A comparison of mechanical properties between different percentage layups of a single-style carbon fibre ankle foot orthosis

Abstract

Background:

Currently, a range of ‘off-the-shelf’ ankle foot orthoses are used in clinical practice, of various functions and designs. Their use relates to immediate control over mild conditions.

Objectives:

To investigate the properties of carbon fibre ankle foot orthoses at different percentage layups and provide a comparison of these through assessment of the (1) elastic properties, (2) deflection about the ankle (including the calculation of stiffness) and (3) failure under compressive forces (dorsiflexion).

Study design:

Experimental, bench test.

Methods:

Literature was reviewed to derive a suitable bench test for mechanical testing of ankle foot orthoses. Two universal Instron machines were used to apply the necessary forces. A pilot device was utilised to establish the range of forces appropriate to confirm the setup chosen was effective. Each test was then carried out on nine ankle foot orthoses (3 × 3 different percentage layups).

Results:

All nine devices had their elastic properties deduced. Stiffness exhibited greater resistance in tension, with angular deflection being greatest in the ‘Lite’ set and least in the Rigid. Failure occurred mainly due to fracture, proximally on the strut; however, this was not consistent among the devices.

Conclusion:

Results confirmed the properties expected of carbon fibre ankle foot orthoses were consistent. This can now be related to functionality and therefore specific device prescription options.

Clinical relevance

This article attempts to increase the understanding and develop the area of mechanically testing ankle foot orthoses. This was achieved by comparing carbon fibre at different percentage layups on an identical design and their resultant structural properties. This article outlines a clear and simple setup for obtaining repeatable results.

Keywords

Ankle foot orthosis orthotic device testing properties ability carbon fibre stock off the shelf prefabricated setup layup stiffness elastic energy return failure destruction mechanical universal testing machine Instron

Background

Ankle foot orthoses (AFOs) are a common treatment for a variety of lower limb conditions to improve stability in stance and/or swing phases during pathological gait.^1–9 Depending on the severity, this will have an impact on the AFO prescription, mainly surrounding its design, geometry and material choice.⁷ Currently, there are no definitive methods for developing a mechanical test for AFOs,⁸ with the decision left to the clinician¹⁰ or in the case of stock devices, the manufacturer.¹¹ Stock products have numerous designs with varying functions, and are commonly used to control flaccid equinus (drop foot)^1,4,6,12 during swing, with minor control over stance elements.⁶

The main aim of this study was to examine essential areas required in a test to develop a baseline testing protocol that can be more universally applied to all AFOs allowing comparable measures of ability between devices to benefit clinical decisions.

Despite no internationally recognised testing standards,⁸ many have documented the main properties to be examined in AFOs to be stiffness, rigidity and flexibility.^{2,3,8–11,13} A review by Kobayashi et al.⁹ noted the differing setups used to obtain a better understanding of the mechanical properties. This should be more effective and afford tailoring patient treatment needs more specifically.³ With no routine objective method for mechanical testing,¹³ most setups are research specific. The most common processes recorded were bench setups,⁹ which are difficult to replicate without specialist equipment and thus reduce options for clinical applications.

Carbon fibre (CF) has become a widely available material choice for orthotic manufacture, being lighter and stiffer with a ‘greater strength-to-weight ratio’⁴ than alternatives. ‘Carbon fibre also has energy-storing capabilities’^4,11 due to its elasticity, so should not yield with forces applied, assuming no limits are exceeded. This article looks at and attempts to prove these characteristics. It was suspected that carbon fibre ankle foot orthoses (CF-AFOs) would exhibit energy-storing capabilities, high stiffness values and failure at the ankle due to dorsiflexion.

Method

No ethical permission was required as this study involved purely mechanical testing of stock AFOs.

It was necessary to design a testing procedure for the different CF-AFOs, as no guidelines existed within the literature. A review of this literature^1–18 suggested the following relevant test areas:

Elastic properties (measure of devices’ ability to exhibit any permanent deformation after a force is applied);

Angular deflection, by calculating stiffness around the ankle;

Failure under compressive forces (dorsiflexion moment). This would be the expected failure mechanism during normal gait and was agreed by the manufacturer.

The manufacturer Orthotic Composites (UK based) provided 10 stock CF-AFOs (partial spiral posterior strut devices, known as Helix AFOs) with different amounts of layering of CF material for each set. Three ‘Lite’ (−30% layup), three ‘Standard’ and three ‘Rigid’ (+30% layup) devices were procured with one additional standard device being provided to develop the initial setup to determine the testing parameters were appropriate.

The following precautions were taken to limit variability between devices:

Manufacture of all devices, undertaken by the same technician at the manufacturers’ workshop with identical design trimlines (only changing the percentage layup for each set).

All devices were identical; left side, medium sized, made from the same master cast.

Consistent materials were used in making each device; the raw pre-impregnated CF came from one supplier and the same epoxy resin used throughout. A 42% resin to fibre ratio was used, curing at 120°C, for the same duration of 2 h.

Measurements were established for a calf-block with hinge, an ankle point and footplate trim locations. The centre point of the calf-block was extended distally to establish the ankle centre. The footplate had locating holes drilled approximately at the metatarsal head section. (All alterations/trims/drilling were kept to a minimum; Figure 1). This became the limb model to mimic the lower leg.

During assembly, the calf-block and attachment plates were hand tightened systematically, with points marked on jubilee clips, screws and C-clamps (established from trial setup), to ensure as thorough consistency as possible.

Figure 1.

AFO setup.

Literature^{1,2,4–9,11,13–15} was reviewed pertaining to AFO testing to deduce the best setup for the most consistent, accurate results. A bench test was chosen, implemented as a static laboratory mechanical test design and external conditions replicated similar to other studies.^{1,2,4–7,9,14} Kobayashi et al.⁹ note this approach provides ‘accurate repetitive measurement of AFO stiffness’⁹ and ‘can perform repeatedly under well-controlled’⁹ conditions when using mechanical testing machines.

To replicate the applied loads, two universal testing machines, the Instron E10000 and 5969, were utilised. Because E10000 had limited crosshead movement of ±30 mm, 5969 was used for failure testing.

Since no universal testing standards for ankle foot orthotics was available, the International Standards Organization (ISO)¹⁶ for testing prosthetic devices around the ankle and foot was reviewed in conjunction with related literature. This formed the basis for testing. In keeping with ISO guidelines for prosthetic feet and ankles, all tests were set out to first replicate plantarflexion (tension) and then dorsiflexion (compression) moments: ‘force initially to the heel and subsequently to the forefoot’.¹⁶ The attempt was to try to represent the gait cycle from initial contact until toe off. This was therefore chosen as the testing order.

The AFOs were mounted, with an artificial wooden calf remaining as a fixture point to the AFO. This also acted as attachment for a proximal rotational hinge and a distal ankle point to be positioned 7 cm proximal to the footplate (Figure 1). An assumption was made that forces were taken directly through the centre of the model tibia and that the theoretical ankle lay directly in line with this point. Final connection between the footplate and base of the machine was established incorporating a modified attachment tackle as previously used in the study by Rooney¹⁴ with the design considered from the work of Major et al.⁵ For this attachment, the device footplate was required to be shortened at a point mimicking the metatarsophalangeal joint. Each AFO was then sandwiched between two metal plates as a pivot point at the distal end, secured with two C-Clamps and then two distal pilot holes drilled. In addition, a moulded template was designed so all drilling and attachment locations were positioned consistently for every device.

For modification, the factors of drilling into CF were considered, with the implications of weakening the structure. As any damage to the fibres can impact overall ability by reducing uniformity, leading to increased stresses, ‘Interrupting the fibres and causing matrix shear’,^17,18 this may add to the chance of delamination. The goal was to reduce this to a minimum, by having the holes positioned 15 mm from the distal end and the medial/lateral sides, at a maximum distance from the strut. The sandwich plates were kept long, ending closely in line with the strut and held in place with two C-clamps. This was hoped to further distribute the forces away from the holes, leaving them as locating points only. From observation this did not appear to impact on the results or the structural stability of the AFOs, allowing for focus around the strut during testing. It was noted that the area surrounding the holes was marginally weakened by partial delamination due to ‘pull up’ and ‘push through’ of the drill bit; however, this was expected.

After preliminary testing with the trial setup, loads of 100 N in tension/compression were considered optimum for the first 2 tests, taking into account the limited crosshead movement.

A set protocol was performed to confirm elastic properties for a Creep test; held for 15 min under tensile forces, followed by a 5-min rest period and then 15 min compressive forces applied, with a final 5-min rest period.

To determine stiffness, moment-angle measures were recorded and calculated (Figure 2), with loads applied in steps of 20 N in either tension or compression.

Figure 2.

Moment and angular measures.

The final test performed was to test mechanism of failure, by continual compressive loading at 200 N/min.

All results were analysed within Excel 2010 using the analysis of variance (ANOVA) statistical method.

Results

Prior to testing, thickness measurements were taken at set points along each AFO (Table 1). Only minor differences were seen within each layup group, although no devices were exactly identical. This may relate to variances seen in later results. These minor differences can be contributed to manufacturing by hand.

Table 1.

Thickness measures.

Thickness (cm)	AFO 1 (mm)	AFO 2 (mm)	AFO 3 (mm)	AFO 4 (mm)	AFO 5 (mm)	AFO 6 (mm)	AFO 7 (mm)	AFO 8 (mm)	AFO 9 (mm)
Along footplate: posterior–anterior
5	3.49	3.46	3.4	3.42	3.67	4.23	4.11	4.32	4.21
10	3.85	3.55	3.57	3.68	3.63	4	3.93	4.15	3.96
Along strut: distal–proximal
5	4.66	4.6	4.88	6.06	5.44	5.43	5.65	6.35	6.04
10	5.43	5.48	5.82	6.56	5.95	6.62	7.13	7.25	7.53
15	4.18	4.41	4.57	5.07	4.63	5.38	6.23	6.48	6.44

The Creep test confirmed the elastic properties pertaining to each AFO (Figure 3(a) to (c)). Both the Lite and Standard AFOs exhibited the greatest consistency of deflection across all ranges of the held times. Within the rest periods, no deformation was exhibited on any device.

Figure 3.

(a) Creep-Lite AFOs, (b) Creep-Standard AFOs and (c) Creep-Rigid AFOs.

All devices proved to be purely elastic in property, having returned to their original positions after Creep testing, which was the expected result for CF. A possible cause for deformation could have been if the resin had fatigued rather than the fibres when stressed. This has a higher possibility of happening depending on resin used. An acrylic matrix¹⁸ which some orthotic devices utilise is known to reduce elastic properties. However, epoxy resin was used within the tested devices, which is known to provide a better matrix for wetting and adhesion to fibres and ‘improves long time strength and fatigue resistance’.^17,18 This suggests epoxy resin should help maintain the elastic properties.

The results of the Creep tests show some energy-storing⁴ properties due to no deformation. More research would be required to quantify this ability, by timing how quickly and with what force the devices return to their natural starting position. These results could then be compared to the force required to deflect each device, the returned energy would be expected to be slightly reduced through dissipation. The percentage force required to return the device to its original state when compared to the force of deflection could be deduced as percentage of stored energy.

The second test was to determine stiffness of the strut as moment versus angular deflection around the ‘ankle point’. Results were measured manually. The load increased in steps of 20 N and then held for a minute, during which a height (l) and perpendicular measure (x) were recorded (Figure 2). These were then used to calculate the moments and deflection angles from the equations

M = F \times d

ϕ = \cos^{- 1} \frac{B^{2} + A^{2} - l^{2}}{2 B A}

Moment-angle data for the AFOs showed near consistent (Figure 4(a) and (b)) linearity in all cases, exhibiting similar results to other studies on CF orthoses.^2,7,8 From the slopes of the graph, rotational ankle stiffness was calculated as

k = \frac{M}{ϕ}

Figure 4.

(a) Moment-angle (stiffness): tension and (b) moment-angle (stiffness): compression (AFOs 1–3 only tested @50 N max).

Approximate stiffness showed that each device withstood better in tension than compression. All devices fell close to the same range that other tests have shown stiffness to commonly exhibit. This is between ‘0.20 and 1.56 Nm/°, depending on the material and design’, as outlined by Kobayashi et al.⁹ in their review of measurement for rigidity.

There appeared to be a high level of consistency among each set as deflection occurs. The average stiffness difference between each group was significant for the majority (Tables 2 and 3). The Rigid AFOs were seen to be 92.17% stiffer than the Standard in tension and 23.93% in compression. The Lite AFOs were 59.93% more flexible than the Standard design in tension and 47.17% in compression.

Table 2.

Approximate stiffness measures.

	AFO 1	AFO 2	AFO 3	AFO 4	AFO 5	AFO 6	AFO 7	AFO 8	AFO 9
Stiffness tension (Nm/°)	0.1835	0.2231	0.1921	0.5752	0.4124	0.5066	0.9503	0.9773	0.9441
Average tension (Nm/°)	0.1996			0.4981			0.9572
Stiffness compression (Nm/°)	0.6358	0.6399	0.5613	1.1373	1.1209	1.2184	1.5759	1.2596	1.4731
Average compression (Nm/°)	0.6123			1.1589			1.4362

Table 3.

p values.

	p values	Result
Hypothesis	Stiffness of the Lite and Rigid AFO sets is significantly different when compared to the Standard set
Null hypothesis	There is no significant difference in stiffness for the Lite and Rigid sets when compared to the Standard set
p value	α < 0.05
Stiffness difference between Lite and Standard AFO–Tension	0.026	Significant
Stiffness difference between Lite and Standard AFO–Compression	0.0001	Significant
Stiffness difference between Rigid and Standard AFO–Tension	0.01	Significant
Stiffness difference between Rigid and Standard AFO–Compression	0.1	Insignificant

Results proved stiffness was significantly different in all cases (p < 0.05), other than Rigid versus Standard in compression. It should be noted that the sample size was very small in all cases. Any study with <10 results per set may not truly reflect stiffness.

The final prerequisite was a failure test. After discussion with the manufacturer and information gathered on stiffness being greater in tension, the devices’ failure method was expected in dorsiflexion (excluding failure out with normal activities). Each device was mounted in the same way as previously and then a continual load was applied at a rate of 200 N/min until failure. This was classed as any delamination, fracture or permanent deformation that would render the device unsafe or preventing its function.

Results (Figure 5 and Table 4) demonstrate a high variance of differing failure points, along with differing methods of failure. All of which would be classed as catastrophic, completely rendering the AFOs unsafe and voiding their function. Failure was predominately by fracture, with some devices exhibiting delamination and one completely fragmenting into two parts. Failure resulted at the proximal points in the strut just distal to the calf band area.

Figure 5.

Failure points (dorsiflexion).

Table 4.

Failure loads (N).

	AFO 1	AFO 2	AFO 3	AFO 4	AFO 5	AFO 6	AFO 7	AFO 8	AFO 9
Failure load (N)	178.977	150.068	152.076	360.606	307.601	387.202	605.540	417.242	560.204
Average failure (N)	160.374			351.803			527.662

Due to a limitation within the setup with a wooden block being used rather than a more lifelike limb model, deflection and dorsiflexion of the device did not truly represent that of normal gait during the failure test. Although unreflective of gait, the test proved the weakest point along the strut was usually in the proximal one-third, although the point was not consistent. Two reasons were concluded for this: minor differences through layup by hand, and for this design, the approximate distal third of the strut had a greater layup. This then caused a secondary rotational point much higher in the strut. This was a contributory factor to the imperfect linearity for failure in the devices as load was applied (Figure 5).

Ultimately, failure occurred at the weakest point along the strut. There was however inconsistent and non-uniform failure points, which would be relevant to look at with additional research. Within the parameters of this study, it is suspected that the causes are discrepant in the hand layup method and the aforementioned trim line/design of the model.

Discussion

This study provided results for all the parameters and aims of testing. Initial creep tests provided evidence of elastic properties. It was noticed in each graph that after the applied force was ramped back to 0 N for the rest period, the devices did not always retain the same initial start position and appeared to be deflecting (within the graphs). A contributing factor was the limitations on both the calf-block design and the Instron function. The calf-block had around 3 mm of possible movement to accommodate for the fit of the marginally different calf bands. This could have effectively changed the position by slipping. The Instron was only able to calibrate either the start position (deflection) or the load to zero. For consistency, no stresses were added during recovery, and 0 N load was selected. This left a small error in position contributing to an apparent deflection after the tests. Looking at each device independently, no fatigue/deformation was observed.

Figure 3(a) shows an abnormality during the initial phase of compression in all devices, most apparent in AFO 3, taking several seconds for the devices to stabilise to the applied load. It was assumed the Instron operation was the cause. To apply the load to the test piece, the Instron uses an electromagnet to calculate the correct resistance; as there was more than one rotation point on the strut, it took a longer period to determine the correct force. This was most apparent in ‘Lite’ devices.

The moment-angle graphs appeared to show minor inconsistency in some cases. This could be attributed to the fact only five measures were taken per device, each for tension or compression. Further measures could have provided more linear results and greater consistency within each set. A limiting factor in the results for the Lite AFOs was the test only being possible in 50 N compression due to the devices’ flexibility and the crosshead limits being reached. This consideration only apparent once testing had commenced and so if repeated another universal tester with greater crosshead movement would be used.

The stiffness measures were not as expected since there was not the expected single rotation point about the ankle, rather multiple points along the strut and a full length lower limb model was not used. This caused the ‘ankle’ point to move anterior or posterior with applied load. This would mean calculations may have exhibited minor inaccuracies in measurement of stiffness. Due to the 100 N loads (50 N for Lite AFOs in compression), only being relatively small forces, A-P ankle deflection was minimal, but would have provided unusable data at greater loads.

The study showed failure inconsistencies among each set. This could be because, without the use of a full model limb, the failure was not truly representative of activity during normal gait. Rather the design setup highlighted the weakest point of the strut.

Overall, this study highlighted the parameters it set out to and allowed a comparison against the different layups. It also suggests areas that could be improved, since this was only a pilot study it acts as a basis for future research. However, it provided valuable information about the specific AFOs used, with the study able to be replicated for other device testing.

This study looked only at the sagittal plane forces, considering it to be the most important plane for analysis with the ‘advantage of using stiffness as a parameter of measurement is that it enables the AFO rigidity to be easily compared with gait or other biomechanical analysis data’.⁹

At initial contact and during late stance of gait, the moments generated about the ankle are 10 and 100 Nm, respectively.¹² Although 10 Nm was not quite achieved in the moment-angle tests, due to the near linear progression, it could be assumed that the devices would be able to tolerate these moments. Further work would also be required to test AFOs at 100 Nm. During swing phase, only around 1 Nm is required¹² to be resisted, due to the weight of the foot. It can be said with confidence all these devices could maintain this, with only about 1°–3° actual deflection.

Conclusion

This study determined the properties of each AFO to be repeatable for their functional requirement through mechanical testing. Differences in layup percentage were not proportional to stiffness. There were differences in both tension and compression for stiffness; however, this had no effect on the properties of the devices, just their limits.

These results should help clinicians and manufacturers further understand the capabilities and limitations of CF-AFOs and the effect of differing levels of layup. The test procedure used within this study could be further developed and improved for future testing to have a standard for similar devices. The hope is for a better estimation of the elasticity, stiffness and failure, leading to better layup precision for the requirements of AFOs.

It is recommended that International Standards be established and published widely to allow mechanical standard testing. This would allow direct comparisons of different products, where many claim to perform the same function without comparable evidence.

Further studies should be performed on a larger scale to investigate the accuracy and repeatability of the outlined procedures.

Footnotes

Author contribution

All authors contributed equally in the preparation of this manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

References

Bielby

Warrick

Benson

et al . Trimline severity significantly affects rotational stiffness of ankle-foot orthosis. J Prosthet Orthot 2010, 22(4): 204–210.

Bregman

DJJ

Rozumalski

Koops

et al . A new method for evaluating ankle foot orthosis characteristics: BRUCE. Gait Posture 2009; 30: 144–149.

Bregman

DJJ

De Groot

Van Diggele

et al . Polypropylene ankle foot orthoses to overcome drop-foot gait in central neurological patients: a mechanical and functional evaluation. Prosthet Orthot Int 2010; 34: 293–304.

Kerr

Moyes

Arnold

et al . Permanent deformation of posterior leaf-spring ankle-foot orthoses: a comparison of different materials. J Prosthet Orthot 2011; 23: 144–148.

Major

Hewart

Macdonald

. A new structural concept in moulded fixed ankle foot orthoses and comparison of the bending stiffness of four constructions. Prosthet Orthot Int 2004; 28: 44–48.

Nagaya

. Shoehorn-type ankle-foot orthoses: prediction of flexibility. Arch Phys Med Rehabil 1997; 78: 82–84.

Singerman

. Design changes in ankle-foot orthosis intended to alter stiffness also alter orthosis kinematics. J Prosthet Orthot 1999; 11(3): 48–55.

Takahashi

Stanhope

. Estimates of stiffness for ankle-foot orthoses are sensitive to loading conditions. J Prosthet Orthot 2010; 22(4): 211–219.

Kobayashi

Aaron

Leung

et al . Techniques to measure rigidity of ankle-foot orthosis: a review. J Rehabil Res Dev 2011; 48(5): 565–576.

10.

Kobayashi

Leung

Hutchins

. Design of a manual device to measure ankle joint stiffness and range of motion. Prosthet Orthot Int 2011; 35(4): 478–481.

11.

Novacheck

TFMD

Beattie

CPT

Rozumalski

AMS

et al . Quantifying the spring-like properties of ankle-foot orthoses (AFOs). J Prosthet Orthot 2007; 19: 98–103.

12.

McHugh

. Analysis of body-device interface forces in the sagittal plane for patients wearing ankle-foot orthoses. Prosthet Orthot Int 1999; 23(1): 75–81.

13.

Sumiya

Suzuki

Kasahara

. Stiffness control in posterior-type plastic ankle-foot orthoses: effect of ankle trimline. Part 1: a device for measuring ankle moment. Prosthet Orthot Int 1996; 20: 129–131.

14.

Rooney

. Investigation how thickness, construction and plastic choice affect the stiffness of thermoplastic AFOs. Glasgow: National Centre for Training and Education in Prosthetics and Orthotics, University of Strathclyde, 2012, p. 51.

15.

Polliack

AAP

Swanson

Landsberger

SES

et al . Development of a testing apparatus for structural stiffness evaluation of ankle-foot orthoses. J Prosthet Orthot 2001; 13: 74–82.

16.

BS EN ISO 22675:2006. Prosthetics: testing of ankle-foot devices and foot units – requirements and test methods.

17.

. Considerations on composites in prosthetics and orthotics with emphasis on carbon fibre composite properties. Melbourne, VIC, Australia: International Society for Prosthetics and Orthotics, 1986.

18.

Klasson

. Fiber composites and carbon fiber. 3rd ed. Solna: LIC Orthopaedic, 1989.