# Fileset

[MINAMI_ISOEN2024_fin.pdf](https://mdr.nims.go.jp/filesets/ae9b75b9-dbad-4cc6-ad3f-878328d9af60/download)

## Creator

[Kosuke Minami](https://orcid.org/0000-0003-4145-1118), Masaaki Matoba, Ryoh Nakakubo, [Genki Yoshikawa](https://orcid.org/0000-0002-9136-8964)

## Rights

© 2024 IEEE.  Personal use of this material is permitted.  Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Detection of acetone in milk through odor towards monitoring of ketosis in dairy cows](https://mdr.nims.go.jp/datasets/4c1badf4-652b-41fb-bfda-5e211269312c)

## Fulltext

Detection of acetone in milk through odortowards monitoring of ketosis in dairy cowsKosuke Minami,†,* Masaaki Matoba,† Ryoh Nakakubo,‡ and Genki Yoshikawa†,§†Research Center for Macromolecules and Biomaterials, National Institute for Materials Science (NIMS), Tsukuba 305-0044 Japan‡Institute of Livestock and Grassland Science, NARO, Tsukuba 305-0901 Japan§Graduate School of Pure and Applied Science, University of Tsukuba, Tsukuba 305-8571 Japan*Email: MINAMI.Kosuke@nims.go.jp; ORCiD: 0000-0003-4145-1118Abstract—Determining subclinical ketosis of dairy cows is ofgreat importance in terms of economic losses because ketoticcows even in subclinical ketosis reduce milk production. Whendairy cows are affected by ketosis, they produce high level ofketone bodies including acetone, therefore, the detection of suchketone bodies is useful for diagnosis of ketosis. In this study, wedemonstrated the detection of acetone in milk samples in therange from 1 to 10 ppm in the vapor phase, which correspondsto the acetone level of subclinical ketosis in a milk sample, usinga nanomechanical Membrane-type Surface stress Sensor (MSS).It was proved that the signal output clearly correlates to thevapor concentration of acetone with ± 0.40 ppm root-mean-square deviation (RMSD). The present study provides a potentialsensor for non-invasive monitoring of ketosis through milk odor.Index Terms—milk, acetone, nanomechanical sensors,Membrane-type Surface stress Sensor (MSS), dairy cows, ketosisI. INTRODUCTIONDairy cows often suffer from ketosis due to dramaticchanges in energy balance before and after calving. Ketosisin dairy cows reduces milk production [1]–[3]. In addition,subclinical ketosis, which is the elevated level of blood ketonebodies (i.e., β-hydroxybutyric acid (BHB), acetoacetic acid,and acetone) in the absence of clinical signs of ketosis, hasbecome a problem in recent farm practice [4] because dairycows in subclinical ketosis have been reported to produce lessmilk than healthy cows [5].To diagnose ketosis, blood BHB levels have been usedbecause measurements of BHB are stable and reproduciblecompared to other ketone bodies [3], [6]–[8]. However, bloodtests require veterinary investigation and are expensive andtime-consuming, making it difficult to quickly link the diag-nosis of subclinical ketosis to treatment. Moreover, the bloodtests are invasive and difficult to perform daily from an animalwelfare perspective. For rapid on-site diagnosis, colorimetricmethods such as test papers have been developed, but they arestill difficult to use for daily monitoring of ketosis on the dairyfarm because of cost and milk sampling issues.In this study, we demonstrate that a nanomechanical sensorcan be utilized for detecting acetone in the range of subclinicalketosis through milk odor. Since BHB and acetoacetic acidare less volatile, we focus on acetone as an indicator ofFig. 1. Schematic illustrations of MSS and its sensing system. (A) Structureof MSS. (B) Photo of whole cow’s milk samples. (C) Sensing system.ketosis for monitoring ketotic cows. We used a nanome-chanical Membrane-type Surface stress Sensor (MSS) as asensing platform (Fig. 1A) [9]. By focusing on the purgingprocess, the acetone can be detected within a couple ofminutes at the concentrations ranging from 1 to 10 ppm, whichcorresponds to the level of subclinical ketosis milk samples.Since the nanomechanical sensors including MSS have variousadvantages on rapid and on-site measurements [10], this studyprovides a promising platform towards daily monitoring ofclinical and subclinical ketosis in dairy cows.II. EXPERIEMENTALA. Preparation of MSSIn this study, we used an MSS as a sensing unit becauseof the high robustness and sensitivity (Fig. 1A) [9]. Theconstruction of the MSS chips and its working principle havebeen previously reported [9]–[11]. Briefly, MSS consists ofa silicon-based membrane suspended by four piezoresistivebeams, composing a full Wheatstone bridge. The membraneis coated with a receptor material, which generates the surfacestress caused by the sorption-induced expansion. The surfacestress on the membrane is transduced to the four sensingbeams as amplified uniaxial stresses, resulting in the changesin the electrical resistance of the piezoresistors embedded inthe beams. The output signal (Vout) can be expressed as [9]Vout =VB4(∆R1R1− ∆R2R2+∆R3R3− ∆R4R4), (1)where VB is the bridge voltage applied to the Wheatstonebridge circuit and ∆Ri/Ri (i = 1–4) is the relative resistancechange in each sensing beam.The MSS chip was provided from Asahi Kasei Corporation.AK02-04, which was provided by Asahi Kasei Corporation,and Tenax [12] were used as receptor layers. Each receptormaterial was coated directly onto the MSS membrane usingan inkjet spotter according to the previous reports [12]–[14].B. Sample preparationTo investigate the different concentrations of acetone inwater and commercial whole cow’s milk, aliquot of acetonewas added to MilliQ water (Merck Millipore) or whole cow’smilk (Milk Y, Yotsuba Milk Products Co., Ltd., Hokkaido,Japan; Milk S, Megmilk Snow Bland Co., Ltd., Kanagawa,Japan; Milk M, Meiji Co., Ltd., Tochigi, Japan) (Fig. 1B)at the concentrations of 0.1, 0.2, 0.3, 0.4, and 0.5 mM. Thecommercial whole milk samples have been pasteurized at least120 °C for 2 s. Averaged nutritional compositions are listedin Table I.To determine the acetone concentration in the vapor phase at25 °C, proton transfer reaction-time of flight-mass spectrometr(PTR-TOF-MS; PTR-TOF 6000 X2, Ionicon Analytik GmbH)equipped with Static Headspace Autosampler was used. Thedetailed measurement conditions are described in the previousreport [12]. The mass spectrum was recorded in the mass rangeof m/z = 9–400, and mass calibration was performed usingtwo ion peaks of known exact masses, i.e., hydronium ionisotope (H318O+; m/z = 21.022) and diiodobenzene fragment(C6H4I2H+; m/z = 203.943). The measured concentrations aresummarized in Table II.TABLE INUTRITIONAL COMPOSITIONS OF WHOLE COW’S MILK SAMPLESUSED IN THIS STUDY.Composition per 100 mL a Whole cow’s milkMilk Y b Milk S b Milk M bNon-fat solids ≥8.5% ≥8.4% ≥8.3%Fat ≥3.7% ≥3.7% ≥3.5%Energy (kJ) 293 287 287Protein (g) 3.50 3.40 3.40Fat (g) 4.05 3.90 3.90Carbohydrate (g) 4.85 4.95 4.95Dietary fibre (g) — 0 —Salt (mg) 105 100 110Calcium (mg) 117 114 114a Average value through a year provided by each company.b Milk Y, Yotsuba; Milk S, Megmilk Snow Bland; Milk M, Meiji.Fig. 2. Evaluation of the acetone concentration range in a subclinical ketoticcow. (A) Photos of colorimetric BHB test papers (Sanketo paper) of healthy,subclinical, and clinical ketosis cows. (B) Results of PTR-TOF-MS. Vaporconcentration of acetone of milk samples from healthy, subclinical ketosis,and clinical ketosis cows. H, healthy; S, subclinical; C, clinical.C. SensingThe MSS chip was placed in an MSS standard measure-ment module produced by the industry-academia-governmentframework called “MSS Aliance” [15]–[17]. The MSS moduleconsists of a Teflon chamber, an aspiration pump, and aswitching valve with sampling and purging lines (Fig. 1C). Allmeasurements by the module were performed in an incubatorkept at 25.00 ± 0.02 °C. The headspace vapor in a 20 mLsample vial, in which 10 mL of each sample was placed, wasflowed by mass flow controller 1 (MFC-1) with flow rate at 30mL min–1. The flow rate of the aspiration pump was set at 20mL min–1. Before measuring MSS signals, pure nitrogen gaswas introduced into the module for at least 2 min to recover theoriginal baseline. Subsequently, the valve was switched to thesampling line for 1 min and then switched back to the purgingline for 4 min. Sensing signals of MSS were measured at thebridge voltage (VB) of –3.0 V and recorded at a sampling rateof 10 Hz.III. RESULTS AND DISCUSSIONTo evaluate the range of vapor phase concentration ofacetone in milk of subclinical ketosis cows, we measuredthe milk samples collected from healthy, subclinical ketosis,and clinical ketosis cows. Dairy cows were examined forketosis based on BHB as an indicator using a test paper(Fig. 2A). The concentrations of acetone in the vapor phaseTABLE IICONCENTRATIONS OF ACETONE IN WATER AND WHOLE MILK SAMPLESDETERMINED BY PTR-TOF-MS.Concentration Concentration in vapor phase [ppm]in liquid [mM] a water Milk Y b Milk S b Milk M b0.00 0.05 0.58 0.60 0.490.10 1.13 1.72 1.86 1.700.20 2.22 2.85 3.18 2.980.30 3.38 4.04 4.47 4.220.40 4.48 5.17 5.80 5.590.50 5.59 6.34 7.08 6.95a Concentration of acetone added to water or milk samples.b Milk Y, Yotsuba; Milk S, Megmilk Snow Bland; Milk M, Meiji.Fig. 3. MSS signal outputs. (A) Signal responses of MSS to aqueous solutionsof acetone. (B) Magnified signal responses to aqueous solutions of acetonewith different concentrations in the purging process at around 67 s. (C)Signal responses to acetone-Milk M mixtures. (D) Magnified signal responsesto acetone-Milk M mixtures with different concentrations of acetone in thepurging process at around 67 s.of milk samples were determined by PTR-TOF-MS. The ionsof acetone (C3H6OH+, m/z = 59.0491; C213CH6OH+, m/z =60.0525) were measured (Fig. 2B). Milk samples from healthydairy cows contained approx. 1 ppm of acetone in the vaporphase, while one from ketotic cows contained more than 10ppm of acetone. In contrast, milk samples from subclinicalketosis dairy cows contained approx. 5 ppm of acetone in thevapor phase. Therefore, acetone should be monitored in therange of 1 to 5 ppm to prevent the development of ketosis.To demonstrate the detection of acetone through odor ofmilk samples, we prepared milk samples with varying con-centration ranging from 0.10 to 0.50 mM by adding aliquotsof acetone into whole cow’s milk as well as into water as areference. The commercial whole milk samples contain someamounts of acetone (approx. 0.5 ppm as shown in TableII). In addition to various other components that may affectthe volatility of acetone, we purchased three milk samplesfrom three different companies (Fig. 1B and Table I; see alsoExperiemental section). The acetone concentration in the vaporphase was measured by PTR-TOF-MS and the results aresummarized in Table II. The same samples were then measuredusing MSS.Fig. 3 shows the signal responses to the aqueous solutionsof acetone and the milk samples. When gas molecules areabsorbed into a receptor layer, the receptor layer generatessurface stress caused by the sorption-induced expansion [9],[18]–[20]. Since all samples, including milk samples, areaqueous solutions, sample vapors contain a large amount ofFig. 4. Plots of signal outputs at 67 s obtained from the whole milk samplesand the aqueous solutions with different concentrations of acetone added asa function of vapor concentration of acetone measured by PTR-TOF-MS.water, resulting that the signal responses are dominated by theresponse to water (Figs. 3A and C). As previously reported[12], water molecules tend to desorb more quickly than othercomponents including acetone, and hence a clear concentrationdependence of a target analyte can often be observed in thepurging process. As shown in Figs. 3B and D, the increasein signal output was observed in the purging process as theacetone concentration increased.To evaluate the concentration dependence, we then extractedthe signal outputs at 7 s after starting the purging process(i.e., 67 s) and plotted them as a function of the vaporphase concentrations of acetone determined by PTR-TOF-MS (Table II). Although the receptor materials have corss-reactivity to various odorous molecules, major volatile organiccompounds in the whole cow’s milk are water and acetonedetermined by PTR-TOF-MS. Therefore, as shown in Fig. 4,the signal outputs show a linear correlation with acetoneconcentrations. The root-mean-square deviation (RMSD) ofacetone concentration is ± 0.40 ppm and the limit of detection,which is defined as the concentration with a signal-to-noiseratio of 3 [21], is 0.46 ppm, demonstrating the possibility topredict the acetone concentration level of subclinical ketosiscows through the odor of milk samples by nanomechanicalsensors.IV. CONCLUSIONWe have demonstrated that a nanomechanical sensor, MSS,can detect acetone through odor of milk at the level ofsubclinical ketosis in dairy cows. Focusing on the desorptionrates between water and acetone molecules from the receptorlayer, the concentration-dependent measurement of acetone inthe range from 1 to 10 ppm was achieved with the RMSD of± 0.40 ppm. This study presents a great potential for non-invasive monitoring of subclinical ketosis, leading to dailymonitoring using automated milking systems.ACKNOWLEDGMENTSWe thank Takahide Farm for providing milk samples fromhealthy, subclinical ketosis, and clinical ketosis dairy cows.We thank Asahi Kasei Co., Ltd. for providing an MSS chipand a receptor material.This work is financially supported by a Grant-in-Aid forScientific Research (A), JSPS, MEXT, Japan (No. 18H04168);a Grant-in-Aid for Scientific Research (C), JSPS, MEXT,Japan (No. 22K05324); a Grant-in-Aid for Challenging Re-search (Exploratory), JSPS, MEXT, Japan (No. 21K18859);the Public/Private R&D Investment Strategic Expansion Pro-gram (PRISM), Cabinet Office, Japan; and Center for Func-tional Sensor & Actuator (CFSN), NIMS.REFERENCES[1] D. E. Bauman and W. B. Currie, ”Partitioning of nutrients duringpregnancy and lactation: a review of mechanisms involving homeostasisand homeorhesis,” J. Dairy Sci., vol. 63, pp. 1514–1529, September1980. DOI: 10.3168/jds.S0022-0302(80)83111-0[2] G. D. Baird, ”Primary ketosis in the high-producing dairy cow: clin-ical and subclinical disorders, treatment, prevention, and outlook,” J.Dairy Sci., vol. 65, pp. 1–10, January 1982. DOI: 10.3168/jds.S0022-0302(82)82146-2[3] T. H. Herdt, ”Ruminant adaptation to negative energy balance: influenceon the etiology of ketosis and fatty liver,” Vet. Clin. North Am. FoodAnim. Pract., vol. 16, pp. 215–230, July 2000. DOI: 10.1016/S0749-0720(15)30102-X[4] L. Andersson, ”Subclinical ketosis in dairy cows,” Vet. Clin. North Am.Food Anim. Pract., vol. 4, pp. 233–251, July 1998. DOI: 10.1016/S0749-0720(15)31046-X[5] J. A. McArt, D. V. Nydam, and G. R. Oetzel, ”Epidemiology ofsubclinical ketosis in early lactation dairy cattle,” J. Dairy Sci., vol.95, pp. 5056–5066, September 2012. DOI: 10.3168/jds.2012-5443[6] G. R. Oetzel, ”Monitoring and testing dairy herds for metabolic dis-ease,” Vet. Clin. North Am. Food Anim. Pract., vol. 20, pp. 651–674,November 2004. DOI: 10.1016/j.cvfa.2004.06.006[7] T. Duffield, ”Subclinical ketosis in lactating dairy cattle,” Vet. Clin.North Am. Food Anim. Pract., vol. 16, pp. 231–253, July 2000. DOI:10.1016/S0749-0720(15)30103-1[8] A. Leduc, S. Souchet, M. Gelé, F. Le Provost, and M. Boutinaud, ”Effectof feed restriction on dairy cow milk production: a review,” J. AnimalSci., vol. 99, pp. 1–12, June 2021. DOI: 10.1093/jas/skab130[9] G. Yoshikawa, T. Akiyama, S. Gautsch, P. Vettiger, and H. Rohrer,“Nanomechanical membrane-type surface stress sensor,” Nano Lett., vol.11, pp. 1044–1048, February 2011. DOI: 10.1021/nl103901a[10] K. Minami, G. Imamura, R. Tamura, K. Shiba, and G. Yoshikawa, “Re-cent advances in nanomechanical membrane-type surface stress sensorstowards artificial olfaction,” Biosensors, vol. 12, p. 762, September 2023.DOI: 10.3390/bios12090762[11] G. Yoshikawa, T. Akiyama, F. Loizeau, K. Shiba, S. Gautsch, T.Nakayama, P. Vettiger, N. F. de Rooij, and M. Aono, “Two dimensionalarray of piezoresistive nanomechanical membrane-type surface stresssensor (MSS) with improved sensitivity,” Sensors, vol. 12, pp. 15873–15887, November 2012. DOI: 10.3390/s121115873[12] K. Minami, H. Kobayashi, M. Matoba, Y. Kamiya, S. Maji, T. Nemoto,M. Tohno, R. Nakakubo, and G. Yoshikawa, “Measurement of volatilefatty acids in silage through odors with nanomechanical sensors,”Biosensors, vol. 13, p. 152, January 2023. DOI: 10.3390/bios13020152[13] K. Minami, K. Shiba, and G. Yoshikawa, “Discrimination of structurallysimilar odorous molecules with various concentrations by using ananomechanical sensor,” Anal. Methods, vol. 10, pp. 3720–3726, June2018. DOI: 10.1039/c8ay01224e[14] K. Minami, G. Imamura, T. Nemoto, K. Shiba, and G. Yoshikawa,“Pattern recognition of solid materials by multiple probe gases,” Mater.Horiz., vol. 6, pp. 580–586, December 2019. DOI: 10.1039/c8mh01169a[15] A. Yaegashi, G. Yoshikawa, T. Yokota, and Y. Aoki, ”Industry-academia-government joint research framework towards defactor standard forolfactory IoT sensors/systems,” J. Jpn. Assoc. Odor. Environ., vol. 49,pp. 323–327, September 2018. DOI: 10.2171/jao.49.323[16] K. Inada, H. Kojima, Y. Cho-Isoda, R. Tamura, G. Imamura, K. Minami,T. Nemoto, and G. Yoshikawa, ”Statistical evaluation of total expiratorybreath samples collected throughout a year: reproducibility and appli-cability toward olfactory sensor-based breath diagnostics,” Sensors, vol.21, p. 4742, July 2021. DOI: 10.3390/s21144742[17] Y. Saeki, N. Maki, T. Nemoto, K. Inada, K. Minami, R. Tamura, G.Imamura, Y. Cho-Isoda, S. Kitazawa, H. Kojima, G. Yoshikawa, and Y.Sato ”Lung cancer detection in perioperative patients’ exhaled breathwith nanomechanical sensor array” Lung Cancer, vol. 190, p. 107514,February 2024. DOI: 10.1016/j.lungcan.2024.107514[18] K. Minami, K. Shiba, and G. Yoshikawa, ”Sorption-induced static modenanomechanical sensing with viscoelastic receptor layers for multistepinjection-purge cycles,” J. Appl. Phys., vol. 129, p. 124503, March 2021.DOI: 10.1063/5.0039045[19] K. Minami, K. Shiba, G. Imamura, and G. Yoshikawa, ”Identificationof gas species and their concentrations by using sorption kinetics ofviscoelastic film,” 2022 IEEE International Symposium on Olfactionand Electronic Nose (ISOEN), Aveiro, Portugal, pp. 1–4, May 2022.DOI: 10.1109/ISOEN54820.2022.9789682[20] K. Minami, Y. Zhou, G. Imamura, K. Shiba, and G. Yoshikawa, ”Sorp-tion kinetic parameters from nanomechanical sensing for discriminationof 2-nonenal from saturated aldehydes,” ACS Sens., vol. 9, pp. 689–698,February 2024. DOI: 10.1021/acssensors.3c01888[21] D. MacDougall and W. B. Crummett, ”Guidelines for data acquisitionand data quality evaluation in environmental chemistry,” Anal. Chem.,vol. 52, pp. 2242–2249, December 1980. DOI: 10.1021/ac50064a004