From chip-in-a-lab to lab-on-a-chip: a portable Coulter counter using a modular platform

The field of microfluidics has been struggling to obtain widespread market penetration. In order to overcome this struggle, a standardized and modular platform is introduced and applied. By providing easy-to-fabricate modular building blocks which are compatible with mass manufacturing, we decrease the gap from lab-to-fab. These standardized blocks are used in combination with an application-specific fluidic circuit board. On this board, electrical and fluidic connections are demonstrated by implementing an alternating current Coulter counter. This multipurpose building block is reusable in many applications. In this study, it identifies and counts 6 and 11 μm beads. The system is kept in a credit card-sized footprint, as a result of in-house-developed electronics and standardized building blocks. We believe that this easy-to-fabricate, credit card-sized, modular, and standardized prototype brings us closer to clinical and veterinary applications, because it provides an essential stepping stone to fully integrated point -of -care devices.

The field of microfluidics has been struggling to obtain widespread market penetration. In order to 12 overcome this struggle, a standardized -modular platform is introduced and applied. By providing 13 easy-to-fabricate modular building blocks which are compatible with mass manufacturing, we 14 decrease the gap from lab-to-fab. These standardized blocks are used in combination with an 15 application specific fluidic circuit board. On this board, electrical and fluidic connections are 16 demonstrated by implementing an alternating current coulter counter. This multipurpose building 17 block is reusable in many applications. In this study it identifies and counts 6 µm and 11 µm beads. 18 The system is kept in a credit card sized footprint, as a result of in-house developed electronics and 19 standardized building blocks. We believe that this easy-to-fabricate, credit card sized, modular and 20 standardized prototype brings us closer to clinical and veterinary applications, because it provides an 21 essential stepping stone to fully integrated point of care devices.Introduction 22

23
After the initial hype in microfluidics almost three decades ago, which occurred after the introduction 24 of the micro total analysis systems (MicroTAS), the field of microfluidics continues to struggle with 25 obtaining widespread market penetration 1,2 . Lack of standards, focus and communication between 26 academics and industry could be the reason for this struggle 3 . Although many lab-on-a-chip devices 27 are presented in the literature, the term chip-in-a-lab fits better with prototypes currently 28 demonstrated in the microfluidic field 4 . Even though there are a few frontier commercialized devices 29 (Abbott i-STAT and DNA PCR machines), most lab-on-a-chip devices are still stuck at a technology 30 readiness level (TRL) of 3 or 4. Figure 1 shows a diagram with different stages of innovation and their 31 corresponding technology readiness levels 5,6 . If these early prototype devices would become 32 available commercially, they could serve several niche markets and start to decrease the discrepancy 33 between academic output and actual market revenue. 34 Currently, most of the microfluidic systems are not compact, due to issues with system integration 7 . 36 Even though auxiliary equipment is part of microfluidic set-ups, it is originally intended for other 37 fields, it is bulky and not easy to miniaturize 8 . Moreover, components for fluid control, visualization 38 and signal detection are not interoperable with each other, thus commercialization is prevented 7,9 . 39 In order to realize true point-of-care (POC) microfluidic devices, this bulky auxiliary equipment should 40 be miniaturized. Decreasing and standardizing the footprint of auxiliary components will be a step in 41 enhancing portability and reducing equipment dependability. Furthermore, it will improve system 42 integration and compactness, ultimately preventing microfluidic systems to end up in the valley of 43 death between academia and industry. 44 Until now standards have not been established in the microfluidic field, which leads to adoption of 45 equipment from other fields, e.g. smart phones for read-out 10 , microscope slides, microtiter plates, 46 and Luer connectors 11 . Therefore it is difficult to assemble them in one compact system, which often 47 manifest itself as a so-called spider web assembly in the lab where many tubes and wires connect the 48 various parts to each other. Moreover, the microfluidic community prefers a wide variety of 49 substrate materials including glass, silicon and polymers, such as polydimethylsiloxane (PDMS). Even 50 though PDMS is a popular material amongst researchers, it lacks high volume and low cost 51 fabrication and does not have long shelf life, making it unsuitable for realizing commercial 52 products 8,12,13 . In order to commercially benefit from already presented microfluidic devices in 53 academia, which are based on a wide variety of materials, a standardized platform is needed. To 54 provide successful system integration, a large consortium of major industrial and academic partners 55 have introduced standardized elements that maintain compatibility and interoperability between MFBBs that provide various functions are assembled onto a credit card size FCB ( Fig. 2a and 2c.1). 82 The printed circuit board (PCB) holds all necessary electronics needed for the pseudo-differential 83 measurement and increases the compactness of the system (Fig. 2b). Interfacing and assembling 84 MFBBs ( Fig. 2c.2) onto the FCB ( Fig. 2c.1) provides functionality and interconnection, respectively. 85 Necessary tools for assembly are shown in Fig. 2c.3. These standardized parts can be put in an online 86 repository to be easily integrated in future projects 22 . The standardization process together with 87 MFBBs facilitates increasing the TRL (Fig. 1) and decreases the gap between the "chip in a lab" 88 concept and a commercial point of care device.

90
Experimental results

91
The AC coulter counter chip integrated onto a FCB is used to detect 6 µm and 11 µm polystyrene 92 beads in phosphate-buffered saline (PBS). Three sets of experiments are conducted. Solutions that 93 contain only 6 µm beads, only 11 µm beads or both of them in 1:1 (v/v) ratio are pumped through 94 the microfluidic channels. The results for these measurements are shown in Table 1. The video data 95 is used to obtain the trajectory of the bead as well as the size of a bead, which is subsequently linked 96 to the electrical signal. An exemplary result of this process is shown in Fig 3b. Also, a video showing 97 this process can be found in the Supplementary Material (Video S1). As Table 1 shows 6 µm beads 98 were detected while running the 11 µm bead solution. Although the system in this case should only 99 detect 11 µm beads as only those kind of beads are present in the bulk solution, some 6 µm beads 100 are still detected from earlier experiments that are left behind in the channels. In order to calculate 101 the detection rate, the number of video detection matched to an electrical signal peak is divided by 102 the total amount of beads detected in the video.

104
The results show that both solutions that contain 11 µm beads have a detection rate of more than 99 105 percent. The detection of 6 µm beads was more challenging, since the peaks generated are close to 106 the noise band (Fig. 3A). 1 mV rms was deemed to be a suitable threshold for peak detection, since the 107 root mean square (RMS) level of the noise was 0.307mV rms . The largest contributor to the noise was 108 the minimal shielding on the FCB. Even though the threshold was set to 3 times of the RMS level of 109 the noise, some false positives still occurred. The threshold was kept fixed to prevent lowering the 110 detection rate. This threshold resulted in detection rates of 86 %, 39 % and 32 % for 6 µm beads at 111 velocities of 110 µm/s, 289 µm/s, 344 µm/s, respectively. From these results it is clear that detection 112 rate was negatively influenced by higher flow speeds, since higher flow speeds result in higher 113 frequency components in the detected peaks. Depending on the filter setting in the lock-in amplifier, 114 removal of these high frequency components leads to a peak amplitude below the threshold for 6 115 µm beads. Increasing the bandwidth of the filters to allow the higher frequency components to pass 116 is not favourable as the noise would also increase. 117 118 Figure 3 -a) Differential impedance measurement with 6 µm and 11 µm beads. In total 101 beads were electrically detected, 46 of them are 6 µm, 46 of them are 11 µm beads and 9 of them are false positives. b) Impedance data of an individual 11 µm bead combined with the video frames used to obtain its trajectory. c) Histograms of the peak amplitudes for a mixed solution.

120
The distribution of detected peak amplitudes for the mixed solution is shown in Fig. 3c. The peak 121 amplitudes for the 11 µm beads are between 1.9 mV rms and 20.7 mV rms , while the majority of peak 122 amplitudes for 6 µm beads is between 1 mV rms and 1.8 mV rms , with an average of 1.6 mV rms . Since 123 the total amplification of the pseudo differential measurement electronics is 100 kV/A, it results in a 124 16nA current difference between the RC reference and the microfluidic chip. A 6 µm bead passing 125 the electrodes generates a difference of 0.3 % in the total current flowing between the electrodes. 126 The wide distribution of the peak amplitudes for 11 µm beads (Fig. 3c) can be explained by size 127 variation of the beads, as well as their position in the non-uniform electric field between the planar 128 electrodes. The coefficients of variation (CV) supplied by the manufacturer are 18 % and 7 % for the 129 11 µm and 6 µm beads, respectively (Fig. 4). This variation results in a small overlap of the 130 populations. Due to this overlap, both populations can generate equal peak amplitudes. Moreover, 131 the peak amplitude of the signal depends on bead volume instead of bead diameter. Therefore, the 132 histogram in Fig. 3c has a broader distribution than the plot in Fig. 4. The two outliers in the 6 µm 133 peak amplitude histogram are 6 µm beads which were falsely linked to an electrical peak of a 11 µm 134 bead. This occurs when a 6 µm bead is not electrically detected and a 11 µm bead is present in the 135 same 100 ms time interval. our standardized platform in order to pave the way for a POC device. This chip consists of coplanar 149 electrodes that are fabricated in a microfluidic channel, as shown in Fig. 5a. An in-house developed 150 lock-in amplifier is used to apply an alternating potential across the electrodes, resulting in an 151 electric field inside the fluidic channel. The impedance of the system increases approximately 0.01%, 152 when an insulating polystyrene bead passes through this electric field. 153 Design of the microfluidic system 154 155 A design philosophy was used where standardization plays a major role. The system consists of 156 standardized MFBBs that provide the functionality of the system along with aforementioned coulter 157 counter chip. A FCB is designed to connect these functional blocks. Interconnect positions between 158 the MFBBs and the FCB are standardized. A footprint and interconnect grid for the MFBB are 159 described in the design guidelines 14-16 . Thanks to this standardization approach, the MFBBs are 160 reusable. Fig. 5a-c show the MFBBs used to assemble the microfluidic system: a coulter counter chip 161 (Fig. 5a), a 1.5 mL reservoir containing the sample (Fig. 5b) and a polymethyl methacrylate (PMMA) 162 block to connect tubing to a waste reservoir (Fig. 5c). An external pressure regulator (Fluigent MFCS-163 4C, France) is used to pump the bead solution through the system. Electronics is developed in-house 164 to amplify and record the impedance signals generated in the coulter counter chip. 165 166 Detection and signal processing electronics 167 168 In order to detect the impedance signals that the coulter counter chip generates, a two-step solution 169 is used. As stated before, only very small changes in impedance of approximately 0.01% generated by 170 the passing beads needs to be detected. Sensitive electronics were developed, including a (pseudo-) 171 differential measurement for detection. Fig. 6 shows a block diagram of this pseudo-differential set-172 up. A resistor and capacitor reference circuit was used to mimic the behaviour of the coulter counter 173 chip. As a result, the small changes in impedance are recorded with more detailed in the dynamic 174 range of the 14-bit analog-to-digital converter (ADC) of the lock-in amplifier. When a bead passes by 175 the electrodes, a difference in current through the coulter counter chip arises. The signal of interest 176 is obtained by subtracting the coulter counter signal from the reference signal. The programmable 177 gain amplifier matches this signal to the dynamic range of the input for the next stage. Additionally, 178 an in-house developed digital lock-in amplifier is used to significantly reduce the amount of noise in 179 the recorded signal. Moreover, the trans-impedance amplifier was placed close to the coulter 180 counter chip on the FCB to prevent issues with noise. All other electronics are soldered on a PCB. 181  The in-house developed lock-in amplifier consists of three main parts: an ADC, a DSP and a direct 183 digital synthesis (DDS) integrated circuit (Fig. 6). Both the ADC and DDS provide an interface between 184 the analog domain and the digital domain. The DDS generates a sinusoidal signal that is applied to 185 the coulter counter chip and the analog output signal of it is digitized by ADC. Moreover, the lock-in 186 process is performed by a DSP to prevent additional noise. 187 In this study, we used three MFBBs: a detection unit (coulter counter chip), an inlet reservoir, and an 188 outlet reservoir. The inlet reservoir MFBB is a gas tight 1.5 mL high performance liquid 189 chromatography (HPLC) vial which is connected to two syringe tips in a PMMA block with a standard 190 footprint 21 . The coulter counter chip is fabricated out of two glass substrates, as described 191 previously 20

198
The FCB is fabricated using two plates of 2 mm thick cyclic olefin copolymer (COC) (DENZ 199 Austria) (Fig. 7). COC is a good candidate for high volume manufacturing, because it can be injection COC plates are exposed to cyclohexane vapour for 3 min. and afterwards immediately pressed 208 together for 15 min. at 2. prevent clustering of the beads. These solutions are thoroughly mixed before they are loaded into 224 the inlet reservoir MFBB. In addition to a height difference of 1 cm between inlet and outlet reservoir, 225 a pressure pump is initially used to introduce fluid into the channel. The current through the coulter 226 counter chip is recorded at 0.3V RMS , 90 kHz in real-time using a LabView program interfacing the lock-227 in amplifier. The total amplification of the electronics is 100kV/A. 228 The recorded data is subsequently processed with a MATLAB (MathWorks, USA) script, in order to 229 detect peaks that represent beads. A moving average filter is applied to the signal and the result of 230 this operation is subtracted from the original signal to remove drift in the signal. Peak amplitudes and 231 corresponding times are obtained from this processed signal and compared to the video data. 232 Video data is obtained during the experiment, using a microscope (Leica DM LM and SUSS MicroTec, 233 Germany) and a high speed camera at 60 fps (Grasshopper 3, Point Grey, USA). A MATLAB script is 234 used to analyse the resulting video. The script detects the amount of pixels the bead occupies, in 235 order to identify different bead sizes. The location and size information for each bead is saved to 236 determine the exact time that the bead passes the electrodes. Subsequently, this time is used to find 237 the corresponding peak in the impedance data. Overall this script provides information about 238 detected bead size, peak amplitude and bead velocity for each bead. In this paper, we demonstrate a low-cost, modular particle counting microfluidic system with a credit 243 card sized footprint. The reusable standardized parts decrease the design and fabrication time, which 244 leads to a reduction in the time-to-market for industrial development. The standard platform 245 provides the possibility of integrating different materials in one system. Here, a COC, a PMMA and a 246 glass device are combined. The low cost of in-house built electronics and lock-in amplifier facilitates a 247 route to commercialization while still keeping system performance up to par. We achieve to measure 248 a 0.3 % change in the impedance signal using a pseudo-differential measurement. As a result, 6 µm 249 and 11 µm polystyrene beads were differentiated. All 11 µm beads were electrically detected, while 250 the detection rates for 6 µm beads were 86 %, 39 % and 32 % depending mostly on the flow rate 251 used. Better control over this flow rate will ensure a higher detection rate in the future. Moreover, 252 better shielding of especially the trans-impedance amplifier on the FCB will reduce the noise levels. 253 Most importantly, we show that this standardized platform bridges the gap from "chip-in-a-lab" to 254 "lab-on-a-chip". The only drawback of the system is the use of an external lock-in amplifier and a 255 pressure regulator. However, we are confident that we can shrink the size of these two components 256 towards obtaining a commercial product. 257