Open Collections

UBC Faculty Research and Publications

Delivering high-resolution landmarks using inkjet micropatterning for spatial monitoring of leaf expansion Wang, Lisheng; Beyer, Simon T; Cronk, Quentin C; Walus, Konrad Jan 25, 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-13007_2011_Article_124.pdf [ 3.13MB ]
JSON: 52383-1.0221421.json
JSON-LD: 52383-1.0221421-ld.json
RDF/XML (Pretty): 52383-1.0221421-rdf.xml
RDF/JSON: 52383-1.0221421-rdf.json
Turtle: 52383-1.0221421-turtle.txt
N-Triples: 52383-1.0221421-rdf-ntriples.txt
Original Record: 52383-1.0221421-source.json
Full Text

Full Text

METHODOLOGY Open AccessDelivering high-resolution landmarks using inkjetmicropatterning for spatial monitoring of leafexpansionLisheng Wang1*, Simon T Beyer1, Quentin CB Cronk2, Konrad Walus1*AbstractBackground: Inkjet micropatterning is a versatile deposition technique with broad applications in numerous fields.However, its application in plant science is largely unexplored. Leaf expansion is one of the most importantparameters in the field of plant science and many methods have been developed to examine differentialexpansion rates of different parts of the leaf lamina. Among them, methods based on the tracking of naturallandmarks through digital imaging require a complicated setup in which the leaf must remain fixed and undertension. Furthermore, the resolution is limited to that of the natural landmarks, which are often difficult to find,particularly in young leaves. To study the fine scale expansion dynamics of the leaf lamina using artificial landmarksit is necessary to place small, noninvasive marks on a leaf surface and then recover the location of those marksafter a period of time.Results: To monitor leaf expansion in two dimensions, at very fine scales, we used a custom designed inkjetmicropatterning system to print a grid composed of c. 0.19 mm2 cells on small developing leaves of ivy (Hederahelix) using 40 μm dots at a spacing of c. 91 μm. The leaves in different growing stages were imaged undermagnification to extract the coordinates of the marks which were then used in subsequent computer-assisted leafexpansion analyses. As an example we obtained quantified global and local expansion information and createdexpansion maps over the entire leaf surface. The results reveal a striking pattern of fine-scale expansion differencesover short periods of time. In these experiments, the base of the leaf is a “cold spot” for expansion, while the leafsinuses are “hot spots” for expansion. We have also measured a strong shading effect on leaf expansion. Wediscuss the features required to build an inkjet printing apparatus optimized for use in plant science, which willfurther maximize the range of tissues that can be printed at these scales.Conclusions: To apply inkjet micropatterning to plant studies, we have successfully delivered landmarks on ivy leafsurfaces and achieved high-resolution, two-dimensional monitoring of leaf expansion at different growing stages.The measurement is capable of reliably identifying the fine scale changes during plant growth. As well asdelivering landmarks, this technology may be used to deliver microscale targeted biological components such asgrowth hormones, and possibly be used to pattern sensors directly on the leaves.BackgroundInkjet printing is a technique involving ejecting tinyliquid droplets in a non-contact manner onto targetobject surfaces to form any desired patterns in highresolution. Historically, it has been used almost exclu-sively for document and image printing, findingconsiderable success in the consumer electronics mar-ket. More recently the potential of this technology hasbeen explored to a greater degree, especially in develop-ing printable electronics. Popular applications includepolymer solar cells, thin film transistors, sensors andorganic light emitting diode arrays [1]. If functionalmaterials and patterns are printed directly on plantbodies, various plant behaviours can be accurately inves-tigated in exquisite detail which raises the possibilityof “sentinel plants” reporting on growth conditions* Correspondence:; konradw@ece.ubc.ca1Department of Electrical and Computer Engineering, The University ofBritish Columbia, 2332 Main Mall, V6T1Z4, Vancouver, BC, CanadaFull list of author information is available at the end of the articleWang et al. Plant Methods 2011, 7:1 METHODS© 2011 Wang et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (, which permits unrestricted use, distribution, and reproduction inany medium, provided the original work is properly cited.(desiccation, gas exchange, photon flux density etc.).The non-contact nature of inkjet printing makes it asuitable tool for depositing materials on plant surfaceswithout causing physical damage. Therefore, we con-sider the application of inkjet printing to plant studies isnot only of great significance but also very feasible.However, this area has scarcely been explored until now.Among various topics in botany, precise understand-ing of the growth of plant organs is important not onlyin the study of morphogenesis but also in the appliedplant sciences, as the expansion of the leaf lamina is animportant factor contributing to plant yield. Change intotal leaf area is easy to measure as the leaf outline iseasily captured. However, differential expansion of dif-ferent parts of the leaf is more difficult to study as nat-ural landmarks within the leaf are difficult to find.The pattern of veins can provide landmarks, particu-larly if the venation is finely reticulate, but at earlystages the veins are poorly formed and the patternchanges rapidly. Nevertheless, for later stages of leafdevelopment the question of whether various parts ofthe lamina expand at the same or at different ratesthroughout development has been answered using theexpansion of small areas delineated by veins [2]. How-ever, this method cannot collect measurements with ahigh spatial resolution.In a more recent study by Walter et al. [3], leaf vena-tion was tracked using advanced digital image sequenceprocessing, resulting in a method with very high tem-poral resolution. However, the spatial resolution of thismethod and the requirement that the leaves remainmechanically fixed to the apparatus during growthmakes it more suitable for large leaves.Finer scale measurements can be achieved with clonalanalysis in which individual cell lineages are followedusing phenotypic markers [4-7] but this depends on theavailability of suitable biological material. It has success-fully been used to study leaf and petal expansion usingvisible changes mediated by, for instance, transposonactivity.Another method involves placing artificial landmarkson the plant organ. Avery [8], in a pioneering study ofleaf expansion, marked a grid on a developing tobaccoleaf using India ink. This technique is only suitable forrelatively large leaves, typically tobacco leaves c. 6 cmlong, and using a fairly coarse initial landmark spacing(c. 5 mm). Numerous subsequent authors have used thesame technique [9-11]. A modern variant of this proce-dure is to use graphite particles scattered randomly [12].This technique can reach much finer scales but requirescareful image analysis to reconstruct the landmarks asthey move apart during expansion.To address some of these problems of artificial land-marks, Hamamoto, Kimura & Yamaguchi [13] usednon-contact inkjet printing to print lines of small dots(0.1-0.2 mm in diameter) on the surfaces of plantorgans. However, the authors only studied the elonga-tion of the organs along one direction. The purpose ofthis paper is to extend the inkjet micropatterningmethod, printing at much finer scales and in twodimensions, using 40 μm dots, in order to study theexpansion of very young leaves. We used ivy (Hederahelix cv) as the experimental material. The complex out-line of its palmate-shaped leaves allowed us to study therole of differential expansion in the morphogenesis ofsuch complex shapes.MethodsPlant material and growthThe plant used in the experiments was English ivy(Hedera helix cv) shown in Figure 1. It was grown in a20 cm-diameter pot positioned with side illuminationfrom a window and kept at a relatively constant tem-perature of c. 25 °C. The plant was able to receive directsunlight from the southeast for approximately five hoursper day throughout the experimental period (summer).It was watered two to three times per week and suppliedwith nutrients using houseplant nutrient spikesFigure 1 Photograph of studied ivy plant. The branches withstudied leaves were supported at the top of the plant andseparated from others.Wang et al. Plant Methods 2011, 7:1 2 of 10(Miracle-Gro®, The Scotts Co.) once before the experi-ment to ensure that leaf growth was unaffected by nutri-ent or water deficits.The ten studied leaves were selected from ten primarybranches of the plant. Each leaf was 17-20 mm long atDay 0 of the experiment, corresponding to a leaf plasto-chron index (LPI) of c. 0.8-0.9 [2] assuming that the LPIof a 10 mm long leaf is set at zero by convention. Theleaves were measured at the 5th, 10th and 15th day, cor-responding to a growth interval of c. 0.3, 0.6 and 0.9plastochrons respectively (where one plastochron is theinterval between the production of successive leaves ona shoot). The branches bearing the leaves under studywere supported at the top of the plant and separatedfrom others to make sure they received ample naturallight distributed as evenly as possible (Figure 1).Pattern printing on leaf surfacesThe inkjet printing was performed using a custom-builtdrop-on-demand micropatterning system. Figure 2shows the schematic diagram of this system. We used a40 μm orifice diameter piezoelectrically-driven nozzle(MJ-AB-63-40; MicroFab Technologies, Inc.). Whenactuated with a voltage pulse, the nozzle was able toeject droplets of ink on the order of picoliters per drop,forming individual dots on the leaves. The voltage levelused for actuation was 29 V with a 92 μs wide sinusoi-dal pulse. A syringe pump (PSD/3 5175-01; HamiltonInc.) combined with an inline pressure sensor (PX139-001D4V; OMEGA, Inc.) was used to control the airpressure in the ink canister in order to maintain consis-tent droplet formation as ink was consumed.To reduce nozzle vibration, the nozzle was kept sta-tionary while the deposited pattern was formed throughthe movement of a two-dimensional stage. Stage move-ment with resolution of 2.5 μm was accomplishedthrough the use of two stepper motors (23D - 6306C;American Precision Industries). Each leaf, along withthe branch, was fastened onto the stage before printingby folding a piece of weakly adhesive tape, applying it tothe back of the leaf and gently pressing to flatten it. Theink used was violet food colouring (Scott Bathgate Ltd.)which provided distinguishable contrast to the greenleaves.All components of the inkjet printer were controlledwith a PC hosting custom C++ based software. Macrobased commands were used to determine the patternsto be printed on the leaves. The pattern used was asquare grid consisting of 56 horizontal lines and 56 ver-tical lines; sufficient to cover the entire surface of eachleaf. Each line was composed of individual dots with acentre-to-centre distance of 87 μm. Line spacing in boththe vertical and horizontal directions was 435 μm.Pattern recordingThe recording of printed patterns was carried out onthe same 2D stage (Figure 2). A USB CMOS camera(EO-1312M; Edmund Optics, Inc.) with a 0.42× zoomlens (HRP042-CMT; Diagnostic Instruments, Inc.) wasmounted beside the nozzle and connected to a compu-ter. Uc480Viewer (Thorlabs, Inc.) was used to capturedigital images. A home-built LED light box was attachedto the top of the stage to provide backlighting duringimaging. Each leaf was placed on the window of thelight box and gently covered with a glass slide to fix andflatten its surface. The stage was stepped in a grid pat-tern to produce a total of 9-16 photographs (dependingon the size of the leaf) of different regions of the leaf.A panoramic image was then created by stitching thosephotographs together using Windows Live Photo Gallery2009 (Microsoft Co.).Pattern extraction and computer-assisted analysisThe method and algorithm to acquire quantitative andvisualized expansion information of one leaf is com-posed of the following steps:1. Manually rotate the leaf image with the initialprinted pattern until the printed lines are approximatelyparallel to the horizontal and vertical axes of the screen.2. Open the rotated image file with ImageJ 1.43j(National Institute of Health of United States; NIH).Manually select the intersecting points of the crossinglines, as well as the intersecting points of the printedlines and the leaf margin, using ImageJ to store theFigure 2 Schematic diagram of the inkjet printing/imagingsystem. The inkjet printing system includes a nozzle, pressuresensor, ink canister, syringe pump, CPU and stage, and the imagingsystem includes a camera, light box, CPU and stage.Wang et al. Plant Methods 2011, 7:1 3 of 10coordinates. If two or more points are very close to eachother, select only one of them. Save the index numbersand coordinates of all selected points in a text file andthose of margin points again in a separate text file.3. The data analysis program, written using MatlabR2008a (The Mathworks, Inc.), is used to import the textfiles. The program takes each point from the text file andlocates its neighbouring points by searching for the near-est one along the ascending x and y axis respectively aswell as any point in its first quadrant to which the dis-tance is no more than 650 μm (a little longer than timesline spacing). Thus, the program will find three neigh-bours for points within the leaf and 0-4 neighbours forpoints along the leaf margin. The program discards thosepoints with no more than one neighbour as well as thosewithout neighbours along both the ascending ¬x and yaxes. It is worth mentioning that the high accuracy of theinkjet printing process makes automatic selection ofneighbouring points a feasible possibility.4. Each point and its neighbouring points specify asmall square or triangular cell. Here, and for theremainder of this paper, a cell is defined as the regionbetween printed lines rather than in the biological sense.The program records the index numbers of those pointsand calculates the area of the cell (A0) using the poly-area function. It also discards those cells which are out-side of the leaf surface (specified by the recordedmargin points) using the inpolygon function. The surfacearea of the entire leaf can be calculated either based onthe polygon specified by the margin points or as thesummation of the areas of all cells. Further study showsthe difference between these two values of all studiedleaves is less than 3%, which is negligible.5. For distorted patterns at Day x after leaf expansion,repeat Steps 1 and 2. Select the points in the exact sameorder as in Step 2. That is, each point is expected tohave the same serial number in the data files before andafter leaf expansion.6. The program calculates the areas of expanded cells(Ax) specified in Step 4 as well as total cell expansion(Ex) and relative expansion rates (GRx-y), defined as:EAAxx=0(1)GRA AA y xx yy xx−=−−( ) (2)7. Lastly the program creates a simple leaf expansionmap using the patch function and a contour map usingthe contourf function.Computer programs and scripts used in this study areavailable upon request.ResultsPattern printing and data acquisitionA total of ten young actively growing ivy leaves wereselected, which were of similar developmental age andsize at the start of the experiments (Table 1). We used acustom inkjet printer for non-contact printing on six ofthe leaves (Leaves 1-6) using commercial food colouringas the printing material.As shown in Figure 3, each leaf is marked with a high-resolution square grid. Each grid line is patterned withdiscontinuous dots of c. 40 μm in diameter. The averagedistance between two adjacent dots is 91 μm and theaverage spacing between two parallel lines is 435 μm(Inset of Figure 3). The entire surface of each leaf isdivided into a grid of 700-1000 small cells depending onthe size of the leaf. Most cells are located on the planarportions of the leaves and are approximately square-shaped with an average area of 0.19 mm2 (region ‘a’ ininset of Figure 3). Some cells are somewhat deformedbecause of the curved surface (Region ‘b’ in inset ofFigure 3) and those near the margin may have otherTable 1 Size information, expansion and relative expansion rate of leaves at different time periodsLeaf No. Initial size Expansion (mm2/mm2) Relative expansion rate (mm2/mm2/day) Experimental detailsLength (mm) Area (mm2) Day 5 Day 10 Day 15 Day 0-5 Day 5-10 Day 10-151 20 184 2.7 3.5 3.7 0.33 0.064 0.012 Patterned.2 18 159 2.6 3.4 3.7 0.32 0.065 0.0133 20 214 1.7 1.8 1.9 0.13 0.019 0.0074 20 195 1.5 1.6 1.6 0.10 0.021 0.0065 17 169 1.3 1.4 1.4 0.05 0.019 0.008 Patterned; partly covered.6 18 202 1.4 1.6 1.7 0.08 0.034 0.0107 19 176 1.9 2.2 2.3 0.17 0.038 0.009 Controls.8 19 180 1.5 1.7 1.7 0.10 0.020 0.0069 20 193 2.3 2.7 2.8 0.25 0.040 0.00710 17 163 2.5 3.3 3.5 0.30 0.067 0.011Wang et al. Plant Methods 2011, 7:1 4 of 10shapes, e.g. triangles, trapezoids or pentagons, etc.(Region ‘c’ in inset of Figure 3). Fortunately, the shape ofthe cells does not influence the quantitative expansionanalysis since we are measuring their relative expansion.To study the influence of the inkjet printing processon leaf viability, four additional leaves were selected asunprinted controls (Leaves 7-10).The studied leaves were monitored three times at fiveday intervals throughout the experiment. In otherexperiments, more frequent measurements over longerintervals could certainly be available if necessary. Thepatterned leaves were non-destructively imaged and thepatterns were extracted and digitized manually for ana-lysis. The intersection points between crossing lines andbetween printed lines and leaf margins were markedand their coordinates recorded. Figure 4 shows 25example points (white circles) and two marker-definedcells (Cell A and Cell B). Those recorded coordinatevalues from each measurement are fundamentals in ourmonitoring system, with which we are able to performvarious analyses using different models of interest. Thefollowing leaf expansion quantification, global expansionmapping and contour plotting are three demonstratingexamples.Quantifying leaf expansionWith the acquired coordinates, we are able to calculatethe area of each marker-defined cell. By recording thedisplacement of each printed point and the expansion ofeach cell, we were able to track the local expansion pat-terns of each leaf.Figure 3 One ivy leaf (Leaf 4) with printed grids on its surface.Each grid line is composed of discontinuous dots. The entiresurface is divided into 700 small cells. Regions a (square), b(deformed square) and c (triangle) represent different shapes of thecells.Figure 4 A small region of Leaf 5 at different growing stagesshowing the change of printed grid. 25 example points (whitecircles) and two marker-defined cells (Cell A and Cell B) werelabelled. Each cell does not experience the same rate of expansion.Cell A is 1.37 times larger at Day 10 than at Day 0 while Cell B is1.62 times larger.Wang et al. Plant Methods 2011, 7:1 5 of 10As an example, Figure 4 shows the morphology of asmall 16 cell region of Leaf 5 over a 10 day span. Theinitial area of this region was 3.08 mm2. At Day 5 andDay 10 the total area increased to 4.05 mm2 (1.3 times)and 4.57 mm2 (1.5 times) respectively. From the imageit is clear that each cell does not experience same rateof expansion. For example, Cell A in Figure 4 is 1.37times larger at Day 10 than at Day 0 while Cell B is 1.62times larger. This difference demonstrates the highlylocal expansion differences on the leaf.The non-uniformity of leaf expansion is also evidentin the deformation of the printed grid. In Figure 4, thetop two horizontal lines are parallel to the bottom hori-zontal line at Day 0, but by Day 10 they are clearly nolonger parallel. We believe this is due to a faster expan-sion rate in the upper right region of the leaf.We are also able to determine the overall expansion ofone leaf by tracking the surface area increase of theentire leaf. The average expansion of the patternedleaves (Leaves 1-4) is 2.1 at Day 5, 2.5 at Day 10 and 2.6at Day 15 while those of controls (Leaves 7-10) are 2.0at Day 5, 2.5 at Day 10 and 2.6 at Day 15. The patternedleaves grew at a similar rate as the controls, suggestingthat the inkjet printing process does not have any signif-icant influence on the expansion of the leaves.The mean relative expansion rates at different timeperiods during the experiment (Table 1) show the tran-sition from fast expansion to slow expansion as the leafnears maturity. As seen in Figure 5 the relationshipbetween the expansion rate and the observation day isapproximately log-linear with an average correlationcoefficient of 0.996 among the ten leaves studied. Thissuggests that the expansion rate of the leaves decreasesexponentially during this phase. More frequent measure-ments could be used to gain a precise quantification ofexpansion rate changes over time.Mapping of leaf expansionBy colouring the marker-defined cells according to theirexpansion values we are able to produce maps showingthe two-dimensional information of leaf expansion. Thecolour map, from blue, to red, indicates an increase inthe rate of expansion.Figure 6 shows the expansion maps for Leaves 1-3 atDay 5 and Day 10. The local expansion of the leaves isclearly expressed through the colour variation of eachcell. Generally speaking, Leaves 1 and 2 have similarexpansion rates throughout the entire experiment, bothof which are larger than Leaf 3. In addition, the expan-sion of Leaves 1 and 2 is generally bilaterally symmetri-cal while Leaf 3 shows a certain degree of asymmetry inits expansion.The non-uniform expansion of the leaves is clearlydemonstrated by the expansion maps. We can see thatthe base of each leaf has the smallest degree of expansion,indicating this area consistently grows the slowest, called“cold spot”. In contrast, the fastest expansion consistentlyoccurs around the sinus areas between the terminal lobeand lateral lobes, called “hot spots”, with an averageexpansion up to 3 times larger than the base area at Day10. It is also clear from the results that the expansionrates along the primary veins, especially the midribs, areslower compared to their surrounding areas.The expansion rate of each tip area is significantlycorrelated with the shape of that tip. A slower expansionrate will lead to a sharp tip, like most tips of the threeleaves. For example, the tips of the terminal lobes showsmaller expansion than their surrounding areas in alltrials. In contrast, a faster expansion rate leads to around-shaped tip.To the best of our knowledge, spatial and temporalpatterns of ivy leaf growth have not been reportedbefore. Existing literature suggests that dicot leaveseither grow with a base to tip growth gradient, or with arelatively uniform spatial growth [3]. In our results thesinus regions appear to have the highest rate of growthand the leaf base and major veins appear to have theslowest, which could be seen as a departure from theFigure 5 Relationship between the observation date andexpansion rate of ten studied leaves. The relationship isapproximately log-linear with an average correlation coefficient of0.996 among the ten leaves studied.Wang et al. Plant Methods 2011, 7:1 6 of 10growth patterns of other studied dicot leaves. However,the gradual bending of the sinus regions of our ivyleaves over the experimental period is indicative of ahigher rate of growth there, and partly validates thefindings described above.To further verify the validity and accuracy of ourmethod, we intentionally covered part of two leaf sur-faces (Leaves 5-6) with aluminium foil which would sup-posedly lead to asymmetric leaf expansion due touneven illumination. Figure 7 shows the mapping resultsof both leaves at Days 5 and 10. It is very clear that theuncovered regions are expanding more quickly (mainlyyellow and red) than the covered regions (blue) at bothDay 5 and 10 leading to the absence of bilateral symme-try in the expansion patterns which commonly exists innormal leaves (Leaves 1-2). Quantitatively, the overallexpansion of the uncovered region is 1.4 and 1.6 timeslarger than the covered region at Day 10 in Leaves 5and 6, respectively. Such a result is consistent with ourhypothesis and in accordance with Sue et al., who char-acterized the expansion patterns of normal and mal-formed Vitis vinifera cv. Ruby Red leaves [14].Contour plotting is another method to describe theexpansion difference over an entire leaf surface. Figure 8(B) shows the colour-filled contour map of Leaf 2 atDay 10. Similar to the corresponding simple map(Figure 8A), it also shows the base to be the slowestexpansion region and sinuses to be the fastest. The rela-tionship between tip shapes and tip expansion rates arealso demonstrated. The slower expansion along theveins (marked by black dots inside the map) becomesmore visible in the contour map.Error analysisAs explained above, the 2D mapping of leaf expansion iscomprised of three steps: pattern printing, opticalrecording and digital extraction. To study the accuracyand resolution of this technique, we performed erroranalysis on a patterned region containing 224 cells(including 190 square cells) for each of the three steps.First, we determined the spatial resolution of the CCDcamera and lens setup by finding the number of pixels permillimetre in an image with known dimensions. We foundthe resolution to be 86 pixels/mm, or 11.6 μm/pixel. Thisis thus the lower resolution limit for this method of expan-sion measurement given the camera used.Because the digital pattern extraction process involvesmanually clicking on the image with a mouse, there isFigure 6 Mapping of Leaves 1-3 in different growing stages.The local expansion of the leaves is expressed through the colorvariation of each cell. The base of each leaf has the smallest degreeof expansion and the sinus areas grow the fastest. Unlike Leaves 1and 2, Leaf 3 shows a certain degree of asymmetry in its expansion.Figure 7 Mapping of Leaves 5-6 in different growing stages.Both leaves were partly covered using alumina foil, including theleft side from the midrib of Leaf 5 and the right basal and laterallobes of Leaf 6. The uncovered regions are expanding more quicklythan the covered regions at both Day 5 and 10.Wang et al. Plant Methods 2011, 7:1 7 of 10some degree of human error. To investigate this errorwe performed the extraction process twice on oneimage, resulting in two sets of 265 dot coordinates.First, the positional deviation between correspondingdots along the X and Y directions was calculated. Statis-tical results (Figure 9A) show that the error values (14.0μm and 12.6 μm along the X and Y directions respec-tively) are larger than the spatial resolution of theimages. Next we compared the areas of 224 cells fromthe two data sets and found a mean deviation of 1.00and a standard error of 0.04 (Figure 9B). These resultsindicate that while the positional error is noticeable, theresulting error in cell area is negligibly small.The accuracy of the physical printing process wasdetermined in two steps. First, 240 pairs of adjacentprinted dots were randomly chosen and their distancesmeasured. The average distance was 91.4 μm with astandard error of 18.3 μm including the pattern extrac-tion error. Figure 9C shows the distribution of the dot-to-dot distances. 91.4 μm is slightly larger than theexpected dot distance of 87 μm, and 18.3 μm is substan-tially larger than the spatial resolution of the image. Wesuspect this error is due to inaccuracy in the movementof the stage, or the migration of the aqueous ink on thehydrophobic surface of the leaf. Although for this pro-ject the absolute positions of single dots are not impor-tant, precise study of surface properties and carefulselection of inks would be necessary if more accurateprinting is needed. The areas of a total of 190 squarecells throughout the investigated region were also mea-sured and their distribution analysed (Figure 9D). Theaverage value is 0.19 mm2 with a standard error of 0.01mm2, which can be neglected.Lastly there is some degree of error associated withimaging the leaves due to the physical action of settingup the camera and loading the sample. To measure thiserror we imaged a printed pattern twice, reloading thesample and reassembling the camera setup each time.We found the deviation in the areas of 224 cells to havea mean value of 0.99 and a standard error of 0.05 (Fig-ure 9E). This error is the combination of both the pat-tern generation and recording processes. Through thisanalysis we have found that the total error is substantiallysmaller than both the overall leaf expansion (Table 1)and the expansion of individual cells (Figures 6 and 7),indicating that error does not have a significant impacton the results of this method.DiscussionInkjet printing can deliver landmarks at fine scales andfast speeds. In the case of this research, by using 40 μmdots, a grid containing one thousand 0.2 mm2 cellscould be precisely printed in 10 min. Finer-scale print-ing can also be achieved using this method. This is aconsiderable improvement over the 25 mm2 gridsmanually printed or stamped on leaves in classicalgrowth studies and can potentially lead to the develop-ment of more sophisticated growth models encompass-ing a greater range of organs. In addition, because thereis no contact between the nozzle and the organ surfacesin our custom built printer the plant organs are notmechanically damaged, as demonstrated by the controlexperiment described before.To further study the influence of printing resolutionon the measurements, we increased the grid cells from0.2 mm2 to 1.8 mm2 in size by evenly selecting one outof three marker dots along both horizontal and verticalprinted lines. The standard errors of deviation of cellareas are 0.01 during the pattern extraction process and0.02 during the pattern recording process, both of whichare substantially smaller than those of previously usedfiner grid (Figures 9B, 9E), indicating that a higher reso-lution would result in an increased measurement error.The mapping and contour plotting results are shown inFigure 8(C, D). Compared to the higher-resolution map-ping results (Figure 8(A, B)), the “hot spots” and “coldspots” are also visible for expansion within one leaf, andFigure 8 Comparison of simple mapping and contour plottingwith different printing resolutions. Comparisons are made onLeaf 2 at Day 10. Black dots outlined the margin and major veins ofthe leaf. Contour plotting provides similar information to simplemapping with the expansion variation between regions morecontinuous. In lower-resolution maps, some important details aremissing.Wang et al. Plant Methods 2011, 7:1 8 of 10the expansion variation from “hot spots” to “cold spots”becomes more continuous, virtually no isolated spots(Figure 8A, B). However, due to the reduced resolutionsome important details are missing. As an example, theslower expansion rate along the veins is not visible. Inaddition, the expansion disparity between “hot spots” and“cold spots” is underestimated. Since high-resolutionprinting provides additional detail while potentiallyincrease measurement error, the choice of an appropriateresolution must be made according to the requirement ofeach study.The main limitations of the system described here areleaf access and surface suitability. Young leaves areoften protected by older leaves and leaf structures suchas stipules. Maneuvering the shoot tip so that youngleaves are accessible to the inkjet printer is a significantchallenge. Our current printing system has a fixed headand moving stage and the machine as a whole cannotbe tilted to access harder to reach parts of a plant.Climbing plants like ivy are therefore ideal as theirshoots can be introduced at any orientation into theprinting chamber. Future innovations should includeredesign of the printer to provide adjustable stage orien-tation to give greater flexibility with regard to printableplant material. Many young leaves are covered withdense hairs and are rolled or folded into one of severalpatterns of space packing or “ptyxis” making these mea-surements challenging. With our present system evenlyspaced landmarking can only be achieved if the leaf isrelatively flat. We flatten leaves with gentle pressure,holding them in place using a weakly adhesive surface.This is often possible with leaves that are folded alongthe midrib (conduplicate). However, leaves that aretightly rolled (supervolute) when young are not treatablein this way and only the exposed surfaces can be land-marked. In addition, leaf flattening is also required dur-ing the pattern recording process, as in many otherexisting techniques. However, leaves of many specieswill experience significant distortion during expansion,leading to undulating surfaces which are difficult to beflattened and imaged.In this study we used an opaque mark consisting offood colouring that provided sufficient contrast whentransilluminated. However, the use of fluorescent mar-kers would potentially be a way of gaining greater con-trast for use with automated image analysis whichwould speed the process of tracking the landmarks.We use non-contact inkjet printing here to place land-marks on leaves. However, the versatility of inkjet tech-nology enables it to be used as a delivery system for awide range of interesting components [15,16]. Examplesrelevant to plant science might include phytohormonesFigure 9 Error analysis of pattern extraction, printing and recording processes.Wang et al. Plant Methods 2011, 7:1 9 of 10delivered in precisely quantifiable amounts (perhapsalong with markers) as quantitative bioassays. Besidesdye, inkjet technology can be used as a particle deliverysystem. Increasingly, sensor technology is being reducedin scale [17,18] and small nanoparticles consisting ofindicator chemicals can in theory be used to monitorphysicochemical conditions on the leaf surface.ConclusionsIn this article, we demonstrate the applicability of inkjetmicropatterning to the study of plant growth andachieved high-resolution and two-dimensional monitor-ing of leaf expansion. Mapping results show that thismethod is able to describe the expansion difference atdifferent regions of a leaf lamina on a very fine scale.The measurement is reliable according to error analysisand control experiments suggest that this technique canreflect the change of conditions during plant growth. Aswell as delivering landmarks, this technology could beused to deliver microscale targeted biological chemicalssuch as growth hormones, and possibly be used to pat-tern sensors directly on the leaves.AcknowledgementsWe acknowledge the financial assistance of a Martha Piper Research Award(UBC). The laboratories of Konrad Walus and Quentin Cronk are funded bythe Natural Sciences and Engineering Research Council of Canada (NSERC)Discovery Grants Program. Gabriel Man, Christine Woollacott and JuliaNowak provided helpful assistance and discussion at an early stage in thedevelopment of this work.Author details1Department of Electrical and Computer Engineering, The University ofBritish Columbia, 2332 Main Mall, V6T1Z4, Vancouver, BC, Canada.2Department of Botany and Faculty of Land and Food Systems, TheUniversity of British Columbia, 6270 University Blvd., V6T1Z4, Vancouver, BC,Canada.Authors’ contributionsLW raised the plant, developed the analysing program, carried out theexpansion analysis and drafted the manuscript. STB built the inkjet printingsystem and carried out printing on the leaves. QCBC participated in thedesign of the study, provided support for the plant science aspects andhelped draft the manuscript. KW conceived the study, participated in itsdesign and coordination and helped draft the manuscript. All authors readand approved the final manuscript.Competing interestsThe authors declare that they have no competing interests.Received: 4 January 2011 Accepted: 25 January 2011Published: 25 January 2011References1. Tekin E, Smith PJ, Schubert US: Inkjet printing as a deposition andpatterning tool for polymers and inorganic particles. Soft Matter 2008,4:703-713.2. Maksymowych R: Analysis of leaf development Cambridge [Eng.]: UniversityPress; 1973.3. Walter A, Silk WK, Schurr U: Environmental Effects on Spatial andTemporal Patterns of Leaf and Root Growth. Annual Review of PlantBiology 2009, 60:279-304.4. Dolan L, Poethig RS: Clonal analysis of leaf development in cotton.American Journal of Botany 1998, 85:315-321.5. Poethig RS, Sussex IM: The cellular-parameters of leaf development intobacco - a clonal analysis. Planta 1985, 165:170-184.6. Poethig RS, Szymkowiak EJ: Clonal analysis of leaf development in maize.Maydica 1995, 40:67-76.7. Rolland-Lagan AG, Bangham JA, Coen E: Growth dynamics underlyingpetal shape and asymmetry. Nature 2003, 422:161-163.8. Avery GS: Structure and development of the tobacco leaf. AmericanJournal of Botany 1933, 20:565-592.9. Poethig RS, Sussex IM: The developmental morphology and growthdynamics of the tobacco leaf. Planta 1985, 165:158-169.10. Erickson RO: Relative elemental rates and anisotropy of growth in area -a computer programme. Journal of Experimental Botany 1966, 17:390-403.11. Granier C, Tardieu F: Spatial and temporal analyses of expansion and cellcycle in sunflower leaves - A common pattern of development for allzones of a leaf and different leaves of a plant. Plant Physiology 1998,116:991-1001.12. Basu P, Pal A, Lynch JP, Brown KM: A novel image-analysis technique forkinematic study of growth and curvature. Plant Physiology 2007,145:305-316.13. Hamamoto H, Kimura M, Yamaguchi I: High resolution analysis of plantorgan growth using non-contact inkjet marking. Environmental andExperimental Botany 2006, 57:62-69.14. Wolf SD, Silk WK, Plant RE: Quantitative Patterns of Leaf Expansion -Comparison of Normal and Malformed Leaf Growth in Vitis-Vinifera CvRuby Red. American Journal of Botany 1986, 73:832-846.15. de Gans BJ, Duineveld PC, Schubert US: Inkjet printing of polymers: Stateof the art and future developments. Advanced Materials 2004, 16:203-213.16. Mabrook MF, Pearson C, Petty MC: An inkjet-printed chemical fuse.Applied Physics Letters 2005, 86:013507.17. Mabrook MF, Pearson C, Petty MC: Inkjet-printed polypyrrole thin films forvapour sensing. Sensors and Actuators B-Chemical 2006, 115:547-551.18. Li B, Santhanam S, Schultz L, Jeffries-El M, Iovu MC, Sauve G, Cooper J,Zhang R, Revelli JC, Kusne AG, et al: Inkjet printed chemical sensor arraybased on polythiophene conductive polymers. Sensors and Actuators B-Chemical 2007, 123:651-660.doi:10.1186/1746-4811-7-1Cite this article as: Wang et al.: Delivering high-resolution landmarksusing inkjet micropatterning for spatial monitoring of leaf expansion.Plant Methods 2011 7:1.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistributionSubmit your manuscript at et al. Plant Methods 2011, 7:1 10 of 10


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items