Refrigerated EV Delivery
Executive Summary
Door openings dump 15-20 kWh of thermal load into cargo spaces over a typical route—more than your entire energy budget. We found three complementary paths to solve this. The quickest win is air curtain + strip door systems ($500-1,500/vehicle) that reduce infiltration losses 70-85%, proven in supermarket refrigeration for 40 years. Combined with vacuum insulation panels (5x conventional performance) and depot-charged phase change materials, total consumption drops to 5-8 kWh for 8-hour routes.
Do you want quick wins with proven technology (air curtains, PCM) or are you ready to invest in depot infrastructure for the paradigm shift (ice slurry)? Start with air curtains regardless—they pay back in weeks.
Solvable With Effort
Multiple proven components exist but haven't been integrated for EV cold chain. Requires validation at fleet scale before confident rollout.
Deploy air curtain + strip door systems fleet-wide immediately ($500-1,500/vehicle) for 30-50% thermal load reduction within weeks. Simultaneously pilot VIP-enhanced cargo bodies with optimized PCM on 5-10 vehicles to validate the integrated system achieving 5-8 kWh consumption. This phased approach de-risks the larger investment while capturing immediate savings.
The Brief
Refrigerated delivery vans for groceries and pharmaceuticals run diesel-powered compressor units that consume 30-40% of total vehicle fuel. Electric vans eliminate the engine but refrigeration still drains 40-60% of battery capacity, cutting delivery range in half. Current solutions are eutectic plates (limited hold time), oversized batteries (cost/weight), or trailer-mounted diesel gensets (defeats the purpose). Need refrigeration approach that maintains -20°C to +5°C for 8-hour delivery routes while consuming <15% of a 60kWh EV battery. Must handle 50+ door openings per route. Retrofit path preferred—fleet operators won't buy new vehicles just for this.
Problem Analysis
Drivers open rear doors 50+ times per route, each time dumping cold air onto the pavement while warm ambient air floods the cargo space. A single 45-second door opening in 35°C weather can inject 1-2 MJ of thermal load—equivalent to running the compressor at full power for 15-20 minutes just to recover. Meanwhile, the refrigeration unit cycles continuously at 2-4 kW, fighting both this infiltration and the steady heat seeping through mediocre foam insulation. The result: 25-35 kWh consumed over an 8-hour route, leaving EVs with half their expected range.
The fundamental challenge is thermodynamic: maintaining a 55°C temperature differential (from -20°C cargo to 35°C ambient) requires continuous energy input to fight entropy. Vapor compression at these temperatures achieves COP of only 1.5-2.0, meaning every watt of cooling requires 0.5-0.7 watts of electrical input. But the real killer is transient loads—door openings inject massive thermal pulses that overwhelm steady-state calculations. A system designed for 500W steady-state heat ingress must handle 5-10 kW peaks during door recovery, forcing oversized equipment that runs inefficiently most of the time.
Q_total = U×A×ΔT + ṁ×cp×ΔT (infiltration) + Q_product
Total thermal load combines wall conduction (U×A×ΔT ~500-1200W), air infiltration during door openings (dominant at 30-50% of total), and product heat load. Reducing any term helps, but infiltration offers the largest opportunity.
The vehicle should distribute cold, not generate it
On-vehicle electricity costs $0.30-0.50/kWh equivalent (battery capacity × range value). Depot electricity costs $0.05-0.10/kWh off-peak. This 6-10x cost differential means every kWh of cooling generated at the depot instead of on-vehicle creates massive economic advantage. The vehicle should be a thermal distribution system carrying pre-manufactured cold, not a mobile refrigeration plant.
Oversized vapor compression units (2-4 kW continuous)
Sized for worst-case recovery, not steady-state. Runs inefficiently at partial load 80% of the time.
Standard polyurethane foam insulation (50-100mm, R-6/inch)
Allows 800-1,200W heat ingress through walls alone. 'Good enough' when diesel was cheap; devastating for EV range.
Traditional eutectic plates (NaCl-water solutions)
Limited to 150-180 kJ/kg latent heat. Insufficient capacity for 8 hours with 50+ door openings.
Bare rear doors with no infiltration mitigation
Door openings account for 30-50% of total thermal load. Supermarkets solved this 40 years ago.
Lidl UK Fleet (500+ trailers)[1]
Eutectic plates + minimal backup compressor, depot charging
45% energy reduction vs continuous vapor compression (~18 kWh/route)
va-Q-tec (pharmaceutical containers)[2]
VIP + PCM passive containers
120+ hour temperature maintenance, zero power
Carrier Transicold Lynx Fleet[3]
IoT thermal monitoring with conventional refrigeration
28-35 kWh per 8-hour route (frozen/chilled combo)
Arktek (PATH/Global Good)[4]
VIP + ice passive vaccine storage
35 days at 2-8°C with zero power input
[1] Lidl UK 2021 Sustainability Report
[2] va-Q-tec product documentation
[3] Carrier white paper, 2022
[4] WHO/UNICEF product information sheet
[1] Lidl UK 2021 Sustainability Report
[2] va-Q-tec product documentation
[3] Carrier white paper, 2022
[4] WHO/UNICEF product information sheet
Constraints
- Temperature compliance: -20°C to +5°C throughout route (regulatory requirement)
- Energy budget: ≤9 kWh over 8 hours (15% of 60kWh battery)
- Retrofit capability: must work with existing van bodies
- 50+ door openings per route (operational reality)
- Multi-zone operation (assumed 30% frozen, 70% chilled—actual ratio varies by route)
- 8-hour continuous operation (some routes may have depot returns)
- Payload capacity preservation (some cargo volume loss may be acceptable for major efficiency gains)
- No mid-route charging opportunity available
- Rear door access (side access with compartmentalization would significantly ease the problem)
- Urban delivery with frequent stops (highway transport has different thermal profile)
- Roof-mount modifications acceptable for retrofit
Total refrigeration energy consumption
Unit: kWh
Temperature compliance
Unit: % time in compliance
Retrofit installation time
Unit: days
Payback period
Unit: years
First Principles Innovation
Instead of asking 'how do we make on-vehicle refrigeration more efficient,' we asked 'how do we eliminate the need for on-vehicle refrigeration entirely.'
Solutions
We identified 6 solutions across three readiness levels.
Start with the Engineering Path. Run R&D in parallel if you need breakthrough potential or competitive differentiation.
Engineering Path
Proven technologies, often borrowed from other industries. The work is adaptation, integration, and validation, not discovery.
Integrated Thermal Management System: VIP + High-Density PCM + Infiltration Control + Right-Sized Active
A defense-in-depth thermal architecture combining VIP insulation, high-density PCM thermal storage, air curtain + strip door infiltration control, and right-sized variable-speed compressor to reduce total refrigeration energy consumption to 5-8 kWh over an 8-hour route.
A defense-in-depth thermal architecture combining four proven technologies into a coherent system. First, vacuum insulated panels (VIP) replace or supplement conventional foam insulation, achieving R-30 per inch versus R-6 for polyurethane—a 5x improvement that reduces wall heat ingress from 800-1,200W to 160-240W. Second, high-density salt hydrate phase change materials (CaCl2·6H2O at 192 kJ/kg for frozen, Na2SO4·10H2O at 254 kJ/kg for chilled buffering) provide 35-45 kWh of thermal storage capacity, handling all transient loads from door openings without active cooling intervention. Third, a retrofit air curtain + strip door system adapted from supermarket practice reduces door opening infiltration by 70-85%. The air curtain creates a laminar airflow barrier at 3-5 m/s across the door opening, while overlapping PVC strips provide a physical barrier that parts around packages. Fourth, a small variable-speed compressor (300-500W) sized for steady-state heat ingress only—not recovery—runs at optimal efficiency when needed rather than cycling inefficiently. The system is orchestrated by IoT thermal monitoring that predicts temperature trajectory based on route data, ambient conditions, and door opening history. The compressor activates only when the thermal model predicts temperature will approach compliance limits, enabling 'thermal coasting' through most of the route.
Each subsystem addresses a different thermal load component at its source. VIP reduces conduction through walls (Q = U×A×ΔT) by reducing U by 5-6x. PCM provides thermal mass to absorb transient loads (door openings) without temperature excursion, exploiting the 334 kJ/kg latent heat of phase change. Air curtains reduce infiltration mass flow (ṁ in Q = ṁ×cp×ΔT) by 70-85%. The small compressor then only needs to handle the residual steady-state load of ~200-400W rather than the 2-4 kW peaks that conventional systems must accommodate. Running a small compressor at optimal load is far more efficient than running an oversized compressor at partial load.
Multiplicative benefits from addressing each thermal load component optimally, rather than oversizing a single system to handle worst-case
Passive House building standard. Buildings achieve 90% energy reduction through integrated design: extreme insulation, air-tightness, heat recovery ventilation, and minimal active systems
Same physics apply—reduce thermal load through passive means, then right-size active systems for residual load
Vendor silos. VIP suppliers sell insulation. PCM suppliers sell thermal storage. Air curtain suppliers sell infiltration control. Refrigeration suppliers sell compressors. No one owns the integrated system optimization problem.
75-85% reduction in refrigeration energy consumption (5-8 kWh vs 25-35 kWh conventional)
12-18 months for integrated prototype, 24-36 months for commercial deployment
$12,000-20,000 per vehicle
- VIP panel puncture during cargo handling causes localized thermal bridging
- Integration complexity increases installation time beyond retrofit window
- PCM cycling degradation over 1000+ freeze-thaw cycles
- Driver bypass of air curtain if perceived as slowing deliveries
Validate VIP + PCM thermal performance in static chamber test
Method: Environmental chamber per ASTM C1363 (hot box method) for wall assembly; PCM capacity per ASTM E793 (DSC method)
Success: Wall assembly U-value <0.04 W/m²·K (equivalent to R-25); PCM latent heat >180 kJ/kg with <10% degradation over 100 cycles
U-value >0.06 W/m²·K or PCM degradation >20% → Investigate alternative VIP suppliers or PCM formulations before vehicle pilot
Retrofit Air Curtain + Strip Door Hybrid System
Adapt supermarket infiltration control technology to delivery vans for 75-90% reduction in door opening thermal losses
Adapt proven supermarket infiltration control technology to delivery van rear doors. A low-velocity air curtain (battery-powered, 50-100W) activates on door opening, creating a laminar airflow barrier at 3-5 m/s. Heavy-duty PVC strip curtains (2mm thick, 200mm wide, overlapping) provide passive physical barrier when the air curtain is off and during package retrieval. Combined effectiveness: 75-90% infiltration reduction based on supermarket data from Foster et al. (2006). This is the highest-impact, lowest-cost intervention available. It addresses the dominant thermal load (door openings at 30-50% of total) with proven technology requiring only mechanical installation. Can be deployed fleet-wide in weeks while longer-term solutions are developed.
Air curtain creates momentum barrier (jet velocity × air density) that opposes buoyancy-driven infiltration (stack effect). At 3-5 m/s jet velocity, the momentum flux exceeds buoyancy forces for typical door dimensions and temperature differentials. Strip curtains add physical barrier that requires displacement to pass through, further reducing air exchange. Foster et al. documented 70-85% infiltration reduction with optimized parameters.
Deploy immediately to all vehicles regardless of other solution paths. This is not an alternative to the primary concept—it's a prerequisite. The 30-50% thermal load reduction from infiltration control makes all other solutions more effective.
Optimized Eutectic Plate System with Salt Hydrate PCM
Replace 1970s NaCl eutectic plates with modern salt hydrate PCMs providing 30-50% higher thermal capacity
Replace conventional NaCl eutectic plates with higher-density salt hydrate PCM (calcium chloride hexahydrate at 192 kJ/kg for frozen zone, sodium sulfate decahydrate at 254 kJ/kg for chilled buffering). Increase plate volume by 30% using thinner but more numerous plates distributed throughout cargo space for improved heat transfer. Depot charging using existing electrical infrastructure overnight. This represents optimization of existing technology rather than paradigm shift. It's the fallback if VIP integration proves too complex or expensive for retrofit, providing 35-45 kWh thermal storage capacity that extends passive operation duration.
Latent heat of phase transition provides isothermal temperature maintenance—the PCM absorbs heat while melting without temperature rise. Higher latent heat per kg means more thermal capacity in same mass/volume. Distributed placement improves heat transfer coefficient by increasing surface area and reducing thermal resistance between air and PCM.
Use as primary thermal storage if VIP retrofit proves impractical due to cargo body constraints or cost. Also appropriate for fleet operators seeking lower upfront investment with proven technology, accepting that active cooling backup will be needed more frequently than with full integrated system.
R&D Path
Fundamentally different approaches that could provide competitive advantage if successful. Pursue as parallel bets alongside solution concepts.
Depot-Charged Ice Slurry Thermal Distribution
Fundamentally reframe the vehicle as a thermal distribution system, not a refrigeration plant. Load pre-made ice slurry (30% ice fraction in water/glycol mixture) at the depot into an insulated tank. Circulate slurry through heat exchangers distributed in the cargo space using a small pump (~100W). Zero on-board refrigeration—all cold is manufactured at the depot using cheap grid electricity. The depot ice slurry generator operates overnight using off-peak electricity at $0.05-0.10/kWh, compared to on-vehicle electrical equivalent of $0.30-0.50/kWh (accounting for battery capacity value and range impact). This 6-10x cost differential fundamentally changes the economics of cold chain logistics. The vehicle carries 200-300 kg of ice slurry providing 60-100 kWh of thermal capacity—far exceeding the 8-hour requirement even with 50+ door openings. The only on-vehicle electrical consumption is the circulation pump and fans, totaling 0.8-1.5 kWh over an 8-hour route. This preserves virtually all EV range for driving.
Ice slurry exploits the highest latent heat of any practical thermal storage medium: 334 kJ/kg for the water-ice phase transition. Unlike solid ice, slurry is pumpable, enabling distributed cooling throughout the cargo space via circulation. The depot generates ice using efficient industrial equipment (COP 3-4) with cheap off-peak electricity, rather than inefficient mobile equipment (COP 1.5-2) with expensive on-vehicle electricity. The vehicle only needs to circulate the pre-made cold, not generate it.
This concept represents the most fundamental reframe of the problem—questioning whether on-vehicle refrigeration is necessary at all. The 6-10x cost differential between depot and mobile electricity creates compelling economics for any fleet with fixed depot infrastructure. If successful, this architecture could become the industry standard for EV cold chain, similar to how containerization transformed shipping logistics.
Cold can be manufactured centrally and distributed as a commodity, just like the pre-mechanical refrigeration ice harvesting industry that moved 25 million tons annually
If it works: Vehicle electrical consumption drops to <2 kWh for refrigeration—essentially solving the EV cold chain problem completely. Full battery capacity available for driving range.
Improvement: 85-95% reduction in vehicle electrical consumption vs conventional systems (0.8-1.5 kWh vs 25-35 kWh)
Separated Frozen Vault + Passive Chilled Architecture
Challenge the assumption that frozen and chilled must share a compartment. Create a small, heavily optimized frozen vault (50-100 liters, VIP walls at R-150 effective, dedicated PCM at -29°C) for the ~20% of cargo requiring -20°C. The main cargo space operates as passive chilled using standard insulation, PCM buffering at +4°C, and minimal intervention. Right-size each zone for actual thermal requirements rather than designing the entire space for worst-case frozen. The frozen vault achieves heat ingress of <20W due to small surface area and extreme insulation. With 20 kg of eutectic plates at -29°C providing 1.1 kWh thermal capacity, the vault can maintain temperature for 15+ hours passively. The chilled zone at +4°C has much lower thermal gradient to ambient (31°C vs 55°C for frozen), making passive operation far more practical.
Ceiling: 1-3 kWh total consumption (backup compressor only for hot days)
Key uncertainty: Frozen vault capacity may be insufficient for frozen-heavy routes (>30% frozen cargo). Requires fleet mix planning with some conventional multi-temp vehicles for outlier routes.
Elevate when: Elevate to primary innovation if cargo composition analysis shows strong skew toward chilled (>80% of routes with <25% frozen). This architecture becomes optimal when frozen is truly minority use case.
Thermal Budget Management with Predictive Coasting
Adopt LNG shipping's boil-off management philosophy: accept gradual temperature drift and intervene minimally. Use IoT sensors and route data to predict thermal trajectory. Allow temperature to drift within compliance band (-25°C to -18°C for frozen), activating cooling only when approaching limits. Start routes with cargo pre-cooled to -25°C, providing 7°C thermal buffer for passive coasting. The system uses a thermal model incorporating ambient conditions, cargo mass, insulation performance, and door opening history to predict temperature trajectory 2 hours ahead. Compressor activates only when the model predicts temperature will exceed -19°C within 30 minutes, enabling 'just-in-time' cooling rather than continuous operation.
Ceiling: 30-50% reduction in active cooling energy through smarter control
Key uncertainty: Prediction errors could cause compliance violations. Requires robust fallback to conventional control and extensive validation across route types.
Elevate when: Elevate to primary if fleet already has IoT infrastructure and thermal monitoring. This is a software upgrade that can be deployed rapidly to existing systems, providing immediate savings while hardware solutions are developed.
Frontier Watch
Technologies worth monitoring.
Magnetocaloric Refrigeration
EMERGING_SCIENCE5
Magnetocaloric materials heat up when magnetized and cool when demagnetized—COP of 5-10 with no compressor or refrigerant gases
If commercialized for transport scale, magnetocaloric could provide 50-75% efficiency improvement over vapor compression with zero refrigerant emissions. The solid-state nature (no compressor, no refrigerant) offers reliability advantages.
Technology readiness level 4-5. Laboratory prototypes exist but commercial systems are limited to wine coolers and small appliances. Cost per watt of cooling is 10-20x vapor compression. Rare earth material supply chain concerns for gadolinium-based systems.
Trigger: Commercial magnetocaloric system >500W cooling capacity at <$50/W cost announced; or major automotive OEM announces magnetocaloric development program for vehicle applications
Earliest viability: 5-7 years for transport applications
Monitor: Cooltech Applications (France), Astronautics Corporation (NASA contractor), Prof. Karl Sandeman at Imperial College London, BASF (La-Fe-Si materials)
Radiative Sky Cooling Roof Integration
EMERGING_SCIENCE6
Metamaterial films that emit infrared radiation to the cold sky, providing 250-800W passive cooling from van roof area
Zero electrical consumption. Provides supplemental cooling that reduces active system load by 10-20%. Technology is commercializing for buildings (SkyCool Systems) and could transfer to vehicles.
Performance varies significantly with humidity and cloud cover—800W in Arizona, 200W in Florida. Moving vehicle complicates thermal coupling to cargo space. Film durability under vehicle conditions (vibration, UV, dirt) unproven. Economics are marginal except in dry climates.
Trigger: SkyCool or competitor announces vehicle-specific product; or cost drops below $25/m² enabling broader economic viability
Earliest viability: 2-3 years
Monitor: SkyCool Systems (Stanford spinout), Prof. Shanhui Fan at Stanford, 3M (building applications)
Liquid Nitrogen Cryogenic Reservoir Revival
PARADIGM7
Revival of 1960s LN2 spray cooling—50-100 kg LN2 provides 40-60 kWh cooling with zero electrical consumption
Zero vehicle electrical consumption for frozen zone. LN2 infrastructure already exists at many food processing depots. Proven technology requiring integration, not invention. Could be combined with passive chilled zone for complete solution.
Requires depot LN2 supply infrastructure. Safety systems needed for oxygen displacement risk (proven solutions exist from food industry). LN2 production consumes ~0.5 kWh/kg, so lifecycle energy benefit depends on depot electricity source. Best suited for frozen-only or frozen-dominant routes.
Trigger: Major food retailer announces LN2 cold chain pilot; or industrial gas supplier launches transport refrigeration product
Earliest viability: 18-24 months
Monitor: Linde/Praxair (LN2 suppliers with food industry relationships), Air Liquide, Dearman/Highview Power (liquid air technology)
Risks & Watchouts
What could go wrong.
VIP durability under commercial delivery abuse is unproven at scale
Pilot program on 5-10 vehicles with intensive monitoring before fleet rollout; develop protective facing and modular replacement system; establish inspection protocols
Fleet operators may resist capital investment given uncertain EV adoption timelines
Offer leasing/service models; demonstrate ROI with pilot data; align with regulatory pressure for zero-emission urban delivery zones
Multi-vendor integration requires coordination across VIP, PCM, air curtain, and compressor suppliers
Identify or become system integrator; develop standardized retrofit kits for common van models; consider partnership with cargo body manufacturer
Temperature compliance documentation requirements may not accommodate predictive coasting approach
Engage with food safety regulators early; demonstrate continuous monitoring provides better compliance evidence than periodic checks; pilot with progressive customers
Driver bypass of infiltration control systems if perceived as slowing deliveries
Automatic activation tied to door sensors; clear training on energy/range benefits; consider incentive programs; design for minimal delivery time impact
Depot infrastructure investment for ice slurry or LN2 approaches requires significant capital
Start with largest depots serving highest-volume routes; demonstrate ROI before broader rollout; consider shared infrastructure models
Self-Critique
Where we might be wrong.
Medium
High confidence in physics and component-level performance based on literature; medium confidence in integrated system performance and real-world durability due to limited fleet-scale validation data
VIP durability under commercial delivery abuse—no fleet-scale data exists; pharmaceutical containers operate in gentler environments
Ice slurry handling operational complexity—mining industry operates at larger scale with different constraints; vehicle integration may surface unexpected issues
Driver adoption of door management systems—behavioral change is harder than technical change; resistance could undermine theoretical savings
Thermal model accuracy for predictive control—building HVAC data may not transfer to mobile applications with frequent door openings
PCM cycling stability over 5+ years—accelerated testing may not capture real-world degradation patterns
Thermoelectric spot cooling for frozen vault—modern materials achieve COP 1.0-1.5; may be viable for small frozen compartment where solid-state reliability is valued
Adsorption refrigeration using EV waste heat—EVs generate waste heat from power electronics that is currently rejected; could potentially drive absorption cycle for supplemental cooling
Compartmentalized access (side doors with curtained sections)—if operational changes are acceptable, never opening full cargo space could reduce infiltration by 80%+
VIP durability under commercial delivery abuse
First validation step includes 6-month pilot with intensive monitoring; protective facing and modular replacement system specified in concept
Driver adoption of door management systems
Mitigated through automatic activation and training, but behavioral change remains inherent risk. Monitoring driver compliance during pilot will provide data.
PCM cycling stability over 5+ years
First validation step covers 100 cycles; should extend to 500+ cycle accelerated aging test before fleet commitment
Ice slurry handling operational complexity
Bench-scale validation specified before vehicle integration; depot infrastructure investment gated on vehicle-side proof
Assumption Check
We assumed your constraints are fixed. If any can flex, here's what changes—and what to reconsider.
If cargo composition analysis shows strong skew, fleet mix strategy with specialized vehicles may outperform universal multi-temp solution. Separated frozen vault concept becomes more attractive.
If operational changes could reduce door openings to 30, thermal management becomes dramatically easier. May be worth investing in route optimization software alongside hardware solutions.
For fleet replacement cycles, purpose-built EV cold chain vehicles with integrated VIP bodies may be more cost-effective than retrofitting conventional vans.
If depot return at 4 hours is possible, simpler eutectic-only solution may be sufficient. Route analysis should identify which routes truly require 8-hour continuous operation.
Final Recommendation
Personal recommendation from the analysis.
If this were my fleet, I'd move in three parallel tracks starting tomorrow.
First, I'd order air curtain + strip door kits for every vehicle in the fleet. This is a no-brainer—$500-1,500 per vehicle, 2-week installation, 30-50% thermal load reduction. The ROI is measured in months, not years. I'd make this mandatory and track compliance through driver feedback and energy monitoring. This alone might get some vehicles close to the 9 kWh target on mild days.
Second, I'd select 5-10 vehicles for a VIP + PCM pilot. I'd work with va-Q-tec or a similar supplier to develop a retrofit kit for my most common van model. The goal is to validate the integrated system achieving 5-8 kWh consumption over a real 8-hour route with 50+ door openings. I'd instrument these vehicles heavily—temperature sensors throughout the cargo space, energy monitoring on every component, GPS-correlated door opening logs. Six months of data would tell me whether to roll out fleet-wide or iterate on the design.
Third, I'd commission a cargo composition analysis across all routes. If the data shows that 80% of routes have <25% frozen cargo, I'd seriously consider the separated frozen vault architecture. This could be even simpler and cheaper than the full integrated system while achieving similar energy performance.
The ice slurry concept is the most exciting long-term, but I wouldn't bet the fleet on it without proving the vehicle-side system works first. Once I have depot infrastructure investment on the table, I need confidence that the thermal distribution approach actually delivers. I'd run a bench-scale validation in parallel with the vehicle pilots.
What I would not do is wait for perfect information. The air curtain intervention is so obviously positive that delaying it to study more options is just leaving money on the table. Start there, learn fast, and iterate.