Lab on a Chip

PAPER

Cite this: DOI: 10.1039/c9lc00759h

Received 5th August 2019,Accepted 18th October 2019

DOI: 10.1039/c9lc00759h

rsc.li/loc

Tunable microfluidic device fabricated byfemtosecond structured light for particle and cellmanipulation†

Kai Hu,‡a Liang Yang,‡a Dongdong Jin,b Jiawen Li, *a Shengyun Ji,a Chen Xin,a

Yanlei Hu, a Dong Wu, *a Li Zhang b and Jiaru Chua

Smart devices made of stimuli-responsive (SR) hydrogel can realize accurate shape control with high

repeatability attributed to their fast swelling and shrinking upon the change of external stimuli. Integrating

these devices into microfluidic chips and utilizing their controllable deformation capability are highly

promising approaches to enrich the functions of microfluidic devices and reduce their external

apparatuses. Herein we propose and demonstrate a tunable microfluidic device (TMFD) by integrating a

pH-sensitive hydrogel microring array into a microchannel. Instantaneous and reversible deformation of

the microrings can be finished in less than 200 ms. The array gaps of the microrings are reversibly switched

to realize the capture or release of microobjects. In addition, a femtosecond laser holographic processing

method is firstly used to pattern and integrate the pH-sensitive hydrogel microrings into a microchannel,

and the pH-responsive properties of the hydrogel affected by laser processing dosages are theoretically

and experimentally investigated. With this method, the height, diameter (6–16 μm), swelling ratio (35–65%),

and diameter change (2–5 μm) can be precisely controlled. As a proof of concept, the filtering of

polystyrene particles with multiple sizes and complete trapping of PS particles and cells are demonstrated

by these TMFDs. The developed TMFD can be easily integrated by the femtosecond laser holographic

processing method, and operates robustly without the need for external precision apparatuses, which hold

great promise in the applications of microobject manipulation and biomedical analysis.

Introduction

Trapping and holding microobjects, such as microparticlesand cells, play an important role in biomedical research andsingle-cell analysis.1–3 Trapping methods based onmicrofluidic chips are an efficient way for single-particle/cellisolation, due to their low reagent consumption, fast sampleprocessing, high integration, portability and low cost.4 Thesemethods can be broadly classified as active and passivetechniques. The active techniques rely on external force fieldslike optics,5 magnetics,6 acoustics,7 dielectrophoresis8 andelectrophoresis9 for function realization. But the apparatusesthat provide external force fields are always complicated,

inconvenient and expensive. The passive techniques rely onchannel geometries and inherent hydrodynamic forces10–12

for functionality without external integrated devices. Recently,several single-particle/cell capture arrays integrated inmicrofluidic devices13–17 have been proposed based onpassive techniques. However, when particles are captured inthis way, a constant pressure is always required to holdparticles. Otherwise, the captured particles could easilyescape, which limits their application. Therefore, a single-particle/cell capture array that is capable of trapping andholding without the need of an external power supplier orcontinuous fluidic pressure is highly demanded.

Hydrogel is a kind of hydrophilic polymer composed ofphysically and/or chemically crosslinked three-dimensional(3D) networks, which can expand in water and hold a largeamount of water while maintaining the structure due to itscross-linking chains. Nowadays, stimuli-responsive (SR)hydrogel has been widely used for controllable actuation andhas revealed great potential for smart devices, which canundergo volume changes upon the change of external stimulilike pH,18 temperatures,19 molecules,20 light,21 solvents,22

electric signals23 and salts.24 Particularly when thedimensions of the SR hydrogel are up to a microscale, the

Lab ChipThis journal is © The Royal Society of Chemistry 2019

a Key Laboratory of Precision Scientific Instrumentation of Anhui Higher

Education Institutes, CAS Key Laboratory of Mechanical Behavior and Design of

Materials, Department of Precision Machinery and Precision Instrumentation,

University of Science and Technology of China, Hefei 230026, China.

E-mail: jwl@ustc.edu.cn, dongwu@ustc.edu.cnbDepartment of Mechanical and Automation Engineering, The Chinese University

of Hong Kong, Hong Kong 999077, China

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9lc00759h‡ These authors contributed equally to this work.

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article OnlineView Journal

Lab Chip This journal is © The Royal Society of Chemistry 2019

hindrance effect of the polymer network becomes very weak,which can significantly shorten the response time.25

Integrating these smart devices into microfluidic chips maybe a promising way for the functionalization of passivecapture devices. However, it is hard to precisely control the3D size of the hydrogel in microchannels due to itsdeformation characteristics with the change of externalstimuli.

Femtosecond laser microfabrication by two-photonpolymerization (TPP) has been used for fabricating hydrogelfunctional microdevices due to its advantages such as its 3Dprocessing capability and high spatial resolution.26–32 It hasbeen proved to be a versatile technique of fabricating 3Dfluidic microstructures like fluid overpasses,33 micromixers,34

microfilters,35 optofluidic microlens arrays,36 and optofluidicsensors,37 and integrating these microstructures intomicrochips. Nevertheless, the fabrication and integration ofSR hydrogel functional devices by TPP in microchips arerarely reported. Moreover, the robustness, portability, andease of use of such TMFDs have not been even discussed indetail in previous studies.

Herein, we utilize a femtosecond laser holographicprocessing method to rapidly integrate a pH-sensitivehydrogel microring array into a “Y” shaped microchannel tofabricate a TMFD, which can realize multi-filtering andcomplete trapping of particles and cells without externaldevices. Ring structures with a high specific surface area aredesigned and fabricated with a femtosecond Bessel beamgenerated by a liquid-crystal spatial light modulator (SLM).These uniform ring-structures swell and shrink in less than200 ms with the change of environmental pH value. Theinfluence of the laser processing dosage on the swellingratio and the diameter of microrings is quantitativelyanalyzed. This structured light based hydrogel fabricationmethod exhibits advantages such as larger shape change(∼20%), faster fabrication speed (∼102) and better surfacequality over conventional single-point direct laser writing.On this basis, microring arrays can be integrated intoTMFDs easily. Finally, the multi-filtering procedure (PSparticles) and complete trapping procedure (PS particles,yeasts and neural stem cells) using the TMFDs aredemonstrated.

Fig. 1 Fabrication of the TMFDs by integrating hydrogel microrings into the fluidic microchannel with the femtosecond laser holographicprocessing method. a) Schematic illustration of the TMFD fabrication procedure. b) Photograph of the TMFDs in comparison with a one ChinaYuan coin. c) The left section is a computer-generated hologram (CGH) with topological charge n = 35 and axicon radius r0 = 500 μm. The middleand the right sections are the simulated intensity distribution at the focal plane of an oil-immersion objective lens (NA = 1.35) and the bright fieldmicroscopy image of the single exposure by the femtosecond Bessel beam, respectively. Scale bar: 5 μm. d) Schematic illustration and bright fieldmicroscopy images of the hydrogel microring, and its response to pH in unlimited space and in the channel. Scale bar: 20 μm. e) 60° tilted SEMimage of the “Archimedes spiral” hydrogel microring array with height decreasing linearly, and the inset is the top view of the SEM image. Scalebar: 20 μm.

Lab on a ChipPaper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online

Lab ChipThis journal is © The Royal Society of Chemistry 2019

Results and discussionFabrication of the TMFDs

The fabrication process of TMFDs is schematically illustratedin Fig. 1a. The TMFDs are fabricated by soft lithography usingpolydimethylsiloxane (PDMS) and the femtosecond laserholographic processing method. First, a Y-shaped channelmold (80 μm in width and 20 μm in height) is fabricated byUV photolithography on SU-8. A liquid PDMS film is cast ontothe mold, cured and peeled off. Then the inlet as well as theoutlet is generated using a small puncher. Next, the film isbonded to glass to form a sealed microfluidic chip. Finally,the pH-sensitive hydrogel is injected into the microchanneland a microring array is fabricated by the femtosecond laserholographic processing method. The image of a TMFD incomparison with a one China Yuan coin is shown in Fig. 1b.The red and blue pipes are used to inject the acid (10−2 M HClsolution) and base (10−2 M NaOH solution), respectively.

It is well known that cells remain approximately spherical insuspension, and in order to reduce the damage to cells, it issignificant to design microstructures that are in contact with thecells in a ring shape. In addition, the specific surface area of themicroring is relatively larger than that of the circle shape, whichexhibits faster absorption of water for deformation. A ring-structure Bessel beam,38 which is generated by a computer-generated hologram (CGH) displayed on the Lcos-SLM (Fig. S1†),is focused with a high numerical aperture (NA = 1.35) 60× oil-immersion objective lens to fabricate microrings. The diameter(6–16 μm) of the Bessel beam is correlated with topologicalcharge n and axicon radius r0 (Fig. S2†).39 The CGH of n = 35and r0 = 500 μm is shown in the left section of Fig. 1c. Themiddle and the right sections are the simulated intensitydistribution at the focus plane and the bright field microscopyimage of the single exposure (13 μm in outer diameter) by thislight field, respectively. By scanning the focused Bessel beamalong the z-direction, height-controllable microrings can befabricated in nearly 1 s. Fig. 1d shows that the microring with a50 μm height is fabricated with a laser power of 60 mW and ascanning speed of 40 μm s−1, which will swell to a 75 μm heightwhen pH > 9. When the microring structure is elongated bypH, the microring structure may squeeze the top wall of themicrochannel, which may result in the distortion of themicroring structure. So it is critical to precisely control the initialheight of the microring structure to prevent contact squeezewhen the microring swell. However, it is difficult to preciselycontrol the height of the microrings in such a complexmicrochannel substrate by UV photolithography because thecapillary force will lead to an uneven photoresist surface (Fig. S3†).As a demonstration of the precise control of height with ourfabrication technique, an “Archimedes spiral” microring array isfabricated by varying the height from 20 μmto 80 μm(Fig. 1e).

Parameter investigations of hydrogel pH-responsiveproperties

The hydrogel adopted here is a pH-sensitive hydrogel with athreshold of pH 9 for expansion and contraction. The

schematic mechanism of its expansion behavior is shown inFig. 2a. The photo-crosslinked hydrogel framework mainlycontains acrylic acid (AAc), N-isopropylacrylamide (NIPAAm)and dipentaerythritol pentaacrylate (DPEPA), and thechemical structures of these components are shown in Fig.S4.† The photo-crosslinked hydrogel contains AAc chains withcarboxylic groups being deprotonated when pH > 9 and thenelectrostatic repulsion between the chains increases, whichmake the gap between the chains larger. Meanwhile, Na+ ionsenter the hydrogel to maintain the charge neutrality, whichcauses an osmotic pressure. As a result, the water moleculespenetrate to the hydrogel and cause drastic swelling.40,41

When pH < 9, the carboxylic groups protonate, which resultsin the shrinking of the hydrogel because of dehydration.Although the polymer network contains thermoresponsiveNIPAAm chains, the temperature has a little effect on theswelling ratio of the hydrogel (Fig. S5 and S6†). The swellingprocedure of the microrings can be completed in less than200 ms (Video S1†), which reaches a sub-second level and isshorter than the liquid diffusion time. The response time ofthe microstructure with a thickness of 3 μm is proportionalto the size (Video S2†), which means that the hindrance ofsolution diffusion from the hydrogel with a thin thickness isweak. So the response time is mainly determined by thediffusion of acid or base.

One unique merit of our fabrication method is that boththe swelling ratio and size of the microrings can be wellcontrolled. It is found that the microrings obtained with ahigher laser processing dosage have a smaller swelling ratiobecause of the stronger extent of polymerization and lesswater molecules penetrating. The definition of the swellingratio can be seen in Fig. S7.† The laser processing dosage canbe defined as the laser intensity received in a certain areawithin a certain period of time, which would increase whenthe laser power increases or the scanning speed slows down.The bright field microscopy images of the microrings withlaser processing dosages of 50 mJ (laser power of 50 mW andscanning speed of 60 μm s−1) and 270 mJ (laser power of 90mW and scanning speed of 20 μm s−1) at pH < 9 and pH > 9are shown in the left part of Fig. 2b. The two kinds ofmicrorings both have a height of 60 μm at pH < 9. However,after being immersed in a liquid environment of pH > 9, themicroring with a lower laser processing dosage expands morethan the one with a higher processing dosage. The swellingratio of the microrings versus scanning speed with differentlaser powers is systematically investigated (Fig. 2bright section). It can be concluded that the swelling ratio ofthe microrings is negatively correlated with the laser powerand positively with the scanning speed.

In addition to the swelling ratio, the diameter of themicrorings can also be controlled by the laser processingdosage. Fig. 2c and d show the bright field microscopyimages of the microrings (height ≈ 20 μm) with laser powersfrom left to right of 90, 80, 70, 60, and 50 mW and scanningspeeds from top to bottom of 20, 30, 40, 50, 60, 70, 80, and90 μm s−1 at pH > 9 and pH < 9, respectively. Detailed

Lab on a Chip Paper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online

Lab Chip This journal is © The Royal Society of Chemistry 2019

statistics about the outside diameters of the microringsunder swelling and shrinking are shown in the right section.It has been observed that the microrings fabricated with ahigher laser processing dosage have a larger outside diameterwhen pH > 9 or pH < 9. It is worth mentioning that at alower laser processing dosage, the thickness of themicrorings becomes thinner and deformation occurs moreeasily when pH < 9. As a result, when the laser power is 50mW and the scanning speed is more than 60 μm s−1, themicrorings can no longer maintain a ring shape at pH < 9,which is similar to single-point direct laser writing (Fig. S8†).The thickness of the microring is mainly determined by thedosage of polymerization, which can be enlarged byincreasing the laser power and decreasing the scanningspeed.27 The performance of reversible switching betweenswelling and shrinking is verified using cycling tests over 50times (Fig. 2e), showing eximious repeatability and stabilityof the microrings fabricated with the pH-sensitive hydrogel.

The gap between the microrings can also be switched dueto the reversible deformation of the microrings with thechange of pH value. Based on this strategy, the microring

array integrated in the TMFDs can realize the capture ofparticles, which can be achieved as long as the followingrelationship is satisfied where L − D1 < d < L − D2. Here, D1

and D2 represent the outside diameter of the microrings atpH > 9 and pH < 9, and d and L represent the diameter ofthe particle and the center gap between the two microrings,as shown in the right part of Fig. 2f. The influence of laserprocessing dosages on the change (ΔD) between D1 and D2 isalso shown in Fig. 2g. Note that the microrings fabricated bythe femtosecond laser holographic processing method have alarger ΔD (∼20%) than those fabricated by single-point directlaser writing (Fig. S9†). It is because 3D structures are writtenpoint by point using single point laser writing and there is anoverlay between adjacent points and layers, which leads tothe higher polymerization density of the hydrogel. However,with the laser holographic method, the laser intensity is welldistributed in the 2D laser pattern and scanning is onlyneeded in the Z direction, which reduces the polymerizationdensity and leads to a large deformation range. It means thatthe microrings fabricated by this method have bettertunability. Meanwhile, since scanning is only needed in the Z

Fig. 2 The effect of dynamic holographic processing parameters on hydrogel pH-responsive properties. a) Schematic mechanism of the pH-sensitive expansion behavior of the hydrogel. b) Bright field microscopy images of the microrings (laser processing dosages are 50 mJ and 270mJ) and the dependence of the swelling ratio on the laser power and scanning speed. Scale bar: 20 μm. c) and d) Bright field microscopy imagesof the microrings and the dependence of the outside diameter on the laser power and scanning speed at pH > 9 and pH < 9, respectively. Scalebar: 20 μm. e) Bright field microscopy images of the microrings at pH > 9 and pH < 9 and repetition test between swelling and shrinking over 50cycles. Scale bar: 10 μm. f) Schematic illustration of the releasing at pH < 9 and capturing at pH > 9 of the particles and the dependence of theoutside diameter change (ΔD = D1 − D2) on the laser processing dosage. All standard deviations were obtained from three parallel tests.

Lab on a ChipPaper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online

Lab ChipThis journal is © The Royal Society of Chemistry 2019

direction in the holographic processing method, thefabrication speed is faster and the fluid disturbance from thehydrogel is reduced, which makes the surface quality of themicrostructure better.

Application in multi-filtering of particles

In general, conventional microporous membranes can onlyfilter particles with one size,35,42,43 whereas a TMFD can filterparticles with multiple sizes due to its tunability. Themicroring array in the TMFD fabricated with the pH-sensitivehydrogel is soft due to the pore network of the material itself,which leads to better swelling and shrinking characteristics.Meanwhile, the microrings in the microchannel have to bestrong enough to resist the pressure from particles and liquid

flow when filtering. In order to improve the resistance of themicrorings, a matryoshka ring with a larger adhesive area tothe microchannel is designed, which contains two nestedrings with different diameters. Meanwhile, due to its up anddown nested structure, the distribution of stress generated bythe flow is more uniform. Bessel beams with differentparameters are employed to fabricate the matryoshka-ringfilter in the microchannel combining a 3D motion stage todetermine the location and height. The fabrication procedureis shown in the left part of Fig. 3a. Outer ring-structures witha 14 μm height are first fabricated on the lower surface of themicrochannel by scanning upward with a Bessel beam of n =35 and r0 = 500, and then inner microrings with a 14 μmheight are fabricated on the upper surface of themicrochannel by scanning downward with a Bessel beam of

Fig. 3 Procedure for multi-filtering of particles with different sizes by changing the pH. a) Schematic illustration of the processing method andconfocal 3D reconstruction of the matryoshka-ring filter. Scale bar: 7 μm. b) Bright field microscopy images of the matryoshka-ring filter at pH < 9and pH > 9. c) and e) Schematic illustration and bright field microscopy images of the three size particle filtering procedure by TMFDs when pH <

9. g) The percentages and bright field microscopy images of each particle at the inlet and outlet of the TMFDs. d) and f) Schematic illustration andbright field microscopy images of the two size particle filtering procedure by TMFDs when pH > 9. h) The percentages and bright field microscopyimages of each particle at the inlet and outlet of the TMFDs. All standard deviations were obtained from three parallel tests. Scale bars: 30 μm,unless specified.

Lab on a Chip Paper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online

Lab Chip This journal is © The Royal Society of Chemistry 2019

n = 5 and r0 = 1000. The scanning speed for all themicrorings is 40 μm s−1, and the laser powers are 60 mW and25 mW, respectively. The right section is the confocal 3Dreconstruction of the matryoshka-ring filter. As shown inFig. 3b, the gap of the matryoshka rings can vary from 11.4μm to 7.2 μm by changing the pH value (Video S3†).

In order to test the function of the matryoshka-ring filter,5, 10, 15 μm-diameter PS particles (1 : 1 : 1 × 104 ml−1) aremixed with acid or base in deionized water, and then themixture is injected into the inlet. Fig. 3c and d schematicallyshow that the 15 μm particles filtered out from threedifferent size particles when pH < 9, and only the 5 μmparticles can pass when pH > 9. The time-lapsed bright fieldmicroscopy images of the filtering procedure are shown in

Fig. 3e and f (Video S4†). When pH < 9, both the 10 μm and5 μm particles pass the filter, and when pH > 9, only the 5μm particles pass the filter. Statistic results and bright fieldmicroscopy images of particle size distributions in the inletand outlet are shown in Fig. 3g and h. 100% of the particleswith sizes of 15 μm and 10 μm have been successfully filteredout at pH < 9 and pH > 9, respectively. The particle sizedistribution in the inlet is not 1 : 1 : 1 due to the differentsettling velocities of different particles (Fig. S10†).

Application for complete trapping of the particles and cells

Traditional microfluidic trapping methods based on 2Dsubstrates can flexibly adjust the array gap, which enable

Fig. 4 Procedure for complete trapping of the particle by changing the pH. a) Schematic illustration and bright field microscopy images of thetrapping and releasing procedure of the particle by TMFDs with a trapezoidal-trap array. b) and c) Bright field microscopy images of the trappingprocedure of the particle by TMFDs with a circular-trap array and rectangle-trap array, respectively. d) Bright field microscopy images of thetrapping procedure of the yeast cells. e) Bright field microscopy images of the deformation of the neural stem cells that are bigger than the arraygap. All scale bars: 20 μm.

Lab on a ChipPaper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online

Lab ChipThis journal is © The Royal Society of Chemistry 2019

trapping varying particles or cells in terms of particlesize.14,15 However, a stable flow is needed when particles arecaptured by this method, otherwise the particles will easilyescape from the array due to the change of fluid. Here, weshow that the microring array in TMFDs can be designed torealize real-time complete trapping of particles and cells. Theconcept of the trapping and releasing procedure by atrapezoidal-trap array is shown in the left part of Fig. 4a.First, 15 μm particles are mixed in a pH < 9 liquid, and theliquid is injected through the red pipe into the TMFDs. Thereis a trapezoidal-trap array consisting of four microrings inthe TMFD that when pH < 9, the gap (∼16.8 μm) in the frontrow is larger than the diameter of the particles, and the rearrow gap (∼8.4 μm) is smaller than the particle diameter.Then the fluid is injected continuously until a particle iscaptured. Next a base is injected through the blue pipe toexpand the microrings, so that the gap in the front rowbecomes smaller (∼13.5 μm), which means that the particleis completely trapped and cannot be flushed away even by areverse flow. The particle can be released by injecting acidthrough the colorless pipe as a reverse flow until themicrorings shrink. The bright field microscopy images of thisprocedure (Video S5†) are shown in the right section. Inaddition to this trapezoidal-trap array, arrays of differentshapes which can be used to trap cells and provide cells withvarious growing spaces can also be flexibly designed. Todemonstrate the feasibility of this arbitrary designation andmicroring capture, a circular-trap array is designed andparticle trapping is realized. The bright field microscopyimages of the trapping procedure (Video S6†) are shown inFig. 4b. It should be mentioned that when the particle istrapped at pH > 9 in the fourth picture, it cannot be releasedby a reverse flow.

It is of great importance to enable clustering of a certainnumber of particles or cells to monitor the interactionsbetween them. For example, when it comes to biologicalapplication like cell communication,44–46 the target cells needto be clustered in a space where other cells cannot enter butnutrients can be exchanged. The difference between single-trap and cluster-trap is the control of the time to inject baseswhen the desired number of particles is trapped in thecluster-trap. Here, it is demonstrated that the TMFDs canachieve the real-time complete trapping of particle clusterswith controlled numbers. As shown in Fig. 4c, four particlesare trapped in a rectangle-trap array one by one (Video S7†).Then, it is shown in the fourth picture that the liquid flowswith pH > 9 are injected and the microrings are expanded,which prevent the entrance of other particles.

Furthermore, the TMFDs can realize the completetrapping of deformable cells. The gap of the rear row needsto be designed smaller than the nucleus rather than the sizeof the cytoplasm, because the nucleus is much more difficultto be deformed than the cytoplasm. A rectangle-trap array isfabricated with small microrings (∼11 μm in diameter),which has a front row gap of 5 μm at pH < 9 and 3 μm at pH> 9. Fig. 4d shows the trapping procedure of yeast cells

(Video S8†), and Fig. 4e shows detailed procedures that theneural stem cell enters the array through its deformation(Video S9†). Due to cytoplasmic deformation and fluid shock,the neural stem cell with a diameter of ∼15 μm can passthrough the front row with a 6 μm gap, and is blocked by therear row with a 3 μm gap.

Conclusion

In summary, the microring array of the pH-sensitive hydrogelcan be integrated into the TMFD rapidly and flexibly by thelaser holographic processing method, which can swell andshrink in less than 200 ms with the change of environmentalpH value. The pH responsive properties of the hydrogel canbe precisely controlled by using different laser processingdosages including swelling ratio (35–65%) and diameterchange (2–5 μm). The TMFDs have demonstrated their abilityin multi-filtering PS particles and completely trapping PSparticles with a specific size. Besides, the efficient trapping ofdeformable yeasts and neural stem cells is demonstrated withthese TMFDs. Both the proposed integration method and theSR hydrogel hold great promise in the applications ofmicroobject manipulation and single-cell biology analysis.

ExperimentalDesign and fabrication of the TMFDs

All the experiments were carried out at room temperature (23°C). The microfluidic chip was laid out by using AutoCADsoftware (Autodesk) and fabricated by soft lithography inPDMS. Briefly, a layer (20 μm thick) of photoresist resin, SU-82025 (MicroChem, Newton, USA), was spin-coated on awafer at 3000 rpm for 20 s and patterned by UV exposurethrough a mask to polymerize the exposed regions to definethe Y-shaped channel (80 μm in width). The PDMS pre-polymer (Sylgard 184, Dow Corning) was mixed with acrosslinker at a ratio of 10 : 1 and degassed under vacuum for15 minutes. Then the mixture was poured onto the wafer,degassed and cured at 65 °C for 1 hour. After curing, themicrochannel-patterned PDMS film was peeled off from thewafer and 3 through-holes were punched to connect with theexternal environment. Next, the PDMS film was bonded toglass by both being treated with oxygen plasma (MinghengPDC-MG, 50 s, 75 W, 100 Pa), and then the microchannelwas treated at 65 °C for 15 minutes to complete the bonding.

Afterwards, the tunable microrings were integrated into amicrochip by the femtosecond laser holographic processingmethod. In this experiment, the pH-sensitive hydrogel wasfirstly injected into the microchannel. The holographicfemtosecond laser writing system source was a mode-lockedTi : sapphire ultrafast oscillator (Chameleon Vision-S,Coherent Inc, USA) with a central wavelength at 800 nm, apulse width at 75 fs, and a repetition rate at 80 MHz. And thelaser beam with 13 mm in diameter was incident on areflection type liquid crystal SLM (Pluto NIR-2, HoloeyePhotonics AG, German) with a reflection rate of 50%, 1920 ×

Lab on a Chip Paper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online

Lab Chip This journal is © The Royal Society of Chemistry 2019

1080 pixels, a pixel pitch of 8 μm and 256 gray levels. TheBessel beam was generated by loading the prepared CGHwith 1080 × 1080 pixels in the center of the SLM and focusedusing a 60× oil-immersion objective lens (Olympus) with highNA (1.35) for high processing quality. The movement of themicrofluidic sample was realized using a three-axispiezoelectric motion stage (E545, from Physik InstrumenteGmbH & Co. KG, Germany) with a resolution of less than 1nm and a 200 μm × 200 μm × 200 μm moving range toprecisely locate the microrings. The microfluidic sample wasdeveloped in isopropanol for 24 h until all of theunpolymerized part was washed away.

Preparation of the hydrogel precursor

Firstly, 1.6 g N-isopropylacrylamide (NIPAAm, 98%), 0.8 mLacrylic acid (AAc, 99%) and 0.15 g polyvinylpyrrolidone (PVP)were added into 1 mL ethyl lactate (EL, 98%) and stirredvigorously. Then the solution was mixed with 0.4 mL dipenta-erythritol pentaacrylate (DPEPA, 98%, American BarkiChemical Inc), 0.5 mL triethanolamine (TEA, 99%), and 0.1mL 4,4′-bisIJdiethylamino)benzophenone (EMK, 97%, ReadingChemical Technology Shanghai Co. Ltd)/N,N-dimethylformamide (DMF, 99.5%) solution (20 wt%) bystirring overnight. AAc, NIPAAm, EL, PVP, DMF and TEA werepurchased from Aladdin Chemicals. The molar ratio ofNIPAAm : AAc :DPEPA is 8.5 : 10.4 : 0.9. PVP was used toincrease the solution viscosity, which enhanced the supportand facilitated the construction of the structure.47

Preparation of PS particles and cells

PS particles with diameters of 5 μm, 10 μm, and 15 μm(BaseLine ChromTech Research Centre, China) were mixed indeionized water solution and the flow speed was about 500μm s−1. In order to meet the different particle concentrationrequirements of the trapping and filtering experiments, theparticles of different sizes were firstly prepared in solutionswith a concentration of 104 ml−1, respectively. If a mixture ofthree particles was required, we can mix the same volume ofthe three particle solutions. And when a mixture of two kindsof particles was required, we can mix the same volume ofthese particle solutions and one PI water. So, theconcentration of particles of various sizes suspended in themixtures was the same. Then HCl solution (0.1 M) or NaOHsolution (0.1 M) was added to adjust the pH value at 4 or 11of the mixture, which ensures that the microrings shrank orexpanded as we designed as the mixture flows through.Finally, the mixed liquid was treated in an ultrasonic bath todisperse particles.

In the trapping experiment, a kind of Saccharomycescerevisiae (Angel Yeast Co., Ltd.) was used and the liquidculture medium was produced by sterilizing the mixture ofsolid granular Sabouraud's glucose broth medium (10.0 g L−1

peptone, 40 g L−1 glucose, pH = 5.6 ± 0.2 at 25 °C, QingDaoHope Biotechnology CO., Ltd.) and distilled water with aconcentration of 50 g L−1 at 120 °C for 20 min. The yeast

medium was produced by dissolving 0.1 g dry yeast in 100mL liquid culture medium. The neural stem cells wereprovided by Stem Cell Bank, Chinese Academy of Science.The neural stem cell culture medium supplemented with10% fetal bovine serum (FBS, Gibco), glutamax (Invitrogen,35050), and nonessential amino acid 100× (Invitrogen, 11140)at 37 °C in a humidified atmosphere of 5% CO2 in abiological operation cabinet.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Key R&D Program ofChina (2018YFB1105400), the National Natural ScienceFoundation of China (No. 51675503, 51875544, 61805230,51805509), the Aeronautical Science Fund (2018ZE78004), theFundamental Research Funds for the Central Universities(WK2090000011, WK2090000013, WK2090090021,WK6030000131, WK6030000103), Youth InnovationPromotion Association CAS (2017495), and the Foundation ofEquipment Development Department (62209140 10901). Weacknowledge the Experimental Center of Engineering andMaterials Sciences at USTC for the fabrication andmeasurement of the samples. This work was partly carriedout at the USTC Center for Micro and Nanoscale Researchand Fabrication.

References

1 F. Zhou, X. Li, W. Wang, P. Zhu, J. Zhou, W. He, M. Ding, F.Xiong, X. Zheng and Z. Li, Nature, 2016, 533, 487.

2 D. E. Wagner, C. Weinreb, Z. M. Collins, J. A. Briggs, S. G.Megason and A. M. Klein, Science, 2018, 360, 981–987.

3 C. Kim, R. Gao, E. Sei, R. Brandt, J. Hartman, T. Hatschek,N. Crosetto, T. Foukakis and N. E. Navin, Cell, 2018, 173,879–893.e813.

4 G. M. Whitesides, Nature, 2006, 442, 368.5 M. P. MacDonald, G. C. Spalding and K. Dholakia, Nature,

2003, 426, 421.6 D. Melville, F. Paul and S. Roath, Nature, 1975, 255, 706.7 F. Petersson, L. Åberg, A.-M. Swärd-Nilsson and T. Laurell,

Anal. Chem., 2007, 79, 5117–5123.8 X. Hu, P. H. Bessette, J. Qian, C. D. Meinhart, P. S.

Daugherty and H. T. Soh, Proc. Natl. Acad. Sci. U. S. A.,2005, 102, 15757–15761.

9 B. Yao, G.-a. Luo, X. Feng, W. Wang, L.-x. Chen and Y.-m.Wang, Lab Chip, 2004, 4, 603–607.

10 P. Sethu, A. Sin and M. Toner, Lab Chip, 2006, 6, 83–89.11 J.-S. Park, S.-H. Song and H.-I. Jung, Lab Chip, 2009, 9,

939–948.12 Z. Wu, Y. Chen, M. Wang and A. J. Chung, Lab Chip,

2016, 16, 532–542.13 W.-H. Tan and S. Takeuchi, Proc. Natl. Acad. Sci. U. S. A.,

2007, 104, 1146–1151.

Lab on a ChipPaper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online

Lab ChipThis journal is © The Royal Society of Chemistry 2019

14 J.-P. Frimat, M. Becker, Y.-Y. Chiang, U. Marggraf, D.Janasek, J. G. Hengstler, J. Franzke and J. West, Lab Chip,2011, 11, 231–237.

15 S. M. McFaul, B. K. Lin and H. Ma, Lab Chip, 2012, 12,2369–2376.

16 M. C. Jo, W. Liu, L. Gu, W. Dang and L. Qin, Proc. Natl.Acad. Sci. U. S. A., 2015, 112, 9364–9369.

17 L. Weng, F. Ellett, J. Edd, K. H. Wong, K. Uygun, D. Irimia,S. L. Stott and M. Toner, Lab Chip, 2017, 17, 4077–4088.

18 C. Ma, T. Li, Q. Zhao, X. Yang, J. Wu, Y. Luo and T. Xie, Adv.Mater., 2014, 26, 5665–5669.

19 J. Kim, J. A. Hanna, M. Byun, C. D. Santangelo and R. C.Hayward, Science, 2012, 335, 1201–1205.

20 A. Cangialosi, C. Yoon, J. Liu, Q. Huang, J. Guo, T. D. Nguyen,D. H. Gracias and R. Schulman, Science, 2017, 357, 1126–1130.

21 E. Wang, M. S. Desai and S.-W. Lee, Nano Lett., 2013, 13,2826–2830.

22 Y.-L. Zhang, Y. Tian, H. Wang, Z.-C. Ma, D.-D. Han, L.-G.Niu, Q.-D. Chen and H.-B. Sun, ACS Nano, 2019, 13(4),4041–4048.

23 L. Zhao, J. Huang, Y. Zhang, T. Wang, W. Sun and Z. Tong,ACS Appl. Mater. Interfaces, 2017, 9, 11866–11873.

24 Y. Zhang, H.-F. Ji, G. M. Brown and T. Thundat, Anal. Chem.,2003, 75, 4773–4777.

25 T. Tanaka and D. J. Fillmore, J. Chem. Phys., 1979, 70,1214–1218.

26 S. Kawata, H. B. Sun, T. Tanaka and K. Takada, Nature,2001, 412, 697–698.

27 L. Yang, A. Münchinger, M. Kadic, V. Hahn, F. Mayer, E.Blasco, C. Barner-Kowollik and M. Wegener, Adv. Opt.Mater., 2019, 1901040.

28 T. Yang, H. Lin and B. Jia, Front. Optoelectron., 2018, 11,2–22.

29 M. Hippler, E. Blasco, J. Qu, M. Tanaka, C. Barner-Kowollik,M. Wegener and M. Bastmeyer, Nat. Commun., 2019, 10, 232.

30 Y.-L. Zhang, Y. Tian, H. Wang, Z.-C. Ma, D.-D. Han, L.-G. Niu,Q.-D. Chen and H.-B. Sun, ACS Nano, 2019, 13, 4041–4048.

31 D. Martella, S. Nocentini, D. Nuzhdin, C. Parmeggiani andD. S. Wiersma, Adv. Mater., 2017, 29, 1704047.

32 A. Tudor, C. Delaney, H. Zhang, A. J. Thompson, V. F. Curto,G.-Z. Yang, M. J. Higgins, D. Diamond and L. Florea, Mater.Today, 2018, 21, 807–816.

33 Y. He, B.-L. Huang, D.-X. Lu, J. Zhao, B.-B. Xu, R. Zhang,X.-F. Lin, Q.-D. Chen, J. Wang and Y.-L. Zhang, Lab Chip,2012, 12, 3866–3869.

34 K. Sugioka, J. Xu, D. Wu, Y. Hanada, Z. Wang, Y. Cheng andK. Midorikawa, Lab Chip, 2014, 14, 3447–3458.

35 L. Amato, Y. Gu, N. Bellini, S. M. Eaton, G. Cerullo and R.Osellame, Lab Chip, 2012, 12, 1135–1142.

36 D. Wu, L.-G. Niu, S.-Z. Wu, J. Xu, K. Midorikawa and K.Sugioka, Lab Chip, 2015, 15, 1515–1523.

37 L. Kelemen, E. Lepera, B. Horváth, P. Ormos, R. Osellameand R. M. Vázquez, Lab Chip, 2019, 1985–1990.

38 L. Yang, A. El-Tamer, U. Hinze, J. Li, Y. Hu, W. Huang, J.Chu and B. N. Chichkov, Appl. Phys. Lett., 2014, 105, 041110.

39 S. Ji, L. Yang, Y. Hu, J. Ni, W. Du, J. Li, G. Zhao, D. Wu andJ. Chu, Small, 2017, 13, 1701190.

40 B. Firestone and R. Siegel, J. Appl. Polym. Sci., 1991, 43,901–914.

41 Y. Qiu and K. Park, Adv. Drug Delivery Rev., 2001, 53,321–339.

42 B. Xu, W.-Q. Du, J.-W. Li, Y.-L. Hu, L. Yang, C.-C. Zhang,G.-Q. Li, Z.-X. Lao, J.-C. Ni and J.-R. Chu, Sci. Rep., 2016, 6,19989.

43 J. A. Hernández-Castro, K. Li, J. Daoud, D. Juncker and T.Veres, Lab Chip, 2019, 19, 589–597.

44 M. Mathieu, L. Martin-Jaular, G. Lavieu and C. Théry, Nat.Cell Biol., 2019, 21, 9.

45 T. Xin, V. Greco and P. Myung, Cell, 2016, 164, 1212–1225.46 W.-J. Rappel, Proc. Natl. Acad. Sci. U. S. A., 2016, 113,

1471–1473.47 D. D. Jin, Q. Y. Chen, T. Y. Huang, J. Y. Huang, L. Zhang

and H. L. Duan, Mater. Today, 2019, DOI: 10.1016/j.mattod.2019.06.002.

Lab on a Chip Paper

Publ

ishe

d on

18

Oct

ober

201

9. D

ownl

oade

d by

Uni

vers

ity o

f Sc

ienc

e an

d T

echn

olog

y of

Chi

na o

n 11

/5/2

019

2:18

:37

AM

. View Article Online