Source DocumentPrevious PageTable Of ContentsNext Page

Application of Diffusive Gradients in Thin-films (DGT) to measure potassium and sulphur availability in agricultural soils

Sean Mason1, Ann McNeill1, Yulin Zhang1, Mike J. McLaughlin1, 2 and Chris Guppy3

1 The University of Adelaide, PMB 1, Waite Campus, Glen Osmond, SA 5064, Australia.
CSIRO Sustainable Agriculture Flagship, CSIRO Land and Water, PMB 2, Glen Osmond, SA 5064, Australia
3 University of New England, Armidale, NSW 2351, Australia


Potential for potassium and sulphur deficiency to become prevalent in broad acre agricultural soils in Australia is increasing due to relatively low application rates in fertilisers and substantial removal in harvested products. It is debatable whether currently available soil test methods provide accurate assessment for available K and S. Thus, new technology may offer an alternative approach. Diffusive gradient in thin films technology (DGT) has been successfully applied to assess P availability in many Australian agricultural soils where conventional methods were unreliable in defining a P pool that correlated with crop P uptake and response to fertiliser. This paper reports on the development of DGT for measuring available K and S in Australian soils. Two separate binding agents were tested, in simple solutions varying in pH (3-9), and these agents demonstrated high affinities and large sink capacities for K and S. DGT-K and standard available K soil measurements were poorly correlated (R2 < 0.2) suggesting the DGT method is measuring a different pool of K that could potentially be more closely associated with plant-available K. Moderate correlations were obtained (R2 = 0.5-0.56) between DGT–S and established soil tests for available S. However, when the soil S test methods were compared with short-term maize S uptake and response to S applications under glasshouse conditions, DGT had the greatest accuracy (R2 = 0.55 (relative yield), 0.80 (uptake) compared to resin S (R2 = 0.48, 0.57), KCL-40 (R2 = 0.34, 0.57) and MCP (R2 = 0.4, 0.29). Further glasshouse and field studies are required to assess the full benefits of using DGT as a tool for measuring plant available K and S in Australian broad acre agriculture.

Key Words

Nutrient availability, sulphur, potassium, soil testing


While Australian grain growers have responded to demand for greater production by utilising best practice agronomy and increased N and P fertiliser inputs, they have had limited information on how to manage K and S on their farms. Recently, two agricultural areas with K deficiency have been identified in Australia, with significant responses to K applications obtained on sandy soil types in WA (Wong et al 2000) and on Vertosols in Queensland (Bell pers. comm.). The diversity of these two soil types offers significant challenges to formulating a single soil test method for K. Sulphur is a major macro nutrient for broad acre crops yet inputs of S into cropping systems in Australia have generally reduced due to a decreased use of single super phosphate fertilisers. Responses to S applications have been observed recently, especially in sensitive crops like canola (Stacey pers. comm.). However, due to a lack of soil and plant testing for S, the areas that may be at critical S levels are not easily identified.

Previous studies assessing the performance of exchangeable K as a measure of plant-available K have shown that differences in soil type can make it difficult to relate exchangeable K to plant K uptake/response and produce one critical soil test value (Romheld and Kirkby 2011). Since it has been shown that exchangeable K and Colwell K measurements are highly correlated for a range of soil types (Wong and Harper 1999) there is a high probability that Colwell K is also not a good measure of plant available K across a wide range of soil types. Previous work aimed at identifying an accurate soil test for S has been mainly performed on pasture systems (e.g. Blair et al. 1991) and has shown only moderate predictive capabilities for plant response to S fertiliser. Recent work has shown that established methods (KCl and MCP) were poor at predicting maize S uptake and responses to S applications in a short-term glasshouse experiment (Guppy 2010). Therefore an improved method for determining available S could have significant benefits.

Strong evidence already exists in the literature that DGT measurements closely mimic plant available P, Zn and Cu, all elements that exhibit diffusional behaviour in soils. Potassium behaviour in soils is very similar to P in that the main contribution to K delivery to the plant root is by diffusion, and so it is highly likely that DGT will accurately assess K availability independent of soil type variations. Therefore the ability of DGT to assess plant available K and predict responses to fertilizer K applications in a range of soil types typical of broadacre cereal production regions in Australia is required. Sulphur supply to plants is not normally diffusion limited due to its low sorption on soil surfaces and therefore the applicability of DGT for measuring S availability is unknown. The paper reports initial development of DGT to enable assessment of K and S in soils and preliminary comparisons of K and S DGT measurements with conventional soil testing methods for available K and S.


DGT method development

Two agents (Amberlite IRP-60 for K, Tandy et al. 2012 (Product A) and product B for S) were selected and cast in polyacrylamide gels using previously published procedures for DGT (Zhang et al. 1995). An infinite sink for the element of choice is required to enable the diffusion theory (Fick’s Law) underlying DGT to be employed, and therefore conditions of deployment must not create saturation of the binding layer. Capacities of the binding layers were determined by deploying DGT devices in separate solutions (3L) at 40 mg/L for K and S for different time intervals up to 24 hours. Capacity was determined when linear uptake of both K and S with time of deployment ceased. Another requirement for the development of DGT binding layers is that it must accurately measure soil solutions under pH conditions relevant to agricultural soils. Tandy et al. (2012) revealed that DGT performance for measuring K was unhindered between pH values 3-9. DGT for S was tested under the same pH range (3-9) using simple solutions (3L) at 5 mg/L S with the pH adjusted using either 1M HCl or 1M NaOH.

DGT K and S measurements compared with conventional soil tests

Using a selection of agricultural soils (20) from across southern Australia the assessment of DGT K was compared to standard soil K tests including Colwell K, exchangeable K and soil solution K obtained using rhizon samplers at 100% WHC. Comparison of DGT S with conventional soil S tests which include KCL-40, MCP and resin S was performed using soils (25) obtained from the New England region of NSW outlined in the glasshouse trial below.

Performance of DGT S with maize response and uptake

Soils were obtained from a glasshouse trial previously described by Guppy et al. (2010). Briefly, soil (0-10 cm) was collected from properties within a 100 km radius of Armidale in the New England region of NSW encompassing 4 main soil type and screened to <1cm. The S applications were applied as solutions of ammonium sulphate and applied at a non-limiting rate of 20 mg S/kg and maize (Zea mays) was grown for a period of 40 days. Soils were analysed for resin-extractable S, MCP- and KCl-40 extractable S (Blair et al. 1991). DGT S measurements (duplicate) were performed using the new binding layer for two deployment times (24 and 48 hours) at 100 % WHC. After removal from the soil DGT devices were rinsed with ultra pure water and the binding layer retrieved and eluted in 1ml 1M HCl. Soil test results were correlated with maize response (% relative yield = Ycontrol/Y@ 20 mg S/kg) x 100 and maize S uptake from the control soils only.

Results and Discussion

DGT method development

Performance of the Amberlite gel for K at different pH and its large capacity for K (Table 1) suggests that it will be capable of measuring readily available K across a range of broad acre agricultural soils. The capacity obtained closely matched that reported by Tandy et al. (2012) and the inclusion of ferrihydrite in the gel to enable easier handling did not appear to hinder the ability of the Amberlite to bind K. Using a cation exchange agent in the gel also enabled the measurement of Ca and Mg using DGT. Competitive studies with Ca and K (data not shown) on the gel revealed a higher selectivity of Ca on the Amberlite and therefore amounts of both Ca and Mg on the gel from a soil deployment need to be considered to determine if infinite sink conditions for K were present. Shorter deployment times could be used to overcome any potential competitive effect.

The binding agent for measurement of S with DGT has a substantially larger binding capacity for P (Table 1) than the iron oxide (ferrihydrite) gel used currently for DGT-P. Elution efficiencies for both S and P using 1M HCl are lower, possibly because the iron component in the iron oxide gel dissolves readily in the eluting acid. DGT performance was unaffected by pH in the range of 3-9 which suggests DGT measurement of S will be effective for soils ranging in chemical properties. The large capacity of DGT for S allows for use of longer deployment times which potentially enables application to measure mineralisation of S.

Table 1. Performance characteristics of two new binding gels for DGT assessment of K, Ca, Mg, S and P. FeOx are the parameters for the established DGT method for P using a ferrihydrite binding layer. Nd – not determined.


Product A - Amberlite IPR-60

Product B









Uptake efficiency %

> 98

> 98

> 98

> 98

> 98


Capacity (on gel)

800 μg

> 800 μg

> 800 μg

500 μg

> 250 μg

12 μg

Elution efficiency %







pH range

3 – 9



3 – 9



DGT K and S measurements compared with conventional soil tests

Only moderate to poor correlations were obtained between DGT K and S with conventional soil testing methods for both these elements (Table 2). For K, DGT was most closely matched with soil solution K concentrations which indicates that the DGT method is measuring the intensity pool of K but potentially there is some quantification of the re-supply factor of K with the depletion of the soil solution pool. This has been shown to be an important parameter for P but the importance of K re-supply for K plant uptake is relatively unknown. Poor relationships between DGT K and both Colwell K and exchangeable K suggests these methods are measuring different pools of K, as has been demonstrated for Colwell P (Mason et al. 2010). These results suggest that further work assessing soil K tests and DGT as K availability measures needs to be performed using representative Australian soils and other crop types.

DGT S measurements were moderately correlated with KCL-40 and resin S but no relationship was obtained with MCP.

Table 2. Correlation coefficients (R2) between DGT and conventional soil tests for both K and S. Nd – not determined.

DGT K with

DGT S with

Colwell K

Exchangeable K

Soil Solution K

KCL – 40


Resin S







Performance of DGT S with maize response and uptake

Comparison of DGT results with maize response to an application of S (Figure 1) provided a slightly better correlation compared to resin S and an improvement with KCl-40 and MCP methods (Table 3). Greater predictive power using DGT was also highlighted when soil test results were compared with maize S uptake on the control soils only (Table 4, Figure 1). When DGT values are calculated as Concentration DGT (CDGT) which is the time averaged concentration calculated at the interface of DGT/soil to account for the mass of S measured on the gel, lower values were obtained for the 2 day deployment compared to the 1 day deployment. This highlights the inability of these soil types to sustain solution S concentrations when the solution S pool is depleted. There was no substantial improvement of either deployment time when values were related to maize response and uptake. The improvement of DGT in predicting plant available S in this study reveals that the method has the potential to measure the short term availability of S. Improvement over the stronger extractant methods could be attributed to the relative short term nature of this glasshouse experiment where the mineralisation component of S (which KCl-40 and MCP attempt to measure) to overall maize uptake would have been small. Therefore it is unclear what benefit DGT will have in relation to assessing S availability over the course of a growing season in the field. Potentially for an accurate DGT S test other soil parameters will need to be assessed due to the chemistry of S in the soil and the potential sources of S that can occur with mineralisation with time.

Table 3. Correlation coefficients (R2) for soil S tests with maize S uptake and maize responses to S applications.


DGT S (1 day)

DGT S (2 day)



Resin S

Relative yield (%)












Figure 1. Correlation coefficients (R2) for DGT (1 day & 2 day deployments) with maize S uptake and maize response to an application of S. Correlations with other S tests can be found in Guppy et al. 2010.


DGT has been developed to measure both K and S in Australian agricultural soils. Promising results correlating DGT measures of S availability with uptake by maize suggest further work is warranted to assess the application of DGT for managing S supply in soils for other crops. Contrasting results between DGT K and standard methods indicate that DGT is measuring a different pool of K but it is unclear whether this pool accurately reflects the plant available pool. Further studies are required before the full benefits of DGT for both K and S are realised. A reliable and correlated soil K and S test will have high significance for the grains industry as it will allow areas with K and S deficiency to be identified with greater confidence. Accurate assessment of available K and S is vitally important in order to make informed decisions about required fertiliser applications and to maximise yields.


Blair GJ, Chinoim N, Lefroy RDB, Anderson GC and Crocker GJ (1991). A soil sulphur test for pastures and crops. Australian Journal of Soil Research 29, 619-626.

Guppy C, Blair G and 2009 Soil411 Class (2010). Predictive value of resin extraction to determine sulfur and phosphorus response of maize in a range of soils from the New England Tablelands of NSW, Australia. Proceeding of the 19th World Congress of Soil Science, Brisbane, Australia, 1 – 6 August 2010.

Mason S, McNeill A, McLaughlin MJ and Zhang H (2010). Prediction of wheat response to an application of phosphorus under field conditions using diffusive gradients in thin-films

(DGT) and extraction methods. Plant and Soil 337, 243-258.

Romheld V and Kirkby A (2010). Research on potassium in agriculture: needs and prospects. Plant and Soil 335, 155-180

Tandy S, Mundus S, Zhang H, Lombi E, Frydenvang J, Holm PE and Husted S (2012) A new method for determination of potassium in soils using diffusive gradients in thin films (DGT). Environmental Chemistry 9, 14-23.

Wong MTF and Harper RJ (1999). Use of on-ground gamma-ray spectrometry to measure plant-available potassium and other topsoil attributes. Australian Journal of Soil Research 37, 267-277.

Zhang H and Davison W (1995). Performance characteristics of diffusion gradients in thin films for the in situ measurement of trace metals in aqueous solution. Analytical Chemistry 67, 3391-3400.

Previous PageTop Of PageNext Page