Plasmodium LDH: Plasmonic Biosensor Applications

• Plasmodium lactate dehydrogenase (pLDH) is a malaria biomarker and metabolic enzyme produced by all plasmodium species, serving as a key target in biosensing applications.
• Plasmonic biosensing platforms, such as divergent‑beam SPR and nanohole metasurfaces, enable label‑free, quantitative detection of pLDH via optical resonance shifts and intensity changes.
• Effective-medium models and Langmuir adsorption approaches link pLDH binding to optical responses, guiding assay calibration, sensitivity, and limit of detection across various formats.

Plasmodium lactate dehydrogenase (pLDH) is a malaria biomarker that has been treated in recent plasmonic biosensing work as a label-free optical target for quantitative detection. It is described as a protein produced during the life cycle of the malaria parasite and as a metabolic enzyme found in all plasmodium species, including the most widespread falciparum. In the current arXiv literature considered here, pLDH is detected through surface-plasmon-mediated perturbations of local refractive index, using angular interrogation in a divergent-beam Kretschmann configuration and spectral or transmission interrogation in nanohole metasurfaces (Kiyumbi et al., 21 Apr 2026, Kiyumbi et al., 6 Aug 2025, Kiyumbi et al., 18 Jun 2025).

1. Biomarker role in malaria sensing

pLDH appears in these studies as a primary malaria biomarker for both single-analyte and multiplexed assays. One work considers a representative dual-biomarker case consisting of pLDH and histidine-rich protein 2 (HRP-2), with spatially separated regions of interest on a single Au film (Kiyumbi et al., 21 Apr 2026). Another reports label-free sensing of plasmodium falciparum LDH (pfLDH) spiked in phosphate-buffered saline, and states that pfLDH concentrations spanned $0$–$10^{3}\,\mathrm{nM}$ in PBS (Kiyumbi et al., 6 Aug 2025).

The experimental nanohole study further states that the dynamic range $(\sim 1$–$10^{3}\,\mathrm{nM})$ and the mid-point $(K_D \approx 71\,\mathrm{nM})$ match many plasmonic immunoassays and cover clinically relevant pLDH levels (Kiyumbi et al., 6 Aug 2025). In the multiplexed SPR study, pLDH is treated as one channel among reference, pLDH, and HRP-2 channels, with resonance tracking performed independently in each region of interest (Kiyumbi et al., 21 Apr 2026). This positions pLDH not only as an analyte for single-channel calibration, but also as a target compatible with multiplexed detector readout.

The nomenclature in the cited work distinguishes the general biomarker term pLDH from pfLDH when the analyte is specifically plasmodium falciparum LDH. A plausible implication is that assay performance claims should be read with attention to the exact analyte designation and biofunctional interface used in each report.

2. Recognition layers and interfacial biochemistry

The sensing literature summarized here treats pLDH detection as an interface-engineering problem in which the capture layer is part of the optical model. In the divergent-beam SPR platform, the functionalized stack for the pLDH region of interest is

$$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$

with two receptor-layer archetypes: aptamer-like, with $d_c = 2\,\mathrm{nm}$ and $n_c = 1.45$, and antibody-like, with $d_c = 15\,\mathrm{nm}$ and $n_c = 1.45$. Table III in that work lists the chosen pLDH recognition pairs as aptamers pL1 and 2008s and antibody 10C4D5, with the last used for LOD analysis (Kiyumbi et al., 21 Apr 2026).

The aluminum metasurface experiment resolves the biochemical stack procedurally. The surface functionalization consists of O$10^{3}\,\mathrm{nM}$0-plasma hydroxylation of the SiO$10^{3}\,\mathrm{nM}$1 passivation layer, silanization with (3-GPS) to yield an epoxide-terminated monolayer, covalent immobilization of Protein A, and site-directed immobilization of anti-pLDH IgG (p-IgG) via Protein A (Kiyumbi et al., 6 Aug 2025).

The gold nanohole effective-model paper compresses a similar chemistry into an optical adlayer. The gold top surface is first coated with a self-assembled monolayer of DSP $10^{3}\,\mathrm{nM}$2 thick), then protein A $10^{3}\,\mathrm{nM}$3, then IgG antibodies $10^{3}\,\mathrm{nM}$4. These biolayers are modeled as a single dielectric adlayer of thickness $10^{3}\,\mathrm{nM}$5 and refractive index $10^{3}\,\mathrm{nM}$6. When pLDH binds, it forms an additional adsorbate layer of thickness $10^{3}\,\mathrm{nM}$7, surrounded by phosphate-buffered saline with $10^{3}\,\mathrm{nM}$8 (Kiyumbi et al., 18 Jun 2025).

Across these reports, the capture chemistry is not treated as an incidental preprocessing step. It is a parameterized dielectric component of the sensing model, and therefore directly affects resonance position, field overlap, and the conversion from concentration to optical shift.

3. Plasmonic transduction architectures

The divergent-beam Kretschmann architecture uses a collimated p-polarized He–Ne (or red diode) laser expanded into a continuous fan of incidence angles by a Powell lens, with a typical fan of approximately $10^{3}\,\mathrm{nM}$9. A $(\sim 1$0 relay composed of two cylindrical lenses images that fan onto the hypotenuse face of an N-SF11 equilateral prism coated with $(\sim 1$1 Au. Surface-plasmon resonance is excited when the in-plane wavevector

$(\sim 1$2

matches the plasmon mode. The reflected fan is imaged, without any mechanical scan, onto a CMOS camera; one camera axis encodes angle $(\sim 1$3, and the orthogonal axis resolves spatial position for the reference, pLDH channel, and HRP-2 channel. In resonance-tracking mode, the SPR dip in each region of interest is fit to determine $(\sim 1$4 (Kiyumbi et al., 21 Apr 2026).

The aluminum metasurface platform uses surface plasmon resonance and extraordinary optical transmission in a perforated metal film. Its biosensor is composed of an aluminum metasurface made from an array of nanoholes, operating in the visible spectral region. The structure comprises a glass substrate with a $(\sim 1$5 TiO$(\sim 1$6 adhesion layer, an aluminum film with $(\sim 1$7–$(\sim 1$8, a $(\sim 1$9 SiO$10^{3}\,\mathrm{nM})$0 conformal coating to arrest oxidation, and a square nanohole lattice with period $10^{3}\,\mathrm{nM})$1 and hole diameter $10^{3}\,\mathrm{nM})$2. The optical setup employs collimated broadband white light $10^{3}\,\mathrm{nM})$3–$10^{3}\,\mathrm{nM})$4, p-polarized selection by a polarizing beamsplitter, two identical $10^{3}\,\mathrm{nM})$5, $10^{3}\,\mathrm{nM})$6 microscope objectives in transmission geometry, and a fiber-coupled CCD spectrometer (CCS200) with integration time $10^{3}\,\mathrm{nM})$7 (Kiyumbi et al., 6 Aug 2025).

The effective microscopic model is anchored to a reference sensor made from a plasmonic metasurface comprised of an array of nanoholes $10^{3}\,\mathrm{nM})$8 in diameter patterned in a $10^{3}\,\mathrm{nM})$9 thick gold film, with period $(K_D \approx 71\,\mathrm{nM})$0 and a $(K_D \approx 71\,\mathrm{nM})$1 indium-tin-oxide adhesion layer on glass $(K_D \approx 71\,\mathrm{nM})$2. Illumination is from below at normal incidence with a TM-polarized plane wave. In the analytical and FEM treatments, two transmission peaks in $(K_D \approx 71\,\mathrm{nM})$3 occur near $(K_D \approx 71\,\mathrm{nM})$4 and $(K_D \approx 71\,\mathrm{nM})$5, and the $(K_D \approx 71\,\mathrm{nM})$6 band is chosen for sensing because it lies most strongly in the low-index PBS side (Kiyumbi et al., 18 Jun 2025).

These architectures differ primarily in the observable used to encode pLDH binding: angular displacement $(K_D \approx 71\,\mathrm{nM})$7 in the prism-coupled geometry, spectral shift $(K_D \approx 71\,\mathrm{nM})$8 in nanohole transmission, and fixed-wavelength intensity change $(K_D \approx 71\,\mathrm{nM})$9 in off-peak interrogation.

4. Effective-medium and adsorption models

A central technical feature of the current pLDH sensing literature is the coupling of biomolecular adsorption to optical response through effective-medium theory and Langmuir equilibrium. In the divergent-beam SPR model, the SP dielectric-side penetration depth is $$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$0. The baseline effective index of the layer before binding is written as

$$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$1

and the change in effective index when the bound-protein adlayer of thickness $$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$2 and effective index $$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$3 is formed is

$$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$4

The bound-protein layer is then treated by Maxwell–Garnett mixing,

$$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$5

with volume fraction $$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$6 linked to bulk biomarker concentration $$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$7 through the Langmuir isotherm

$$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$8

The resonance-angle shift is then expressed as

$$\text{prism} \;|\; \text{Au }(45\,\mathrm{nm}) \;|\; \text{baseline biointerface }(d_c) \;|\; \text{analyte-adlayer }(d_a) \;|\; \text{buffer }(n_b = 1.335),$$9

A channel-referenced observable subtracts the reference region of interest,

$d_c = 2\,\mathrm{nm}$0

(Kiyumbi et al., 21 Apr 2026)

The effective microscopic model for the gold nanohole sensor uses a closely related formalism. Above the antibody layer, the bound-protein plus buffer mixture is treated as a two-component composite. For volume fraction $d_c = 2\,\mathrm{nm}$1 of protein in the layer, the Maxwell–Garnett relation is

$d_c = 2\,\mathrm{nm}$2

The equilibrium binding law is

$d_c = 2\,\mathrm{nm}$3

with $d_c = 2\,\mathrm{nm}$4 for the pLDH–IgG interaction in that model. The field-weighted refractive index seen by the surface plasmon is then

$d_c = 2\,\mathrm{nm}$5

where $d_c = 2\,\mathrm{nm}$6 is the SP field penetration into the dielectric (Kiyumbi et al., 18 Jun 2025).

The aluminum metasurface experiment uses a directly fitted Langmuir adsorption model for the calibration curve,

$d_c = 2\,\mathrm{nm}$7

and reports $d_c = 2\,\mathrm{nm}$8, implying $d_c = 2\,\mathrm{nm}$9 (Kiyumbi et al., 6 Aug 2025).

The reported $n_c = 1.45$0 values differ substantially across these studies. This suggests that the binding parameter is assay- and interface-dependent within these modeling and experimental frameworks, rather than a universal constant for all pLDH capture configurations.

5. Sensitivity, calibration, and limit of detection

The divergent-beam Kretschmann study benchmarks its N-SF11/Au $n_c = 1.45$1 baseline against published water/glycerol data. For water $n_c = 1.45$2 and $n_c = 1.45$3 glycerol $n_c = 1.45$4, the simulated resonance positions are $n_c = 1.45$5 and $n_c = 1.45$6, compared with experimental $n_c = 1.45$7 and $n_c = 1.45$8. A linear fit of $n_c = 1.45$9 versus bulk refractive index in the malaria-relevant range $d_c = 15\,\mathrm{nm}$0–$d_c = 15\,\mathrm{nm}$1 yields a bulk angular sensitivity

$d_c = 15\,\mathrm{nm}$2

For the pLDH/10C4D5 antibody channel, assuming an angular noise floor equal to the pixel calibration $d_c = 15\,\mathrm{nm}$3 and using a $d_c = 15\,\mathrm{nm}$4 rule, the predicted limit of detection is

$d_c = 15\,\mathrm{nm}$5

which, using pLDH mass $d_c = 15\,\mathrm{nm}$6, corresponds to $d_c = 15\,\mathrm{nm}$7 (Kiyumbi et al., 21 Apr 2026).

The aluminum nanohole experiment reports a measured EOT $d_c = 15\,\mathrm{nm}$8 peak shift corresponding to a spectral sensitivity of approximately $d_c = 15\,\mathrm{nm}$9, in excellent agreement with FEM simulations $n_c = 1.45$0. Its intensity sensitivity at $n_c = 1.45$1 is approximately $n_c = 1.45$2. For pLDH sensing, the spectral LOD is approximately $n_c = 1.45$3, intensity sensing at $n_c = 1.45$4 gives approximately $n_c = 1.45$5, and choosing the optimal off-peak probe near $n_c = 1.45$6 lowers the LOD to approximately $n_c = 1.45$7, equivalent to $n_c = 1.45$8 of pfLDH (Kiyumbi et al., 6 Aug 2025).

The effective microscopic model for the gold nanohole sensor reports a local spectral sensitivity

$n_c = 1.45$9

a fixed-$10^{3}\,\mathrm{nM}$00 intensity sensitivity

$10^{3}\,\mathrm{nM}$01

and a low-concentration sensitivity

$10^{3}\,\mathrm{nM}$02

The idealized calculation gives $10^{3}\,\mathrm{nM}$03, while a more conservative use of the directly fitted calibration gives an LOD $10^{3}\,\mathrm{nM}$04, or $10^{3}\,\mathrm{nM}$05 for pLDH MW $10^{3}\,\mathrm{nM}$06 (Kiyumbi et al., 18 Jun 2025).

Platform	Reported pLDH performance	arXiv id
Divergent-beam Kretschmann SPR	$10^{3}\,\mathrm{nM}$07; $10^{3}\,\mathrm{nM}$08	(Kiyumbi et al., 21 Apr 2026)
Aluminum nanohole SPR/EOT metasurface	$10^{3}\,\mathrm{nM}$09; $10^{3}\,\mathrm{nM}$10 of pfLDH	(Kiyumbi et al., 6 Aug 2025)
Gold nanohole effective microscopic model	$10^{3}\,\mathrm{nM}$11; $10^{3}\,\mathrm{nM}$12	(Kiyumbi et al., 18 Jun 2025)

These values are not directly interchangeable. One report gives a model-based detection limit, one gives an experimental limit of detection, and one gives a theoretical limit of detection. This suggests that absolute numerical comparison requires attention to interrogation mode, noise model, and whether the estimate is experimental, model-based, or theoretical.

6. Multiplexing, detector resolvability, and diagnostic context

The divergent-beam SPR architecture explicitly addresses multiplexing. In the antibody-functionalized pLDH channel, the resonance lies at $10^{3}\,\mathrm{nM}$13 in buffer and shifts to $10^{3}\,\mathrm{nM}$14 with a full analyte layer of $10^{3}\,\mathrm{nM}$15. All sensing states—reference, pLDH, and HRP-2—remain within the camera working window of approximately $10^{3}\,\mathrm{nM}$16–$10^{3}\,\mathrm{nM}$17, corresponding to approximately $10^{3}\,\mathrm{nM}$18 pixels, or approximately $10^{3}\,\mathrm{nM}$19 on chip. The smallest nearest-neighbour shift between channels after binding, approximately $10^{3}\,\mathrm{nM}$20, still spans approximately $10^{3}\,\mathrm{nM}$21 pixels, or approximately $10^{3}\,\mathrm{nM}$22, so the multiplexed regions of interest are fully resolvable on the detector (Kiyumbi et al., 21 Apr 2026).

The aluminum metasurface experiment is not multiplexed in the same way, but it does describe practical assay operation: a silicone gasket wet chamber $10^{3}\,\mathrm{nM}$23, static incubation, equilibration within $10^{3}\,\mathrm{nM}$24–$10^{3}\,\mathrm{nM}$25 minutes upon injection of pfLDH, rinsing with PBS after each injection, and occasional regeneration with $10^{3}\,\mathrm{nM}$26 NaCl to remove bound antigen (Kiyumbi et al., 6 Aug 2025). The same study notes that commercial mRDTs for pLDH typically achieve $10^{3}\,\mathrm{nM}$27–$10^{3}\,\mathrm{nM}$28 LOD $10^{3}\,\mathrm{nM}$29–$10^{3}\,\mathrm{nM}$30 but require labeled antibodies and have limited quantitation (Kiyumbi et al., 6 Aug 2025).

Suggested enhancements in the aluminum metasurface report include intensity interrogation with narrow-line lasers at the optimal off-peak wavelength $10^{3}\,\mathrm{nM}$31–$10^{3}\,\mathrm{nM}$32 for sub-nM LOD, integration with microfluidic flow-over designs to accelerate binding kinetics and lower assay time, multiplexed arrays to detect alternative biomarkers such as aldolase or discriminate species, high-resolution spectrometers or phase-sensitive readout to resolve sub-$10^{3}\,\mathrm{nM}$33 shifts, exploration of quantum-light interrogation to reduce shot noise below classical limits, and testing in whole-blood lysate or finger-prick samples (Kiyumbi et al., 6 Aug 2025). The effective microscopic model paper adds that the framework is quite general and could be applied to other types of plasmonic biosensors and biomarkers of different diseases (Kiyumbi et al., 18 Jun 2025).

Taken together, these studies place pLDH at the intersection of interface biophysics, effective-medium optics, and detector-limited transduction. The resulting picture is not of a single canonical pLDH assay, but of a family of plasmonic sensing formulations in which the measured pLDH signal depends on the chosen recognition layer, the optical observable, and the degree to which the binding process is embedded in the forward model.