotinic
acid which changes during the roasting process. Nicotinic acid compounds give a typical bitter
taste to coffee. Research on the effect of coffee roasting on trigonelline and nicotinic acid
concentrations in coffee has been carried out by Taguchi et al., (1985). Protein in coffee is an
ingredient of degradation and formation changes that occur during roasting (Shibamoto, 1983).
This reaction occurs between free amino acids and reducing sugars. This reaction has a
contribution to the formation of taste (Janzen, 2010). The Kjeldahl method is one of the common
methods used to determine total nitrogen levels (Winarno, 2004).
The lipid fraction in roasted coffee contains coffee oil that carries the volatile flavor in
coffee and vitamins that dissolve in lipids, this content contributes to the texture of the flavor of
coffee (Oestreich-Janzen, 2010). The composition of lipids in coffee are triglycerides, sterols,
tocopherols, diterpene and phospholipids. Lipid in coffee serves as a creator of flavor in
steeping, lipid levels in robusta coffee (green bean) is about 10% on average (Janzen, 2010).
Analysis of total lipid content using the Soxhlet method according to the AOAC method.
Research by Speer and Kolling (2006) used petroleum ether solvents to determine the lipid
fraction in coffee beans.
Based on the above background, an analysis of volatile component content, N-total
content and lipid content of dry and wet processed coffee, those analysis also applied for green
bean and roasting coffee. The extraction method of volatile compounds used is steam distillation
in combination with solvent extraction which will then be analyzed using GC-MS. The total N-
level analysis will be carried out using the Kjehdhal method and the lipid level analysis uses the
Soxhlet method. The aim of this research is to study the characteristics of non-roasted and
roasted robusta on dry and wet processing coffee beans involve volatile components, total N-
level and lipid level contained therein.

MATERIAL AND METHODS

Material
The materials used in this study were dry and wet robusta coffee beans before and after
roasting, distilled water, diethyl ether, petroleum ether, magnesium sulfate anhydrous (MgSO4),
H2SO4 (pa), 40% NaOH, 0.1 M NaOH, KHC8H4O4 0.05 M, pp indicator, 0.1 M HCl, methyl red-
methyl blue indicator, Na2SO4, CuSO4, and 4% H3BO3.

Tools
The tools used in this study are measuring glass, beaker glass, stirring rods, measuring
flasks, drop pipettes, mohr pipettes, erlenmeyer, burettes, condensers, round bottom flasks,
three neck flasks, heating mantles, statives and clamps, hot plates, measuring flasks ,
aluminum foil, desiccator, filter paper, ball pipette, spray bottle, analytical balance, distillation
set, kjeldahl tool set, oven, set of sloth equipment, GC-MS (GCMS-QP2010S SHIMADZU).

Material Preparation
Dry and wet processing robusta coffee used in green bean and roasted. Each coffee
bean is smoothed into coffee grounds. This coffee powder is then sieved with a 35 mesh sieve.

Water Content Determination

The coffee grounds used for determined in advance the water content of it. The
porcelain cup as a container is dried in an oven for 30 minutes then cooled in a desiccator and
weighed. Weigh 2.0 grams coffee grounds and place in a porcelain cup and then dried in an
oven at 105oC for 5 hours. Then it is cooled in a desiccator and weighed. The sample is then re-
ovened for 1 hour, cooled in a desiccator and weighed. This procedure is repeated until it has a
constant mass and the mass obtained is recorded. This measurement is repeated three times
(AOAC, 1970). This procedure is used on every samples of coffee.
Extraction of Volatile Compounds with Steam Distillation Method
Robusta coffee grounds was weighed 200 grams and put into a three-neck flask. The
distilled water is put in a round bottom flask. The distillation process is carried out approximately
5 hours and is set at a boiling point of water (maximum at 102 oC). Destilat is stored in a
separating funnel and left for a while until two phases (aqueous phase and organic phase) are
obtained. The distillate is then extracted with diethyl ether. The mixture is evaporated until the
solvent runs out then added with magnesium sulfate anhydrous. Extraction of volatile
compounds was also carried out again on all samples, each of which was repeated three times.

GC-MS (Gas Cromatograpy Mass Spectrometry) Analysis

Destilat is taken as much as 1 mL injected into the mobile phase of He. Injection
temperature 300oC and column temperature 70oC. The component then goes to the detector.
The results of the separation from the column then proceed to the Mass Spectrometer (MS).
Components are fired with 70 eV electrons until fragmentation occurs as m / z. The results of
analysis of the components of coffee volatile compounds in the form of chromatograms and
mass spectra, are supplemented with retention times and % of the area of each detected
compound.

Determination of N-total levels with the Kjeldahl Method

Destruction
1,000 grams of coffee grounds were put into the kjeldahl flask and then 2,0000 grams of
Na2SO4: CuSO4 (8: 1) catalyst mixture were added. The blank treatment is also made without
using a sample. The mixture then added with 12.0 mL concentrated H2SO4 followed by shaking
the mixture. The mixed flask is then placed on the stand of the destruction device in the fume
hood. Destruction is carried out on low heat and the volume of fire is raised after the smoke is
gone. Destruction is stopped after the liquid becomes clear and greenish. The destruction tool is
turned off and the pumpkin is cooled slowly to reach room temperature.

Distillation
Destruction results cooled down then it were added to 25.0 mL of distilled water and
placed on a set of distillation devices. Erlenmeyer is prepared and filled with 10.0 mL 4%
H3BO3 and 6 drops of methyl blue-blue indicator so that the solution is blue. 40% NaOH was
added in a Kjeldahl flask until the mixture was alkaline (brownish) just before it was distilled then
the distillation was done until the distillate volume filled the connecting pipe and changed the
blue color of H3BO3 to green. The results of the distillation are then titrated.

Titration
Samples and blanks were then titrated with standardized 0.0800 M HCl to endpoints..
Total N levels can be obtained according to the formula% N (AOAC, 2001). This treatment was
carried out again on all samples, each repeated 3 times.

Determination of Lipid Levels with Soxhlet Method

4 grams of coffee grounds wrapped in filter paper, then put into thimble and set a series
of soxhlet tools. Petrolum. 60 mL of petroleum ether solvent is added into the soxhlet flask.
Soxhletation is carried out for 3 hours (25 cycles) or until the solvent is no longer colored. The
solvent is evaporated clean using a rotary evaporator. The resulting lipids are oven treated at
45oC, dried and cooled in a desiccator. This treatment was repeated in all sample, each
repeated 3 times. Lipid levels were calculated using the% lipid formula according to (AOAC,
2006) and (AOAC, 1998).

RESULT AND DISCUSSION

Composition of Volatile Coffee Compounds
Extraction of volatile compounds uses steam distillation method which lasts for 5
hours. The distillate obtained is extracted using diethyl ether then concentrated. The Destilat of
each sample is determined by its yield. Non-roasted and roasted dry coffee yields were 0.02%
and 0.08% respectively, while for non-roasted and roasted wet coffee, respectively 0.16% and
0.08%. The increase in coffee yield due to roasting process occurs due to the formation of
volatile compounds in coffee which are also extracted in the steam distillation process. The
results of GCMS analysis on dry-processed coffee contained 37 compounds in non-roasted
coffee and 40 compounds that appeared in roasted coffee, but the one that was identified was a
compound that had 85% similiarity index (SI) of 21 compounds and 14 compounds. Non-
roasted coffee contains 3 aromatic compounds which are not found in roasted coffee, these
results indicate that the aromatic compound is lost during the roasting process. These results
are not in accordance with the literature. This study produced only 9 new compounds due to
roasting process, the results were very different from Farah's (2012) study which produced 950
new compounds. GCMS results in this study indicate the emergence of high boiling non volatile
compounds in the results of the analysis of volatile compounds. Carboxylic acid and long chain
alkane compounds dominate and it is known that one of the carboxylic acid groups such as
hexadecanoic acid has a high boiling point.
GC-MS results in wet processing coffee showed that the compounds that appeared were
not only volatile compounds, but non-volatile compounds were also identified in non-roasted
and roasted wet coffee. The number of compounds that appeared on non-roasted wet coffee
were 83 compounds and only 30 compounds were identified. Whereas there were 31
compounds in roasted coffee and 21 compounds were identified. The compounds that are
identified and dominated in dry and wet coffee are compounds with very large boiling points.
This occurs because the closed system at the time of extraction produces considerable
pressure, so the temperature inside also increases, as a result long chain compounds and have
a high boiling point can evaporate. The group of many volatile compounds that were not found
in distilled roasted coffee in this study. Like the pyrazine class that was not found in this study.
Furan compounds are identified in roasted coffee, but with small amounts. This result is not in
accordance with the literature, according to Clarke and Macrae (1985) and the study by
Sunarharum et al (2014) of furan and pyrazine groups are most commonly found in roasted
coffee. This occur because the extraction process is not going well. In the extraction process,
the aroma from coffee spreads out of the system, it is suspected that many volatile compounds
that evaporate. Many low boiling points which are not found in the results of this research,
because these compounds disappear or evaporate during roasting, during extraction and during
evaporation of the solvent. The other biggest possibility why it gone, is when coffee distillate
storage takes 2 months for GC-MS analysis.
An interesting fact occurs in a group of non-volatile compounds in coffee distillate if wet,
which has some similarity in the name of the compound with the results of research conducted
by Dong et al (2015). Such compounds include tetradekana, pentadecane, hexadecane,
heptadecane, and eikosana. Data analysis of volatile compounds obtained in the research of
Sunarharum et al (2014) did not exist at all in the volatile component of the results of this study.
Similarly, compounds from pyrazine, pyridine, thiol, tiophene, thiazole, terpene and sulfur-
containing compounds were not found in the results of this research. This is occurs due to
several factors such as the loss of the compound during the roasting process, storage of coffee
after roasting, and the extraction process until the coffee distillate analysis process.

Lipid Contain in Robusta Coffee

The lipid levels of non-roasted and roasted dry processing coffee were 11.34% and
16.41% respectively. Lipid content in non roasted wet coffee is 7.7% and roasted is 24.2%.
These results indicate that the roasting process affects the lipid content of coffee. Previous
research by Oestreich-Janzen (2013) show that the roasting process does not change the
amount of triacylglycerol in coffee. The high temperature roasting process will result in the
transfer of triacylglycerol in coffee beans to the surface of coffee beans (sweating). The
triacylglycerol compounds that have moved to the surface of the coffee are more easily
extracted by petroleum ether. The displacement of the triacylglycerol location resulted in the
appearance of a creamy effect on roasted coffee. According to Jansen (2010) the content of
triacylglycerol compounds is known to be found mostly in coffee, which is about 75.2% of the
total lipid in coffee. This is consistent with the results of the study which showed that the lipid
content of roasted coffee was higher than the level of non-roasted coffee.
The level of non-roasted coffee lipids is smaller because of the presence of wax on non-
roasted coffee beans. Wax on the surface of non-roasted coffee beans are carboxylic acid-5-
hydroxytryptamides (C-5-HT) (figure 4.4). These compounds cannot be extracted using
petroleum ether and can only be extracted using chlorinated organic solvents (Speer and
Kolling, 2006). The study by Viani and Horman (1975) explained that when the roasting process
occurs the C-5-HT pyrolysis reaction produces alkaloid compounds. The C-5-HT pyrolysis
reaction during roasting also resulted in the new short chain compounds in roasted coffee. The
roasting causes the thermolysis reaction of C-5-HT according to Folstar et al., (1975) produces
5-hydroxyindole, 3-methyl-5-hydroxyndole, and various types of n-alkanes, n-nitrilalkanes, and
n-acidamidealkanes. The process of soxhletation on roasted coffee using petroleum ether
solvents can extract alkaloid compounds and short compounds from C-5-HT pyrolysis. This
statement is in accordance with the results of increased dry and wet coffee lipid levels resulting
from the roasting process.

OH
NH (CH2)n R
O
N
H
1. R = CH3 ; n = 16
2. R = CH3 ; n = 18
3. R = CH3 ; n = 20
4. R = CH3 ; n = 22
5. R = CH3 ; n = 16
6. R = CH2OH ; n = 18
7. R = CH2OH ; n = 20
Picture 4.4 Structure C-5-HT in Coffee (Source: Oestreich-Janzen, 2013)

Total Nitrogen in Robusta Coffee

N-total levels in roasted and non-roasted dry coffee processing were 2.68% and 2.65%,
respectively, while the total N-levels in non-roasted and roasted wet coffee were 2.57% and
2.53% respectively. These results indicate that there was a decrease in N-total levels between
the two types of coffee, although not so significant. This decrease occurs as a result of amino
acids contributing to maillard reactions or non-enzymatic browning reactions in the roasting
process. This is consistent with Jansen (2010) that the roasting process will largely change
nitrogen-containing compounds. Maillard reactions will produce volatile compounds that
contribute to the aroma and flavor of coffee. Janzen (2010) also reported that the compounds
produced by the maillard reaction include thiophene, furfurylthiol, diketone groups, furfuril
mercaptan, pyrolysis and alkylpirazine compounds which are classified as volatile compounds.
This is also explained in the research of Lee et al. (2015).
The main compounds containing Nitrogen in coffee such as caffeine, trigonelline and
proteins undergo changes during the roasting process. The results of changes in N-total levels
due to roasting process are the same as Taguchi et al., (1985), through the reaction of
trigonelline pyrolysis in coffee to nicotinic acid compounds whose results will decrease with
increasing heat of roasting temperature. Nicotinic acid in coffee decreases due to high heating
but the main ingredient, trigonellin, has been reacted so that the total N-content in roasted
coffee has decreased. Rizzi and Sanders (1996) reported that trigonellin pyrolysis produces
several compounds through two types of reaction pathways. The result of pyrolysis reaction is
one of pyridine compounds. Pyridine volatile compounds are compounds that can be released
during the roasting process, this fact supports the results that the total N-level of roasted coffee
decreases due to roasting. These results are in accordance with the research of Buffo and
Cardelli (2004) and are described in Nursten's (2005) book on the role of protein in the Maillard
reaction.
Farah (2012) explained that other nitrogen compounds that changes due to roasting are
trigonelin. These trigonelin compounds will produce a variety compounds include nicotinic acid
(vitamin B3) and volatile compounds such as pyrrole, pyridine, and methyl nicotinic. In Clarke
and Macrae (1985) and Janzen (2010), it is explained that nicotinic acid is formed through the
demethylation of trigonelin. Nitrogen-containing and volatile compounds such as pyrrole,
pyridine and pyrazine are not found in the results of the analysis of volatile compounds. This
occurs because the extraction process of volatile compounds that are less effective or
imperfect, so that the compounds are not found in the data of components of volatile coffee
compounds.

CONCLUSION
The conclusion obtained based on the research that has been carried out is the results
of GC-MS analysis showed that non-roasted and roasted dry coffee were 21 and 14 compounds
respectively, which were dominated by long-chain alkanes and carboxylic acids, whereas for
non-roasted wet coffee and roasted successively 30 and 21 compounds. Identified compounds
are not only volatile compounds, but also non-volatile compounds. N-total levels of non-roasted
and roasted dried coffee were respectively 2.68% and 2.65%, while those in non-roasted and
roasted coffee were 2.57% and 2.53% respectively. Lipid levels obtained in non-roasted dry
coffee were 11.34% and experienced an increase in roasted coffee, the value obtained was
16.41%. Lipid content in non roasted wet coffee is 7.7% and roasted is 24.2%.

DAFTAR PUSTAKA
AOAC, 1970. Official Methods of Analysis of The Association of Analytical Chemists,
Washington D.C.

AOAC, 1998. Official Methods of Analysis of The Association of Analytical Chemists,

Washington D.C.
AOAC, 2001. Official Methods of Analysis of The Association of Analytical Chemists,
Washington D.C.

AOAC. 2006. Official Methods of Analysis of The Association of Analytical Chemist. Washington
D.C.

Andueza, S., Maetzu, L., Pascual, L., Ibanez, C., de la Pena, M.p. dan Cid, C. 2003. Influence
of Extraction Temperature On the Final Quality of Espresso Coffee. J. Science of Food
and Agr.. 83, 240-248.

Buffo, R.A. dan Cardelli-Freire, C. 2004. Coffee Flavour: an overview. Flavour and Fragrance
Journal. 19(2), 99-104.

Clarke, R.J., dan R. Macrae. 1985. Coffee Volume I : Chemistry. London: Elsevier Applied
Science Publishers.

Dong, W., Tan, L., Zhao, J., Hu, R., dan Lu, M. 2015. Characterization of Fatty Acid, Amino, and
Volatile Compound Compositions and Bioactive Components of Seven Coffee (Coffee
robusta) Cultivars Grown in Hainan Province, China. 20: 16687-16708.

Farah, Adriana. 2012. Coffee: Emerging Health Effects and Disease Prevention, First Edition.
John Wiley & Sons, Inc. Institute of Food Technologist: Wiley-Blackwell.

Folstar, P., Pilnik, W., de Heus, J. G & van der Plas, H. C. 1975. The Composition of Fatty Acidin
Coffee Oil and Wax. Lebens-mittelwiss. Tehnol. 8: 286-8.

Fowlis, Ian A. 1998. Gas Chromatography Analytical Chemistry by Open Learning. John Wiley &
Sons Ltd: Chichester.

Franca, A. S., L. S. Oliveira., R. C. S. Oliveira., P. C. M. Agresti., dan R. Augusti. 2009. A

preliminary evaluation of the effect of processing temperature on coffee roasting degree
assessment. Journal of Food Engineering. 92: 345-352.

Illy, A dan R. Viani 1995. Espresso Coffee: The Chemistry of Quality Academic London: Press
Limited.

Janzen, S. Oestreich. 2010. Chemistry of Coffee. Hambrurg, Germany: Elsevier.

Kementerian Pertanian Republik Indonesia. 2014. Statistik Perkebunan Indonesia : Kopi 2013-
2015. Jakarta: Direktorat Jenderal Perkebunan.

Kumuzawa, Kenji. and Masuda, Hideki. 2003. Identification of Odor-Active 3-Mercapto-3-

methylbutyl Acetate in Volatile Fraction of Roasted Coffee Brew Isolated by Steam
Distillation under Reduce Pressure. J. Agr. Food Chem. 51. 3079-3082.

Lee, L. W., Cheong, M. W., Curan, P., Yu, B., dan Liu, S. Q. 2015. Coffee Fermentation and
Flavor- An intricate and delicate relationship. Food Chemistry. 185: 182-191.

Nursten, Harry. 2005. The Maillard Reaction: Chemistry, Biochemistry and Implications. UK: the
University of Reading.
Oestreich-Jenzen, S. 2013. Chemistry of Coffee. Germany: Elsevier Inc.

Pavia, D.L. 2006. Introduction to Organic Laboratory Techniques. United state : Brooks
Cole/Thomson
Rahardjo, Pudji. 2012. Panduan Budi Daya dan Pengolahan Kopi Arabika dan Robusta.
Jakarta: Penebar Swadaya.

Rizzi, G.P., Sanders, R.A. 1996. Mechanism of Pyridine Formation from Trigonelline Under
Coffee Roasting Condition. p. 206-210. Flavour Science: Recent Development. USA : The
Procter and Gamble Company.

Speer, Karl., Kolling, Speer, Isabelle. 2006. The Lipid Fraction of Coffee Bean. p. 201-216. Plant
Physiol. 18(1). Germany: Institute of Food Chemistry, Technische Universitat Dresden.

Sunarharum, W. B., Williams, D. J, dan Smyth H, E. 2014. Complexity of Coffee Flavor: A

Compositional and Sensory Perspective. Food Research. International. 60: 315-325.

Stoffelsma, J., G. Sipma., D. K. Kettenes, and J. Pypker. 1968. New volatile components of
rosted coffee. J. Agr. Food Chem. 16 (6): 1000-1004.

Shibamoto, T. 1983. Heterocyclic Compounds in Browning and Browning/Nitrite Model Systems:

Occurrence, Formation Mechanisms, Flavor Characteristics and Mutagenicactivity. New
York: Academic Press.

Taguchi, H., Sakaguchi, M., Shimabayashi, Y. 1985. Trigonelline Content of Coffee Beans and
the Thermal Conversion of Trigonelline into Nicotinic Acid During the Roastinh of Coffee
Beans. Agric. Biol. Chem. 49(12). 3467-3471.

Underwood, G.E. dan Deatherage, F.E. 1952. Nitrogen Compounds of Coffee. Ohio: The Ohio
State University.

Viani, R. & Horman, I. 1975. Determination of Trigonelline in Coffee. In: 7th International
Colloquium on the Chem. of Coffee. 273-8. Paris: ASIC.

Vitzthum, O.G, dan P. Werkhoff. 1976. Steam volatile constituen of roasted coffee: natural
fraction. 160: 277-291.

Winarno, F. G. 2004. Kimia Pangan dan Gizi. Jakarta: Gramedia Pustaka Utama.