Stable isotope and multielement composition of coffee beans from different geographic origin

P.10

113

Stable isotope and multielement composition of coffee beans from different

geographic origin

Alessandro Santato, Daniela Bertoldi, Matteo Perini, Federica Camin, Andreas Rossmann,

Roberto Larcher

1

IASMA Fondazione Edmund Mach, Via Mach 1, 38010 San Michele all’Adige (Tn), Italy

Summary: This work combines the mineral and stable isotope profile to determine the geographic

origin of 62 green coffee beans collected in 4 coffee growing region. The Canonical Discriminant analysis permits to achieve a good separation between samples of South and Central America, Africa and Asia with a 100% correct reclassification.

Keywords: coffee, IRMS, ICP-MS.

1 Introduction

Coffee is one of the world’s most popular beverage and its market is one of the biggest worldwide. Depending on quality, taste and geographical origin the price of coffee beans, can range from 2 up to 77 $/kilogramme [1]. Fraud, where cheap coffee is sold as the more expensive one, is an important issue that

concerns the coffee industry. In this

framework, it is important to define an analytical/chemometric methods in order to

decipher the geographical origin of coffee beans. Several authors have been trying to establish the provenance of coffee, either using their elemental concentration [1, 2] or using their isotope ratios [3, 4].

In this work we aim to couple the mineral profile (ICP-MS analysis) and the stable isotope ratios (IRMS analysis) of 62 green coffee beans endeavouring to find a tool for the determination of their geographical origin.

2 Experimental

Sixty-two samples of green coffee of known origin coming from the principal worldwide coffee-growing areas were collected. In detail, 17 samples were from Central America (Mexico, 5; Guatemala, 4; Honduras, 2; El Salvador, 2; Costa Rica, 2; Dominican Republic, 2); 14 from South America (Colombia, 4; Brazil, 7; Uruguay, 3); 14 from Africa (Côte d'Ivoire, 2; Cameroon, 2; Congo, 2; Central African Republic, 3; Ethiopia, 5) and 17 from Asia (India, 9; Indonesia, 8). The major part of the coffee bean were of the more valuable and appreciate arabica variety whereas 2 Central American, 1 South American, 7 African and 10 Asian samples were from the robusta variety.

The ratios 13C/12C, 15N/14N and 34S/32S were directly measured on grinded coffee beans using an IRMS (DELTA V, Thermo Scientific, Germany) following total combustion in an elemental analyser (EA Flash 1112, Thermo Scientific or Vario EL III, Elementar

Analysensysteme GmbH). The ratios 2H/1_H

and 18O/16O were measured using an IRMS

(Finnigan DELTA XP, Thermo Scientific)

coupled with a Pyrolyser (Finnigan TC/EA, Thermo Scientific).

The isotope values were expressed in !‰ against international standards: Vienna-Pee

Dee Belemnite (V-PDB) for !13C, Air for !15N,

Vienna–Standard Mean Ocean Water

(V-SMOW) for !2H and !18O, Vienna-Canyon

Diablo Triolite (V-CDT) for !34_S.

The coffee seeds were grinded and mineralized

with HNO3 in closed vessels using a

microwave digester (MarsXpress, CEM, USA; max temperature 200°C). Analysis of Li, Be, B, Na, Mg, Al, P, K, Ca, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Rb, Sr, Y, Mo, Pd, Ag, Cd, Sn, Sb, Te, Cs, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Dy, Er, Tm, Yb, Ir, Pt, Au, Hg, Re, Tl, Pb, Bi and U was carried out using an

ICP-MS (Agilent 7500ce Agilent

Technologies, Tokyo, Japan) equipped with an Octopole Reaction System for the removal of principal polyatomic interferences.

Data analysis (Canonical Discriminant

analysis) was performed with Statistica 8.0 (StatSoft Inc., USA) on standardized data.

114

3 Results

The !13C values ranged form 29.4‰ to

-25.3‰; !15N values from 0.6‰ to 7.1‰, !18O

values from 21.1‰ to 37.2‰, !2H from -83‰

to -20‰ and !34S from 2.6‰ to 12.8‰. As,

Ge, Tm, Au, Tl and U were quantifiable in less than 50% of the samples whereas Pd and Sn in less than 10% of the samples. Te (<0.135 µg/kg dry weight), Ir (<0.622 µg/kg dw), Re (<0.055 µg/kg dw), Bi (<2.43 µg/kg dw) and Hg (<19.0 µg/kg dw) were always below the method’s detection limit. Regarding the element present in at least an half of the

samples: P, K, Mg and Ca median

concentration was > 1 g/kg dw; Rb, Fe, Mn, Cu, B, Al, Na, Sr, Ba and Zn ranged between 1000 and 1 mg/kg dw; Ti, Ni, Mo, Co, Cr, Cs, Se, Pb, Ce, Li, la, Cd, Nd, V, Y, Ag and Ga ranged between 1000 and 1 µg/kg dw whereas Pr, Gd, Sm, Dy, Sb, Eu, Er, Yb, Be, and Pt were below 1 µg/kg dw. These orders of magnitude were generally in agreement with the literature [3, 5, 6].

The coffee beans of Central America were

characterized by low values of !15N, !18O

and

!2_{H, highcontent of Sr and Ba and low level}

of Co, Mo and rare earth elements. The South

American samples showed high values of !13C,

Mn and Co and low content of Mo. Coffee beans growth in Africa presented high values

of !15N,!2_{H and K and low level of Co and}

Mn. Finally, the Asian samples had high content of Co, K, Li and Na and low level of

d15N, d18O, Mg and Mn. Manganese in

particular, but also Al, Co, Cs, Na and Rb were considered suited as origin indicators by different authors [1, 5].

To assess the discrimination efficiency for coffee beans origin, a Canonical Discriminant

analysis was performed using d13C, d15N, d18O,

d2H, Li, B, Na, Mg, P, K, Mn, Co, Ga, Se, Sr,

Mo, Cd, Sb, Cs, Ba, La, Ce and U data (Fig. 1). The combination of the first 2 canonical variables accounted for 88% of variability assuring a good discrimination of the samples of the 4 growth regions with a correct re-classification of 100% of the samples.

Fig. 1. Canonical Discriminant analysis of the isotopic and elemental composition of coffee beans from the 4 growing regions: scatterplot of the first two canonical variables.

4 Conclusions

This research shows that isotope and mineral composition can specifically characterize green coffee beans from the world’s largest producing areas, proving that a combined use of this parameters can effectively trace their

geographic origin, at least at continental level. The limited availability of genuine samples per each country did not permit the use of statistical analysis to further differentiate coffees at a more subtle level.

References

[1] K.A. Anderson, B.W. Smith; Journal of Agricultural and Food Chemistry, 50 (2002), pp 2068-2075. [2] A.P. Fernandes, M.C. Santos, S.G. Lemos, M.M.C. Ferreira, A.R.A. Nogueira, J.A. Nóbrega;

Spectrochimica Acta Part B, 60 (2005), pp 717-724.

[3] C.I. Rodrigues, R. Maia, M. Miranda, M. Ribeirinho, J.M.F. Nogueira, C. Máguas; Journal of Food

Composition and Analysis, 22 (2009), pp 463-471.

[4] F. Serra, C. G. Guillou, F. Reniero, L. Ballarin, M.I. Cantagallo, M. Wieser, S.S. Iyer, Károly Héberger, F. Vanhaecke; Rapid Communications in Mass Spectrometry, 19 (2005), pp 2111-2115.

[5] V. Krivan, P. Barth, A. Feria Morales; Mikrochimica Acta, 110 (1993), pp 217-236. [6] M.J. Martin, F. Pablos, A.G. Gonzalez; Analytica Chimica Acta, 358 (1998), pp 177-183.

s

t

l

u

s

e

R

3

e h T !13_{C v} rm fo ed g ran es u al v -29.4‰ o t ‰ - d15N, d18O, b l i art an g an M . Mn d an Mg , Rb d N Cs Co l A al t n i e es an Rb ‰; 3 . 5 2 !15_N m fro es u al v to -20‰ an u A m, T , e G f o % 0 5 an h t s es l than 1 µg/kg dry w 5 5 0 . 0 (< µg/ 0 . 9 1 < ( g H ’s d h met ‰, 1 . 7 o t ‰ 6 . 0 m fro es u al v N ‰, 2 . 7 3 o t ‰ 1 . 1 2 m !2_Hfrom d an !34_{S from 2.6‰ to 12.8} e l ab fi i t an u q e er w U d an l T , u an Pd reas e h w es l mp a s e h t f ( e T . es l amp s e h t f o % 0 1 2 2 6 . 0 < ( Ir ), t h g ei w µg/kg d 3 4 . 2 < ( Bi , ) w d g k / µg/kgd µg/kg dw) were always bel i rd Reg t mi i l i ect et d ‰, !18O m -83‰ . ‰ As, s es l n i e n i Sn d an 5 3 1 . 0 < ( Re , ) w d d an ) w d e h t w o el h t u b ar, l cu i art p ered d i s n co t au t fferen i d s es s as o T s ean b ffee co as w s i s y al an d2H,Li,B,N Sb , Cd , Mo ). 1 The comb ac es l ab ari v Rb d an a N , Cs , Co , l A o s al t u o cat i d n i n i g ri o as ed t i u s ]. 5 , [1 rs o h t en ci effi n o i at n mi cri s i d e h t cri s i D cal i n o an C a , n i g ri o s g n i s u ed rm fo er p as d13C, d15N , a G , Co , Mn , K P, , Mg a, N at d U d an e C a, L a, B , Cs , can 2 t rs fi e h t f o n o i at n i mb % 8 8 r fo ed t n u co ac of vari re e w Rb y b rs o r fo cy en t an n mi cri , N d18O, , Sr e, S . g Fi ( a at cal i n o can y t i l i ab ari ’s d o h met es r p t emen el : es l amp s P o i rat t cen n co N , l A B, , Cu 1 d an 0 0 0 1 Ce, , Pb Se, w et b ed g ran m, S , d G Pr, w o el b re e w n i rd a Reg . t mi i l n o i ect et d f al h an t eas l at n i t en es a C d an Mg , K , P Fe , Rb ; w d g k / g 1 > as w n o b ed g ran n Z d an Ba Sr, a, N , Co , Mo , i N , i T ; w d g k / mg g A , Y , V , d N , Cd a, l , i L 1 d an 0 0 0 1 en e w µg/kgdww D m, y, Sb, Eu, Er, Yb, Be, 1 w µg/kg dw. These ord e h t g n e h t f o an i med , Mn , Fe een w et b , Cs Cr, a G d an eas er h w Pt d an f o ers rd g a g n ri u s as ro g 4 e h t f o n o i cat fi i s as cl s e h t f o n o i at n mi cri s i d d o o g rr co a h t i w s n o i reg h t w ro a s e h t f o % 0 0 1 f o n es l amp s e r ect rr -. es l mp a w e d u t i n mag re u erat t i l e h t e ffe co e h T zed eri ct ara ch !2_{H, highco} an Mo , Co f o amp s an c meri A Co d an Mn t w ro g s ean b f !15N, !2H µg g en reem ag n i y l al er en g ere w . ] 6 , 5 , [3 re ca ri me A al r t Cen f o s ean b f o es u al v w o l y b zed !15_{N !,} 18 Ba d an Sr f o t en t n co and lo h T . s t men e el h rt ea rare d an o es u al v h g i h ed w o h s es l amp . Mo f o t en t n co w o l d an Co h g i h ed t en es r p a c fri A n i h t f el l l d K d H h t i w t en re e w ca 18 O

a

nd

el ev l w o h t u So e f o !13C, ee fffe Co es u al v h d Co . g Fi 1. Can d n a c i p o op t o s i r g 4 e h t m o r ffr a c i n o n a c o w ttw s i s y lly a n a t n a n i m i r c s i D l a c i n o ff o of c f o of n o i t i s o p m mp o c l a t n e m e l e fffee f o of t o l p rrp e t t a c s : s n o i g e r g n i w o r . s e l b a i r a v l e h t f o of s n a e b e e t s r i ffi e h t f o !15N, !2H y l al n Fi . Mn Co f o t en t n co

i

s

u

l

c

on

C

4

arch e res s i h T n o i t i s o mp co ean b ffee co areas g n ci u d ro p aram p s i h t f o

f

R

f o el ev l w o l d an K d an H s an i s A e h t , y amples had l w o l d an a N d an i L , K , Co

s

on

i

d an e p o t o s i at h t s w o h s arch z eri aract ch y l cal fi eci p s can n ’s d rl o w e h t m fro s ean v ro p , areas ing thata combin rac t y el v i fect efff an c ers et aram d an Co h g i h ad f o el ev l eral n mi reen g e z t es arg l e s u ed n r ei h t e rac o c i h rap g eo g ed t mi i l e h T t n u co each an cal i t s i at t s coffeesata mo al t en n i t n co at t eas l at , n i g ri o amp s e n i u en g f o y t i l i ab l ai av e h t t mi er p t o n d i d ry er fffe i d er h rt fu o t s i s y al an . el ev l e l t b u s re mo . el ev l al es l amp per f o e s u e at i t eren

s

e

c

n

e

r

e

ffe

e

R

] 1 [ K.A.A ] 2 [ A.P. r t c e p S Sp ] 3 [ C.I. R p m mp o C ] 4 [ F.Ser n a V . F ] 5 [ V.Kri ] 6 [ M.J.M

s

; h t i m S . W . B , n o s r e d n A JJoourna . S , s o t n a S . C . M , s e d n a n r e F B t r Pa a t Ac a c i m i h c o r , 60(200 n a r i M . M , a i a M . R , s e u g i r d o R s i s y lly a An d n a n o i t i s o ,22(2009) , o r e i n e R . F , u o l l i u G . G . C , a r r ; e k c e a h n RaappidCommunication e l a r o M a i r e F . A , h t r a B . P , n a v i e l a z n o G . G . A , s o l b a P . F , n i t r a M ry e h C d o Fo d n a l a r u t l u c i r Ag f o of l a r i e r r e F . C . M . M , s o m Le . G 7 1 7 p p , ) 5 0 -724. o N . F . M . J , o h n i r i e b i R . M , a d n 3 6 4 p p , ) -471. M , o l l a g a t n a C . I . M , n i r a l l a B . L , y r t e m o r t c e p S Sp s s a M Ma n i s n 19(2 ; s MiMikkrrochimicaActa,110(199 ; z e AnallyyticaChimicaActa,35 y rry t s i m e ,50(2002),pp2068-2075. N . A . J , a r i e u g o N . A . R . A , a s a u g á M . C , a r i e u g o ; JJoournalo , r e y I . S . S , r e s e i W . M KárolyH 1 1 1 2 p p , ) 5 0 0 -2115. 7 1 2 p p , ) 3 9 -236. , ) 8 9 9 1 ( 8 pp177-183. 2075. ; a g e r b ó N d o Fo f o of , r e g r e b é