Rheology, commonly simplified by the property of viscosity, describes the flow of all fluids, from food and plastics, to coatings, adhesives, and 3D printing inks. While viscometers adequately probe Newtonian (constant) viscosity, most fluids have complex viscosity, requiring tests over multiple shear rates, and transient measurements. As a result, rheometers are typically large, expensive, and require additional infrastructure (e.g., gas lines), rendering them inaccessible for regular use by many individuals, small organizations, and educators. Here, we introduce a low-cost (under USD$200 bill of materials) Open Source Rheometer (OSR), constructed entirely from thermoplastic 3D printed components and off-the-shelf electromechanical components. A sample fluid rests in a cup while a microstepping motor rotates a tool inside the cup, applying strain-controlled shear flow. A load cell measures reaction torque exerted on the cup, and viscosity is calculated. To establish the measurement range, the viscosity of four Newtonian samples of 0.1–10 Pa.s were measured with the OSR and compared to benchmark values from a laboratory rheometer, showing under 23% error. Building on this, flow curves of three complex fluids – a microgel (hand sanitizer), foam (Gillette), and biopolymer solution (1% xanthan gum) – were measured with a similar error range. A further stress relaxation test was demonstrated on the biopolymer solution. The OSR cost is ∼1/25th that of commercially available devices with comparable minimum torque (200 µN.m), and provides a platform for further innovation in open-source rheometry.
Article
A Low-Cost, Open-Source Cylindrical Couette Rheometer
https://doi.org/10.21203/rs.3.rs-4590232/v1
This work is licensed under a CC BY 4.0 License
You are reading this latest preprint version
Viscosity, η, is the most fundamental rheological property of fluids, and it denotes the resistance to shearing flow. However, outside of a classroom this property is not typically constant, and in fact varies with time or applied stress. Additional rheological properties such as shear-thinning behavior, yield stress, hysteresis, aging, and thixotropy all substantially affect the flow behavior of most fluids (besides water, oil, and simple solvents) [1] [2] [3] [4]. Shear thinning occurs when viscosity decreases with increasing shear rate, and is instrumental to extrusion-based 3D printing of viscous inks [5], or for injection of concentrated antibody solutions through small needles [6]. In both cases, yield stress, then dictates whether the material will be able to stay in a desired location once ejected [7]. For another example, a fluid with a non-zero yield stress behaves as a soft solid when subjected to stresses below its yield stress. This phenomenon allows, toothpaste, for illustration, to be squeezed out of a toothpaste tube (shear flow) yet remain sitting on top of toothbrush bristles without sinking down (as a simple solid) [8] [9] [10], or for hair gel to be spread (shear flow) and then remain in place (solid), mechanically reinforcing a hairstyle of choice [11].
Everyday examples where viscosity and other rheological properties are important include concrete, painting, injection molding, food, and personal care products. As another specific example, hand sanitizers are carefully engineered to be both shear-thinning and to have a yield stress [11]. An ideal behavior is to have a high viscosity or yield stress to reduce phase separation during storage and transportation, as well as sufficient yield stress when dispensed onto a surface or hand, while allowing dispensing without requiring excessive force. When hand sanitizer is pumped, the narrow nozzle inlet applies a high shear rate, effectively lowering the viscosity of the shear-thinning fluid [11]. Therefore, the formulation must tailor the rheology while maintaining medical efficacy [12].
Understanding of viscosity is also instrumental to successful thermoplastic 3D printing by fused filament fabrication. If the viscosity of the filament is too high, the nozzle of the 3D printer can clog, and if the viscosity is too low, then material can leak out of the nozzle at a rate different than desired [13], leading to defects like stringing, surface roughness, or internal voids. Ideally, a 3D printed material would be shear thinning [13] so it stays in place after being deposited. The material used to make the prototype rheometer, poly-lactic acid (PLA), is shear thinning [14]. These properties remain important in applications of bioprinting, where often yield stress and thixotropy are used rather than temperature-based solidification to retain shape.
These and other rheological properties of fluids are essential to everyday life.
However, the ability to make accurate and robust rheological measurements is restricted for schools, hobbyists, and professionals, and even within research labs due to cost and requirements for facilities (e.g., air lines), space, and expertise, often precluding familiarity with rheology in general. A low-cost and open-source rheometer would allow students to explore and gain a hands-on understanding of these concepts, and would allow hobbyists to incorporate fluid properties locally and immediately into their playful designs of new materials e.g., for homemade hand sanitizers, foods, and locally-sourced recyclables [15].
Many existing commercial rheometers, such as the Discovery Hybrid Rheometer 3 (DHR-3; TA Instruments), follow similar design concepts. This includes a motor with controllable speed driving a spindle mounted on bearings, along with sensors to measure the torque [16]. Commercial, research-grade rheometers use high quality magnetic or air bearings and have delicate torque sensing in the instrumentation head or base [16], and can cost over USD$50k for initial purchase. Other companies (e.g., Brookfield, Grace) offer lower cost options such as the DVNext Rheometer (Brookfield) [17]. Most of these operate from the same design principles as the research-grade rheometers, but use lower-cost components, which typically reduces measurement ranges [18]. Even when priced in the thousands of dollars, these rheometers are still expensive and often still use proprietary software.
Within the field of rheology, there are several designs for low-cost and home-built measurement devices. These range from the simplest 50¢ rheometer to measure yield stress via the slump test [19] to a variety of custom devices for measuring highly specific rheological parameters, such as the critical stress required to remove viscoelastic material from a surface [20] or a method to measure extensional viscosity via a camera focused on fluid dripping from a nozzle [21]. A variety of 3D printed tools have been developed to use with commercial rheometers for tests of specific fluids [22] and flowing powders [23] [24].
Low-cost, open source designs for full rheometers are less widespread. Among existing prototypes, multiple approximate torque by measuring the electrical current of the driving direct current motor [25] [26] [27]. One such current draw device used a 3D printed cup with a mounted off-the shelf encoded DC motor, saving cost, but electrical noise limited the measurement range to viscosity > 100 Pa.s, barring measurement of most fluids [28]. B. Cherrington, et al developed a rheometer with a similar concept using external torque sensing instead of current draw, but it is large in size and sample volume, manufactured from metal using machining, and its performance was only characterized for glycerin, a Newtonian fluid [29]. Another low-cost rheometer was based on a vertically oscillating spring-mass-damper system, but this was found to be unreliable for measuring non-Newtonian fluids [30], likely due to elastic effects interfering with the oscillations.
The OSR presented in this article combines the external torque sensing concept of B. Cherrington’s work, with the addition of a flexure for massive noise reduction. (Please see [31] for an introduction to flexures.) This approach has been used, for example, in a high-resolution device to torsion test micro-wires [32]. By detailed mechanical design and reduction of electrical noise, the OSR offers a significant improvement in accuracy and versatility over these existing low-cost rheometer designs. In what follows we present the design, construction, and testing of our prototype open-source rotary rheometer costing less than $200, which can rotate at speeds ranging from 10 to 60 RPM and has been tested with fluids ranging from 0.1 to 10 Pa.s along with an array of non-Newtonian fluids.
2.1 Major Design and Layout
Rotary rheometers contain a sample fluid between the surface of two solid objects and rotate one versus the other, shearing the enclosed fluid to make a measurement. One of the most typical geometries for this uses concentric cylinders, in which one rotates concentrically inside another hollow cylinder, the cup, as shown in Fig. 1. Other typical measurement geometries include parallel plates, which use a lower stationary plate and an upper rotating disk, and the cone-and-plate geometry, which uses a lower stationary plate and an upper rotating cone with a small angle to keep shear rate constant per radius. For concentric cylinders, the shear rates and shear stresses can be calculated from the applied and measured rotation rate, torque, and known measurement geometry [33]. Viscosity (η) in SI units of Pascal-seconds is defined as the ratio of shear stress (σ) with units of Pascals and shear rate (˙γ) with units of inverse seconds as shown in Eq. (1) [34].
While the shear rate, shear stress, and viscosity are used for developing material models, what is measured at the machine level is the torque and tool rotation rate, and then conversion factors are used based on geometry and flow state to extract the material values. Thus, the error of measurement is due both to the ability to measure raw values precisely and accurately, and the ability to accurately convert data into material units. We focus first on the former problem. The error due to the measurement geometry depends on the error of motion of the rotating surface, which varies with the angular runout of the spindle. The runout will be similar for concentric cylinders, in which the measurement surface is concentric with the axis of rotation of the spindle, and for parallel plates or cone-and-plate setups, for which the measurement surface is orthogonal to the axis. However, the error is compounded by influence of the free surface, or sample-air interface, which strongly influences measurements in parallel plate geometries, but is mostly removed in concentric cylinder setups, which are fully submerged. As a second major consideration, the torque measurement is proportional to the surface area and scales typically cubically with the radius of the rotating element, both of which are easier to make large in concentric cylinder geometries compared to other standard tools. For these reasons, the base OSR was designed to use a rotating cylindrical geometry in a stationary cylindrical cup.
Next, we consider the problem of accurate unit conversion. Shear rate and shear stress for a Newtonian fluid were calculated for this geometry with Eq. (2) and Eq. (3) respectively [33].
Here, Ω is the rotation rate in radians per second, T is the torque in Newton-meters required to rotate the cylinder at Ω, and RCylinder, RCup, and LCylinder represent the radius of the cylinder and cup, and the length of the cylinder respectively, in meters. For Eq. (3) to hold, the cylinder should be submerged in the fluid with fluid volumes both above and below the extent of the cylinder so there are no measurement artifacts associated with horizontal boundaries.
The tools used to contact the fluid, either a cylinder as shown or a vane fixture, were also designed to be 3D printed. Such methods have successfully generated other test geometries for rheometry [33] [35].
2.2 Mechanical Design
The measurement design is based on applying an input rotation to the sample and measuring the reaction torque, also known as a “controlled strain” measurement paradigm. To accomplish this, the cup has to have one degree of freedom alone: rotation about the z-axis. The cruciform flexure design shown in Fig. 3 was chosen for this purpose. Then, as the testing cylinder rotates in the sample, the flexure rotates freely, and an externally mounted loadcell measures the reaction torque.
This results in the overall design of the OSR (Fig. 1, 2), which has a sample sheared between concentric cylinders, using a strain-controlled motor above the sample, and a flexure and loadcell below and outside the cup for measurement of resulting torque, with 11 major components.
The base is printed as one part with the cup and flexure attached to prevent assembly misalignment errors. The motor mount slides into a circular pocket on the base for horizontal alignment. In this version screws were used to rigidly attach the motor mount to the base. The attachment points also serve to vertically align the motor mount and to align the cylinder within the cup.
A stepper motor was selected to obtain a wide usable speed range. Using a stepper motor required addressing concerns about non-continuous rotation. To simulate continuous rotation, this setup utilizes a gearbox and micro stepping [36]. To reduce the runout, the drive train consists of a flexible shaft coupler with 2 bearings to reduce the motor’s influence on runout. To give this rheometer a wide selection of adequate commercial bearings to choose from, this design uses an 8mm shaft compatible with the widely available 608 “skateboard” bearings.
As shown in Fig. 1, a ball-point set screw is threaded into the loadcell to give a well-defined contact point. It is secured with Loctite so the set screw does not move within the threaded hole during testing. The ball then interacts with a hard ceramic surface (a glass mirror) glued onto a tab on the printed cup. The mirror provides a hard surface for the ball to rest against. This prevents the risk of long-term creep indenting the surface and adding error to measurements, as would be expected with less-hard metal or plastic surfaces. To ensure constant contact, the set screw is tightened an additional half-turn after contacting the mirror to apply a small preload. A computer-rendered video of the assembly process is included in the Supporting Information as S1 Video.
The flexure was designed to be much less stiff in rotation than the 70 Nm/rad loadcell. The performance of the flexure was verified (Fig S-1), ensuring small stiffness and linear response over the intended range of deformation (≈ 0.1 rad), though the actual flexure had half the stiffness of the FEA prediction (0.32 vs 0.57 Nm/rad), likely due to variations in the true material properties and mild porosity and dimensional change in the fabricated part. As a result, the flexure is fragile, even more so due to it being printed vertically, meaning the layer lines are in the weakest orientation for this part. However, this weakness is required to enable high resolution measurements with the loadcell. To protect the flexure when not testing, locks were designed to mount on the 4 dovetail joints on the base. These locks can be seen in Fig. 2c.
2.3 Electrical Components
The main electrical components included in the rheometer design are an Arduino Uno, HX711 loadcell amplifier, beam loadcell, TMC2209 stepper controller, and geared stepper motor. The HX711 takes the analog signal from the loadcell, amplifies it, converts it to a digital signal with a 24 bit analog to digital converter, and sends it to the Arduino via a serial connection [37]. It is unknown exactly what units the data the HX711 sends to the Arduino is in, but it is unimportant for this endeavor as the conversions are linear and the calibration process ensures the data is converted into the correct units. The Arduino also controls the stepper controller which controls the stepper motor. A schematic of the electrical connections is included as Fig S-2.
3.1 Printing and Assembly
The custom components of this design were mostly 3D printed on a Prusa Mini printer, with the motor mount printed on an Ultimaker S5 due to its larger size. Components in contact with the sample were printed with 4 shells to prevent leaking by increasing the number of fully filled layers before the infill. Support material was used for the cup, motor mount, and flexure locks. The slicer-generated supports did not adequately constrain the flexure, so custom supports were added directly the computer-aided design (CAD) file. After printing, the cylinder was reamed using a mill to just over 8mm to fit the shaft and the motor mount was reamed to 22mm to fit the bearings. To ensure the bearing flanges were resting against a flat surface, the faces of the motor mount in contact with the bearings were faced on a mill. The set screw holes in the cylinder also had to be reamed and tapped to accept 10–24 set screws. an assembly and use guide is included in Supporting Information as S1 File and 3D-Printable STEP files are included in Supporting Information as S2 File.
3.2 Testing
With the method shown in Fig. 4, seven fluids were tested: four viscosity standards with
Newtonian (constant) viscosities of 10, 1, 0.5, and 0.1 Pa.s, as well as three non-Newtonian fluids: Hydra Pearl hand sanitizer, Gillette Original shaving foam, and a 1% xanthan gum solution in water. These viscosity standards were chosen to determine the accuracy of this device and the hand sanitizer, shaving foam, and xanthan gum were chosen to demonstrate how this device can be used with more realistic (and interesting) non-Newtonian fluids. The viscosities from these tests were found using Eq. (1), Eq. (2), and Eq. (3) with rotation rate and loadcell data as well as Eq. (4) from the loadcell calibration testing.
The OSR was secured to a wooden base for stability, on a padded film for vibration damping. To test a fluid with the OSR, the steps in Fig. 4 were established and followed. First, the fluid under test is filled to or slightly above the fill line on the cup. Second, the flexure locks are removed. Third, the motor mount is screwed onto the 3D printed base. Finally, for testing, the ambient temperature of fluids was measured because viscosity is temperature dependent. The temperature was found to be around 19°C within our laboratory and the sample viscosity values used for comparison were adjusted or measured at 19°C with a laboratory DHR-3 rheometer. For a testing cycle, the OSR program code tares the value read from the loadcell and then ramps the motor speed up to 60 RPM before running through a sequence of up to 5 different speeds per decade decreasing from 60 to 0.25 RPM, dwelling at each speed for 15 seconds. Only 9 seconds of the 15 seconds of data are used for the viscosity calculations because the first 5 seconds and last second of each speed were removed to ensure that the sample was at steady state and to avoid transient fluctuating data from changes in motor speed. Due to oscillations discussed in Design Limitations 5.1, only the data from 60 RPM to 10 RPM is usable. A video of the testing process is included in Supporting Information as S2 Video. Sample code to run this test is provided in Supporting Information S3 File and S4 File with more detailed analysis code in MATLAB provided in Supporting Information as S5 File.
3.3 Loadcell Calibration
To interpret the raw loadcell data output, the loadcell was calibrated by two independent methods. First, it was calibrated by hanging weights of different values when the loadcell was horizontal. The loadcell was secured to a base, and using a thin gauge wire, an increasing amount of weight was suspended from the set screw that would normally be in contact with the mirror during viscosity testing. The results are shown in Fig. 5a. The fit is very close to linear with a coefficient of determination r2 = 0.9998. To assess the conversion factor for the finally assembled OSR, the sensing head of a DHR3 rheometer was directly used in conjunction with the sensing base of the OSR, in which the torque was simultaneously measured using both machines’ sensors through an intermediary fluid (Fig. 5b). Tests were conducted with a 0.5 Pa.s calibration fluid for sequential constant rotation rates of 0.01 − 10 rad/s in 7 log-spaced steps. This provided calibration down to the noise floor level of the OSR (Fig. 5a). Based on the 95% confidence interval of the best fit equations, this conversion equation is not statistically different from the one found by hanging weights off the loadcell, although it proceeds to much lower torque and loadcell values, indicating the impressive range of measurement of both the OSR and DHR compared to typically accessible weights in the lab. The conversion equation in general for the torque T is
T = Ls (4)
for loadcell value s and calibration slope L. The calibration value L would be
expected to change for different loadcells, different mountings of the same loadcell, and changes in cup geometry. In combination with Eq. (3) from geometric factors of the design, Eq. (4) allows calculation of the material shear stress from the output loadcell value.
From hanging weights off the loadcell, L = 4.6 × 10− 8 ± 0.05 × 10− 8 Nm while the DHR test gave L = 4.7 × 10− 8 ± 0.08 × 10− 8 Nm. Eq. (4), with Torque in Nm, from the DHR loadcell calibration testing was used for analysis since it accounts for manufacturing and assembly variations in the lever arm to the loadcell. If a secondary rheometer or weights are not available, the rheometer could also be calibrated by measuring a fluid of known viscosity, such as glycerol. This is possible because the loadcell response is known to be linear over a wide range of torque.
Four constant-viscosity samples were tested, measuring shear stress and viscosity as a function of shear rate (Fig. 6). For all four samples, the average measured viscosity was within 20% of the actual viscosity. For the 0.5 Pa.s sample it was within 1%. The RMSE of the viscosity was less than 23% of the actual viscosity for every sample except for the 10 Pa.s sample which was 43%. The 0.1 Pa.s sample only had one usable data point because at speeds lower than 40 RPM, the torque value could not be differentiated from the system noise of 3 × 10− 5 N.m, which for our design and geometry sets a noise floor of 2 Pa.
Next, we tested three non-Newtonian fluids: hand sanitizer, shaving cream, and a solution of xanthan gum (Fig. 7). With the cylindrical geometry, all three samples showed signs of slip which skewed the averages for the slower shear rates and narrowed the range of usable data. Slip occurs when the fluid moves relative to the testing bob. This presents as a falsely low shear stress at lower shear rates due to the sliding of the sample along the smooth cylinder rather than following robust non-slip boundary conditions [33]. For all samples the RMSE of the OSR data compared to the DHR fit was ≤ 7.6 for shear stress and ≤ 2.5 for viscosity. The yield stress was also estimated using a model fit. The hand sanitizer and shaving cream OSR estimates were within 93% of the DHR estimate and the xanthan gum OSR estimate was within 260% of the DHR estimate. The hand sanitizer sample was also tested with an 8-arm vane geometry which prevented slip. This also improved accuracy with the yield stress changing from
1.6 to 32.6 Pa compared to the DHR estimate of 24.3 Pa.
Figure 8 shows the results of a transient step strain and stress relaxation test with the xanthan gum. This data is all in the linear range of stress relaxation so it cannot be compared to previous literature evaluating damping functions [38], but it shows good repeatability with the two tests at 63% strain and with the progression of relaxation for each test.
5.1 Design Overview
When compared to selected commercial rheometers and other prototypes (Table 1), the OSR achieves a moderate range of speeds, and is most impressive at measuring relatively low torques for relatively low cost. While there is room for improvement with further decrease of the minimum torque measurement and improvement in widening the accessible speed range, it still has shown adequate performance within typical measurement ranges. We detail below specific limitations of the current design and suggestions for further improvement to continue to push the limits without driving up inordinate cost.
| OSR (here) | Current Draw | DVNext | DHR-3 | ARES-G2 | Target | |
|---|---|---|---|---|---|---|
| Speed (RPM) | 10–60 | 20–90 | 0.01–250 | 0–2,800 | 10− 5 − 2,800 | 0.01–100 |
| Min Torque (Nm) | 7 × 10− 5 | 7 × 10− 2 | 7 × 10− 5 | 6 × 10− 9† | 10− 7 | 10− 5 |
| Cost ($USD) | < 200 | < 150 | > 5,000 | > 50,000 | > 100,000 | 200 |
| Source | − − − | [28] | [17] | [16] | [39] | − − − |
| †Practical Limit: 2 × 10− 6 | ||||||
5.2 Data validation
The measurements taken by the rheometer as used in this paper were compared to those made by a commercial DHR-3 rheometer for partial calibration and error analysis. Potential users of the OSR may not have access to secondary measurements. In these cases, testing of a viscosity standard available at the grocery store such as glycerol would be recommended, or testing non-Newtonian samples such as Gillette Original shaving cream and a 1% xanthan gum solution as demonstrated in this paper, which typically have reproducible properties. The design and process currently do not have feedback for data quality, and testing in the lab has shown that innocuous steps such as cleaning the cup can affect the measured signal if any fluid becomes trapped between the loadcell and the ceramic mirror surface. In addition, a currently manual data analysis process adds time to processing, making the OSR results fully reproducible but potentially challenging for a novice rheologist to use. In the case that a standard rheometer is not available for data validation, we recommend testing a viscosity standard for each new setup of the instrument to verify performance, and for users to repeat measurements multiple times, discarding datasets that behave differently, until the user gains more familiarity.
These inherent needs for some expertise to use prototype designs is offset by the customizability of the 3D-printable open-source design. A team at the University of Western Ontario has fully recreated and tested our design. When they found issues for their particular applications, they were able to quickly make small design changes: altering the width and thickness of the flexure, and separating the loadcell and cup into separate parts for ease of cleaning (data not shown). They then printed their new version of the OSR with customized usability and functionality to fit their applications. Thus, this OSR design should be considered a usable base to be modified to fit for any specialized applications.
5.3 System Noise
Fig S-3a displays the raw data from the loadcell during a flow curve measurement, omitting only the initial ramp up to the initial speed of 60 RPM, for a test with a 10 Pa.s constant-viscosity liquid. Ideally, the loadcell value should be a step function, and it is, except for some oscillations occurring at each step due to device runout. Fig S-3b focuses on the oscillation on the first step, at a rotation rate of 60 RPM. Though magnified and clearly visible, these have an amplitude < 0.5% of the signal.
The oscillation frequency of approximately 1 Hz correlates with the motor rotation frequency of 1 rotation per second at 60 RPM. Fig S-3b shows how the loadcell reading consistently varies with the angular position of the cylinder. The oscillations in the other steps also happen once per rotation. These oscillations are likely due to small eccentricities in the drive train and cup. Due to the oscillations, data collected for speeds slower than 6.7 RPM is less reliable because slower speeds do not complete a full rotation during the 9 seconds of data collection that are analyzed and small torque values could be obscured by the oscillations. It is also possible that the unreadable data was caused by the stepper not rotating smoothly enough at small rotation rates with only eight micro steps per step (i.e. 1600 micro steps per full rotation) to generate a steady shear flow.
5.4 Improving Data Quality
We suggest potential refinements to further reduce mechanical noise in the system. Both the rotational runout and vibrations in the system likely contribute. To address these, parts where the bearing holes in the motor mount and the shaft hole in the cylinder should be machined on a lathe. A stepper motor without a gearbox should be tested, and the finest usable micro stepping setting should be used for each rotation rate to ensure the smoothest possible rotation. The stepper motor should also be mounted with more isolation from the measurement system (e.g., using rubber washers). Padding below the rheometer could also help isolate from the less significant environmental noise, and the loadcell and flexure may be fully enclosed at least during data collection.
The openness of this design allows for easy modification to create new test bob geometries matching the application and improving data quality for each specific use case. Potentially by increasing force-multiplying potential or resistance to slip. The design can be further modified to work with a parallel plate geometry for testing lower viscosity samples, or for samples with lower volume. With that geometry, runout is more disvantageous, though a wider gap may be used. Similarly, concentric cylinders with a larger radius and a smaller gap would improve the measurement threshold but also require lower runout of the system and cylinder surfaces with post-machining to ensure their cylindricity compared to an as-printed part.
5.5 Future Work
Currently, the OSR design utilizes a few large parts to limit stack-up of error from assembling the parts, but an ideal final design may be printable entirely on smaller 3D printers and have smaller print times to reduce the material cost and time of 3D printing replacement parts, especially for the fragile flexure, and to allow for faster innovation of new attachments and geometries for test. We would also recommend a future design have the ability to wash the cup without taking the flexure and 3D printed base assembly off the wooden base. Common procedures like cleaning the cup ideally could be accomplished without additional tools. To this end, the 3D printed parts should be split into smaller components, but this needs to be done carefully and gradually as changes could increase the noise, failure points, and sources of non-concentricity in the system, adding unwanted error to measurements in a tradeoff for usability.
In addition, there is a large potential to combine open-source designs like this with other common open-source tools for other types of measurements, including Raman spectroscopy with OpenRAMAN [40] for rheo-Raman tests, microscopy with OpenFlexure [41] for visualizing the microstructure of test liquids under flow, liquid dispensing [42] for automated loading, and more, as well as adding additional motors for orthogonal superposition rheometry, testing effects of novel cup and tool geometries, easily adjusting tool sizes to handle different fluids, and performing low-cost tests of irreversibly hardening or destructive materials for which the cup and cylinder used for measurement can be treated as disposable and easily replaced. Industries that deal with more volatile compounds such as pyrotechnics research can even consider the entire rheometer disposable when necessary.
We introduced and demonstrated the design of a low-cost Open Source Rheometer (OSR) constructed entirely from thermoplastic 3D printed components and off-the-shelf electromechanical components. We determined the areas of the design most critical for robust measurements, and highlighted areas still in need of improvement. We showed that the OSR measured viscosities between 10 and 0.1 Pa.s with the average being within 23% of baseline values from the manufacturer or values measured independently on a commercial laboratory DHR-3 rheometer. In addition, we showed that the OSR can measure three shear-thinning fluids with a RMSE of ≤ 2.5 Pa.s for viscosity. This OSR design can be used as-is, or is suitable for engineers, hobbyists, and tinkerers to modify for specific needs and applications. For example, a future version of this design could prove to be an inexpensive, more customizable replacement or additional tool-set for research-grade rheometers, while also lowering the cost of rheometry to a level acceptable for many schools and hobbyists.
Acknowledgments
Research was supported in part by MIT’s Sandbox Entrepreneurial Fund. M.E. was supported by funding from MIT’s Undergraduate Research Opportunities Program. This work was further supported by the machining facilities at MIT’s Laboratory for
Manufacturing and Productivity. C.E.O. was supported by a MathWorks Engineering Fellowship from MIT. The authors thank Gareth McKinley for helpful discussions about the rheology. The authors thank Joshua Pearce, Apoorv Kulkarni, Anjutha Selvaraj, and Lewis Klug from the University of Western Ontario for the valuable feedback they provided as the built, tested, modified, and retested designs based on those in this paper.
Author Contributions
C.E.O. conceived the presented idea. C.E.O., M.E., A.J.H, and D.T. contributed to the equipment design. M.E. performed the parts selection, program, and data analysis.
C.E.O. interpreted the data and designed the experiments to test the initial prototype.
C.E.O., M.E., A.J.H, and D.T. contributed to the final manuscript.
Additional Information
The authors CEO and AJH declare US utility patent US11698330B2 on fractal vane structures. The authors otherwise declare no conflict of interest.
Data Availability
The datasets used and analyzed during the current study are available from the corresponding author upon reasonable request.
- Larson RG, Wei Y. A review of thixotropy and its rheological modeling. Journal of Rheology. 2019;63(3):477–501. doi:10.1122/1.5055031.
- Coussot P, Nguyen QD, Huynh HT, Bonn D. Viscosity bifurcation in thixotropic, yielding fluids. Journal of Rheology. 2002;46(3):573–589. doi:10.1122/1.1459447.
- Coussot P, Tocquer L, Lanos C, Ovarlez G. Macroscopic vs. local rheology of yield stress fluids. Journal of Non-Newtonian Fluid Mechanics. 2009;158(1):85–90. doi:10.1016/j.jnnfm.2008.08.003.
- Malkin AY, Isayev A. Rheology - Concept, Methods, and Applications (3rd Edition). 3rd ed. ChemTec Publishing; 2017. Available from: https://app-knovel-com.libproxy.mit.edu/kn/resources/kpRCMAE012/ toc?cid=kpRCMAE012.
- Wei P, Leng H, Chen Q, Advincula RC, Pentzer EB. Reprocessable 3D-Printed Conductive Elastomeric Composite Foams for Strain and Gas Sensing. ACS Applied Polymer Materials. 2019;1(4):885–892. doi:10.1021/acsapm.9b00118.
- Goswami S, Wang W, Arakawa T, Ohtake S. Developments and Challenges for mAb-Based Therapeutics. Antibodies. 2013;2(3):452–500. doi:10.3390/antib2030452.
- Milazzo M, Fitzpatrick V, Owens CE, Carraretto IM, McKinley GH, Kaplan DL, et al. 3D Printability of Silk/Hydroxyapatite Composites for Microprosthetic Applications. ACS Biomaterials Science & Engineering. 2023;doi:10.1021/acsbiomaterials.2c01357.
- Toothpaste Rheology Methods and Approaches; 2023. Available from: https://www.rheologylab.com/articles/cosmetic-personal-care/ toothpaste-rheology/.
- Viscosity measurement in the personal care industry | Anton Paar Wiki; 2023. Available from: https://wiki.anton-paar.com/us-en/ viscosity-measurement-in-the-personal-care-industry/.
- Toksoy MO, Tugcu-Demir¨oz FN, Tirnaksiz F. Evaluation of rheological and textural properties of toothpastes. 2013;38:135–141.
- The Viscosity of Hand Sanitisers - Perception, performance and packaging; 2022. Available from: https://www.rheologylab.com/articles/household/hand-sanitisers/.
- Silva AF, Wood TA, Hodgson DJM, Royer JR, Thijssen JHJ, Lips A, et al. Rheological design of thickened alcohol-based hand rubs. Rheologica Acta. 2022;61(8):571–581. doi:10.1007/s00397-022-01347-y.
- Azad MA, Olawuni D, Kimbell G, Badruddoza AZM, Hossain MS, Sultana T. Polymers for Extrusion-Based 3D Printing of Pharmaceuticals: A Holistic Materials–Process Perspective. Pharmaceutics. 2020;12(2):124. doi:10.3390/pharmaceutics12020124.
- Nofar M, Salehiyan R, Sinha Ray S. Rheology of poly (lactic acid)-based systems. Polymer Reviews. 2019;59(3):465–509. doi:10.1080/15583724.2019.1572185.
- Materiom. Materiom : Home; 2023. Available from: https://materiom.org/.
- Discovery HR-3 - TA Instruments; 2016. Available from: https://www.tainstruments.com/dhr-3/.
- DVNext Rheometer; 2023. Available from: https://www.brookfieldengineering.com/products/viscometers/ laboratory-viscometers/dvnext-rheometer.
- M5600 Brochure.pdf; 2023. Available from: https://gic.com/brochures/M5600%20Brochure.pdf.
- Pashias N, Boger DV, Summers J, Glenister DJ. A fifty cent rheometer for yield stress measurement. Journal of Rheology. 1996;40(6):1179–1189. doi:10.1122/1.550780.
- Fernandes RR, Suleiman N, Wilson DI. In-situ measurement of the critical stress of viscoplastic soil layers. Journal of Food Engineering. 2021;303:110568. doi:10.1016/j.jfoodeng.2021.110568.
- Dinic J, Zhang Y, Jimenez LN, Sharma V. Extensional Relaxation Times of Dilute, Aqueous Polymer Solutions. ACS Macro Letters. 2015;4(7):804–808. doi:10.1021/acsmacrolett.5b00393.
- Bikos DA, Mason TG. Customizable tool geometries by additive manufacturing for mechanical rheometry of soft matter. Journal of Rheology. 2016;60(6):1257–1267. doi:10.1122/1.4963802.
- Kadhim K. Twin screw wet granulation of pharmaceutical powders: Effects of drying on granules and tablet attributes [phd]. University of Sheffield; 2021. Available from: https://etheses.whiterose.ac.uk/29634/.
- Moose RC. Improvements in Powder Rheometry Through Novel 3D Printed Tools [Thesis]. Massachusetts Institute of Technology; 2022. Available from: https://dspace.mit.edu/handle/1721.1/146072.
- Seckin T, Kormaly SM. An Easy-to-Build Rotational Viscometer with Digital Readout. Journal of Chemical Education. 1996;73(2):193. doi:10.1021/ed073p193.
- Rabani A, Challis R. A low-cost viscometer based on a permanent magnet dc motor. Measurement Science and Technology. 2013;24(3):035304. doi:10.1088/0957-0233/24/3/035304.
- Souza R, K¨onig A. MANUFACTURING OF A VISCOMETER FOR MONITORING OIL CONTENT IN POLLUTED WATER. HOLOS. 2012;6. doi:10.15628/holos.2012.1094.
- Erni MF. Electro-Mechanical Design and Testing of a Prototype Low-Cost Rheometer [Undergraduate]. Massachusetts Institute of Technology; 2022. Available from: https://dspace.mit.edu/handle/1721.1/144832?show=full.
- Cherrington B, Rothstein J, Carnasciali MI, Daniels SD. Building and Validating a Rotational Viscometer; 2015. Available from: https://www.semanticscholar.org/paper/Building-and-Validating-a-Rotational-Viscometer-Cherrington-Rothstein/ 6b10c05956638a2ddd94abe64c0d737b457fcdcd.
- Brown University. Low-Cost Rheometer; 2022. Available from: https://www.instructables.com/Low-Cost-Rheometer/.
- Hopkins JB, Culpepper ML. Synthesis of precision serial flexure systems using freedom and constraint topologies (FACT). Precision Engineering. 2011;35(4):638–649. doi:10.1016/j.precisioneng.2011.04.006.
- Fick T, Harrison P, King K, Miller G. A Novel Device for Tension-Torsion Testing of Micro-Wires. 2008;.
- Owens CE, Hart AJ, McKinley GH. Improved rheometry of yield stress fluids using bespoke fractal 3D printed vanes. Journal of Rheology. 2020;64(3):643–662. doi:10.1122/1.5132340.
- Boyes W, editor. Instrumentation reference book. 4th ed. Amsterdam ; Boston: Butterworth-Heinemann/Elsevier; 2010.
- Bikos DA, Mason TG. Customizable tool geometries by additive manufacturing for mechanical rheometry of soft matter. Journal of Rheology. 2016;60(6):1257–1267. doi:10.1122/1.4963802.
- Mitra D. How to Improve Motion Smoothness and Accuracy of Stepper Motors; 2022. Available from: https://www.ti.com/lit/an/sloa293a/sloa293a.pdf? ts=1635012807439&ref_url=https%253A%252F%252Fwww.google.com%252F.
- SparkFun Load Cell Amplifier - HX711 - SEN-13879 - SparkFun Electronics; 2023. Available from: https://www.sparkfun.com/products/13879.
- Jaishankar A, McKinley GH. A fractional K-BKZ constitutive formulation for describing the nonlinear rheology of multiscale complex fluids. Journal of Rheology. 2014;58(6):1751–1788. doi:10.1122/1.4892114.
- ARES-G2 - TA Instruments; 2015. Available from: https://www.tainstruments.com/ares-g2/.
- OpenRAMAN. OpenRAMAN; 2023. Available from: https://www.open-raman.org/.
- Sharkey JP, Foo DCW, Kabla A, Baumberg JJ, Bowman RW. A one-piece 3D printed flexure translation stage for open-source microscopy. Review of Scientific Instruments. 2016;87(2):025104. doi:10.1063/1.4941068.
- Keesey R, LeSuer R, Schrier J. Sidekick: A Low-Cost Open-Source 3D-printed liquid dispensing robot. HardwareX. 2022;12:e00319. doi:10.1016/j.ohx.2022.e00319.
Competing interest reported. The authors CEO and AJH declare US utility patent US11698330B2 on fractal vane structures. The authors otherwise declare no conflict of interest.