# Vacuum System Troubleshooting: Why Your Liquid Ring Pump or Steam Ejector Won't Pull Deep Vacuum

# A Diagnostic Engineering Guide to Quantifying Air Ingress Leak Rates, Eliminating Summer Seal-Water Cavitation, and Restoring Multi-Stage Steam Jet Ejector Performance

# Executive Summary

In Active Pharmaceutical Ingredient (API) synthesis, specialty chemical manufacturing, and solvent recovery, process vacuum is as vital as steam or electricity. Vacuum operations dictate solvent stripping temperatures, reaction distillation rates, vacuum tray drying kinetics, and overhead condenser thermal performance.

When an industrial vacuum system stalls—for example, pulling only $80\text{ to }120\text{ mbar}$ instead of the design $2\text{ to }5\text{ mbar}$ —the plant consequences are immediate and severe:

Batch Cycle Extensions: Distillations that should finish in 4 hours drag on for 24 to 36 hours.
Thermal Degradation of APIs: Elevated boiling temperatures cause heat-sensitive pharmaceutical intermediates to decompose, generating out-of-specification related substances.
Solvent Condensation Failure: Low vacuum reduces vapor velocity and condenser logarithmic mean temperature difference ( $\Delta T_{lm}$ ), causing volatile solvents to bypass chilled overhead condensers and vent directly into the atmosphere or scrubber.

When vacuum degrades, plant operators frequently misdiagnose the problem: they replace pump mechanical seals, switch motors, or blame the utility department. In reality, vacuum failures are governed by rigorous physical phenomena: mass leak rates from air ingress, seal fluid vapor pressure limits ( $P_v$ ), sonic shock detachment in steam nozzles, or non-condensable vapor blanketing.

This technical troubleshooting guide provides the thermodynamic equations, leak rate quantification methods, mechanical failure diagnostic trees, and a step-by-step worked industrial case study to systematically restore deep plant vacuum.

# 1. Quantifying System Air Ingress: The Pressure Rise (Drop) Test

Before disassembling pumps or blaming ejectors, an engineer must first answer the foundational diagnostic question: Is the vacuum deficit caused by mechanical equipment failure, or by excessive ambient air in-leakage through piping and reactor flanges?

                 [ Reactor System Isolated Under Vacuum: P0 ]
                                      │
                                      ▼ Clock Starts (t = 0)
                 [ System Pressure Rises Naturally: P(t) ]
                                      │
                                      ▼ Clock Stops (t = 15 - 30 min)
┌──────────────────────────────────────────────────────────────────────────┐
│ Leak Rate (mbar·L/s) = [ System Volume (V) × ΔP ] / Isolation Time (Δt)  │
└──────────────────────────────────────────────────────────────────────────┘

# 1.1 The Mathematical Leak Rate Formula

The standard drop test (pressure rise method) isolates the evacuated vessel and overhead condenser train from the vacuum skid by closing the primary isolation valve, then recording pressure rise over time:

Q_{leak} = \frac{V_{system} \cdot (P_2 - P_1)}{\Delta t}

\dot{m}_{air} = \frac{V_{system} \cdot M_{air}}{R \cdot T} \cdot \left( \frac{dP}{dt} \right) \approx \frac{V_{system} \cdot (P_2 - P_1)}{\Delta t} \times \left( \frac{29.0}{83.14 \cdot T} \right)

Where:

$Q_{leak}$ : Volumetric leak rate ( $\text{mbar}\cdot\text{L/s}$ or $\text{kg/h}$ )
$V_{system}$ : Total internal volume of reactor, vapor duct, and condenser shell ( $\text{L}$ or $\text{m}^3$ )
$P_1, P_2$ : Absolute pressure at start and end of isolation ( $\text{mbar}$ )
$\Delta t$ : Elapsed isolation duration ( $\text{seconds}$ )
$T$ : Absolute temperature of vessel ( $\text{K}$ )
$\dot{m}_{air}$ : Equivalent mass flow rate of ambient dry air entering the system ( $\text{kg/h}$ )

# 1.2 HEI Air Ingress Acceptable Standards (Heat Exchange Institute)

The Heat Exchange Institute (HEI) standards establish maximum allowable air leakage rates for industrial chemical equipment:

Total System Volume ( $V_{system}$ )	Acceptable Leak Rate (Clean Plant)	Investigation Required	Immediate Action Mandatory (Severe Leak)
$\le 2.0\text{ m}^3$ ( $2,000\text{ L}$ )	$< 1.5\text{ kg/h}$ ( $0.4\text{ mbar}\cdot\text{L/s}$ )	$1.5\text{–}3.0\text{ kg/h}$	$> 3.5\text{ kg/h}$ (Blown gasket or open valve)
$5.0\text{ m}^3$ ( $5,000\text{ L}$ )	$< 2.5\text{ kg/h}$ ( $0.7\text{ mbar}\cdot\text{L/s}$ )	$2.5\text{–}5.0\text{ kg/h}$	$> 6.0\text{ kg/h}$
$10.0\text{ m}^3$ ( $10,000\text{ L}$ )	$< 4.0\text{ kg/h}$ ( $1.1\text{ mbar}\cdot\text{L/s}$ )	$4.0\text{–}8.0\text{ kg/h}$	$> 10.0\text{ kg/h}$

📌 Important

Diagnostic Rule of Thumb: If the pressure rises rapidly in the first 2 minutes and then levels off into a flat horizontal plateau, you do not have an air leak. You have solvent outgassing / boiling off from a wet reactor heel. True air in-leakage displays a strictly linear pressure rise ( $dP/dt = \text{constant}$ ) up to atmospheric pressure.

# 2. Liquid Ring Vacuum Pump (LRVP) Failure Modes: Summer Cavitation

Liquid Ring Vacuum Pumps (LRVPs) are the most rugged vacuum machines in the chemical industry because they handle wet, condensable vapors without mechanical damage. However, their ultimate achievable vacuum is strictly limited by the vapor pressure of the circulating liquid sealant.

    Circulating Seal Water Temp Rises: 20°C ──► 32°C (Summer Cooling Tower)
                                       │
                                       ▼
    Seal Water Vapor Pressure (Pv) Spikes: 23 mbar ──► 48 mbar!
                                       │
                                       ▼
┌────────────────────────────────────────────────────────────────────────┐
│ P_suction Approaches Pv ──► CAVITATION: Violent Vapor Bubble Implosion │
│ • Severe Mechanical Vibration (Rattling "Gravel" Sound in Pump Casing) │
│ • Impeller Vane Pitting & Erosion                                      │
│ • Ultimate Vacuum Limited to ~60 mbar (Loss of High Vacuum Function)   │
└────────────────────────────────────────────────────────────────────────┘

# 2.1 The Seal Water Vapor Pressure Barrier

The theoretical maximum vacuum an LRVP can achieve ( $P_{suction,min}$ ) is fundamentally constrained by the vapor pressure of the seal liquid ( $P_v(T_{seal})$ ):

P_{suction,min} \ge P_v(T_{seal}) + \Delta P_{cavitation}

Where $\Delta P_{cavitation} \approx 10\text{ to }15\text{ mbar}$ is the hydrodynamic margin required to prevent flashing at the impeller blade tip suction eye.

Seal Water Temperature ( $T_{seal}$ )	Water Vapor Pressure ( $P_v$ )	Theoretical Max Vacuum	Realistic Safe Operating Vacuum Limit
$15^\circ\text{C}$ (Chilled Water)	$17.0\text{ mbar}$	$27\text{ mbar}$	$30\text{ mbar}$
$25^\circ\text{C}$ (Design Spring/Autumn)	$31.7\text{ mbar}$	$42\text{ mbar}$	$45\text{ mbar}$
$32^\circ\text{C}$ (Summer Cooling Tower)	$47.6\text{ mbar}$	$60\text{ mbar}$	$65\text{ mbar}$ (Severe Cavitation!)
$38^\circ\text{C}$ (Fouled Heat Exchanger)	$66.3\text{ mbar}$	$80\text{ mbar}$	$85\text{ mbar}$ (Total Vacuum Loss)

# 2.2 Sizing the Cavitation Relief Valve

When operating near the seal fluid vapor pressure, an LRVP emits a distinct, violent sound akin to pumping gravel or marble balls. This mechanical vibration shatters carbon-ceramic mechanical seal faces in days.

Immediate Operating Fix: Crack open the atmospheric air ballast valve (cavitation relief valve) located on the suction manifold. Introducing a small, controlled stream of air ( $1\text{–}2\text{ kg/h}$ ) raises the local suction pressure slightly above $P_v$ , cushioning the imploding bubbles and immediately stopping cavitation erosion.
Permanent Engineering Fix: Route chilled water ( $10\text{–}15^\circ\text{C}$ ) to a plate heat exchanger on the pump closed recirculation loop, maintaining sealant temperature below $20^\circ\text{C}$ year-round.

# 3. Steam Jet Ejector Troubleshooting: Sonic Flow & Diffuser Detachment

For deep vacuum requirements ( $0.5\text{ to }10\text{ mbar}$ ), pharmaceutical plants deploy multi-stage steam jet ejector systems (typically 2-stage or 3-stage with inter-condensers).

Steam ejectors contain zero moving parts, making their operation purely aerodynamic: high-pressure motive steam expands through a convergent-divergent de Laval nozzle, accelerating to supersonic velocities (Mach 3 to 4, $\approx 1,200\text{ m/s}$ ), entraining suction gas via momentum transfer, and recompressing through a diffuser.

Motive Steam (4 - 6 bar) ──► [ Convergent-Divergent Nozzle (Mach 3.5) ]
                                            │ High-Velocity Jet
Suction Gas (P_suction)  ──────────────► [ Mixing Chamber ]
                                            │ Momentum Transfer
                                            ▼
                             [ Supersonic Diffuser Throat ] ──► Discharge (P_discharge)

# 3.1 The 4 Fatal Steam Ejector Failure Mechanisms

Failure Mode	Physical Root Cause	Diagnostic Indicator	Corrective Action
1. Low Motive Steam Pressure	Motive pressure falls below design ( $P_{motive} < 0.95 \times P_{design}$ ). Shock wave detaches from diffuser throat.	Suction vacuum suddenly collapses into unstable, violent cycling (gauge needle swings).	Increase header steam pressure to $105\%$ of design stamp; inspect upstream steam PRV.
2. Wet Motive Steam (Droplet Entrainment)	Steam quality is $<99\%$ . Moisture droplets erode nozzle throat and flash prematurely.	Ejector body feels cool/cold to the touch; clicking/chattering noises; rapid nozzle wear.	Install an oversized condensate knockout separator with a high-capacity inverted bucket trap immediately upstream of the ejector.
3. Nozzle Throat Erosion or Scale Buildup	Silica/boiler carryover scales the nozzle throat, reducing throat area ( $A_{throat}$ ), or erosive wet steam enlarges it.	Steam consumption increases while entrainment capacity drops; unable to reach ultimate vacuum.	Measure nozzle throat diameter with pin gauges; clean scale with ultrasonic bath or replace nozzle.
4. Inter-Condenser Barometric Leg Flooding	Condensate leg back-siphons into ejector discharge because the barometric leg height is $<10.5\text{ m}$ or seal pot overflows.	Second-stage ejector discharge turns cold; high liquid carryover in exhaust vent; complete loss of vacuum.	Verify barometric drop leg is strictly vertical and submerged $\ge 300\text{ mm}$ below seal pot overflow weir.

# 4. Overhead Condenser Troubleshooting: Non-Condensable Thermal Blanketing

A frequent root cause of vacuum loss during high-boilup reactions is overhead condenser failure, mistakenly diagnosed as vacuum pump malfunction.

When condensable solvent vapors (e.g., Methanol, Toluene, Acetone) mix with in-leaking ambient air or reaction off-gases ( $CO_2$ , $HCl$ , $H_2$ ), the vapor mixture flows across the shell-and-tube or plate condenser:

  Condensable Vapor + Air Film ──► [ Cold Tube Wall ]
                                          │
                                          ▼
  Solvent Condenses Rapidly ──► Air Molecule Boundary Layer Remains (Thermal Blanket!)
                                          │
                                          ▼
┌────────────────────────────────────────────────────────────────────────┐
│ Heat Transfer Coefficient (U) Collapses from 600 W/m²·K down to 35 W/m²·K│
│ Overhead Condenser Chokes ──► Huge Vapor Flow Bypasses into Vacuum Skid│
└────────────────────────────────────────────────────────────────────────┘

# 4.1 The Thermal Blanketing Equation

The presence of even $1.5\text{ to }3.0\text{ wt\%}$ of non-condensable inert gas (air) in the incoming vapor stream creates a stagnant gas film over the condensing liquid film. The effective heat transfer coefficient collapses:

\frac{1}{U_{effective}} = \frac{1}{h_{condensate}} + \frac{1}{h_{gas\_film}} + R_{wall} + \frac{1}{h_{coolant}}

Where $h_{gas\_film} \approx 20\text{–}50\text{ W/m}^2\cdot\text{K}$ (compared to $h_{condensate} \approx 1,500\text{–}2,500\text{ W/m}^2\cdot\text{K}$ ).

🛑 Caution

Symptom of Air Blanketing: The overhead condenser shell feels cold on the lower half, but hot on the top and vent nozzle. The vent line going to the vacuum pump is warm (

>40^\circ\text{C}

), carrying massive quantities of un-condensed solvent vapor into the vacuum pump, overloading its suction stage.

# 5. Worked Industrial Case Study: Diagnosing a 6,000 L Reactor Vacuum Stall

# 5.1 The Plant Failure Baseline

Equipment: 6,000 L Glass-Lined Reactor (GLR) coupled to a primary shell-and-tube condenser ( $25\text{ m}^2$ Hastelloy C-22), vent secondary condenser ( $6\text{ m}^2$ ), and a 2-stage vacuum skid (Roots blower booster + Liquid Ring Vacuum Pump).
Process Operation: Vacuum concentration of a sensitive intermediate in Ethyl Acetate ( $BP_{normal} = 77.1^\circ\text{C}$ ).
Design Operating Vacuum: $50\text{ mbar}$ absolute ( $T_{boiling} \approx 18^\circ\text{C}$ , maintaining batch temperature below $25^\circ\text{C}$ to prevent degradation).
The Failure: During summer operations, the system was unable to pull below $135\text{ mbar}$ . Batch temperature rose to $46^\circ\text{C}$ , causing product degradation and extending the distillation duration from 3 hours to 19 hours.

# 5.2 Step-by-Step Diagnostic Protocol

# Step 1: Drop Test (Pressure Rise Rate)

The reactor and condenser train were evacuated to $135\text{ mbar}$ , then isolated by closing the primary 6-inch pneumatic butterfly valve:

Total system volume ( $V_{system}$ ): $6.0\text{ m}^3\text{ (reactor)} + 1.2\text{ m}^3\text{ (condensers + piping)} = 7.2\text{ m}^3 = 7,200\text{ L}$ .
Starting pressure ( $P_1$ ): $135\text{ mbar}$ .
Pressure after $10\text{ minutes}$ ( $600\text{ s}$ ): $P_2 = 188\text{ mbar}$ .
Pressure Rise Rate:

\frac{dP}{dt} = \frac{188 - 135}{600\text{ s}} = \frac{53\text{ mbar}}{600\text{ s}} = 0.0883\text{ mbar/s}

Calculated Air Ingress Mass Flow ( $\dot{m}_{air}$ ):

\dot{m}_{air} = \frac{7,200\text{ L} \times 0.0883\text{ mbar/s}}{1,000} \times \left( \frac{29.0 \times 3,600}{83.14 \times 298} \right) \approx \mathbf{2.68\text{ kg/h of air in-leakage}}

Diagnostic Assessment: A leak rate of $2.68\text{ kg/h}$ for a $7.2\text{ m}^3$ system is slightly above the clean plant threshold ( $2.5\text{ kg/h}$ ), but wholly insufficient to cause a vacuum collapse from $50\text{ mbar}$ to $135\text{ mbar}$ . Air ingress was ruled out as the primary root cause.

# Step 2: Liquid Ring Vacuum Pump Audit

Measured seal-water temperature entering the LRVP: $34.5^\circ\text{C}$ (Supplied from a fouled summer cooling tower loop).
Pure water vapor pressure at $34.5^\circ\text{C}$ :

P_v(34.5^\circ\text{C}) = \mathbf{54.8\text{ mbar}}

The pump suction pressure was operating at $135\text{ mbar}$ , but deep metallic rattling and cavitation vibrations were detected on the pump casing.
Furthermore, analysis of the seal water revealed $8.2\text{ wt\% dissolved Ethyl Acetate}$ , carried over from condenser bypass.
The vapor pressure of an Ethyl Acetate/Water binary mixture at $34.5^\circ\text{C}$ is $128\text{ mbar}$ !