Food Chemistry 130 (2012) 702–709

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Intra- and interspecific mineral composition variability of commercial instantcoffees and coffee substitutes: Contribution to mineral intake

Marta Oliveira a, Susana Casal b,⇑, Simone Morais a,⇑, Cláudia Alves b, Filipa Dias b, Sandra Ramos c,Eulália Mendes b, Cristina Delerue-Matos a, M. Beatriz P.P. Oliveira b

a REQUIMTE/Instituto Superior de Engenharia do Porto, Departamento de Engenharia Química, Rua Dr. António Bernardino de Almeida 431, 4200-472 Porto, Portugalb REQUIMTE/Laboratório de Bromatologia e Hidrologia, Faculdade de Farmácia da Universidade do Porto, Rua Aníbal Cunha 164, 4099-030 Porto, Portugalc Instituto Superior de Engenharia do Porto, Departamento de Matemática, Rua Dr. António Bernardino de Almeida 431, 4200-472 Porto, Portugal

a r t i c l e i n f o

Article history:Received 9 March 2011Received in revised form 14 June 2011Accepted 26 July 2011Available online 3 August 2011

Keywords:Soluble coffeesCoffee substitutesCoffee surrogatesMineralsHigh-resolution continuum source atomicabsorption spectrometry

0308-8146/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.foodchem.2011.07.113

⇑ Corresponding authors. Tel.: +351 2220789228340500x1918 (S. Morais); fax: +351 222003977 (159 (S. Morais).

E-mail addresses: sucasal@ff.up.pt (S. Casal), sbm@

a b s t r a c t

The present paper reports the amount and estimated daily mineral intake of nine elements (Ca, Mg, K, Na,P, Fe, Mn, Cr and Ni) in commercial instant coffees and coffee substitutes (n = 49). Elements were quan-tified by high-resolution continuum source flame (HR-CS-FAAS) and graphite furnace (HR-CS-GFAAS)atomic absorption spectrometry, while phosphorous was evaluated by a standard vanadomolybdophos-phoric acid colorimetric method.

Instant coffees and coffee substitutes are rich in K, Mg and P (>100 mg/100 g dw), contain Na, Ca and Fein moderate amounts (>1 mg/100 g), and trace levels of Cr and Ni. Among the samples analysed, plaininstant coffees are richer in minerals (p < 0.001), except for Na and Cr. Blends of coffee substitutes (barley,malt, chicory and rye) with coffee (20–66%) present intermediate amounts, while lower quantities arefound in substitutes without coffee, particularly in barley.

From a nutritional point of view the results indicate that the mean ingestion of two instant beveragesper day (total of 4 g instant powder), either with or without coffee, cannot be regarded as importantsources of minerals to the human diet, although providing a supplementation of some minerals, partic-ularly Mg and Mn in instant coffees. Additionally, and for authentication purposes, the correlationsobserved between some elements and the coffee percentage in the blends, with particular significancefor Mg amounts, provides a potential tool for the estimation of coffee in substitute blends.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Coffee has a relevant importance in human society for at least1200 years (Muriel & Arauz, 2010) and is amongst the most eco-nomically important agricultural products in international trade,particularly in developing countries, representing the main sourceof income for millions of people worldwide (Pendergrast, 2009).Coffee fills approximately 400 billion cups a year and is estimatedto be regularly consumed by more than 40 percent of the world’spopulation (Fitter & Kaplinsky, 2001). Its widespread consumptionis due primarily to the sensorial characteristics achieved, followedby a variety of factors of both social and economic nature. Althoughsome negative effects have been suggested over decades, includingaddiction, reproductive and cardiovascular adverse effects, in-

ll rights reserved.

02 (S. Casal), tel.: +351S. Casal), fax: +351 22 83 21

isep.ipp.pt (S. Morais).

creased susceptibility to certain cancers, etc. (Higdon & Frei,2006). Recent investigations increasingly suggest that regularcoffee consumption (about 2–4 cups per day) is associated withsubstantially lower mortality risk, colorectal cancer development,hepatic injury and cirrhosis, as well as degenerative, progressiveand chronic diseases (including Alzheimer’s and Parkinson’sdisease, type 2 diabetes and coronary heart disease) (Alves, Casal,& Oliveira, 2009; Debry, 1994; Higdon & Frei, 2006; Milon,Guidoux, & Antonioli, 1988, chap. 4; Nkondjock, 2009; Tan, Fook-Chong, Lum, Chai, Chung, et al., 2003).

Coffee beverages are usually prepared with arabica beans, plainor blended with robusta beans, the two most important commer-cial coffee species. Coffee is available to consumers as roastedgrain, ground or soluble/instant powder. Instant coffee is morepractical and ‘‘clean’’ to prepare, which is an important factor inconsumer preferences (Geel, Kinnear, & Kock, 2005). In the produc-tion of instant coffee, robusta coffee is preferred since, in additionto its economical attractiveness, the extraction yield of soluble sol-ids in the manufacturing process is superior to arabica coffee(Clarke & Macrae, 1987, chap. 6). According to the European Coffee

Table 1Identification and labelled composition of the samples under study.

Sample Composition (%)

Barley Chicory Malt Rye Coffee

Instant coffees1–8 – – – – 100

Mixtures with coffee9 – 50 – – 5010 – 50 – – 5011 – 60 – – 4012 – 60 – – 4013 – 60 – – 4014 – 66.7 – – 33.315 – 80 – – 2016 55 25 – – 2017 x x – – 2018 x x – – 2019 x x – – 2020 55 25 – – 2021 x x – – 2022 x x – – 2023 55 25 – – 2024 55 25 – – 2025 55 25 – – 20

Mixtures without coffee26 40 x 30 5 –27 35 25 35 5 –28 35 25 35 5 –29 35 25 35 5 –30 35 25 35 5 –31 35 25 35 5 –32 35 25 35 5 –33 x x x x –34 x x x x –35 x x x x –36 x x x x –37 x x – x –38 x x – x –

Chicory39 – 100 – – –

Barley40–49 100 – – – –

x: no detail given on the amounts.

M. Oliveira et al. / Food Chemistry 130 (2012) 702–709 703

Report (2009), soluble coffee exports and imports to and from non-European destinations represent currently around 39 and 41 thou-sand tons per year in Europe, respectively. Also according to thisreport, the soluble coffee consumption in Europe has been rela-tively stable in the last few years, being particularly important inEastern Europe and in the UK.

Coffee substitutes or surrogates are prepared from roasted veg-etables, being used to produce a product that, in the presence of hotwater, results in a drink similar to coffee. People look increasinglyfor these products due their inferior or absent caffeine contentand because they are cheaper than coffee. They can be commercia-lised in the form of roasted cereal or, more frequently, as solublepowders based on mixture of several cereals, often with some per-centage of soluble coffee. In Portugal, they are based mostly onroasted chicory and some cereals, namely barley, rye and theirmalts, thus contributing for energy and nutritional intake, with alarge amount of important bioactive substances, essential to theproper functioning of the body (Milon et al., 1988, chap. 4).

Food minerals play an important role in human nutrition. Onthe basis of the relative amounts in the human body and quantitiesneeded per day (more or less that 100 mg/day), essential mineralsare usually classified into macro and microelements (Nabrzyski,2007, chap. 5). These minerals are critical for the growth and for-mation of bones, synthesis of vitamins, enzymes and hormones,as well as for the healthy functioning of the nervous system, bloodcirculation and cellular integrity, if maintained at required levels(McDowell, 2003).

Several reports regarding the mineral composition of instantcoffees are found in literature. Gillies and Birkbeck (1983) esti-mated the daily mineral intake by ingestion of instant coffeesamples using atomic absorption spectrometry. In 1998, a marketsurvey was performed in the UK on the concentrations of metalsand other elements in selected snack and convenience foods,including instant coffee samples (MAFF, 1998). The levels of sev-enteen elements were determined in Brazilian soluble coffees byinductively coupled plasma atomic emission spectrometry (San-tos and de Oliveira, 2001) while Grembecka, Malinowska, andSzefer (2007) analysed fourteen elements in coffee samples mar-keted in Poland, including some instant coffee ones. Minor sur-veys were also performed in five brands of instant coffeepowder available in the Indian market by atomic absorptionspectrometry (Suseela, Bhalke, Kumar, Tripathi, & Sastry, 2001),in eight brands of soluble/instant coffee commercialised in Mex-ico and United States by neutron activation analysis (Vega-Carril-lo, Iskander, & Manzanares-Acuna, 2002), and two samples fromPakistan (Zaidi, Fatima, Arif, & Qureshi, 2006) evaluated by thesame methodology.

Although coffee substitutes should theoretically be also re-garded as good sources of some minerals, only one study (Suseelaet al., 2001) reported the mineral content for chicory-blended cof-fee (n = 4). Furthermore, only a limited number of studies werepublished concerning the application of a promising techniquevery recently developed, high-resolution continuum source atomicabsorption spectrometry (with flame – HR-CS-FAAS- or graphitefurnace detection – HR-CS-GFAAS), in food matrices, namely in yo-gurt, wine, beans, grain products and mineral waters (Welz et al.,2010).

The aim of this work was (i) to evaluate the content of someminerals in instant coffee-based and coffee substitutes samples,(ii) to estimate the daily mineral intake promoted by the consump-tion of theses beverages in order to study their nutritional signifi-cance, and (iii) to provide an analytical tool for the estimation ofcoffee in instant coffee substitute blends. This study also includedthe characterisation of the analytical performance of high-resolu-tion continuum source atomic absorption spectrometry when ap-plied to instant coffees and coffee substitutes.

2. Materials and methods

2.1. Reagents

A Milli-Q water purification system (Millipore, Molsheim,France) was used to obtain ultrapure water (18.2 MO cm�1 resis-tivity) for quantitative analysis. Standard solutions of Ca, Mg, Fe,Mn, Cr and Ni were prepared from the correspondent 1000 mg/lstock solutions (Panreac, Barcelona, Spain). K, Na and P standardsolutions were obtained by potassium chloride (99.5%, Riedel-deHaën, Seelze, Germany), sodium chloride (99.8%, Riedel-de Haën,Seelze, Germany) and potassium dihydrogen phosphate (99.5%,Riedel-de Haën, Seelze, Germany) dissolution in ultrapure water,respectively. Standards and samples were acidified with 1% (v/v)suprapur hydrochloric (for HR-CS-FAAS analyses, Sigma–Aldrich,Steinheim, Germany) or nitric acid (for HR-CS-GFAAS determina-tions, Sigma–Aldrich, Steinheim, Germany), except for P analysis.For quantification by HR-CS-FAAS, cesium chloride 1% (w/v; Sig-ma–Aldrich, Steinheim, Germany) was used as ionisation chemicalsuppressor. For Cr and Ni analyses by HR-CS-GFAAS, the appliedmatrix modifier was 0.05% (v/v) Mg(NO3)2�6H2O (p.a; Panreac, Bar-celona, Spain).

The colour development reagent for P determination was pre-pared as described in 4500-P standard (Greenberg, Clesceri, &Eaton, 1992) by the addition of ammonium molybdate tetrahy-drated (99.0%, Merck, Darmstadt, Germany) and ammonium

704 M. Oliveira et al. / Food Chemistry 130 (2012) 702–709

metavanadate (99.0%, Merck, Darmstadt, Germany). All glasswareand polyethylene material were soaked in nitric acid (10% v/v) atleast 24 h, thoroughly rinsed with ultrapure water and dried be-fore use.

2.2. Samples

A total of 49 commercial samples were collected from differentsupermarkets in Oporto metropolitan area (NW Portugal). In accor-dance with their labelled composition, samples were organised infive different groups: 100% coffee (n = 8), blends of coffee substi-tutes with coffee (n = 17), blends of coffee substitutes without cof-fee (n = 13), plain barley samples (n = 10) and one sample of plainchicory. Coffee percentage was labelled, as legally required, butonly some brands detailed the percentage of each substitute inthe blend (Table 1). These were based mostly on barley, malt chic-ory and rye.

2.3. Brews preparation

Samples were prepared as recommended by most manufactur-ers by dissolving 2.0 g of powder in hot ultrapure water. The finalvolume was adjusted to 30.0 ml. Samples were thoroughly mixed,and centrifuged for 30 min at 4000 rpm (Sartorius 2–16, Sigma) be-fore analysis. Appropriate dilutions were prepared with ultrapurewater.

2.4. Sample analysis

Ca, Mg, Na, K, Fe and Mn quantification were carried out usingan Analytik Jena ContrAA 700 High-Resolution Continuum SourceFlame Atomic Absorption Spectrometer equipped with a xenonshort-arc lamp XBO 301 (GLE, Berlin, Germany) with a nominalpower of 300 W operating in a hot-spot mode as a continuum radi-ation source. This recent equipment presents a compact high-res-olution double-Echelle grating monochromator correspondent toa spectral band with <2 pm per pixel in the far ultraviolet rangeand a charge coupled (CCD) array detector. Air-acetylene (Air Li-quid, Portugal) oxidising flame was used for atomization. Cr andNi were analysed by using the same equipment but using theGraphite Furnace module equipped with an MPE60 autosampler(Analytik Jena, Germany) and argon as inert gas. Transversal andpyrolitically coated graphite tubes with integrated platform wereused. In order to obtain maximum absorbance and minimum back-ground values, operational parameters were optimised and arepresented in Table 2. P measurements were performed in a dualbeam UV visible spectrophotometer (Evolution 300, Thermo Scien-

Table 2Optimised instrumental analytical parameters for determination of the selected elements bGFAAS) atomic absorption spectrometry.

HR-CS-FAAS Ca Mg N

Acetylene (L/h) 50 65 9Flow ratio C2H2/air 0.190 0.141 0Wavelength (nm) 422.6728 285.2125 5Height burner (mm) 5 8 6Pixels 5 5 5

HR-CS-GFAAS CrHeating program T �C (ramp time (s), hold time (s))Drying 1 80 (6, 28.2)Drying 2 90 (3, 23.3)Drying 3 110 (5, 14.0)Ashing 1100 (300, 13.3)Atomization 2400 (1500, 4.9)Cleaning 2450 (500, 4.1)

tific, USA) according to standard 4500-P at 420 nm (Greenberg,Clesceri, and Eaton, 1992).

External calibration curves were constructed daily based on, atleast, five standards. Analytical blanks and standards were ana-lysed daily and regularly along with samples to check instrumentperformance. All measurements were performed, at least, intriplicate.

2.5. Data analysis

Data were expressed as mean ± standard deviation and the sta-tistical significance was accessed using the SPSS (v.13, SPSS Inc.,USA) and Statistica software (v. 7, StatSoft Inc., USA). The signifi-cance of the differences between instant coffee and substituteswith and without coffee was tested by the one-way ANOVA fol-lowed by post-hoc Tukey HSD test. The assumption of normal dis-tribution of the amount of mineral in the different groups wasassessed by Kolmogorov–Smnirnov test with Liliefors correction.The Levene test was applied to study the homogeneity of the ana-lysed samples. Dependence of mineral concentration and the per-centage of coffee were examined using Pearson correlationcoefficient. Statistical significance was defined as p 6 0.05 (p-valueat 95% confidence level). Univariate analyses were carried out byconventional methods, as is the case of means, standard deviationsand coefficients of variation.

3. Results and discussion

3.1. Analysis by High-resolution continuum source atomic absorptionspectrometry

In order to obtain maximum absorbance and minimum back-ground values, operational HR-CS-FAAS and HR-CS-GFAAS param-eters were optimised (Table 2). The visibility of the spectralenvironment of the analytical line at high resolution greatly helpsin method development and to avoid spectral interferences. Theadvanced simultaneous background correction automatically elim-inates lamp flicker noise and continuous background absorption.Fine-structured background absorption can be removed using areference spectrum and a least-squares algorithm (Welz et al.,2010). An initial analytical screening was performed on some se-lected samples in order to evaluate the minerals present and theirapproximate amounts. Based on these results, nine elements wereselected, and the sample dilutions adjusted. The analytical charac-teristics i.e, calibration data, detection (LOD) and quantification(LOQ) limits were evaluated for each element and are presentedin Table 3. LOD and LOQ were calculated as the minimumdetectable amount of analyte with a signal-to-noise ratio of 3:1

y high-resolution continuum source flame (HR-CS-FAAS) and graphite furnace (HR-CS-

a K Fe Mn

0 60 50 60.196 0.243 0.130 0.13088.9953 766.4905 248.3270 279.4817

8 6 65 5 5

NiT �C (ramp time (s), hold time (s))80 (6, 29.2)90 (3, 23.3)110 (5, 14.0)1200 (300, 13.6)2200 (1500, 5.7)2450 (500, 4.5)

Table 3Calibration data obtained (n P 5) for the selected elements.

Element Calibration range (mg/l) Regression equationª R2 LOD (mg/l) LOQ (mg/l)

Ca 0.150–2.500 y = 0.070x + 0.002 0.9972 0.14 0.45Mg 0.080–1.100 y = 1.014x + 0.104 0.9983 0.07 0.22Na 0.350–6.200 y = 0.151x + 0.055 0.9976 0.26 0.85K 0.300–2.000 y = 0.193x � 0.038 0.9995 0.05 0.17P 0.350–8.000 y = 0.052x + 0.002 0.9990 0.32 1.05Fe 0.100–2.000 y = 0.068x + 0.001 0.9999 0.03 0.09Mn 0.060–1.300 y = 0.162x � 0.001 0.9991 0.05 0.17

Element Calibration range (lg/l) Regression equationª R2 LOD (lg/l) LOQ (lg/l)

Cr 2.50–10.00 y = 0.019x + 0.060 0.9997 0.26 0.86Ni 2.50–20.00 y = 0.008x + 0.016 0.9982 0.13 0.43

a y – absorbance; x – concentration (mg/l for HR-CS-FAAS or lg/l for HR-CS-GFAAS).

M. Oliveira et al. / Food Chemistry 130 (2012) 702–709 705

and 10:1, respectively, (Miller & Miller, 2000) and expressed asmineral concentration in beverages (mg/l and lg/l). LODs between0.13 lg/l for Ni and 0.26 mg/l for Na were obtained, withcorresponding LOQs in the range 0.43 lg/l for Ni and 0.85 mg/lfor Na.

Since no certified or standard reference material is available forsoluble coffee brew, the analytical procedures were validated bysystematic recovery experiments performed at three fortificationlevels using two different samples for each group (except for chic-ory, n = 1). Acceptable recovery values ranging from 91.1 ± 2.7 forFe to 109.1 ± 0.8% for Ca were reached. Global precision wasascertained accordingly with IUPAC recommendations by intra-and inter-assay studies (Thompson, Ellison, & Wood, 2002).Regarding the reproducibility of the optimised methodologies,expressed as relative standard deviation, values were lower than2.8% and 9.4%, respectively for intra- (each metal concentrationwas assessed twenty times during the day) and inter-day (twodifferent concentrations observed during five days) precision. Theseresults comply with the criterion of acceptability (RSD 6 10%;Thompson et al., 2002).

3.2. Mineral composition of instant coffees and surrogates

The results obtained for the major, minor and trace elements inthe instant coffees, coffee blends (coffee, chicory, malt and barley),mixtures without coffee (barley, rye, chicory and malt) and barleysamples are presented in Table 4. A detailed discussion on theirnutritional significance will be postponed to Section 3.3. The re-sults obtained from the one-way ANOVA analysis and from thepost-hoc Tukey’s HSD test for all the groups are also presented.Only one commercial plain chicory sample was found in the Oportometropolitan area (Portugal), although being of Spanish origin, en-abling its statistical mineral comparison with other type of sam-ples. Also, no plain instant malt or rye was commercially available.

The mineral content of the instant coffees and substitutes ishighly variable (Table 4). A great heterogeneity for the nine ele-ments between and among the various groups is observed. Thesevariations reflect differences in the industrial manufacturing pro-cess, including the water used, and/or factors that influence thecharacteristics of the raw materials, such as the type of soil, theuse of fertilizers with different chemical compositions and ambientconditions (Clarke & Macrae, 1987, chap. 6).

The assumption of normal distribution of the mineral amount inthe different groups was assessed by Kolmogorov–Smnirnov testwith Lilliefors correction and revealed, accordingly with Grembe-cka et al. (2007), that data should not be represented by a normaldistribution, given the limited number of analytical results. TheLevene’s test was applied to study the mean variances of each ele-ment and showed that there is no homogeneity among the samplesfor all the selected elements.

Aware of the high variability regarding instant coffee data, andknowing that these variations might be due to agricultural, techno-logical, as well as analytical differences, and in the absence of dataon Portuguese samples, a representative number of plain coffeesamples were included in order to have a comparison base for dis-cussion of the substitute mineral composition. Therefore, theresults present herein will be discussed in two groups: coffee first,with a detailed comparison with literature data, and then coffeesubstitutes, aiming to identify the major differences regardingtheir mineral content.

Detailing the major elements in instant coffees, the range of val-ues obtained for Ca (Table 4) is in the lower range of results pre-sented by Grembecka et al. (2007) (65.6–265.0 mg/100 g) for 27instant coffee samples acquired in Gdansk, Poland, and slightlylower than those reported by Santos et al. (2001) (106.0–189.0 mg/100 g) for 21 Brazilian samples, as in a UK market survey(MAFF, 1998) (96.7–129.0 mg/100 g), in Mexican/USA samples(Vega-Carrillo et al., 2002) (132.0–197.0 mg/100 g), and by Debry(1994) (150.0 mg/100 g). In opposition, the experimental valuesfor Mg in the instant coffee samples are slightly higher than thevalues reported by Santos et al. (2001) (212–415 mg/100 g),Grembecka et al. (2007) (84.8–451.0 mg/100 g), Vega-Carrilloet al. (2002) (3.2–45.8 mg/100 g) and Debry (1994) (160.0–310.0 mg/100 g). The high variability presented herein for Nacontent is in accordance with the values presented by Santoset al. (2001) (27.4–666.5 mg/100 g), Vega-Carrillo et al. (2002)(1.4–46.4 mg/100 g), Grembecka et al. (2007) (2.78–347.0 mg/100 g) and with the 50.0 mg/100 g mean reported by Debry(1994). Regarding K, the major element, similar levels are reportedby Debry (1994) (4000.0 mg/100 g), Vega-Carrillo et al. (2002)(135.0–3990.0 mg/100 g) and Santos et al. (2001) (3250.0–5170.0 mg/100 g), being slightly higher than those reported byGrembecka et al. (2007) (1871.0–3324.0 mg/100 g). The P contentin instant coffee results are comparable to the findings of Santoset al. (2001) (223.0–410.0 mg/100 g), Grembecka et al. (2007)(348.0–598.0 mg/100 g) and Debry (1994) with 350.0 mg/100 g.

The Fe and Mn values for soluble coffee samples are also inaccordance with the results obtained by Santos et al. (2001)(1.4–45.1 mg/100 g for Fe and 0.4–3.9 mg/100 g for Mn), by Grem-becka et al. (2007) (1.9–5.9 mg/100 g and 0.9–3.5 mg/100 g,respectively), in the UK market survey (MAFF, 1998) (2.4–4.1 mg/100 g and 1.6–2.7 mg/100 g) in Mexican/USA samples (Vega-Carrillo et al., 2002) (2.0–7.5 mg/100 g, Fe and 0.1–3.8 mg/100 g,Mn) and by Debry (1994) (4.5 mg/100 g Fe). The values obtainedfor Cr were in accordance with all the previous discussed studiesand were not increased in stainless steel bottled samples. Ni exper-imental values in instant coffee samples are higher than the resultsreported by MAFF (1998) (0.06–0.130 mg/100 g), and similar withthose presented by Santos et al. (2001) (<0.250 mg/100 g), andGrembecka et al. (2007) (0.03–0.34 mg/100 g), with one sample

Table 4Mineral content (mg/100 g) in commercial instant coffee and substitutes.

Sample Coffee (%) Ca Mg Na K P Fe Mn Cr Ni

Instant coffees1 100 70.1 ± 3.4 546.8 ± 4.9 43.2 ± 0.3 3948.5 ± 229.0 420.5 ± 8.4 7.11 ± 0.36 2.09 ± 0.01 0.013 ± 0.001 0.593 ± 0.012 100 65.4 ± 1.5 520.7 ± 4.7 46.4 ± 1.9 4647.4 ± 223.1 371.7 ± 2.9 6.16 ± 0.33 2.24 ± 0.03 0.004 ± 0.001 0.332 ± 0.013 100 75.6 ± 0.8 501.9 ± 4.0 289.4 ± 1.2 4535.6 ± 267.6 379.5 ± 2.8 2.44 ± 0.19 1.98 ± 0.02 ND 0.341 ± 0.014 100 81.0 ± 2.0 306.9 ± 0.6 29.6 ± 0.4 6149.6 ± 180.3 415.8 ± 0.6 3.00 ± 0.09 0.99 ± 0.04 0.002 ± 0.001 0.045 ± 0.015 100 88.6 ± 2.2 554.4 ± 3.3 25.9 ± 0.4 4454.3 ± 129.2 407.9 ± 4.7 4.90 ± 0.19 2.00 ± 0.02 0.003 ± 0.001 0.306 ± 0.016 100 75.4 ± 1.4 474.0 ± 2.8 34.1 ± 0.4 4350.9 ± 295.9 260.8 ± 2.3 10.81 ± 0.17 3.99 ± 0.22 0.002 ± 0.001 0.070 ± 0.017 100 90.9 ± 2.8 460.5 ± 1.4 83.9 ± 3.0 3716.6 ± 234.1 417.6 ± 5.4 4.70 ± 0.27 2.42 ± 0.05 0.005 ± 0.001 0.398 ± 0.018 100 96.5 ± 1.1 513.6 ± 1.5 126.2 ± 0.9 6062.3 ± 169.7 429.2 ± 6.5 1.85 ± 0.08 1.73 ± 0.02 0.005 ± 0.001 0.031 ± 0.01Mean 80.4 ± 10.8a 484.9 ± 78.8a 84.8 ± 89.4a 4733.2 ± 901.8a 387.9 ± 55.2a 5.12 ± 2.93a 2.18 ± 0.85a 0.005 ± 0.004a 0.264 ± 0.199a

Mixtures with coffee9 50 51.9 ± 1.7 258.3 ± 1.3 188.6 ± 3.0 3221.2 ± 67.6 291.7 ± 3.2 8.22 ± 0.37 1.60 ± 0.05 0.008 ± 0.001 0.081 ± 0.00310 50 88.5 ± 1.8 318.4 ± 1.6 213.1 ± 4.0 2900.4 ± 150.8 328.0 ± 4.5 6.31 ± 0.30 1.37 ± 0.02 0.017 ± 0.001 0.117 ± 0.00211 40 97.5 ± 2.0 376.0 ± 2.3 248.6 ± 4.0 3944.7 ± 55.2 314.8 ± 3.8 4.14 ± 0.24 1.83 ± 0.04 0.007 ± 0.001 0.115 ± 0.00112 40 55.6 ± 1.8 308.4 ± 3.7 313.2 ± 3.8 3324.8 ± 206.1 237.3 ± 1.2 3.76 ± 0.27 1.79 ± 0.03 0.007 ± 0.001 0.069 ± 0.00113 40 47.1 ± 0.7 317.0 ± 0.6 267.4 ± 2.7 3780.2 ± 102.1 313.3 ± 5.9 6.81 ± 0.20 1.68 ± 0.03 0.009 ± 0.001 0.071 ± 0.00214 33.3 113.8 ± 4.8 331.5 ± 1.0 301.0 ± 7.2 2672.0 ± 138.9 211.2 ± 2.9 7.87 ± 0.48 1.45 ± 0.01 0.025 ± 0.001 0.044 ± 0.00815 20 67.2 ± 0.1 215.6 ± 2.4 357.9 ± 2.1 2296.3 ± 41.3 140.2 ± 1.0 3.75 ± 0.07 1.90 ± 0.03 0.007 ± 0.001 0.137 ± 0.00316 20 51.7 ± 1.0 175.0 ± 0.2 429.0 ± 0.9 2749.8 ± 71.5 199.2 ± 1.2 2.61 ± 0.08 1.35 ± 0.01 0.007 ± 0.001 0.139 ± 0.00517 20 58.3 ± 0.4 159.9 ± 1.0 332.3 ± 3.3 3596.5 ± 194.2 163.8 ± 2.9 2.73 ± 0.09 1.39 ± 0.03 0.002 ± 0.001 0.088 ± 0.00318 20 52.3 ± 0.5 168.3 ± 1.7 248.0 ± 9.7 1733.2 ± 12.1 158.1 ± 2.0 1.37 ± 0.01 0.97 ± 0.01 0.004 ± 0.001 0.095 ± 0.00319 20 62.4 ± 0.6 190.0 ± 1.1 151.6 ± 5.5 2386.3 ± 100.2 159.0 ± 3.8 2.59 ± 0.13 0.98 ± 0.01 0.015 ± 0.001 0.103 ± 0.00120 20 66.3 ± 0.4 246.6 ± 0.7 147.8 ± 7.8 2481.1 ± 117.7 106,6 ± 1.3 3.27 ± 0.05 1.24 ± 0.01 0.011 ± 0.001 0.077 ± 0.00221 20 43.4 ± 1.1 195.0 ± 0.8 143.0 ± 9.9 4521.0 ± 171.8 172.6 ± 2.6 2.99 ± 0.13 1.00 ± 0.01 0.016 ± 0.001 0.051 ± 0.00222 20 43.9 ± 0.6 171.8 ± 1.9 149.8 ± 0.9 4371.6 ± 122.4 154.9 ± 0.6 2.79 ± 0.02 0.99 ± 0.02 0.016 ± 0.001 0.061 ± 0.00123 20 45.0 ± 0.1 234.1 ± 1.9 134.6 ± 1.3 2676.2 ± 131.1 132.0 ± 2.7 3.33 ± 0.05 1.16 ± 0.01 0.020 ± 0.001 0.046 ± 0.00824 20 57.7 ± 0.2 199.3 ± 1.0 160.9 ± 2.1 2223.8 ± 68.9 162.6 ± 2.8 3.20 ± 0.03 0.95 ± 0.01 0.026 ± 0.001 0.038 ± 0.00225 20 74.0 ± 1.0 171.4 ± 1.2 164.4 ± 0.8 4482.9 ± 130.0 163.2 ± 1.1 4.98 ± 0.06 0.92 ± 0.01 0.035 ± 0.001 0.057 ± 0.001Mean 63.3 ± 20.0a,b 237.5 ± 68.9b 232.4 ± 89.3b 3139.0 ± 855.50b 200.5 ± 70.5b 4.16 ± 1.99a,b 1.33 ± 0.34b 0.014 ± 0.009b 0.082 ± 0.032b

Mixtures without coffee26 – 48.0 ± 0.2 99.5 ± 1.2 158.0 ± 2.8 2184.3 ± 155.1 119.7 ± 1.1 1.31 ± 0.02 1.09 ± 0.02 0.001 ± 0.001 0.023 ± 0.00127 – 38.5 ± 0.7 105.5 ± 0.5 170.3 ± 0.3 1262.0 ± 36.6 128.4 ± 2.1 1.38 ± 0.01 1.00 ± 0.03 0.005 ± 0.001 0.020 ± 0.00128 – 37.5 ± 0.4 89.5 ± 0.6 134.2 ± 5.0 1570.5 ± 64.4 98.6 ± 1.1 2.44 ± 0.03 0.90 ± 0.02 0.008 ± 0.001 0.029 ± 0.00129 – 33.0 ± 0.6 83.3 ± 1.2 197.3 ± 1.0 1076.4 ± 39.8 96.0 ± 2.2 2.45 ± 0.03 1.05 ± 0.01 0.007 ± 0.001 0.037 ± 0.00230 – 29.3 ± 1.0 94.3 ± 0.5 175.1 ± 0.3 890.6 ± 8.9 111.0 ± 1.4 2.79 ± 0.08 1,09 ± 0.02 0.011 ± 0.001 0.036 ± 0.00331 – 30.1 ± 1.3 60.3 ± 0.3 56.8 ± 0.2 1221.7 ± 67.2 82.4 ± 1.4 3.86 ± 0.10 1.20 ± 0.02 0.005 ± 0.001 0.031 ± 0.00232 – 32.5 ± 0.7 108.1 ± 0.3 154.0 ± 0.1 1547.0 ± 71.2 119.8 ± 1.3 2.63 ± 0.03 0.90 ± 0.02 0.012 ± 0.001 0.009 ± 0.00133 – 72.1 ± 0.4 159.7 ± 1.0 163.0 ± 2.8 1450.1 ± 94.3 145.4 ± 1.0 3.61 ± 0.07 1.32 ± 0.02 0.009 ± 0.001 0.010 ± 0.00834 – 17.8 ± 0.7 94.4 ± 0.8 164.2 ± 6.2 1077.3 ± 33.4 73.4 ± 1.5 4.40 ± 0.06 1.53 ± 0.01 0.008 ± 0.001 0.028 ± 0.00135 – 58.4 ± 0.4 87.4 ± 0.3 121.5 ± 2.4 924.4 ± 16.6 129.4 ± 0.9 2.77 ± 0.03 0.90 ± 0.01 0.012 ± 0.001 0.033 ± 0.00236 – 56.0 ± 0.4 85.0 ± 1.0 139.4 ± 1.5 1277.0 ± 39.6 149.5 ± 2.9 3.14 ± 0.08 0.96 ± 0.03 0.010 ± 0.001 0.010 ± 0.00137 – 52.3 ± 0.5 107.7 ± 0.6 189.1 ± 4.5 3404.2 ± 173.6 137.9 ± 1.6 1.41 ± 0.01 1.28 ± 0.01 0.002 ± 0.001 0.028 ± 0.00138 – 33.7 ± 0.6 95.8 ± 0.4 520.7 ± 4.2 1770.5 ± 46.0 90.0 ± 1.5 1.40 ± 0.01 1.03 ± 0.01 0.001 ± 0.001 0.014 ± 0.002Mean 41.5 ± 14.9c 97.7 ± 22.5c 180.3 ± 108.2a,b,c 1512.0 ± 672.1c 114.0 ± 24.4c 2.58 ± 1.01b,c 1.10 ± 0.19b,c 0.007 ± 0.004b,c 0.024 ± 0.010a,c

Chicory39 – 46.0 ± 1.0 89.6 ± 0.3 162.3 ± 2.3 2724.7 ± 43.6 65.4 ± 0.9 9.05 ± 0.18 1.02 ± 0.01 0.020 ± 0.002 0.012 ± 0.001

706M

.Oliveira

etal./Food

Chemistry

130(2012)

702–709

Tabl

e4

(con

tinu

ed)

Sam

ple

Cof

fee

(%)

Ca

Mg

Na

KP

FeM

nC

rN

i

Barl

ey40

–25

.8±

0.8

88.4

±1.

354

.5±

0.6

889.

7±

26.7

144.

8±

2.6

1.79

±0.

040.

65±

0.01

ND

0.00

9±

0.00

141

–34

.1±

0.3

87.8

±0.

756

.3±

3.2

1546

.5±

52.6

89.5

±2.

10.

58±

0.02

0.80

±0.

010.

008

±0.

001

0.01

9±

0.00

142

–34

.3±

0.5

69.7

±0.

549

.5±

2.9

1169

.7±

15.2

142.

2±

0.4

0.94

±0.

020.

74±

0.04

0.01

0±

0.00

20.

022

±0.

001

43–

16.6

±0.

474

.4±

0.4

48.8

±1.

012

87.6

±56

.712

6.4

±1.

60.

72±

0.03

0.90

±0.

010.

005

±0.

001

0.00

8±

0.00

144

–18

.2±

0.4

66.4

±0.

473

.4±

2.2

890.

6±

7.1

118.

9±

2.3

1.56

±0.

040.

58±

0.01

0.00

6±

0.00

10.

021

±0.

001

45–

22.6

±0.

853

.1±

0.5

86.2

±0.

977

4.2

±20

.912

4.6

±0.

40.

74±

0.01

0.77

±0.

020.

011

±0.

001

0.02

0±

0.00

146

–26

.6±

0.6

117.

7±

0.9

59.6

±0.

619

42.4

±77

.712

7.1

±3.

20.

57±

0.01

0.76

±0.

020.

010

±0.

001

0.01

9±

0.00

147

–20

.0±

0.2

86.5

±0.

421

.2±

0.1

937.

0±

36.5

134.

1±

3.1

0.62

±0.

020.

77±

0.01

0.00

4±

0.00

10.

016

±0.

001

48–

37.8

±0.

281

.0±

0.2

51.0

±1.

463

4.9

±31

.113

1.0

±2.

60.

83±

0.01

1.01

±0.

040.

011

±0.

001

0.02

2±

0.00

349

–23

.1±

0,3

87.3

±1.

052

.9±

0.3

1918

.9±

51.8

68.6

±1.

20.

84±

0.01

0.81

±0.

010.

011

±0.

001

0.02

0±

0.00

2M

ean

(ran

ge)

25.9

±7.

3c,d

81.2

±17

.3c,

d55

.3±

16.9

a,d

1199

.1±

466.

5c,d

120.

7±

23.8

c,d

0.92

±0.

42c,

d0.

78±

0.12

c,d

0.00

8±

0.00

3b,c

,d0.

018

±0.

005a,

b,c

,d

Mea

nda

tafo

llow

edby

diff

eren

tle

tter

s(a

–d)

for

each

elem

ent

are

sign

ifica

ntl

ydi

ffer

ent

acco

rdin

gto

AN

OV

Aat

p<

0.00

1;N

D:

not

dete

cted

.

M. Oliveira et al. / Food Chemistry 130 (2012) 702–709 707

slightly above maximum legal limits (5.00 mg/kg) (Santos et al.,2001). The higher Ni amounts in five coffee samples were not asso-ciated with metal type packages.

In comparison with plain instant coffee, blends of substituteswith coffee (20–60%) presented similar contents of Ca and Cr,and significantly lower amounts of all the remaining minerals ex-cept for Na (Table 4). The mixtures without coffee follow the samepattern, with a significant reduction in Ca, Mg, K, P and Ni. Whenthe plain barley samples are observed, they have a statisticallysimilar content of all the minerals analysed (p < 0.001) with themixtures without coffee except for Na (not statistically differentfrom instant coffee samples). Although only one sample of instantchicory was available, apparently this substitute can be regarded asbeing richer than coffee for Na, Fe and Cr, being therefore interest-ing to extent this study to other chicory samples, particularlyregarding the nutritional importance of Na. On the opposite handto chicory, and taking the overall results, coffee is comparativelya higher source of Ca, Mg, K, P, Mn and Ni. Only one report wasfound for similar mixtures regarding four samples from the Indianmarket (Suseela et al., 2001), being our results in total agreement.For the other coffee surrogates groups, no comparison can be madesince no study regarding their mineral characterisation was found.Nevertheless, it can generally be concluded that all the mineralsare in smaller amounts than in coffee, exception made for Naand Cr (and Fe in the chicory sample). K is always the major min-eral compound and the barley samples presented the lowestamounts for all the minerals analysed, except for P, Cr and Ni.

These results highlight that coffee substitutes are generallypoorer in minerals than instant coffees. Also, the differences ob-served between groups have statistical significance, namely be-tween mixtures with coffee, mixtures without coffee and plainbarley samples (Table 4). Generally all sample can be characterisedby high amounts of K, Mg and P, but the content of each metal bybeverage group can be represented by K > Mg > P > Na > Ca > -Fe > Mn > Ni > Cr for instant coffee, with an inversion of Na and Porder for mixtures with coffee, K > Na > P > Mg > Ca > Fe > Mn > -Ni > Cr for mixtures without coffee, K > P > Mg > Na > Ca > -Fe > Mn > Ni > Cr for barley samples and K > Na > Mg > P > Ca >Fe > Mn > Cr > Ni for plain chicory sample.

3.3. Estimation of dietary mineral intake

Despite the differences observed in the mineral contents, partic-ularly between samples with and without coffee, a discussion onthe significance of the ingestion of these beverages to the dairymineral intake is relevant from a nutritional point of view. On thisbasis, the essential elements analysed should be divided into mac-roelements (Ca, Mg, P, K, Na), with requirements superior to100 mg/day, and micro or trace elements (Fe, Mn, Cr) (Nabrzyski,

Table 5Pearson’s correlation analysis of the mean concentrations of Ca, Mg, Na, K, Fe, Mn, P,Cr and Ni, and labelled coffee percentage (from 20% to 100%).

Element Coffee blends

Pearson’s correlation coeficient p

Ca 0.489 0.013*

Mg 0.913 <0.001*

Na �0.517 0.002*

K 0.664 <0.001*

P 0.333 0.104Fe 0.648 <0.001*

Mn 0.907 <0.001*

Cr 0.601 0.001*

Ni �0.490 0.015*

* Significant correlation at a 95% confidence level.

Table 6Estimated daily mineral intake (DMI, %) of each element by the consumption of two cups of instant coffees or surrogates (4 g).

DMI (%)a

Element RDA/AIb (mg/day) Instant coffees (n = 8) Mixtures with coffee (n = 17) Mixtures without coffee (n = 13) Barley (n = 10) Chicory (n = 1)

Ca 800 0.4 0.3 0.2 0.1 0.2Mg 375 5.2 2.5 1.0 0.9 1.0Na 1500 0.2 0.6 0.5 0.1 0.4K 2000 9.5 6.3 3.0 2.4 5.4P 700 2.2 1.1 0.7 0.8 0.4Fe 14 1.5 1.2 0.7 0.3 2.6Mn 2⁄ 4.4 2.7 2.2 1.6 2.0Cr 0.040 0.5 1.4 0.7 0.8 2.0Ni 0.3c 3.5 1.1 0.3 0.2 0.2

a Calculated as DMI (%) = C � 100/(RDA, AI or UL), where C is the element amount (in mg) present in 4 g of instant powder.b According to EEC (2008) and FNIC (2010).c According to EMEA (2008), with a permitted daily exposure of 300 lg/day.

708 M. Oliveira et al. / Food Chemistry 130 (2012) 702–709

2007). Recommended daily allowances (RDA) are regularly up-dated by health authorities for each element, on the basis on scien-tific evidences or, when inconsistent, adequate intakes (AI) areindicated, being this the actual case for K, Na, Ca, Mn, and Cr (FNIC,2010). Ni, although included in the list of recent essential minerals,should be observed with caution due to its toxicity, with recom-mendation regarding restricted tolerable upper intake level (UL).Indeed, despite being recognised as beneficial in trace amounts,Ni biochemical role and functions are not defined (EMEA, 2008;FNIC, 2010).

The daily mineral intake (DMI, %) of each element achieved bythe consumption of instant coffees or surrogates was estimated.Results are presented in Table 6. The following assumptions havebeen made: (i) an average of two cups of beverage are consumeddaily; (ii) two grammes of instant powder is used in the prepara-tion of one cup of coffee (from 30 to 100 ml), being equivalent toa full tea spoon per cup. DMI was calculated as DMI = C � 100/(RDA, AI or UL); where C is the element concentration (in mg) for4 g of instant powder.

The highest global mineral daily intake ranges are, by descend-ing order, 2.4–9.5% for K, 0.9–5.2% for Mg, 1.6–4.4% for Mn,0.2–3.5% for Ni, 0.3–2.6% for Fe, 0.4–2.2% for P, 0.5–2.0% for Cr,0.1–0.6% for Na and 0.1–0.4% for Ca. These figures are reducedfor all the elements, including Ni, although up to 5% can beachieved for Mg and Mn. Suseela et al. (2001) also reported thatthe contribution for total dietary intake from five Indian samples(one instant coffee and four chicory-blended coffee samples) wasless than or around 1% for Ca, K, Mg and Mn, and 3.0% for Cr,1.7% for Fe and 3.2% for Ni.

Instant coffees provide higher nutritional intakes of Ca, Mg, K, P,Mn and Ni, with values that range from 0.2% in Na to 9.7% in K.These values are in accordance with Gillies and Birkbeck (1983).Chicory contributes with the highest nutritional intake of Cr andFe, with 2.0% and 2.6%, respectively, aware that only one samplewas available. As previously observed for mineral content, instantcoffees contribute more to the mineral intake than coffee blends,followed by mixtures without coffee, and finally, barley samples.The only chicory sample analysed presents the highest Cr contents,followed by barley. Indeed, barley has been frequently mentionedin the literature regarding its beneficial chromium species, partic-ularly chromium picolinate, and potential health effect in severalpathologies, including diabetes (Volaufova et al., 2006). Exceptfor Cr, as mentioned, for Fe and P, barley samples present the low-est DMI contributions. Regarding Na, associated with several po-tential health effects, the values are also of no concern. Despitethe higher values provided by substitute blends the figures aregenerally low.

These results highlight that the ingested amounts are generallysmall, and no mineral can be labelled as significant in a single dose,

since none achieves 15% of the RDA, as recommended for labellingpurposes (EEC, 2008). If the minerals present in the water used forbeverage preparation are taken also into account, these figures willincrease slightly, but will still be reduced. Generally, all coffee sub-stitutes present lower mineral amounts then instant coffee bever-ages. In the particular case of instant coffee, the ingested amountscan be slightly higher than those obtained from the ingestion offresh coffee beverages. In fact, since instant coffee is prepared fromconcentrated extracts, extracted under pressure at higher temper-ature than those used to prepare fresh coffee brews, 2 g of solublepowder are equivalent to a higher amount of roasted coffee (4.4 g,ICO 2009).

It is also important to take into account the presence of otherbioactive compounds in the beverages that can influence thebioavailability of the minerals ingested in the beverage or pres-ent in the foods consumed simultaneously (Gilles & Birkbeck1983), as in the case of being consumed with milk or after ameal.

3.4. Mineral composition as a tool to estimate the amount of coffee

Despite being mandatory to declare the presence and coffeeamount in label, the possibility of fraud exists, particularly regard-ing true coffee amounts. Indeed, coffee is more expensive thanthose used as coffee surrogates, such as barley, malt, chicory andrye. Once these products are subjected to water extraction fol-lowed by technological processes that convert them into instantsoluble powders, all details on their cellular structure are lost(Clarke & Macrae, 1987, chap. 6). Therefore, the chemical composi-tion is frequently regarded as the last possibility for authenticationpurposes. Nevertheless, even caffeine, the classical coffee marker,is highly variable within coffee, being particularly dependent uponthe coffee species used, with robusta coffee having almost doubleamounts of caffeine than Arabica coffee. Indeed Robusta coffee isthe more frequent raw coffee for instant coffee production due toits higher solid soluble yield, higher caffeine amount and, certainly,lower price (Clarke & Macrae, 1987, chap. 6). Estimation of coffeeamounts based solely on the caffeine levels has a small precision.A prediction of coffee level based on the mineral amounts andcomposition, could complement this tool.

The dependence of mineral concentrations on the labelled cof-fee percentage was evaluated. Accordingly, diagrams of dispersionof results were built up and Pearson correlation coefficients weredetermined to evaluate the intensity of the linear combination ofthe two variables (mean detected concentration of each elementand percentage of coffee (20–100%) – Table 5. Results of linearregression analysis showed that significant positive relationshipsexist between all the elements and labelled coffee (%), except forNa. These results corroborate the previous discussions, with all

M. Oliveira et al. / Food Chemistry 130 (2012) 702–709 709

minerals increasing with the coffee amounts, except for Na and Ni,these two with negative correlations. The correlation value (r2)values are of particular significance for Mg and Mn (higher than0.8) Mg, being present in higher amounts, and being therefore eas-ier to analyse, can be of particular usefulness for the estimation ofcoffee amounts in cereal blends.

4. Conclusions

It was found that high-resolution continuum source atomicabsorption spectrometry (with flame or graphite furnace detection)is an appropriate technique for routine analyses of coffee-based andsubstitutes beverages. The most obvious advantage is the need foronly one single lamp for all elements and wavelengths, comparedto one lamp for each element, as it is necessary in line source atomicabsorption spectrometry.

Instant coffee substitutes, particularly barley, are poorer in min-erals (p < 0.001) than plain instant coffees. Therefore, the amountof coffee in the blend will contribute to the mineral amount, withMg providing a potential estimator for the presence of coffee in thesubstitute blends.

The contribution of a regular intake of soluble coffee and coffeesubstitutes for the dairy essential mineral requirements is reduced,particularly with coffee substitutes, being therefore important tobe complemented by a balanced diet, including other plant-derivedproducts, to achieve the adequate ingestion of essential minerals.

References

Alves, R. C., Casal, S., & Oliveira, B. (2009). Health benefits of coffee: Myth or reality.Química Nova, 32(8), 2169–2180.

Clarke, R. J., & Macrae, R. (1987). Instant Coffee. Coffee: Technology (vol. 2). London:Elsevier Applied Science.

Debry, G. (1994). Coffee and Health: Composition, Consumption and Effect upon Health.Paris: John Libbey Eurotext.

EEC (2008). Commission Directive 2008/100/EC of 28 October 2008, amendingCouncil Directive 90/496/EEC on nutrition labelling for foodstuffs as regardsrecommended daily allowances, energy conversion factors and definitions.Official Journal of the European Union, L 285/9.

EMEA (2008). Guideline on the specification limits for residue of metal catalysts ormetal reagents, Doc. Ref. EMEA/CHMP/SWP/4446/2000, p.22, EMEA - EuropeanMedicines Agency. London, UK.

European Coffee Report (2009). European Coffee Federation, The Netherlands.Fitter, R., & Kaplinsky, R. (2001). ‘‘Who gains from product rents as the coffee

market becomes more differentiated? A value chain analysis’’, IDS Bulletin. TheValue of Value Chains, 32(3), 69–82.

FNIC (2010). Dietary Reference Intakes (DRIs): Recommended Dietary Allowancesand Adequate Intakes, Elements; Food and Nutrition Board, Institute ofMedicine, National Academy of Sciences.

Geel, L., Kinnear, M., & Kock, H. L. (2005). Relating consumer preferences to sensoryattributes of instant coffee. Food Quality and Preference, 16, 237–244.

Gillies, M. E., & Birkbeck, J. A. (1983). Tea and coffee as sources of some minerals inthe New-Zealand diet. American Journal of Clinical Nutrition, 38(6), 936–942.

Grembecka, M., Malinowska, E., & Szefer, P. (2007). Differentiation of market coffeeand its infusions in view of their mineral composition. Science of the TotalEnvironment, 383(1–3), 59–69.

Greenberg, A. E., Clesceri, L. S. & Eaton, A. D. (1992). Standard Methods for theExamination Of Water and Wastewater (4500-P), USA.

Higdon, J. V., & Frei, B. (2006). Coffee and health: A review of recent humanresearch. Critical Reviews in Food Science and Nutrition, 46(2), 101–123.

MAFF (1998) Concentrations of metals and other elements in selected snack andconvenience foods, UK Ministry of Agriculture, Fisheries and Food (MAFF).

McDowell, L. R. (2003). General introduction, chapter 1. In L. R. McDowell (Ed.),Minerals in animal and human nutrition (2nd ed.). The Netherlands: ElsevierScience.

Milon, H., Guidoux, R., & Antonioli, J. A. (1988). Physiological effects of coffee and itscomponents. In R. J. Clarke & R. Macrae (Eds.). Coffee: Physiology (Vol. 3).London: Elsevier Applied Science.

Miller, J. N., & Miller, J. C. (2000). Statistics and Chemometrics for Analytical Chemistry.Edinburgh Gate, England: Pearson Education Limited.

Muriel, P., & Arauz, J. (2010). Coffee and liver diseases (review). Fitoterapia, 81,297–305.

Nabrzyski, M. (2007). Functional role of some minerals in foods. In J. O. Nriagu & P.Szefer (Eds.), Mineral Components in Foods. CRC Press.

Nkondjock, A. (2009). Coffee consumption and the risk of cancer: An overview.Cancer Letters, 277(2), 121–125.

Pendergrast, M. (2009). Coffee: Second to oil? Tea & Coffee Trade Journal (April),38–41.

Santos, E. J., & de Oliveira, E. (2001). Determination of mineral nutrients and toxicelements in Brazilian soluble coffee by ICP-AES. Journal of Food Composition andAnalysis, 14(5), 523–531.

Suseela, B., Bhalke, S., Kumar, A. V., Tripathi, R. M., & Sastry, V. N. (2001). Dailyintake of trace metals through coffee consumption in India. Food Additives andContaminants, 18(2), 115–120.

Tan, E. K., Tan, C., Fook-Chong, S. M. C., Lum, S. Y., Chai, A., Chung, H., et al. (2003).Dose-dependent protective effect of coffee, tea, and smoking in Parkinson’sdisease: A study in ethnic Chinese. Journal of the Neurological Sciences, 216(1),163–167.

Thompson, M., Ellison, S. L. R., & Wood, R. (2002). Harmonized guidelines for singlelaboratory validation of methods of analysis (iupac technical report). Pure andApplied Chemistry, 74(5), 835–855.

Vega-Carrillo, H. R., Iskander, F. Y., & Manzanares-Acuna, E. (2002). Elementalcontent in ground and soluble/instant coffee. Journal of Radioanalytical andNuclear Chemistry, 252(1), 75–80.

Volaufova, J., Matthews, D. E., Cefalu, W. T., Martin, J., Wang, Z. Q., Zhang, X. H., et al.(2006). Chromium picolinate supplementation attenuates body weight gain andincreases insulin sensitivity in subjects with type 2 diabetes. Diabetes Care, 29,1826–1832.

Welz, B., Vale, M. G. R., Florek, S., Okruss, M., Huang, M.-D., & Becker-Ross, H.(2010). High-resolution continuum source atomic absorption spectrometry –theory and applications. In Encyclopedia of Analytical Chemistry. John Wiley &Sons Ltd..

Zaidi, J. H., Fatima, I., Arif, M., & Qureshi, I. H. (2006). Determination of traceelements in coffee beans and instant coffee of various origins by INAA. Journal ofRadioanalytical and Nuclear Chemistry, 267(1), 109–112.