Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Analysis of milk by FT-Raman spectroscopy Sylwester Mazurek a,n, Roman Szostak a,n, Tomasz Czaja a, Andrzej Zachwieja b a b
Department of Chemistry, University of Wrocław, 14F. Joliot-Curie, 50-383 Wrocław, Poland Faculty of Biology and Animal Sciences, Wrocław University of Environmental and Life Sciences, 38C Chełmońskiego, 50-630 Wrocław, Poland
art ic l e i nf o
a b s t r a c t
Article history: Received 29 October 2014 Received in revised form 9 March 2015 Accepted 15 March 2015
Fat, protein, carbohydrates and dry matter were quantified in commercial bovine milk samples, with the relative standard errors of prediction (RSEP) in the 3.4–6.1% range, using the partial least squares (PLS) method based on Raman spectra of liquid milk samples. Results of a better quality were obtained from a PLS model derived from IR spectra registered using single reflection ATR diamond accessory, which yielded RSEP values of 2.4–4.4%. The data indicated IR single reflection ATR spectroscopy and Raman spectroscopy in combination with multivariate modelling using the PLS method, allowed for the reliable, simultaneous quantitative determination of macronutrients in milk. The low signal to noise ratio of Raman spectra affects the quality of fat quantification especially for strongly defatted milk samples. & 2015 Elsevier B.V. All rights reserved.
Keywords: Milk analysis Raman spectroscopy ATR Quantitative analysis Multivariate calibration
1. Introduction Milk is one of the most important dietary products which contains nearly all the nutrients necessary to sustain life [1]. Being a complex natural fluid in which water is a major constituent, milk contains varying quantities of lipids, proteins and carbohydrates as well as smaller amounts of minerals and other fat-soluble or watersoluble components [2]. A number of analytical methods have been developed to determine the chemical composition of milk, providing a detailed analysis of fat, protein, carbohydrates and dry matter which is required for quality control purposes and consumer information. The traditional methods used for milk, i.e. Röse-Gottlieb for fat determination, Kjeldahl for proteins and polarimetry or gravimetry for lactose quantification, have been replaced by modern techniques that could significantly reduce cost and time needed for milk analysis [3,4]. Considering the global production scale of milk and milk-derived products, simplification of analytical procedures would be extremely advantageous. Fifty years ago, infrared (IR) spectroscopy was found to be a valuable technique allowing for fast and accurate determination of the chemical composition of milk and several IR techniques utilising both mid- (MIR) and near infrared (NIR) spectral ranges were used [5–9]. Infrared spectroscopy was adopted for use in commercially-available milk analysers and has become a standard tool for the quantification of milk components [3,5]. n
Corresponding authors. Tel.: þ 48 71 3757 238; fax: þ 48 71 3757 420. E-mail addresses:
[email protected] (S. Mazurek),
[email protected] (R. Szostak).
Another vibrational spectroscopy technique, Raman spectroscopy, is suitable for the analysis of solid and liquid samples, providing qualitative and quantitative data [10–12]. It is a non-destructive analytical method, where studied objects require no special preparation and the spectra are recorded for samples in their native state, significantly simplifying analyses. An important advantage to this method is that Raman data can be collected for substances placed in glass and polymer packaging [13], and may have potential applications in the milk industry. Raman spectroscopy has been used for quality control and quantitative analysis of powdered milk constituents [14,15], and to screen samples adulterated with whey [16,17] and melamine [18,19]. In the case of liquid milk samples, Raman spectroscopy in combination with gel filtration was utilised to detect whey proteins in milk [20]. However, the widespread application of Raman analysis to liquid milk samples is hampered by the low intensity of its spectra, which are dominated by water bands. The spectral contributions of the remaining constituents in milk, accounting for little more than a few percent of the liquid mass, are weak. Raman spectra of liquid milk samples are usually characterised by a poor signal to noise (S/N) ratio which makes traditional univariate analyses difficult to perform. To overcome these limitations and extract relevant information from spectral data, chemometric methods can be utilised [21]. The use of multivariate data analysis techniques often yield robust calibration models, even for noisy systems poorly modelled by univariate techniques [22,23]. Raman spectroscopy in conjunction with PLS modelling was applied to quantify fat in liquid milk samples [24].
http://dx.doi.org/10.1016/j.talanta.2015.03.024 0039-9140/& 2015 Elsevier B.V. All rights reserved.
Please cite this article as: S. Mazurek, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.03.024i
S. Mazurek et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
This study used FT-Raman spectroscopy to simultaneously quantify fat, protein, carbohydrates and dry matter, the main components of milk. These results were compared to data obtained using a single reflection ATR accessory.
2. Materials and methods 2.1. Sample preparation Samples (75) were prepared by mixing various commerciallyavailable milk, including evaporated, containing 0%, 0.5%, 1.5%, 2.0%, 3.2% and 7.5% of fat. When necessary, small quantities of water were also added. Concentrations of 0.5–3.9%, 2.2–4.9% and 3.4–7.0% were obtained for fat, protein and carbohydrates, respectively (a detailed composition of the samples, determined using the reference IR method, is presented in Table S1 of supplementary material). Each sample was divided into 3 portions and quantified by three independent analytical methods i.e. Raman, ATR and reference IR. Spectroscopic measurements were performed without any physical or chemical pretreatment of samples, except for maintaining the samples at 40 °C. 2.2. FT-Raman measurements A Nicolet Magna 860 FT-IR spectrometer (Thermo Nicolet, Madison, WI, USA) interfaced with a FT-Raman accessory equipped with CaF2 beamsplitter and indium–gallium–arsenide (InGaAs) detector was used to carry out the measurements. Milk samples, placed in a typical NMR glass tube, were illuminated by a 1.064 mm Nd:YVO4 laser with the power of 400 mW at the sample, and backscattered radiation was collected. The interferograms were averaged over 512 scans, Happ-Genzel apodized and Fourier transformed using a zerofilling factor of 2 to yield spectra in the 100–3700 cm 1 range at a resolution of 8 and 16 cm 1. The interferograms were averaged over 768 scans for spectra recorded with a resolution of 32 cm 1. Under such conditions, each spectrum required approximately 6–10 min to complete. 2.3. FTIR ATR measurements ATR spectra were recorded using a single reflection Golden Gate (Specac, Slough, UK) diamond accessory mounted in the Nicolet iS50 FTIR spectrometer (Thermo Fisher Scientific, Madison, WI, USA). A KBr beamsplitter and DTGS detector were used for infrared measurements. The spectra of milk samples in the 400–4000 cm 1 range with a resolution of 4 cm 1 were obtained (128 scans). Similar to Raman experiments, Happ-Genzel apodization and Fourier transformation of interferograms using a zero filling factor of 2 were applied to obtain IR spectra. Under such conditions, it took approximately 2–3 min to register the spectrum. ATR measurements were performed in a similar manner to those described by Iñón et al. [4], though a single reflection diamond accessory was used instead of a 6-reflection ZnSe accessory. 2.4. Reference analysis Comparative analyses of the chemical composition of milk were performed with the B-150 IR milk analyser (Bentley Instruments, Chaska, MI, USA). Fat, protein, carbohydrates and dry matter content was determined according to the ISO 9622 procedure [25]. 2.5. Software and numerical data treatment
models and to perform the quantitative analysis of macronutrients and dry matter in commercial cow milk samples. Principal component analysis (PCA) was performed using PLS Toolbox (ver. 6.2, Eigenvector Research, Wenatchee, WA, USA) in Matlab (ver. 7.0.4, MathWorks, Natwick, MA, USA) environment. Fast Fourier transform (FFT) filter was applied to de-noise Raman spectra recorded with the resolution of 8 cm 1 (an example of the original and smoothed spectrum is shown in Fig. S1 of supplementary material). Spectral data were mean-centred and normalised using the MVN algorithm [26]. To characterise and compare the predictive abilities of calibration models, the relative standard errors of prediction (RSEP) were calculated according to the following formula:
RSEP(%) =
∑in=1 (Ci−CiA )2 2
∑in=1 CiA
× 100 (1)
A
where C is the component concentration determined by the reference method, C is the concentration found from the model and n is the number of samples. The RSEPcal and RSEPval errors were calculated for the calibration and validation data sets, respectively. The RSEPtest error for each chemical component was computed for the commercial milk samples. Cross-validation using the leave-4-out technique was performed to estimate the performance of the models, and the root-mean-square error of crossvalidation (RMSECV) was calculated to select an optimal number of factors for the PLS models [21].
3. Results and discussion The chemical composition of milk can vary considerably depending on several factors including species, breed, individuals, stage of lactation, age, health status and feeding regimen [1]. Raw bovine milk usually contains 3–5% of fat, 3.5–4% protein, 4.5–5% lactose and 12% of dry matter, while commercial milk samples were characterised by a lower level of fat due to industrial recovery of this component. Twenty one local milk samples containing 1.5–3.2% of fat and typical amounts of protein and carbohydrates were selected for analysis. 3.1. FT-Raman analysis Representative Raman milk spectra are shown in Fig. 1. These spectra were dominated by water scattering, while the remaining milk constituents were responsible for a number of low intensity
6
Raman intensity
2
4
f e d
2
c b a
0 3500
3000
2500
2000
1500
1000
500
-1
Wavenumber [cm ]
Nicolet TQ Analyst Version 7 (Thermo Fisher Scientific, Madison, WI, USA) chemometric software was used to construct the PLS
Fig. 1. FT Raman spectra of commercial milks containing (a) 0%, (b) 0.5%, (c) 1.5%, (d) 2.0%, (e) 3.2% and (f) 7.5% (evaporated) of fat; spectra are offset for clarity.
Please cite this article as: S. Mazurek, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.03.024i
S. Mazurek et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
bands present in different regions of the spectra. The S/N values obtained for milk Raman spectra were at least 10-fold lower than those found for spectrum of pure lactose (FT-Raman spectra of water, lactose, fat and casein are shown in Fig. S2 of supplementary material). All PLS calibration models were constructed using the Raman spectra of the 75 laboratory prepared milk samples. Sixty samples were used for training purposes while 15 constituted an external validation set (see Table S1 of supplementary material), and the selection of validation samples was supported by the PCA. Principal component analysis was used to assess spectral uniformity between calibration samples and commercial samples, and Fig. 2 shows the score plots obtained as a result of PCA decomposition of the Raman intensity matrix. The positions of real milk samples in the variability space of the few first principal components confirmed that spectra of the calibration mixtures reflect the spectral variability characterising commercial milk samples. These data also indicated that the laboratory mixtures reflect the chemical structure of real milk samples. A preliminary analysis performed for samples prepared by dissolving lactose, casein and selected other milk constituents in water resulted in milk-like samples poorly mimicking liquid milk. These samples were outliers in the PCA score plots obtained from spectral data.
Scores on PC 2 (4.57%)
2
1
0
-1
-2
-6
-4
-2
0
2
4
6
4
6
Scores on PC 1 (40.42%)
Scores on PC 3 (3.55%)
2
1
0
-1
3
Table 1 Calibration parameters for milk component determination. Component
Parameter
FT Raman
FTIR ATR
8 cm 1
16 cm 1
32 cm 1
Fat
R Rcv RSEPcal (%) RSEPval (%)
0.994 0.964 5.09 6.41
0.993 0.969 4.66 5.91
0.989 0.968 6.27 7.05
0.995 0.977 3.28 3.47
Protein
R Rcv RSEPcal (%) RSEPval (%)
0.985 0.849 3.37 4.47
0.980 0.873 3.43 3.80
0.930 0.771 6.32 6.79
0.991 0.988 1.86 2.28
Carbohydrates
R Rcv RSEPcal (%) RSEPval (%)
0.983 0.811 3.19 4.35
0.980 0.889 2.82 3.10
0.938 0.803 5.59 5.79
0.993 0.990 1.86 1.93
Dry matter
R Rcv RSEPcal (%) RSEPval (%)
0.976 0.897 4.62 5.19
0.994 0.971 2.13 2.65
0.987 0.957 3.24 3.42
0.995 0.990 2.36 2.47
Spectral ranges of 370–860, 1020–1809 and 2800–2953 cm 1, including the most characteristic milk bands were selected for modelling purposes. The number of latent PLS variables were determined on the basis of RMSECV plots (Fig. S3 of supplementary material) and were set to 3, except for protein and carbohydrates in the case of spectra recorded with the resolution of 8 cm 1, when 4 PLS factors were used. The detailed PLS modelling results are summarised in Table 1. Although a difference in the quality of models based on Raman spectra registered with different resolution is visible, it is not pronounced. Correlation coefficient for the obtained calibration curves varied in the range 0.930-0.994. The RSEP errors were generally lower for models based on Raman spectra registered with the resolution of 16 cm 1. These values varied in the 2.1–4.7% and 2.6–5.9% range for calibration and validation samples, respectively (Table 1). Raman spectra measured with the resolution of 32 cm 1 are characterised by a higher S/N ratio in comparison with those registered with a better resolution and they can be recorded faster. Unfortunately, apparently too much of the specific spectral information is lost and final calibration models, based on these spectra, give higher quantification errors. On the other side, spectra registered with the resolution of 8 cm 1 are too noisy which results in less robust calibration models. Calibration curves, based on Raman spectra recorded with the resolution of 16 cm 1, and plots of the relative errors obtained for milk components are presented in Fig. 3. Twenty one commercial milk samples containing 1.5% (n ¼7), 2.0% (n ¼7) and 3.2% (n ¼7) fat were analysed applying the developed calibration models. The RSEPtest errors in the range of 5.3–5.8% for fat, 5.6–6.1% for protein, 3.5–4.8% for carbohydrates and 3.4–4.8% for dry matter were found (Fig. 4). Attempts to determine amount of fat in milk samples containing less than 1% fat resulted in unacceptable high quantification errors due to the weakness of respective bands in the Raman spectra.
-2
3.2. Repeatability of quantification -6
-4
-2
0
2
Scores on PC 1 (40.42%) Fig. 2. PCA analysis: score plots for milk samples with the 95% confidence interval; squares-calibration samples, circles-validation set, triangles-commercial samples.
In a separate step, a set of 64 samples of commercial milk samples containing 3.2% of fat were analysed using the developed calibration models (the scores plots of PCA modelling for these samples are
Please cite this article as: S. Mazurek, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.03.024i
S. Mazurek et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
4
fat R=0.9934 10
3 Difference [%]
Calculated C [%]
4
2
0
-10
1
1
2
3
4
1
5
4
3
2 Actual C [%]
Actual C [%]
protein R=0.9805
Difference [%]
Calculated C [%]
5 4
3
0
-5
2 2
3
4
5
2
3
7
4
5
Actual C [%]
Actual C [%] carbohydrates R=0.9803
Difference [%]
Calculated C [%]
5 6
5
0
-5
4
4
5
6
7
4
16
6
5
Actual C [%]
7
Actual C [%]
dry matter R=0.9937
6
Difference [%]
Calculated C [%]
3 12
0
-3
8 -6 8
12
Actual C [%]
16
8
12
16
Actual C [%]
Fig. 3. Calibration curves and relative errors for milk components on the basis of Raman spectra registered with the resolution of 16 cm 1; squares-calibration samples, circles-validation samples.
Please cite this article as: S. Mazurek, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.03.024i
S. Mazurek et al. / Talanta ∎ (∎∎∎∎) ∎∎∎–∎∎∎
-1
Raman 8 cm -1 Raman 16 cm -1 Raman 32 cm FTIR ATR
RSEP [%]
6
5
another effective tool for the simultaneous quantitative determination of macronutrients in milk. Fat, protein, carbohydrates and dry matter were quantified in 21 commercial milk samples with RSEPtest errors of 3.4–5.6% relative to the values found applying the IR reference method. Results of a better quality were obtained from a separate PLS model based on IR spectra registered using a single reflection ATR diamond accessory, yielding RSEPtest values of 2.4–4.4%.
3
Acknowledgments Authors thank Ms. A. Łoza and Ms. M. Lenarska from the Wrocław University of Environmental and Life Sciences for performing the reference analysis of milk samples. 0 fat
protein
carbohydrates
dry matter
Fig. 4. Raman and ATR RSEPtest quantification errors for macronutrients and dry matter in commercial milk samples (n¼21).
shown in Fig S4 of supplementary material). Milk components were quantified with the mean relative error values in the range 4.2–5.8% for fat, 3.8–4.6% for protein, 3.9–4.8% for carbohydrates and 2.4–2.8% for dry matter (Fig. S5 of supplementary material).
Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.talanta.2015.03. 024.
References 3.3. FTIR ATR analysis Simultaneous with Raman measurements, single reflection ATR spectra of the same milk samples were recorded (two sets of milk IR spectra, raw and after water spectrum subtraction, are shown in Figs. S6 and S7 of supplementary material, respectively). Applying methodology analogous to that used during Raman data analysis, PLS modelling was performed (calibration curves and relative errors plots for milk constituents were shown in Fig. S8 of supplementary material). The following spectral ranges were utilised: 800–1017, 1063–1634 and 1732–1760 cm 1. Correlation coefficient values determined on the basis of ATR data varied from 0.991 to 0.995, while RSEP errors were 1.9–3.3% and 1.9–3.5% for calibration and validation samples, respectively. Representative results for the ATR PLS modelling are shown in Table 1. Quantitative analyses of commercial samples resulted in the RSEPtest error values of 2.4–4.4% (Fig. 4). 3.4. Comparison of the techniques The relatively low intensity of Raman signals given by the milk samples suggested that longer collection times, in comparison with the IR technique, were required to obtain acceptable S/N ratios. Having in mind the quality differences between the Raman and ATR spectra (see Fig. 1 and Figs. S6 and S9 of supplementary material) it is not surprising that the prediction errors for carbohydrates quantification based on the Raman spectra appeared to be 50% and for protein determination even two times higher when compared with those obtained from ATR data (Fig. 4). The obvious advantage of Raman spectroscopy over IR technique is that it can be applied to the on-line monitoring of the chemical composition of milk at different stages of the industrial milk processing.
[1] R.G. Jensen, Handbook of Milk Composition, Academic Press, London, 1995. [2] A. Varnam, J.P. Sutherland, Milk and Milk Products: Technology, Chemistry and Microbiology, Chapman & Hall, New York, 1994. [3] D.A. Biggs, J. Dairy Sci. 50 (1967) 799–803. [4] F.A. Iñón, S. Garrigues, M. de la Guardia, Anal. Chim. Acta 513 (2004) 401–412. [5] J.D.S. Goulden, J. Dairy Sci. 31 (1964) 273–284. [6] R. Tsenkova, S. Atanassova, K. Itoh, Y. Ozaki, K. Toyoda, J. Anim. Sci. 78 (2000) 515–522. [7] H. Soyeurt, P. Dardenne, F. Dehareng, G. Lognay, D. Veselko, M. Marlier, C. Bertozzi, P. Mayeres, N. Gengler, J. Dairy Sci. 89 (2006) 3690–3695. [8] B. Aernouts, E. Polshin, J. Lammertyn, W. Saeys, J. Dairy Sci. 94 (2011) 5315–5329. [9] A. Melfsen, E. Hartung, A. Haeussermann, J. Dairy Sci. 95 (2012) 6465–6476. [10] E.C.Y. Li-Chan, Trends Food Sci. Technol. 7 (1996) 361–370. [11] R. Szostak, S. Mazurek, Analyst 127 (2002) 144–148. [12] L.G. Thygesen, M.M. Løkke, E. Micklander, S.B. Engelsen, Trends Food Sci. Technol. 14 (2003) 50–57. [13] A.M. Tudora, C.D. Melia, J.S. Binns, P.J. Hendra, S. Church, M.C. Davies, J. Pharm. Biomed. Anal. 8 (1990) 717–720. [14] J. Moros, S. Garrigues, M. de la Guardia, Anal. Chim. Acta 593 (2007) 30–38. [15] C.M. McGoverin, A.S.S. Clark, S.E. Holroyd, K.C. Gordon, Anal. Chim. Acta 673 (2010) 26–32. [16] M.R. Almeida, K. Oliveira, R. Stephani, L.F.C. de Oliveira, J. Raman Spectrosc. 42 (2011) 1548–1552. [17] J. Qin, K. Chao, M.S. Kim, Food Chem. 138 (2013) 998–1007. [18] S. Okazaki, M. Hiramatsu, K. Gonmori, O. Suzuki, A.T. Tu, Forensic Toxicol. 27 (2009) 94–97. [19] Y. Cheng, Y. Dong, J. Wu, X. Yang, H. Bai, H. Zheng, D. Ren, Y. Zou, M. Li, J. Food Compos. Anal. 23 (2010) 199–202. [20] M. Liang, V.Y.T. Chen, H.L. Chen, W. Chen, Talanta 69 (2006) 1269–1277. [21] B.M. Wise, B.R. Kowalski, in: F. McLennan, B.R. Kowalski (Eds.), Process in Analytical Chemistry, Chapman & Hall, London, 1995, pp. 259–312. [22] R. Kramer, Chemometric Techniques for Quantitative Analysis, Marcel Dekker, New York, 1998. [23] I. Stefanov, V. Baeten, B. De Baets, V. Fievez, Talanta 112 (2013) 101–110. [24] R.M. El-Abassy, P.J. Eeravuchira, P. Donfack, B. von der Kammer, A. Materny, Appl. Spectrosc. 66 (2012) 538–544. [25] PN-ISO 9622:2006 (IDF 141), Milk and Liquid Milk Products—Guidelines for the Application of Mid-infrared Spectrometry, Polish Committee for Standardization, Warsaw, 2006. [26] R. Szostak, S. Mazurek, Vib. Spectrosc. 49 (2009) 298–302.
4. Conclusions The results indicated that multivariate modelling using the PLS method based on Raman spectra of liquid milk samples was
Please cite this article as: S. Mazurek, et al., Talanta (2015), http://dx.doi.org/10.1016/j.talanta.2015.03.024i