Search tips
Search criteria 


Logo of chromsciLink to Publisher's site
J Chromatogr Sci. 2016 April; 54(4): 523–530.
Published online 2015 December 11. doi:  10.1093/chromsci/bmv176
PMCID: PMC4885384
Editor's choice

A Rapidly Fabricated Microfluidic Chip for Cell Culture


Microfluidic chips (μFC) are emerging as powerful tools in chemistry, biochemistry, nanotechnology and biotechnology. The microscale size, possibility of integration and high-throughput present huge technical potential to facilitate the research of cell behavior by creating in vivo-like microenvironments. Here, we have developed a new method for rapid fabrication of μFC with Norland Optical Adhesive 81 (NOA81) for multiple cell culture with high efficiency. The proposed method is more suitable for the early structure exploration stage of μFC than existing procedures since no templates are needed and fast fabrication methods are presented. Simple PDMS-NOA81-linked microvalves were embedded in the μFC to control or block the fluid flow effectively, which significantly broadened the applications of μFC. Various types of cells were integrated into the chip and normal viabilities were maintained up to 1 week. Besides, concentration gradient was generated to investigate the cells in the μFC responded to drug stimulation. The cells appeared different in terms of shape and proliferation that strongly demonstrated the potential application of our μFC in online drug delivery. The high biocompatibility of NOA81 and its facile fabrication (μFC) promise its use in various cell analyses, such as cell–cell interactions or tissue engineering.


In recent years, new cell culturing techniques are highly significant for cell biology (1). As we know, traditional cell culture methods are relatively mature and mostly carried out in petri plates or flasks; however, the multistep operations increase the probability of contamination and abundant reagent consumption is an extravagant waste. Moreover, compared with the microenvironment of human body, the obtained results cannot truly reflect the biological characteristics of cells (2).

The development of microfluidic devices has the potential to revolutionize the whole processes for modern biology and chemistry laboratory. More and more attention has been paid to microfluidic devices due to their better analytical performance with integration, high-throughput and reduced amounts of reagents in the field of life-sciences, particularly in cell research (35). This economic and advanced technology provides extensive opportunities for various cell-related applications especially in making multidesign gradient generators superior for drug delivery (6, 7). Furthermore, microstructure can realize different biological models such as co-culture (8, 9) or observe the interactions between cells and the microenvironment (10, 11) or real-time monitor and record, each subtle change during cell culture (12).

During a quick survey of microfluidic chips (μFC) for cell researches, it can be found that μFC is mostly fabricated by polydimethylsiloxane (PDMS) because of its gas permeability and rapid prototyping. However, it must be noted that PDMS possesses poor adherent viabilities for mammalian cells. Furthermore, the permeable and hydrophobic features also limited its use for commercial purposes (13). The fabrication methods for PDMS μFC have been well developed (14) though abundant time is still needed to reproduce a template and even slight changes can influence the whole designs turn to be useless. Therefore, different materials and processing techniques have been continuously reported to make μFC fabrication more operable (15). For example, Ren et al. (16) have fabricated whole-Teflon μFCs with better advantages compared with PDMS and various biological cells have been cultured in the whole-Teflon channel with normal viabilities. Nowadays, a material–Norland Optical Adhesive 81 (NOA 81) gaining popularity with high optical and mechanical capabilities is employed to fabricate multifunctional μFC (1719). After exposure to ultraviolet light (UV light), the mercaptoacrylic ester can be rapidly cured by an inert material. While the fabrication methods still rely on photolithography and molding technique which seem complicated as ever and no helps with the changing designs of the μFC that cause wastage of templates (20, 21).

In this study, a new method was proposed for rapid fabrication of μFC with NOA81 in order to provide different cell cultures with a time-saving and less laborious method. The complicated flow in the multilayer structure with different channels could be easily and efficiently controlled by the internally embedded microvalves. Various cell lines were seeded into the prepared μFC with NOA 81 to investigate the biocompatibilities and survival rates. The obtained results will offer a template-free process to re-build the μFC with only modifications of the mask, which is particularly suitable for the early stage of structure study in μFC.

Materials and methods

Cells, materials and reagents

Human neuroblastoma cell line SH-SY5Y was obtained from Xuanwu Hospital (Beijing, China) as a gift; human glioblastoma multiform cell line U87, uterine cervical cancer cell line Hela, embryonic kidney cell line HEK 293t and colon carcinoma cell line HCT 116 were purchased from Peking Union Medical College (Beijing, China).

NOA81 was purchased from Norland Products (Cranbury, NJ). PDMS Sylgard 184 was purchased from Dow Corning (Midland, MI). Petri dishes were from Corning (Inc, Acton, MA).

Ethanol AR (Cat No: 64-17-5) and acetone AR (Cat No: 67-64-1) were purchased from Beijing Chemical Works (Beijing, China). The standard culture medium, Dulbecco's Modified Eagle's Medium (Gibco®, São Paulo, Brazil), Dulbecco's Modified Medium (HyClone™, Basingstoke, UK), fetal bovine serum (FBS, Gibco®) and 100 units mL−1 of penicillin and streptomycin (Gibco®) were used for different kinds of cells culturing. Trypsin–EDTA (0.25%) was purchased from Life Technology, USA. Polylysine (PLL, Cat No: P2100-10), 4′,6-diamidino-2-phenylindole (DAPI, Cat No: 28718-90-3) and propidium iodide (PI, Cat No: 25535-16-4) were purchased from Solarbio (Beijing, China). MTT was purchased from Genview (Houston, TX, USA; Cat No: JT343). DMSO and dopamine hydrochloride (DA, H8502) were purchased from Sigma (Cat No: D8428).

Fabrication of μFC

Photomaskers were drawn by Adobe Illustrator (AI) and printed into films for preparation. The number of photomaskers depended on the number of layers that were prepared to be fabricated. The fabrication steps are shown in Figure 1A where (1) a simplest three-layer μFC with PDMS cover plate was presented for long-term cell culture and (2) an illustration of PDMS-NOA 81-linked microvalves’ insert progress was also presented and detailed explanation has been given in “Simple microvalve fabrication” section. The bottom layer was fabricated on the glass slide directly. NOA 81 uncured liquid was pulled into a square of wire and exposed to the UV light with a photomasker upside. As a result, NOA81 glued under the photomaskers in the transparent place became solid. In contrast, the dark place remained liquid and could be removed by organic solvent (4 : 1 mixture of acetone and alcohol) easily. This mechanism helped the pattern of the photomasker to transform into the fabricated layer. The middle and third layers were fabricated on chrome plates so that the layers could be peeled off easily. The thickness of all the layers was determined by the thickness of wire. The exposure time depending on the volume of the glue and fineness of the photomaskers usually changed from 6 to 40 s. When the photomasker was peeled off after exposure, a little glue remained uncured on the layers, which formed stickiness, and it is the key element to help conglutinate between layers. The three layers were crosslinked by UV-light and bounded to the PDMS cover plate with plasma oxide. The inlet and outlet were connected with a polytetrafluoroethylene tube for cell seeding and nutrient exchanging through the transfer of the sticky NOA 81 glue into solid. Finally, the chip was placed onto the oven and cured at 50°C for at least 2 h to improve the mechanical strength and transparency by crosslinking.

Figure 1.
Sketches of NOA 81 μFC. (A) Fabrication steps for simplest basic μFC: (1) a three-layer μFC with a PDMS cover plate and (2) progress to insert the microvalves; (B) layered schematic representation of the μFC and (C) integral ...

Simple microvalve fabrication

Although NOA81 is a promising material for fabricating μFC, because of high mechanical strength and low drawing coefficient, there are few reports on microvalve fabrication and rare application (22). NOA81 and elastomer PDMS were combined to a simple microvalve that was embedded into μFC for different applications (23). The brief fabrication step is shown in Figure 1A-(2) and the principle is presented in Figure 2A. A PDMS membrane (spin coated on a silicon wafer and thickness 0.4 mm) was embedded into the NOA 81 chip by plasma oxide. There were two functions of the membrane: (1) it acted as the elastomer to respond to the screw pressure and block the fluid flow; and (2) it acted as the cover plate to steal the chamber. A little screw (diameter 1 mm) was inserted into the microfluidic chip and immobilized by the cured NOA 81. Figure 2B shows a photograph of μFC with seven embedded microvalves, and different food dye or water was filled into the channels to investigate its performance. For the first time, 500 μL/h fluid flew through the channel without any block. When the screw was twisted down, the PDMS elastomer was pushed down and blocked the corresponding fluid flow.

Figure 2.
Easily fabricated microvalves. (A) The theory of the simple microvalve; (B) μFC with seven microvalves and (C) cell culture chip with a microvalve (coin diameter 25 mm). This figure is available in black and white in print and in color at JCS ...

Cell culture

The culture medium contained 90% DMEM (for SH-SY5Y, Hela, Hek 293t and HCT 116) or MEM (for U87) and 10% FBS, and 100 units mL−1 penicillin and streptomycin were prepared for different kinds of cells culture. Multiple cell types were cultured separately at 37°C in a humidified incubator containing 5% CO2 for 2–3 days prior to microfluidic experiments. When cells became confluent, they were detached by trypsin–EDTA and centrifuged at 1,000 rpm for 5 min. All the experiments in μFC were carried out while the cells were in the exponential growth phase. The generation of SH-SY5Y with enhanced green fluorescence protein (EGFP) was as same as in the previous studies (24).

Material biocompatibility test

To characterize NOA81, a droplet of glue was dripped onto a potassium bromide slice for infrared analysis. A piece of UV-cured NOA81 material was also prepared for the contact angle test. The glue NOA81 was directly poured and cured in 24 well-plates with the same treatment as it was fabricated into chips but there was no pattern on the photomasker. As an example, SH-SY5Y was seeded into the plate with or without NOA81 (PDMS as a contrast) substrate at the same density and then incubated at 37°C with 5% CO2 incubator for 24 h. The cured NOA81 was cut into pieces and soaked into the culture medium with a percentage of 0.2 mg/mL for different periods. After collected, the impregnated liquid was added into the cells of the same density and incubated for 24 h. The material's toxicity to the SH-SY5Y cells was detected using the Thiazolyl blue (MTT) test.

Cell seeding and culture in μFCs

Before the chip was used, 75% disinfectant alcohol was introduced into the chip at the flow rate of 1,000 μL/h for 30 min, which avoided the influence of any remaining uncured glue on cell viability. Subsequently, the chip was treated with 100 μg/mL PLL for 5 min and incubated at 60°C for 1 h to improve cell adherence. In addition, before cell seeding, the whole chip was exposed to ultraviolet germicidal irradiation for 30 min to ensure a sterile environment. Then, phosphate-buffered saline (PBS) and the culture medium were introduced into the chip. Each cell type was digested and suspended from the culture dish and seeded into the chip separately at a density of 105–106 cells/mL. The chip was then placed into the incubator. After 2 h for cell settlement and adherence, a micropump (LongerPump, Baoding, China) began to work with a flow rate of 30 μL/h for 1 h and rest for 1 h alternatively for nutrient and metabolic waste exchange.

Cell staining and drug stimulation

The cell conditions in the chips were investigated via staining with DAPI and PI. The flow rate was set at 20 μL/min (10 min). PI (10 μg/mL) was first induced into the chip to stain the dead cells. Then, the cells were immobilized by paraformaldehyde. Ultimately, DAPI was injected for stain cell nucleus. During each experimental process just mentioned earlier, there was always a step for PBS washing for removing excess reagents. For the drug simulation session, after 8 h, when SH-SY5Y cells became adherent and stable under optical views, the medium containing dopamine (DA, 240 μM/L) was replaced and introduced into the chip at the same flow rate as used for cell culture. After 24 h, the cell was stained by PI and studied under a fluorescent inverted microscope.


Device fabrication

Figure 1A shows the progress of fabricating a multilayer μFC as described in the “Materials and methods” section. Cell culture chamber is 7 mm long × 5 mm wide, and the channel is 200 μm wide. Each layer is 200 μm thick. The chip contained a cell seeding inlet, a medium transport inlet and a waste conveying outlet. After cell seeding, the seeding inlet could be easily blocked by the microvalve (Figure 2C) or NOA81 cured glue (cured by UV-light). The advantage of the present method for fabricating μFC is its template-free development and it can be used to make a spatial and stereo structure in a short time by adding different layers together, which develop various types of 3D structures of chips and achieve numerous applications (Figures 1C, C,2B2B and C and Supplementary Material, Figure S1). The proposed method is highly suitable for the early structure exploration stage of μFC.

Function of the simple microvalve

Figure 2B delineates a μFC with seven microvalves. The channel was 500 μm wide and 200 μm high. A 2 × 2 mm chamber was fabricated for observation. By blocking selected pathways, different dyes passed through the observation chamber (Figure 3). One simplest test for the microvalve is to block the cell seedling inlet (Figure 2C) that efficiently prevents cross-contamination. With the combination of easily fabricated microvalves in μFC for further scenarios, the functions of basic channel closure and change in the direction of flow can be achieved, which might be used in micro-mixer, microreactor or other certain fields.

Figure 3.
Different dyes passing through the observation chamber during blockage of different pathways (the top channel was pink dye, the middle was water and the lower was blue dye). (A–C) show the sketch map, and crosses represent blocking the flow path ...

NOA81's feature assay

According to official and article records, NOA81 is a mercaptoester and its polymerization is based on thiol-ene crosslinking (25). Infrared spectroscopy was used as mentioned in Supplementary Material, Figure S2, to point out the changes after NOA81 was cured and baked. Prior studies were conducted to make sure about the effect of NOA81's components on the cell viability. For this purpose, some basic experiments were carried to investigate the toxicity and properties of NOA81. Compared with original culture dishes, the contact angle of NOA81 was 53° (Figure 4D) meaning the material was hydrophilic and approached by culture dish (46.6° not shown). When compared with PDMS substrate mentioned in Figure 4B, the cells on NOA81 substrate (Figure 4C) were more similar to normal culture dish (Figure 4A). SH-SY5Y on NOA 81 did not change in terms of proliferation or viability, which proves cells' viabilities on NOA81 in sufficient medium. This may further indicate that the material is suitable for cell culture. The MTT results (Figure 4E) indicated that as time passed, the cell maintained high survival rate compared with the control group which means, with proper medium exchange in μFC, the cell ability can be ensured as explained below.

Figure 4.
SH-SY5Ycells on different substrates (2 × 104/mL, 100×): (A) conventional culture dish substrate; (B) PDMS substrate; (C) NOA 81 substrate; (D) NOA 81's contact angle (53°) and (E) cell viability assay (MTT, 24 h) for NOA81 impregnated ...

Cell viability in chip

Different cell lines were used to observe their abilities in the chips. All cells in the chip had normal shape and favorable proliferation. Figure 5 shows that SH-SY5Y with EGFP in the chip under optical (Figure 5A–D) and fluorescent (Figure 5E, 3 day) microscopes (200×) with the same view (C) at different time periods (1d–4d). The cells maintained normal shape and high viability. With the passage of time, obvious proliferation was observed, which can be directly reflected by the picture in the line chart (Figure 5F). The standard error bar indicates the variation of five chip areas. The cell growth curve did not show obvious stationary phase or logarithmic phase as in the routine test. There are two reasons: first, in the regularly changed culture medium that maintained sufficient nutrients, cells entered at a rapid growth mode. Second, after the third day, SH-SY5Y grew too fast that limited space restricted its proliferation. SH-SY5Y cells in μFC kept good viability for more than 7 days (Supplementary Material, Figure S3). This may give convenient opportunities for long-term cell studies. The direct staining pictures on chips are also given. In a merged picture (Figure 6F), DAPI was blue whereas PI was only red in the nucleus of dead cells. According to the picture, a few dead cells were labeled red, which could also prove the cells' high viability on chip. Cells can be counted and tested with the help of online staining on chips. Furthermore, U87, Hela, HCT 116 and Hek 293t were also tested in the chip with the same procedures. The results demonstrated that all these animal adherent cell lines showed normal shape and favorable proliferation (Figure 6A–D).

Figure 5.
Different time periods, SH-SY5Y cells with EGFP are under optical (A–D, from 1 to 4 days) and fluorescent (E, 3 days) views in the same area. The crosses are designed to ensure that the microscope can collect the same area in different days. (F) ...
Figure 6.
Different cells on the third day (200×): (A) U87, (B) HTC 116, (C) Hela, (D) Hek 293t, (E) bright field of fixed SH-SY5Y with EGFP, (F) merged picture of SH-SY5Y with EGFP cellular staining with DAPI and PI in the same view of (E). This figure ...

To prove the immediate applications of μFC, the cells' response to drug stimulation was subsequently tested. The SH-SY5Y cells exposed that different concentrations of DA appeared different in terms of shape and proliferation (Figure 7). Supplementary Material, Figure S4 presents the μFC design of a drug concentration gradient generator and according to the fluid simulation, the center area was selected as the observation place to maximally reduce the shear stress and ensure that the performance of the cell was mostly influenced by the drug level. The cells closer to the main channel were influenced by the high concentration of drug and became round or died, while the cells below maintained normal viability because the low concentration of drug by diffusion. The results directly proved the feasibility of μFC for the application of drug delivery, and we may predict its potential applications in cell–cell interactions and many other related cell research fields.

Figure 7.
SH-SY5Y (with EGFP) cells were different in terms of shape and proliferation due to DA concentration gradient. This figure is available in black and white in print and in color at JCS online.


Compared with other methods that were usually mentioned or utilized for fabricating the μFC, the feature of our progress is to provide an easy way to fabricate the μFC without any template and it is greatly fit for the early stage of μFC structure exploration. Besides, the multilayer fabricated technique can make a complex 3D structure. The common chip made by different materials such as PDMS can usually form not more than four layers, while according to our experiments, the NOA81 μFC can be built up to seven layers with clearer flow passage. It is convenient for the demand of deep channel or chamber. The disadvantage of the μFC is the relatively low precision compared with some of the fabrication methods, and we are now working on to improve it. We have first proposed of fabricating PDMS-NOA81-linked microvalves that may greatly broaden the application of NOA 81 μFC since the track of the fluid can be controlled and converted easily. According to our experiments, various cells can keep normal proliferation in the chip with better biocompatibility than PDMS for cell adherent. The outcome may admit the μFC potential in cell biology, which gives us a direct evidence for cell response to the drug stimulation. We believe that this kind of rapidly fabricated μFC has its extensive applications not only for cell manipulation, but also for various areas such as electrophoresis, chemical reaction and qualitative analysis. Since each application of the chip may undergo the exploration progress until best-matched design was proposed, the μFC may be a time-saving method with better stability and biocompatibility.


We have presented a rapid method to fabricate μFCs without any templates and to be helpful for making spatial and stereo structure by adding different layers in a few hours. In this way, various patterns of chips can be easily fabricated. The proposed PDMS-NOA81-linked microvalves definitely expand the various applications of μFCs. It was observed that SH-SY5Y on both NOA 81 substrate and μFC did not show any change in terms of proliferation or viability, which proves cells' viabilities on NOA81 μFCs in sufficient medium. To make sure the accuracy of this chip, U87, Hela, HCT 116 and Hek 293t were also tested with the same procedures. Furthermore, all cell lines showed normal shape and favorable proliferation after 5 days, which may give convenient opportunities for long-term cell study. The drug test was also carried out to observe the cells' response in chip. Our results show high biocompatibility for various cell lines in vitro and normal response for drug delivery. The outcomes may confirm that this kind of fabrication of μFC can be used as a promising tool for the study of cell manipulation and various biological applications in vitro study.


This work was supported by the National Key Scientific Instrument and Equipment Development Project of China (No. 2012YQ040140).


1. Huang C.-W., Lee G.-B.; A microfluidic system for automatic cell culture; Journal of Micromechanics and Microengineering, (2007); 17(7): 1266.
2. El-Ali J., Sorger P.K., Jensen K.F.; Cells on chips; Nature, (2006); 442(7101): 403–411. [PubMed]
3. Munoz-Pinedo C., Green D.R., van den Berg A.; Confocal restricted-height imaging of suspension cells (CRISC) in a PDMS microdevice during apoptosis; Lab on a Chip, (2005); 5(6): 628–633. [PubMed]
4. Whitesides G.M.; The origins and the future of microfluidics; Nature, (2006); 442(7101): 368–373. [PubMed]
5. Mehling M., Tay S.; Microfluidic cell culture; Current Opinion in Biotechnology, (2014); 25: 95–102. [PubMed]
6. Ye N., Qin J., Shi W., Liu X., Lin B.; Cell-based high content screening using an integrated microfluidic device; Lab on a Chip, (2007); 7(12): 1696–1704. [PubMed]
7. Huang C., Ramadan Q., Wacker J.B., Tekin H.C., Ruffert C., Vergères G. et al. ; Microfluidic chip for monitoring Ca2+ transport through a confluent layer of intestinal cells; RSC Advances, (2014); 4(95): 52887–52891.
8. Majumdar D., Gao Y., Li D., Webb D.J.; Co-culture of neurons and glia in a novel microfluidic platform; Journal of Neuroscience Methods, (2011); 196(1): 38–44. [PMC free article] [PubMed]
9. Menon N.V., Chuah Y.J., Cao B., Lim M., Kang Y.; A microfluidic co-culture system to monitor tumor-stromal interactions on a chip; Biomicrofluidics, (2014); 8(6): 064118. [PubMed]
10. Stroock A.D., Fischbach C.; Microfluidic culture models of tumor angiogenesis; Tissue Engineering Part A, (2010); 16(7): 2143–2146. [PMC free article] [PubMed]
11. Huang H., Jiang L., Li S., Deng J., Li Y., Yao J. et al. ; Using microfluidic chip to form brain-derived neurotrophic factor concentration gradient for studying neuron axon guidance; Biomicrofluidics, (2014); 8(1): 014108. [PubMed]
12. Baudoin R., Griscom L., Prot J.M., Legallais C., Leclerc E.; Behavior of HepG2/C3A cell cultures in a microfluidic bioreactor; Biochemical Engineering Journal, (2011); 53(2): 172–181.
13. Wang L., Sun B., Ziemer K.S., Barabino G.A., Carrier R.L.; Chemical and physical modifications to poly (dimethylsiloxane) surfaces affect adhesion of Caco-2 cells; Journal of Biomedical Materials Research Part A, (2010); 93(4): 1260–1271. [PubMed]
14. Anderson J.R., Chiu D.T., Jackman R.J., Cherniavskaya O., McDonald J.C., Wu H. et al. ; Fabrication of topologically complex three-dimensional microfluidic systems in PDMS by rapid prototyping; Analytical Chemistry, (2000); 72(14): 3158–3164. [PubMed]
15. Ren K., Chen Y., Wu H.; New materials for microfluidics in biology; Current Opinion in Biotechnology, (2014); 25: 78–85. [PubMed]
16. Ren K., Dai W., Zhou J., Su J., Wu H.; Whole-teflon microfluidic chips; Proceedings of the National Academy of Sciences, (2011); 108(20): 8162–8166. [PubMed]
17. Bartolo D., Degré G., Nghe P., Studer V.; Microfluidic stickers; Lab on a Chip, (2008); 8(2): 274–279. [PubMed]
18. PhillipáLee A.; Rapid microfabrication of solvent-resistant biocompatible microfluidic devices; Lab on a Chip, (2008); 8(6): 983–987. [PubMed]
19. Arayanarakool R., Le Gac S., van den Berg A.; Low-temperature, simple and fast integration technique of microfluidic chips by using a UV-curable adhesive; Lab on a Chip, (2010); 10(16): 2115–2121. [PubMed]
20. Weinhausen B., Köster S.; Microfluidic devices for X-ray studies on hydrated cells; Lab on a Chip, (2013); 13(2): 212–215. [PubMed]
21. Morel M., Bartolo D., Galas J.-C., Dahan M., Studer V.; Microfluidic stickers for cell-and tissue-based assays in microchannels; Lab on a Chip, (2009); 9(7): 1011–1013. [PubMed]
22. Naito T., Arayanarakool R., Le Gac S., Yasui T., Kaji N., Tokeshi M. et al. ; Temperature-driven self-actuated microchamber sealing system for highly integrated microfluidic devices; Lab on a Chip, (2013); 13(3): 452–458. [PubMed]
23. Hulme S.E., Shevkoplyas S.S., Whitesides G.M.; Incorporation of prefabricated screw, pneumatic, and solenoid valves into microfluidic devices; Lab Chip, (2009); 9(1): 79–86. [PMC free article] [PubMed]
24. Ying Z., Lin F., Gu W., Su Y., Arshad A., Qing H. et al. ; α-Synuclein increases U251 cells vulnerability to hydrogen peroxide by disrupting calcium homeostasis; Journal of Neural Transmission, (2011); 118(8): 1165–1172. [PubMed]
25. Wägli P., Homsy A., de Rooij N.; Norland optical adhesive (NOA81) microchannels with adjustable wetting behavior and high chemical resistance against a range of mid-infrared-transparent organic solvents; Sensors and Actuators B: Chemical, (2011); 156(2): 994–1001.

Articles from Journal of Chromatographic Science are provided here courtesy of Oxford University Press