Property Data for Low-GWP Refrigerants:

What Do We Know and

What Don’t We Know?

Mark O. McLinden

Thermophysical Properties Division

National Institute of Standards and Technology

Boulder, Colorado USA

ASHRAE Winter Meeting, Las Vegas, NV

Seminar 6—Removing Barriers for Low-GWP Refrigerants

January 30, 2011

Removing Barriers to Low-GWP RefrigerantsLearning Objectives

Describe the climate change issue associated with high-GWP

refrigerants and the leading low-GWP options available today

Explain the refrigerant thermophysical property requirements needed

for new low-GWP refrigerants and how property data may be used

Be able to explain the challenges with measuring the flammability

properties of refrigerants that are only marginally flammable and

options to make these measurements

Explain the development history of hydrofluoroolefin low-GWP

refrigerants such as HFO-1234yf and why this new class of

compounds has unique properties

Describe the codes and regulations in the U.S., Europe, and Japan

that govern the use of low-GWP refrigerants, such as CO2, ammonia,

hydrocarbons, and HFOs, and the barriers in current standards for

their potential use

Apply learnings from seminar to begin selecting low-GWP refrigerants

for specific applications and begin designing new HVAC&R systems

with these refrigerants

Outline

What data do we need?

What are the candidate fluids?

Assessment of the data

How are the new fluids different?

Safety

toxicity (acute and chronic)

flammability

Environmental

ozone depletion potential (ODP)

green house warming potential (GWP)

atmospheric life (impacts ODP & GWP)

Materials

compatibility with metals, seals, etc.

lubricant

stability (hydrolysis, polymerization, etc.)

Performance

thermodynamic properties

transport properties

What Data Do We Need on a Refrigerant?

What Counts as “Low-GWP”?

R134a: GWP = 1430

(relative to CO2 w/ 100-year time horizon)

must have GWP << 1430

EU regulation for automotive A/C:

GWP < 150

(takes effect 2011 for new models)

North American Proposal to Montreal Protocol:

85 % phase-down by 2033 (0.15 1430 = 215)

Low-GWP Options (Current Fluids)

A31–24.8E170Ether

A1<1100.0718 (water)H2O

B2L<1–33.3717 (ammonia)NH3

A11–56.6744CO2

A320–0.5600 (butane)

A320–11.7600a (isobutane)

A320–42.1290 (propane)HC

B17727.8123HCFC

A2124–24.0152a

A11430–26.1134a

N/A (flam)12–37.6161

A2L675–51.732

N/A (flam)92–78.341HFC

ASHRAEclass

(tox/flam)

GWP[CO2 = 1]

NBP(˚C)Fluid

Hydro-Fluoro-Olefin (HFO)

C=C double bond

fluorine-containing

hydrogen-containing

Low-GWP Refrigerants—New Possibilities

2L(pending)6–19.01234ze(E)

A2L4–29.51234yfHFO

ASHRAEclass

(tox/flam)

GWP[CO2 = 1]

NBP(˚C)Fluid

This is the major “What We Don’t Know”

Additional HFOs Under Development

Announced in talks and conference papers

but not identified

Many (most?) are in patent or chemical literature

molecules themselves are not patentable

(production processes are patentable)

Blends w/ HFOs also being developed

azeotropic blends are patentable

zeotropic blends may be patentable

?

? ?

Low-GWP Refrigerants—Additional Possibilities

Interest in fluorinated ethers in 1990s

Examples among specialty solvents and heat-transfer fluids:

–fluorinated ethers

especially “segregated” ethers (all H on a single carbon)

–fluorinated ketones

non-flammable

low(ish) GWP

commercially available

(but boiling points higher than typical refrigerants)

Property Data and

Equations of State

Fluid Properties—Why Should ASHRAE Care?

Condenser

Evaporator

CompressorExpansion Valve

1

23

4

S2 = S1P2 = P3

q = 0T3 = TcondP3 = Psat(Tcond)

h4 = h3P4 = P1

q = 1T1 = TevapP1 = Psat(Tevap)

COPR =h1 h4h2 h1

= f Tevap ,Tcond , properties( )

Cycle analysis IS properties

Enthalpy

Pressure

1

23

4

Equation of State(Thermodynamic Properties)

Equation of State—Helmholtz Energy Form

A thermodynamically consistent representation of the properties of a fluid

=A

RT=

ideal + Niti dk

i

+ Njt j d j exp aj

l j( )j

+ Nktk dk exp ak k( )

lk

k

exp k k( )mk

where : = crit , = Tcrit T

“Gaussian terms” (critical region)

“traditional terms”

All other properties by differentiation:

p = RT 1 +

r

, CV = R 2

2

2

Equation of State—Mixture Form

Sum pure component contributions + “excess” (mixing) term

“excess” contribution

ideal solution terms

(ideal gas & “real gas”)

Evaluate at reduced T, (not mixture T, ) given by “reducing

parameters”

mix = xi i0 ,( )+ ln xi( )[ ]

i

+ xi ir ,( )

i

+ xij=i+1i

x jFij Nkik jk

k

“mixing function”

T*= xiTi

crit

i=1

n

+ xij=i+1

n

i

n 1

x j ij 1* =

xi

icrit

i=1

n

+ xij =i +1

n

i

n 1

xi ij

= T* T =*

Adjustable parameters , , Fij + mixing function give great flexibility in

fitting mixtures with limited or extensive data

Data Needed to Establish Equation of State(Thermodynamic Properties)

Critical parameters (Tcrit, pcrit, crit)

Vapor pressure as f(T)

Density (liquid, vapor, supercritical)

Ideal-gas heat capacity [Cp as p 0]

Caloric data (liquid): Cp or CV or w

Recent R1234ze(E) Data

225

275

325

375

425

Tem

pera

ture

(K

)

0 200 400 600 800 1000 1200 1400

Density (kg•m–3)

saturation

filling 3

filling 2

filling 1

0.0

1.0

2.0

3.0

4.0

Pre

ssur

e (M

Pa)

260 280 300 320 340 360 380

Temperature (K)

Critical region data, Higashi, et al.,

J. Chem. Eng. Data (2010)

p- -T and vapor pressure data

McLinden, et al., Int. Ref. Conf. Purdue (2010)

Liquid-phase speed of sound

Lago, et al. (2010)

200

300

400

500

600

700

800

w (

m/s

)

250 275 300 325 350 375

Temperature (K)

p = 10 MPa

p = 2 MPa

Liquid-phase heat capacity

Tanaka, et al., J. Chem. Eng. Data (2010)

Equation of State—Fitting Process(R1234ze(E) as Example)

Non-linear in the parameters

must start with initial guess for EOS

propane equation with 14 terms

Objective function:

minimize sum of squares

Fit numerical coefficients

Fit exponents on temperature terms

(0 ~ 5), max for R1234ze is 2.5

Fixed exponents on density terms

integers 1

Highly iterative process

(1000’s of iterations)

Equation of State—Fitting Process (Constraints)

Numerous further constraints

e.g., shape of critical isotherm

EOS with correct shape:

small addition to sum of squares

EOS with incorrect shape:

large penalty to sum of squares

for < crit :

p> 0;

2 p2 < 0;

3p3 > 0;

4 p4 < 0;

for > crit :

p> 0;

2 p2 > 0;

3p3 > 0;

4 p4 > 0;

add expp

to sum of squares

0.0

2.0

4.0

6.0

8.0

Pre

ssur

e (M

Pa)

0 200 400 600 800 1000 1200 1400

Density (kg/m3)

T = Tcrit

saturatedliquid

saturatedvapor

criticalpoint

calculatedpoints

Data Comparisons and

Assessments

Data Comparisons—EOS vs. p- -T Data, R1234ze(E)

1.31Grebenkov (vapor)

0.52Grebenkov (liquid)

Source (%)

McLinden et al. 0.032

Tanaka + 0.087

Grebenkov (sat’n) 0.22

–0.30

–0.20

–0.10

0.00

0.10

0.20

0.30

100

(e

xp –

E

OS)/

EO

S o

r 10

0 (p

exp

– p

EO

S)/

p EO

S

225 275 325 375 425

Temperature (K)

–0.30

–0.20

–0.10

0.00

0.10

0.20

0.30

0 200 400 600 800 1000 1200 1400

Density (kg m–3)

Tcrit

crit

Data Comparisons—EOS vs. psat Data, R1234ze(E)

1.78Grebenkov

Source (%)

McLinden et al. 0.028

Tanaka + 0.039

–0.20

–0.10

0.00

0.10

0.20

100

(pe

xp –

pE

OS)/

p EO

S

240 280 320 360 400

Temperature (K)

Data Comparisons—EOS vs. Cp and Sound Speed Data

R1234ze(E)

0.49Lago (w)

Source (%)

Tanaka (Cp) + 2.16

–8.00

–6.00

–4.00

–2.00

0.00

2.00

4.00

100

(Xe

xp –

XE

OS)/

XE

OS

260 300 340 380

Temperature (K)

Data Summary

limited datagood (limited data)1234ze(E)

limited datagood (limited data)1234yfHFO

HFC

fair (limited data)fair (limited data)E170DME

excellentexcellent718 (water)H2O

goodfair (!)*717 (ammonia)NH3

very goodexcellent744CO2

very goodvery good600 (butane)

very goodvery good600a (isobutane)

very goodexcellent290 (propane)HC

fair (visc)–good (t.c.)fair (old form)123HCFC

goodfair (old form)152a

goodgood134a

goodgood32

Transport(visc. & t.c.)

Eqn. of State(thermo)Fluid

*new EOS under development

Thermo Data Summary—Blends

available, generally good,

but dated (2001)

predictive model for

refrigerant blends

proprietary data (azeotropes disclosed)

Arakawa et al. (2010): R32/1234yf*HFO + HFC

very good data and models (natural gas)HC blends

virtually no dataHFC + ethers

limited pairs, generally only VLE dataHFC + HC

numerous mixture pairs, but generallyonly VLE data (i.e., no (T,p,x))

other HFC blends

very good data and modelHFC 32/125/134a

Data AssessmentBlend Class

*additional work in progress in Japan0.12

0.18

Pressure

0 1Composition

azeotrope

0.200

0.400

0.600

0.800

1.000

Pre

ssu

re

0 1

Composition

bubble line (liquid)

dew line (vapor)

Refrigerant Molecules Are Getting More Complex

Increased complexity:

changes fundamental shape of thermodynamic surface

increased flash losses

The cycle may need to be modified

CFC-12 HFC-134a HFO-1234yfNBP: (–29.8 ˚C) (–26.1 ˚C) (–29.5 ˚C)

Enthalpy (mass basis)

Pre

ssu

re

R12

R134a

R1234yf Te

mp

era

ture

Entropy (mass basis)

R12

R134a

R1234yf

Conclusions

Adequate data for a range of traditional and new low-GWP fluids

But much work remains:

–Additional candidates are proprietary

–Blend data w/ HFOs are very limited

New fluids may require modified cycles

Questions?