Fourier Transform Infrared Spectroscopy (FT-IR) - MDPI

48 downloads 106 Views 3MB Size Report
Jun 22, 2017 - dividing the intensity of the individual band in the 1800–600 cm−1 region ... mid infrared (mid-IR) region of 1800–600 cm−1 are anticipated for ...
sensors Article

Fourier Transform Infrared Spectroscopy (FT-IR) and Simple Algorithm Analysis for Rapid and Non-Destructive Assessment of Developmental Cotton Fibers Yongliang Liu 1, * and Hee-Jin Kim 2 1 2

*

Cotton Structure & Quality Research Unit, United States Department of Agriculture (USDA), Agricultural Research Service (ARS), New Orleans, LA 70124, USA Cotton Fiber Bioscience Research Unit, United States Department of Agriculture (USDA), Agricultural Research Service (ARS), New Orleans, LA 70124, USA; [email protected] Correspondence: [email protected]; Tel.: +1-504-286-4445

Received: 15 May 2017; Accepted: 20 June 2017; Published: 22 June 2017

Abstract: With cotton fiber growth or maturation, cellulose content in cotton fibers markedly increases. Traditional chemical methods have been developed to determine cellulose content, but it is time-consuming and labor-intensive, mostly owing to the slow hydrolysis process of fiber cellulose components. As one approach, the attenuated total reflection Fourier transform infrared (ATR FT-IR) spectroscopy technique has also been utilized to monitor cotton cellulose formation, by implementing various spectral interpretation strategies of both multivariate principal component analysis (PCA) and 1-, 2- or 3-band/-variable intensity or intensity ratios. The main objective of this study was to compare the correlations between cellulose content determined by chemical analysis and ATR FT-IR spectral indices acquired by the reported procedures, among developmental Texas Marker-1 (TM-1) and immature fiber (im) mutant cotton fibers. It was observed that the R value, CIIR , and the integrated intensity of the 895 cm−1 band exhibited strong and linear relationships with cellulose content. The results have demonstrated the suitability and utility of ATR FT-IR spectroscopy, combined with a simple algorithm analysis, in assessing cotton fiber cellulose content, maturity, and crystallinity in a manner which is rapid, routine, and non-destructive. Keywords: Fourier transform infrared spectroscopy; attenuated total reflection; cellulose; fiber secondary cell wall biosynthesis; principal component analysis; algorithm analysis

1. Introduction As one of the most important and widely grown crops in the world, cotton is a well-traded agricultural commodity primarily for its naturally produced textile fiber [1]. Commercial cotton fibers are harvested from cotton plants. In biological terms, cotton fibers are the dried cell walls of formerly living cells. They initiate from an ovary of the flower and grow into a mature seed-containing cotton boll within approximately 1.5~2 months. Fiber growth consists of four overlapping but distinctive phases: initiation, primary cell wall (PCW) formation for fiber elongation, secondary cell wall (SCW) biosynthesis for cellulose deposition and cell wall thickening, and maturation [2,3]. The day of flowering is referred to as anthesis, and the term “days post anthesis” (DPA) is commonly used to describe the cotton fiber growth (Figure 1). The fiber cells initiate at 0 DPA and then elongate to reach a fiber length of 22~35 mm within 20 to 25 DPA. The secondary cell wall synthesis starts around 15 to 22 DPA and continues for an additional 30 to 40 days until the maturation phase, when the fibers dehydrate and collapse into flattened and twisted ribbons. The apparent cellulose amount in cotton fibers increases with cotton fiber growth, leading to significant differences in chemical, physical, and Sensors 2017, 17, 1469; doi:10.3390/s17071469

www.mdpi.com/journal/sensors

Sensors 2017, 17, 1469

2 of 13

Sensors 2017, 17, 1469

2 of 13

ribbons. The apparent cellulose amount in cotton fibers increases with cotton fiber growth, leading to significant differences in chemical, physical, and compositional attributes during cotton fiber development. Over the during years, cotton diversified and well-defined testing methods, as high compositional attributes fiber development. Over fiber the years, diversified andsuch well-defined volume instrument (HVI) advanced fiber information (AFIS), have been developed to fiber testing methods, suchand as high volume instrument (HVI)system and advanced fiber information system reflect routinelytoinreflect the cotton (AFIS),these havechanges been developed theseindustry changes [4]. routinely in the cotton industry [4].

Figure1.1.Schematic Schematicof ofgeneral generalcotton cottonfiber fibergrowth growthat atvarious variousdays dayspost postanthesis anthesis (DPA). (DPA). Figure

Mature Mature fibers fibers are are composed composed mostly mostly of of cellulose cellulose (88.0–96.5%), (88.0–96.5%), followed followed by by such such non-cellulosic non-cellulosic constituents as proteins (1.0–1.9%), waxes (0.4–1.2%), pectins (0.4–1.2%), inorganics (0.7–1.6%), constituents as proteins (1.0–1.9%), waxes (0.4–1.2%), pectins (0.4–1.2%), inorganics (0.7–1.6%), and and other other substances substances(0.5–8.0%) (0.5–8.0%)[5]. [5].Analysis Analysisof ofcell cellwall wallcompositions compositionsof offibers fibersfrom from the the early early stages stages of of elongation through the period of the SCW formation was achieved by a set of extraction, separation, elongation through the period of the SCW formation was achieved by a set of extraction, separation, and and isolation isolation steps, steps, prior prior to to chemical chemical and and instrumental instrumental determination determination of of targeted targeted components components [6–9]. [6–9]. Cotton Cotton fiber fiber cellulose cellulose isis not not easily easily dissolved dissolved in in most most solvent, solvent, and and the the extraction extraction and and separation separation process process in in conventional conventional methods methods experience experience significant significant drawbacks drawbacks that that include includethe thetedious tediousprocedures procedures of optimal solvent and temperature selection as well as extracted specimen identification. of optimal solvent and temperature selection as well as extracted specimen identification. As As the the need need for number of of available available analytical analytical techniques, techniques, including for more more rapid rapid approaches approaches inceases, inceases, aa number including Fourier Fourier transform transform infrared infrared (FT-IR) (FT-IR) spectroscopy, spectroscopy, differential differential scanning scanning calorimeter calorimeter (DSC), (DSC), thermogravimetric thermogravimetric analysis and pyrolysis-gas pyrolysis-gas chromatography/mass chromatography/mass spectroscopy analysis (TGA), (TGA), and spectroscopy (GC/MS) (GC/MS)methods methods[10–15], [10–15], have been applied to identify cellulose and non-cellulose components in cottons. These measurements have been applied to identify cellulose and non-cellulose components in cottons. These measurements provide components in developing cotton fibersfibers through the onset of SCW provide clear clearevidence evidenceofofvarious various components in developing cotton through the onset of −1 − 1 synthesis. For example, the appearance of FT-IR bands at 1733bands cm (C=O stretching originating from SCW synthesis. For example, the appearance of FT-IR at 1733 cm (C=O stretching −1 (NH2 deformation − 1 (NH deformation esters or amides) 1534 to proteins or amino acids) reflects originating fromand esters orcm amides) and 1534 cmcorresponding corresponding to proteins 2 the presence of such non-cellulosic components as esters and proteins in cotton fibers [11,13]. Among or amino acids) reflects the presence of such non-cellulosic components as esters and proteins in cotton these the ATR FT-IR spectroscopy method hasspectroscopy evolved as an important alternative to fiberstechniques, [11,13]. Among these techniques, the ATR FT-IR method has evolved as an examine cotton fiber development, because it requires minimal sample preparation by ATR sampling important alternative to examine cotton fiber development, because it requires minimal sample on a small by bundle of cotton on fibers as bundle little asof0.5 mg, fibers permits routine analysis rapidly and preparation ATR sampling a small cotton as little as 0.5 mg, permits routine non-destructively, and is easy to operate [10,11,13–20]. analysis rapidly and non-destructively, and is easy to operate [10,11,13–20]. Different Different spectral spectral interpretation interpretation approaches approaches have have been been applied applied to to acquire acquire useful useful information information from from FT-IR FT-IR measurement. measurement. ItIt includes includes the the direct direct use useof of1-band 1-bandintensity, intensity, the the estimation estimation of of 22- or or 3-band 3-band intensity ratios, and the implementation of a chemometrical or multivariate tool known as principal intensity ratios, and the implementation of a chemometrical or multivariate tool known as principal component component analysis analysis (PCA) (PCA) [10,11,13–21]. [10,11,13–21]. The Theresults resultshave haveindicated indicated that that both both simple simple algorithms algorithms and and PCA PCA patterns patternscan canbe beused usedto tomonitor monitorthe thetransition transitionfrom fromPCW PCWto toSCW SCWbiosyntheses biosynthesesin incotton cottonfibers. fibers. Interestingly, Interestingly, simple simple algorithms algorithms were were observed observed to to have have the the ability ability to to detect detect the the subtle subtle discrepancies discrepancies in in fibers fibers older older than than25 25DPA DPA among among respective respective fibers fibers grown grown in in planta planta or or in in culture culture [17]. [17]. In In the the current current work, we performed the PCA and simple algorithm examination of ATR FT-IR spectra representing work, we performed the PCA and simple algorithm examination of ATR FT-IR spectra representing the the developmental immature mutant and its near-isogenic wildTexas type Marker-1 Texas Marker-1 developmental immature fiberfiber (im) (im) mutant fibersfibers and its near-isogenic wild type (TM-1) (TM-1) fibers, and compared the correlations content determined by conventional fibers, and compared the correlations betweenbetween cellulosecellulose content determined by conventional chemical chemical analysis and respective FT-IR spectralacquired responses acquiredspectral by various spectral analysis and respective ATR FT-IRATR spectral responses by various interpretation interpretation strategies. Notably, simple algorithm analysis of ATR FT-IR spectra has been strategies. Notably, simple algorithm analysis of ATR FT-IR spectra has been developed to estimate developed to maturity estimate and fibercrystallinity cellulose maturity and crystallinity simultaneously in developing fiber cellulose simultaneously in developing cottons [14–17]. cottons [14–17].

Sensors 2017, 17, 1469

3 of 13

2. Materials and Methods 2.1. Texas Marker-1 (TM-1) and Immature Fiber (im) Mutant Fibers Two cotton near-isogenic lines (NILs), TM-1 and im, were grown side by side in a field of USDA-ARS (New Orleans, LA, USA) in 2011. At day of post anthesis (DPA), cotton flowers were tagged. Two biological replicates of cotton bolls were taken from different cotton plants at 10, 17, 24, 28, 33, 37 and 44 DPA. Fibers at each DPA were collected from 10 to 30 bolls in fifty plants for each biological replication by manually removing the seeds, prior to drying in 40 ◦ C incubator. The soil type was Aquent dredged over alluvium in an elevated location to provide adequate drainage. Harvested fibers were kept in a dark storage room with a constant temperature (23 ± 1 ◦ C) and relative humidity (50 ± 10%), and their ATR FT-IR spectra were collected in March 2014. The ~2.5 years gap between the harvest of fibers and the collection of FT-IR spectra was due to the beginning of collaboration between two authors in March 2014. 2.2. Cellulose Content Fiber cellulose content at each developmental stage was measured by the modified Updegraff method [22]. Briefly, 10 mg of cut fibers were placed into 5 mL of reacti-vials. Non-cellulosic materials in fibers were hydrolyzed with acetic-nitric reagent (a mixture of 73% acetic acid, 9% nitric acid and 18% water). The remaining cellulose was hydrolyzed with 67% sulfuric acid (v/v) and measured by a colorimetric assay with anthrone and by the use of Avicel PH-101 (FMC, Rockland, ME, USA) as a cellulose standard. The average cellulose content for each fiber was obtained from three replications. It took at least 2 days to measure cellulose content for each sample mostly due to slow hydrolysis process of fiber cellulose component. 2.3. ATR FT-IR Spectral Collection and Data Analysis All fibers were scanned by an FTS 3000MX FT-IR spectrometer (Varian Instruments, Randolph, MA, USA) equipped with a ceramic source, KBr beam splitter, and deuterated triglycine sulfate (DTGS) detector and attenuated total reflection (ATR) attachment. The ATR sampling device utilized a DuraSamplIR single-pass diamond-coated internal reflection accessory (Smiths Detection, Danbury, CT, USA), and a consistent contact pressure was applied by way of a stainless steel rod and an electronic load display. During the data collection, cautions were taken to make sure that the window (2 mm in diameter) of the ATR sampling devise was covered completely by fiber samples. At least five measurements for individual fiber samples, by re-sampling at different locations across entire sample, were collected over the range of 4000–600 cm−1 at 4 cm−1 and 16 co-added scans. All spectra were given in absorbance units and no ATR baseline correction was applied. Importing the spectra to the GRAMS IQ application in Grams/AI (Version 9.1, Thermo Fisher Scientific, Waltham, MA, USA), the mean spectrum was taken for each sample and then was smoothed with a Savitzky–Golay function (polynomial = 2 and points = 11). The spectra were normalized by dividing the intensity of the individual band in the 1800–600 cm−1 region with the average intensity in this 1800–600 cm−1 region, and subsequent PCA characterization was performed in the 1800–600 cm−1 IR region, with mean centering (MC) and Savitzky–Golay first-derivative (2 degrees and 13 points) spectral pretreatment, as well as with the leave-one-out cross-validation method. With the use of the Grams/AI program, integrated intensities of 4 bands at 2900, 1372, 895, and 664 cm−1 were estimated in the respective ranges of 3000 to 2800, 1410 to 1290, 910 to 875, and 684 to 650 cm−1 from normalized spectra. Separately, the spectral set was loaded into Microsoft Excel 2007 to execute simple algorithm analysis.

Sensors 2017, 17, 1469 Sensors 2017, 17, 1469

4 of 13 4 of 13

3. Results Results and and Discussion Discussion 3. 3.1. DPA-Dependent DPA-DependentCellulose CelluloseContent ContentofofDevelopmental DevelopmentalFibers Fibers 3.1. The DPA-dependent DPA-dependentcellulose cellulose content content between between the the two two sets sets of of developmental developmental fibers fibers in in Figure Figure 22 The shows aa similar similar increasing increasing pattern pattern between between the the two two types types of fibers at at various various developmental developmental stages. shows For either either TM-1 or im from 10 10 to to 37 37 DPA, as For im fiber, fiber, cellulose cellulosecontent contentincreases increasesrapidly rapidlyand andlinearly linearly from DPA, expected. as expected.

Cellulose Content (%)

100

● TM-1 ○ im

80

60

40

20

0 5

10

15

20

25

30

35

40

45

50

DPA (days) Figure 2. 2. Cellulose Cellulose content content against against DPA DPA for for developmental developmental Texas TexasMarker-1 Marker-1(TM-1) (TM-1)((●) and immature immature Figure ) and fiber (im) (○) fibers by the modified Updegraff method [22]. fiber (im) (#) fibers by the modified Updegraff method [22].

3.2. ATR ATR FT-IR FT-IRSpectral SpectralCharacteristics CharacteristicsofofDevelopmental DevelopmentalTM-1 TM-1Cotton CottonFibers Fibers 3.2. With fiber fiberDPA DPAprogressing progressing(Figure (Figure1), 1),apparent apparentspectral spectralintensity intensity increases increases or or decreases decreases in in the the With −1 mid infrared infrared(mid-IR) (mid-IR)region regionof of1800–600 1800–600cm cm−1 are anticipated for for these these developmental developmental TM-1 TM-1 cotton cotton mid are anticipated fibers (Figure 3). A thorough examination of ATR FT-IR spectral feature reveals that TM-1 fiber fibers (Figure 3). A thorough examination of ATR FT-IR spectral feature reveals that TM-1 fiber exhibits a nearly spectral identicalpattern spectraltopattern fiberdevelopmental at same developmental point. TM-1 and aexhibits nearly identical im fibertoatim same point. TM-1 fibers andfibers im fibers, im fibers, alike, demonstrate growth results in the dominant production of major common chemical alike, demonstrate growth results in the dominant production of major common chemical component component in cotton fibers, cellulose. Spectral intensity changes in Figure 3 are in good with agreement in cotton fibers, cellulose. Spectral intensity changes in Figure 3 are in good agreement those with those reported earlier [11,13,17], and characteristic band assignments have been summarized reported earlier [11,13,17], and characteristic band assignments have been summarized in Table in 1. Table 1.the Briefly, the vibration at−1740 cm−1 is assigned to the C=O stretching of carbonyl 1 is assigned Briefly, vibration at 1740 cm to the C=O stretching mode ofmode carbonyl groups groups due to −1 is mainly attributed to the OH bending mode due to and lipids, and a band broadcentered band centered 1620 −1 cm lipids, a broad at 1620atcm is mainly attributed to the OH bending mode of of adsorbed water. Bands region 1500–1200 representthe thecontributions contributionsof ofboth bothCH CH2 −1−1represent adsorbed water. Bands in in thethe region of of 1500–1200 cmcm 2 −1 region originate deformations and C–O–H bending vibrations, and those bands in the 1200–900 cm − 1 deformations and C–O–H bending vibrations, and those bands in the 1200–900 cm region originate fromthe thecoupling couplingmodes modesof ofC–O C–Oand andC–C C–Cvibrations. vibrations.The Thebands bandsbetween between800 800and and700 700cm cm are likely likely −−1 1 are from due to two crystal forms (I α and Iβ) of cotton cellulose. In addition, there are intense absorptions due to two crystal forms (Iα and Iβ ) of cotton cellulose. In addition, there are intense absorptions −1 regions that are assignable to the O–H and C–H stretching betweenthe the3600 3600 1 regions between andand 27502750 cm−cm that are assignable to the O–H and C–H stretching vibrations vibrations (not shown). (not shown). A comparison comparison of of of these bands between shorter DPADPA (for A of the theincreasing increasingorordecreasing decreasingintensity intensity these bands between shorter example, 10 DPA) and longer DPA (for example, 37 DPA) fibers is tabulated in Table 1. In general, (for example, 10 DPA) and longer DPA (for example, 37 DPA) fibers is tabulated in Table 1. In general, intensities of ofthe thebands bandsatat1740, 1740,1620, 1620,1545, 1545,1405, 1405,and and1236 1236cm cm aswell wellas asthose thoseininthe the850–750 850–750cm cm −1−1as −−1 1 intensities region decrease, while those at 1425, 1365, 1335, 1315, 1200, 1158, 1104, 1055, 1028, 985, 895 and 662 region decrease, while those at 1425, 1365, 1335, 1315, 1200, 1158, 1104, 1055, 1028, 985, 895 and cm−1cm increase. Subjective interpretation of these spectra in Figure 3 cannot be be applied to compare or −1 increase. 662 Subjective interpretation of these spectra in Figure 3 cannot applied to compare assess the degree of fiber secondary wall biosynthesis in a semi-quantitative way. or assess the degree of fiber secondary wall biosynthesis in a semi-quantitative way.

Sensors 2017, 17, 1469 Sensors 2017, 17, 1469

5 of 13 5 of 13

Figure 3. 3. Attenuated Attenuatedtotal totalreflection reflectionFourier Fourier transform infrared (ATR FT-IR) spectral response of transform infrared (ATR FT-IR) spectral response of TM-1 TM-1 to various fibers fibers to various DPA. DPA. Table DPA fibers fibers [11,13,17] [11,13,17] *.*. Table 1. 1. Characteristic Characteristic ATR ATR FT-IR spectral bands of shorter and longer DPA

10 DPA Fiber 10 DPA Fiber 1740 (s) 1740 (s) 1620 (s)

37 DPA Fiber 37 DPA Fiber 1740 (w)

Band Assignment Band Assignment C=O stretching 1740 (w) C=O stretching HOH bending of adsorbed water + amide I 1620 (s) HOH bending of adsorbed water + amide I 16201620 (w) (w) HOHHOH bending of adsorbed water bending of adsorbed water 1545 15451545 (w) (w) Amide II II 1545(s) (s) Amide 1425(w) (w) 2 scissoring 1425 14251425 (s) (s) CHCH 2 scissoring 1405 (s) 1405 (w) O–H deformation 1405 (s) 1405 (w) O–H deformation 1365 (w) 1365 (s) C–H bending 1365 (w) 1365 (s) C–H bending 1335 (w) 1335 (s) CH 2 wagging 1335 13351315 (s) (s) CHCH 2 wagging 1315(w) (w) 2 wagging 1236 (s) O–H deformation or N–H deformation 1315 (w) 13151236 (s) (w) CH2 wagging 1200(s) (w) C–O 1236 1236 1200 (w) (s) O–H deformation orstretching N–H deformation 1158 (w) 1158 (s) C–O–C stretching 1200 (w) 12001104 (s) (s) C–OC–O stretching 1104 (w) stretching 1158 (w) 1158 (s) C–O–C stretching 1055 (w) 1055 (s) C–O stretching 1028(w) (w) stretching 1104 11041028 (s) (s) C–OC–O stretching 985(w) (w) stretching 1055 1055 985 (s) (s) C–OC–O stretching 895 (w) 895 (s) β-glycosidic linkage 1028 (w) 1028 (s) C–O stretching 662 (w) 662 (s) O–H out-of-plane bending 985 (w) 985 (s) C–O stretching * Parentheses indicate the relative ATR FT-IR spectral intensities between 10 and 37 DPA fibers: s = strong, w = weak. 895 (w) 895 (s) β-glycosidic linkage 662 (w) 662 (s) O–H out-of-plane bending 3.3. Correlation between Cellulose Content and ATR FT-IR Spectral Response of Developmental Fibers * Parentheses indicate the relative ATR FT-IR spectral intensities between 10 and 37 DPA fibers: = strong, w and = weak. 3.3.1.s PC1 Score the Correlation with Cellulose Content

As usual, PCA were performed to understand the similarity or dissimilarity of ATR FT-IR 3.3. Correlation between Cellulose Content and ATR FT-IR Spectral Response of Developmental Fibersspectra that are indicative of fiber growth for TM-1 and im fibers. The plot of the first principal component (PC1)PC1 score vs. and DPAthe in Correlation Figure 4 provides a good Content visualization of the sample distribution among 3.3.1. Score with Cellulose two sets of fibers. The first two PCs accounted for 89.5% of the total variation, with the PC1 explaining As usual, PCA were performed to understand the similarity or dissimilarity of ATR FT-IR 80.4% of the variation. For both TM-1 and im fibers, PC1 scores increase rapidly between 10 and 24 DPA spectra that are indicative of fiber growth for TM-1 and im fibers. The plot of the first principal before reaching the relatively constant PC1 score. It implies that the 10 DPA fibers are composed component (PC1) score vs. DPA in Figure 4 provides a good visualization of the sample distribution of PCW constituents, whereas the 24 DPA fibers consist of SCW components. The observation is among two sets of fibers. The first two PCs accounted for 89.5% of the total variation, with the PC1 consistent with accumulated knowledge suggesting that elongating fibers at 10 DPA contain no SCW explaining 80.4% of the variation. For both TM-1 and im fibers, PC1 scores increase rapidly between cellulose whereas thickening fibers at 24 DPA are composed of more SCW cellulose. Notably, the PC1 10 and 24 DPA before reaching the relatively constant PC1 score. It implies that the 10 DPA fibers are scores are nearly independent of fiber DPA when DPA is more than 25 days, suggesting that the overall composed of PCW constituents, whereas the 24 DPA fibers consist of SCW components. The contributions from spectral intensity variations in the 1800–600 cm−1 region during this period of fiber observation is consistent with accumulated knowledge suggesting that elongating fibers at 10 DPA maturation are probably insignificant. contain no SCW cellulose whereas thickening fibers at 24 DPA are composed of more SCW cellulose. Notably, the PC1 scores are nearly independent of fiber DPA when DPA is more than 25 days,

Sensors 2017, 17, 1469

6 of 13

suggesting that the overall contributions from spectral intensity variations in the 1800–600 cm−1 6 of 13 region during this period of fiber maturation are probably insignificant.

Sensors 2017, 17, 1469

0.4

● TM-1 ○ im

PC1 Score

0.2

0.0

-0.2

-0.4

-0.6 5

10

15

20

25

30

35

40

45

50

DPA (days) Figure4.4.First First principal component (PC1)against score DPA against from spectra ATR FT-IR spectra of Figure principal component (PC1) score fromDPA ATR FT-IR of developmental developmental TM-1 (●) and im (○) fibers. TM-1 ( ) and im (#) fibers.

The plot of relating cellulose content to PC1 score for developmental TM-1 and im fibers in The plot of relating cellulose content to PC1 score for developmental TM-1 and im fibers in Figure 5 offers a better visualization of relationship between two parameters. In general, cellulose Figure 5 offers a better visualization of relationship between two parameters. In general, cellulose content increases with PC1 score up to 24 DPA by referring to the Figure 4. Remarkably, the content increases with PC1 score up to 24 DPA by referring to the Figure 4. Remarkably, the correlation correlation between the two parameters could be described more accurately by a polynomial (with between the two parameters could be described more accurately by a polynomial (with the order of 2) the order of 2) regression than by a linear regression, as shown in Figure 5. It implies that cellulose regression than by a linear regression, as shown in Figure 5. It implies that cellulose content has a content has a weak linear relationship with PC1 score. weak linear relationship with PC1 score. 0.3

Polynomial regression y = -9E-05x2 + 0.016x - 0.527 R² = 0.92

0.2 0.1

PC1 Score

0.0 -0.1

Linear regression y = 0.0062x - 0.384 R² = 0.82

-0.2 -0.3 -0.4 -0.5 -0.6 0

10

20

30

40

50

60

70

80

90

100

Cellulose Content (%) Figure5.5.PC1 PC1 score vs. cellulose content of developmental TM-1 and Linear im fibers. Linear and Figure score vs. cellulose content of developmental TM-1 and im fibers. and polynomial polynomial (with the order of 2)were regressions to all fiber samples. (with the order of 2) regressions appliedwere to allapplied fiber samples.

3.3.2. IR IR maturity maturity and and Crystallinity Crystallinity and andthe theCorrelation Correlationwith withCellulose CelluloseContent Content 3.3.2. Compared to tothe thePCA PCAapproach approachthat thatutilized utilizedaanumber numberof of624 624datapoints datapointsor orvariables variables(from (fromthe the Compared −1 −1 − 1 − 1 1800to to600 600cm cm IR with 1.949 1.949 cm cm interval proposed 3-band 3-band based based 1800 IR region region with interval in in this this study), study), previously previously proposed ratios or algorithms were assessed on the same spectral dataset. The three bands at 1800, 1315, and ratios or algorithms were assessed on the same spectral dataset. The three bands at 1800, 1315, and −11 were used to assess the ratio R values and these values were found to be related with fiber − 1236 cm 1236 cm were used to assess the ratio R values and these values were found to be related with fiber growth[17]. [17].The Theratio ratioR Rvalues values were assessed by the algorithm R 1315 = (A− 1315 A1800 )/(A 1236 − A1800), growth were assessed by the algorithm of Rof = (A A−1800 )/(A 1236 − A1800 ), where A1800 , A1315 , and A1236 represent the respective band intensities centered at 1800, 1315, and

Sensors Sensors 2017, 2017, 17, 17, 1469 1469 Sensors 2017, 17, 1469

77of of13 13 7 of 13

Values RR Values

where A1800, A1315, and A1236 represent the respective band intensities centered at 1800, 1315, and 1236 where A1800 , A1315, and−1A1236 represent the respective band intensities centered at 1800, 1315, and 1236 1 [17]. −1 band −1 [17]. −1 band 1236 cm−The The cmwas offset zero insubjectively. intensity subjectively. The 1315arises cm−1 cm 1800 cm1800 band offsetwas to be zeroto inbe intensity The 1315 cm cm−1 [17]. The 1800 cm−1 band was offset to be zero in intensity subjectively. The 1315 cm−1 band arises −1 band band the arises from the CH2mode wagging that is increasing in intensity withwhile DPA, the while the cm 1236 cm−1 from CH 2 wagging thatmode is increasing in intensity with DPA, 1236 from the CH2 wagging mode that is increasing in intensity with DPA, while the 1236 cm−1 band band assigned to O–H/N–H deformation decreases in intensity. shown Figure theRR value value assigned to O–H/N–H deformation decreases in intensity. As As shown in in Figure 6, 6,the assigned to O–H/N–H deformation decreases in intensity. As shown in Figure 6, the R value increases along with the DPA, is increases DPA, both both directly directlyand andobviously. obviously.The Thepattern patternsuggests suggeststhat thatthis thisalgorithm algorithm increases along with the DPA, both directly and obviously. The pattern suggests that this algorithm capable ofof reflecting thethe continuous cellulose production in fibers older than 24 DPA, which waswas not is capable reflecting continuous cellulose production in fibers older than 24 DPA, which is capable of reflecting the continuous cellulose production in fibers older than 24 DPA, which was not apparent from score of PCA the PCA result in Figure apparent from the the PC1PC1 score of the result in Figure 4. 4. not apparent from the PC1 score of the PCA result in Figure 4. 2.5 2.5 2.3 2.3 2.1 2.1 1.9 1.9 1.7 1.7 1.5 1.5 1.3 1.3 1.1 1.1 0.9 0.9 0.7 0.7 5 5

● TM-1 TM-1 ○● im ○ im

10 10

15 15

20 20

25 25

30 30

35 35

DPA (days) DPA (days)

40 40

45 45

50 50

Figure (○) fibers. Figure6.6.RRvalue valueagainst againstDPA DPA from from ATR ATR FT-IR spectra of developmental TM-1 ((●)) and im (#) fibers. Figure 6. R value against DPA from ATR FT-IR spectra of developmental TM-1 (●) and im (○) fibers.

Unlike the trend in Figure 5, there is no visual disparity in Figure 7 between aa polynomial Unlike Figure Unlike the the trend trend in in Figure Figure 5, 5, there there is is no no visual visual disparity disparity in in Figure 77 between between a polynomial polynomial 2 regression and aalinear linearregression. regression.AA strong linear correlation = 0.96) is within the expectation 2 (R regression and strong linear correlation (R = 0.96) is within the expectation since 2 regression and a linear regression. A strong linear correlation (R = 0.96) is within the expectation since two independent methods should be consistent in representing cottonfiber fibergrowth. growth. This This result two independent methods should be consistent in representing cotton result since two independent methods should be consistent in representing cotton fiber growth. This result demonstrates that ATR FT-IR technique, along with this simple algorithm, can be used for demonstrates that the the demonstrates that the ATR ATR FT-IR FT-IR technique, technique, along along with with this this simple simple algorithm, algorithm, can can be be used used for for monitoring fiber cellulose content and fiber maturity, rapidly and non-destructively. monitoring fiber cellulose content and fiber maturity, rapidly and non-destructively. monitoring fiber cellulose content and fiber maturity, rapidly and non-destructively. 2.5 2.5 2.3 2.3

Polynomial regression regression + 0.852 y = Polynomial -1E-05x2 + 0.015x 2 + 0.015x + 0.852 y = -1E-05x R² = 0.96 R² = 0.96

2.1 2.1

Values RR Values

1.9 1.9 1.7 1.7 1.5 1.5

Linear regression regression yLinear = 0.014x + 0.872 y = 0.014x + 0.872 R² = 0.96 R² = 0.96

1.3 1.3 1.1 1.1 0.9 0.9 0.7 0.7 0 0

10 10

20 20

30 30

40 40

50 50

60 60

70 70

Cellulsoe Content (%) Cellulsoe Content (%)

80 80

90 90

100 100

Figure 7. R value vs. cellulose content of developmental TM-1 and im fibers. Linear and polynomial Figure 7. 7. R R value value vs. vs. cellulose content of developmental TM-1 and im im fibers. fibers. Linear and polynomial polynomial (with the order of 2) regressions were applied to all fiber samples. (with the order of 2) regressions were applied to all fiber samples. (with the order of 2) regressions were applied to all fiber samples.

A further interest was to estimate IR crystallinity (CIIR) by using the respective IR bands at 708, A further further interest interest was was to to estimate IR IR crystallinity crystallinity (CI (CIIR)) by using respective IR bands at A using the theto respective IRin bands at 708, 708, 730, and 800 cm−1 [15–17]. Twoestimate bands at 708 and 730 cm−1 IR areby assignable Iβ cellulose crystalline −11 [15–17]. Two bands at 708 and 730 cm− −11 are assignable to Iβ cellulose in crystalline − 730, and 800 cm 730, and [15–17]. Two bands at 708 and 730 cm are assignable to Iβ cellulose in crystalline −1 band was part and 800 to Icm α cellulose in amorphous part, respectively, whereas the 800 cm used to part and to Iα cellulose in amorphous part, respectively, whereas the 800 cm−1 band was used to

Sensors 2017, 17, 1469

8 of 13

Sensors 2017, 17, 1469 Sensors 2017, 17, 1469

8 of 13 8 of 13

part and to Iα cellulose in amorphous part, respectively, whereas the 800 cm−1 band was used to offset offset in intensity subjectively. CIIR against DPA in Figure 8 reveals a steady CIIR increase for in intensity subjectively. Plot ofPlot CI of against DPA in Figure 8 reveals a steady CIIR increase for each offset in intensity subjectively. PlotIRof CIIR against DPA in Figure 8 reveals a steady CIIR increase for each set, fiberasset, as anticipated. The general in Figure 8 is similar to in that in Figure 6 more so than fiber anticipated. The general trendtrend in Figure 8 is similar to that Figure 6 more so than that each fiber set, as anticipated. The general trend in Figure 8 is similar to that in Figure 6 more so than that which is in Figure 4, when examining CI IR values for the fibers older than 24 DPA. which is in Figure 4, when examining the CIthe values for the fibers older than 24 DPA. IR CIIR values for the fibers older than 24 DPA. that which is in Figure 4, when examining the

Crystallinity(CI (CI IRIRCrystallinity )) IRIR

100 100

● ● ○ ○

80 80

TM -1 TM -1 im im

60 60 40 40 20 20 0 0

5 5

10 10

15 15

20 20

25 25

30 30

35 35

40 40

45 45

50 50

DPA (days) DPA (days) Figure 8. 8. IR crystallinity crystallinity (CIIR)) against from ATR ATR FT-IR spectra spectra of developmental developmental TM-1 ((●) and Figure against DPA DPA from ) and Figure 8. IR IR crystallinity (CI (CIIR IR) against DPA from ATR FT-IR FT-IR spectra of of developmental TM-1 TM-1 (●) and im (○) fibers. im im (#) (○) fibers. fibers.

Figure 9 shows a high linear relationship (R222 = 0.92) between the degree of CIIR and cellulose Figure 99 shows shows aahigh highlinear linearrelationship relationship(R (R == 0.92) 0.92) between between the the degree degreeof ofCI CIIRIR and and cellulose cellulose Figure content. There is an insignificant difference in Figure 9 between a polynomial regression and a linear content. There is is an an insignificant insignificant difference difference in in Figure Figure 99 between between aa polynomial polynomial regression regression and and aa linear linear content. regression, and the pattern in Figure 9 is much closer to that in Figure 7, than that which is in Figure 5. regression, and the pattern in Figure 9 is much closer to that in Figure 7, than that which is in Figure 5. regression, and the pattern in Figure 9 is much closer to that A strong linear correlation (R222 = 0.92) is much expected, because cotton fiber crystallinity increases A strong stronglinear linearcorrelation correlation(R (R == 0.92) 0.92) is is much much expected, expected, because because cotton cotton fiber fiber crystallinity crystallinity increases increases A with cotton fiber growth [14,16–18]. This result verifies that the ATR FT-IR technique, along with this with cotton fiber growth [14,16–18]. This result verifies that the ATR FT-IR technique, along with this this with cotton fiber growth [14,16–18]. This result verifies that the ATR FT-IR technique, along with simple algorithm, can also be used for assessing fiber crystallinity, rapidly and non-destructively. simple algorithm, algorithm,can canalso alsobe beused usedfor forassessing assessingfiber fibercrystallinity, crystallinity,rapidly rapidlyand andnon-destructively. non-destructively. simple

Crystallinity(CI (CI IRIRCrystallinity )) IRIR

100 100

Polynomial regression Polynomial regression y = -0.0025x22 + 0.838x + 10.91 y = -0.0025x + 0.838x + 10.91 R² = 0.93 R² = 0.93

80 80

60 60

40 40

Linear regression Linear regression y = 0.592x + 14.62 y = 0.592x + 14.62 R² = 0.92 R² = 0.92

20 20

0 0

0 0

10 10

20 20

30 30

40 40

50 50

60 60

70 70

80 80

90 90

100 100

Cellulose Content (%) Cellulose Content (%) Figure9. 9. IR crystallinity crystallinity (CIIR IR) vs. cellulose of developmental developmental TM-1 and and im fibers. fibers. Linear and and Figure cellulose content content Figure 9. IR IR crystallinity(CI (CIIR)) vs. vs. cellulose content of of developmental TM-1 TM-1 and im im fibers. Linear Linear and polynomial (with the order of 2) regressions were applied to all fiber samples. polynomial polynomial(with (withthe theorder orderof of2)2)regressions regressionswere wereapplied appliedtotoall allfiber fibersamples. samples.

3.3.3. The 1372 and 2900 cm−1 Band Based IR Crystallinity and the Correlation with Cellulose Content 3.3.3. The 1372 and 2900 cm−1 Band Based IR Crystallinity and the Correlation with Cellulose Content Two bands at 1372 and 2900 cm−1 based IR Ratio 1372/2900 have been reported to indirectly Two bands at 1372 and 2900 cm−1 based IR Ratio 1372/2900 have been reported to indirectly estimate the percentage crystallinity in cellulose materials [14,21]. The 1372 cm−1 band is attributed to estimate the percentage crystallinity in cellulose materials [14,21]. The 1372 cm−1 band is attributed to the C–H bending vibration, while the 2900 cm−1 band is assigned to the C–H stretching mode. As the C–H bending vibration, while the 2900 cm−1 band is assigned to the C–H stretching mode. As shown in Figure 10, the IR ratio 1372/2900 elevates rapidly from 10 to 28 DPA, and is consistent with shown in Figure 10, the IR ratio 1372/2900 elevates rapidly from 10 to 28 DPA, and is consistent with

Sensors 2017, 17, 1469

9 of 13

3.3.3. The 1372 and 2900 cm−1 Band Based IR Crystallinity and the Correlation with Cellulose Content Two bands at 1372 and 2900 cm−1 based IR Ratio 1372/2900 have been reported to indirectly estimate the percentage crystallinity in cellulose materials [14,21]. The 1372 cm−1 band is attributed to the2017, C–H17,bending vibration, while the 2900 cm−1 band is assigned to the C–H stretching mode. Sensors 1469 9 of 13 As shown 10, the IR ratio 1372/2900 elevates rapidly from 10 to 28 DPA, and is consistent Sensors 2017, in 17, Figure 1469 9 of 13 with different cotton varieties reported by other researchers [14]. The pattern in Figure 10 resembles different cotton varieties reported by other researchers [14]. The pattern in Figure 10 resembles the different cotton reported [14]. The pattern the PC1 score of varieties PCA representation inother Figure 4, other those in Figures 6Figure and PC1 score of PCA representation inby Figure 4,researchers other thanthan those in Figures 6 in and 8. 8.10 resembles the PC1 score of PCA representation in Figure 4, other than those in Figures 6 and 8. 1.0 1.0

Ratio 1372/2900 IR IR Ratio 1372/2900

0.8 0.8

● TM-1 ○● im TM-1 ○ im

0.6 0.6 0.4 0.4 0.2 0.2 0.0 0.0 5 5

10

15

20

35

40

45

50

10

15

20DPA 25 (Days) 30 35

25

30

40

45

50

DPA (Days) Figure Figure 10. 10. IR IR Ratio Ratio 1372/2900 1372/2900against againstDPA DPAfrom fromATR ATR FT-IR FT-IR spectra spectra of of developmental developmental TM-1 TM-1 (●) ( ) and and im im −1. −im Figure 10. IR Ratio 1372/2900 against DPA from ATR FT-IR spectra of developmental TM-1 (●) 1. (○) fibers. The ratios were calculated from integrated intensities of two bands at 1372 and 2900 cm (#) fibers. The ratios were calculated from integrated intensities of two bands at 1372 and 2900 and cm (○) fibers. The ratios were calculated from integrated intensities of two bands at 1372 and 2900 cm−1.

Much like the tendency in Figure 5, cellulose content increases with the IR Ratio 1372/2900 in Much the tendency ininFigure 5, cellulose content increases with the the IR Ratio 1372/2900 in the Much like like tendencycurve Figure cellulose content with IR Ratio 1372/2900 in the manner of a the polynomial over a5,linear curve (Figureincreases 11). manner of a polynomial curve over a linear curve (Figure 11). the manner of a polynomial curve over a linear curve (Figure 11). 1.0

Band Ratio (172/2900) IR IR Band Ratio (172/2900)

1.0 0.8 0.8

Polynomial regression 2 + 0.0158x + 0.0401 y = -8E-05x Polynomial regression R² 0.98 + 0.0401 y = -8E-05x2 +=0.0158x R² = 0.98

0.6 0.6

Linear regression yLinear = 0.0076x + 0.163 regression R² = 0.92 y = 0.0076x + 0.163 R² = 0.92

0.4 0.4 0.2 0.2 0.0 0.0 0 0

10

20

10

20

30

80

90

100

30 40 Content 50 60 (%) 70 80 Cellulose

40

50

60

70

90

100

Cellulose Content (%) Figure 11. IR Ratio 1372/2900 vs. cellulose content of developmental TM-1 and im fibers. Linear and Figure 11. 11. IR IR Ratio 1372/2900 vs.regressions cellulose content of developmental fibers. Linear and and Figure Ratio 1372/2900 TM-1 and im fibers. polynomial (with the order of 2) were applied to all fiber samples. polynomial (with (with the the order order of of2) 2)regressions regressionswere wereapplied appliedto toall allfiber fibersamples. samples. polynomial

3.3.4. Integrated Intensity of the 895 cm−1 Band and the Correlation with Cellulose Content 3.3.4. Integrated Intensity of the 895 cm−1 Band and the Correlation with Cellulose Content The band at 895 cm−1 has been assigned to the β-glycosidic linkage in cellulose, and its −1 has been assigned to the β-glycosidic linkage in cellulose, and its The band at 895 integrated intensity wascm found to be linearly correlated with the percentage of cellulose [13] and of integrated intensity was found to from be linearly correlated in with the percentage of cellulose and of crystallinity [14]. Slightly differing the relationship a previous report [13], Figure 12[13] indicates −1 crystallinity [14]. Slightly differing from the relationship in a previous report [13], Figure 12 indicates an obvious linear response of integrated intensity of the 895 cm band to fiber growth. −1 band to fiber growth. an obvious of integrated of the 895 cm A greatlinear linearresponse relationship (R2 = 0.95) intensity is observed between the integrated intensity of the 895 cm−1 2 A great linear relationship (R = 0.95) is observed between the integrated intensity of the 7895 cm band and cellulose content (Figure 13). The pattern in Figure 13 is identical to those in Figures and 9.−1 band and cellulose content (Figure 13). The pattern in Figure 13 is identical to those in Figures 7 and 9.

Sensors 2017, 17, 1469

10 of 13

3.3.4. Integrated Intensity of the 895 cm−1 Band and the Correlation with Cellulose Content The band at 895 cm−1 has been assigned to the β-glycosidic linkage in cellulose, and its integrated intensity was found to be linearly correlated with the percentage of cellulose [13] and of crystallinity [14]. Slightly differing from the relationship in a previous report [13], Figure 12 indicates an obvious linear response of integrated intensity of the 895 cm−1 band to fiber growth. 1 Sensors 1469 relationship (R2 = 0.95) is observed between the integrated intensity of the 89510 of−13 A2017, great17,linear cm Sensors 2017, 17, 1469 10 of 13 band and cellulose content (Figure 13). The pattern in Figure 13 is identical to those in Figures 7 and 9.

-1) -1 Intergrated Intensity (895 cmcm Intergrated Intensity (895 )

0.004 0.004

0.003 0.003

● TM-1 ● ○ TM-1 im ○ im

0.002 0.002

0.001 0.001

0.000 0.000 5 5

10 10

15 15

20 20

25 25

30 30

DPA (days) DPA (days)

35 35

40 40

45 45

50 50

Figure 12. Integrated intensity of the 895 cm−1 band against DPA from normalized ATR FT-IR spectra Figure 12. 12. Integrated Integrated intensity intensity of of the the 895 895 cm cm−−11band against DPA from normalized ATR FT-IR FT-IR spectra spectra Figure of developmental TM-1 (●) and im (○) fibers. band against DPA from of developmental developmental TM-1 TM-1 ((●) and im im (#) (○) fibers. of ) and

-1) -1 Integrated Intensity (895 cmcm Integrated Intensity (895 )

0.004 0.004

Polynomial regression regression + 0.0008 y =Polynomial 2E-08x2 + 4E-05x 2 + 4E-05x + 0.0008 y = 2E-08xR² = 0.88 R² = 0.88

0.003 0.003

0.002 0.002

Linear regression regression yLinear = 3E-05x + 0.0002 y = 3E-05x + 0.0002 R² = 0.95 R² = 0.95

0.001 0.001

0.000 0.000 0 0

10 10

20 20

30

40

50

60

70

80

90

100

30 40 50 60 (%) 70 80 90 100 Cellulose Content Cellulose Content (%) Figure 13. 13. Integrated Integratedintensity intensityof ofthe the895 895cm cm−−11 band of developmental developmental TM-1 TM-1 and and Figure band vs. vs. cellulose cellulose content content of Figure 13. Integrated intensity of the 895 cm−1 band vs. cellulose content of developmental TM-1 and im fibers. fibers. Linear Linear and and polynomial polynomial(with (withthe theorder orderof of2) 2)regressions regressionswere wereapplied appliedtotoall allfiber fibersamples. samples. im im fibers. Linear and polynomial (with the order of 2) regressions were applied to all fiber samples.

3.3.5. Integrated Integrated Intensity Intensity of of the the 664 664 cm cm−−11Band 3.3.5. Bandand andthe theCorrelation Correlationwith withCellulose CelluloseContent Content 3.3.5. Integrated Intensity of the 664 cm−1 Band and the Correlation with Cellulose Content The band bandat at664 664cm cm−−11 has bending mode mode [13,14]. [13,14]. Its Its integrated integrated The has been been ascribed ascribed to to OH OH out-of-plane out-of-plane bending The band at 664 cm−1 has been ascribed to OH out-of-plane bending mode [13,14]. Its integrated intensity had had aa good good linear linear relationship relationship with with cellulose cellulose content content [13] [13] and and the the percentage percentage of of crystallinity crystallinity intensity intensity had a good linear relationship with cellulose content [13] and the percentage of crystallinity in cotton cotton fibers fibers [14]. [14]. The The trend trend in in Figure Figure 14 14 is is much much similar similar to to the the PC1 PC1 score score of of PCA PCA result result in in Figure Figure 44 in in cotton fibers [14]. The trend in Figure 14 is much similar to the PC1 score of PCA result in Figure 4 and the IR Ratio 1372/2900 in Figure 10, and also agrees with that for other cotton varieties [13]. and the IR Ratio 1372/2900 in Figure 10, and also agrees with that for other cotton varieties [13]. and the IR Ratio 1372/2900 in Figure 10, and also agrees with that for other cotton varieties [13]. Figure 15, being dissimilar from the patterns in Figures 7, 9 and 13, reveals a similar tendency to Figure 15, being dissimilar from the patterns in Figures 7, 9 and 13, reveals a similar tendency to that which is in Figures 5 and 11, when relating the integrated intensity of the 664 cm−1 band to that which is in Figures 5 and 11, when relating the integrated intensity of the 664 cm−1 band to cellulose content. cellulose content.

Sensors 2017, 17, 1469

11 of 13

Figure 15, being dissimilar from the patterns in Figures 7, 9 and 13, reveals a similar tendency to that which1469 is in Figures 5 and 11, when relating the integrated intensity of the 664 cm−1 band to Sensors 11 Sensors 2017, 2017, 17, 17, 1469 11 of of 13 13 cellulose content.

IntergratedIntensity Intensity(664 (664cm cm-1-1)) Intergrated

0.010 0.010

●● TM-1 TM-1 ○○ im im

0.008 0.008

0.006 0.006

0.004 0.004

0.002 0.002

0.000 0.000 55

10 10

15 15

20 20

25 25

30 30

35 35

40 40

45 45

50 50

DPA DPA (days) (days) Figure 14. intensity of the 664 cm against DPA from normalized ATR spectra Figure14. 14. Integrated Integratedintensity intensityof ofthe the664 664cm cm−−11band band ATR FT-IR FT-IR spectra spectra Figure Integrated band against against DPA DPA from from normalized normalized ATR FT-IR of developmental TM-1 (●) and im (○) fibers. ofdevelopmental developmentalTM-1 TM-1((●) andim im(#) (○) fibers. fibers. of ) and −1

IntegratedIntensity Intensity(664 (664cm cm-1-1)) Integrated

0.010 0.010

Polynomial Polynomial regression regression 2 yy == 2E-08x 4E-05x ++ 0.0008 0.0008 2E-08x2 ++ 4E-05x R² R² == 0.88 0.88

0.008 0.008

0.006 0.006

Linear Linear regression regression yy == 8E-05x 8E-05x ++ 0.0013 0.0013 R² R² == 0.90 0.90

0.004 0.004

0.002 0.002

0.000 0.000 00

20 20

40 40

60 60

80 80

100 100

Cellulose Cellulose Content Content (%) (%) −1 Figure 15. Integrated intensity of the 664 cm TM-1 and Figure band vs. vs. cellulose cellulose content content of of developmental developmental Figure15. 15. Integrated Integratedintensity intensityof ofthe the664 664cm cm−−11 band developmental TM-1 TM-1and and im fibers. Linear and polynomial (with the order of 2) regressions were applied to all fiber samples. im fibers. Linear and polynomial (withthe theorder orderof of2) 2)regressions regressionswere wereapplied appliedtotoall allfiber fibersamples. samples.

4. 4. Conclusions Conclusions 4. Conclusions The present study demonstrates FT-IR spectroscopy, in conjunction The present present study study demonstrates demonstrates the the great great potential potential of of ATR ATR FT-IR FT-IR spectroscopy, spectroscopy,in inconjunction conjunction The the great potential of ATR with simple algorithm analysis, in sensing the cellulose deposition during cotton fiber development. with simple algorithm analysis, in sensing the cellulose deposition during cotton fiber development. with simple algorithm analysis, in sensing the cellulose deposition Upon comparing the correlations between cellulose content determined by assay analysis and the Upon comparing comparing the the correlations correlations between between cellulose cellulose content content determined determined by by assay assay analysis analysis and and the the Upon ATR FT-IR spectral response acquired by multivariate PCA and the proposed simple algorithms, itit ATRFT-IR FT-IRspectral spectralresponse responseacquired acquiredby bymultivariate multivariatePCA PCAand andthe theproposed proposedsimple simplealgorithms, algorithms, it ATR −1 −1 − 1 has become evident that the R value, CI IR , and the integrated intensity of the 895 cm band can be has become evident that the R value, CI IR , and the integrated intensity of the 895 cm band can be has become evident that the R value, CIIR , and the integrated intensity of the 895 cm band can be utilized to estimate cotton fiber cellulose content, maturity in rapid, routine, and utilized to to estimate estimate cotton cotton fiber fiber cellulose cellulose content, content, maturity maturity and and crystallinity, crystallinity, in in aaa rapid, rapid, routine, routine, and and utilized and crystallinity, non-destructive manner. non-destructive manner. non-destructive manner. Acknowledgments: The authors thank Tracy Condon of in Acknowledgments: thank Tracy Condon of USDA-ARS-SRRC for the for technical assistanceassistance in collecting Acknowledgments:The Theauthors authors thank Tracy Condon of USDA-ARS-SRRC USDA-ARS-SRRC for the the technical technical assistance in the experimental samples. Mention of a product or of specific equipment does notequipment constitute adoes guarantee or warrantyaa collecting the samples. Mention or not collecting the experimental experimental samples. Mention of aa product product or specific specific equipment does not constitute constitute guarantee guarantee or or warranty warranty by by the the U.S. U.S. Department Department of of Agriculture Agriculture and and does does not not imply imply its its approval approval to to the the exclusion exclusion of of other other products products that that may may also also be be suitable. suitable. Author Author Contributions: Contributions: Y. Y. Liu Liu conceived conceived the the research, research, collected collected the the ATR ATR FT-IR FT-IR spectra, spectra, interpreted interpreted the the results, results, and wrote the manuscript. H.J. Kim conceived the research, coordinated the fiber chemical/property and wrote the manuscript. H.J. Kim conceived the research, coordinated the fiber chemical/property analysis, analysis, analyzed analyzed the the results results and and co-wrote co-wrote the the manuscript. manuscript. All All authors authors read read and and approved approved the the final final manuscript. manuscript.

Sensors 2017, 17, 1469

12 of 13

by the U.S. Department of Agriculture and does not imply its approval to the exclusion of other products that may also be suitable. Author Contributions: Y. Liu conceived the research, collected the ATR FT-IR spectra, interpreted the results, and wrote the manuscript. H.J. Kim conceived the research, coordinated the fiber chemical/property analysis, analyzed the results and co-wrote the manuscript. All authors read and approved the final manuscript. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2.

3.

4.

5. 6. 7. 8. 9.

10.

11. 12. 13. 14. 15. 16.

17.

Robertson, W.C.; Roberts, B.A. Integrated Crop Management for Cotton Production in the 21st Century. In Cotton: Technology for the 21st Century; Wakelyn, P.J., Chaudhry, M.R., Eds.; International Cotton Advisory Committee: Washington, DC, USA, 2010; pp. 63–97. Wakelyn, P.J.; Bertoniere, N.R.; French, A.D.; Thibodeaux, D.P.; Triplett, B.A.; Rousselle, A.A.; Goynes, W.R., Jr.; Edwards, J.V.; Hunter, L.; Mcalister, D.D.; et al. (Eds.) Biosynthesis of Cotton. In Cotton Fiber Chemistry and Technology; CRC Press: Boca Raton, FL, USA, 2007; Chapter 2; pp. 11–14. Kim, H.J. Fiber Biology. In Cotton, 2nd ed.; Fang, D.D., Percy, R.G., Eds.; American Society of Agronomy, Inc.; Crop Science Society of America, Inc.; Soil Science Society of America, Inc.: Madison, WI, USA, 2015; pp. 97–127. Frydrych, I.; Thibodeaux, D.P. Fiber quality evaluation-current and future trends/intrinsic value of fiber quality in cotton. In Cotton: Technology for the 21st Century; Wakelyn, P.J., Chaudhry, M.R., Eds.; International Cotton Advisory Committee: Washington, DC, USA, 2010; pp. 251–295. Hsieh, Y.L. Chemical Structure and Properties of Cotton. In Cotton: Science and Technology; Gordon, S., Hsieh, Y.L., Eds.; Woodhead Publishing Limited: Cambridge, UK, 2007; pp. 3–34. Meinert, M.C.; Delmer, D.P. Changes in biochemical composition of the cell wall of the cotton fiber during development. Plant Physiol. 1977, 59, 1088–1097. [CrossRef] [PubMed] Huwyler, H.R.; Franz, G.; Meier, H. Changes in the composition of cotton fiber cell walls during development. Planta 1979, 146, 635–642. [CrossRef] [PubMed] Abidi, N.; Hequet, E.; Cabrales, L. Changes in sugar composition and cellulose content during the secondary cell wall biogenesis in cotton fibers. Cellulose 2010, 17, 153–160. [CrossRef] Wang, C.; Lv, Y.; Xu, W.; Zhang, T.; Guo, W. Aberrant phenotype and transcriptome expression during fiber cell wall thickening caused by the mutation of the Im gene in immature fiber (im) mutant in Gossypium hirsutum L. BMC Genom. 2014, 15, 94. [CrossRef] [PubMed] Abidi, N.; Hequet, E.; Cabrales, L.; Gannaway, J.; Wilkins, T.; Wells, L.W. Evaluating cell wall structure and composition of developing cotton fibers using Fourier transform infrared spectroscopy and thermogravimetric analysis. J. Appl. Phys. 2008, 107, 476–486. [CrossRef] Abidi, N.; Cabrales, L.; Hequet, E. Fourier transform infrared spectroscopic approach to the study of the secondary cell wall development in cotton fiber. Cellulose 2010, 17, 309–320. [CrossRef] Hartzell-Lawson, M.M.; Hsieh, Y.-L. Characterizing the noncellulosics in developing cotton fibers. Text. Res. J. 2010, 70, 810–819. [CrossRef] Abidi, N.; Cabrales, L.; Haigler, C.H. Changes in the cell wall and cellulose content of developing cotton fibers investigated by FTIR spectroscopy. Carbohydr. Polym. 2014, 100, 9–16. [CrossRef] [PubMed] Abidi, N.; Manike, M. X-ray diffraction and FTIR investigations of cellulose deposition during cotton fiber development. Text. Res. J. 2017. [CrossRef] Liu, Y.; Thibodeaux, D.; Gamble, G. Development of Fourier transform infrared spectroscopy in direct, non-destructive, and rapid determination of cotton fiber maturity. Text. Res. J. 2011, 81, 1559–1567. Liu, Y.; Thibodeaux, D.; Gamble, G.; Bauer, P.; VanDerveer, D. Comparative investigation of Fourier transform infrared (FT-IR) spectroscopy and X-ray diffraction (XRD) in the determination of cotton fiber crystallinity. Appl. Spectrosc. 2012, 66, 983–986. [CrossRef] [PubMed] Liu, Y.; Kim, H.J. Use of attenuated total reflection Fourier transform infrared (ATR FT-IR) spectroscopy in direct, non-destructive, and rapid assessment of developmental cotton fibers grown in planta and in culture. Appl. Spectrosc. 2015, 69, 1004–1010. [CrossRef] [PubMed]

Sensors 2017, 17, 1469

18.

19. 20.

21.

22.

13 of 13

Liu, Y.; Kim, H.J. Characterization of developmental immature fiber (im) mutant and Texas Marker-1 (TM-1) cotton fibers by attenuated total reflection Fourier transform infrared (ATR FT-IR) spectroscopy. Appl. Spectrosc. 2017. [CrossRef] [PubMed] Santiago, C.M.; Hinchliffe, D.J. FT-IR examination of the development of secondary cell wall in cotton fibers. Fibers 2015, 3, 30–40. Lee, C.M.; Kafle, K.; Belias, D.W.; Park, Y.B.; Glick, R.E.; Haigler, C.H.; Kim, S.H. Comprehensive analysis of cellulose content, crystallinity, and lateral packing in Gossypium hirsutum and Gossypium barbadense cotton fibers using sum frequency generation, infrared and Raman spectroscopy, and X-ray diffraction. Cellulose 2015, 22, 971–989. [CrossRef] Nelson, M.L.; O’Connor, R.T. Relation of certain infrared bands to cellulose crystallinity and crystal lattice type. Part II. A new infrared ratio for estimation of crystallinity in cellulose I and II. J. Appl. Polym. Sci. 1964, 8, 1325–1341. [CrossRef] Updegraff, D.M. Semimicro determination of cellulose in biological materials. Anal. Biochem. 1969, 32, 420–424. [CrossRef] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).