Micro-assembled multi-chamber thermal cycler for low-cost reaction chip thermal multiplexing

https://doi.org/10.1016/S0924-4247(02)00384-9Get rights and content

Abstract

This paper presents a miniaturised multi-chamber thermal cycler capable of thermal multiplexing for high throughput polymerase chain reaction (PCR) of nucleic acids, using low-cost reaction chip. The thermal cycler has been fabricated in a micro-assembly manner using flip–chip bonding technique, which is batch manufacturable with good reproducibility. Silicon heating blocks in the multi-chamber array are mounted on a printed-circuit-board (PCB) substrate, with the later attached to a metal plate heat sink. Thermal cross-talk has been minimised by using thin and low thermal conductivity PCB. The preferred reaction chip containing multiple chambers is made of low-cost plastics, while other PCR compatible materials are also possible. The preliminary experiments show that with up to 20 μl sample in the thermally formed plastic chip, a full speed of 8 min for 30-cycle PCR is achievable. Thermal cross-talk of as less as 0.2% is obtained with the very thin PCB substrate (500 μm, FR4) and the plastic chip (100 μm PET), in a standard format of multi-chamber array. A temperature fluctuation of ±0.1 °C has been achieved during thermal multiplexing of up to 16 chambers, with each chamber consuming an average heating power of no more than 1.2 W. Finite element analysis (FEA) is conducted to optimise the thermal performance of the cycler. Experiments are in well agreements with the simulations.

Introduction

Microfabricated thermal cyclers for nucleic acid amplifications by using polymerase chain reaction (PCR) technology have been developed with silicon/glass micromachining for several yeas, with improved cycling speed together with decreased sample volume consumption [1], [2], [3], [4], [5], [6], [7]. Multi-chamber thermal cycler was also reported for parallel processing with independent thermal protocols for 16 chambers [8], using multi-layer plastic process including metallisation/lithography and bonding, which makes it difficult for the reaction chip to be of low cost. Most of the developed PCR chips developed so far are not disposable due to the expensive fabrication cost for the large-size devices of at least a few microlitres of sample volume (e.g. 5–25 μl).

On the other hand, the conventional plastic PCR micro-plate can be of low cost, but the thermal speed is normally slow due to the thermal diffusion delay in the sample that is usually in a cone-shaped tube. The thermal diffusion distance in the sample and the plastic wall is usually a few millimetres above and hence the resulted thermal delay is very long (e.g. more than 20–30 s). Consequently, after reaching each setpoint during thermal cycling, one still needs to wait such a long time until reaching the balanced uniform temperature in the sample. This is the key reason why conventional PCR machines need 2–3 h to finish a 30–40 cycles reaction, even for a moderate sample volume of 5–25 μl.

In summary, disposable or low-cost (e.g. less than a few cents per reaction chamber) multi-chamber reaction chip with accelerated thermal pace (e.g. less than 15 min for 30 cycles) is still not available although highly preferred by the bio-analysis R&D/industry.

In this study, the reaction chip is made with a low-cost plastic process, while the thermal cycler for the temperature control is outside the reaction chip and thus reusable. The multiple chambers in the thermal cycler can be thermally isolated from each other by the low thermal conductivity printed-circuit-board (PCB) substrate and plastics. Each reaction chamber on the micro-chip is sitting on top of a high thermally conductive silicon block containing heater and temperature sensor. The plastic chip and the reaction chamber are so thin that the uniform temperature in the silicon block produces a same temperature in the reaction chamber in short time (e.g. less than 2 s). Thus, low cost and high performance can be obtained in the same time with this design. As a first prototype, the volume of each reaction chamber is in a range of 5–20 μl, as preferred by users. Less volume is achievable with faster thermal speed, but sample handling may be more difficult. While larger volume will drastically slow down the thermal speed (thermal delay is proportional to square of the thermal distance formed by the sample plus the plastic wall).

This paper starts with the description of the design of the multi-chamber thermal cycler, including thermal balancing and thermal isolation, followed by analytical and finite element analyses to predict and optimise the thermal model. Then the fabrication processes of the thermal cycler and the reaction micro-chip are introduced. After that, the device and system characterisations and comparison with the predictions will be illustrated, followed by a discussion and conclusion.

Section snippets

Thermal delay

The thermal delay in a one-side heating scheme can be illustrated in Fig. 1. High thermally conductive plate 1 can be regarded as a uniform heating source of constant temperature (as heating/sensing elements are embedded in this plate). The time period from immediately reaching the target temperature in plate 1 to reaching uniform temperature in the whole of plate 2 is defined as thermal delay td. According to diffusion theory, the delay time can be expressed asL∼Dtdwhere L and D (=λ/ρc with λ,

Fabrication

The fabrication of the thermal cycler can be illustrated in Fig. 6. The silicon wafer process starts with a dielectric thermal oxidation, followed by aluminium deposition and patterning to form heaters and temperature sensors. (1) Here aluminium alloy (with Si and Cu) of as thick as 2.5 μm has been used for good reliability. This is important especially for heater that will carry large current during thermal cycling. Then a passivation layer of silicon nitride is deposited using plasma enhanced

Experiments and discussions

The prototype of the 16-chamber thermal cycler is characterised. All temperature sensors are calibrated and show high and consistent temperature coefficient of resistance (TCR) in a range of 4060±32 ppm/°C. Resistance values for heaters and temperature sensors are 32±0.5 and 560±12 Ω at 20 °C. The control system includes the data acquisition (DAQ) cards communicating with the control circuit that can handle up to 32 W power at no more than 10 V power supply, a PC with LabView interface and PID tools

Conclusions

A multi-chamber thermal cycler, using low-cost reaction chip and of high performance, is designed, fabricated and characterised. The thermal cycler is made by micro-assembly technique using flip–chip bonding. Multiple chambers are mounted on a thermally non-conductive substrate for thermal isolation to each other, thus thermal multiplexing of multiple chambers is achievable. Reaction micro-chip can be any material whichever is of low cost in fabrication process. Very low-cost plastic micro-chip

Acknowledgements

The authors would like to thank all colleagues at Institute of Microelectronics, Singapore, for the wafer fabrication, package and device characterisation, and team members in the project for the long-term collaboration. This project is under Biosensor-Focused-Interest-Group (BFIG) which is financially supported by the National Science and Technology Board (NSTB) of Singapore.

References (8)

  • M.A. Northrup, et al., DNA amplification with a microfabricated reaction chamber, in: Proceedings of the Seventh...
  • T.W. Adam

    Functional integration of PCR amplification and capillary electrophoresis in a microfabricated DNA analysis device

    Anal. Chem.

    (1996)
  • S. Poser, et al., Chip elements for fast thermocycling, in: Proceedings of the Eurosensors X, Leuven, Belgium,...
  • B. Volker, et al., Miniaturised multi-chamber thermocycler, US Patent 5939312...
There are more references available in the full text version of this article.

Cited by (60)

  • Shunting microfluidic PCR device for rapid bacterial detection

    2020, Talanta
    Citation Excerpt :

    There are three main types of microfluidic PCR devices, namely: chamber-based stationary PCR; continuous-flow PCR; and shunting PCR. In all of these types the primary objective is to achieve faster thermal cycling and use of reduced volumes of expensive sample/reagents [10]. A number of chamber-based stationary PCR [3,5–8,11] and continuous-flow PCR microfluidic devices [12–18] have been reported.

  • Lab-on-a-chip: Emerging analytical platforms for immune-mediated diseases

    2013, Autoimmunity Reviews
    Citation Excerpt :

    Nevertheless, the materials and the fabrication processes are quite expensive. However, during the last decade there has been increased focus on polymer-based chips, including polydimethylsiloxane (PDMS) [14], polycarbonate (PC) [15], polymethylmethacrylate (PMMA) [16], polyimide [17], polyethylene terephthalate (PET) [18], SU-8 [19], polycyclic olefin [20], epoxy [21], and gene frame [22]. These materials are cheaper than silicon and more suitable for the production of disposable devices.

  • Miniaturized nucleic acid amplification systems for rapid and point-of-care diagnostics: A review

    2012, Analytica Chimica Acta
    Citation Excerpt :

    Certain parameters in microPCR systems like thermal mass, conductivity, flow rate, and channel cross-sectional area, which if optimized can lead to amplification in less than 5–10 min compared to 30 min to 2 h in conventional PCR systems. Major advantages of static chamber microPCR over the conventional PCR are rapid gene amplification, lower cost of reagents and system components, high throughput, and possibility of integration as a portable or hand-held molecular diagnostic device [29–31]. Flow-through microPCR does not require temperature cycling, resulting in a much faster amplification than static chamber microPCR system [32].

View all citing articles on Scopus
View full text