Document [original]

Phosphorus-containing (bio) compounds in combination with

expandable graphite: Flame retardancy and smoke/toxicity

suppression of flexible polyurethane foams through

synergistic effects

vorgelegt von

M. Sc.

Yin Yam Chan

an der Fakultät III – Prozesswissenschaften

der Technischen Universität Berlin

zur Erlangung des akademischen Grades

Doktorin der Naturwissenschaften

- Dr. rer. nat. -

genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Dr.-Ing. Sören Müller

Gutachter: Prof. Dr.-Ing. Dietmar Auhl

Gutachter: Prof. Dr. habil. Bernhard Schartel

Tag der wissenschaftlichen Aussprache: 16. Dezember 2022

Berlin 2023

To my parents

I can do all this through Him who gives me strength.

Bible (NIV) – Philippians 4:13

Acknowledgements

First of all, I would like to thank God. He has given me strength and encouragement and has

accompanied me through the ups and downs.

This endeavor would not have been possible without my esteemed supervisor - Prof. Dr. habil. Bernhard

Schartel. I am extremely grateful that he gave me an opportunity to start my Ph.D. in his group. Although

I am not so familiar with the topic at the beginning, thanks for his patience in teaching and guiding me

over the years. I know sometimes I was not confident enough in my work, but he always encourages me

and believes in me to get the work done.

I would like to express my deepest appreciation to Prof. Dr.-Ing Dietmar Auhl for agreeing to supervise

my work and progress, providing me with invaluable feedback and reviewing this doctoral dissertation.

Thanks should also go to all my colleagues in the working group 7.5 at BAM over the years, Dr.

Alexander Battig, Weronika Tabaka, Sandra Falkenhagen, Dr. Robert Gleuwitz, Dr. Benjamin Zirnstein,

Dr. Yi Tan, Dr. Xuebao Lin, Vitus Hupp, Dr. Lars-Hendrik Daus, Sebastian Goller, Maria Jauregui,

Daniel Rockel and Bettina Strommer... I appreciate for their friendly help with my project and, scientific

and non-scientific discussions. I really had great memories with them over the years. I would like to

extend my sincere thanks to Patrick Klack for providing the technical support and the workshop team,

especially Michael Schneider and Tobias Lauterbach, for cutting the samples.

I would like to offer my special thanks to Dr. Simone Krüger and Tina Raspe for their assistance with

the smoke density chamber measurements and Dietmar Schulze for his contribution to the mechanical

testing for the materials.

I had the pleasure of collaborating with our Chinese partners at the University of Science and

Technology China (USTC), especially Prof. Yuan Hu, Dr. Chao Ma and Feng Zhou. They provided me

a lot of resources and I learned more about the topic from them during my 5-month stay in China.

In addition, I would like to express gratitude to Dr. Doris Pospiech and Andreas Korwitz from Leibniz

Institute of Polymer Research Dresden (IPF Dresden) for supporting my research ideas and giving me

precious advice on my idea during my two stays in Dresden.

Lastly, I would like to acknowledge my family, especially my parents. They have given me a lot of

support and freedom to achieve my dreams and goals in Germany. I would also want to give a big thank

to my best friend Kayee Poon who is always there for me, always encourages me and believes in me.

May Almighty God continue to bless you all.

Zusammenfassung

Polyurethan-Weichschaum (FPUF) ist aufgrund seiner offenzelligen Struktur anfällig für Brände. FPUF

besteht hauptsächlich aus Kohlenwasserstoffen, die brennbar sind. Um den Flammschutz zu verbessern,

wurden in der Industrie häufig halogenierte Flammschutzmittel, wie chlor- und bromhaltige

Verbindungen, in FPUF eingesetzt. Da halogenierte Flammschutzmittel giftig für Lebewesen und die

Umwelt sind, werden sie durch halogenfreie Flammschutzmittel ersetzt. Daher synthetisierte der

Kooperationspartner eine Reihe neuartiger phosphor-stickstoffhaltiger 9,10-Dihydro-9-oxa-10-

phosphaphenanthren-10-oxid (DOPO)-Derivatsalze und ein neuartiges flüssiges phosphorhaltiges

Flammschutzmittel (P-FR) - Bis([dimethoxyphosphoryl]methyl)phenylphosphat (BDMPP) und setzte

sie in FPUF ein. Das Brandverhalten dieser FPUF wurde im ersten Teil dieser Arbeit untersucht.

Diese wissenschaftliche Arbeit konzentriert sich hauptsächlich auf die Erhöhung der Flammwidrigkeit

und die Unterdrückung des Rauches von FPUF während der Verbrennung durch die Nutzung des

Synergieeffekts zwischen halogenfreien (biobasierten) phosphorhaltigen Flammschutzmitteln und

expandierbarem Graphit (EG). Der Clou ist, dass ein phosphorgepfropftes Polyol auf Sojabasis

synthetisiert und zusammen mit EG in FPUF eingesetzt wurde. Der Brandrückstand zeigt, dass es einen

synergistischen Effekt von Phosphor und EG in der kondensierten Phase gibt, und die Konzentration

der Phosphorverbindung zeigt ein nichtlineares Verhalten in Bezug auf die Flammhemmung. Neben

dem mit Phosphor gepfropften Polyol auf Sojabasis wurde dem FPUF auch BDMPP zusammen mit EG

zugesetzt. Die Ergebnisse der Brandtests von FPUF, die BDMPP und EG enthalten, zeigen ebenfalls

einen hervorragenden Synergieeffekt zwischen P-FR und EG.

Da die übermäßige Ausbeutung von Erdöl in den letzten Jahrzehnten ein heißes Thema ist, ist die

Verwendung erneuerbarer Ressourcen wie Pflanzenöle ein Ansatz, um die Verwendung von

petrochemischen Produkten zu reduzieren. Neben Sojaöl wurde auch Rizinusöl verwendet, um den

biologischen Gehalt von FPUF zu erhöhen. In Kombination mit anderen handelsüblichen

Flammschutzmitteln und Additiven wurden das Brandverhalten und das Rauchverhalten von FPUF

untersucht. Die Kombination von P-FR und EG zeigt auch in diesen Systemen einen vielversprechenden

Synergieeffekt in FPUF.

Das Verständnis der Brandphänomene von Polyurethanschaum (PUF) ist ein Schlüssel zur

Verbesserung seines Flammschutzes. Daher wurden in der Arbeit die chemischen Komponenten und

die physikalische Struktur von PUF klar dargestellt und das Brennverhalten von PUF vorgeschlagen,

wenn es einem Wärmestrom von oben ausgesetzt wird. Der Synergieeffekt zwischen P-FR und EG

wurde detailliert beschrieben, und die Verwendung von erneuerbaren Rohstoffen in PUF wurde

diskutiert, um Einblicke in das Thema Nachhaltigkeit zu geben.

Die Kombination von P-FR und EG ist in der Regel auch die erste Wahl in der Industrie, um die

Flammhemmung von PUF extrem zu verbessern. Drei industrielle Benchmark-PUFs, die P-FR und EG

iii

enthalten und in verschiedenen Anwendungen eingesetzt werden, um die höchsten Anforderungen zu

erfüllen, wurden untersucht und analysiert. Es wurde somit der Aktuelle Stand der Technik diskutiert.

Die Kombination von P-FR und EG stellt die klassenbeste Mehrkomponente in Bezug auf Flammschutz

und Rauchverhalten in PUFs ist.

Anhand verschiedener Brandtests wurde in dieser Arbeit der Synergieeffekt zwischen P-FR und EG in

PUF, insbesondere FPUF, umfassend untersucht, um in Zukunft wissensbasiert ein effektiveres

Mehrkomponenten-Flammschutzsystem für PUF zu entwickeln.

Abstract

Flexible polyurethane foam (FPUF) is prone to fire due to its open-cell structure, and the chemical

composition of FPUF is mainly composed of hydrocarbons which are flammable. To improve the flame

retardancy, halogenated flame retardants, such as chlorine- and bromine-containing compounds, were

commonly used in FPUF in the industry. Since halogenated flame retardants are poisonous to the living

organisms and the environment, they are being replaced by halogen-free flame retardants. Therefore,

the cooperation partner synthesized a series of novel phosphorus-nitrogen-containing 9,10-dihydro-9-

oxa-10-phosphaphenanthrene-10-oxide (DOPO) derivative salts and a novel liquid phosphorus-

containing flame retardant (P-FR) - bis([dimethoxyphosphoryl]methyl) phenyl phosphate (BDMPP) and

applied them in FPUF. The fire behavior of these FPUF foams were investigated in the first part of this

work.

This scientific work mainly focuses on enhancing the flame retardancy and suppressing smoke of FPUF

during burning by using the synergistic effect between halogen-free (bio-based) P-FRs and expandable

graphite (EG). The highlight is that a phosphorus-grafted soybean-based polyol was synthesized and

was applied in FPUF along with EG. The fire residue reveals that there is a synergistic effect of P-FR

and EG in the condensed phase, and the concentration of phosphorus compound exhibits non-linear

behavior in terms of flame retardancy. Apart from the phosphorus-grafted soybean-based polyol,

BDMPP was also added into FPUF along with EG. The results from the fire tests of FPUF containing

BDMPP and EG also show an excellent synergistic effect between P-FR and EG.

As overexploitation of petroleum is a hot topic in recent decades, using renewable resources such as

plant oils is an approach to reduce the use of petrochemical products. Besides soybean oil, castor oil was

used to increase the biological content of FPUF. Combined with other commercial flame retardants and

additives, fire performance and smoke behavior of the FPUF were investigated. The combination of P-

FR and EG shows also in these systems a very promising synergistic effect in FPUF.

Understanding the fire phenomena of polyurethane foam (PUF) is a key to improve its flame retardancy.

Therefore, the work clearly depicted the chemical components and the physical structure of PUF and

proposed the burning behavior of PUF when it is subjected to a heat flux on top. The synergistic effect

between P-FR and EG was described in detail, and the use of renewable feedstocks in PUF was discussed

to provide insights into the topic of sustainability.

The combination of P-FR and EG is generally the first choice used in the industry for greatly enhancing

flame retardancy of PUF. Three industrial benchmark PUFs containing P-FR and EG used in different

applications to fulfill challenging demands were investigated and analyzed. Thus, the current state-of-

the-art was discussed. The combination of P-FR and EG is the best-in-class multicomponent in terms of

flame retardancy and smoke behavior in PUFs.

Through different fire tests, this work comprehensively investigated the synergistic effect between P-

FR and EG in PUF, especially FPUF, to enable evidence-based development of more effective

multicomponent flame retardant system for PUF in the future.

Index of abbreviations

ATH Aluminum trihydrate

BDMPP Bis([dimethoxyphosphoryl]methyl) phenyl phosphate

CAS Castor oil

CO Carbon monoxide

CuO Copper (II) oxide

DOPO 9,10-Dihydro-9-oxa-10-phosphaphenanthrene-10-oxide

DTG The first derivative of TG

EG Expandable graphite

EHC Effective heat of combustion

FPUF Flexible polyurethane foam

FTIR Fourier transform infrared spectroscopy

HCN Hydrogen cyanide

HRR Heat release rate

LOI Limiting oxygen index

MA Melamine

MARHE Maximum average heat emission

MHD Magnesium hydroxide

NMR Nuclear magnetic resonance spectroscopy

OP Exolit® OP560 polyol

P-FR Phosphorus-containing flame retardant

PHRR Peak heat release rate

PUF Polyurethane foam

RPUF Rigid polyurethane foam

SDC Smoke density chamber

SEM Scanning electron microscopy

vii

2,4-TDI 2,4-Toluenediisocyanate

2,6-TDI 2,6-Toluenediisocyanate

TDI Toluenediisocyanate

TG Thermogravimetry

TGA Thermogravimetry analysis

THR Total heat release

TSR Total smoke release

UL 94 HBF Underwriters Laboratories 94 horizontal burning test

Table of contents

Acknowledgements…………………………………………………………………………………..….i

Zusammenfassung……………………………………..…………………………………………….ii-iii

Abstract……………………………………………………………………………………………....iv-v

Index of abbreviations...……………………………………………………………..…...…………vi-vii

1 Introduction ………………………………………………………………………………….…..…1-3

2 Scientific background – Fire behavior and flame retardancy ………………………………………4-6

2.1 Flame retardancy of polyurethane foams……………………………………………………………7

2.1.1 Factors affect the flammability of polyurethane foams…………………………………………7-8

2.2 Synergistic effect of flame retardants……………………………………………….………….…8-9

2.3 Materials…………………………………………………………………………………………9-14

2.4 Methods………………………………………………………………………………...………14-17

2.4.1 Nuclear magnetic resonance (NMR)……………………………………………...……………..14

2.4.2 Thermogravimetry analysis (TGA)……………………………………………………...………15

2.4.3 Thermogravimetry coupled with Fourier transform infrared spectroscopy (TG-FTIR)…….…..15

2.4.4 Limiting oxygen index (LOI) …………………………………………………………...………15

2.4.5 Underwriters Laboratories 94 horizontal burning test (UL 94 HBF)……………………………16

2.4.6 Cone calorimeter………………………………………………………………………..……16-17

2.4.7 Scanning electron microscopy (SEM) ………………………………...………………….……..17

2.4.8 Mechanical test………………………………………………………………………….……….17

2.4.9 Smoke density chamber (SDC) coupled with FTIR….………………………………………… 17

3 Discussion of the results………………………………………………………………….………18-19

3.1 Main messages from the publications……………………………………………………….…18-19

4 Publications……………………………………………………………………………………..20-163

4.1 Effects of novel phosphorus-nitrogen-containing DOPO derivative salts on mechanical properties,

thermal stability and flame retardancy of flexible polyurethane foam……………………………..20-32

4.2 Synthesis of ethyl (diethoxymethyl)phosphinate derivatives and their flame retardancy in flexible

polyurethane foam: Structure-flame retardancy relationships………………………………..……33-43

4.3 A liquid phosphorous flame retardant combined with expandable graphite or melamine in flexible

polyurethane foam………………………………………………………………………………….44-59

4.4 Flame retardant flexible polyurethane foams based on phosphorous soybean-oil polyol and

expandable graphite………………………………………………………………………..……….60-77

4.5 Flame retardant combinations with expandable graphite/phosphorus/CuO/castor oil in flexible

polyurethane foams…………………………………………………………………………..…...78-117

4.6 It takes two to tango: Industrial benchmark PU-foams with expandable graphite/P-flame retardant

combinations……..………………………………….……..………………………………….…118-142

4.7 It takes two to tango: Synergistic expandable graphite – phosphorus flame retardant combinations

in polyurethane foams…………………………………………………………………………...143-167

5 Summary……………………………………………………………………………………….168-169

6 References……………………………………………………………………………..………170-174

1 Introduction

Synthetic polymeric materials can be seen almost everywhere in people’s daily life due to their low cost,

ease of processing and versatility. There are different kinds of polymeric materials used in various

applications according to their different physical and mechanical properties. Many household items

around us are made of synthetic polymers, such as flexible polyurethane foams (FPUFs). FPUF is a

material widely used in consumer goods with excellent cushioning effect such as mattresses, upholstered

furniture, car seats and packaging. [1] In a fire situation, the high thermal inertia of flexible polyurethane

foam (FPUF) because of its chemical composition and physical structure in nature results in rapid

heating of the first layer, which reduces ignition time and accelerates flame spread, limiting the

application of FPUF. In addition, burning FPUF releases large amounts of toxic gases, such as carbon

monoxide (CO) and hydrogen cyanide (HCN), which can cause human death. [2-3] Therefore, the goal

of this research was set to enhance the flame retardancy and to reduce the smoke toxicity in flexible

polyurethane foams (FPUFs) by designing, constructing, and modifying the polyols and flame retardants

based on the theory and technology of synergism, catalytic carbonization, and catalytic conversion. This

scientific work is summarized in papers that present different approaches on the flame retardancy

enhancement of FPUF to enable a wide range of applications and to reduce potential fire risk to people.

This scientific work is summarized in 7 papers as listed below.

1. Shicong Ma, Yuling Xiao, Feng Zhou, Bernhard Schartel, Yin Yam Chan, Oleg P.

Korobeinichev, Stanislav A.Trubachev, Weizhao Hu, Chao Ma and Yuan Hu. Effects of novel

phosphorus-nitrogen-containing DOPO derivative salts on mechanical properties, thermal

stability and flame retardancy of flexible polyurethane foam. Polym Degrad Stabil. 2020, 177.

Doi: 10.1016/j.polymdegradstab.2020.109160.

2. Feng Zhou, Chao Ma, Kang Zhang, Yin Yam Chan, Yuling Xiao, Bernhard Schartel, Manfred

Doring, Bibo Wang, Weizhao Hu and Yuan Hu. Synthesis of ethyl

(Diethoxymethyl)phosphinate derivatives and their flame retardancy in flexible polyurethane

foam: Structure-flame retardancy relationships. Polym Degrad Stabil. 2021, 188. Doi:

10.1016/j.polymdegradstab.2021.109557.

3. Yin Yam Chan, Chao Ma, Feng Zhou, Yuan Hu and Bernhard Schartel. A liquid phosphorous

flame retardant combined with expandable graphite or melamine in flexible polyurethane foams.

Polym Advan Technol. 2022, 33, 326-339. Doi: 10.1002/pat.5519.

4. Yin Yam Chan, Chao Ma, Feng Zhou, Yuan Hu and Bernhard Schartel. Flame retardant flexible

polyurethane foams based on phosphorous soybean-oil polyol and expandable graphite. Polym

Degrad Stabil. 2021, 191. Doi: 10.1016/j.polymdegradstab.2021.109656.

5. Yin Yam Chan, Andreas Korwitz, Doris Pospiech and Bernhard Schartel. Flame retardant

combinations with expandable graphite/phosphorus/CuO/castor oil in flexible polyurethane

foams (This article was submitted to ACS Applied Polymer Materials on 11.11.2022.)

6. Yin Yam Chan and Bernhard Schartel. It takes two to tango: Industrial benchmark PU-foams

with expandable graphite/P-flame retardant combinations (This article was accepted by

Kautschuk Gummi Kunststoffe on 23.10.2022 and is being published.)

7. Yin Yam Chan and Bernhard Schartel. It takes two to tango: Synergistic expandable graphite –

phosphorus flame retardant combinations in polyurethane foams. Polymers. 2022, 14, 2562.

Doi:10.3390/polym14132562.

The simplest way to enhance the flame retardancy of materials is to physically mix the flame retardant

additives into the polymer matrix. In the past, the addition of halogenated flame retardants into polymers

was the most common approach to improve the flame retardancy. Due to the toxicity of halogenated

flame retardants, they have been replaced by halogen-free flame retardants, such as phosphorus-

containing flame retardant (P-FR). [4] Paper 1 and Paper 2 are studies related to the flame retardancy of

the additive novel P-FR in FPUF. A novel 9,10-dihydro-9-oxa-10-phosphaphenanthrene-10-oxide

(DOPO) derivative salts and a novel liquid phosphorous flame retardant,

bis([dimethoxyphosphoryl]methyl) phenyl phosphate (BDMPP), are synthesized by the cooperation

partner in Paper 1 and Paper 2, respectively. They were simply used as single additive flame retardants

to improve the fire performance of FPUF in both gas phase and the condensed phase. The investigation

of the fire behavior, in particular the cone calorimeter analysis, was my work contribution in Paper 1

and Paper 2.

The flame retardancy of P-FR is limited when its concentration is further increased. [5] Therefore,

system combining P-FR with other flame retardants is an effective approach. In Paper 3, BDMPP was

combined with expandable graphite (EG) or melamine (MA) in FPUF. BDMPP is effective in both gas

phase and condensed phase and cooperated well with EG and MA in the polymer matrix to enhance the

flame retardancy of FPUF. The combination of P-FR and MA created a layered residue which provides

thermal protection to the material underneath during burning. The synergistic effect between P-FRs and

EG was clearly recognizable from the results of fire tests.

Due to the growing focus on sustainability issues, more and more industries have paid more attention

and started to use renewable resources instead of petroleum-derived raw materials for over a decade. [6]

A renewable resource is a supply of a substance that can be replenished. Since the raw materials used in

conventional FPUF are derived from petroleum, the synthesis of reactive flame-retarded bio-based

polyols is a solution that kills two birds with one stone to produce FPUF for a wide range of applications.

[7] On the one hand, the renewable biomass content is increased. On the other hand, the flame retardancy

is improved. Paper 4 presents a good example in terms of the need for sustainability and flame retardancy

in FPUFs. Modified vegetable oils, such as palm oil, sunflower oil, linseed oil, rapeseed oil, can be used

as substitute for polyol production due to their certain similarities in chemical structure. [8] In Paper 4,

soybean oil was successfully modified into reactive phosphorus-containing polyol. In the formulation

of FPUF, the petrochemical polyether polyol was partially replaced by the phosphorus-containing

soybean oil-based polyol up to 80 %. The foam samples with modified soybean oil-based polyol still

show excellent physical and mechanical properties. Furthermore, the combination of phosphorus-

containing bio-based polyol and EG in FPUF provided excellent synergistic effect in charring, resulting

in improved flame retardancy.

The objective of Paper 5 was to find out the synergism among different combinations of commercial

flame retardants and additives (EG, P-FR, copper (II) oxide) in terms of flame retardancy and smoke

behavior in FPUF. The results showed that the combination of EG and F-FR greatly improved the char

yield, thereby providing better flame retardancy for FPUF. Regarding sustainability, the petrochemical

polyol was partially replaced by castor oil in the formulation of FPUF to enhance the bio-based content.

Paper 5 provides a multi-component strategy in FPUF that not only improves the flame retardancy but

also simultaneously reduce the smoke and toxic gas emission.

Paper 6 and Paper 7 present the state-of-the-art of the flame retardancy of PUF. The most prominent

highlight among the papers is the synergistic effect between P-FRs and EG in FPUFs. The binding effect

by P-FR strengthens the structure of expanded graphite residue, thereby providing impressive protection

to the material underneath during burning. According to the results of the work from Paper 3 to Paper

5, combination of P-FR and EG in FPUF is an effective way to significantly increase flame retardancy

and reduce amount of smoke released. Three industrial benchmark P-FR/ EG polyurethane foam

products used in different applications (Railways/ships, buildings, and lightweight structure) were

investigated in Paper 6, showing remarkable results in terms of flame retardancy. In addition, many

published research studies in recent years have found a synergistic effect between P-FR and EG. [9-13]

However, comprehensive studies on the interaction between P-FR and EG in PUF during burning are

insufficient. Therefore, the feature article (Paper 7) reveals in detail the burning process of the

synergistic effect between P-FRs and EG in PUFs, providing a better understanding for the successful

development of flame-retardant PUF. To gain a deeper understanding and modify the flame retardancy

of PUFs according to their burning behavior, the temperature-thickness relationship of PUFs at different

burning stages is illustrated. Moreover, current and future topics related to the use of renewable

feedstocks to increase the bio-based content in PUF and isocyanate-free approaches are discussed in the

article to provide potentially sustainable ways for future PUF production. [14, 15]

2 Scientific background – Fire behavior and flame retardancy

Various synthetic polymers are generally high molecular weight hydrocarbon, and most of them are

highly flammable, raising public concern about potential fire risks. [16-18] Among different fire safety

measures, actively enhancing the flame retardancy of polymeric materials is one of the approaches to

deal with the fire risk problem. Accordingly, many studies in recent decades have focused on improving

the flame retardancy of polymers through flame retardants to reduce the probability of fire accidents.

Fuel, oxygen, and heat are indispensable for a continuous combustion. Flame retardant is a substance

inhibits or disrupts the combustion cycle. In the past, halogen-based flame retardants were commonly

used as additives because of their excellent performance in improving the fire retardancy of polymer

even at low concentrations. Nonetheless, the use of halogen-based flame retardants in consumer goods

was already banned in some countries due to their toxicity, persistence, and bioaccumulation. Inhalation

of halogens, especially bromine, has adverse effects on environment and human health. Therefore, more

halogen-free flame retardants have been developed and used in polymers. Furthermore, scientists found

that two or more flame retardants used in a polymeric system may induce synergistic effect to enhance

the flame retardancy. In order to understand the fire behavior to better design promising flame-retardant

systems, the fire development of polymeric materials during burning is discussed in the next paragraphs.

Figure 1 depicts the typical four stages of a fire scenario. They are ignition, developing fire, fully

developed fire and decay. [19] Once the polymer is ignited, the heat from the ignition source on the

polymeric materials causes the rising of temperature. When the temperature is high enough, the chemical

bonds keep breaking and volatile fragments keep generating to a certain concentration. If the products

are flammable as burning fuels and at the same time a sufficient concentration is reached, a flame forms

and may initially grow by flame spread to a continuation of flaming combustion in the stage of

developing fire. The temperature gets higher as more and more material melts and volatilizes to fuel the

fire. In the stage of fully developed fire, which is the penultimate stage of the fire growth, the flame may

stabilize to become a steady state. When the fuel is consumed so that the flame cannot sustain, the stage

of decay occurs and reduces the overall temperature.

Figure 1 Stages in a fire scenario

Due to incomplete combustion of polymeric materials, not only carbon dioxide and water, but also CO

and smoke are the main products produced.

To improve fire safety, flame retardants are commonly used in polymers, textiles, and coatings to inhibit

or prevent the development of the propagation of fire. However, since compatibility and flame-retardant

interactions vary from one polymer to another, there is no all-purpose flame retardants suitable for all

polymeric materials. For FPUF, there are some commercial flame retardants such as metal hydroxides,

P-FRs, MA, and intumescent products commonly used. These flame retardants act differently under

various modes of action, and they are generally categorized into three approaches: Condensed phase,

endothermic and gas phase modes of action. [20-21]

Condensed phase mode of action:

Non-flammable char is produced in the polymer through the dehydration of the flame retardant to

generate double bonds during the pyrolysis process. Carbonaceous char layer is formed by cyclization,

cross-linking, aromatization, and graphitization to act as a physical hinderance preventing the heat and

mass transfer from the gas phase to the unburned material underneath. [22] In addition, the intumescent

effect of char is triggered in the presence of three basic ingredients: acid source, charring agent and

blowing agent. The gaps trapped in the intumescent char provide superior flame retardancy to the system.

Endothermic mode of action:

Heat is absorbed by endothermic decomposition of metal hydroxides to prevent reaching the pyrolysis

temperature of the material. [21] Aluminum trihydrate (ATH) and magnesium hydroxide (MHD) are

common mineral-based flame retardants that undergoing endothermic reaction under high temperature.

During the endothermic decomposition of these mineral fillers, heat is absorbed, and non-flammable

water vapor is generated to further quench and slow down the pyrolysis process of the material. The

released water vapor also effectively dilutes surrounding combustible gases such as oxygen in the flame

zone. The endothermic decomposition reactions of ATH and MHD are described in equation (1) and

equation (2), respectively. [23-25]

2Al(OH)3 → Al2O3 + 3H2O (1)

Mg(OH)2 → MgO +H2O (2)

Gas phase mode of action:

Gas phase flame retardants reduce the heat released in the gas phase by scavenging reactive free radicals,

such as H‧ and OH‧, during combustion. [26] Halogen-based flame retardants provide the most effective

flame retardancy to various polymers in the gas phase. Halogen can be readily released under fire

condition to inhibit free radical reactions, which are exothermic to maintain the flame. Bromine is taken

as an example:

H‧ + Br‧ → HBr

H‧ + HBr → H2 + Br‧

HBr +OH‧ → H2O +Br

Nonetheless, halogen-based flame retardants release toxic and corrosive gases, especially hydrogen

halides, and large amount of smoke during burning. Therefore, less toxic alternatives such as phosphorus

compounds are used as flame retardants in polymers. [4] Phosphorus-containing compounds is a

halogen-free flame retardant which is a less toxic alternative. Phosphorus derivatives act not only a char

promoter via dehydration to induce cyclization, crosslinking, aromatization/ graphitization in the

condensed phase, but also in the gas phase, depending on their chemical structures and their interactions

with the polymer. [27] Radical scavengers such as PO‧, HPO‧ and PO2‧ combined with free radicals in

the gas phase to reduce the heat released during burning. However, polymeric materials with P-FRs

increase the release of smoke and CO during burning.

H‧ + PO‧ → HPO

H‧ + HPO → H2 + PO‧

A lot of above-mentioned flame retardants are additives. Since these additives are generally required in

high amounts to achieve satisfactory flame retardancy, they can adversely affect the final mechanical

properties of the polymer in practical applications. To solve this problem, reactive flame retardants have

been developed and applied to polymers. Reactive flame retardants can be chemically joined to the

polymer backbone to reduce the impact on the mechanical properties of the material. Reactive flame

retardants with specific functional groups should be tailored for certain types of polymers.

2.1 Flame retardancy of polyurethane foams

Polyurethane is a large molecule composes of multiple urethane groups in the molecular backbone.

Figure 2 shows the typical synthetic route of polyurethane by the hydroxyl groups from polyol and

isocyanate group to form the repeating urethane linkage via polyaddition reaction. Polyurethane can be

linear, branched, or crosslinked depending on the chemical structure of polyols and isocyanates. [28]

Using different ratios of different types of polyols and isocyanates and different kinds of additive,

polyurethanes can be produced for different applications ranging from rigid and flexible foams, coating,

adhesives, sealants to elastomers. [29]

Figure 2 Typical synthetic route of polyurethane

2.1.1 Factors affect the flammability of polyurethane foams

The chemical composition of polyurethane, shown in Figure 2, is mainly hydrocarbons, and its cellular

structure, low density and large surface area make PUF vulnerable to fire. In a real fire scenario, burning

droplets of PUF can ignite surrounding objects to increase the spread of the fire. Due to incomplete

combustion, the entire material is not consumed. Tiny and lightweight unburned particles called smoke

are emitted in the air. Apart from smoke, toxic gases such as HCN and CO released by PUFs are the

leading cause of death due to the inhalation in fires. [3]

Among a wide range of applications, PUF occupies the largest market share of polyurethane products.

Polyurethane foam is generally divided into two categories: rigid polyurethane foam (RPUF) and

flexible polyurethane foam (FPUF). RPUF has extremely low thermal conductivity due to its closed-

cell structure serving as a perfect thermal insulation material for buildings and refrigeration. Due to the

flexibility of FPUF by its open-cell structure, it is commonly used as a cushioning material for furniture,

car seats and packaging. [6]

RPUF has better flame retardancy than FPUF because RPUF has higher apparent density and crosslink

density. More char is produced in RPUF during burning to protect the material underneath. [30] On the

other hand, FPUF collapses at the start of burning and then turns into a pool fire. [31] After the flame

was extinguished, there were only fragile tiny fragments reminded in Figure 3.

Figure 3 SEM image of residue from FPUF.

The flammability of PUF is largely determined by its main chemical components - polyol and

diisocyanate. [32] The chemical structure of polyol have great influence on flammability. The

functionality of polyol directly affects crosslinking density of PUF. Higher functionality of polyol means

that a greater number of hydroxyl group reacts with isocyanate group to form more urethane linkage.

More urethane linkages in PUF result in higher crosslink density and require more energy to break the

crosslinks. Therefore, the flammability of PUF is reduced with higher crosslink density. The flame

retardancy of PUFs with aromatic based polyols is significantly higher compared to PUFs based on

aliphatic polyols. It is because aromatic hydrocarbons have higher hydrolytic and thermal stability

compared to aliphatic hydrocarbons. Aromatic structures are more easily charred during burning. [33]

Isocyanate index is the equivalent ratio of isocyanate to alcohol in the formulation. If a significant excess

of isocyanate is present in the polymer matrix, trimerization of the excess isocyanate forms isocyanurate

rings, shown in Figure 4. Isocyanate rings promote charring during burning to improve flame retardancy

by providing protective layers for unburned parts.

Figure 4 Chemical structure of isocyanurate.

2.2 Synergistic effect of flame retardants

Most of the additive flame retardant may influence adversely the mechanical properties of polymers. In

order to achieve satisfactory fire performance, high loading of flame retardants, especially metal

hydroxides, is usually necessary to be added to the polyurethane foam. [34] When the combination of

two or more flame retardants in a polymer matrix performs better overall flame retardancy than the sum

of the effects of individual flame retardants, it is called synergistic effect. [35] Due to the synergistic

effect of combination of flame retardants, less flame retardants can be added to reduce the adverse

impact on mechanical properties of the material. Systems such as Br-Sb2O3, Br-NH3 and P-N are

advantageous combinations of synergistic effect in polymers. [36-37] The non-halogenated combination

P-FR/EG is a perfect match to trigger strong synergistic effect in FPUF. P-FRs decomposed into glassy

polyphosphate that binds the fluffy expanded graphite together to form a reinforced thermal protective

char layer for the unburned material underneath. EG alone significantly reduced peak heat release rate

(PHRR) in FPUF, but PHRR is even lower in the combination of P-FR and EG. The char residue yield

after burning greatly increased. The increase content of phosphorus at the same amount of EG enhanced

the weight percentage of char residue, but greater amount of phosphorus increased the total smoke

release (TSR).

2.3 Materials

i. Polyols and isocyanates

Polyols and isocyanates are the main chemical components for the polymerization of FPUF and RPUF.

Polyol shown in Figure 5 is an organic compound containing multiple OH functional groups to build

the soft segment of PUF. Isocyanate is a major component commonly used to build the hard segments

of PUF. However, the type of polyol and isocyanates used in FPUF and RPUF are different due to their

physical and mechanical properties. For example, the isocyanate used in FPUF is usually an 80:20

mixture of 2,4-toluenediisocyanate (2,4-TDI) and 2,6-toluenediisocyanate (2,6-TDI) as displayed in

Figure 6. Apart from polyols and isocyanates, some additives also need to be added for successful

foaming, such as blowing agents, surfactants, and catalysts. Since FPUF is the main material of this

study, the composition of FPUF is detailly described as follows. The mechanical properties of FPUF are

highly influenced by the types of polyols and isocyanates. The functionality of polyol used for FPUF is

about 2-3 because FPUF has a lower crosslink density. Long chain polyols contribute to the flexibility

and elasticity of the foam. Isocyanate forms the hard segments in the polymer network to provide

strength and rigidity to the foam. [38] The open-cell structure shown in Figure 7 containing cell windows,

struts, and strut joins provides a cushioning function in nature.

Figure 5 Chemical structure of petrochemical polyol.

Figure 6 Chemical structure of 2,4-toluenediisocyanate and 2,6-toluenediisocyanate.

Figure 7 SEM image of cell structure of FPUF.

ii. Catalyst, foaming stabilizer and blowing agent

Besides polyol and diisocyanate, catalyst, foaming stabilizer and blowing agent are also important to

the foaming process. [39] Amine catalysts and organic tin complex catalysts are most widely used

catalyst in polyurethane foam production. Amine catalysts balance the reaction, help to reduce foam

defects, and improve the structural stability of the final foam. Organic tin complex catalysts facilitate

the gel reaction in the foaming process. Foaming stabilizer is used to stabilize the bubble formation to

control the cell size. It is used to maintain the cell structure by increasing the resilience of the cell wall.

Foaming stabilizers are organic/ inorganic silicone-based complexes. There are two types of blowing

agent: physical and chemical. Physical blowing agents form gas bubbles by introducing inert gas (e.g.

carbon dioxide, nitrogen) or low-boiling-point liquid (e.g. cyclopentane, hydrofluorocarbon,

dichloromethane) in the polymer matrix through gas release or evaporation of the liquid. Water is usually

used as a chemical blowing agent in polyurethane. As shown in Figure 8, it is a condensation reaction

of water and isocyanate to release carbon dioxide (CO2) gas as a blowing agent in FPUF.

Figure 8 Condensation reaction between isocyanate group and water

iii. Synthesis of flexible polyurethane foams

FPUF is typically prepared by one-pot method with two components (Component A and component B)

shown in Figure 9. Component A consists of polyol, amine catalyst, tin catalyst, foaming stabilizer and

blowing agent and component B is TDI. Component A is mixed homogenously by a high-speed

mechanical mixer. Component B (TDI) is then poured into the mixed component A and stirred

subsequently at high speed for 5 s. The mixture is transferred to a mold during the subsequent expansion.

Afterwards, the foam is placed in an oven at 80 °C for 24 h to complete curing.

Figure 9 Schematic diagram of the synthesis of FPUF

iii. Plant oils

Soybean oil and castor oil, which are shown in Figure 10, were used in this work. A plant oil molecule

composed of a triacylglycerol and linked to three fatty acids, which has structural similarities to

petrochemical polyol. (See Figure 5). [40] Therefore, the plant oil is an alternative source for

polyurethane production as it is a renewable resource. Soybean oil consists of unsaturated bonds on fatty

acids which can be modified into polyols by simple chemical reactions such as epoxidation and

hydrolysis. Castor oil has hydroxyl groups in nature. It can be replaced directly with the petrochemical

polyol in the formulation. Since castor oil is not suitable for human consumption, it does not compete

with edible oil, and thus it is more sustainable.

Figure 10 Chemical structure of soybean oil and castor oil.

iv. Flame retardants and smoke suppressants

Due to the flammability of FPUF, adding flame retardants is a common way to reduce the fire risk. For

this purpose, several flame retardants and smoke suppressants (EG, P-FRs, MA, and nano metal oxide)

were added to FPUF in this work.

Expandable graphite (EG)

Graphite is in layered structure in natural. As shown in Figure 11, Intercalant such as sulfuric acid is

added between the layers. Sulfuric acid decomposes into gases under elevated temperature. These gases

force to increase the distance between the graphite layers and thus expanding the graphite into several

hundred times of its original size. Instead of the decomposition of sulfuric acid, Camino et al. [41]

suggested that the expansion of EG is due to the reaction of carbon in the graphite with sulfuric acid to

form blowing gases, as shown in the chemical equation below. As shown in Figure 12, expanded

graphite has a porous and bulky worm-like structure which provides excellent thermal insulation for the

unburned part. Sufficient EG provides excellent flame retardancy, which greatly reduces the fire risk.

Moreover, EG is an effective smoke suppressant for FPUF by increasing the residence time of smoke

precursors in the pyrolysis zone into aromatic char. [42]

C + 2H2SO4 → CO2 + 2H2O +2SO2 [41]

Figure 11 Chemical structure of expandable graphite

Figure 12 SEM image of the surface of expanded graphite

Phosphorus-containing flame retardants (P-FR)

P-FRs generally act in the gas phase and condensed phase. In the gas phase, phosphorous compounds

decompose into free phosphorous radicals to quench the other free radicals generated to inhibit the

combustion process during burning of materials. In the condensed phase, the presence of phosphorus

compounds in a polymer matrix promotes carbonization to stabilize the char. [27] The

phosphocarbonaceous char is formed by linking polyaromatic macromolecules and phosphate groups to

act as a protective layer for the material underneath. [43] However, the increasing concentration of P-

FRs exhibit nonlinear behavior in terms of flame retardancy. [5, 44] Therefore, the flame retardancy of

P-FR is optimized in a certain amount of phosphorus content (See Figure 13). The effectiveness of P-

FRs in higher concentration is limited when used alone. It usually gives synergy in multicomponent

flame retardant systems, for instance P-EG which is mentioned in chapter 2.2.

Figure 13 Flame retardancy against phosphorus content

Melamine

Melamine (MA), an organic compound containing high content of nitrogen, acts as a heat sink to

increase the heat capacity of the system and inert diluent released in the flame during the decomposition

process. [45] MA effectively increases the time to ignition. Apart from enhancing the flame retardancy,

Price et al. [46] found that MA reduces the amount of aromatic smoke precursors thereby suppressing

smoke in PFUF. However, the increasing amount of MA increases the viscosity and thus decreases the

growth and rising height of FPUF during foaming. Therefore, the physical structure and the mechanical

properties is further degraded due to the embedding of higher concentration of MA in the FPUF matrix.

[47]

Nano metal oxides

Transition metal nanoparticles can catalyze the pyrolysis gas product such as smoke, HCN and CO at

high temperature. [48-50] Metal compounds such as copper, zinc, nickel, and iron have good adsorption

and catalytic conversion capabilities for HCN which can be converted into N2, CO2 and H2O etc.

2.4 Methods

Different measurements and tests were performed to characterize the materials from different

perspectives, such as morphology of the unburned and burned materials, mechanical properties,

decomposition behavior, flammability, dripping behavior, fire performance and smoke behavior.

2.4.1 Nuclear magnetic resonance spectroscopy (NMR)

Nuclear magnetic resonance (NMR) is a spectroscopy analysis to determine the molecular structure of

chemical substances based on the interaction of nuclear spins in a magnetic field. NMR was mainly used

to determine the hydrogen and phosphorus spectra of the synthesized polyol in Paper 4.

2.4.2 Thermogravimetric analysis (TGA)

Pyrolysis is the thermal decomposition process of materials at elevated temperature in an inert

atmosphere. During pyrolysis of organic materials, volatile products are released in the gas phase while

solid residue remains in the condensed phase as carbon char. 10 mg powdered samples in an alumina

crucible were heated up at a steady heating rate of 10 K min-1 under constant nitrogen flow (i.e. 30

mL/min). Figure 14 displays the curve of thermogravimetry (TG) and the first derivative of the TG curve

(DTG). The mass loss curve of FPUF generally consists of two decomposition steps. The first

decomposition step is attributed to the hard segment derived from the isocyanates. The second

decomposition step at higher temperature is related to the soft segment from polyols.

Figure 14 TG and DTG curves of FPUF

2.4.3 Thermogravimetry coupled with Fourier transform infrared spectroscopy

The combination of TGA and Fourier transform infrared spectroscopy (FTIR) allows simultaneous

analysis of gaseous reaction products produced during the pyrolysis of materials. It is a useful tool for

understanding the chemical reactions between flame retardants and polymers during thermal

decomposition that take place in the pyrolysis zone. FTIR measures the wavelength range in the infrared

region that a sample absorbs.

2.4.4 Flammability – limiting oxygen index (LOI)

Limiting oxygen index (LOI) is done to characterize the flammability according to ISO 4589-2. A test

specimen with the dimensions of 100 mm x 10 mm x 10 mm is burned vertically at the upper end in a

glass cylinder at room temperature under a constant flow of mixture of oxygen and nitrogen at specific

concentration. Varying the oxygen concentration until its specimen burns at the lowest oxygen

concentration. The lowest oxygen concentration represents the LOI value. FPUF is very flammable due

to its chemical composition and physical structure. The general LOI value for FPUF ranged from 16 to

19 vol.-%. [51]

2.4.5 Underwriters Laboratories 94 horizontal burning test (UL 94 HBF)

The test uses a small flame source to determine the burning behavior of FPUF specimens in the

horizontal direction according to ISO 9772. Specimen dimensions of 150 mm x 50 mm x 10 mm is used

for testing. Measurement determines the tendency of the material to extinguish or spread the flame once

the specimen is ignited. The dripping behavior is observed whether the cotton pad is ignited by burning

drops. [52, 53] The setup of UL94 HBF is shown in Figure 15.

Figure 15 Setup of UL 94 HBF test

2.4.6 Cone calorimeter

The burning behavior of the materials under forced flaming conditions were studied in accordance with

ISO 5660. Cone calorimeter, shown in Figure 16, is a tool used to study the fire performance of materials

in a fire scenario of developing fire under dynamic airflow condition. The specimen size for foam is 50

mm x 100 mm x 100 mm. The bottom and sides of foam specimens are wrapped with a single layer of

aluminum foil. To obtain meaningful data for comparison, a specimen placed horizontally on a sample

holder is exposed horizontally to a heat flux 25 kW m-2 with a spark ignitor in 25 mm spacing with a

cone heater. The load cell under the sample holder records the mass loss throughout the test. The

effluents released from the specimen are piped and analyzed for O2, CO2 and CO content. Moreover,

there are thermocouple, pressure sensor, smoke measurement and a sample probe to collect different

valuable data such as heat release rate (HRR), total heat release (THR), apparent effective heat of

combustion (EHC), total smoke release (TSR), carbon monoxide yield, char yield and maximum average

heat emission (MARHE) can be collected in the measurement. [19, 54, 55]

Figure 16 Burning sample on the load cell of cone calorimeter

2.4.7 Scanning electron microscopy (SEM)

The morphology of the foam before burning and its residue after burning was investigated by SEM.

SEM images samples in nanometer scale by emitting electron beam. There is an interaction between the

emitted electron beam and the sample, resulting in different signals depending on the surface topography.

These signals are received and processed to present the morphology of samples. The specimens are

sputter-coated with 15 nm of gold to increase the conductivity of the samples, resulting in higher

resolution images.

2.4.8 Mechanical test

The compression strength, tensile strength as well as elongation at break of FPUF were measured by a

Universal Testing Machine. The compression strength was determined by following ISO 3386-1. The

specimen size for compression is 40 mm x 30 mm x 10 mm. Tensile strength and elongation at break

were evaluated using type 1A specimens with a thickness of 10 mm according to ISO 1798.

2.4.9 Smoke density chamber (SDC) coupled with FTIR

SDC is a sealed test chamber equipped with a photometric instrument. [56] SDC measures the specific

optical density under conditions of static accumulation in accordance with ISO 5659. The specimen size

for foam samples is 75 mm x 75 mm x 25 mm in the SDC measurement. The foam specimen is exposed

to a heat flux of 25 kW m-2 in a separation of 25 mm to the cone heater, with or without the use of a

pilot flame depending on the fire scenario. The SDC is combined with FTIR to analyze qualitatively and

quantitively the gases evolved from materials during burning in such mentioned condition.

3 Discussion of the results

3.1 Main messages from the publications

Due to the high flammability of FPUF, several flame-retardant systems were designed and prepared to

improve the flame retardancy of FPUF for a wide range of applications. Five first-authorship and two

co-authorship articles were published detailing the studies on different flame-retardant FPUF systems.

Paper 1 and Paper 2 from the academic partner, University of Science and Technology of China (USTC),

investigated the morphology, mechanical properties, thermal stability, and flame retardancy of FPUF

with self-synthesized P-FRs. The novel flame retardants synthesized by our academic partner performs

well in both gas phase and condensed phase due to the presence of phosphorus.

Paper 3 summarizes a study on a liquid phosphorous flame retardant –

bis([dimethoxyphosphoryl]methyl) phenyl phosphate (BDMPP) combined with EG or MA incorporated

in FPUF. The results show that BDMPP worked in gas phase with reduced value of EHC. MA alone

already greatly reduced the peak heat release rate as it undergoes endothermic reaction to increase the

heat capacity of the system. EG is an effective smoke suppressant that significantly reduces the TSR

from 392 m2 m-2 to 52 m2 m-2. EG also reduced the MARHE from 320 kW m-2 to 109 kW m-2. The char

yield from the combination of BDMPP and EG in PFUF is greater than the sum of the individual effects.

The combination of BDMPP and MA in FPUF produces layered structure residue, which is an excellent

fire protection during burning. Therefore, the combination of BDMPP with EG or MA provided

synergistic effect on flame retardancy of FPUF.

In Paper 4, phosphorus-containing soybean-oil-based polyol was synthesized to enhance the flame

retardancy and bio-based content of FPUF. Apart from replacing petrochemical polyol with flame-

retardant bio-based counterpart, another flame retardant, EG was also added to the system. Notably, the

synergistic effect between the phosphorus-containing soybean-oil-based polyol and EG provides

excellent char yield during burning. The char yield percentage is one of the decisive factors regarding

the flame retardancy of a material. The increase in char residue provided a protective layer for the

underlying unburned FPUF, thereby significantly improve the flame retardancy. The main mechanism

behind is that the phosphorus-containing soybean-oil-based polyol acts as a gluing agent to bind and

strengthen the loose expanded graphite. Therefore, the structural integrity is maintained. Even if only

20 wt.% of polyether polyol was replaced by phosphorus-containing soybean-oil-based polyol and

additional 10 wt.% of EG in FPUF, the char residue was twice than that of the sample with only 10 wt.%

of EG. For the sample containing 80 wt.% of phosphorus-containing soybean-oil-based polyol and an

additional 10 wt.% of EG, the char residue was 3.6 times higher than the sample with only 10 wt.% of

EG. In addition to flame retardancy, EG also significantly reduced the smoke emission, as shown by the

result from cone calorimeter and smoke density chamber.

In Paper 5, several flame retardants and additives were used in FPUF. The results showed the

combination of P-FR and EG was the most effective system in flame retardancy among different flame

retardant combinations. The synergistic effect between P-FR and EG in FPUF created a superior thermal

barrier resulting in high char yield after burning in cone calorimeter measurement due to incomplete

pyrolysis. EG acted as a smoke suppressant and reduced the amount of toxic gases (CO and HCN)

released. The presence of castor oil (CAS) enhanced the bio-based content and maintain the physical

and mechanical properties of FPUF.

Paper 6 presents a scientific analysis of fire performance and smoke behavior on three industrial P-EG

PUFs. These materials show extremely low heat release rate and smoke emission which means less fire

hazard. Since the value of MARHE is less than 90 kW m-2 or even lower, they meet the requirement of

EN 45545. The high char yield of expanded graphite was used as a protective layer to shield the

underlying material from heat, resulting in insufficient pyrolysis temperature and eventually incomplete

burning. The combination of P-FR and EG was proved to be the state-of-the-art in PUF flame retardant

systems for various applications such as damping, construction and lightweight structure.

The main highlight from a series of scientific works is the synergistic effect between P-FR and EG in

FPUF. Therefore, a comprehensive study on the combination of EG and P-FRs in PUFs is presented in

Paper 7. The feature article first introduces the general fire hazard of PUFs, the difference between FPUF

and RPUF, the difference between polyisocyanurates foam and PUF, the flammability and smoke

toxicity during burning of PUFs, and the commercial flame retardants for PUFs. The burning process of

FPUF and RPUF is described in detail with texts and diagrams. FPUF and RPUF behave differently

during burning due to crosslink density, apparent density, and cellular structure. The char formed due

to the high crosslinking density of RPUF provides protection for the underlying material. However, due

to the low crosslink density of FPUF collapses easily into pool fire and burns fiercely with almost none

of the residue remained after burning. Therefore, the RPUF has better flame retardancy than FPUF.

Polyol and diisocyanate are the main components for polyurethane synthesis. The crosslink density

depends on the chemical properties and the amount of polyol and diisocyanate in PUFs. PUFs with EG

and P-FR were discussed individually in terms of burning behavior. The burning behavior of PUF with

the combination of EG and P-FR was discussed in depth. The combination of PUF with EG and

phosphorous compounds has such excellent flame retardancy because the phosphorous compounds turn

into glassy polyphosphates that act as binder to maintain the integrity of the carbonaceous char.

Moreover, the phosphorus functions as a bridge to bind the aromatic char into a larger molecule. As a

result, this enhanced carbonaceous char provides better protection for the material underneath and less

mass and heat is transferred. Furthermore, current and future topics regarding green solution for flame

retardants and PUFs were discussed. However, there are still some challenges to be overcome regarding

the recycling of flame-retardant PUF, as well as the potential health concerns of P-FRs.

4 Publications

4.1 Effects of novel phosphorus-nitrogen-containing DOPO derivative salts on mechanical

properties, thermal stability and flame retardancy of flexible polyurethane foam

Shicong Ma, Yuling Xiao, Feng Zhou, Bernhard Schartel, Yin Yam Chan, Oleg P. Korobeinichev,

Stanislav A. Trubachev, Weizhao Hu, Chao Ma and Yuan Hu. Polym Degrad Stabil. 2020, 177. Doi:

10.1016/j.polymdegradstab.2020.109160.

Status: This article was accepted and published.

https://doi.org/10.1016/j.polymdegradstab.2020.109160

Authors’ contributions:

• Shicong Ma

o Formal analysis

o Conceptualization

o Writing – Original draft

• Yuling Xiao

o Software

o Validation

• Feng Zhou

o Software

• Bernhard Schartel

o Methodology

• Yin Yam Chan

o Investigation

• Oleg P. Korobeinichev

o Data curation

• Stanislav A. Trubachev

o Supervision

• Weizhao Hu

o Resources

• Chao Ma

o Visualization

o Writing – review & editing

• Yuan Hu

o Supervision

Abstract

In this work, a series of 9,10-dihydro-9-oxa-10-phosphaphenanthrene-10-oxide (DOPO) derivative salts

containing phosphorus and nitrogen were synthesized, and their effects on mechanical properties,

thermal stability and flame retardancy of flexible polyurethane foam (FPUF) were investigated. Studies

have shown that the addition of DOPO derivatives will increase the tensile strength, compression set,

and compression hardness of FPUF, but it will lead to a decrease in elongation at break.

Thermogravimetric analysis showed that the initial decomposition temperature of FPUF containing

DOPO derivatives was reduced, but the char residue was significantly improved. A series of combustion

tests indicated that the addition of DOPO derivative salts can improve the flame retardancy of FPUF, of

which 10-hydroxy-9,10-dihydro-9-oxa-10-phosphaphenanthrene-10-oxide dicyandiamide salt (D-

DICY) exhibited the best flame retardancy. When the load of D-DICY was 20 phr, the limiting oxygen

index (LOI) of foam reached 24.5%, and the peak heat release rate and total heat release were decreased

by 55.7% and 52.9%, respectively. Furthermore, based on the analysis of the gas phase combustion

products and the char residue of the condensed phase, the possible flame retardant mechanism was

proposed.

Effects of novel phosphorus-nitrogen-containing DOPO derivative

salts on mechanical properties, thermal stability and ﬂame retardancy

of ﬂexible polyurethane foam

Shicong Ma

, Yuling Xiao

, Feng Zhou

, Bernhard Schartel

, Yin Yam Chan

Oleg P. Korobeinichev

, Stanislav A. Trubachev

, Weizhao Hu

, Chao Ma

Yuan Hu

State Key Laboratory of Fire Science, University of Science and Technology of China, 96 Jinzhai Road, Hefei, Anhui, 230026, PR China

Bundesanstalt für Materialforschung und -prüfung (BAM), Unter den Eichen 87, 12205, Berlin, Germany

Voevodsky Institute of Chemical Kinetics and Combustion, Russian Academy of Science, Siberian Branch, Novosibirsk, 630090, Russia

Novosibirsk State University, Novosibirsk, 630090, Russia

article info

Article history:

Received 8 February 2020

Received in revised form

28 March 2020

Accepted 29 March 2020

Available online 3 April 2020

Keywords:

Flexible polyurethane foam

DOPO derivative salts

Mechanical properties

Thermal stability

Flame retardancy

abstract

In this work, a series of 9,10-dihydro-9-oxa-10-phosphaphenanthrene-10-oxide (DOPO) derivative salts

containing phosphorus and nitrogen were synthesized, and their effects on mechanical properties,

thermal stability and ﬂame retardancy of ﬂexible polyurethane foam (FPUF) were investigated. Studies

have shown that the addition of DOPO derivatives will increase the tensile strength, compression set, and

compression hardness of FPUF, but it will lead to a decrease in elongation at break. Thermogravimetric

analysis showed that the initial decomposition temperature of FPUF containing DOPO derivatives was

reduecd, but the char reside was signiﬁcantly improved. A series of combustion tests indicated that the

addition of DOPO derivative salts can improve the ﬂame retardancy of FPUF, of which 10-hydroxy-9,10-

dihydro-9-oxa-10-phosphaphenanthrene-10-oxide dicyandiamide salt (D-DICY) exhibited the best ﬂame

retardancy. When the load of D-DICY was 20 phr, the limiting oxygen index (LOI) of foam reached 24.5%,

and the peak heat release rate and total heat release were decreased by 55.7% and 52.9%, respectively.

Furthermore, based on the analysis of the gas phase combustion products and the char residue of the

condensed phase, the possible ﬂame retardant mechanism was proposed.

1. Introduction

Flexible polyurethane foam (FPUF) is one of the most important

components of polyurethane materials, accounting for about 40% of

its market share [1,2]. Based on its low density, high resilience,

excellent gas permeability and low thermal conductivity, FPUF is

widely used in bedding, shoes, architectural decoration, textile in-

dustry and other ﬁelds [3]. However, FPUF is highly ﬂammable, and

its limiting oxygen index is only 16e18% [4]. The large surface area

and good permeability of foam will accelerate the spread of ﬁre

during combustion, which greatly limits its application [5]. There-

fore, it is important to ﬁnd or synthesize a suitable ﬂame retardant

to impart ﬂame retardancy to FPUF.

At present, the commonly used FPUF ﬂame retardants can be

divided into two types: additive and reactive ﬂame retardants. The

former mainly includes triethyl phosphate, dimethyl methyl

phosphate (DMMP), tris(2-chloroisopropyl) phosphate ester

(TCPP), trichloroethyl phosphate (TCEP), melamine, expandable

graphite, and aluminum hydroxide [6,7]. The latter mainly includes

ﬂame retardant polyols and ﬂame retardant isocyanates [8]. The

additive ﬂame retardants are introduced into the foam through

simple physical and mechanical mixing, which has little effect on

the foaming formula and exhibit good ﬂame retardancy under low

load. Therefore, they are widely used in FPUF [9]. At present, the

most widely used FPUF ﬂame retardants are halogen-based ﬂame

retardants, which have high thermal stability, low price, and have

little effect on the mechanical properties of foam [10]. However,

halogen-based ﬂame retardants have many disadvantages that

*Corresponding author.E-mail

** Corresponding author.

E-mail addresses: [email protected] (C. Ma), [email protected] (Y. Hu).

Contents lists available at ScienceDirect

Polymer Degradation and Stability

journal homepage: www.elsevier.com/locate/polydegstab

https://doi.org/10.1016/j.polymdegradstab.2020.109160

Polymer Degradation and Stability 177 (2020) 109160

cannot be ignored. First, they will produce a large amount of smoke

and corrosive hydrogen halide gas during combustion. Second,

halogen-based ﬂame retardants may generate extractable organic

halides (EOX), polybrominated dibenzodioxins (PBDD) and poly-

brominated dibenzofurans (PBDF). Among them, EOX is an envi-

ronmentally persistent organic compound, which is lipophilic,

hydrophobic, and difﬁcult to degrade. And PBDD and PBDF are easy

to accumulate in the human body and damage the human immune

and regeneration system [11]. Therefore, the use of halogen-free

ﬂame retardants has become the future development trend [12].

In the ﬁeld of halogen-free ﬂame retardants, compounds con-

taining single ﬂame retardant element are increasingly unable to

meet application requirements, hence, the development of ﬂame

retardants is increasingly focused on the synergy of multiple ﬂame

retardant elements. Among them, phosphorus-nitrogen synergistic

ﬂame retardants have become one of the main research directions

[13]. Generally, they have the advantages of heat insulation, low

toxicity, low smoke generation and low corrosion [14]. Liang et al.

synthesized a series of organic phosphorus compounds and studied

the ﬂame retardancy of different compounds on FPUF. The results

showed that the synthesized phosphoramidates exhibited higher

ﬂame retardancy than the corresponding phosphates [5]. Chen

et al.synthesized a series of phosphorus-containing melamine salts

and applied to FPUF. Among them, 2-carboxyethyl(phenyl)phos-

phinic acid melamine salt (CMA) showed the best ﬂame retardancy.

When its load was 20 phr, the LOI of the foam reached 23.5% [15].

Rao et al. prepared a melamine salt with high ﬂame retardancy

from diphenylphosphinic acid (DPPA) and melamine (MA). When

its addition amount was 20 phr, the LOI of FPUF increased to 24.5%

[16]. However, the 2-carboxyethyl (phenyl) phosphinic acid and the

diphenylphosphinic acid mentioned above have certain limitations

in practical applications, mainly due to the relatively expensive

price and less supply in industry. In comparison, as a commonly

used and industrialized phosphorus-containing ﬂame retardant,

DOPO has great advantages in practical applications [17]. There are

some researches on the use of DOPO and its derivatives to ﬂame

retard FPUF [18,19]. Przystas et al. evaluated the ﬂame retardancy of

several bridged DOPO on FPUF. When the addition amount of them

was 7.5%, the horizontal combustion test level of FPUF was HF-1. In

contrast, the horizontal combustion level of FPUF containing con-

ventional ﬂame retardants TCPP and TCEP was HF-2 [20]. Sabya-

sachi et al. investigated the ﬂame retardancy of novel DOPO

phosphoramidate to FPUF. When its load was 10 php, the horizontal

combustion test level of FPUF can reach HF-1 [21].

In this work, DOPO was ﬁrst oxidized to 10-hydroxy-9,10-

dihydro-9-oxa-10-phosphaphenanthrene-10-oxide (DOPO-OH),

then DOPO-OH was reacted with nitrogen-containing compounds

(melamine, dicyandiamide, and urea) through a hydrothermal re-

action to synthesize a series of phosphorus-nitrogen-containing

DOPO derivative salts and apply it to FPUF. Their effects on the

structure, mechanical properties, thermal stability, and ﬂame

retardancy of FPUF was evaluated. By analyzing the products in the

gas phase and the condensed phase, the ﬂame retardant mecha-

nism is explained.

2. Experimental

2.1. Materials

9,10-Dihydro-9-oxa-10-phosphaphenanthrene-10-oxide

(DOPO) was provided by Shandong Mingshan Fine Chemical In-

dustry Co. Ltd (Shandong, China). Melamine, dicyandiamide, urea

and dibutyltin dilaurate (DBTDL) were provided by Sinopharm

Chemical Reagent Co., Ltd (Shanghai, China). Ethanol and hydrogen

peroxide were supplied by Aladdin Holdings Group Ltd. Polyether

polyol 330, triethylenediamine (A33, 33%), silicone surfactant and

toluene diisocyanate (TDI 80/20) were supplied by Jiangsu Lvyuan

New Material Co., Ltd., China. Distilled water was obtained from our

laboratory.

2.2. Synthesis of DOPO derivatives

DOPO-OH was synthesized according to reference [22]. 108 g

(0.5 mol) of DOPO was added to a three-necked ﬂask, and then

100 mL of 30% hydrogen peroxide solution was slowly dropped into

the three-necked ﬂask at 60



C. After the addition was completed,

the temperature was raised to 80



C and the reaction was continued

for 8 h. The product was ﬁltered and washed 3 times with acetone,

and dried in a vacuum oven at 80



C for 24 h.

The synthesis method of DOPO derivative salts was similar.

Herein, the reaction between DOPO-OH and dicyandiamide was

taken as an example. 58 g (0.25 mol) of DOPO-OH and 300 mL of

mixed solvent of water and ethanol (1: 1 by volume) were added to

a three-necked ﬂask. After DOPO-OH was completely dissolved,

21 g (0.25 mol) of dicyandiamide was added in portions at room

temperature. Then the reaction was continued for 6 h. Finally, the

mixture was evaporated under reduced pressure and the obtained

product was dried in a vacuum oven at 70



C for 24 h. The obtained

DOPO derivative salts were all white solid powders. The yields of D-

Mel, D-DICY, and D-Urea were 92.7%, 85.1%, and 81.5%, respectively.

The abbreviations of the salts produced by the reaction of

DOPO-OH with melamine, dicyandiamide, and urea are D-Mel, D-

DICY, and D-Urea respectively. The chemical structure of the DOPO

derivative salts are shown in Fig. 1.

2.3. Preparation of FPUF

The pure and DOPO derivatives-containing FPUF were prepared

by the one-pot and free-rise method. Brieﬂy, polyol 330, distilled

water, catalyst (DABCO and DBTDL), silicone oil and ﬂame retardant

were added into a 1 L plastic cup and thoroughly mixed by me-

chanical stirring. TDI 80/20 was then added to the plastic cup with

vigorous agitation for 6s, and the mixture was immediately poured

into an open plastic mold to create a free foaming foam. The foam

was heated at 80



C for 24 h. The NCO/OH ratio is 1.05 and the

formulation of FPUF is shown in Table 1. The abbreviations of the

three DOPO derivatives-containing FPUF are FPUF/D-D, FPUF/D-M,

and FPUF/D-U.

2.4. Measurements

Fourier transform infrared spectra (FTIR) was measured on a

Nicolet 6700 spectrometer (Nicolet Instrument Company, U.S.) in a

wavenumber range of 4000 to 500 cm

1

using KBr disk method.

Nuclear magnetic resonance (NMR) spectra was recorded on a

Bruker AV400 NMR spectrometer (400 MHz) using deuterated

dimethyl sulfoxide (DMSO‑d

) as the solvent.

The density of FPUF was tested according to the standard of ISO

845:2006. The sample was left more than 72 h before testing and

the volume of sample was larger than 100 cm

Thermogravimetric analysis (TGA) was performed using a

Q5000 thermal analyzer (TA Co., U.S.). The test conditions were a

rise from room temperature to 800



C at an increase rate of 20



min under nitrogen atmosphere.

The limiting oxygen index (LOI) was tested in accordance with

ISO 4589-1:1996 and tested on an HC-2 oxygen index meter (Jiang

Ning Analytical Instrument Company, China). The size of the test

sample was 130 10 10 mm

Vertical combustion test was performed on a CFZ-2 instrument

(China Jiangning Analytical Instrument Co., Ltd.) according to

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 1091602

California Technical Bulletin 117 (TB 117e2000). The size of the test

sample was 12 30.5 inch

Cone calorimetry test was performed on a ﬁre-resistant test

technical equipment according to ISO 5660-1 standard under an

external heat ﬂux of 25 kW/m

. The sample size was

100 100 25 mm

Scanning electron microscope (SEM) images of the samples

were taken on a FEI Sirion 200 scanning electron microscope under

high vacuum conditions with an acceleration voltage of 10 kV.

Tensile properties was performed according to ISO 1798: 2008,

and the tensile speed was 500mm/min

1

. The width and thickness

of the sample were 10 mm, and the gauge length was 40 mm.

Compression set test was performed according to ISO 1856:

2000. The foam was compressed between two metal plates to 50%

of its original thickness and left at 70



C for 22 h. The size of the test

sample was 50 50 25 mm

. 25% of compression hardness was

measured in accordance with ISO 2439: 1997.

The Nicolet 6700 FTIR spectrometer was combined with a TGA

Q5000IR thermogravimetric analyzer to characterize the ther-

mogravimetric analyzer-Fourier transform infrared spectrum (TG-

IR). Under nitrogen atmosphere, the heating rate was 20



C/min

1

and the ﬂow rate was 30mL/min

1

Laser Raman spectroscopy (LRS) measurement was performed

on a SPEX-1403 laser Raman spectrometer (SPEX Co., United States)

at room temperature. Wave numbers range from 2000 to 500 cm

1

3. Results and discussion

3.1. Characterization of DOPO derivatives

The chemical structure of the synthesized DOPO derivatives was

characterized by FTIR and NMR. The FTIR spectrum of DOPO and its

derivatives are shown in Fig. 2 and the corresponding NMR spec-

trum is shown in Fig. 3. In the FTIR spectrum of DOPO (Fig. 2a), the

absorption peak at 1450-1650 cm

1

corresponds to the skeletal

vibration of the benzene ring, and the peak at 2385 cm

1

attributed to the stretching vibration of the P-H bond [23]. It can be

seen from the spectrum of DOPO-OH (Fig. 2a) that the characteristic

peak of the P-H bond disappears while the characteristic peak of P-

OH bond appears at 927 cm

1

, indicating that DOPO is oxidized to

DOPO-OH [24]. From the FTIR spectrum of D-Mel (Fig. 2b), the

absorption peak at 3389 cm

1

is attributed to the asymmetrical

stretching vibration of -NH

[25]. The stretching vibration of C¼Nin

triazine ring is observed at 1660 cm

1

. The characteristic peak at

893 cm

1

corresponds to the deformation vibration of triazine ring

[26]. The spectrum of D-DICY shows absorption peaks at 2209/

2160 cm

1

and 1638 cm

1

, which are related to the tensile vibration

of the C≡N and C¼N bonds [27]. The adsorption peaks of D-Urea at

3440 cm

1

and 3342 cm

1

are attributed to the antisymmetric and

symmetrical tensile vibration of -NH

group [28]. The combined

absorption peak at 1610 cm

1

is attributed to the tensile vibration

of C¼O bond and the bending vibration of -NH

group [29]. The

NMR (Fig. 3a) and

P NMR (Fig. 3b) spectrum of DOPO and DOPO-

OH are shown in Fig. 3. Chemical shifts from 7.36 to 8.26 ppm

represent the resonance of phenyl groups [30]. The disappearance

of hydrogen on the P-H bond (8.89 ppm and 6.50 ppm) and the

appearance of hydrogen on the -OH bond (11.42 ppm) indicate that

DOPO has been oxidized to DOPO-OH. In addition, the change in the

chemical shift of the single peak of phosphorus also conﬁrms that

DOPO has reacted [22]. The

H NMR (Fig. 4a and b and c) and

NMR spectra (Fig. 4d and e and f) of DOPO derivative salts are

displayed in Fig. 4. It can be seen that the chemical shift of hydrogen

on the P-OH bond does not appear in the spectrum of the DOPO

derivative salts, and the chemical shift of phosphorus has shifted.

The above results prove that P-OH has reacted with amine groups.

3.2. Morphology and mechanical properties of FPUF

In order to study the inﬂuence of DOPO derivative salts on the

structure of FPUF, the surface morphology of FPUF was character-

ized by SEM. It can be seen that the surface of pure FPUF is smooth

(Fig. 5e), while the agglomeration can be observed from the images

of DOPO derivatives-containing FPUF (Fig. 5f, g and h). From Fig. 5a,

b, c and d, the foam size of the FPUF modiﬁed by the DOPO deriv-

ative is almost the same as the pure FPUF, which indicates that the

structure of the FPUF is maintained after the addition of DOPO

derivative salts.

The tensile and compressive properties of pure FPUF and DOPO

derivative-added FPUF were tested to evaluate the effects of addi-

tives on the mechanical properties of the foam. The relevant data

are listed in Table 2 and the stress-strain curve of FPUF is displayed

in Fig. 6. Compared with pure FPUF, the tensile strength, 50%

Fig. 1. The chemical structure of D-Mel (a), D-DICY (b) and D-Urea (c).

Table 1

Formulations of pure FPUF and DOPO derivatives-containing FPUF.

Sample polyol 330 (php) Flame retardants (php) H

O (php) DABCO (php)\ DBTDL (php) SZ580 (php) TDI (80/20) (php)

Pure FPUF 100 0 3 0.6 0.10 0.5 41

FPUF/D-M-10 100 10 3 0.6 0.10 0.5 41

FPUF/D-M-20 100 20 3 0.7 0.12 0.6 41

FPUF/D-D-10 100 10 3 0.9 0.12 0.6 41

FPUF/D-D-20 100 20 3 1.2 0.15 0.7 41

FPUF/D-U-10 100 10 3 1.2 0.13 0.6 41

FPUF/D-U-20 100 20 3 1.5 0.16 0.7 41

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 109160 3

Fig. 2. FTIR spectrum of DOPO and DOPO-OH (a), D-Mel, D-DICY and D-Urea (b).

Fig. 3.

H (a) and

P (b) NMR spectrum of DOPO and DOPO-OH.

Fig. 4.

H and

P NMR spectrum of D-Mel (a and d), D-DICY (b and e) and D-Urea (c and f).

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 1091604

compression set and 25% compression hardness of FPUF modiﬁed

by DOPO derivative salts are improved, and the elongation at break

is reduced. The reasons for the above results can be summarized as

follows. DOPO derivative salts dispersed in FPUF can act as rein-

forcing particles to improve the strength of the matrix and hinder

the movement of the polyurethane molecular chain, thus

improving the tensile strength, compression set and compression

hardness of FPUF [31]. However, the DOPO derivative salts in the

polyurethane matrix will cause agglomeration due to the problem

of dispersibility, which will cause the foam to break more easily

during the stretching process. Therefore, the elongation at break of

the foam decreases [32]. The density of FPUF is also showed in

Table 3. It can be observed that the addition of DOPO derivatives

will increase the density of FPUF. This is because when the additives

are mixed into the FPUF, the quality of the foam is increased

without affecting the volume of the foam, thereby increasing the

density of the foam [21].

3.3. Thermal stability of the DOPO derivatives and FPUF

TGA and DTG curves of DOPO and its derivatives are shown in

Fig. 7a and the relevant characteristic data are listed in Table 3.It

can be seen that the initial decomposition temperature (T

) of the

DOPO derivative salts are all higher than 270



C, showing good

thermal stability. And the char residue is also signiﬁcantly

improved compared to DOPO. Because the thermal decomposition

temperature of nitrogen-containing substances in DOPO derivative

salts is low, their maximum decomposition rate temperature (T

max

)

is slightly lower than that of DOPO-OH [33]. The TGA and DTG

curves of pure FPUF and FPUF with DOPO derivatives are shown in

Fig. 7b and the corresponding data are listed in Table 4. All curves

present typical two-step decomposition process. The ﬁrst

maximum mass loss occurs at 220e280



C, which is due to the

degradation of hard segments. The second stage of maximum mass

loss appears at 360e420



C, which corresponds to the thermal

decomposition of soft segments [34]. The results show that the T

FPUF modiﬁed by DOPO derivative salts is slightly lower than that

of pure FPUF, which is due to the premature decomposition of the

additives [35]. Furthermore, the introduction of additives can in-

crease the char residue of FPUF. This is mainly because the phos-

phoric acid, polyphosphoric acid, etc. generated by DOPO derivative

salts will cover the surface of the FPUF [36]. These foam-covering

compounds reduce the rate of thermal decomposition of the car-

bon skeleton and suppress the release of volatile gases. In addition,

it can also provide additional covalent bonds to catalytic the cross-

linking of the skeleton to form a carbon layer network [37]. In

Fig. 5. SEM images of pure FPUF (a and e), FPUF/D-M-20 (b and f), FPUF/D-D-20 (c and g), FPUF/D-U-20 (d and h).

Table 2

Mechanical properties and density of pure FPUF and DOPO derivatives-containing FPUF.

Sample Tensile strength (kPa) Elongation break (%) 50% compression set (%) 25% compression hardness (N) Density (kg/m

)

Pure FPUF 110 ±8 300 ±13 6.0 ±0.19 95 ±332±1

FPUF/D-M-10 142 ±6 258 ±8 7.5 ±0.25 127 ±635±1

FPUF/D-M-20 148 ±8 267 ±10 7.9 ±0.31 136 ±439±2

FPUF/D-D-10 162 ±4 251 ±12 7.4 ±0.15 126 ±635±1

FPUF/D-D-20 168 ±5 260 ±9 8.0 ±0.20 141 ±538±1

FPUF/D-U-10 131 ±5 282 ±14 6.9 ±0.27 115 ±736±2

FPUF/D-U-20 140 ±8 295 ±13 7.3 ±0.08 120 ±639±1

Fig. 6. Stress-strain curves of pure FPUF and DOPO derivatives-containing FPUF.

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 109160 5

summary, the addition of DOPO derivatives has little effect on the

thermal decomposition temperature of FPUF and it can promote the

formation of char residue.

3.4. Flame retardancy of FPUF

The LOI of the samples was tested by limiting oxygen index

meter and the results are listed in Table 5. It can be seen that the LOI

of FPUF increases as the load of additives increase. Among them, D-

DICY has the best ﬂame retardancy. When the addition amount is

20phr, the LOI of FPUF reaches 24.5%. The vertical burning test is an

important indicator for evaluating the ﬂame retardancy of poly-

mers. In this work, vertical burning test was carried out in accor-

dance with the TB 117e2000 standard [38]. The digital photos of

the vertical burning test of FPUF are shown in Fig. 8 and the relevant

data are listed in Table 5. Pure FPUF burns quickly after ignition and

Table 3

The characteristic TGA data of DOPO and its derivatives.

Sample T

(



C) T

max

(



C) Char residue at T

max

(%) Char residue at 800



C (%)

DOPO 186 261 31 2

DOPO-OH 274 391 23 3

D-Mel 273 374 38 5

D-DICY 274 357 38 7

D-Urea 272 350 40 4

: Temperature at which the material decomposes to 5%.

max

: Temperature at maximum decomposition rate.

Fig. 7. TGA and DTG curves of DOPO and its derivatives (a and b), pure FPUF and DOPO derivatives-containing FPUF (c and d).

Table 4

The characteristic TGA data of pure FPUF and DOPO derivatives-containing FPUF.

Sample T

(



C) T

max

(



C) Char residue at T

max

(%) Char residue at 800



C (%)

Pure FPUF 238 365 24 1

FPUF/D-M-10 228 375 29 3

FPUF/D-M-20 228 380 31 6

FPUF/D-D-10 229 375 32 5

FPUF/D-D-20 228 378 36 9

FPUF/D-U-10 227 374 29 2

FPUF/D-U-20 226 376 33 5

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 1091606

the ﬁre spreads rapidly. At the same time, the dripping of melt is

very obvious. After the addition of the DOPO derivatives, the

combustion of FPUF is suppressed and no ﬂame spread, and the

dripping phenomenon is signiﬁcantly reduced. It is worth noting

that all of the DOPO derivatives-containing FPUF can pass the

vertical combustion test.

Cone calorimetry (CC) is a widely used method for investigating

the combustion behavior of materials. In this work, the combustion

performance of the samples was quantitatively analyzed by studying

parameters such as time to ignition (TTI), peak heat release rate

(pHRR), total heat release (THR), total smoke production (TSP) and

average effective heat of combustion (Av-EHC) [39]. The HRR and

THR curves of FPUF are shown in Fig. 9 and the corresponding data

are listed in Table 6. The ignition time of pure FPUF is only about 2s,

the combustion is rapid and a large amount of heat is released after

ignition. Its pHRR and THR values are 317.5 kW/m

and 24.2 MJ/m

respectively. With the addition of the DOPO derivatives, the TTI of

FPUF is extended to more than 6s and the heat release is signiﬁcantly

reduced. The above results indicate that D-DICY has the best ﬂame

retardancy under the same load. When the addition amount of D-

DICY was 20 phr, the pHRR and THR of FPUF decreased by 55.7% and

52.9%, respectively. However, the introduction of DOPO derivatives

will increase the TSP of FPUF. This is due to the increase in the

density of FPUF. On the other hand, this is also related to the

decomposition of DOPO derivatives, which will release gas-phase

products to play a ﬂame retardant effect. Furthermore, the Av-EHC

and the burn rate index (FIGRA) of FPUF have also been reduced.

Av-EHC reﬂectes the combustion degree of volatile gases, and a

decrease in Av-EHC value indicates that the additives have gas phase

ﬂame retardant effect [40]. FIGRA is a derived parameter equal to the

maximum value of HRR/time, reﬂecting the maximum combustion

effect of the material as it burns [41]. This value decreased by 64.3%

when 20 phr D-DICY was incorporated into FPUF. In summary, CC

results show that the addition of DOPO derivatives can improve the

ﬂame retardancy of FPUF.

3.5. Flame retardant mechanism

The gaseous products of pure FPUF and DOPO derivatives-

containing FPUF at the maximum decomposition rate were

Table 5

The vertical burning and LOI test results of pure FPUF and DOPO derivatives-

containing FPUF.

Sample Pass/No (TB117-2000) LOI (%)

Pure FPUF No 17.5

FPUF/D-M-10 Pass 23.5

FPUF/D-M-20 Pass 24.0

FPUF/D-D-10 Pass 24.0

FPUF/D-D-20 Pass 24.5

FPUF/D-U-10 Pass 23.0

FPUF/D-U-20 Pass 23.5

Fig. 8. Digital images of pure FPUF (a), FPUF/D-M-20 (b), FPUF/D-D-20 (c), FPUF/D-U-20 (d) during vertical burning.

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 109160 7

analyzed by TG-IR. The FTIR spectrum of the volatile pyrolysis

products of FPUF/D-M-20 at different temperature are shown in

Fig. 10a and d. At 220



C, the peak at 3387 cm

1

is attributed to the

asymmetrical stretching vibration of -NH

group. The absorption

peaks at 2356 and 2312 cm

1

are due to the stretching vibration of

. The peak at 1660 cm

1

is attributed to the stretching vibration

of C¼N bond [25,26]. When the temperature reached 280



C, no

obvious new absorption peaks appeared. When the temperature

Fig. 9. HRR (a) and THR (b) curves of pure FPUF and DOPO derivatives-containing FPUF.

Table 6

Cone calorimetry data of pure FPUF and DOPO derivatives-containing FPUF.

Sample TTI (s) pHRR (kW/m

) THR (MJ/m

) TSP (m

) Av-EHC (MJ/kg) FIGRA Char residue (%)

Pure FPUF 2 ±1 318 ±17 24.2 ±2.1 1.58 ±0.10 24 ±3 3.20 ±0.13 2 ±1

FPUF/D-M-10 6 ±1 171 ±10 13.8 ±1.5 4.57 ±0.14 11 ±1 1.36 ±0.08 6 ±1

FPUF/D-M-20 7 ±1 166 ±8 12.5 ±0.8 5.30 ±0.25 12 ±2 1.32 ±0.05 9 ±1

FPUF/D-D-10 7 ±1 159 ±13 12.1 ±1.1 4.31 ±0.18 10 ±2 1.29 ±0.03 7 ±1

FPUF/D-D-20 7 ±1 141 ±9 11.4 ±0.6 4.91 ±0.13 9 ±1 1.14 ±0.06 10 ±1

FPUF/D-U-10 6 ±1 265 ±14 21.5 ±1.9 4.80 ±0.14 18 ±3 2.10 ±0.04 4 ±1

FPUF/D-U-20 6 ±1 232 ±11 18.4 ±1.4 5.46 ±0.31 16 ±2 1.84 ±0.11 7 ±1

Fig. 10. FT-IR spectra at maximum weight loss rate of FPUF/D-M-20 (a and d), FPUF/D-D-20 (b and e), FPUF/D-U-20 (c and f).

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 1091608

increased to 380



C, the characteristic absorption peaks of -N-H

group (3120-3370 cm

1

) and benzene (2978 and 1370 cm

1

) can be

observed. The absorption peaks at 2180 and 2130 cm

1

are attrib-

uted to the stretching vibration of C≡N bond. In addition, the peak

at 1439 cm

1

corresponds to the stretching vibration of C-N bond.

The absorption peak at 946 cm

1

represents -NH

group. The

characteristic peaks at 1200 cm

1

and 1117 cm

1

are related to the

stretching vibration of P¼O and P-O bond [16]. As can be seen from

Fig. 10b and e, FPUF/D-D-20 has characteristic absorption peaks of

-NH group (3185e3438 cm

1

) and C≡N bond (2200 and 2177 cm

1

)

at 220



C[27]. When the temperature reached 380



C, the peak at

946 cm

1

represents the stretching vibration of -NH

group, and

the absorption peaks at 1200 and 1117 cm

1

are due to the

stretching vibration of P¼O and P-O bond. From Fig.10c and f, FPUF/

D-U-20 has a characteristic absorption peak of the -C¼N bond at

220



C. And when the temperature reached 380



C, the absorption

peaks of C≡N (2178/2107 cm

1

), -NH

(946 cm

1

), P¼O(1210cm

1

)

and P¼O(1117cm

1

) appeared. The above results indicate that

DOPO derivatives-containing FPUF decomposes with increasing

temperature, and the pyrolysis products include phosphorus and

nitrogen-containing compounds.

The SEM was used to analyze the char residue of FPUF. Because

pure FPUF has almost no char residue, its SEM image is not taken. The

images of char residue of DOPO derivatives-containing FPUF are dis-

played in Fig.11.FromFig.11a and b, a continuous carbon layer can be

observed. However, in Fig. 11c, there are holes and cracks in the sur-

face of carbon layer. In order to further study the characteristics of

char residue, Raman spectroscopy was performed on the char residue

[42]. The corresponding Raman spectra of FPUF are shown in Fig. 11.

The D band at 1360 cm

1

and the G band at 1600 cm

1

correspond to

carbon vibrations from disordered carbon-containing compounds

and graphite-containing compounds, respectively. The integrated

intensity ratio(I

) of DandG bands is usedto evaluatethedegree of

graphitization [43]. It can be seen that FPUF/D-M-20 and FPUF/D-D-

20 have lower I

than FPUF/D-U-20, indicating that D-Mel and D-

DICY have better efﬁciency in improving the graphitization of FPUF.

Based on the above analysis, the possible ﬂame retardant

mechanism was proposed. Taking FPUF/D-D-20 as an example, D-

DICY decomposes with increasing temperature, and it is broke

down into DOPO-OH and dicyandiamide. As the temperature in-

creases, DOPO-OH can produce gaseous phosphorus-containing

products such as P$and PO$, which can quench highly active

hydrogen and hydroxyl radicals in the ﬂame, thereby interrupting

the chain reaction of combustion. The non-ﬂammable NH

pro-

duced by dicyandiamide can dilute ﬂammable gases. On the other

hand, phosphoric acid, polyphosphoric acid, etc. generated by

DOPO-OH will cover the surface of the FPUF, which can reduce the

thermal decomposition rate of the carbon skeleton and catalyze it

to form a carbon layer network. Furthermore, D-DICY promotes the

formation of char residue and improves the graphitization of them.

The continuous and compact carbon layer can suppress the transfer

of heat and the overﬂow of gases.

4. Conclusions

In this work, novel phosphorus-nitrogen-containing DOPO de-

rivative salts were synthesized and applied to FPUF. The test results

show that the structure of FPUF modiﬁed by DOPO derivative salts

is maintained. The initial decomposition temperature of the foam is

slightly reduced, the maximum decomposition rate temperature is

increased, and the char formation performance is signiﬁcantly

improved. The density of foam increases and the tensile strength

and compression hardness are enhanced. The ﬂame retardancy of

FPUF is improved and D-DICY has the best ﬂame retardancy. When

its addition amount is 20 phr, the LOI of FPUF can reach 24.5%.

Based on the analysis of the gas phase and the condensed phase,

the ﬂame retardant mechanism was proposed. In the gas phase,

DOPO derivatives can produce phosphorus-containing products

such as P$and PO$, which can interrupt the combustion reaction by

capturing highly active hydrogen and hydroxyl radicals in the

ﬂame. Furthermore, the non-combustible NH

can dilute the

ﬂammable gases. In the condensed phase, DOPO derivatives can

Fig. 11. The surface SEM images and Raman spectra of the char residues of FPUF/D-M-20 (a and d), FPUF/D-D-20 (b and e), FPUF/D-U-20 (c and f).

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 109160 9

catalyze the formation of the carbon network and improve its

graphitization degree. The dense and continuous carbon layer

effectively prevents heat transfer and gas release during combus-

tion. The synthesis steps in this work are simple, low cost, envi-

ronmental friendly, and have broad prospects of practical

application.

Declaration of competing interest

The authors declare that they have no known competing

ﬁnancial interests or personal relationships that could have

appeared to inﬂuence the work reported in this paper.

Acknowledgment.

CRediT authorship contribution statement

Shicong Ma: Formal analysis, Conceptualization, Writing -

original draft. Yuling Xiao: Software, Validation. Feng Zhou: Soft-

ware. Bernhard Schartel: Methodology. Yin Yam Chan: Investi-

gation. Oleg P. Korobeinichev: Data curation. Stanislav A.

Trubachev: Supervision. Weizhao Hu: Resources. Chao Ma: Visu-

alization, Writing - review &editing. Yuan Hu: Supervision.

This work was ﬁnancially supported by the National Natural

Science Foundation of China (51761135113, 51874266, U1833113,

51911530127).

References

[1] M.A. Garrido, R. Font, Pyrolysis and combustion study of ﬂexible polyurethane

foam, J. Anal. Appl. Pyrol. 113 (2015) 202e215.

[2] Y. Hou, Z. Xu, Y. Yuan, L. Liu, S. Ma, W. Wang, Y. Hu, W. Hu, Z. Gui, Nanosized

bimetal-organic frameworks as robust coating for multi-functional ﬂexible

polyurethane foam: rapid oil-absorption and excellent ﬁre safety, Compos. Sci.

Technol. 177 (2019) 66e72.

[3] X. Wang, Y.-T. Pan, J.-T. Wan, D.-Y. Wang, An eco-friendly way to ﬁre retardant

ﬂexible polyurethane foam: layer-by-layer assembly of fully bio-based sub-

stances, RSC Adv. 4 (86) (2014) 46164e46169.

[4] M.-J. Chen, Z.-B. Shao, X.-L. Wang, L. Chen, Y.-Z. Wang, Halogen-free ﬂame-

retardant ﬂexible polyurethane foam with a novel nitrogenephosphorus

ﬂame retardant, Ind. Eng. Chem. Res. 51 (29) (2012) 9769e9776.

[5] S. Liang, M. Neisius, H. Mispreuve, R. Naescher, S. Gaan, Flame retardancy and

thermal decomposition of ﬂexible polyurethane foams: structural inﬂuence of

organophosphorus compounds, Polym. Degrad. Stabil. 97 (11) (2012)

2428e2440.

[6] C.Q. Wang, F.Y. Ge, J. Sun, Z.S. Cai, Effects of expandable graphite and dimethyl

methylphosphonate on mechanical, thermal, and ﬂame-retardant properties

of ﬂexible polyurethane foams, J. Appl. Polym. Sci. 130 (2) (2013) 916e926.

[7] F. Zhou, T. Zhang, B. Zou, W. Hu, B. Wang, J. Zhan, C. Ma, Y. Hu, Synthesis of a

novel liquid phosphorus-containing ﬂame retardant for ﬂexible polyurethane

foam: combustion behaviors and thermal properties, Polym. Degrad. Stabil.

171 (2020), 109029.

[8] S. Batool, R. Gill, C. Ma, G.C.S. Reddy, W. Guo, Y. Hu, Epoxy-based multilayers

for ﬂame resistant ﬂexible polyurethane foam (FPUF), J. Appl. Polym. Sci.

48890.

[9] W.-H. Rao, W. Liao, H. Wang, H.-B. Zhao, Y.-Z. Wang, Flame-retardant and

smoke-suppressant ﬂexible polyurethane foams based on reactive

phosphorus-containing polyol and expandable graphite, J. Hazard Mater. 360

(2018) 651e660.

[10] C.Q. Wang, H.N. Lv, J. Sun, Z.S. Cai, Flame retardant and thermal decomposi-

tion properties of ﬂexible polyurethane foams ﬁlled with several halogen-free

ﬂame retardants, Polym. Eng. Sci. 54 (11) (2014) 2497e2507.

[11] S.V. Levchik, E.D. Weil, A review of recent progress in phosphorus-based ﬂame

retardants, J. Fire Sci. 24 (5) (2006) 345e364.

[12] M.-J. Chen, C.-R. Chen, Y. Tan, J.-Q. Huang, X.-L. Wang, L. Chen, Y.-Z. Wang,

Inherently ﬂame-retardant ﬂexible polyurethane foam with low content of

phosphorus-containing cross-linking agent, Ind. Eng. Chem. Res. 53 (3) (2014)

1160e1171.

[13] X. Qian, L. Song, Y. Bihe, B. Yu, Y. Shi, Y. Hu, R.K.K. Yuen, Organic/inorganic

ﬂame retardants containing phosphorus, nitrogen and silicon: preparation

and their performance on the ﬂame retardancy of epoxy resins as a novel

intumescent ﬂame retardant system, Mater. Chem. Phys. 143 (3) (2014)

1243e1252.

[14] G. Sun Gaan, K. Hutches, M.H. Engelhard, Effect of nitrogen additives on ﬂame

retardant action of tributyl phosphate: phosphorus-nitrogen synergism,

Polym. Degrad. Stabil. 93 (1) (2008) 99e108.

[15] W.-H. Rao, Z.-Y. Hu, H.-X. Xu, Y.-J. Xu, M. Qi, W. Liao, S. Xu, Y.-Z. Wang, Flame-

retardant ﬂexible polyurethane foams with highly efﬁcient melamine salt,

Ind. Eng. Chem. Res. 56 (25) (2017) 7112e7119.

[16] M.-J. Chen, Y.-J. Xu, W.-H. Rao, J.-Q. Huang, X.-L. Wang, L. Chen, Y.-Z. Wang,

Inﬂuence of valence and structure of phosphorus-containing melamine salts

on the decomposition and ﬁre behaviors of ﬂexible polyurethane foams, Ind.

Eng. Chem. Res. 53 (21) (2014) 8773e8783.

[17] Q. Dong, K. Chen, X. Jin, S. Sun, Y. Tian, F. Wang, P. Liu, M. Yang, Investigation

of ﬂame retardant ﬂexible polyurethane foams containing DOPO immobilized

titanium dioxide nanoparticles, Polymer 11 (1) (2019) 75.

[18] S.S. Ray, M. Kuruma, Types of Flame Retardants Used for the Synthesis of

Flame-Retardant Polymers, Halogen-free Flame-Retardant Polymers,

Springer, 2020, pp. 15e45.

[19] J. Rao, M.P. Fern

andez-Ronco, M. Vong, S. Gaan, Enhanced ﬂame-retardancy

and controlled physical properties of ﬂexible polyurethane foams based on

a shear-responsive internal network, RSC Adv. 7 (69) (2017) 44013e44020.

[20] A. Przystas, M. Jovic, K. Salmeia, D. Rentsch, L. Ferry, H. Mispreuve, H. Perler,

S. Gaan, Some key factors inﬂuencing the ﬂame retardancy of EDA-DOPO

containing ﬂexible polyurethane foams, Polymer 10 (10) (2018) 1115.

[21] S. Gaan, S. Liang, H. Mispreuve, H. Perler, R. Naescher, M. Neisius, Flame

retardant ﬂexible polyurethane foams from novel DOPO-phosphonamidate

additives, Polym. Degrad. Stabil. 113 (2015) 180e188.

[22] Y. Lin, S. Jiang, Z. Gui, G. Li, X. Shi, G. Chen, X. Peng, Synthesis of a novel highly

effective ﬂame retardant containing multivalent phosphorus and its applica-

tion in unsaturated polyester resins, RSC Adv. 6 (89) (2016) 86632e86639.

[23] J. Wang, C. Ma, X. Mu, W. Cai, L. Liu, X. Zhou, W. Hu, Y. Hu, Construction of

multifunctional MoSe2 hybrid towards the simultaneous improvements in

ﬁre safety and mechanical property of polymer, J. Hazard Mater. 352 (2018)

36e46.

[24] L. Yu, L. Chen, L.-P. Dong, L.-J. Li, Y.-Z. Wang, Organiceinorganic hybrid ﬂame

retardant: preparation, characterization and application in EVA, RSC Adv. 4

(34) (2014) 17812e17821.

[25] Z. Zhou, X. Zhang, L. Xing, J. Liu, A. Kong, Y. Shan, Copper-assisted thermal

conversion of microporous covalent melamine-boroxine frameworks to hol-

low B, N-codoped carbon capsules as bifunctional metal-free electrode ma-

terials, Electrochim. Acta 298 (2019) 210e218.

[26] N.E. Mircescu, M. Oltean, V. Chis¸ , N. Leopold, FTIR, FT-Raman, SERS and DFT

study on melamine, Vib. Spectrosc. 62 (2012) 165e171.

[27] K.-X. Xu, M.-H. Guo, L.-Q. Ren, W. Huang, J.-J. Sun, Spectroscopic and elec-

trochemical properties of Cu-dicyandiamide complex, Sci. China Chem. 61 (3)

(2018) 360e367.

[28] B. Xue, K. Yang, X. Wang, Q. Chi, Y. Jiang, The role of potassium chlorate on

expansion of dickite layers and the preparation of a novel TiO 2 impregnated

dickite photocatalyst using expanded dickite as carrier, RSC Adv. 6 (12) (2016)

9803e9811.

[29] Y. Feng, J. Yang, F. Zhao, Synthesis and characterization of long-chain dialkyl

carbonate, Asian J. Chem. 26 (4) (2014) 1232.

[30] X. Chen, J. Ye, L. Yuan, G. Liang, A. Gu, Multi-functional ladderlike poly-

siloxane: synthesis, characterization and its high performance ﬂame retarding

bismaleimide resins with simultaneously improved thermal resistance,

dimensional stability and dielectric properties, J. Mater. Chem. A. 2 (20) (2014)

7491e7501.

[31] Y. Pan, H. Pan, B. Yuan, N. Hong, J. Zhan, B. Wang, L. Song, Y. Hu, Construction

of organiceinorganic hybrid nano-coatings containing

-zirconium phosphate

with high efﬁciency for reducing ﬁre hazards of ﬂexible polyurethane foam,

Mater. Chem. Phys. 163 (2015) 107e115.

[32] C.Q. Wang, F.Y. Ge, J. Sun, Z.S. Cai, Effects of expandable graphite and dimethyl

methylphosphonate on mechanical, thermal, and ﬂame-retardant properties

of ﬂexible polyurethane foams, J. Appl. Polym. Sci. 130 (2) (2013) 916e926.

[33] G. Yang, W.-H. Wu, Y.-H. Wang, Y.-H. Jiao, L.-Y. Lu, H.-Q. Qu, X.-Y. Qin, Syn-

thesis of a novel phosphazene-based ﬂame retardant with active amine

groups and its application in reducing the ﬁre hazard of Epoxy Resin, J. Hazard

Mater. 366 (2019) 78e87.

[34] Y. Pan, J. Zhan, H. Pan, W. Wang, G. Tang, L. Song, Y. Hu, Effect of fully biobased

coatings constructed via layer-by-layer assembly of chitosan and lignosulfo-

nate on the thermal, ﬂame retardant, and mechanical properties of ﬂexible

polyurethane foam, Acs. Sustain. Chem. Eng. 4 (3) (2016) 1431e1438.

[35] K.A. Salmeia, F. Flaig, D. Rentsch, S. Gaan, One-pot synthesis of P (O)-N con-

taining compounds using N-chlorosuccinimide and their inﬂuence in thermal

decomposition of PU foams, Polymer 10 (7) (2018) 740.

[36] Q. Li, P. Jiang, P. Wei, Thermal degradation behavior of poly (propylene) with a

novel silicon-containing intumescent ﬂame retardant, Macromol. Mater. Eng.

290 (9) (2005) 912e919.

[37] Y. Pan, J. Zhan, H. Pan, W. Wang, H. Ge, L. Song, Y. Hu, A novel and effective

method to fabricate ﬂame retardant and smoke suppressed ﬂexible poly-

urethane foam, RSC Adv. 5 (83) (2015) 67878e67885.

[38] S. Nazar

e, R.D. Davis, K. Butler, Assessment of factors affecting ﬁre perfor-

mance of mattresses: a review, J. Fire Sci. 1 (1) (2012) 2.

[39] Y. Xiao, Z. Jin, L. He, S. Ma, C. Wang, X. Mu, L. Song, Synthesis of a novel

graphene conjugated covalent organic framework nanohybrid for enhancing

the ﬂame retardancy and mechanical properties of epoxy resins through

synergistic effect, Compos. B Eng. 182 (2020), 107616.

[40] R. Jian, P. Wang, L. Xia, X. Zheng, Effect of a novel P/N/S-containing reactive

ﬂame retardant on curing behavior, thermal and ﬂame-retardant properties of

epoxy resin, J. Anal. Appl. Pyrol. 127 (2017) 360e368.

[41] X.-L. Fu, X. Wang, W. Xing, P. Zhang, L. Song, Y. Hu, Two-dimensional

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 10916010

cardanol-derived zirconium phosphate hybrid as ﬂame retardant and smoke

suppressant for epoxy resin, Polym. Degrad. Stabil. 151 (2018) 172e180.

[42] J. Wang, D. Zhang, Y. Zhang, W. Cai, C. Yao, Y. Hu, W. Hu, Construction of

multifunctional boron nitride nanosheet towards reducing toxic volatiles (CO

and HCN) generation and ﬁre hazard of thermoplastic polyurethane, J. Hazard

Mater. 362 (2019) 482e494.

[43] W. Cai, Y. Hu, Y. Pan, X. Zhou, F. Chu, L. Han, X. Mu, Z. Zhuang, X. Wang,

W. Xing, Self-assembly followed by radical polymerization of ionic liquid for

interfacial engineering of black phosphorus nanosheets: enhancing ﬂame

retardancy, toxic gas suppression and mechanical performance of poly-

urethane, J. Colloid Interface Sci. 561 (2020) 32e45.

S. Ma et al. / Polymer Degradation and Stability 177 (2020) 109160 11

4.2 Synthesis of ethyl (Diethoxymethyl)phosphinate derivatives and their flame retardancy in

flexible polyurethane foam: structure-flame retardancy relationships

Feng Zhou, Chao Ma, Kang Zhang, Yin Yam Chan, Yuling Xiao, Bernhard Schartel, Manfred Doring,

Bibo Wang, Weizhao Hu and Yuan Hu. Polym Degrad Stabil. 2021, 188. Doi:

10.1016/j.polymdegradstab.2021.109557.

Status: This article was accepted and published.

https://doi.org/10.1016/j.polymdegradstab.2021.109557

Authors’ contributions:

• Feng Zhou

o Investigation

o Formal analysis

o Writing – original draft

• Chao Ma

o Validation

o Writing – review & editing

• Kang Zhang

o Data curation

• Yin Yam Chan

o Visualization

• Yuling Xiao

o Software

• Bernhard Schartel

o Methodology

• Manfred Doring

o Conceptualization

• Bibo Wang

o Resources

• Weizhao Hu

o Project administration

• Yuan Hu

o Funding acquisition

Abstract

Three novel liquid ethyl (diethoxymethyl)phosphinate derivatives (EDPs) were synthesized and

incorporated into flexible polyurethane foams (FPUFs). The flame retardancy of FPUFs were evaluated

by limiting oxygen index (LOI), vertical burning and cone calorimetry tests, and the results indicated

the structure-flame retardancy relationship of EDPs. Among these EDPs, P-(diethoxymethyl)-N-

phenylphosphonamidate (EDPPA) exhibited the best flame retardant effect, methyl 3-

((diethoxymethyl)(ethoxy)phosphoryl)propanoate (EDPMA) the second, and ethyl phenyl (di-

ethoxymethyl)phosphonate (EDPPO) the worst. When the incorporation of EDPPA was 10 wt%, the

FPUFs could self-extinguish and pass the vertical burning test. Meanwhile, the LOI value of FPUF-PA

increased to 23.6% with 20 wt% loading of flame retardant. According to the investigation of volatiles

during the thermal degradation of FPUFs and the morphologies of char residues after cone test, we

inferred the possible flame retardant mechanism. The results indicated that EDPs could release

phosphorus-containing compounds in the gas phase, which would generate phosphorus-containing

radicals and play the role of radical scavenger. In the condensed phase, EDPs can promote the formation

of dense, intact and thermal stably char layer on the surface of FPUFs. Moreover, we found that the

structure influence on flame retardancy was attributed to the atoms linked to the central phosphorus.

Our results indicate that these EDPs are promising flame retardants in FPUFs that can be applied to

improve the flame retardancy of FPUFs in various practical applications.

Polymer Degradation and Stability 188 (2021) 109557

Contents lists available at ScienceDirect

Polymer Degradation and Stability

journal homepage: www.elsevier.com/locate/polymdegradstab

Synthesis of Ethyl (Diethoxymethyl)phosphinate Derivatives and Their

Flame Retardancy in Flexible Polyurethane Foam: Structure-ﬂame

Retardancy Relationships

Feng Zhou

a , 1

, Chao Ma

a , 1

, Kang Zhang

, Yin yam Chan

, Yuling Xiao

, Bernhard Schartel

Manfred Doring

, Bibo Wang

, Weizhao Hu

a , ∗, Yuan Hu

a , ∗

State Key Laboratory of Fire Science, University of Science and Technology of China, 96 Jinzhai Road, Hefei 230026, P. R. China

Bundesanstalt Mat Forsch & Prufung BAM, Unter Eichen 87, D-11205 Berlin, Germany

Fraunhofer Institute for Structural Durability and System Reliability LBF, Division Plastics, Schlossgartenstr. 6, 64289, Darmstadt, Germany

a r t i c l e i n f o

Article history:

Received 24 December 2020

Revised 5 March 2021

Accepted 12 March 2021

Available online 18 March 2021

Keywords:

Ethyl (diethoxymethyl)phosphinate

derivatives

Flame retardant

Flexible polyurethane foam

Structure-ﬂame retardancy relationship

a b s t r a c t

Three novel liquid ethyl (diethoxymethyl)phosphinate derivatives (EDPs) were synthesized and

incorporated into ﬂexible polyurethane foams (FPUFs). The ﬂame retardancy of FPUFs were

evaluated by limiting oxygen index (LOI), vertical burning and cone calorimetry tests, and

the results indicated the structure-ﬂame retardancy relationship of EDPs. Among these EDPs,

P-(diethoxymethyl)-N-phenylphosphonamidate (EDPPA) exhibited the best ﬂame retardant effect,

methyl 3-((diethoxymethyl)(ethoxy)phosphoryl)propanoate (EDPMA) the second, and ethyl phenyl (di-

ethoxymethyl)phosphonate (EDPPO) the worst. When the incorporation of EDPPA was 10 wt%, the FPUFs

could self-extinguish and pass the vertical burning test. Meanwhile, the LOI value of FPUF-PA increased

to 23.6% with 20 wt% loading of ﬂame retardant. According to the investigation of volatiles during the

thermal degradation of FPUFs and the morphologies of char residues after cone test, we inferred the pos-

sible ﬂame retardant mechanism. The results indicated that EDPs could release phosphorus-containing

compounds in the gas phase, which would generate phosphorus-containing radicals and play the role of

radical scavenger. In the condensed phase, EDPs can promote the formation of dense, intact and thermal

stably char layer on the surface of FPUFs. Moreover, we found that the structure inﬂuence on ﬂame retar-

dancy was attributed to the atoms linked to the central phosphorus. Our results indicate that these EDPs

are promising ﬂame retardants in FPUFs that can be applied to improve the ﬂame retardancy of FPUFs in

various practical applications.

1. Introduction

Flexible polyurethane foams (FPUFs) are polymeric materials

that are synthesized from polyols and isocyanates. They are used

wildly in household for their excellent resilience and cushioning

properties.[ 1 , 2 ] However, as polymeric materials with open-cell

structure, they are easily ignited and highly ﬂammable. In recent

years, ﬁre accidents frequently happen in the buildings and cause

large damages because of the ignition of the cushioning materi-

als in the furniture. [3–5] In order to improve the ﬂame retar-

dancy of FPUFs and eliminate the ﬁre safety threats, researchers

have developed many kinds of ﬂame retardants and ﬂame retard-

∗Corresponding authors. State Key Laboratory of Fire Science, University of Sci-

ence and Technology of China, 96 Jinzhai Road, Hefei 230026, P. R. China

E-mail addresses: hwz1[email protected] (W. Hu), [email protected] (Y. Hu).

1 These authors contributed equally to this work.

ing strategies for FPUFs. [6–11] Among the many types of ﬂame

retardants, organophosphorus compounds (like phosphonate, phos-

phate, phosphonamidate and phosphoramidate, etc) exhibit high

ﬂame retardant eﬃciency and good compatibility in the poly-

mer matrix. [12–18] Moreover, there are no releasing of halogen-

containing gases during the combustion, which means they are

lower toxic and more environmental friendly as compared with

halogenated ﬂame retardants. [19] Over the past decade, many

researches on this area have been reported. Rao et al. had syn-

thesized a phosphorus-containing melamine salt (DPMMA) and

found high ﬂame-retarding eﬃciency in FPUFs. In their research,

FPUFs that contained merely 5 php ﬂame retardant exhibited

self-extinguish in TB 117-20 0 0 vertical burning. [11] Zhou et al.

synthesized a novel liquid phosphonate (BDMPP) and compared

its ﬂame-retarding effect to the commercialized ﬂame retardant

dimethyl methylphosphonate (DMMP). Their investigation indi-

https://doi.org/10.1016/j.polymdegradstab.2021.109557

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

cated that FPUFs-BD samples showed higher limiting oxygen index

(LOI) values, vertical burning level and lower heat release at the

same ﬂame retardant loading because of the existence of ( + 3) va-

lence phosphorus. [20] Based on the previous investigations, the

ﬂame retardant effects of organophosphorus can been concluded.

Generally, the ﬂame retardant effects are divided into two types:

gas phase and condensed phase effects. The gas phase effect is at-

tributed to the released phosphorus-contained species which can

play the radical scavenging effects and thus promote the ﬂame in-

hibition. [21–23] In the condensed phase, the decomposition prod-

ucts of organophosphorus ﬂame retardants (like phosphoric and

polyphosphoric acid) are able to form protective char layers on the

surface of polymer matrix, which can hinder the transfer of heat,

oxygen and combustible gas, and thus protect the inner polymer.

[24–27]

In recent years, more attentions are concentrated on the inﬂu-

ence of phosphorus-containing compounds’ chemical structure on

the ﬂame retardancy of FPUFs. At the very beginning, Liang et al.

[12] synthesized a series of dimethyl phosphite derivatives with

systematic structure variations to investigate their structure-ﬂame

retardancy relationship on FPUFs. Their results of ﬂame tests re-

vealed that phosphonates and phosphoramidates performed better

than the corresponding phosphates in improving the ﬂame retar-

dancy of FPUFs. Moreover, ﬂame retardants that containing allyl

show better ﬂame retardancy than the corresponding propyl con-

taining ones. Compared with the other one, phosphonates, phos-

phoramidates and allyl terminated ﬂame retardants were more ef-

fective in generating phosphorus-containing radicals, thus exhib-

ited higher ﬂame retardancy in FPUFs. Basing on these results,

Matthias Neisius et al. [28] focused on the structure inﬂuence of

phosphoramidates on FPUFs’ ﬂame retardancy. As shown in the

ﬂame tests, dimethyl phosphoramidates presented higher ﬂame re-

tardancy than the phenyl ones, monoallyl derivatives were better

than the other one in the same class phosphoramidates. Moreover,

they had investigated the inﬂuence of the structure on the ﬂame

retardancy of 9,10-dihydro-9,10-oxa-10-phosphaphenanthrene-10-

oxide (DOPO) based phosphonamidate. [29] In the ﬁre tests,

bis DOPO phosphonamidate/foam formulations show the highest

ﬂame retardant level as compared to the other foam compos-

ites. This results was attributed to the better thermal stability of

bis DOPO phosphonamidate that it can prolong the effective gas

phase ﬂame retardant effect during the combustion. Chen et al.

[30] had synthesized four melamine salts with different phospho-

rus valence, investigate their inﬂuence on the decomposition and

ﬂame retardancy of FPUFs. Among these four melamine salts, CMA

and MHPA the had the + 1 phosphorus valence, performed bet-

ter ﬂame retardant effect than MPOA ( + 3) and MPyP ( + 5). Fur-

thermore, they found that with the increasing of the phospho-

rus valence, the char yield was increased while the phosphorus-

containing compounds decreased in the gas phase.

Easily obtained from the industrialized and recyclable hy-

pophosphorous acid (H

), ethyl (diethoxymethyl)phosphinate

(EDP) is a potential ﬂame retardant intermediate. Through the re-

actions of P–H, numerous organophosphorus compounds can be

synthesized. However, to the best of our knowledge, there are

no reports on the ﬂame retardancy of EDP and its derivatives

(EDPs). Therefore, it is essential to investigate their ﬂame retar-

dancy and the corresponding structure-ﬂame retardant proper-

ties relationships of EDPs. In this paper, a series of ethyl (di-

ethoxymethyl)phosphinate derivatives (EDPs) were synthesized ac-

cording to the Atherton-Todd or Michael addition reactions, and

used into FPUFs. The aim of our work was to improve the ﬂame

retardancy of FPUFs and investigate EDPs’ structure inﬂuence on

ﬂame retardancy. The ﬂame retardancy of FPUF-EDPs composites

were evaluated through LOI, vertical burning and cone calorime-

try tests. Thermogravimetric analysis/infrared spectrometry was

applied to investigate the volatiles of these FPUFs during the

thermal degradation. Scanning electron microscopy and Raman

spectroscopy were used to characterized the morphology of char

residues after cone calorimetry test. According to these investiga-

tions, we inferred the possible ﬂame retardant mechanisms. The

results indicated that the structure inﬂuence on ﬂame retardancy

was attributed to the atoms linked to the central phosphorus.

2. Experiment

2.1. Materials

Hypophosphorous acid aqueous solution (30.0-35.0%), triethyl

orthoformate (TEO), phenylamine, phenol, methyl acrylate, tri-

ethylamine (TEA), chloroform, dichloromethane, dibutyltin dilau-

rate (DBTDL) and tetrahydrofuran (THF) were purchased from

Sinopharm Chemical Reagent Co. Ltd., China. Triﬂuoroacetic acid

(TFA) was purchased from Shanghai Aladdin Biochemical Technol-

ogy Co. Ltd., China. Polyether polyol 330 (hydroxyl value = 56

mg KOH/g, number average molecular weight = 30 0 0 g/mol, av-

erage functionality = 3), triethylenediamine (A33, 33%), silicone

surfactant and toluene diisocyanate (TDI 80/20, 4:1 mixture of

2,4-toluene diisocyanate and 2,6-toluene diisocyanate) were kindly

provided by Jiangsu Lvyuan New Material Co. Ltd., China. Distilled

water was used as a chemical blowing agent and made by our lab-

oratory. Hypophosphorous acid aqueous solution was distilled un-

der vacuum at 30 °C to get the anhydrous hypophosphorous acid.

The other chemicals were used as received.

2.2. Synthesis of EDP

The synthesis of EDP had been reported by Cécile Fougère.

[31] At room temperature, 0.02 mol TFA and given amount of an-

hydrous hypophosphorous acid (0.1 mol, 6.6 g) and TEO (0.2 mol,

29.6 g) were added into a 500 mL three-necked round-bottom ﬂask

equipped with a magnetic stirrer under nitrogen atmosphere. After

3 h stirring, the reaction mixture was concentrated by rotary evap-

oration and dissolved into chloroform, then washed by saturated

aqueous NaHCO

and dried with magnesium sulfate. The solvent

was removed by rotary evaporation and the ﬁnal product was ob-

tained and puriﬁed via reduced pressure distillation. The synthetic

route was illustrated in Scheme 1 and nuclear magnetic resonance

(NMR), fourier transform infrared (FTIR) spectra of product were

shown in Fig. 1 .

Analytical Data for EDP: colorless, transparent liquid.

NMR (CDCl

, Fig. 1 a-1) δ(ppm): 1.27 (6H, CHOCH

), 1.39

(3H, POCH

), 3.71-3.85 (4H, CHO CH

), 4.23-4.24 (2H,

PO CH

), 4.70 (1H, CH OCH

), 6.90 (1H, P–H ).

P NMR

(CDCl

, Fig. 1 a-2) δ(ppm): 27.82. FTIR (KBr, Fig. 1 e): 2979, 2933,

2902 cm

−1 (–CH

–, –CH

), 2385 cm

−1 (P–H), 1230 cm

−1 (P = O),

1165, 1114 cm

−1 (C–O–C), 1062 cm

−1 (P–O–C), 773 cm

−1 (P–C).

2.2. Synthesis of EDPPA and EDPPO

EDPPA and EDPPO were synthesized through the Atherton–

Todd reaction between EDP and phenylamine or phenol. Brieﬂy,

EDP and carbon tetrachloride (1.1 equiv.) were dissolved into

dichloromethane in a 500 mL three-necked round-bottom ﬂask

equipped with a dripping funnel and magnetic stirrer. Subse-

quently, given amount of phenylamine or phenol and TEA (1.1

equiv.) were added dropwise into this ﬂask during 1 h. The reac-

tion temperature was controlled at 0-5 °C, and the reaction solution

was stirred at room temperature for 8 h. After the reaction was

completed, the solution was ﬁltrated and the ﬁltrate was washed

by hydrochloric acid solution (1 M), saturated aqueous NaHCO

so-

lution and saturated aqueous NaCl solution. Then, the organic layer

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

Scheme 1. Synthetic routes of EDP (a), EDPPA (b), EDPPO (c) and EDPMA (d).

Fig. 1. NMR spectra of EDP (a), EDPPA (b), EDPPO (c) and EDPMA (d), FT-IR spectrum (e).

was dried over anhydrous MgSO

and the solvent was removed by

rotary evaporator. The ﬁnal product was obtained after drying at

60 °C under reduced pressure. The synthetic routes were illustrated

in Scheme 1 and the NMR, FTIR spectra of products were shown in

Fig. 1 .

Analytical Data for EDPPA: light brown, transparent liquid.

H NMR (CDCl

, Fig. 1 b-1) δ(ppm): 1.19 (6H, CHOCH

1.34 (3H, POCH

), 3.65-3.86 (4H, CHO CH

), 4.14-4.28 (2H,

PO CH

), 4.83 (1H, CH OCH

), 6.78 (1H, N –H ), 6.92-7.22 (5H,

Ar-H ).

P NMR (CDCl

, Fig. 1 b-2) δ(ppm): 15.27. FTIR (KBr,

Fig. 1 e): 2981, 2933, 2900 cm

−1 (–CH

–, –CH

), 3037, 3085 cm

−1

(Ar–H), 3365 cm

−1 (N–H), 1286 cm

−1 (C–N), 1215 cm

−1 (P = O),

1170, 1111 cm

−1 (C–O–C), 1050 cm

−1 (P–O–C), 887 cm

−1 (P–N),

750 cm

−1 (P–C).

Analytical Data for EDPPO: colorless, transparent liquid.

NMR (CDCl

, Fig. 1 c-1) δ(ppm): 1.25 (6H, CHOCH

), 1.34 (3H,

PO CH

), 3.70-3.88 (4H, CHO CH

), 4.30 (2H, PO CH

4.92 (1H, CH OCH

), 7.15-7.33 (5H, Ar–H ).

P NMR (CDCl

Fig. 1 c-2) δ(ppm): 10.46. FTIR (KBr, Fig. 1 e): 2983, 2931, 2898

−1 (–CH

–, –CH

), 3071, 3046 cm

−1 (Ar–H), 1209 cm

−1 (P = O),

1165, 1117 cm

−1

(C–O–C), 1060 cm

−1

(P–O–C), 934 cm

−1

(P–O–Ar),

763 cm

−1 (P–C).

2.3. Synthesis of EDPMA

EDPMA was synthesized from EDP and methyl acrylate accord-

ing to the Michael addition reaction. Generally, EDP, methyl acry-

late (2.2 equiv.) and TEA (1.1 equiv.) were added into a 250 mL

ﬂask and stirred at 80 °C for 12 h. After removing TEA and methyl

acrylate through the rotary evaporator, the transparent, colorless,

liquid ﬁnal product was dried overnight under vacuum at 60 °C. The

synthetic route was illustrated in Scheme 1 and NMR, FTIR spectra

of product were shown in Fig. 1 .

Analytical Data for EDPMA: colorless, transparent liquid.

NMR (CDCl

, Fig. 1 d-1) δ(ppm): 1.26 (6H, CHOCH

), 1.33 (3H,

POCH

), 2.13 (2H, P CH

) 2.67 (2H, CH

COOCH

), 3.69 (4H,

CHO CH

), 3.84 (3H, CH

COO CH

), 4.20 (2H, PO CH

), 4.69

(1H, CH OCH

P NMR (CDCl

, Fig. 1 d-2) δ(ppm): 44.06. FTIR

(KBr, Fig. 1 e): 2979, 2933, 2902 cm

−1 (–CH

–, –CH

), 1740 cm

−1

(C = O), 1229 cm

−1

(P = O), 1170, 1114 cm

−1

(C–O–C), 1051 cm

−1

(P–

O–C), 761 cm

−1 (P–C).

2.4. Preparation of pure and ﬂame retardants FPUFs

The FPUFs used in this paper were prepared through a one-pot

free rising foaming process and the detailed formulations of these

FPUFs were listed in Table S1. In a plastic cup, polyether polyols,

ﬂame retardants, DBTDL, A33, distilled water, and silicone surfac-

tant were mixed using vigorous mechanical, and stirring for 1 min.

After adding given amount of TDI (NCO/OH = 1.05), the mixture

was vigorously stirred for 5 s, then was poured into an aluminum

mold to pre-polymerization. The FPUFs were obtained after further

cured at 80 °C for 24 h.

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

Fig. 2. SEM images of pure FPUF (a), FPUF-20PA (b), FPUF-20PO (c) and FPUF-20MA (d).

2.5. Characterization

Fourier transform infrared (FTIR) spectra of EDP and its deriva-

tives were obtained from a Nicolet 6700 spectrometer (Nicolet In-

strument Company, U.S.). Nuclear magnetic resonance (NMR) was

performed on a Bruker AV400 NMR spectrometer (400MHz) using

deuterated chloroform (CDCl

). The densities of these FPUFs were

measured manually according to ISO 845 standard, and the val-

ues were summarized in Table S1. The dimension of test sample

was 100 ×100 ×25 mm

. Scanning electron microscopy (SEM)

photographs were recorded on a FEI Sirion 200 scanning electron

microscope at an acceleration voltage of 10 kV. Thermogravimetric

analysis (TGA) was conducted under nitrogen atmosphere using a

Q50 0 0 thermal analyzer (TA Co., U.S.). The foam sample was cut

into small pieces and put in the pan, with a heating rate of 10 °C

/min from room temperature to 600 °C. Limiting oxygen index (LOI)

was obtained from a HC-2 oxygen index meter (Jiangning Analysis

Instrument Company, China) according to ISO 4589-1:1996. The di-

mension of the test samples was 150 ×10 ×10 mm

. Vertical burn

test was performed on a CFZ-2 type instrument (Jiangning Analy-

sis Instrument Co., China) based on California technical bulletin 117

(TB 117-20 0 0), the dimension of the test samples was 12 ×3 ×0.5

inch

. Cone calorimetry test was carried out on a cone calorimeter

according to ISO Standard 5660-1. The heat ﬂux was 25 kW/m

and the samples were of dimension 100 ×100 ×25 mm

. Raman

spectra were measured on a LabRAM-HR Confocal Raman Micro-

probe (JobinYvon Instruments, France) in the wave number range

of 20 0 0 to 50 0 cm

−1

3. Results and discussion

3.1. Microstructure of FPUFs

The microstructure of FPUFs was investigated using SEM, and

the SEM images were presented in Fig. 2 . As shown in Fig. 2 a, pure

FPUF exhibited an open-cell cellular structure with uniformed pore

size. In addition, the surface of the pores was smooth and homoge-

neous under high magniﬁcation. After incorporated with EDPs, the

microstructures of the modiﬁed FPUFs were not changed ( Fig. 2 b–

2 d). All EDPs modiﬁed foams maintained similar cellular structure

and uniformed pore sizes as compared with pure foam. Moreover,

no wrinkles could be observed on the surface of pores which in-

dicated the good compatibility of these liquid EDPs and FPUFs.

Nevertheless, the pore sizes of modiﬁed samples were decreased,

which might be attributed to the effect of EDPs on the activity of

foaming catalysts. The density of pure FPUF was measured as 34.4

kg/m

. After ﬂame retardant modiﬁcation, the densities of FPUFs

were increased. The densities of FPUF-10FR and FPUF-20FR sam-

ples were increased to about 38 kg/m

3 and 40 kg/m

, respectively.

This was because of the decrease of pore sizes that the FPUFs be-

came more denser as compared with pure foam. The investigation

of microstructure indicated that these EDPs were compatible with

FPUFs and would not affect the microstructure of FPUFs.

3.2. Thermal degradation of FPUFs

Thermogravimetric analysis (TGA) under nitrogen atmosphere

was conducted to investigate the thermal degradation of FPUFs.

The resultant curves of TGA and differential thermogravimetry

(DTG) were presented in Fig. 3 and related key parameters like the

initial decomposition temperature ( T

), the maximum weight loss

rate ( R

max

), the temperature at R

max

( T

max

) and the char yield after

thermal decomposition ( CY ) were summarized in Table 1 . The ther-

mal degradation of pure and EDPs incorporated FPUFs was similar

and could be divided in two steps. The ﬁrst weight loss stage was

attributed to the thermal decomposition of urethane bond and the

second one belonged to the thermal decomposition of the remain-

ing soft segments. For pure FPUF, the initial decomposition temper-

ature was 250 °C and it showed two maximum weight loss rate at

280 °C and 365 °C. After incorporated with EDPs, the T

and T

max1

values of modiﬁed FPUFs were decreased indicating that the ﬂame

retardant FPUFs were less thermal stable than pure FPUF. This phe-

nomenon was attributed to the lower thermal stability of these

ﬂame retardants that the thermal decomposition products of EDPs

may react with the urethane bond, promoting the decomposition

of polyurethane matrix. (TGA curves of EDPs were shown in Fig.

S2) In addition, among these modiﬁed samples, the T

values of

FPUF-MA were lower than the others. This was because that aro-

matic ring was more thermal stable than alkyl chain. Considering

max2

, the values of modiﬁed FPUFs shifted to higher temperatures.

However, the increase was not obvious. When the thermal decom-

position was ﬁnished at 600 °C, there was almost no CY for pure

FPUF. In contrast, all modiﬁed FPUFs presented higher CY , which

also increased with the increasing of ﬂame retardant amount. The

improvement of T

max2

and CY demonstrated that the ﬂame retar-

dant FPUFs were more thermal stable than pure FPUF at the sec-

ond thermal degradation stage. Based on the TGA results, it might

be suggested that EDPs predominantly decomposed at the ﬁrst

weight loss stage and the decomposition products of EDPs could

improve the char yield.

3.3. Flame retardancy of FPUFs

The ﬂame retardancy of FPUFs were evaluated by LOI, vertical

burning, UL-94 horizontal burning and cone calorimetry tests. LOI

represents the minimum oxygen content that needed for combus-

tion. Normally, materials with higher LOI are hard to combust. As

listed in Table 2 , the LOI value of pure FPUF was 17.5%. After in-

corporated with EDPs, higher LOI values were obtained for modi-

ﬁed foams and the LOI values increased slowly with the increasing

of ﬂame retardant addition. Among these modiﬁed foams, under

the same ﬂame retardant addition, FPUF-PA exhibited the highest

LOI values and the values of foam composites were in the order:

FPUF-PA > FPUF-MA > FPUF-PO. Moreover, vertical burning tests

of FPUFs were conducted to investigate the ﬂame spreading ac-

cording to TB117-20 0 0. The digital photos of FPUFs at 5s, 10s, 12s

and 15s during the vertical burning tests were presented in Fig. 4 .

For pure foam ( Fig. 4 a), it was easy to ignite and burned fast and

completely after ignition. When EDPs were incorporated, the ﬂame

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

Fig. 3. TGA (a) and DTG (b) curves of FPUFs.

Table 1

TGA results of FPUFs.

Sample T

( °C) T

max1

( °C) R

max1

(%/ °C) T

max2

( °C) R

max2

(%/ °C) CY (%)

Pure FPUF 250 280 0.583 365 2.05 1.8

FPUF-10PA 219 274 0.516 366 1.86 3.7

FPUF-20PA 192 275 0.466 368 1.62 4.1

FPUF-10PO 226 270 0.543 363 1.76 4.1

FPUF-20PO 206 268 0.551 361 1.54 4.4

FPUF-10MA 214 273 0.554 368 1.78 3.3

FPUF-20MA 190 276 0.448 368 1.59 3.8

Table 2

Data of cone calorimetry tests of pure FPUF, FPUF-PA, FPUF-PO and FPUF-MA.

Sample Pure FPUF FPUF- 10PA FPUF- 20PA FPUF-10PO FPUF-20PO FPUF-10MA FPUF- 20MA

TTI (s) 2 ±0 4 ±2 4 ±1 4 ±1 3 ±1 2 ±1 3 ±0

(s) 75 ±3 93 ±3 99 ±2 81 ±5 90 ±3 84 ±3 93 ±1

PHRR (kW/m

) 340 ±10 282 ±5 298 ±9 311 ±5 322 ±11 291 ±8 304 ±10

FIGRA 4.53 ±0.05 3.03 ±0.03 3.01 ±0.04 3.84 ±0.02 3.58 ±0.01 3.46 ±0.04 3.27 ±0.06

THR (MJ/m

) 22.8 ±0.5 20.2 ±0.7 20.6 ±0.8 21.1 ±0.7 21.7 ±1.0 20.4 ±0.3 20.9 ±0.6

THR/TML (MJ

•m

−2

•g

−1

) 2.69 ±0.02 2.19 ±0.05 2.12 ±0.07 2.34 ±0.02 2.33 ±0.05 2.20 ±0.07 2.18 ±0.04

TSP (m

) 1.33 ±0.06 2.03 ±0.07 2.32 ±0.04 1.81 ±0.05 2.19 ±0.05 1.88 ±0.03 2.28 ±0.07

CY (%) 1.4 ±0.4 3.2 ±0.5 3.7 ±0.3 3.4 ±0.2 3.9 ±0.3 2.8 ±0.1 3.4 ±0.1

LOI (%) 17.5 ±0.1 21.4 ±0.1 23.6 ±0.3 20.9 ±0.2 22.1 ±0.3 21.2 ±0.4 22.5 ±0.2

Vertical burning (Pass or Fail) Fail Pass Pass Fail Pass Fail Pass

UL94-HB No HF2 HF1 HF2 HF1 HF2 HF1

Fig. 4. Digital photos of pure FPUF (a), FPUF-10PA (b), FPUF-10PO (c) and FPUF-10MA (d) during vertical burning.

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

Fig. 5. PHRR (a) and THR (b) curves of FPUFs.

propagation became slower as compared with pure FPUF. Besides,

there were fewer melting droplets for the modiﬁed FPUFs. With

10 wt% loading of ﬂame retardant, only EDPPA modiﬁed FPUF can

pass the vertical burning test and exhibited self-extinguish behav-

ior ( Fig. 4 b). As shown in these pictures, the ﬂame zone become

smaller during vertical burning test. After combustion, the charred

region did not reach the clamp which indicated good ﬂame retar-

dancy of the EDPPA. When the addition increased to 20 wt%, all

ﬂame retardant FPUFs can pass the vertical burning (Fig. S1). Be-

sides, in UL 94-HB tests, FPUFs can reach HF-2 level with 10 wt%

ﬂame retardant, and HF-1 with 20 wt% ﬂame retardant. The re-

sults of LOI, vertical burning and horizontal burning tests indicated

that FPUF-PA showed the best ﬂame retardancy while the ﬂame

retardancy of FPUF-PO was the worst at the same ﬂame retardant

amount. This phenomenon was attributed to the structure inﬂu-

ence on the ﬂame retardancy of EDPs. The detailed ﬂame retar-

dant mechanisms of these EDPs would be discussed in the follow-

ing sections.

Cone calorimetry test is now recognized as the best method to

investigate the combustion properties of materials. In order to in-

vestigate the combustion properties of FPUFs and further evaluate

their ﬂame retardancy, cone calorimetry test with a heat ﬂux of

25 kW/m

was conducted according to ISO 5660-1. The time to ig-

nition (TTI), the peak heat release rate (PHRR), the time to PHRR

(t-PHRR), the total heat release (THR), the total smoke production

(TSP), the char yield (CY), the ﬁre growth rate (FIGRA) and the to-

tal heat release per total mass loss (THR/TML) were obtained from

cone calorimetry test measurements and values were summarized

in Table 2 . Besides, HRR and THR curves versus temperature were

presented in Fig. 5 . In cone calorimetry test, pure FPUF ignited and

burned fast. The values of its TTI and t-PHRR were only 2s and

72s, respectively. After incorporated with EDPs, the TTI and t-PHRR

values were increased, which meant the ignition and combustion

of FPUFs became slower as compared with pure FPUF. Moreover,

FIGRA of EDPs modiﬁed FPUFs were lower than pure FPUF too. FI-

GRA is deﬁned as the PHRR/t-PHRR, lower FIGRA represents pro-

longed time to the maximum heat release. [32] Therefore, the de-

crease of FIGRA could also indicated the ﬂame was slowing down.

PHRR and THR are key parameters in cone calorimetry test that

low PHRR and THR values indicate good ﬂame retardancy of ma-

terials. For pure FPUF, it exhibited one heat release peak in Fig. 5 ,

it’s PHRR and THR values were measured as 340 kW/m

2 and 22.8

MJ/m

, respectively. After ﬂame retardant modiﬁcation, the PHRR

and THR values of FPUFs became lower. Nevertheless, PHRR and

THR values of FPUFs bounced as more EDPs was incorporated.

This bounce was attributed to the increasing of foams’ densities

that more ﬂammable ingredients were involved into the combus-

tion. [ 33 , 34 ] In addition, it’s could be obviously noticed that FPUF-

10PA presented the lowest PHRR and THR values among the FPUF-

10FR samples, with a decrease of 17.1% and 11.4% respectively as

compared with pure FPUF; FPUF-10MA the second and FPUF-10PO

worst. When 20 wt% EDPs were incorporated, the decreasing trend

was the same. This results further indicated the structure inﬂuence

on ﬂame retardant effect of these EDPs.

THR/TML indicates the effective combustion heat releasing from

the volatiles, and its reduction is attributed to the gas phase ﬂame

retardant effects. [35] As seen in Table 2 , the THR/TML value of

pure FPUF was 2.69 MJ

•m

−2

•g

−1 and this value decreased after

EDPs was incorporated. Moreover, among the EDPs modiﬁed FPUFs,

the decrease of THR/TML were in the order: FPUF-PA > FPUF-MA

> FPUF-PO, which reﬂected the corresponding increasing in the

gas phase ﬂame retardant effect of EDPs. Besides, the increasing

of TSP values could also indicate the gas phase ﬂame retardant ef-

fect of EDPs, which presented the same growth trend with that of

THR/TML. The condensed phase ﬂame retardant effect is attributed

to the protective char residues that high CY would indicate high

ﬂame retardant effect. In cone calorimetry test, the CY values

of ﬂame retardant modiﬁed FPUFs were increased as compared

with pure sample. The value of FPUF-10PA, FPUF-20PA, FPUF-10PO,

FPUF-20PO, FPUF-10MA and FPUF-20MA was 3.2%, 3.7%, 3.4%, 3.9%,

2.8% and 3.4%, respectively. However, these increase were limited

due to the open-cell structure of FPUFs. Moreover, the increasing

of CY was corresponded with that in the TGA results.

Based on the above results and discussions, it could be con-

cluded that these as designed EDPs was able to improve the ﬂame

retardancy of FPUFs and the ﬂame retardancy of these EDPs in

FPUF decreases in the order: EDPPA > EDPPO > EDPMA.

3.4. Flame retardant mechanism of FPUFs

Generally, the char residues after combustion, and volatiles dur-

ing the thermal pyrolysis can reﬂect the ﬂame retardant mech-

anism of materials. SEM photographs, digital photos and Raman

spectra of the char residues after cone calorimetry tests were il-

lustrated in Fig. 6 . It was intuitively seen in Fig. 6 a that there were

almost no char residues for pure FPUF, while the modiﬁed samples

exhibited more char residues as compared with pure FPUF. Besides,

the char residues of FPUF-PO and FPUF-PA were more intact and

denser than that of FPUF-MA, as illustrated in Fig. 6 b. These results

suggested that the combustion products of FPUF-PO and FPUF-PA

can form dense and intact char layer to isolate the heat and oxy-

gen transfer and protect the inner foam. In addition, Raman spec-

troscopy was applied to evaluate the thermal stability of the char

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

Fig. 6. Digital photos, SEM images and the corresponding Raman spectra of char residues after cone calorimetry tests. (a, pure FPUF; b, FPUF-20PA; c, FPUF-20PO; d, FPUF-

20MA).

Fig. 7. FTIR spectra of volatiles at different temperature and the corresponding 3D absorbing curves of FPUFs. (a, pure FPUF; b, FPUF-20PA; c, FPUF-20PO; d, FPUF-20MA)

residues. In Raman spectrum, the peaks at 1390 cm

−1 and 1604

−1 belongs to the disordered graphite (D band) and the ordered

graphic structure (G band), respectively. The degree of graphitiza-

tion (I

) is deﬁned as the ratio of the integrated intensities of

D band to G band. Generally, lower I

value indicates higher

graphitization degree of char residues, which can serve as more

protective layers. [ 4 , 36–38 ] As presented in Fig. 6 c, I

value of

pure FPUF was 5.03. This value decreased to 3.08, 2.91 and 3.16

for FPUF-20PA, FPUF-20PO and FPUF-20MA, respectively, indicat-

ing more protective layers were generated by incorporating EDPs.

Besides, it was clearly seen that aromatic EDPs containing foams

(FPUF-PA, FPUF-PO) were able to generate more protective char

layers as compared with FPUF-MA. Therefore, the condensed phase

ﬂame retardancy of FPUFs was attributed to the ability to gen-

erate intact, dense and protective char layers during combustion.

Meanwhile, because of the denser, more intact and protective char

residues, FPUF-PO and FPUF-PA exhibited better ﬂame retardancy

than FPUF-MA.

To identify the gas phase ﬂame retardant mechanism of FPUFs,

TG-IR was conducted to investigate the volatiles during pyroly-

sis. The resultant IR spectra at the typical thermal degradation

temperature and corresponding 3D absorbing maps were illus-

trated in Fig. 7 . For pure FPUF, the absorption peaks at T

max1

were mainly accumulated at 2360 cm

−1 and 2322 cm

−1 due to

the vibration of CO

and –NCO groups, which was ascribed to

the decomposition of the hard segments. At T

max2

, the remain-

ing soft segments started to decompose and characteristic peaks of

methyl (3110 cm

−1

, 1462 cm

−1

), methylene or phenyl (2976 cm

−1

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

2884 cm

−1

), C = O (1742 cm

−1

), N = N (1376 cm

−1

), C–O (916 cm

−1

)

and NH

(1122 cm

−1

) were observed. [ 39 , 40 ] After incorporating

EDPs, phosphorus-containing compounds were detected in the py-

rolysis products of FPUFs. As marked in Fig. 7 b-c, the identical

peaks at 2346 cm

−1 and 1216 cm

−1 were because of the vibra-

tion of P–H and P = O, indicating the releasing of phosphine and

its derivatives. [ 41 , 42 ] During the combustion, these phosphorus-

containing compounds could generate phosphorus-containing rad-

icals and play the role of radical scavenger in the gas phase.

[ 36 , 43 ] In addition, it was obviously seen that no peaks of the

phosphorus-containing compounds were founded in IR spectrum

of FPUF-PO T

( Fig. 7 c). This phenomenon was attributed to the

reason that the bond energy of P–O is higher than P–C and P–

N, which means EDPPA and EDPMA are easier to pyrolysis and

generate phosphorus-containing compounds than EDPPO. Thus, the

gas phase ﬂame retardancy of FPUFs was attributed to the abil-

ity to generate phosphorus-containing radicals during the combus-

tion. Meanwhile, because of more phosphorus-containing radicals

existed in the gas phase, FPUF-PA and FPUF-MA exhibited better

ﬂame retardancy than FPUF-MA.

4. Conclusions

Three novel EDPs with similar structure were successfully syn-

thesized from hypophosphorous acid and incorporated into FPUFs.

The result of SEM revealed that these EDPs were compatible with

FPUFs and would not affect the structure. In TGA tests, the mod-

iﬁed FPUFs presented lower thermal stability as compared with

pure FPUF. Nevertheless, they showed better ﬂame retardancy in

the LOI, vertical burning and cone calorimetry tests. Besides, it was

clearly seen that EDPPA exhibited the best ﬂame retardant effect in

FPUFs. With 10 wt% EDPPA, the modiﬁed FPUF can pass the verti-

cal burning and show high LOI value of 23.6%. In addition, FPUF-

10PA showed the lowest PHRR and THR values among these FPUFs,

with a decrease of 17.1% and 11.4% respectively as compared with

pure FPUF. This phenomenon was because of the structure inﬂu-

ence on ﬂame retardancy that FPUF-PA possessed the ability to

form denser, more intact, protective char layers and more radical

scavengers as compared with the other two FPUF-EDPs.

Declaration of Competing Interest

The authors declare that they have no known competing ﬁnan-

cial interests or personal relationships that could have appeared to

inﬂuence the work reported in this paper.

CRediT authorship contribution statement

Feng Zhou: Investigation, Formal analysis, Writing - original

draft. Chao Ma: Validation, Writing - review & editing. Kang

Zhang: Data curation. Yin yam Chan: Visualization. Yuling Xiao:

Software. Bernhard Schartel: Methodology. Manfred Doring: Con-

ceptualization. Bibo Wang: Resources. Weizhao Hu: Project ad-

ministration. Yuan Hu: Funding acquisition.

Acknowledgments

The work was ﬁnancially supported by the National Natural Sci-

ence Foundation of China ( 51803204 , 51761135113 , U1833113 and

51874266 ) and Fundamental Research Funds for the Central Uni-

versities ( WK2320 0 0 0 043 ).

References

[1] K.Y. Guo , Q. Wu , M. Mao , H. Chen , G.D. Zhang , L. Zhao , J.F. Gao , P. Song ,

L.C. Tang , Water-based hybrid coatings toward mechanically ﬂexible, super-hy-

drophobic and ﬂame-retardant polyurethane foam nanocomposites with high-

-eﬃciency and reliable ﬁre alarm response, Compos. Part B Eng. 193 (2020)

108017 .

[2] H.W. Engels , H.G. Pirkl , R. Albers , R.W. Albach , J. Krause , A. Hoffmann , H. Cas-

selmann , J. Dormish , Polyurethanes: Versatile materials and sustainable prob-

lem solvers for today’s challenges, Angew. Chemie - Int, Ed. 52 (2013)

9422–9441 .

[3] Y. Zhang , W. Tian , L. Liu , W. Cheng , W. Wang , K.M. Liew , B. Wang , Y. Hu ,

Eco-friendly ﬂame retardant and electromagnetic interference shielding cotton

fabrics with multi-layered coatings, Chem. Eng. J. 372

(2019) 1077–1090 .

[4] J. Wang , D. Zhang , Y. Zhang , W. Cai , C. Yao , Y. Hu , W. Hu , Construction of

multifunctional boron nitride nanosheet towards reducing toxic volatiles (CO

and HCN) generation and ﬁre hazard of thermoplastic polyurethane, J. Hazard.

Mater. 362 (2019) 4 82–4 94 .

[5] Y. Zhang , W. Cheng , W. Tian , J. Lu , L. Song , K.M. Liew , B. Wang , Y. Hu ,

Nacre-Inspired Tunable Electromagnetic Interference Shielding Sandwich Films

with Superior Mechanical and Fire-Resistant Protective Performance, ACS Appl.

Mater. Interfaces. 12 (2020) 6371–6382 .

[6] Q. Wu , Q. Zhang , L. Zhao , S.N. Li , L. Bin Wu , J.X. Jiang , L.C. Tang , A novel and

facile strategy for highly ﬂame retardant polymer foam composite materials:

Transforming silicone resin coating into silica self-extinguishing layer, J. Haz-

ard. Mater. 336 (2017) 222–231 .

[7] G. Laufer , C. Kirkland , A .A . Cain , J.C. Grunlan , Clay-chitosan nanobrick walls:

Completely renewable gas barrier and ﬂame-retardant nanocoatings, ACS Appl.

Mater. Interfaces 4 (2012) 1643–1649 .

[8] Y. Pan , W. Cai , J. Du , L. Song , Y. Hu , H. Zhao , Lanthanum phenylphospho-

nate–based multilayered coating for reducing ﬂammability and smoke produc-

tion of ﬂexible polyurethane foam, Polym. Adv. Technol. (2020) 1330–1339 .

[9] M. Zhi , Q. Liu , Y. Zhao , S. Gao , Z. Zhang , Y. He , Novel MoS2-DOPO Hybrid for

Effective Enhancements on Flame Retardancy and Smoke Suppression of Flex-

ible Polyurethane Foams, ACS Omega 5 (2020) 2734–2746 .

[10] P. Chen , Y. Zhao , W. Wang , T. Zhang , S. Song , Correlation of montmoril-

lonite sheet thickness and ﬂame retardant behavior of a chitosan-montmoril-

lonite nanosheet membrane assembled on ﬂexible

polyurethane foam, Poly-

mers (Basel) 11 (2019) .

[11] W.H. Rao , Z.Y. Hu , H.X. Xu , Y.J. Xu , M. Qi , W. Liao , S. Xu , Y.Z. Wang , Flame-Re-

tardant Flexible Polyurethane Foams with Highly Eﬃcient Melamine Salt, Ind.

Eng. Chem. Res. 56 (2017) 7112–7119 .

[12] S. Liang , M. Neisius , H. Mispreuve , R. Naescher , S. Gaan ,Flame retardancy and

thermal decomposition of ﬂexible polyurethane foams: Structural inﬂuence of

organophosphorus compounds, Polym. Degrad. Stab. 97 (2012) 2428–2440 .

[13] N.M. Neisius , M. Lutz , D. Rentsch , P. Hemberger , S. Gaan , Synthesis of

DOPO-based phosphonamidates and their thermal properties, Ind. Eng. Chem.

Res. 53 (2014) 2889–2896 .

[14] J. Wang , C. Ma , P. Wang , S. Qiu , W. Cai , Y. Hu , Ultra-low phosphorus loading to

achieve the superior ﬂame retardancy of epoxy resin, Polym. Degrad. Stab. 149

(2018) 119–128 .

[15] X. Wang , J. Zhan , W. Xing , X. Wang , L. Song , X. Qian , B. Yu , Y. Hu , Flame re-

tardancy and thermal properties of novel UV-curable epoxy acrylate coatings

modiﬁed by a silicon-bearing hyperbranched polyphosphonate acrylate, Ind.

Eng. Chem. Res. 52 (2013) 5548–5555 .

[16] M. Zhang , Y. Li , C. Bai , X. Guo , J. Han , S. Hu , H. Jiang , W. Tan , S. Li , L. Ma , Syn-

thesis of Microporous Covalent Phosphazene-Based Frameworks for Selective

Separation of Uranium in Highly Acidic Media Based on Size-Matching Effect,

ACS Appl. Mater. Interfaces. 10 (2018) 28936–28947 .

[17] L. Ding , C. Wang , L. Lin , Z. Zhu , One-pot sequential ring-opening metathe-

sis polymerization and acyclic diene metathesis polymerization synthesis

of unsaturated block polyphosphoesters, Macromol. Chem. Phys. 216 (2015)

761–769

[18] P. Wen , Q. Tai , Y. Hu , R.K.K. Yuen , Cyclotriphosphazene-Based Intumescent

Flame Retardant against the Combustible Polypropylene, Ind. Eng. Chem. Res.

55 (2016) 8018–8024 .

[19] H.Y. Ma , L.F. Tong , Z. Bin Xu , Z.P. Fang , Functionalizing carbon nanotubes by

grafting on intumescent ﬂame retardant: nanocomposite synthesis, morphol-

ogy, rheology, and ﬂammability, Adv. Funct. Mater. 18 (2008) 414–.

[20] F. Zhou , T. Zhang , B. Zou , W. Hu , B. Wang , J. Zhan , C. Ma , Y. Hu , Synthesis of

a novel liquid phosphorus-containing ﬂame retardant for ﬂexible

polyurethane

foam: Combustion behaviors and thermal properties, Polym. Degrad. Stab. 171

(2020) 109029 .

[21] H. Xie , X. Lai , H. Li , X. Zeng , Synthesis of a novel macromolecular charring

agent with free-radical quenching capability and its synergism in ﬂame retar-

dant polypropylene, Polym. Degrad. Stab. 130 (2016) 68–77 .

[22] F. Xu , L. Zhong , C. Zhang , P. Wang , F. Zhang , G. Zhang ,Novel High-Eﬃciency

Casein-Based P-N-Containing Flame Retardants with Multiple Reactive Groups

for Cotton Fabrics, ACS Sustain. Chem. Eng. 7 (2019) 13999–14008 .

[23] T.T. Li , M. Xing , H. Wang

, S.Y. Huang , C. Fu , C.W. Lou , J.H. Lin , Nitro-

gen/phosphorus synergistic ﬂame retardant-ﬁlled ﬂexible polyurethane foams:

Microstructure, compressive stress, sound absorption, and combustion resis-

tance, RSC Adv 9 (2019) 21192–21201 .

[24] Z.M. Zhu , L.X. Wang , L.P. Dong , Inﬂuence of a novel P/N-containing oligomer

on ﬂame retardancy and thermal degradation of intumescent ﬂame-retardant

epoxy resin, Polym. Degrad. Stab. 162 (2019) 129–137 .

[25] T. Zhang , Y. Tao , F. Zhou , H. Sheng , S. Qiu , C. Ma , Y. Hu , Synthesis of a hy-

perbranched phosphorus-containing polyurethane as char forming

agent com-

bined with ammonium polyphosphate for reducing ﬁre hazard of polypropy-

lene, Polym. Degrad. Stab. 165 (2019) 207–219 .

F. Zhou, C. Ma, K. Zhang et al. Polymer Degradation and Stability 188 (2021) 109557

[26] I.D. Carja , D. Serbezeanu , T. Vlad-Bubulac , C. Hamciuc , A. Coroaba , G. Lisa ,

C.G. López , M.F. Soriano , V.F. Pérez , M.D. Romero Sánchez , A straightforward,

eco-friendly and cost-effective approach towards ﬂame retardant epoxy resins,

J. Mater. Chem. A. 2 (2014) 16230–16241 .

[27] W. Zhao , J. Liu , H. Peng , J. Liao , X. Wang , Synthesis of a novel PEPA-substi-

tuted polyphosphoramide with high char residues and its performance as an

intumescent ﬂame retardant for epoxy resins, Polym. Degrad. Stab. 118 (2015)

120–129 .

[28] M. Neisius , S. Liang , H. Mispreuve , S. Gaan , Phosphoramidate-containing

ﬂame-retardant ﬂexible polyurethane foams, Ind. Eng. Chem. Res. 52 (2013)

9752–9762 .

[29] K.A. Salmeia , F. Flaig , D. Rentsch , S. Gaan , One-pot synthesis of P(O)-N con-

taining compounds using N-chlorosuccinimide and their inﬂuence in thermal

decomposition of PU foams, Polymers (Basel) 10 (2018) .

[30] M.J. Chen , Y.J. Xu , W.H. Rao , J.Q. Huang , X.L. Wang , L. Chen , Y.Z. Wang , In-

ﬂuence of valence and structure of phosphorus-containing melamine salts on

the decomposition and ﬁre behaviors of ﬂexible polyurethane foams, Ind. Eng.

Chem.

Res. 53 (2014) 8773–8783 .

[31] C. Fougère , E. Guénin , J. Hardouin , M. Lecouvey , Rapid and eﬃcient synthesis

of unsymmetrical phosphinic acids r



t(o)ohr



, European J. Org. Chem. (2009)

6048–6054 .

[32] D. Wang , P. Wen , J. Wang , L. Song , Y. Hu , The effect of def ect-rich molybde-

num disulﬁde nanosheets with phosphorus, nitrogen and silicon elements on

mechanical, thermal, and ﬁre behaviors of unsaturated polyester composites,

Chem.

Eng. J. 313 (2017) 238–249 .

[33] W.H. Rao , H.X. Xu , Y.J. Xu , M. Qi , W. Liao , S. Xu , Y.Z. Wang , Persistently

ﬂame-retardant ﬂexible polyurethane foams by a novel phosphorus-containing

polyol, Chem. Eng. J. 343 (2018) 198–206 .

[34] A.H. Kang , K. Shang , D.D. Ye , Y.T. Wang , H. Wang , Z.M. Zhu , W. Liao , S.M. Xu ,

Y.Z. Wang , D.A. Schiraldi , Rejuvenated ﬂy ash in poly(vinyl alcohol)-based com-

posite aerogels with high ﬁre safety and smoke suppression, Chem. Eng. J. 327

(2017) 992–999 .

[35] N. Wu , G. Fu , Y. Yang , M. Xia , H. Yun , Q. Wang ,Fire safety enhancement of a

highly eﬃcient ﬂame retardant poly(phenylphosphoryl phenylenediamine) in

biodegradable poly(lactic acid), J. Hazard. Mater. 363 (2019) 1–9 .

[36] Y. Shi , B. Yu , L. Duan , Z. Gui , B. Wang , Y. Hu , R.K.K. Yuen , Graphitic carbon ni-

tride/phosphorus-rich aluminum phosphinates hybrids as smoke suppressants

and ﬂame retardants for polystyrene, J. Hazard. Mater. 332 (2017) 87–96 .

[37] Y. Yuan , W. Wang , Y. Shi , L. Song , C. Ma , Y.

Hu , The inﬂuence of highly dis-

persed Cu2O-anchored MoS2 hybrids on reducing smoke toxicity and ﬁre haz-

ards for rigid polyurethane foam, J. Hazard. Mater. 382 (2020) 121028 .

[38] C. Ma , B. Yu , N. Hong , Y. Pan , W. Hu , Y. Hu , Facile Synthesis of a Highly Eﬃ-

cient, Halogen-Free, and Intumescent Flame Retardant for Epoxy Resins: Ther-

mal Properties, Combustion Behaviors, and Flame-Retardant Mechanisms, Ind.

Eng. Chem. Res. 55 (2016) 10868–10879 .

[39] F. Gao , L. Tong , Z. Fang , Effect of a novel phosphorous-nitrogen containing in-

tumescent ﬂame retardant on

the ﬁre retardancy and the thermal behaviour

of poly(butylene terephthalate), Polym. Degrad. Stab. 91 (2006) 1295–1299 .

[40] L. Chen , L. Song , P. Lv , G. Jie , Q. Tai , W. Xing , Y. Hu , A new intumescent ﬂame

retardant containing phosphorus and nitrogen: Preparation, thermal properties

and application to UV curable coating, Prog. Org. Coatings. 70 (2011) 59–66 .

[41] K. Cao , S.L. Wu , S.L. Qiu , Y. Li , Z. Yao , Synthesis of N-alkoxy hindered amine

containing silane as a multifunctional ﬂame retardant synergist and its appli-

cation in intumescent ﬂame retardant

polypropylene, Ind. Eng. Chem. Res. 52

(2013) 309–317 .

[42] W. Xu , R. Chen , Y. Du , G. Wang , Design water-soluble phenolic/zeolitic im-

idazolate framework-67 ﬂame retardant coating via layer-by-layer assembly

technology: Enhanced ﬂame retardancy and smoke suppression of ﬂexible

polyurethane foam, Polym. Degrad. Stab. 176 (2020) 109152 .

[43] Y. Yuan , H. Yang , B. Yu , Y. Shi , W. Wang , L. Song , Y. Hu , Y. Zhang , Phosphorus

and Nitrogen-Containing Polyols: Synergistic Effect on the Thermal Property

and Flame Retardancy of Rigid Polyurethane Foam Composites, Ind. Eng. Chem.

Res. 55 (2016)

10813–10822 .

4.3 A liquid phosphorous flame retardant combined with expandable graphite or melamine in

flexible polyurethane foam

Yin Yam Chan, Chao Ma, Feng Zhou, Yuan Hu and Bernhard Schartel. Polym Advan Technol. 2022,

33, 326-339. Doi: 10.1002/pat.5519.

Status: This article was accepted and published. https://doi.org/10.1002/pat.5519

First author contribution:

• Validation

• Methodology

• Investigation

• Data curation

• Writing – Original draft, review and editing

• Visualization

• Project administration

Contributions from other authors:

• Chao Ma

o Conceptualization

o Investigation

o Resources

• Feng Zhou

o Conceptualization

o Investigation

o Resources

• Yuan Hu

o Conceptualization

o Resources

o Supervision

o Project administration

o Funding acquisition

• Bernhard Schartel

o Methodology

o Resources

o Writing – Original draft, review and editing

o Supervision

o Project administration

o Funding acquisition

Abstract

A systematic series of flexible polyurethane foams (FPUF) with different concentrations of flame

retardants, bis([dimethoxyphosphoryl]methyl) phenyl phosphate (BDMPP), and melamine (MA) or

expandable graphite (EG) was prepared. The mechanical properties of the FPUFs were evaluated by a

universal testing machine. The pyrolysis behaviors and the evolved gas analysis were done by

thermogravimetric analysis (TGA) and TGA coupled with Fourier-transform infrared (TG-FTIR),

respectively. The fire behaviors were studied by limiting oxygen index (LOI), UL 94 test for horizontal

burning of cellular materials (UL 94 HBF), and cone calorimeter measurement. Scanning electronic

microscopy (SEM) was used to examine the cellular structure’s morphology and the postfire char residue

of the FPUFs. LOI and UL 94 HBF tests of all the flame retarded samples show improved flame

retardancy. BDMPP plays an essential role in the gas phase because it significantly reduces the effective

heat of combustion (EHC). This study highlights the synergistic effect caused by the combination of

BDMPP and EG. The measured char yield from TGA is greater than the sum of individual effects. No

dripping phenomenon occurs during burning for FPUF-BDMPP-EGs, as demonstrated by the result of

the UL 94 HBF test. EG performs excellently on smoke suppression during burning, as evident in the

result of the cone calorimeter test. MA reduces the peak heat release rate (pHRR) significantly. The

synergistic effect of the combination of BDMPP and EG as well as MA offers an approach to enhance

flame retardancy and smoke suppression.

RESEARCH ARTICLE

A liquid phosphorous flame retardant combined with

expandable graphite or melamine in flexible polyurethane foam

Yin Yam Chan

| Chao Ma

| Feng Zhou

| Yuan Hu

| Bernhard Schartel

Bundesanstalt für Materialforschung und

-prüfung (BAM), Berlin, Germany

State Key Laboratory of Fire Science,

University of Science and Technology of

China, Hefei, China

Correspondence

Bernhard Schartel, Bundesanstalt für

Materialforschung und -prüfung (BAM), Unter

den Eichen 87, 12205 Berlin, Germany.

Email: [email protected]

Funding information

Deutsche Forschungsgemeinschaft, Grant/

Award Number: Scha 730/19-1; National

Natural Science Foundation of China, Grant/

Award Number: 51761135113

Abstract

A systematic series of flexible polyurethane foams (FPUF) with different concentra-

tions of flame retardants, bis([dimethoxyphosphoryl]methyl) phenyl phosphate

(BDMPP), and melamine (MA) or expandable graphite (EG) was prepared. The

mechanical properties of the FPUFs were evaluated by a universal testing machine.

The pyrolysis behaviors and the evolved gas analysis were done by ther-

mogravimetric analysis (TGA) and TGA coupled with Fourier-transform infrared (TG-

FTIR), respectively. The fire behaviors were studied by limiting oxygen index (LOI),

UL 94 test for horizontal burning of cellular materials (UL 94 HBF), and cone calorim-

eter measurement. Scanning electronic microscopy (SEM) was used to examine the

cellular structure's morphology and the postfire char residue of the FPUFs. LOI and

UL 94 HBF tests of all the flame retarded samples show improved flame retardancy.

BDMPP plays an essential role in the gas phase because it significantly reduces the

effective heat of combustion (EHC). This study highlights the synergistic effect cau-

sed by the combination of BDMPP and EG. The measured char yield from TGA is

greater than the sum of individual effects. No dripping phenomenon occurs during

burning for FPUF-BDMPP-EGs, as demonstrated by the result of the UL 94 HBF test.

EG performs excellently on smoke suppression during burning, as evident in the

result of the cone calorimeter test. MA reduces the peak heat release rate (pHRR) sig-

nificantly. The synergistic effect of the combination of BDMPP and EG as well as MA

offers an approach to enhance flame retardancy and smoke suppression.

KEYWORDS

bis([dimethoxyphosphoryl]methyl) phenyl phosphate, expandable graphite, flexible

polyurethane foam, melamine, phosphorous flame retardant

1|INTRODUCTION

Generally, polyurethanes (PU) are a class of copolymers composed of

soft and hard segments. Usually, the soft segment is a polyester or a

polyether polyol, whereas the hard segment is composed of isocyanate

and maybe chain extender if needed.

The soft segment determines

elasticity, while the hard segment provides strength and rigidity. PU

with desired physical and mechanical properties can be produced by

altering the ratio of the soft segment to the hard segment. For the past

several decades, PU has been used frequently because of its wide range

of applications in products such as foams, elastomers, adhesives, paints,

and coatings. The PU foams are usually classified into rigid, semi-rigid,

and flexible types, depending primarily on the density and degree of

openness of the cells. Closed-cell rigid PU foam is used mainly for

Received: 28 May 2021 Revised: 23 August 2021 Accepted: 22 September 2021

DOI: 10.1002/pat.5519

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,

provided the original work is properly cited.

Polym Adv Technol. 2021;1–14. wileyonlinelibrary.com/journal/pat 1

thermal insulation in buildings and refrigerators. Open-cell flexible PU

foam (FPUF) is used as a cushion material in furniture, vehicles, and

packaging. It is easy to ignite PU foams readily by a small flame source,

and they burn quickly with a high rate of heat release because of their

cellular structure, low density, and high hydrocarbon content.

2–5

reduce the possibility of fire, commercial additive-type flame retardants

available on the market are simply physically mixed with the polymer

matrix. Metal hydroxides, halogenated compounds, phosphorous com-

pounds, melamine cyanurate, and intumescent products are commonly

used as flame retardants in polymers.

The concerns regarding health

and environmental problems caused by halogenated flame retardants

aroused interest among scientists in developing non-halogenated flame

retardants. Interestingly, incorporating more than one type of flame

retardant into the polymer matrix may bring a synergistic effect to

flame retardancy.

7,8

A synergistic effect means that the overall flame

retardancy is even better than the superposition of the individual com-

ponent's effects. Wilke et al. studied the synergetic effect between

phosphorus and expandable graphite (EG) in thermoplastic styrene-eth-

ylene-butylene-styrene elastomers.

Rao et al. found that EG and phos-

phorus contributed to the compactness of char residue in FPUF.

This

is due to the gluing effect on to the fluffy expanded graphite exerted

by the phosphorous compound. They concluded that the intensity of

synergism, which provides better flame retardancy to the material,

increased significantly when the proper ratio of EG to phosphorus was

applied. Feng et al. explored the synergistic effect between phosphorus

and EG as well.

They showed that the system remarkably increased

residual char yield and intensely reduced the fire parameters compared

to the one using a single flame retardant. Synergistic action is apparent

in the polymer matrix with phosphorus and nitrogen compounds. Yuan

et al. synthesized phosphorous and melamine-derived polyol for rigid

polyurethane foam and found out that the appropriate ratio between

these components greatly improved the material's fire performance.

Phosphorus-nitrogen synergism in cotton cellulose was a focus of the

work by Gaan and his coworkers.

They proposed that the formation

of a protective layer during the burning process was an observable

effective synergism between phosphorus and nitrogen. The voluminous

protective layer acted as a shield to prevent further burning of the

underlying materials. By taking advantage of synergism, a smaller quan-

tity of flame retardants can be used to maintain the physical and

mechanical properties of the material. Among different phosphorous

flame retardants, dimethyl methylphosphonate (DMMP) is an effective

flame retardant because it contains a rather high content of phosphorus

(25 wt%) compared to other phosphorous flame retardants, such

as 9,10-dihydro-9-oxa-10-phosphaphenanthrene-10-oxide, triphenyl

phosphate, triethyl phosphate, and aluminum diethylphosphinate.

In this work, a liquid phosphorous flame retardant, bis

([dimethoxyphosphoryl]methyl) phenyl phosphate (BDMPP), was com-

bined with two commercial flame retardants, melamine (MA) and EG in

FPUF. Zhou et al. proposed and compared BDMPP with dimethyl

methylphosphonate (DMMP) in terms of fire properties; they found

that BDMPP is more advantageous to flame retardancy than DMPP.

BDMPP still contains a very high phosphorus content (22 wt%).

BDMPP contains two kinds of phosphorus, phosphonate, and phos-

phate, since it is synthesized from dimethyl (hydroxymethyl)

phosphonate and phenyl dichlorophosphate. The presence of aromatic

rings in BDMPP contributed to a better charring effect and an

increased char yield in the condensed phase during burning. The molec-

ular weight of BDMPP is higher than that of DMMP. Hence, BDMPP

provided superior retainability to inhibit migration and volatilization.

The current work focusses on the flame retardancy and fire perfor-

mance of FPUF incorporated with BDMPP, MA, and EG.

2|EXPERIMENTAL

2.1 |Materials

The chemicals used in the foaming formulation are described in

Table 1. The materials listed in Table 1, except deionized water, were

provided by Jiangsu Lvyuan New Material Co., Ltd. (Jiangsu, China).

For the flame retardants, EG flakes with an expansion ratio of

150–200 and particle size of 30–50 μm were supplied by Qingdao

Xingyuan Colloidal Graphite Co., Ltd. (Qingdao, China). MA was pur-

chased from Anhui Jinhe Chemical Co., Ltd (Shanghai, China). BDMPP

was prepared by the State Key Laboratory of Fire Science, the Univer-

sity of Science and Technology of China, following the procedures

described in Reference 14. The attention of readers interested in

corresponding information on BDMPP is also turned to the original

paper published on its synthesis.

A set of foam samples were prepared at the State Key Laboratory

of Fire Science, University of Science and Technology of China. Three

additive-type flame retardants were used in the FPUF. One of the

flame retardants is a liquid phosphorous flame retardant, shown in

Figure 1, called bis([dimethoxyphosphoryl]methyl) phenyl phosphate

TABLE 1 Foaming formulation of flexible polyurethane foam

Material Weight (g)

Component

Polyether polyol (330, hydroxyl

value =56 mg KOH/g, number

average molecular

weight =3000 g/mol, average

functionality =3)

62.5

Grafted polyether polyol (2045,

prepared by the in situ

polymerization of acrylonitrile and

styrene in a polyether polyol,

hydroxyl value =20 mg KOH/g,

number average molecular

weight =8400 g/mol,

functionality =3)

20.83

Silicone oil 0.92

Stannous octoate 0.15

Triethylenediamine (A33, 33%) 0.23

Dichloromethane 2.92

Deionized water 2.67

Component

Toluene diisocyanate (TDI 80/20,

80:20 mixture of 2,4-toluene

diisocyanate and 2,6-toluene

diisocyanate)

38.33

2CHAN ET AL.

(BDMPP). The rest are all commercial flame retardants, MA and

EG. Each sample, except the reference sample FPUF, contains 20-phr

of the polyether polyol and the grafted polyether polyol blended with

a sole flame retardant or adual flame retardant.Table 2 shows the type

and amount of flame retardants used in each sample. The additive

flame retardants were weighted based on parts per hundred of the

sum of polyether polyol and grafted polyether polyol.

2.2 |Sample preparation

The preparation of FPUF was conducted by mixing component A and

component B using the one-pot method. The foams were prepared at

60C for 20 min in a temperature controlled closed mold

200 200 100 mm (lengthwidththickness) in size. First, compo-

nent A was stirred uniformly in a disposable polypropylene cup by a

high-speed stirrer for 3 min. Afterward, component B was mixed with

the blended component A for few seconds under a high stirring rate.

Then the mixture was discharged into the mold. The foam was cured

for 24 h in an oven at 80C to complete the polymerization.

2.3 |Measurements and characterization

2.3.1 | Morphological characterization

The micrographs of the foams and their char residues were examined

using a scanning electron microscope (SEM) Zeiss EVO 10 (Oberkochen,

Germany). The acceleration voltage was set to 10 kV. The specimens

were sputter-coated with 15 nm of gold to reduce the chances of elec-

trostatic charging. Only the foam structure at the surface, the direct

interface with the mold, shows a skin accompanied with a thin layer of

smaller cell size and higher density. We prepared and investigated only

specimens cut out from the inner homogenous part of the foams.

2.3.2 | Physical and mechanical properties

measurements

The apparent density of the specimen was measured according to ISO

845. A Universal Testing Machine Zwick Z010 (Ulm, Germany) was used

to evaluate the tensile strength, elongation at break, and compression

strength. A load cell with 500 N was used, the speed of the power-

actuated grip was 500 mm/min for the tensile test. The strain rate was

100 mm/min for the compression test. The tensile strength and elonga-

tion at break were measured using specimens with a thickness of

10 mm cut as test piece type 1A following ISO 1798, while the com-

pressive strength was measured according to ISO 3386-1. The specimen

size for determining the compression was 40 30 10 mm

(lengthwidththickness). 3 cycles of compression were performed for

the compression test. Four and three test specimens were measured for

each material in the tensile test and the compression test, respectively.

2.3.3 | Pyrolysis: Mass loss and evolved gases

Thermogravimetric analysis (TGA) records the change in the mass of a

sample as a function of time under a nitrogen atmosphere using a TG

209 F1 Iris from Netzsch Instruments (Selb, Germany), thereby deter-

mining the thermal decomposition behavior of the samples. The alu-

mina crucible with 10 mg of the powdered sample was then put on

the thermo-microbalance of the TGA device. The samples were sub-

jected to a heating program under a constant nitrogen gas flow of

30 ml/min at a steady heating rate of 10 K/min. Simultaneously, TGA

coupled with Fourier transform infrared spectroscopy (FTIR), Brucker

Tensor 27 FT-IR (Ettlingen, Germany) analyzes the gaseous pyrolysis

products evolved from the specimens in the furnace.

2.3.4 | Fire behavior

Before the measurements, all test specimens were stored at 23Cand

50% relative humidity for a minimum of 48 h. The limiting oxygen index

(LOI) of the specimens 150 10 10 mm (lengthwidththickness)

in size was determined at room temperature according to ISO 4589-2.

The fire behavior was analyzed using a cone calorimeter from Fire Test-

ing Technology Limited (West Sussex, United Kingdom), and the test

was carried out in accordance with ISO 5660. The specimen for the

FIGURE 1 Bis([dimethoxyphosphoryl]methyl) phenyl phosphate

(BDMPP)

TABLE 2 Content of flame retardants in the samples

Sample Flame retardants (in phr

)

FPUF —

FPUF-20BDMPP 20-phr BDMPP

FPUF-20MA 20-phr MA

FPUF-20EG 20-phr EG

FPUF-5BDMPP-15MA 5-phr BDMPP and 15-phr MA

FPUF-10BDMPP-10MA 10-phr BDMPP and 10-phr MA

FPUF-15BDMPP-5MA 15-phr BDMPP and 5-phr MA

FPUF-5BDMPP-15EG 5-phr BDMPP and 15-phr EG

FPUF-10BDMPP-10EG 10-phr BDMPP and 10-phr EG

FPUF-15BDMPP-5EG 15-phr BDMPP and 5-phr EG

phr, Parts per hundred of the sum of polyether polyol and grafted

polyether polyol.

CHAN ET AL.3

cone calorimeter is 100 100 50 mm (lengthwidththickness) in

size. The specimen placed in an aluminum foil container was exposed in

the horizontal orientation to a radiative heat flux of 25 kW m

2

with a

distance of 25 mm between the cone heater and the surface of the

specimen. All materials showed and impressive repeatability and thus

were only measured twice in the cone calorimeter. UL 94 HBF is used

to measure the burning rate and evaluate the tendency of the materials

to either extinguish or spread the flame when the specimen has

been ignited. It was performed to determine the horizontal burning

characteristics following ISO 9772 with a specimen size of

150 50 10 mm (lengthwidththickness). The number of speci-

mens for LOI and UL 94 were according to the standards.

3|RESULTS AND DISCUSSION

3.1 |Morphological characterization and

mechanical properties measurements

The morphology is a critical factor that influences the physical and

mechanical properties of FPUF. A scanning electron microscope

(SEM) was used to observe the morphology of foams, and the SEM

images are displayed in Figure 2. All foams show a semi-open cell

structure. The FPUF in Figure 2A has smoother edges on the blow-

holes, uniform cell size, and even cell distribution. Generally, high

loading of additive-type flame retardants must be incorporated into

the polymer matrix for better flame retardancy. Therefore, the high

content of additive-type flame retardants is often detrimental to the

mechanical properties, as they function mainly as a nucleating agent.

The SEM images show the influence of flame retardants on cellular

structures. FPUF-20BDMPP, and FPUF-20MA in Figure 2B,C, respec-

tively, show a similar structure with cells slightly larger than in FPUF.

The SEM images of FPUF-20EG and FPUF-BDMPP-EGs show less

continuous and less regular spherical cell structure. This phenomenon

was attributed to nucleation triggered by EG particles in the polymer

matrix.

15–16

EG is a kind of solid particle that affects bubble nucleation

and bubble growth in the foaming process, thereby damaging the

foam's structure to some extent. Especially for FPUF-10BDMPP-

10EG shown in Figure 2I, the struts are somewhat thicker, and the

cellular structure is somewhat less complete than the others.

Density is a major parameter affecting foam flexibility and support.

The apparent density of the flame retardant samples shown in Table 3

FIGURE 2 SEM images of (A) FPUF, (B) FPUF-20BDMPP, (C) FPUF-20MA, (D) FPUF-20EG, (E) FPUF-5BDMPP-15MA, (F) FPUF-10BDMPP-

10MA, (G) FPUF-15BDMPP-5MA, (H) FPUF-5BDMPP-15EG, (I) FPUF-10BDMPP-10EG, (J) FPUF-15BDMPP-5EG

4CHAN ET AL.

ranges from 33.7 to 42 kg m

3

, which is comparable to the FPUF.

FPUF-20BDMPP exhibits the highest density among the FPUFs with

other single flame retardants. For the FPUFs containing two flame retar-

dants, FPUF-10BDMPP-10EG shows the highest density. FPUF-

15BDMPP-5MA and FPUF-15BDMPP-5EG show the lowest density.

Additive flame retardants often show an adverse effect on the density of

foams. Here in this study, no significant change occurred in the density

of the foams.

The mechanical properties of FPUFs were evaluated by measur-

ing their tensile strength, elongation at break, and compression. The

data are listed in Table 3. Apart from foam density, the additives

themselves influence the mechanical properties. The value for the

tensile strength of FPUF-20EG is the highest of all the samples,

which means the EG improved the tensile strength of FPUF. This

effect of EG is also obvious in FPUF-5BDMPP-15EG and FPUF-

10BDMPP-10EG. Among the FPUFs with 20-phr of a single flame

retardant, FPUF-20MA has the lowest tensile strength and elonga-

tion at break, but still has good mechanical properties. The mela-

mine particles weakened the structurebystiffeningthecellular

network.

ThiseffectofMAcanalsobeobservedinFPUF-

5BDMPP-15MA and FPUF-10BDMPP-10MA. For the FPUFs with

the combination of two flame retardants, the incorporation of

15-phr BDMPP with either 5-phr MA or 5-phr EG in FPUF signifi-

cantly reduces tensile strength and compression stress/

straincharacteristic at 40% compression (CV

) because of the lower

density and detrimental effects to the mechanical properties caused

by the two flame retardants. Apart from FPUF-15BDMPP-5MA and

FPUF-15BDMPP-5EG, all the other seven foams showed very simi-

lar mechanical properties compared to FPUF. Thus, the mechanical

properties of the foam depend on the amount and the type of flame

retardant added.

3.2 |Pyrolysis: Mass loss

The thermal decomposition behavior of the FPUFs was investigated.

The pyrolysis of organic content typically generates volatile products

and leaves mostly carbonaceous char as residue. TGA measured the

mass loss of the FPUFs during the pyrolysis process under a nitrogen

atmosphere. Figure 3 shows the mass and the first derivative of mass

loss (DTG) curves, and Table 4 records the selected data. T

and T

max

are the temperatures where 5-wt% mass loss and maximum mass loss

occurred, respectively. All FPUFs went through two distinguishable

decomposition steps based on the chemical structure of the FPUF.

There is no additional separated minor decomposition step happening

at lower temperatures, an early decomposition or release of BDMMP

were ruled out. The first weight loss is related to the urethane bond's

cleavage, while the second step is attributed to the decomposition of

the hydrocarbon chains.

19,20

FUPF-20BDMPP shows the lowest T

at approximately 220C, which is due to the interaction between

polyurethane decomposition and the phosphorous compounds of

BDMPP decomposition.

2,21

Among the FPUFs incorporated with a

sole retardant, 8.2 and 7.5-wt% of char yield were measured for

FPUF-20BDMPP and FPUF-20EG, respectively, while there was no

significant degree of charring for FPUF-20MA. It is evident that

BDMPP and EG worked in the condensed phase, while MA did not.

Phosphorous compounds enhanced carbonization, while EG yielded

intumescence and remained in the crucible because EG cannot evapo-

rate during thermal decomposition.

22,23

The amount of char residue

of all FPUF-BDMPP-EGs is greater than that of either FPUF-

20BDMPP or FPUF-20EG. The synergistic effect caused by the com-

bination of BDMPP and EG was attributed to the phosphorous flame

retardant, which generates char-forming catalysts, increasing the char

yield during thermal decomposition.

10,24

FPUF-20EG shifted the first

and second decomposition steps towards the highest temperature

max

#1 =313C and T

max

#2 =383C) among all the samples, which

indicates that 20-phr EG enhanced the thermal stability of the system.

3.3 |Pyrolysis: Evolved gas analysis

The evolved gaseous products during thermal decomposition under

nitrogen atmosphere were determined using TG-FTIR. The TG-FTIR

were performed to characterize how the single flame retardant work,

TABLE 3 Mechanical test results of the samples

Sample Tensile strength (kPa) Elongation at break (%)

Compression stress valueat

40% compression (CV

) (kPa) Apparent density (kg m

3

)

FPUF 126 ± 13 138 ± 9 6.61 ± 0.13 34.2 ± 0.4

FPUF-20BDMPP 140 ± 16 140 ± 16 6.37 ± 0.21 40.1 ± 2.2

FPUF-20MA 102 ± 13 91 ± 15 6.34 ± 0.12 34.4 ± 0.7

FPUF-20EG 156 ± 9 111 ± 11 6.08 ± 0.20 35.9 ± 2.3

FPUF-5BDMPP-15MA 94 ± 11 87 ± 16 6.67 ± 0.87 37.9 ± 2.3

FPUF-10BDMPP-10MA 93 ± 10 96 ± 11 5.30 ± 0.29 40.2 ± 0.5

FPUF-15BDMPP-5MA 54 ± 9 104 ± 14 1.99 ± 0.19 33.8 ± 0.7

FPUF-5BDMPP-15EG 139 ± 4 151 ± 4 5.98 ± 0.75 38.9 ± 3.4

FPUF-10BDMPP-10EG 112 ± 4 145 ± 8 4.22 ± 0.09 42.0 ± 8.8

FPUF-15BDMPP-5EG 54 ± 14 83 ± 17 2.84 ± 0.53 33.7 ± 0.9

CHAN ET AL.5

particular which one is releasing phosphorus in the gas phase. Apart

from that, the limited change in TG curves yields evolved gas analysis

being of minor importance for understanding the fire behavior. As

shown in Figure 4A, some characteristic bands were detected during

the pyrolysis process. The noisy signals in the ranges of 2150–1250

and 4000–3400 cm

1

are related to the water vapor produced during

thermal decomposition. The peaks of CO

(2360 and 670 cm

1

) were

observed.

The peak at 2276 cm

1

is attributed to the stretching

vibration of N═C═O. From about 340C, the transmittance intensity

at 2276 cm

1

of NCO disappeared. This indicates that N═C═Ois

the main product generated in the first stage of decomposition.

26,27

The broad peaks at 3000–2850 cm

1

at 340 to 440C are attributed

to hydrocarbons.

Therefore, the major products in the second stage

are hydrocarbons. The peaks at 1734 and 1363 cm

1

correspond to

FIGURE 3 Thermogravimetry-(A) mass curves and (B) DTG curves of FPUF-BDMPP-MAs; (C) mass curves and (D) DTG curves of FPUF-

BDMPP-EGs

TABLE 4 Selected thermogravimetry results obtained from the mass and DTG curves of FPUFs

Material T

(C) T

max

#1 (C) T

max

#2 ( C) Mass change #1 (wt%) Mass change #2 (wt%) Residue at 700C (wt%)

FPUF 263 297 380 31.2 65.7 1.6

FPUF-20BDMPP 220 288 360 29.3 60.5 8.2

FPUF-20MA 260 290 381 34.7 62.7 2.0

FPUF-20EG 265 313 383 27.4 64.4 7.5

FPUF-5BDMPP-15MA 252 284 379 32.5 62.2 5.7

FPUF-10BDMPP-10MA 242 282 369 31.6 61.0 6.9

FPUF-15BDMPP-5MA 229 281 362 30.2 60.8 7.9

FPUF-5BDMPP-15EG 251 287 372 27.1 60.4 11.8

FPUF-10BDMPP-10EG 238 295 356 29.3 59.2 11.6

FPUF-15BDMPP-5EG 228 289 356 29.3 59.9 9.9

Abbreviations: T

, the temperature at 5% mass loss; T

max

#1, the first maximum mass loss rate; T

max

#2, The second maximum mass loss rate.

6CHAN ET AL.

the stretching vibration of the carbonyl compound.

The peaks at

1272 and 1110 cm

1

are attributed to C O stretching. Figure 4B

shows that the peaks at 1218 and 880 cm

1

are attributed to P═O

and P O, respectively.

The phosphorous moieties are released

mainly during the second stage of decomposition. As a result, the frag-

ments composed of phosphorus in the gas phase provide a radical

quenching effect during combustion. As shown in Figure 4C, the peak

at 1660 cm

1

is attributed to the triazine ring of melamine.

The

nitrogen from MA produced by combustion acts as an inert diluent,

diluting the fuel gases. There are smaller peaks in Figure 4D at 2400–

2300 cm

1

and 670 cm

1

related to CO

in the range from 383 to

400C. This is probably due to the presence of EG, which may pro-

duce more stable char residue to change the CO

release. In conclu-

sion, the results agree with the decomposition pathway of FPUF

described in the literature.

3.4 |Fire behavior: Reaction to the small flame

The limiting oxygen index (LOI) is a measure of the minimum percent-

age of oxygen in a mixture of oxygen and nitrogen gases required to

support the combustion of materials in a candle-like setup. Table 5

lists the LOI value of the samples. All the FPUFs with flame retardants

show improvement in the LOI values. The improvements are some-

what limited prospecting for further development. Nevertheless, con-

sidering that foams were investigated and there is no significant

difference in morphology and density the increase in LOI is assessed

to be meaningful. When 20-phr BDMPP, MA, or EG was added to

FPUF, the LOI values increase, and their values are very similar: 20.0,

20.4, and 20.8 vol%, respectively. However, FPUFs with a single flame

retardant exhibit typical LOI values of foams, limiting flame retardancy

to FPUF. The LOI value of FPUFs containing different proportions of

BDMPP and MA remains the same as those with single flame retar-

dants. Nevertheless, it is notable that FPUF-5BDMPP-15EG achieved

the highest LOI value among all the samples. With an increase of 3.6

to 22.2 vol%. FPUF-10BDMPP-10EG shows the second highest LOI

value, 21.8 vol%. The higher EG content in FPUF-BDMPP-EGs pro-

vides the polymer with better fire protection, as observed in this case.

Therefore, the optimal ratio of BDMPP to EG in FPUF-BDMPP-EGs

that produces the greatest synergy is 1: 3.

The results in Table 6 show that all the modified samples exhibit

different degrees of improvement to flame retardancy in the UL

FIGURE 4 TG-FTIR spectra of the gas phase in the thermal degradation of (A) FPUF, (B) FPUF-20BDMPP, (C) FPUF-20MA, and (D) FPUF-

20EG at different pyrolysis temperatures

CHAN ET AL.7

94 HBF test as compared to FPUF. FPUF burned entirely with the

fastest burning rate and significant dripping behavior. The burning

rate of all the foams with 20-phr of either a single or dual flame

retardant(s) decreased compared to that of FPUF (120 mm/min). The

EG contributed effectively to slowing down the burning rate. Due to

the formation of a dense carbon layer, all samples with EG completely

stopped the melt dripping. The expanded graphite served as a carrier

to retain the polymer melt. Compared to FPUF-20EG, FPUF-

5BDMPP-15EG exhibited a lower burning rate and exhibited self-

extinguishing behavior, even though less EG was added. This is

because the synergistic effect was exerted by BDMPP and EG. The

gluing effect of BDMPP strengthened the integrality and continuity of

the EG char layer.

32,33

The burning rate of FPUF-5BDMPP-15MA was

lower than that of FPUF-20BDMPP and FPUF-20MA. This indicates

that 5-phr of BDMPP with either 15-phr of EG or 15-phr MA in the

FPUF is enough to yield a remarkable synergistic effect in flame

retardancy.

3.5 |Fire behavior: Cone calorimeter

A cone calorimeter is used to evaluate a comprehensive set of fire

properties such as time of ignition (t

), peak heat release rate (PHRR),

total heat release (THR), average effective heat of combustion (EHC),

residue, total smoke released (TSR) and the maximum average rate of

heat emission (MARHE) in the fire scenarios of developing fires. Heat

release rate (HRR) and THR curves of FPUF-BDMPP-MAs and FPUF-

BDMPP-EGs are displayed in Figure 5. The measured data are pres-

ented in Table 7.

Typically, the HRR curve of FPUF consists of the two peaks asso-

ciated with two-step decomposition, which is concluded from the

result of TGA in accordance with the literature.

34,35

The first decom-

position step corresponds to the breaking of the urethane bond in PU,

while the soft segment dominates the second step. The HRR curve of

FPUF shows two peaks of heat release rate, where the second peak

(pHRR at 503 kW m

2

) is higher than that of the first peak (pHRR at

294 kW m

2

). The t

is usually very short because of the low heat

conductivity of FPUFs. At the first peak, the material was ignited, and

the cellular structure started to collapse, thus producing volatile and

liquid fragments. These liquid fragments produced more heat by oxi-

dation and quickly developed a feedback loop. Subsequently, this

feedback loop at the second peak with a high HRR formed a pool fire,

and the sample burned intensively. After that, as the fuel was con-

sumed, HRR dropped rapidly until the flame went out.

36,37

Figure 5

(A1) shows that the HRR curves of the FPUFs with flame retardant(s)

are similar to those of FPUF, with two explicit HRR peaks and the sec-

ond peak higher than the first.

The EHC monitored in the cone calorimeter is a product of the

effective heat of combustion of the volatiles and the combustion effi-

ciency of the flame. The fuel dilution, reducing the effective heat of

combustion of the volatiles, and flame inhibition, reducing mainly the

combustion efficiency in the flame, reduce the EHC. Therefore, EHC

is an important parameter to measure the activity of flame retardants

in the gas phase. A reduction in EHC is observed for all flame retarded

samples in Table 7, indicating that all flame retardants exerted differ-

ent degrees of flame retardant effects in the gas phase. Among the

FPUFs containing a single flame retardant, the best result in terms of

EHC was achieved by FPUF-20BDMPP (22 MJ kg

1

). The EHC of

FPUF-20BDMPP reduces by more than 14% when compared to the

that of FPUF. This showed that the BDMPP plays an important role

as a flame retardant through flame inhibition in the gas phase.

TABLE 5 LOI measurement

Sample LOI/vol%

FPUF 18.6 ± 0.2

FPUF-20BDMPP 20.0 ± 0.2

FPUF-20MA 20.4 ± 0.1

FPUF-20EG 20.8 ± 0.2

FPUF-5BDMPP-15MA 20.0 ± 0.2

FPUF-10BDMPP-10MA 20.2 ± 0.1

FPUF-15BDMPP-5MA 20.6 ± 0.1

FPUF-5BDMPP-15EG 22.2 ± 0.2

FPUF-10BDMPP-10EG 21.8 ± 0.2

FPUF-15BDMPP-5EG 20.6 ± 0.1

TABLE 6 Result of UL 94 horizontal burning tests

Material Burning time (s) Distance burned (mm) Burning drops Burning rate (mm/min)

FPUF 50 100 Yes 120

FPUF-20BDMPP 63 100 Yes 95

FPUF-20MA 108 100 Yes 56

FPUF-20EG 46 15 No 20

FPUF-5BDMPP-15MA 131 100 Yes 46

FPUF-10BDMPP-10MA 69 100 Yes 87

FPUF-15BDMPP-5MA 70 100 Yes 86

FPUF-5BDMPP-15EG 34 5 No 9

FPUF-10BDMPP-10EG 205 100 No 29

FPUF-15BDMPP-5EG 72 100 No 83

8CHAN ET AL.

concluded from the evolved gas analysis, radical scavenging occurred

because phosphorus was released from BDMPP.

The presence of

EG also showed a significant reduction in EHC. We hypothesize three

contributions that could explain this phenomenon. EG created a pro-

tection layer, which caused incomplete pyrolysis in the second stage

of burning. The second stage of burning generally corresponds mainly

to the second step of pyrolysis. In PU the second stage of burning is

usually characterized by a higher EHC.

Reducing the contribution of

the second stage of burning results in a reduced contribution of the

second decomposition step to the pyrolysis products, and thus

reduces EHC. The second reason is the strongly increased charring in

FPUF-20EG,which means that mainly graphitized carbon is stored

with a higher effective heat of combustion than the PU. Thus, in the

case of PU, increased charring goes along with emitting volatiles with

a lower EHC than PU.

The last minor reason is that EG is treated

with sulfuric acid as an intercalation reagent. During burning, the oxi-

dation reaction of H

releases inert gases such as CO

,SO

and,H

O, which dilute the combustible gas.

16,40,41

Compared to other

flame retardants, MA only shows a small effect on EHC through some

fuel dilution. The FPUFs with the combination of BDMPP and EG,

FIGURE 5 (A1) Heat release rate and (A2) total heat release rate of FPUF-BDMPP-MAs; (B1) heat release rate and (B2) total heat release rate

of FPUF-BDMPP-EGs

TABLE 7 Table of cone calorimeter results

Sample

/s ± 1 s

PHRR

/kW m

2

THR

/MJ m

2

TML/g

Av.

EHC/MJ kg

1

Residue/

wt%

TSR /m

2

MARHE/

kW m

2

FPUF 6 503 ± 20 43.3 ± 0.1 16.9 ± 0.4 25.6 ± 0.6 0 ± 0.3 392 ± 4 320 ± 13

FPUF-20BDMPP 7 586 ± 52 42.3 ± 0.8 19.2 ± 0.8 22.0 ± 0.5 2.9 ± 1.9 946 ± 132 325 ± 7

FPUF-20MA 4 391 ± 20 40.3 ± 0.9 16.6 ± 0.2 24.3 ± 0.3 2.2 ± 0.8 299 ± 2 276 ± 11

FPUF-20EG 5 183 ± 19 16.2 ± 4.5 6.8 ± 1.7 23.6 ± 0.6 63.1 ± 7.7 52 ± 26 109 ± 10

FPUF-5BDMPP-15MA 5 523 ± 5 41.8 ± 0.9 18.2 ± 0.2 23.0 ± 0.1 4.6 ± 0.1 641 ± 22 310 ± 13

FPUF-10BDMPP-10MA 5 564 ± 18 43.1 ± 0.8 18.8 ± 0.2 22.9 ± 0.2 5.3 ± 0.4 757 ± 8 314 ± 8

FPUF-15BDMPP-5MA 4 559 ± 24 35.6 ± 0.4 16.1 ± 0.4 22.1 ± 0.3 4.7 ± 0.4 789 ± 36 328 ± 9

FPUF-5BDMPP-15EG 4 213 ± 27 22.4 ± 5.8 10.3 ± 2.0 21.4 ± 1.4 47.4 ± 4.7 204 ± 73 130 ± 17

FPUF-10BDMPP-10EG 5 252 ± 23 40.4 ± 8.0 18.3 ± 3.6 22.0 ± 0.1 12.3 ± 1.8 987 ± 291 173 ± 15

FPUF-15BDMPP-5EG 4 346 ± 42 31.6 ± 0.7 15.7 ± 0.2 20.1 ± 0.2 3.8 ± 1.5 998 ± 37 280 ± 28

Abbreviations: av, average; EHC, effective heat of combustion; MARHE, maximum average rate heat emission; PHRR, peak heat release rate; t

, time to

ignition; THR, total heat release; TML, total mass loss; TSR, total smoke release.

CHAN ET AL.9

especially FPUF-15BDMPP-5EG (20.1 MJ kg

1

), exhibit significant

reduction in EHC values. The EHC of FPUF-15BDMPP-5EG reduced

more than 21% compared with that of FPUF. As can be seen from this

phenomenon, there was a synergistic effect between BDMPP and EG

in the gas phase.

The pHRR of FPUF-20MA was reduced by 22% because MA acts

as a heat sink to increase the heat capacity of the system and limits

the increase in surface temperature of FPUF. Hence, MA reduces the

generation rate of volatile fuel and decreases the pHRR effectively.

Considering Figure 5(A2), the burning time and THR of FPUF-

15BDMPP-5MA are reduced significantly compared to those of either

FPUF-20BDMPP or FPUF-20MA. Therefore, a synergistic effect was

observed between 15-phr BDMPP and 5-phr MA in FPUF.

Even with the low loadings of EG presented in the system, the

shape of the HRR curve was clearly changed. The addition of higher EG

content led to a profound reduction in the PHRR, as the peaks became

significantly smaller, with the curves much flatter. EG provides for sig-

nificant charring, leading to the formation of a thermal insulating barrier

which slows down the transfer of heat and mass within the pyrolysis

zone for further decomposition in the condensed phase. The expanded

graphite increases the thickness of the char layer to prolong the time of

burning. This can be explained in detail with the cone calorimeter data

FIGURE 6 Fire residue

images of (A) FPUF, (B) FPUF-

20BDMPP, (C) FPUF-20MA,

(D) FPUF-20EG, (E) FPUF-

5BDMPP-15MA, (F) FPUF-

10BDMPP-10MA, (G) FPUF-

15BDMPP-5MA, (H) FPUF-

5BDMPP-15EG, (I) FPUF-

10BDMPP-10EG, (J) FPUF-

15BDMPP-5EG

10 CHAN ET AL.

and the fire behavior of FPUF-20EG. At the beginning of the curve, the

polymer matrix was under heat exposure, and thus the HRR increased

swiftly. After the first sharp peak, the HRR dropped quickly because

the EG expanded underheat, and acted as an excellent protective layer

to save the material underneath.

THR is a measure of the entire amount of heat energy evolved

during the burning time of the material. The THR decreased drastically

from 43.3 MJm

2

for the non-flame retarded foam to 16.2 MJ m

2

for FPUF-20EG. The reduced value indicates that the expanded

graphite formed protective layer, providing an excellent shielding

FIGURE 7 (A1) Surface and (A2) side

view of FPUF-20BDMPP; (B1) surface

and (B2) side view of FPUF-20MA;

(C1) surface and (C2) side view of FPUF-

5BDMPP-15MA; (D1) surface and

(D2) side view of FPUF-10BDMPP-

10MA; (E1) surface and (E2) side view of

FPUF-15BDMPP-5MA

CHAN ET AL.11

effect to the material underneath and leading to incomplete combus-

tion. As combined with the results from the EHC and the yield of resi-

due, the charring effect of EG in the condensed phase caused the

major reduction in THR.

MARHE is one of the fire hazard indices of developing fires under

a real-scale fire scenario. This is used to determine the combustibility

of a material. The MARHE value of FPUF-20EG is reduced remark-

ably, to almost onethird that of the non-flame retarded foam. The

FPUF-BDMPP-MAs show no changes in MARHE, but the FPUF-

BDMPP-EGs lowered the value significantly. Based on the results for

MARHE, EG is apparently an effective additive to reduce the MARHE.

The TSR of FPUF-20BDMPP is 946 m

2

, which indicates that

BDMPP generates a large amount of smoke. This value is nearly 2.5

times that of FPUF. FPUF-20MA and FPUF-20EG greatly suppressed

the smoke. Melamine acts as an inert diluent in the gas phase to

reduce smoke emission.

D. Price et al. showed a chemical interac-

tion between the melamine and the isocyanate at temperaturesover

250C through the reaction between NH

and NCO. This interac-

tion suppressed the smoke produced from isocyanate.

The great

charring ability of EG produced a compact carbonaceous char that

could limit the release of aromatic hydrocarbons to form smoke from

the condensed phase into the gas phase.

3.6 |Fire residues

Figure 6 shows the images for the char residue of FUPFs after cone

calorimeter measurement. FPUF in Figure 6A was consumed

completely after burning and almost no residue remained. Figure 6B,

C,E, F,G display a thin layer of fragile inorganic residue that remained

in the aluminum foil tray. The micrographs of the FPUFs with EG

(Figure 6D, H, I, J) show that the char layer formed during burning

provided a thermal insulation barrier to protect the inner polymer

matrix and to prevent further decomposition. According to Figure 6H,

I, J, the integrity of expanded graphite residue was retained due to the

presence of a sufficient amount of BDMPP.

The gluing effect by the

phosphorous compound reinforced the integrity and continuity of the

char layers, resulting in an enhanced barrier formed in the condensed

phase.

32,45

Figure 7 shows SEM images of the fire residue of FPUF-

BDMPP20, FPUF-20MA, FPUF-5BDMPP-15MA, FPUF-10BDMPP-

10MA, and FPUF-15BDMPP-5MA. FPUF-20BDMPP produced an

intact and dense char residue which acted as a protective layer to the

materials underneath during burning. FPUF-20MA resulted in a thin,

layered residue with more holes on the surface. When 5-phr of MA

from FPUF-20MA was replaced with 5-phr BDMPP, fire residue chan-

ged significantly on the surface and in the crosssection. Interestingly,

FPUF-5BDMPP-15MA has a bumpy surface with a random size of

holes and bubbles. Both FPUF-10BDMPP-10MA and FPUF-

15BDMPP-5MA show a closed char surface. FPUF-10BDMPP-10MA

exhibits a layer of tiny, compacted bubbles in its crosssection, while

FPUF-15BDMPP-5MA consists of multiple layers with small bubbles

and holes. The layered structure exhibited excellent protection against

fire during burning.

Hence, the THR of FPUF-15BDMPP-5MA was

significantly reduced.

4|CONCLUSIONS

In this work, a set of flame retarded FPUF samples was prepared to

understand the interaction between BDMPP and MA/EG. In each

flame retarded sample, the total amount of additives was 20 phr.

From the result of LOI and UL 94 HBF tests, all flame retarded sam-

ples showed reduced flammability and a lower burning rate. In the

systems with a single flame retardant, both 20-phr BDMPP and

20-phr EG enhanced flame retardancy in the gas phase. 20-phr MA

and 20-phr EG reduced the pHRR significantly. EG is a great smoke

suppressant, according to the result of TSR from the cone calorimeter.

FPUF-20EG produced only 13% of the amount of smoke released by

FPUF. FPUF-20BDMPP and FPUF-20EG exhibited high char yield

after a pyrolysis process. As to the systems with dual flame retardants,

the overall flame retardancy of FPUF-BDMPP-EGs was better than

that of FPUF-BDMPP-MAs. The synergistic effect between BDMPP

and EG, mainly due to BDMPP contributing gluing effect to expanded

graphite, improved the char yield and stopped dripping. Among FPUF-

BDMPP-EGs, FPUF-5BDMPP-15EG showed the best flame ret-

ardancy properties according to the LOI value and the burning rate in

UL 94 HBF. Self-extinguishing behavior was also observed for FPUF-

5BDMPP-15EG from the UL 94 HBF test.

In summary, the thermal pyrolysis and fire performance indicate

that the combination of BDMPP and EG actively improves the fire

behavior of PFUF by synergistic effects in the gas and condensed

phases.

ACKNOWLEDGEMENTS

The authors appreciate the contribution of the mechanical tests from

D. Schulze. This project was financed by the DFG (Deutsche

Forschungsgemeinschaft) (SCHA 730/19-1) and the NSFC (National

Natural Science Foundation of China) (51761135113).

Open access funding enabled and organized by Projekt DEAL.

CONFLICT OF INTEREST

The authors declare no potential conflict of interest.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the

corresponding author upon reasonable request.

ORCID

Yuan Hu https://orcid.org/0000-0003-0753-5430

Bernhard Schartel https://orcid.org/0000-0001-5726-9754

REFERENCES

1. Jiang L, Ren Z, Zhao W, Liu W, Liu H, Zhu C. Synthesis and

structure/properties characterizations of four polyurethane model

hard segments. R Soc Open Sci. 2018;5(7):180536.

12 CHAN ET AL.

2. Günther M, Levchik SV, Schartel B. Bubbles and collapses: fire phe-

nomena of flame-retarded flexible polyurethane foams. Polym Adv

Technol. 2020;31(10):2185-2198.

3. Günther M, Lorenzetti A, Schartel B. Fire phenomena of rigid polyure-

thane foams. Polymers. 2018;10(10):1166.

4. Günther M, Lorenzetti A, Schartel B. From cells to residues: flame-

retarded rigid polyurethane foams. Combust Sci Technol. 2020;

192(12):2209-2237.

5. Lefebvre J, Le Bras M, Bastin B, Paleja R, Delobel R. Flexible polyure-

thane foams: flammability. J Fire Sci. 2003;21(5):343-367.

6. Weil ED, Levchik SV. Flame Retardants for Plastics and Textiles Practi-

cal Applications. Hanser Publishers; 2016.

7. Lorenzetti A, Besco S, Hrelja D, et al. Phosphinates and layered sili-

cates in charring polymers: the flame retardancy action in polyure-

thane foams. Polym Degrad Stabil. 2013;98(11):2366-2374.

8. Schartel B, Wilkie CA, Camino G. Recommendations on the scientific

approach to polymer flame retardancy: part 2 Concepts. J Fire Sci.

2017;35(1):3-20.

9. Wilke A, Langfeld K, Ulmer B, et al. Halogen-free multicomponent

flame retardant thermoplastic styrene-ethylene-butylene-styrene

elastomers based on ammonium polyphosphate-expandable graphite

synergy. Ind Eng Chem Res. 2017;56(29):8251-8263.

10. Rao W-H, Liao W, Wang H, Zhao H-B, Wang Y-Z. Flame-retardant

and smoke-suppressant flexible polyurethane foams based on reac-

tive phosphorus-containing polyol and expandable graphite. J Hazard

Mater. 2018;360:651-660.

11. Feng F-F, Qian L-J. The flame retardant behaviors and synergistic

effect of expandable graphite and dimethyl Methylphosphonate in

rigid polyurethane foams. Polym Compos. 2014;35(2):301-309.

12. Yuan Y, Yang H, Yu B, et al. Phosphorus and nitrogen-containing

polyols: synergistic effect on the thermal property and flame Ret-

ardancy of rigid polyurethane foam composites. Ind Eng Chem Res.

2016;55(41):10813-10822.

13. Gaan S, Sun G, Hutches K, Engelhard M-H. Effect of nitrogen addi-

tives on flame retardant action of tributyl phosphate: phosphorus-

nitrogen synergism. Polym Degrad Stabil. 2008;93(1):99-108.

14. Zhou F, Zhang T, Zou B, et al. Synthesis of a novel liquid phosphorus-

containing flame retardant for flexible polyurethane foam: combus-

tion behaviors and thermal properties. Polym Degrad Stabil. 2020;171:

109029.

15. Zhang L-Q, Zhang M, Zhou Y-H, Hu L-H. The study of mechanical

behavior and flame retardancy of castor oil phosphate-based rigid

polyurethane foam composites containing expanded graphite and

triethyl phosphate. Polym Degrad Stabil. 2013;98(12):2784-2794.

16. Lorenzetti A, Dittrich B, Schartel B, Roso M, Modesti M. Expandable

graphite in polyurethane foams: the effect of expansion volume and

intercalants on flame retardancy. J Appl Polym Sci. 2017;134(31):

45173.

17. König A, Fehrenbacher U, Hirth T, Kroke E. Flexible polyurethane

foam with the flame-retardant melamine. J Cell Plast. 2008;44(6):

469-480.

18. Sut A, Metzsch-Zilligen E, Grosshauser M, Pfaendner R, Schartel B.

Rapid mass calorimeter as a high-throughput screening method for

the development of flame-retarded TPU. Polym Degrad Stabil. 2018;

156:43-58.

19. Ugarte L, Saralegi A, Fernandez R, Martin L, Corcuera MA, Eceiza A.

Flexible polyurethane foams based on 100% renewably sourced poly-

ols. Ind Crop Prod. 2014;62:545-551.

20. Yang H-Y, Liu H-Y, Jiang Y-P, Chen M-F, Wan C-J. Density effect on

flame Retardancy, thermal degradation, and combustibility of rigid

polyurethane foam modified by expandable graphite or ammonium

polyphosphate. Polymers. 2019;11(4):668.

21. Wendels S, Chavez T, Bonnet M, Salmeia KA, Gaan S. Recent devel-

opments in organophosphorus flame retardants containing P-C bond

and their applications. Materials. 2017;10(7):784.

22. Ming G, Chen S, Sun Y-J, Wang Y-X. Flame Retardancy and thermal

properties of flexible polyurethane foam containing expanded graph-

ite. Combust Sci Technol. 2017;189(5):793-805.

23. Wang WJ, Dong K-H-Q, Zhu N, et al. Synergistic effect of aluminum

hydroxide and expandable graphite on the flame Retardancy of

Polyisocyanurate-polyurethane foams. J Appl Polym Sci. 2014;131(4):

39936.

24. Yang S, Wang J, Huo S-Q, Wang M, Wang J-P, Zhang B. Synergistic

flame-retardant effect of expandable graphite and phosphorus-

containing compounds for epoxy resin: strong bonding of different

carbon residues. Polym Degrad Stabil. 2016;128:89-98.

25. Luo F-B, Wu K, Lu M-G. Enhanced thermal stability and flame ret-

ardancy of polyurethane foam composites with polybenzoxazine

modified ammonium polyphosphates. RSC Adv. 2016;6(16):13418-

13425.

26. Liu W, Tang Y, Li F, Ge X-G, Zhang Z-J. TG-FTIR characterization of

flame retardant polyurethane foams materials. Iop Conf Ser-Mat Sci.

2016;137:012033.

27. Chen D-M, Zhao Y-P, Yan J-J, Chen L, Dong Z-Z, Fu W-G. Prepara-

tion and properties of halogen-free flame retardant polyurethane

foams. Adv Mater Res-Switz. 2012;418–420:540-543.

28. Lin F, Lin H, Ke J, Liu J, Bai X, Chen D. Preparation of reactive and

additive flame retardant with different oxidation state of phosphorus

on the thermal and flammability of thermoplastic polyurethane. Com-

pos Mater. 2019;3(2):43-53.

29. Zhu H, Xu S-A. Preparation of flame-retardant rigid polyurethane

foams by combining modified melamine-formaldehyde resin and

phosphorus flame retardants. ACS Omega. 2020;5(17):9658-9667.

30. Ravey M, Pearce EM. Flexible polyurethane foam. I. Thermal decom-

position of a polyether-based, water-blown commerical type of flexi-

ble polyurethane foam. J Appl Polym Sci. 1997;63(1):47-74.

31. Lewin M. Synergistic and catalytic effects in flame retardancy of poly-

meric materials—an overview. J Fire Sci. 1999;17(1):3-19.

32. Liu D-Y, Zhao B, Wang J-S, Liu P-W, Liu Y-Q. Flame retardation and

thermal stability of novel phosphoramide/expandable graphite in rigid

polyurethane foam. J Appl Polym Sci. 2018;135(27):46434.

33. Hu Y, Zhou Z, Li S, Yang D, Zhang S, Hou Y. Flame retarded rigid poly-

urethane foams composites modified by aluminum

diethylphosphinate and expanded graphite. Front Mater. 2021;7:471.

34. Sut A, Metzsch-Zilligen E, Grosshauser M, Pfaendner R, Schartel B.

Synergy between melamine cyanurate, melamine polyphosphate and

aluminum diethylphosphinate in flame retarded thermoplastic poly-

urethane. Polym Test. 2019;74:196-204.

35. Rabe S, Chuenban Y, Schartel B. Exploring the modes of action of

phosphorus-based flame retardants in polymeric systems. Materials.

2017;10(5):455.

36. Baguian AF, Ouiminga SK, Longuet C, et al. Influence of density on

foam collapse under burning. Polymers. 2021;13(1):13.

37. Kramer RH, Zammarano M, Linteris GT, Gedde UW, Gilman JW. Heat

release and structural collapse of flexible polyurethane foam. Polym

Degrad Stabil. 2010;95(6):1115-1122.

38. Schartel B. Phosphorus-based flame Retardancy mechanisms-old hat

or a starting point for future development? Materials. 2010;3(10):

4710-4745.

39. Lorenzetti A, Modesti M, Gallo E, Schartel B, Besco S, Roso M. Syn-

thesis of phosphinated polyurethane foams with improved fire behav-

ior. Polm Degrad Stabil. 2012;97(11):2364-2369.

40. Modesti M, Lorenzetti A, Simioni F, Camino G. Expandable graphite

as an intumescent flame retardant in polyisocyanurate-polyurethane

foams. Polym Degrad Stabil. 2002;77(2):195-202.

41. Camino G, Duquesne S, Delobel R, Eling B, Lindsay C, Roels T. Mech-

anism of expandable graphite fire retardant action in polyurethanes.

Abstr Pap Am Chem S. 2000;220:U333-U334.

42. Price D, Liu Y, Milnes GJ, Hull R, Kandola BK, Horrocks AR. An inves-

tigation into the mechanism of flame retardancy and smoke

CHAN ET AL.13

suppression by melamine in flexible polyurethane foam. Fire Mater.

2002;26(4–5):201-206.

43. Xu Q-W, Zhai H-M, Wang G-J. Mechanism of smoke suppression by

melamine in rigid polyurethane foam. Fire Mater. 2015;39(3):271-282.

44. G

omez-Fernández S, Günther M, Schartel B, Corcuera MA, Eceiza A.

Impact of the combined use of layered double hydroxides, lignin and

phosphorous polyol on the fire behavior of flexible polyurethane

foams. Ind Crop Prod. 2018;125:346-359.

45. Li J, Mo X-H, Li Y, Zou H-W, Liang M, Chen Y. Influence of expand-

able graphite particle size on the synergy flame retardant property

between expandable graphite and ammonium polyphosphate in semi-

rigid polyurethane foam. Polym Bull. 2018;75(11):5287-5304.

46. Rao W-H, Hu Z-Y, Xu H-X, et al. Flame-retardant flexible polyure-

thane foams with highly efficient melamine salt. Ind Eng Chem Res.

2017;56(25):7112-7119.

How to cite this article: Chan YY, Ma C, Zhou F, Hu Y,

Schartel B. A liquid phosphorous flame retardant combined

with expandable graphite or melamine in flexible polyurethane

foam. Polym Adv Technol. 2021;1-14. doi:10.1002/pat.5519

14 CHAN ET AL.

4.4 Flame retardant flexible polyurethane foams based on phosphorous soybean-oil polyol and

expandable graphite

Yin Yam Chan, Chao Ma, Feng Zhou, Yuan Hu and Bernhard Schartel. Polym Degrad Stabil. 2021, 191.

Doi: 10.1016/j.polymdegradstab.2021.109656.

Status: This article was accepted and published. https://doi.org/10.1016/j.polymdegradstab.2021.109656

First author contribution:

• Conceptualization

• Methodology

• Validation

• Investigation

• Data curation

• Writing – Original draft, review and editing

• Visualization

• Project administration

Contributions from other authors:

• Chao Ma

o Conceptualization

o Investigation

o Resources

• Feng Zhou

o Conceptualization

o Investigation

• Yuan Hu

o Conceptualization

o Resources

o Supervision

o Project administration

o Funding acquisition

• Bernhard Schartel

o Conceptualization

o Methodology

o Resources

o Writing – Original draft, review and editing

o Supervision

o Project administration

o Funding acquisition

Abstract

A phosphorous soybean-oil-based polyol was derived via epoxidation and ring opening reaction as an

alternative to petrochemical-based polyol for the synthesis of flexible polyurethane foams (FPUFs). 5-wt.%

and 10-wt.% of expandable graphite (EG) were added to further improve flame retardancy. The mechanical

properties (tensile strength and compression stress) of the foams were investigated. Thermogravimetric

analysis (TGA) coupled with Fourier-transform infrared (FTIR) were conducted to evaluate the pyrolysis;

limiting oxygen index (LOI), UL 94 and cone calorimeter were performed to analyze the fire performance

of the foams; smoke density chamber was used to investigate the smoke released during burning. When 10-

wt.% of EG was used, the flame retardancy of the foams was much enhanced due to the synergistic effect

between phosphorus and EG. The char yield was three times higher (54-wt.%). The fire load MARHE

approached 100kWm-2, half of the value expected for a superposition. The combination of phosphorous

polyols and EG is proposed as strategy for future flame retarded PFUFs.

Polymer Degradation and Stability 191 (2021) 109656

Contents lists available at ScienceDirect

Polymer Degradation and Stability

journal homepage: www.elsevier.com/locate/polymdegradstab

Flame retardant ﬂexible polyurethane foams based on phosphorous

soybean-oil polyol and expandable graphite

Yin Yam Chan

, Chao Ma

, Feng Zhou

, Yuan Hu

b , ∗, Bernhard Schartel

a , ∗

Bundesanstalt für Materialforschung und -prüfung (BAM), Unter den Eichen 87, 12205 Berlin, Germany

State Key Laboratory of Fire Science, University of Science and Technology of China, Hefei, Anhui 230027, PR China

a r t i c l e i n f o

Article history:

Received 10 March 2021

Revised 11 June 2021

Accepted 14 June 2021

Available online 18 June 2021

Keywords:

Phosphorous soybean-oil–based polyol

Flexible polyurethane foam

Expandable graphite

Flame retardancy

Smoke measurement

a b s t r a c t

A phosphorous soybean-oil–based polyol was derived via epoxidation and ring opening reaction as an

alternative to petrochemical-based polyol for the synthesis of ﬂexible polyurethane foams (FPUFs). 5-wt.%

and 10-wt.% of expandable graphite (EG) were added to further improve ﬂame retardancy. The mechanical

properties (tensile strength and compression stress) of the foams were investigated. Thermogravimetric

analysis (TGA) coupled with Fourier-transform infrared (FTIR) were conducted to evaluate the pyrolysis;

limiting oxygen index (LOI), UL 94 and cone calorimeter were performed to analyze the ﬁre performance

of the foams; smoke density chamber was used to investigate the smoke released during burning. When

10-wt.% of EG was used, the ﬂame retardancy of the foams was much enhanced due to the synergistic

effect between phosphorus and EG. The char yield was three times higher (54-wt.%). The ﬁre load MARHE

approached 100 kWm

−2

, half of the value expected for a superposition. The combination of phosphorous

polyols and EG is proposed as strategy for future ﬂame retarded FPUFs.

1. Introduction

Polyurethane (PU) presents in different parts of our daily life

due to its versatility. The production of ﬂexible polyurethane foam

(FPUF) is the largest sector of the worldwide PU market, and the

investigation showed that demand for it continues to increase [1] .

FPUF is usually used in upholstered furniture, mattresses, packag-

ing, and all kind of seats in transportation, because of its excellent

properties such as its light weight and high resilience [2] . Most of

the FPUFs available on the market are produced from petroleum

derivatives. However, petroleum is a non-renewable resource. To

reduce the consumption of petroleum, scientists and the public are

shifting their attention toward using alternative raw materials. One

of the appropriate approaches is to use vegetable oil as a substitute

in PU production.

Petroleum-derived polyols and diisocyanates are usually the

main constituents of FPUF for building crosslinked networks. Gen-

erally speaking, petrochemical polyols for FPUF production have a

functionality of 3 and a molecular weight between 30 0 0 and 60 0 0,

with a hydroxyl value (OH value) of 56 - 28 mg KOH/g [3] . The

hydroxyl functional groups react with diisocyanates to form ure-

thane bonds in PU. If there is water used as a blowing agent, di-

∗Corresponding authors.

E-mail addresses: [email protected] (Y. Hu), [email protected] (B.

Schartel).

isocyanates react with the water to generate urea linkages in the

foam. A certain similarity in chemical structure can be found be-

tween vegetable oils and petrochemical polyols. Vegetable oils are

composed of triglycerides, which contain a glycerol with three long

carbon chains of fatty acid. Several methods of synthesizing of bio-

based polyols have been reported in the literature; the polyols de-

rived from vegetable oil such as soybean oil [4–6] , palm oil [ 7 , 8 ],

castor oil [9] and rapeseed oil [10] could potentially replace the

petrochemical ones completely or in part. Hence, modiﬁed veg-

etable oils with the appropriate functionality, molecular weight

and OH value might be a good substitute for producing FPUF.

As compared to other kinds of vegetable oil, soybean oil dom-

inates the market in vegetable oil production [11] , has competi-

tively low costs [12] , and could serve as a raw material for bio-

based industrial products [13] . Soybean oil consists of fatty acids

which can easily be modiﬁed into polyol. These fatty acids are nat-

urally unsaturated and are reactive to form oxirane rings by epoxi-

dation. The hydroxyl groups are introduced by a ring-opening reac-

tion of epoxidized soybean oil. However, soybean-oil–based polyol

has its disadvantages, including a low molecular weight which af-

fects the mechanical properties, and the absence of primary hy-

droxyl groups which results in a slow curing process. Apart from

their sustainability advantages over ingredients from limited non-

renewable resources, using bio-based polyols usually shows bene-

ﬁts for biodegradability and tensile strength [ 3 , 14 ].

https://doi.org/10.1016/j.polymdegradstab.2021.109656

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

FPUFs catch ﬁre easily, mostly because of their low-density

open-cell structure, which boosts ﬁre growth velocity and can

cause detriment to human lives and the economy. Therefore, it is

important to enhance their ﬁre retardancy to reduce the occur-

rence of tragedies [15] . Flame retardants can be generally catego-

rized into two types: reactive type and addition type. Reactive type

ﬂame retardants form covalent bonds with the polymer chain. Ad-

ditive type ﬂame retardant can simply be mixed physically with

polymeric materials. Reactive type ﬂame retardants typically show

better compatibility with polymers and often better mechanical

properties [16–18] . The combination of two or more types of ﬂame

retardant in the polymer may provide synergistic effects, making

overall ﬂame retardancy higher than the sum of the effects of the

single ﬂame retardants [ 19 , 20 ]. Yuan et al. synthesized rigid PU

foam by using phosphorous polyol and nitrogenous polyol [21] . The

combination of phosphorous and nitrogenous ﬂame retardants in

the polymer yields a synergistic effect [22] . This effect improved

the thermal stability and ﬂame retardancy of the rigid PU foam.

Wilke et al. analyzed the synergy between phosphorus ﬂame re-

tardant and EG in a styrene-ethylene-butylene-styrene elastomer

system. They concluded that the phosphorus content inﬂuences

the degree of volume expansion of graphite layer by gluing ﬂuffy

residue together [23–26] .

The aim of this work was to synthesize ﬂame retarded bio-

based FPUF by replacing petrochemical polyol with phosphorous

bio-based polyol. Dimethyl phosphonate was used as a phospho-

rous source to graft on the backbone of soybean oil because the

epoxy groups on epoxidized soybean oil are in the secondary po-

sition. The P-H group of dimethyl phosphonate has high reactiv-

ity to attack the epoxy groups to undergo ring opening reaction

[ 27 , 28 ]. Phosphorus promotes carbonization and inhibits combus-

tion [ 29 , 30 ]. Soybean oil was chemically modiﬁed into phosphorus-

grafted polyol, which then partially substituted the petrochemical

polyol in the formulation of FPUF. The formulation was modiﬁed

and optimized to produce a ﬂame retarded bio-based FPUF with

physical and mechanical properties that are acceptable in compar-

ison to the reference FPUF. Furthermore, EG, which was added to

the polymer matrix physically, contributed to ﬂame retardancy and

smoke suppression. During burning, loose “worm-like” char lay-

ers, developed rapidly by graphite expansion, shielded the material

from heat transfer to the inner matrix and reduced the compo-

nents’ rates of decomposition [30–33] . The morphologies, mechan-

ical properties, pyrolysis, ﬁre behaviors and smoke properties were

studied.

2. Experimental

2.1. Materials

Soybean oil (99%) and dimethyl phosphonate (98%) were pur-

chased from Shanghai Macklin Biochemical Co., Ltd (Shanghai,

China). Hydrogen peroxide 30% aqueous solution ( ≥30.0%), ethyl

acetate ( ≥99.5%) and anhydrous sodium sulfate ( ≥99%) were

obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai,

China). Anhydrous formic acid ( ≥98.5%) was produced by Tian-

jin Guangfu Fine Chemical Research Institute (Tianjin, China).

Polyether polyol (330, hydroxyl value = 56 mg KOH/g, num-

ber average molecular weight = 30 0 0 g/mol, average function-

ality = 3), graft polyether polyol (2045, prepared by the in

situ polymerization of acrylonitrile and styrene in a polyether

polyol, hydroxyl value = 20 mg KOH/g, number average molecu-

lar weight = 8400 g/mol, average functionality = 3), stannous oc-

toate, triethylenediamine (A33, 33%), dichloromethane, silicone sur-

factant and toluene diisocyanate (TDI 80/20, 80:20 mixture of 2,4-

toluene diisocyanate and 2,6-toluene diisocyanate) were supplied

by Jiangsu Lvyuan New Material Co., Ltd., (Jiangsu, China). Deion-

ized water and dichloromethane were used as a chemical blowing

agent and physical blowing agent in the foaming process respec-

tively. EG with an expansion ratio of 150 –200 was supplied by

Qingdao Xingyuan Colloidal Graphite Co., Ltd. (Qingdao, China). All

the chemicals were used without further puriﬁcation.

2.2. Preparation of phosphorous soybean-oil–based polyol

2.2.1. Epoxidation of soybean oil

250 g of soybean oil and 15.2 g of anhydrous formic acid were

poured into a 500 ml round bottomed ﬂask, which was equipped

with a magnetic stirrer and a pressure equalizing dropping fun-

nel ﬁlled with 45.1 g of 30% hydrogen peroxide. Under continu-

ous stirring, the system was heated up to 55 °C in the oil bath.

With agitation, hydrogen peroxide was then gradually added drop

by drop. After all of the hydrogen peroxide had been added, the

reaction was maintained at 65 °C for a further 6 h. After cooling,

the reaction mixture was dissolved in ethyl acetate and washed

several times with deionized water from a separation funnel. The

remaining water in the mixture was removed by adding the excess

amount of anhydrous sodium sulfate. This sodium sulfate was then

ﬁltered out through suction ﬁltration. The ethyl acetate was re-

moved using a rotatory evaporator when the temperature reached

75 °C. The washed product was stored in a vacuum oven overnight

at 70 °C. The synthetic route is illustrated in Fig. 1 .

2.2.2. Opening of oxirane rings in epoxidized soybean oil

150 g of epoxidized soybean oil and 65.7 g of dimethyl phos-

phonate (DMP) were poured into a 500 ml round bottomed ﬂask

which was equipped with a magnetic stirrer. The reaction was

maintained under stirring for 6 h at 80 °C. After cooling, the mix-

ture was dissolved in ethyl acetate and washed several times with

deionized water from a separation funnel. The ethyl acetate was

removed using a rotatory evaporator at 75 °C. The washed prod-

uct was stored in a vacuum oven overnight at 70 °C. The syn-

thetic route is illustrated in Fig. 2 . The OH value of the synthesized

polyol is 90 mg KOH/g and its viscosity is favorable for foaming.

The product is a mixture because it is not possible to conﬁrm that

all of the phosphorus groups attached to speciﬁc locations on the

carbon chain.

2.3. Preparation of ﬂexible polyurethane foams

The FPUFs were prepared using a one pot method, and formed

inside a closed mold 200 mm ×200 mm ×100 mm in size

(length ×width ×thickness) set to a controlled temperature of

60 °C for 20 min. The composition of soybean-oil–based ﬂexible

polyurethane foams is tabulated in Table 1 . The additives consisted

of 0.92 g of silicon oil, 0.15 g of stannous oil, 0.23 g of A-33, 2.91 g

dichloromethane and 3.2 g deionized water. First component A

(the synthesized soybean-oil–based polyol, polyether polyol, graft

polyether polyol and the additives) was stirred in a disposable

plastic cup using a high-speed stirrer at a uniform rate for 3 min.

Then the weighted component B (TDI) was added to the stirred

component A and blended under high speed in 5 s. The mixture

was poured into the mold during the subsequent expansion. After-

ward the foam was cured in an oven at 80 °C for 24 h to com-

plete the polymerization. The obtained foams were designated as

FPUF20, FPUF40, FPUF60 and FPUF80, respectively, designating the

percentage of polyether polyol replaced by phosphorous soybean-

oil–based polyol. It was our goal to increase the intrinsic ﬂame re-

tardancy as well as the amount of renewable starting material of

ﬂame retarded FPUF; nevertheless these materials contains reac-

tive phosphorus ﬂame retardants in terms of California AB 2998.

The FPUFs with an additional 5-wt.% or 10-wt.% of EG were pre-

pared in the same way. The only difference was that the weighted

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 1. Synthesis of partially epoxidized soybean oil.

Table 1

Composition of soybean-oil–based ﬂexible polyurethane foams.

Synthesized polyol (g) Polyether polyol (g) Graft polyether polyol (g) Additives (g) TDI (g) EG (g)

FPUF 0 62.5 20.83 7.41 38.33 0

FPUF20 12.5 50 20.83 7.41 38.33 0

FPUF40 25 37.5 20.83 7.41 38.33 0

FPUF60 37.5 25 20.83 7.41 38.33 0

FPUF80 50 12.5 20.83 7.41 38.33 0

FPUF-5EG 0 62.5 20.83 7.41 38.33 6.43

FPUF20-5EG 12.5 50 20.83 7.41 38.33 6.43

FPUF40-5EG 25 37.5 20.83 7.41 38.33 6.43

FPUF60-5EG 37.5 25 20.83 7.41 38.33 6.43

FPUF80-5EG 50 12.5 20.83 7.41 38.33 6.43

FPUF-10EG 0 62.5 20.83 7.41 38.33 12.85

FPUF20-10EG 12.5 50 20.83 7.41 38.33 12.85

FPUF40-10EG 25 37.5 20.83 7.41 38.33 12.85

FPUF60-10EG 37.5 25 20.83 7.41 38.33 12.85

FPUF80-10EG 50 12.5 20.83 7.41 38.33 12.85

amount of EG listed in Table 1 was ﬁrst incorporated homoge-

nously into the component A through high-speed stirring for 3 min

before foaming.

2.4. Measurements

2.4.1. Characterization of phosphorous soybean-oil–based polyol

Nuclear magnetic resonance (NMR) spectroscopy analysis was

used to determine the molecular structure, the chemical con-

tent, and the purity of chemicals.

H and

P NMR spectra were

recorded by Bruker AVANCE AV 400 (Fällanden, Switzerland) using

deuterated chloroform CDCl

as the solvent.

2.4.2. Morphological characterization

The morphologies of the foams and their char obtained from

the ﬁre test were studied by scanning electron microscopy (SEM).

The specimens were sputter coated with a 15 nm of gold to avoid

electrostatic charging during examination. The electron high ten-

sion (EHT) value was set to 10 kV. The images were obtained from

Zeiss EVO 10 (Oberkochen, Germany).

2.4.3. Foam mechanical and physical properties measurements

The apparent density of the samples was measured in accor-

dance with the ISO 845 standard. For mechanical tests, tensile

strength and elongation at break of the foams were determined

using a Universal Testing Machine Zwick Z010 (Ulm, Germany) fol-

lowing ISO 1798. The compressive strength was measured accord-

ing to ISO 3386-1.

2.4.4. Thermal properties

Thermogravimetric analysis (TGA) measures the change in the

mass of a sample over time as a function of temperature in an

inert atmosphere. TG 209 F1 Iris from Netzsch Instruments (Selb,

Germany) was used as a heating source to measure the thermal

degradation behavior of FPUF. The samples were grinded to a ﬁne

powder using CryoMill from Retsch (Haan, Germany) before test-

ing. 10 mg of powdered sample was weighed and placed in the

aluminum oxide crucible before it was put in the furnace. The

sample was then heated up from 40 °C to 950 °C under nitrogen at

a heating rate of 10 K/min. TGA coupled with Fourier transform in-

frared spectroscopy (FTIR), Brucker Tensor 27 FT-IR (Ettlingen, Ger-

many), allowed the gaseous reaction products produced in the TGA

chamber to be investigated simultaneously.

2.4.5. Fire behavior measurements

All the specimens for ﬁre behavior measurements were condi-

tioned at 23 °C and 50% relative humidity for at least 48 h before

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 2. Synthesis of phosphorous soybean-oil–based polyols.

measuring. Limiting oxygen index (LOI) was determined at ambi-

ent temperature according to ISO 4589-2, and the size of the spec-

imens was 150 mm ×10 mm ×10 mm (length ×width ×thick-

ness). The measurement was done by combusting a specimen

with a mixture of a certain concentration of oxygen and nitro-

gen and then reducing the percentage of oxygen until a crit-

ical level was reached. The burning behavior was character-

ized by cone calorimeter manufactured by Fire Testing Technol-

ogy Limited, United Kingdom. The test was performed follow-

ing the ISO 5660 standard. A specimen with the dimensions of

100 mm ×100 mm ×50 mm (length ×width ×thickness) was

placed in an aluminum foil tray and exposed horizontally to an

external heat ﬂux of 25 kW m

−2 located 25 mm away from the

cone heater. Considering the low phosphorus content in the foams,

the rather typical low heat ﬂux for the development of foams was

chosen. The foams should burn not too intense to get meaning-

ful comparisons. All samples were run twice; a third measurement

was done if any key result deviated by more than 10%. The graphs

use the data from the single measurement with the most repre-

sentative data, the tables summarize the averaged results; the un-

certainty is estimated using the observed deviation from the av-

eraged value. UL 94 HBF tests were conducted according to ISO

9772. The measurement determines the tendency of the materials

to either extinguish or spread the ﬂame once the specimen has

been ignited. A specimen 150 mm ×50 mm ×10 mm in size

(length ×width ×thickness) was used for this test.

2.4.6. Smoke and toxic gas measurements

The smoke density chamber (SDC) from Fire Testing Technol-

ogy Limited, United Kingdom, contains a sealed test chamber with

photometric equipment. It measured the speciﬁc optical density

of smoke generated by samples and determined the content of

smoke. The measurement was done according to the ISO 5659

standard with a specimen size of 75 mm ×75 mm ×25 mm

(length ×width ×thickness), exposing the specimens to an exter-

nal heat ﬂux of 25 kW m

−2

. The FTIR-spectrometer coupled with

the smoke density chamber provided the qualitative and quanti-

tative data on the composition of gases in the smoke. The FTIR-

spectrometer was calibrated according to the standard ISO 19702.

3. Results and discussion

3.1. Characterization of phosphorous soybean-oil–based polyol

To comﬁrm the synthesis of the soybean-oil–based polyol, the

chemical structure was characterized by

H NMR and

P NMR

spectroscopy.

P NMR was used to determine the presence of

phosphorus in the synthesized polyol. Fig. 3 a and b show the

and

P NMR spectra of dimethyl phosphonate and soybean-oil–

based polyol. It was mentioned above that the product was a mix-

ture. Thus, two polyol molecules labelled with numbers are shown

in Fig. 2 for a better explanation of the peaks in the

H NMR spec-

trum of phosphorous soybean-oil–based polyol in Fig. 3 a, which

depicts different types of hydrogen in the synthesized polyol and

their corresponding signals. In Fig. 3 a, a signal at about 3.4 ppm

refers to the hydrogen generated from epoxy ring opening. The

peak at 5.34 ppm corresponds to a carbon double bond, RCH

CHR.

It indicates that a peak at 2.04 ppm belongs to the hydroxyl group

formed from expoxy ring opening. The highest signal at 1.25 ppm

is attributed to the hydrogen atoms which bond with the rest

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 3. (a)

H and (b)

P NMR spectra of dimethyl phosphonate and phosphorous soybean-oil–based polyol.

Table 2.

Apparent density and mechanical test results.

Apparent density

(kg m

−3

)

Tensile strength

(kPa)

Elongation at break

(%)

Compression stress at 40%

compression (kPa)

Compression stress at 65% compression

(kPa)

FPUF 34.77 ±1.79 116 ±10.69 86.36 ±8.16 3.82 ±0.11 8.51 ±0.42

FPUF20 29.13 ±1.64 143.3 ±6.21 148.2 ±6.01 4.75 ±0.39 10.54 ±1.48

FPUF40 31.28 ±2.08 133.7 ±4.55 133.1 ±1.65 6.68 ±0.35 15.17 ±1.43

FPUF60 34.90 ±1.97 99.6 ±16.21 88.96 ±15.22 4.82 ±0.27 12.41 ±1.54

FPUF80 31.45 ±0.56 86.53 ±7.88 69.37 ±7.31 4.57 ±0.24 12.89 ±1.21

FPUF-5EG 32.41 ±1.51 114.2 ±6.61 92.02 ±7.98 6.32 ±0.49 14.27 ±1.64

FPUF20-5EG 31.75 ±1.37 113.7 ±20.66 118.1 ±22.79 4.82 ±0.19 11.51 ±0.80

FPUF40-5EG 30.03 ±1.28 101.7 ±7.27 120.4 ±10.42 4.27 ±0.11 10.09 ±0.33

FPUF60-5EG 32.95 ±1.22 96.74 ±7.18 102.4 ±7.59 5.41 ±0.22 14.51 ±0.66

FPUF80-5EG 37.50 ±1.81 52.21 ±4.69 48.74 ±2.27 4.31 ±0.14 13.52 ±0.57

FPUF-10EG 40.33 ±3.73 133.2 ±8.34 120.9 ±14.31 6.83 ±0.16 18.94 ±0.64

FPUF20-10EG 31.05 ±0.45 84.23 ±10.66 84.71 ±9.39 4.74 ±0.27 11.72 ±1.54

FPUF40-10EG 37.48 ±1.52 98.79 ±22.09 82.18 ±16.26 7.20 ±0.16 18.42 ±0.62

FPUF60-10EG 36.98 ±2.44 101.5 ±11.37 84.38 ±5.44 6.15 ±0.26 18.92 ±1.72

FPUF80-10EG 40.83 ±1.47 55.15 ±2.87 39.31 ±0.85 6.59 ±1.23 32.44 ±8.94

of the carbon on the glycerol backbones. In Fig. 3 b, the single

peak for phosphorus at 10.46 ppm in dimethyl phosphonate shifted

to 20.5 ppm in phosphorous soybean-oil–based polyol. The sig-

nal at 20.5 ppm corresponding to phosphate group indicated that

dimethyl phosphonate reacted successfully with epoxy groups of

epoxidized soybean oil to produce phosphorous soybean-oil–based

polyol.

3.2. Foam morphologies

The apparent density and the cell structure, which are shown in

Table 2 and Fig. 4 , are important information to evaluate the physi-

cal properties of FPUF. All the FPUFs showed similar apparent den-

sity of around 30–40 kg m

−3

. The cell size and cell size distribu-

tion are comparable for FPUF, FPUF20, FPUF40, FPUF-5EG, FPUF20-

5EG, FPUF40-5EG, FPUF-10EG, FPUF20-10EG, and FPUF40-10EG. For

higher contents of soybean-oil-based polyol, the cell sizes tended

to increase, what is more the cellular structure was affected by the

cell nucleation and cell growth during foaming. The SEM images

in Fig. 4 reveal that more defects and a less regular cell structure

were formed with increasing amounts of substitution of synthe-

sized polyol. The structure becomes spongier; the increased poros-

ity is expected to increase the air ﬂow through the foam also re-

ducing foam recovery due the pneumatic effect. The addition of EG

up to 10-wt.% is characterized by remaining integrity throughout

the entire cell structure.

3.3. Mechanical properties

According to Table 2 , the compression stress increases due to

the increasing amount of EG in FPUFs. The presence of EG gen-

erally increased the apparent density. The compression stress is

highly related to the density of the samples. The foams with higher

density showed the higher compression stress. As a result, EG en-

hances supporting properties.

Considering the samples without EG, the tensile strength and

the elongation at break of FPUF20 and FPUF40 are higher than

those of FPUF. Hydrogen bonds between segments in urethane

groups are one of the factors affecting the mechanical proper-

ties. The molecular weight of the synthesized bio-based polyol

(1256 g/mol) is lower than that of its petrochemical counterpart.

Typically, the low molecular weight polyol with higher hydroxyl

value induces more reactions between isocyanates and hydroxyl

groups, so that more urethane and urea linkages are formed. This

causes a greater number of hydrogen bonds and thus higher tensile

strength [34] . The synthesized polyol contained dangling carbon

chains which acted as a plasticizer, but also reduced the crosslink

density [3] . Petrochemical polyols are polyether triols with hy-

droxyls at the terminal position, whereas the synthesized polyol

has only secondary hydroxyl groups. Furthermore, the molecular

weight of synthesized polyol is lower than that of the petrochem-

ical one. After reacting with isocyanate, the length of the soft

segment between the branching points of synthesized polyol is

shorter than that in petrochemical polyol. Therefore, with a low

content of the synthesized polyol, the elongation at break was en-

hanced, as the dangling carbon chains acted as a plasticizer. When

the amount of synthesized polyol was further increased, the dan-

gling carbon chains further reduced the crosslink density, and the

shorter soft segments reduced the ﬂexibility of the polymer. Mean-

while, the crosslinking density increases with the amount of syn-

thesized polyol due to higher OH number. According to the results

from SEM images, at higher content of polyol leading the foam

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 4. SEM images of (a) FPUF, (b) FPUF20, (c) FPUF40, (d) FPUF60, (e) FPUF80, (f) FPUF-5EG, (g) FPUF20-5EG, (h) FPUF40-5EG, (i) FPUF60-5EG, (j) FPUF80-5EG, (k) FPUF-

10EG, (l) FPUF20-10EG, (m) FPUF40-10EG, (n) FPUF60-10EG, (o) FPUF80-10EG.

Table 3

Selected results obtained from TG and DTG curves of the samples.

Material T

5% T

max1 Tmax

2 Difference between Tmax

and Tmax

2 Residue at 700 °C Mass change 1 Mass change 2

°C °C °C °C % % %

FPUF 264.6 293.9 379.6 85.7 2.6 −30.5 −66.7

FPUF20 263.2 292.6 391.8 99.2 1.9 −28.3 −69.3

FPUF40 259.0 292.2 394.9 102.7 2.5 −27.6 −69.4

FPUF60 261.4 289.4 400.5 111.2 3.0 −26.4 −70.4

FPUF80 259.0 287.1 400.1 113.0 3.4 −24.9 −71.2

FPUF-5EG 263.1 296.1 380.1 84.0 4.0 −30.2 −65.3

FPUF20-5EG 263.6 295.0 392.7 97.8 3.3 −27.7 −68.6

FPUF40-5EG 262.9 293.0 398.7 105.7 5.5 −27.7 −66.5

FPUF60-5EG 261.5 290.9 400.8 110.0 7.0 −25.4 −66.9

FPUF80-5EG 263.4 293.7 402.0 108.3 6.3 −27.1 −66.2

FPUF-10EG 261.2 292.9 380.7 87.8 9.5 −28.1 −61.8

FPUF20-10EG 259.0 292.4 394.2 101.9 9.9 −26.1 −63.3

FPUF40-10EG 258.7 292.8 398.2 105.4 7.7 −27.9 −64.0

FPUF60-10EG 261.9 293.0 404.0 111.1 10.8 −25.6 −63.1

FPUF80-10EG 260.4 289.8 405.2 115.4 9.9 −25.2 −64.5

∗T

: The onset temperature at 5% mass loss.

#Tmax

: The ﬁrst maximum thermal decomposition rate; Tmax

: The second maximum thermal decomposition rate.

structure defects will most likely cause the observed decrease in

mechanical properties. As a result, the tensile strength decreased

after a further increase in the amount of synthesized polyol. The

trend for the mechanical properties of FPUF-5EGs is similar to that

of FPUFs without EG, but the mechanical properties of FPUF-5EGs

are generally inferior to those of FPUFs without EG. The incorpo-

ration of EG into FPUF causes inferior mechanical properties as

the EG deteriorates cellular structure of the foams. The results for

FPUF-10EGs indicated that the higher loading of EG and the syn-

thesized polyol compromised the performance in terms of tensile

strength and elongation at break.

3.4. Pyrolysis

3.4.1. Thermogravimetry analysis

The mass loss of the FPUFs during pyrolysis under nitrogen was

measured by thermogravimetry (TGA), and the selected character-

istic data are recorded in Table 3 . The thermal decomposition be-

haviors of the FPUFs were investigated. All samples went through

a two-stage decomposition and, accordingly, there are two peaks

in the DTG curve (the ﬁrst derivative of the TGA curve) in the

range from 200 to 500 °C. Generally, the ﬁrst weight loss was

attributed to the rupture of the polyurethane bond and thus it

is accompanied with the release of TDI-derived products, while

the second stage was attributed mainly to the decomposition of

the hydrocarbon chains of the polyol [35–38] . As illustrated in

Fig. 5 a1–a3, the decomposition of the ﬁrst stage of all samples

showed quite similar behavior. The incorporation of synthesized

polyol and EG both increased the thermal stability of the second

stage of decomposition. The FPUFs with higher content of synthe-

sized polyol and EG shifted to the right in the second stage of

the TGA curve, which means that the soft segments with higher

phosphorus content started to decompose at higher temperatures.

It indicates that the thermal stability of FPUFs containing synthe-

sized polyol is higher than that of FPUF. This may be because the

phosphorus increased the residue in the ﬁrst decomposition step

of FPUFs containing synthesized polyol, since phosphorous ﬂame

retardants generate char, forming a catalyst and thus increasing

char yield during decomposition. The higher the synthesized polyol

content, the greater the thermal stability. It was observed that

the FPUFs with EG but without any synthesized polyol generally

started to degrade at lower temperatures in the second stage, at

around 380 °C. This indicates that they have lower thermal sta-

bility at the second stage of the decomposition. In contrast to the

rest of the samples, they started to degrade between 391.8 °C and

405.1 °C in the second stage.

According to the DTG curves in Fig. 5 b1–b3, the second peaks

of DTG shifted to higher temperatures with higher content of syn-

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 5. TG curves of (a1) FPUFs, (a2) FPUF-5EGs and (a3) FPUF-10EGs; DTG curves of (b1) FPUFs, (b2) FPUF-5EGs and (b3) FPUF-10EGs.

thesized polyol. The separation between the ﬁrst and second peaks

was greater when the FPUFs contained higher phosphorus content.

3.4.2. Evolved gas analysis

To investigate the pyrolysis components of FPUFs evolved in

the gas phase, the volatile products produced during thermal de-

composition were evaluated using TG-FTIR. The TG-FTIR spectra of

FPUF are similar to that of FPUF80. Common peaks are exhibited

for both FPUF and FPUF80. The spectrum in Fig. 6 a shows sharp

absorption peaks at 2361 cm

−1 and 2278 cm

−1 at 294 °C, indi-

cating the vibration absorption of CO

and –NCO groups, respec-

tively. At 340 °C, the absorption peaks at 2978 cm

−1

, 2933 cm

−1

and 2890 cm

−1 show the stretching vibrations of –CH

and –

. Moreover, the absorption peaks at 1744 cm

−1

, 1100 cm

−1

and

914 cm

−1 represent C = O, C–O and NH

, respectively. Although the

peaks are not obvious, at 259 °C Fig. 6 b shows absorption peaks at

2320 cm

−1 and 900 °C, which were attributed to the P–H bond. A

peak at 1205 cm

−1 corresponds to the stretching vibration of P = O.

The phosphorous compounds were released as a gaseous prod-

uct from the beginning of the thermal decomposition (259 °C -

398 °C).

3.4. Flame retardancy

3.4.1. Limiting oxygen value (LOI)

The results of LOI are listed in Table 4 . Referring to the group

of samples without EG, the LOI value of FPUF was only 18.8 vol.-

%, whereas that of FPUF20 increased to 20.7 vol.-%. The LOI value

of FPUF40 reached 21 vol.-%, which is the highest value among

this group of samples. The phosphorus content in the synthesized

polyol enhanced the ﬂame retardancy. When 60-wt.% and 80-wt.%

of the polyether polyol were replaced with the synthesized polyol

in the polymer matrix, the results showed a trend of decreased

LOI values (19.3 vol.-% and 19.2 vol.-% for FPUF60 and FPUF80, re-

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 6. TG-FTIR spectra of the gas phase in the thermal degradation of (a) FPUF and (b) FPUF80 at different pyrolysis temperatures.

Table 4

LOI results.

Sample LOI (vol.-%)

FPUF 18.8 ±0.1

FPUF20 20.7 ±0.3

FPUF40 21 ±0.1

FPUF60 19.3 ±0.3

FPUF80 19.2 ±0.2

FPUF-5EG 19.2 ±0.2

FPUF20-5EG 19.8 ±0.2

FPUF40-5EG 19.8 ±0.2

FPUF60-5EG 20.2 ±0.2

FPUF80-5EG 20.4 ±0.2

FPUF-10EG 19.8 ±0.2

FPUF20-10EG 21.2 ±0.2

FPUF40-10EG 21.8 ±0.2

FPUF60-10EG 21.6 ±0.2

FPUF80-10EG 21.6 ±0.2

spectively). This phenomenon can be explained by the existence of

open-ended dangling aliphatic chains from the synthesized polyol.

The larger amount of these dangling chains was brought into the

polymer matrix when a higher content of synthesized polyol was

used. These dangling hydrocarbon chains served as a good com-

bustion fuel. Therefore, the ﬁre behavior was weakened despite the

higher phosphorus content in FPUF60 and FPUF80.

Moreover, 5-wt.% of EG was added as an additive type ﬂame

retardant. During burning of the FPUF, the EG ﬂakes expanded

and developed a “worm-like” structure that limited the heat and

mass transfer from the polymer to the heat source [30] . EG pro-

vided additional ﬂame retardancy to the system and the LOI value

increased with increasing amounts of phosphorous soybean-oil–

based polyol.

As a rule, the higher the loading of EG added, the higher the

LOI value. The FPUFs with 10-wt.% EG exhibited better ﬁre perfor-

mance (with LOI values of up to 21.8 vol.-%) than the ones with

5-wt.% EG added (with LOI values of up to 20.4 vol.-%).

3.5.2. UL94 horizontal burning test

The horizontal burning characteristics of the samples were

measured by UL 94 HBF test. The test was performed under the de-

ﬁned conditions with the specimens placed horizontally on a sup-

port gauze, with the speciﬁed gas ﬂow rate and line pressure for

the ﬂame. Self-extinguishing and dripping behaviors were taken

into consideration for the test. The results are listed in Table 5 .

The samples with EG exhibited anti-dripping behavior because

they all showed no sign of releasing burning drops that could ig-

nite the cotton underneath. The FPUF with a higher loading of EG

showed a lower burning rate. The burned distance of the samples

with 10-wt.% EG were reduced remarkably to 5 mm. Hence, FPUF-

10EGs showed self-extinguishing behavior. The samples containing

synthesized polyol and 10-wt.% EG exhibited better ﬂame retar-

dancy in the test. The burning rate of these samples was reduced

to 8 mm/min.

3.5.3. Fire behavior: cone calorimeter

The cone calorimeter is used to evaluate the ﬁre performance in

ﬁre scenarios of developing ﬁres. Several parameters such as peak

heat release rate (PHRR), time of ignition ( t

), total heat release

(THR), effective heat of combustion (EHC), total smoke released

(TSR), char yield at ﬂameout and the maximum average rate of

heat emission (MARHE) are summarized in Table 6 . Fig. 7 a–c show

the HRR curves of FPUFs without EG, with 5-wt.% of EG added,

and with 10-wt.% of EG added, respectively. Fig. 8 a–c display the

heat release rates of selected FPUFs with different weight percent-

ages of EG added. The char residue of the FPUFs, the relationship

between the percentage of residue and the phosphorus content in

residue, and the total heat release of FPUFs are plotted in Fig. 9 a–c,

respectively. Fig, 9b is taken from [39] .

In general, FPUF collapsed rapidly during the ﬁrst stage of burn-

ing and subsequently a liquid pool ﬁre developed [ 23 , 40 ]. The val-

ues of t

recorded in Table 6 did not show any signiﬁcant change

when the content of EG and the synthesized polyol was varied,

probably due to the uncertainty. According to Fig. 7 a, FPUF exhibits

two peaks of heat release rate (HRR). The second peak (pHRR at

492 kW m

−2

) is higher than the ﬁrst (pHRR at 330 kW m

−2

). In

the ﬁrst peak, the material was ignited and the cellular structure

started to collapse. The pool ﬁre formed in the second peak and

it mainly dominated the burning process. It also promoted the de-

composition by establishing a feedback loop to the substance start-

ing to burn and may have enhanced the ﬂame spread to adjacent

materials. To a certain extent, the phenomenon is associated with

the two step decomposition of polyurethane (PU) that was ob-

served in thermogravimetric analysis (TGA). The ﬁrst decomposi-

tion step corresponds to the breaking of the urethane bonds ac-

companied by the release of TDI-derived products, whereas the

second step is dominated by the decomposition of the hydrocar-

bon chains of polyol. The HRR curve of the remaining FPUFs with-

out EG is similar to FPUF, with quite a similar shape and burning

time. However, the HRR curve of the remaining FPUFs without EG

does not show an obvious ﬁrst pHRR. The HRR of the ﬁrst stage of

these FPUFs remained at around 200 kW m

−2 and then climbed

up to a pHRR at 60 0–70 0 kW m

−2 in the second stage of burning.

The pHRR of the remaining FPUFs without EG is higher than that of

FPUF. The incorporation of synthesized polyol into the FPUF caused

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 7. Heat release rates of FPUFs with different replacement percentages of phosphorous soybean-oil–based polyol and (a) without expandable graphite added, (b) with

5-wt.% expandable graphite added, (c) with 10-wt.% expandable graphite added.

Fig. 8. Heat release rates of FPUFs with different weight percentages of additional expandable graphite; (a) FPUF, (b) FPUF40 and (c) FPUF80.

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Table 5

UL 94 HBF results.

Material Burning time (s) Distance burned (mm) Burning drops Burning rate (mm/min)

FPUF 55 100 Yes 109

FPUF20 61 100 Yes 98

FPUF40 63.8 100 Yes 94

FPUF60 54.7 100 Yes 110

FPUF80 51.4 100 Yes 117

FPUF-5EG 53 100 No 113

FPUF20-5EG 123.4 100 No 48

FPUF40-5EG 73.1 100 No 82

FPUF60-5EG 105.6 100 No 57

FPUF80-5EG 118.6 100 No 51

FPUF-10EG 61 40 No 39

FPUF20-10EG 37.4 5 No 8

FPUF40-10EG 41.4 10 No 14.4

FPUF60-10EG 38 10 No 15.8

FPUF80-10EG 32 10 No 19

Fig. 9. (a) Char residue of FPUFs, (b) relationship between percentage of residue and the phosphorus content in residue, and (c) total heat release of FPUFs.

slight decrease in HRR during the ﬁrst stage of burning. This may

be due to the presence of phosphorus in the synthesized polyol.

In the second stage, the pHRR of the FPUFs without EG (FPUF20,

FPUF40, FPUF60 and FPUF80) increased compared to the pHRR of

FPUF due to the incorporation of the P-containing polyol. This phe-

nomenon may be explained by the low content of phosphorus in

the FPUFs being consumed during the ﬁrst stage of burning, and by

the dangling carbon chains in the soybean-oil–based polyol acting

as an effective fuel.

The effective heat of combustion (EHC) is measured in the cone

calorimeter as a product of the effective heat of combustion of the

volatiles and the combustion eﬃciency of the ﬂame [ 41 ]. EHC is

a tool to assess the eﬃciency of a ﬂame retardant in terms of

ﬂame inhibition action and gas phase activity [ 37 , 39 ]. The aver-

age EHC of FPUF-5EGs did not evidence any signiﬁcant ﬂame re-

tardancy in the gas phase as compared to that of FPUFs without

EG. We hypothesize that ﬂame retardancy occurs mainly in the

condensed phase. However, the average EHC of volatiles of FPUF-

10EGs decreased compared to that of FPUFs and FPUF-5EGs. This

phenomenon was attributed to the ﬂame retardant mechanism of

the phosphorus component in the gas phase [ 42 , 43 ]. For the FPUF,

the EHC at the ﬁrst stage was 20.7 MJ kg

−1

, whereas that at the

second stage was 26.2 MJ kg

−1

. Although the EHC at the ﬁrst stage

of FPUF is somewhat lower than what was reported for comparable

systems [ 37 , 44 ], overall EHC and the difference in EHC for the two

stages is well comparable with literature. We believe that the sep-

aration in burning stages mainly controlled by different decompo-

sition steps, the difference in composition of the PFUF, and differ-

ent degrees of overlapping of the decomposition stages explain the

results. Hereby some systems showed three separate stages, some

only one, the EHC vary between 20 and 32 MJ kg

−1

. The EHC of

FPUFs always showed the lowest value at the ﬁrst stage. At the

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Table 6

Cone calorimeter data.

Sample t

(s) t

(s) PHRR (kW m

−2

) THR (MJ m

−2

) Average EHC # (MJ kg

−1

) EHC

(MJ kg

−1

) EHC

(MJ kg

−1

) EHC

(MJ kg

−1

) Char residue (wt.%) TSR (m

−2

) MARHE (kW m

−2

)

FPUF 5.5 ±0.5 99 ±1 492 ±24 44.9 ±2.4 25.7 ±0.1 20.7 ±1.7 26.2 ±2.9 N.A. −1.0 ±1.3 383 ±11.7 311 ±10.4

FPUF20 5.5 ±0.5 85 ±3 600 ±17 42.3 ±3 28.3 ±1.6 23.1 ±0.4 31.5 ±0.9 N.A. −0.3 ±0.3 418 ±23 326 ±10.1

FPUF40 6.5 ±0.5 84 ±2 682 ±19 44.5 ±1.1 26.8 ±0.3 20.4 ±1.0 28.7 ±0.8 N.A. 0.3 ±0.2 528 ±13.2 328 ±2.8

FPUF60 7.5 ±0.5 85 ±3 581 ±32 44.1 ±0.9 26.6 ±0.1 20.4 ±0.4 28.7 ±0.1 N.A. 0.0 ±0.1 603 ±51.8 314 ±2.9

FPUF80 6.5 ±0.5 81 ±3 597 ±7 43.0 ±1.1 26.9 ±0.1 20.3 ±0.5 29.3 ±0.1 N.A. −0.2 ±0.4 660 ±28.8 311 ±4.2

FPUF-5EG 4.5 ±0.5 20 ±1 287 ±12 38.2 ±1.7 25.8 ±0.1 24.3 ±0.4 26.0 ±0.5 28.7 ±0.4 9.2 ±0.5 222 ±31.2 189 ±2.8

FPUF20-5EG 5.5 ±1.5 24 ±4 297 ±25 38.6 ±0.9 26.4 ±0.2 24.5 ±0.3 26.3 ±0.1 30.7 ±0.1 8.3 ±0.1 369 ±8.6 203 ±15

FPUF40-5EG 5 ±0 21 ±1 257 ±8 36.7 ±0.4 26.7 ±0.1 24.2 ±0.1 26.1 ±0.1 31.7 ±0.7 9.3 ±0.5 373 ±64.7 176 ±8.7

FPUF60-5EG 6 ±1 22 ±1 262 ±10 41.4 ±1.3 26.9 ±0.3 24.4 ±0.7 26.4 ±0.1 31.5 ±0.8 9.7 ±0.2 573 ±0.7 183 ±9.5

FPUF80-5EG 6

±2 25 ±3 279 ±5 43.3 ±0.12 26.9 ±0.1 24.9 ±0.1 28.9 ±0.1 N.A. 9.1 ±0.1 710 ±4.74 194 ±0.4

FPUF-10EG 5.5 ±0.5 15 ±1 222 ±13 41 ±3.26 23.8 ±0.5 23.8 ±0.5 N.A. N.A. 15.0 ±0.2 55,1 ±6.8 127 ±2.1

FPUF20-10EG 5 ±1 15 ±1 194 ±3 26.9 ±0.7 24.5 ±0.3 24.5 ±0.3 N.A. N.A. 29.8 ±2 62.7 ±2.1 121 ±7

FPUF40-10EG 4.5 ±0.5 16 ±1 193 ±11 25.6 ±0.5 23.4 ±0.1 23.4 ±0.1 N.A. N.A. 43.8 ±1.4 65.4 ±1.6 117 ±7.5

FPUF60-10EG 4 ±2 15 ±1 204 ±17 22.9 ±0.39 23.7 ±0.5 23.7 ±0.5 N.A. N.A. 51.2 ±0.3 72.2 ±2.7 119 ±7.8

FPUF80-10EG 3.5 ±2.5 15 ±1 211 ±6 22.4 ±2.6 23.8 ±0.8 23.8 ±0.8 N.A. N.A. 54.2 ±4.5 91.2 ±3.4 124 ±0.6

+ t

corresponds to the time to peak of heat release rate (PHRR).

# EHC corresponds to the effective heat of combustion of the deﬁned peak (Total heat release/ Total mass loss), (i.e. EHC

, EHC

and EHC

correspond to the EHC of peak 1, peak 2 and peak 3, respectively).

second stage, higher values of EHC were recorded. The EHC value

at the third stage was the worst.

The addition of EG caused a signiﬁcant reduction in pHRR and

MARHE values. As the EG provided great charring ability to form a

thermally insulating barrier, it prevented further decomposition in

the condensed phase and prolonged the time of burning. In Fig. 7 b,

all FPUFs containing 5-wt.% EG exhibit a similar ﬁrst pHRR. Refer-

ring to the ﬁrst pHRR, the addition of 5-wt.% EG resulted in a re-

duction of approximately 44% as compared to the reference FPUF.

After the ﬁrst sharp peak, the HRR dropped rapidly from around

280 kW m

−2 to 125 kW m

−2

, due to the EG acting as an excel-

lent protective layer against heat transfer and thermally protect-

ing the underlying polymer. All FPUF-5EGs have second or third

peaks that are lower than the ﬁrst pHRR. A sharp peak appeared

at the beginning because the material facing the heating source

was not suﬃciently protected by the EG. After consuming the very

top layer, the expanded graphite accumulated to the next layers of

material. Over time, more and more expanded graphite remained.

The thicker layer of expanded graphite contributed to better ﬂame

retardancy. However, there were still second and third pyrolysis

zones in FPUF-5EGs, due perhaps to the char from EG cracking,

leaving the underlying materials exposed to the ﬂame. This hap-

pens because the char layers are usually fragile and do not have

enough adhesion to each other. Compared to FPUF-5EG, the sec-

ond or third pHRR of FPUF-5EGs with synthesized polyol is more

signiﬁcant and earlier. This is probably because the dangling car-

bon chain in the synthesized polyol acted as a fuel source, enhanc-

ing ﬂammability. MARHE is used as an index for the hazard of de-

veloping ﬁres. The lower the value of MARHE, the better the ﬁre

performance. From Table 6 , the value of MARHE is reduced when

the amount of EG is increased. For FPUFs without EG, the value of

MARHE is around 315 kW m

−2

. When 5-wt.% and 10-wt.% of EG

were added, the MARHE values decreased to around 190 kW m

−2

and 120 kW m

−2

, respectively.

In Fig. 7 c, the EG further reduced the ﬁrst pHRRs in FPUF-10EGs

(pHRRs at around 200 kW m

−2

) compared to that in FPUF-5EGs

(pHRRs at around 280 kW m

−2

). The HRR surged to a sharp peak

ﬁrst and then spread over a wide area. The FPUF-10EG exhibited

second and third peaks, but as these are signiﬁcantly smaller and

ﬂatten, they are not considered to be a ﬁre concern. Thus, the time

to ﬂameout was much longer. The FPUF-10EGs with the synthe-

sized polyol exhibited only one obvious pHRR. The amount of EG in

the FPUF-10EGs with the synthesized polyol was suﬃcient to build

a better char layer with phosphorus, and protected against heat

transfer and stopped further pyrolysis better than the FPUF-5EGs

did. What is more this ﬂame retardancy becomes strongly syn-

ergistic for FPUF-10EGs when increasing the P-containing polyol.

The increase in char yield, the decrease in THR, and the decrease

in burning time outperformed what can be expected for a super-

position of the effects of EG and P-containing polyol. The FPUF-

10EGs with the synthesized polyol left a great amount of char

residue. Hardly any char residue was left after burning FPUF with-

out EG, indicating that most of the materials were converted into

volatiles. For FPUF-5EGs, voluminous worm-like char was formed

and around 9% of char residue remained. As depicted in Fig. 9 a,

with the higher amount of phosphorus in FPUF-10EG, char yield

at ﬂameout increased signiﬁcantly, up to 54-wt.%. This indicated

that the phosphorus in the synthesized polyol induced the forma-

tion of a stronger char layer with the presence of EG to protect

the material underneath during burning in the condensed phase.

Compared to the schematic curve in Fig. 9 b [39] , char yield ver-

sus phosphorus content in the residue, it can be assumed that the

curve of FPUF-10EG exhibits the key part, strong increase and lev-

eling off, of the S-shaped curve concluded for the transition of a

charring material with limited protective layer effect to a charring

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 10. Images of the ﬁre residue (a) FPUF, (b) FPUF20, (c) FPUF40, (d) FPUF60, (e) FPUF80, (f) FPUF-5EG, (g) FPUF20-5EG, (h) FPUF40-5EG, (i) FPUF60-5EG, (j) FPUF80-5EG,

(k) FPUF-10EG, (l) FPUF20-10EG, (m) FPUF40-10EG, (n) FPUF60-10EG, (o) FPUF80-10EG.

material with a protective layer effect eﬃcient enough to result in

incomplete pyrolysis.

Fig. 9 a and c show that the results of the char yield and THR are

comparable. The THR of FPUFs and FPUF-5EGs remained around

40 MJ m

−2

. The char residue of FPUF-0EGs and FPUF-5EGs was also

kept to 0-wt.% and 9-wt.%, respectively, although the content of the

synthesized polyol was increased. However, the THR of FPUF-10EGs

was reduced by almost 50% with higher content of synthesized

polyol, decreasing from 41.3 MJ m

−2 to 22.4 MJ m

−2

. This behav-

ior is related to the char yield of FPUF-10EGs. With the increasing

content of the synthesized polyol, the char yield of FPUF-10EGs in-

creased noticeably and the THR was reduced accordingly. This phe-

nomenon indicated that the thermal stability and insulation of the

char residue was adequate to prevent the materials from burning

completely.

Considering Fig. 8 a–c, the peak in the second stage for FPUF

with increasing amount of EG became ﬂatter, and the time to

ﬂameout became longer. The time to the highest peak of heat re-

lease rate (PHRR) decreased with increasing EG content.

3.5.4. Fire residue

The photographs of the ﬁre residue from the samples after cone

calorimetry measurements are shown in Fig. 10 . Fig. 11 shows

the SEM images of the ﬁre residue. The FPUF collapsed into a

pool and was completely consumed during the cone calorimetry

measurement. Therefore, there was nearly no residue after burn-

ing. Thin layers of glassy inorganic residue were obtained from

FPUF20, FPUF40, FPUF60 and FPUF80. This glassy layer was de-

ﬁned as polyphosphate ash. The phosphorus group from the syn-

thesized polyol decomposed into polyphosphoric acid, and ﬁnally

degraded to glassy polyphosphate [ 45 , 46 ]. For FPUFs with EG, the

formed char layer provided a thermal barrier protecting the ma-

terials underneath, reducing the rate of further decomposition.

The higher the content of EG, the more compact the expanded

graphite, and thus the more superior the intumescent effect. As

concluded from the char yield, a synergistic effect took place be-

tween phosphorus and EG. The char yield increased notably with

the increased amount of phosphorus in FPUF-10EG. The integrity of

the expanded graphite residue was retained because of the adhe-

sive effect of the phosphorous compound [23] . It was reported that

there is strong interfacial bonding between different char residues,

which strengthened the integrity and continuity of the intumes-

cent carbonization layers, leading to a stronger barrier effect in the

condensed phase [ 47 , 48 ].

3.6. Smoke and toxic gas measurements

The smoke produced in ﬁre poses a major threat to victims

of accidental ﬁre. Smoke particles hinder the visibility of escape

routes and may retard rescue operations because of their light

absorbing and scattering properties. Also, most ﬁre casualties are

caused by smoke inhalation, not by burns from the ﬂames. The

toxic gases and soot generated by burning polyurethane cause poi-

soning and suffocation to human beings. Carbon monoxide (CO)

and hydrogen cyanide (HCN) are the two major asphyxiants re-

leased during the burning of polyurethane. Hence, assessments of

smoke and toxic gases are critical for evaluating the ﬁre safety of

materials. The cone calorimeter and smoke density chamber are

small scale ﬁre tests; these were used to determine the smoke

density and smoke hazard of the materials.

3.6.1. Cone calorimeter

Rate of smoke release (RSR), total smoke release (TSR) and car-

bon monoxide production rate (COP) curves, illustrated in Fig. 12 a–

c, respectively, provide the data measured by cone calorimeter for

evaluating the emissions of smoke and toxic gases. Phosphorus

usually inhibits ﬂames eﬃciently by increasing CO production. This

is because phosphorus free radicals presented in the gas phase lead

to incomplete combustion, which inhibits the conversion from CO

to CO

[49] . Although, relevant ﬂame inhibition was not unambigu-

ously proven in the cone calorimeter, the release of P-containing

volatiles was shown by the evolved gas analysis and is proposed

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 11. SEM images of ﬁre residue of (a) FPUF20, (b) FPUF40, (c) FPUF60, (d) FPUF80, (e) FPUF-5EG, (f) FPUF20-5EG, (g) FPUF40-5EG, (h) FPUF60-5EG, (i) FPUF80-5EG, (j)

FPUF-10EG, (k) FPUF20-10EG, (l) FPUF40-10EG, (m) FPUF60-10EG, (n) FPUF80-10EG.

Fig. 12. (a) Rate of smoke release (RSR), (b) total smoke release (TSR) and (c) carbon monoxide production rate (COP) curves for selected samples.

to enhance the CO-production. Compared to FPUF, FPUF80 obvi-

ously produces more smoke, because of the phosphorous soybean-

oil–based polyol is a char promotor and larger decomposition frag-

ments are produced due to the incomplete pyrolysis. 10-wt.% of

EG in the foams predominantly contributed the effectiveness of

smoke suppression and since the RSR and TSR were reduced re-

markably. The 10-wt.% of EG generally reduced more than 80% of

the total smoke production. This amount of accumulated expanded

graphite was suﬃcient to produce a compact carbonaceous char

structure that increases the residence time of the smoke particles

in the pyrolysis zone. It provides a higher retardancy by enhanc-

ing the charring of aromatics instead of releasing smoke particles.

Furthermore, Fig. 12 c shows that the CO production of FPUF-10EG

was reduced by more than 60% as compared to that of FPUF. The

adequate amount of EG reduced the production of CO eﬃciently.

Therefore, the EG in the polymer matrix played the main role in

reducing total smoke production and the production of CO.

3.6.2. Smoke density chamber

The smoke density chamber is an elementary tool for investi-

gating the ﬁre safety of materials. In the smoke density test, spe-

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

Fig. 13. (a) Speciﬁc optical density and (b) transmission of the selected FPUFs burned with a pilot ﬂame.

Fig. 14. (a) Carbon monoxide and (b) hydrogen cyanide released from the selected samples burned with a pilot ﬂame.

ciﬁc optical density (D

) and light transmission (T) are the impor-

tant parameters measured by a photometric element to determine

the amount of smoke released from the materials during burn-

ing. Speciﬁc optical density is a dimensionless parameter of the

amount of smoke generated per unit area by a material during

burning. Light transmission (T) is detected by a photomultiplier

which integrates the light intensity over the visible spectrum [50] .

The photometric scale is comparable to the optical scale of human

vision. Accordingly, a higher transmission of light beams means

that people have higher chance of escaping a ﬁre [51] . Specimens

75 mm ×75 mm ×25 mm in size were exposed to a radiant heat

source of 25 kW m

−2

in the chamber with the use of a pilot ﬂame.

FPUF shows an obvious decrease in D

s when only the weight

percent of EG is increased in Fig. 13 a. EG worked as an ef-

fective smoke suppressant in FPUF. To compare between FPUF

and FPUF80, D

s decreased with increasing phosphorus content.

However, for the samples with the same content of EG (FPUF-

5EG/FPUF80-5EG and FPUF-10EG/FPUF80-10EG), higher phospho-

rus content increased the D

. The data of transmission are dis-

played in Fig. 13 b. The percentage of transmission corresponds to

the visibility for ﬁre evacuation. FPUF-10EG showed the greatest

percentage of light transmission among the selected samples.

Fig. 14 a and b show the amount of CO and HCN, respectively,

released during the burning of selected FPUFs. In all of the selected

samples, the results show that phosphorus reduces CO and HCN

emission. Interestingly, the HCN emission of FPUF-5EG, FPUF80-

5EG, FPUF-10EG and FPUF80-10EG increased steadily, while for

FPUF and FPUF80 it increased drastically at the beginning then

leveled off. This suggests that the presence of EG in PFUFs may

change the mechanism of HCN emission.

Data on smoke production from the cone calorimeter are ob-

tained using the dynamic air ﬂow method, while those from a

sealed smoke density chamber are accessed under conditions of

static accumulation [ 52 , 53 ]. The quality and quantity of smoke

are readily inﬂuenced by different combustion conditions [54] . As

the two tests are performed under different ventilation conditions,

ﬂow and accumulation, the results do not coincide with each other.

Therefore, two sets of the data are discussed separately.

4. Conclusion

Phosphorus-grafted soybean-oil–based polyol for FPUFs was

successfully synthesized. The use of reactive bio-based ﬂame re-

tardants is an attractive approach to achieve two aims at once in

polymer matrix, improving ﬁre performance and providing an al-

ternative to reduce the use of petroleum byproducts. The ﬂame re-

tardant FPUFs were synthesized by replacing petrochemical polyol

with synthesized soybean-oil–based polyol and adding EG. FPUFs

were prepared with different amounts of synthesized soybean-oil–

based polyol and EG in order to study their ﬂame retardancy. The

rigidity of the foam was increased to some extend and the foam-

ability were somewhat affected when using high amounts of syn-

thesized polyol. These limitations might be balanced out adjust-

ing the formulation in future works. With higher loadings of EG,

the results from LOI, UL 94 HBF and the cone calorimeter con-

sistently showed greater enhancement of ﬂame retardancy. The

THR and char yield from cone calorimeter testing revealed that

EG and phosphorus work exceptionally well together in the con-

densed phase. A nearly 50% reduction in the THR for FPUF-10EGs

was realized with substitution of 80-wt.% of phosphorous synthe-

sized polyol. The char yield of FPUFs with 10-wt.% EG surged when

phosphorus content was increased. Thus, the ratio of EG to phos-

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

phorous soybean-oil–based polyol is crucial for good ﬂame retar-

dancy performance. Apart from the improvement in ﬂame retar-

dancy, EG reduced the amount of smoke and toxic gases released

during burning with pilot ﬂame in a smoke density chamber. TGA

results indicated that the FPUF with phosphorus-grafted soybean-

oil–based polyol was more thermally stable than the reference

FPUF.

To summarize, the cooperation of phosphorus-grafted soybean

oil with EG provided ﬂame retardancy to FPUF by facilitating the

formation of char residues with outstanding barrier effect. Increas-

ing the P-content of the phosphorus-grafted soybean oil may be

proposed as promising strategy for future development.

Declaration of Competing Interest

The authors declare that they have no known competing ﬁnan-

cial interests or personal relationships that could have appeared to

inﬂuence the work reported in this paper.

CRediT authorship contribution statement

Yin Yam Chan: Conceptualization, Methodology, Validation, In-

vestigation, Data curation, Writing – original draft, Writing –re-

view & editing, Visualization, Project administration. Chao Ma:

Conceptualization, Investigation, Resources. Feng Zhou: Concep-

tualization, Investigation. Yuan Hu: Conceptualization, Resources,

Supervision, Project administration, Funding acquisition. Bernhard

Schartel: Conceptualization, Methodology, Resources, Writing –

original draft, Writing –review & editing, Supervision, Project ad-

ministration, Funding acquisition.

Acknowledgements

The authors thank D. Schulze for his contribution to the me-

chanical tests of the materials. Also, the help from T. Raspe and Dr.

S. Krüger with the smoke density chamber coupled with FTIR was

highly appreciated. This project was ﬁnanced by the DFG ( Deutsche

Forschungsgemeinschaft ) ( SCHA 730/19-1 ) and the NSFC ( National

Natural Science Foundation of China ) ( 51761135113 ).

References

[1] J.O. Akindoyo, M.D.H. Beg, S. Ghazali, M.R. Islam, N. Jeyaratnam, A.R. Yuvaraj,

Polyurethane types, synthesis and applications - a review, RSC Adv. 6 (115)

(2016) 114453–114482, doi: 10.1039/c6ra14525f .

[2] Y.C. Tu, G.J. Suppes, F.H. Hsieh, Water-blown rigid and ﬂexible polyurethane

foams containing epoxidized soybean oil triglycerides, J. Appl. Polym. Sci. 109

(1) (2008) 537–544, doi: 10.1002/app.281.53 .

[3] Z.S. Petrovic, Polyurethanes from vegetable oils, Polym. Rev. 48 (1) (2008) 109–

155, doi: 10.1080/15583720701834224 .

[4] K.B. Zhang, Y.B. Hong, N.G. Wang, Y.H. Wang, Flame retardant polyurethane

foam prepared from compatible blends of soybean oil-based polyol and phos-

phorus containing polyol, J. Appl. Polym. Sci. 135 (5) (2018) 45779, doi: 10.

1002/app.45779 .

[5] Z.S. Petrovic, A. Guo, I. Javni, I. Cvetkovic, D.P. Hong, Polyurethane networks

from polyols obtained by hydroformylation of soybean oil, Polym. Int. 57 (2)

(2008) 275–281, doi: 10.1002/pi.2340 .

[6] Z.S. Petrovic, L.T. Yang, A. Zlatanic, W.

Zhang, I. Javni, Network structure and

properties of polyurethanes from soybean oil, J. Appl. Polym. Sci. 105 (5)

(2007) 2717–2727 26346, doi: 10.1002/app.26346 .

[7] P.K.S. Pillai, S.J. Li, L. Bouzidi, S.S. Narine, Metathesized palm oil & novel polyol

derivatives: structure, chemical composition and physical properties, Ind. Crop.

Prod. 84 (2016) 205–223, doi: 10.1016/j.indcrop.2016.02.008 .

[8] A. Prociak, E. Malewska, M. Kuranska, S. Bak, P. Budny, Flexible polyurethane

foams synthesized with palm oil-based bio-polyols obtained with the use of

different oxirane ring opener, Ind. Crop. Prod. 115 (2018) 69–77, doi: 10.1016/j.

indcrop.2018.02.008 .

[9] M.C.C. Ferrer, D. Babb, A.J. Ryan, Characterisation of polyurethane networks

based on vegetable derived polyol, Polymer 49 (15) (2008) 3279–3287, doi: 10.

1016/j.polymer.2008.05.017 .

[10] P. Rojek, A. Prociak, Effect of different rapeseed-oil-based polyols on mechani-

cal properties of ﬂexible polyurethane foams, J. Appl. Polym. Sci. 125 (4) (2012)

2936–2945, doi: 10.10 02/app.3650 0 .

[11] M.A.R. Meier, J.O. Metzger, U.S. Schubert, Plant oil renewable resources as

green alternatives in polymer science, Chem. Soc. Rev. 36 (11) (2007) 1788–

1802, doi: 10.1039/b703294c .

[12] P. Alagi, S.C. Hong, Vegetable oil-based polyols for sustainable polyurethanes,

Macromol. Res. 23 (12) (2015) 1079–1086, doi: 10.1007/s13233- 015- 3154- 6 .

[13] L. Singaram, Biodiesel: an eco-friendly alternate fuel for the future –a review,

Therm. Sci. 13 (3) (2009) 185–199, doi: 10.2298/Tsci0903185s .

[14] A. Prociak, P. Rojek, H. Pawlik, Flexible polyurethane foams modiﬁed with

natural oil based polyols, J. Cell. Plast. 48 (6) (2012) 489–499, doi: 10.1177/

0021955x12446210 .

[15] A.B. Morgan, Revisiting ﬂexible polyurethane foam ﬂammability in furniture

and bedding in the United States, Fire Mater. 45 (1) (2021) 68–80, doi: 10.1002/

fam.2848 .

[16] A.B. Morgan, J.W. Gilman, An overview of ﬂame retardancy of polymeric ma-

terials: application, technology, and future directions, Fire Mater. 37 (4) (2013)

259–279, doi:

10.1002/fam.2128 .

[17] J.Q. Wang, W.K. Chow, A brief review on ﬁre retardants for polymeric foams, J.

Appl. Polym. Sci. 97 (1) (2005) 366–376, doi: 10.1002/app.21758 .

[18] S. Bourbigot, S. Duquesne, Fire retardant polymers: recent developments

and opportunities, J. Mater. Chem. 17 (22) (2007) 2283–2300, doi: 10.1039/

b702511d .

[19] B. Schartel, C.A. Wilkie, G. Camino, Recommendations on the scientiﬁc ap-

proach to polymer ﬂame retardancy: part 2 – concepts, J. Fire Sci. 35 (1) (2017)

3–20, doi: 10.1177/0734904116675370 .

[20] S. Gomez-Fernandez, M. Günther, B. Schartel, M.A. Corcuera, A. Eceiza, Impact

of the combined use of layered double hydroxides, lignin and phosphorous

polyol on the ﬁre behavior of ﬂexible polyurethane foams, Ind. Crop. Prod. 125

(2018) 346–359, doi: 10.1016/j.indcrop.2018.09.018 .

[21] Y. Yuan, H. Yang, B. Yu, Y. Shi, W. Wang, L. Song, Y. Hu, Y. Zhang, Phospho-

rus and nitrogen-containing polyols: synergistic effect on the thermal property

and ﬂame retardancy of rigid polyurethane foam composites, Ind. Eng. Chem.

Res. 55 (41) (2016) 10813–10822, doi: 10.1021/acs.iecr.6b02942 .

[22] M.J. Chen, Z.B. Shao, X.L. Wang, L. Chen, Y.Z. Wang, Halogen-free ﬂame-

retardant ﬂexible polyurethane foam with a novel nitrogen-phosphorus ﬂame

retardant, Ind. Eng. Chem. Res. 51 (29) (2012) 9769–9776, doi: 10.1021/

ie301004d

[23] A. Wilke, K. Langfeld, B. Ulmer, V. Andrievici, A. Hörold, P. Limbach, M. Bas-

tian, B. Schartel, Halogen-free multicomponent ﬂame retardant thermo-

plastic styrene-ethylene-butylene-styrene elastomers based on ammonium

polyphosphate-expandable graphite synergy, Ind. Eng. Chem. Res. 56 (29)

(2017) 8251–8263, doi: 10.1021/acs.iecr.7b01177 .

[24] P. Acuna, J. Zhang, G.Z. Yin, X.Q. Liu, D.Y. Wang, Bio-based rigid polyurethane

foam from castor oil with excellent ﬂame retardancy and high insulation ca-

pacity via cooperation with carbon-based materials, J. Mater. Sci. 56 (3) (2021)

2684–2701, doi: 10.1007/s10853- 020- 05125- 0 .

[25] J.S. Xu, Y.Q. Wu, B.L. Zhang, G.L. Zhang, Synthesis and synergistic ﬂame-

retardant effects of rigid polyurethane foams used reactive DOPO-based poly-

ols combination with expandable graphite, J. Appl. Polym. Sci. 138 (16) (2021)

e50223, doi: 10.1002/app.50223 .

[26] C. Liu, P. Zhang, Y. Shi, X. Rao, S. Cai, L. Fu, Y. Feng, L. Wang, X. Zheng,

W. Yang, Enhanced ﬁre safety of rigid polyurethane foam via synergistic ef-

fect of phosphorus/nitrogen compounds and expandable graphite, Molecules

25 (20) (2020) 4741, doi: 10.3390/molecules25204741 .

[27] M. Sobkowski, A. Kraszewski, J. Stawinski, Recent advances in H-phosphonate

chemistry. Part 1. H-phosphonate esters: synthesis and basic reactions, Top.

Curr. Chem. 361 (2018) 137–177, doi: 10.1007/128

_ 2014 _ 562 .

[28] K.D. Troev, Reactivity of H-phosphonates, Chem. Appl. H-Phosphonates (2006)

23–105, doi: 10.1016/B978-04 4 452737-0/50 0 04-1 .

[29] M. Günther, A. Lorenzetti, B. Schartel, From cells to residues: ﬂame-retarded

rigid polyurethane foams, Combust. Sci. Technol. 192 (12) (2020) 2209–2237,

doi: 10.1080/00102202.2019.1634060 .

[30] S.V. Levchik, E.D. Weil, Thermal decomposition, combustion and ﬁre-

retardancy of polyurethanes –a review of the recent literature, Polym. Int. 53

(11) (2004) 1585–1610, doi: 10.1002/pi.1314 .

[31] F.F. Feng, L.J. Qian, The ﬂame retardant behaviors and synergistic effect of

expandable graphite and dimethyl methylphosphonate in rigid polyurethane

foams, Polym. Compos. 35 (2) (2014) 301–309, doi: 10.1002/pc.22662 .

[32] A. Lorenzetti, B. Dittrich, B. Schartel, M. Roso, M. Modesti, Expandable graphite

in polyurethane foams: the effect of expansion volume and intercalants on

ﬂame retardancy, J. Appl. Polym. Sci. 134 (31) (2017) 45173, doi: 10.1002/app.

45173 .

[33] M. Modesti, A. Lorenzetti, F. Simioni,

G. Camino, Expandable graphite as an

intumescent ﬂame retardant in polyisocyanurate-polyurethane foams, Polym.

Degrad. Stabil. 77 (2) (2002) 195–202, doi: 10.1016/S0141- 3910(02)00034- 4 .

[34] T. Kattiyaboot, C. Thongpin, Effect of natural oil based polyols on the proper-

ties of ﬂexible polyurethane foams blown by distilled water, Enrgy Proced. 89

(2016) 177–185, doi: 10.1016/j.egypro.2016.05.024 .

[35] M. Heinen, A.E. Gerbase, C.L. Petzhold, Vegetable oil-based rigid polyurethanes

and phosphorylated ﬂame-retardants derived from epoxydized soybean oil,

Polym. Degrad .Stabil. 108 (2014) 76–86, doi: 10.1016/j.polymdegradstab.2014.

05.024 .

[36] A. Sut, E. Metzsch-Zilligen, M. Grosshauser, R. Pfaendner, B. Schartel, Syn-

ergy between melamine cyanurate, melamine

polyphosphate and aluminum

diethylphosphinate in ﬂame retarded thermoplastic polyurethane, Polym. Test.

74 (2019) 196–204, doi: 10.1016/j.polymertesting.2019.01.001 .

[37] A. Sut, E. Metzsch-Zilligen, M. Grosshauser, R. Pfaendner, B. Schartel, Rapid

Y.Y. Chan, C. Ma, F. Zhou et al. Polymer Degradation and Stability 191 (2021) 109656

mass calorimeter as a high-throughput screening method for the development

of ﬂame-retarded TPU, Polym. Degrad. Stabil. 156 (2018) 43–58, doi: 10.1016/j.

polymdegradstab.2018.08.004 .

[38] M. Ravey, E.M. Pearce, Flexible Polyurethane Foam. 1. Thermal Decom-

positionof a Polyether-based, Water-blown Commercial Type of Flexible

Polyurethane Foam, J. Appl. Polym. Sci. 63 (1997) 47–74, doi: 10.1002/(SICI)

1097-4628(19970103)63:1

 47::AID- APP7  3.3.CO;2- 3 .

[39] S. Rabe, Y. Chuenban, B. Schartel, Exploring the modes of action of

phosphorus-based ﬂame retardants in polymeric systems, Materials (5) (2017)

10, doi: 10.3390/ma10050455 .

[40] M. Günther, S.V. Levchik, B. Schartel, Bubbles and collapses: ﬁre phenomena

of ﬂame-retarded ﬂexible polyurethane foams, Polym. Advan. Technol. 31

(10)

(2020) 2185–2198, doi: 10.1002/pat.4939 .

[41] Y. Hu, X. Wang, Flame Retardant Polymeric materials: a Handbook (Series in

Materials Science and Engineering), CRC Press, Boca Raton, 2020, doi: 10.1201/

b22345 .

[42] W.H. Rao, Z.M. Zhu, S.X. Wang, T. Wang, Y. Tan, W. Liao, H.B. Zhao, Y.Z. Wang, A

reactive phosphorus-containing polyol incorporated into ﬂexible polyurethane

foam: self-extinguishing behavior and mechanism, Polym. Degrad. Stabil. 153

(2018) 192–200, doi: 10.1016/j.polymdegradstab.2018.04.029 .

[43] A. Lorenzetti, M. Modesti, E. Gallo, B. Schartel, S. Besco, M. Roso, Synthesis of

phosphinated polyurethane foams with improved ﬁre behaviour, Polym. De-

grad. Stabil. 97 (11) (2012) 2364–2369,

doi: 10.1016/j.polymdegradstab.2012.07.

026 .

[44] R.H. Krämer, M. Zammarano, G.T. Linteris, U.W. Gedde, J.W. Gilman, Heat re-

lease and structural collapse of ﬂexible polyurethane foam, Polym. Degrad. Sta-

bil. 95 (6) (2010) 1115–1122, doi: 10.1016/j.polymdegradstab.2010.02.019 .

[45] J. Li, X.H. Mo, Y. Li, H.W. Zou, M. Liang, Y. Chen, Inﬂuence of expandable

graphite particle size on the synergy ﬂame retardant property between ex-

pandable graphite and ammonium polyphosphate in semi-rigid polyurethane

foam, Polym. Bull. 75 (11) (2018) 5287–5304, doi: 10.10 07/s0 0289- 018- 2309- y .

[46] J.C. de Haro, D. Lopez-Pedrajas, A. Perez, J.F. Rodriguez, M. Carmona, Synthe-

sis of rigid polyurethane foams from phosphorylated biopolyols, Environ. Sci.

Pollut. R. 26 (4) (2019) 3174–3183, doi: 10.1007/s11356- 017- 9765- z .

[47] D.Y. Liu, B. Zhao, J.S. Wang, P.W. Liu, Y.Q. Liu, Flame retardation and thermal

stability of novel phosphoramide/expandable graphite in rigid polyurethane

foam, J. Appl. Polym. Sci. 135 (27) (2018) 46434, doi: 10.1002/app.46434 .

[48] M. Günther, A. Lorenzetti, B. Schartel, Fire phenomena of rigid polyurethane

foams, Polymers 10 (10) (2018) 1166, doi: 10.3390/polym10101166 .

[49] B. Schartel, Phosphorus-based ﬂame retardancy mechanisms –old hat or a

starting point for future development? Materials 3 (10) (2010) 4710–4745 3,

doi: 10.3390/ma3104710 .

[50] J. Tissot, M. Talbaut, J. Yon, A. Coppalle, A. Bescond, Spectral study of the

smoke optical density in non-ﬂaming condition, Procedia Engineer 62 (2013)

821–828, doi: 10.1016/j.proeng.2013.08.131 .

[51] W.H. Rao, W. Liao, H. Wang, H.B. Zhao, Y.Z. Wang, Flame-retardant and

smoke-suppressant ﬂexible polyurethane foams based on reactive phosphorus-

containing polyol and expandable graphite, J. Hazard. Mater. 360 (2018) 651–

660, doi: 10.1016/j.jhazmat.2018.08.053 .

[52] Q.W. Xu, H.M. Zhai, G.J. Wang, Mechanism of smoke suppression by melamine

in rigid polyurethane foam, Fire Mater. 39 (3) (2015) 271–282, doi: 10.1002/

fam.2249 .

[53] K. Salasinska, M. Borucka, M. Leszczy

nska, W. Zatorski, M. Celi

nski, A. Gajek,

J. Ryszkowska, Analysis of ﬂammability and smoke emission of rigid

polyurethane foams modiﬁed with nanoparticles and halogen-free ﬁre re-

tardants, J. Therm. Anal. Calorim. 130 (1) (2017) 131–141, doi:

10.1007/

s10973- 017- 6294- 4 .

[54] X. Liu, J.W. Hao, S. Gaan, Recent studies on the decomposition and strategies

of smoke and toxicity suppression for polyurethane based materials, RSC Adv.

6 (78) (2016) 74742–74756, doi: 10.1039/c6ra14345h .

4.5 Flame retardant combination with expandable graphite/phosphorus/CuO/castor oil in flexible

polyurethane foams

Yin Yam Chan, Andreas Korwitz, Doris Pospiech and Bernhard Schartel. ACS Appl. Polym. Mater.

2023. Doi: 10.1021/acsapm.2c01969.

Status: This article was accepted and published. https://doi.org/10.1021/acsapm.2c01969

Reprinted with permission from Yin Yam Chan, Andreas Korwitz, Doris Pospiech and Bernhard

Chemical Society.

First author contribution:

• Conceptualization

• Methodology

• Validation

• Investigation

• Writing – Original draft, review and editing

• Visualization

• Project administration

Contributions from other authors:

• Andreas Korwitz

o Methodology

o Resources

• Doris Pospiech

o Conceptualization

o Methodology

o Resources

o Writing – Review and editing

o Supervision

• Bernhard Schartel

o Conceptualization

o Methodology

o Resources

o Writing – Original draft, review and editing

o Supervision

o Project administration

o Funding acquisition

Abstract

A series of flexible polyurethane foams (FPUFs) was prepared with single and different

combinations of flame retardants and additives. Expandable graphite (EG), phosphorous polyol

(OP), copper (II) oxide (CuO) and/or castor oil (CAS) was/were added to FPUF during the foam

preparation in a one-step process. The purpose of the study is to evaluate the synergistic effects

of the flame retardants, additives, and the presence of bio-based content on the mechanical

properties, flame retardancy, and smoke behavior of FPUFs. The combination of 10 wt.% EG

and 5 wt.% OP in FPUF significantly improves the char yield. In the cone calorimeter

experiment, the char yield is nearly three times higher than that with 10 wt.% EG alone. The

smoke behavior is additionally evaluated in a smoke density chamber (SDC). Comparing the

samples with a single flame retardant, 10 wt.% of EG in FPUF greatly reduces the amount of

smoke released and the emission of toxic gases. Replacing the amount of 10 wt.% polyether

polyol in FPUF with CAS maintains the physical and mechanical properties and fire behavior

and enhances the bio-based content. The presence of 0.1 wt.% CuO in FPUF effectively reduces

the emission of hydrogen cyanide. As result, this study proposes a multi-component flame

retardant strategy for FPUF to enhance the biomass content and address the weaknesses in flame

retardancy, smoke and toxic gas emissions.

Accepted manuscript

Flame retardant combinations with expandable

graphite/phosphorus/CuO/castor oil in flexible

polyurethane foams

Yin Yam Chana, Andreas Korwitzb, Doris Pospiechb, Bernhard Schartela*

aBundesanstalt für Materialforschung und -prüfung (BAM), Unter den Eichen 87, 12205

Berlin, Germany

bLeibniz Institute for Polymer Research Dresden, Hohe Str. 6, 01069, Dresden, Germany

*Corresponding author. E-mail address: [email protected] (B. Schartel); Tel.: +49-

30-8104-1021

Keywords: flexible polyurethane foam, flame retardancy, synergistic effect, smoke behavior,

expandable graphite, bio-based

Abstract

A series of flexible polyurethane foams (FPUFs) was prepared with single and different

combinations of flame retardants and additives. Expandable graphite (EG), phosphorous polyol

(OP), copper (II) oxide (CuO) and/or castor oil (CAS) were added to FPUF during the foam

preparation in a one-step process. The purpose of the study is to evaluate the synergistic effects

of the flame retardants, additives, and the presence of bio-based content on the mechanical

properties, flame retardancy, and smoke behavior of FPUFs. The combination of 10 wt.% EG

and 5 wt.% OP in FPUF significantly improves the char yield. In the cone calorimeter

experiment, the char yield is nearly three times higher than that with 10 wt.% EG alone. The

smoke behavior is additionally evaluated in a smoke density chamber (SDC). Comparing the

samples with a single flame retardant, 10 wt.% of EG in FPUF considerably reduces the amount

of smoke released and the emission of toxic gases. Replacing the amount of 10 wt.% polyether

polyol in FPUF with CAS maintains the physical and mechanical properties and fire behavior

and enhances the bio-based content. The presence of 0.1 wt.% CuO in FPUF effectively reduces

the emission of hydrogen cyanide. As result, this study proposes a multi-component flame

retardant strategy for FPUF to enhance the biomass content and address the weaknesses in flame

retardancy, smoke and toxic gas emissions. A starting point is disclosed for future product

development.

1. Introduction

Flexible polyurethane foam (FPUF) is prone to fire. It burns rapidly due to its porous open-cell

structure (i.e., high surface-to-mass ratio), making it difficult to be effectively flame retarded.

Improving the flame retardancy of FPUF is critical to prevent serious fire hazards that

jeopardize human lives, because FPUF is ubiquitous in our daily surroundings, used not only

in upholstered furniture, mattresses, and seats in transportation, [1] but also as quasi universal

lightweight and insulation material all over the world. To improve the flame retardancy of

FPUF, flame retardants are added as additives to the polymer matrix by mechanical mixing

during the preparation process. However, not all flame retardants are suitable for incorporation

in FPUF formulations, as they may disrupt the balance between the gelling and blowing

reactions due to increased viscosity, leading to a collapse of the foam structure. [2]

In this work, expandable graphite (EG) and the commercially available phosphorous polyol

Exolit® OP 560 (OP) are used as flame retardants for FPUF. Many studies have demonstrated

that EG is an effective intumescent flame retardant for FPUF. [2-7] EG is a kind of graphite

intercalated with acid (usually sulfuric acid) in the presence of an oxidizing agent. After being

heated up to 200 °C, the EG expands hundreds of times to generate a “worm-like” structure,

acting as a protective layer for the unburned material underneath. [8-10] Unlike EG, OP is a

reactive flame retardant with terminal hydroxyl groups, forming chemical bonds directly with

the polymer network. Phosphorous compounds in polyurethane can generally function both in

the gas phase and the condensed phase. [11-13] However, according to the related research

studies on FPUFs with OP, excess OP destroys the cellular structure of FPUF because the

excess is not chemically linked to the network. [14,15] Hence, in this work, only 5 wt.% of

polyether polyol was replaced by OP to avoid foam collapse during the foaming process.

Different studies have reported that there is a synergistic effect between EG and phosphorus in

FPUF during burning. [4, 16-19] The system of EG and phosphorus not only improves flame

retardancy, but also reduces the smoke emission of FPUF during combustion, which is critical

since toxic fumes have always been regarded as the major killer in fires. Copper (I) oxide (Cu2O)

has been commonly chosen as an inhibitor for smoke and toxicants due to its adsorption and

catalytic conversion capabilities in polyurethane foams. [20-22] Copper (II) oxide (CuO) has

seldom been reported in the literature as an additive to suppress the release of smoke and

toxicants. In this work, CuO was used in FPUF to investigate its effectiveness in reducing

smoke as well as toxic gases.

Environmental issues associated with synthetic polymers have always been a major concern,

since most of them are made from petrochemical derivatives. Using natural resources such as

plant oils is an alternative to produce polyurethane in a sustainable manner. [23-28] In this study,

castor oil (CAS) is used to replace part of the petrochemical polyol and thereby increase the

renewable bio-based content in FPUF in order to improve biodegradability and eco-friendliness.

CAS is a non-edible renewable resource extracted from Ricinus communis. It is a natural polyol

as it consists of terminal hydroxyl groups on its alkyl chains. Due to the high hydroxyl value of

CAS, it can replace only some of the petrochemical polyols in this formulation of FPUF to

avoid collapse of the cellular foam structure. [29-31] The purpose of this work was to

investigate the mechanical properties, flame retardancy and smoke behavior among different

combinations of the selected additives (EG, OP, CuO and CAS) in FPUF through a series of

fire testing methods and observe how the bio-based ingredient influences the properties of the

resulting FPUFs. Using commercially available materials, this feasibility study addresses

multicomponent flame retardant systems that are close to application and may be used as

starting point for evidence-based future product development.

2. Experimental

2.1 Materials

Polyether polyol (VORANOLTM 3322 Polyol; average OH number: 48 mg KOH/g; molecular

weight: 3506 g/mol; nominal functionality: 3) was provided by Dow Europe GmbH (Horgen,

Switzerland). Dimethylethanolamine (Dabco® DMEA), tin-II-isooctotate (100%, Kosmos® T9)

and polyether polysiloxane (Tegostab® BF 2370) were supplied by Evonik Operations GmbH

(Essen, Germany). Toluylene diisocyanate (mixture of isomers) for synthesis (TDI) was

purchased from Merck Chemical GmbH (Darmstadt, Germany). Deionized water was used as

a chemical blowing agent in the foaming procedure. Expandable graphite (EG) with an

expansion ratio of 270 to 325 and a particle size of +50 mesh (>300 μm, ≥75% minimum) was

purchased from Sigma-Aldrich Chemie GmbH (Taufkirchen, Germany). Exolit® OP 560 polyol

(OP) (hydroxyl number: 400–500 mg KOH/g; phosphorus content: 10–13% wt/wt; Acid

number: max. 2 mg KOH/g) was provided by Clariant AG (Muttenz, Switzerland). Nano copper

oxide powder (CuO) (99+%) and castor oil (CAS) (100%) were purchased from abcr GmbH

(Karlsruhe, Germany) and Thermo Fisher (Kandel) GmbH (Kandel, Germany), respectively.

The metal oxide was used as a nanoparticle to achieve high activity. The selection of materials

was oriented toward the feasible production of a systematically varied set of FPUFs acting as

representative case example. The complex multicomponent approach is based on commercially

available raw materials and may be used as direct starting point for evidence-based future

product development There is plenty of room for optimization, such as by varying the flame

retardant polyol, for instance using a polyol with a higher phosphorus content such as Exolit®

OP 550 (16–18% wt/wt), using an alternative expandable graphite (different size, or coated),

[32] or optimizing the metal oxide with regard to cost efficiency.

2.2 Sample preparation

The foams were prepared by one-pot method and formed in an open paper mold 125 mm x

125 mm x 160 mm (height x width x height) in size to allow free foaming. The type and amount

of flame retardant/additives used in each FPUF sample are listed in Table 1. Part A (polyether

polyol, dimethylethanolamine, tin-II-isooctotate, polyether polysiloxane, water and the

additives (EG/ OP/ CuO/ CAS)) was stirred for 5 min in a 500 mL disposable plastic cup using

a mechanical stirrer at 2000 rpm. Then the weighted part B (TDI) was added to the blended part

A and the mixture was stirred at 2000 rpm for 7 s. The mixture was then poured into the mold,

where the foam started to rise freely. Afterward, the foam was cured at 80 °C in an oven for 24

h. The specimens used in the following tests were cut from the core of the freely risen foams.

The surface and interface parts of the foams were segregated, so that only specimens with

homogenous morphologies were investigated.

Table 1. Formulation of the samples

Sample

Polyether

polyol (g)

Surfactant

and

catalysts*

(g)

Water

(g)

CuO

(g)

CAS

(g)

TDI

(g)

FPUF

90.9

0.54

3.66

46.8

FPUF-10EG

90.9

0.54

3.66

13.38

46.8

FPUF-0.1CuO

90.9

0.54

3.66

0.13

46.8

FPUF-5OP

86.36

0.54

3.66

4.54

46.8

FPUF-5OP-10EG

86.36

0.54

3.66

13.38

4.54

46.8

FPUF-5OP-0.1CuO

86.36

0.54

3.66

4.54

0.13

46.8

FPUF-5OP-10EG-

0.1CuO

86.36

0.54

3.66

13.38

4.54

0.13

46.8

FPUF-10CAS

81.81

0.54

3.66

9.09

46.8

FPUF-5OP-10CAS

77.27

0.54

3.66

4.54

9.09

46.8

FPUF-5OP-

10CAS-10EG

77.27

0.54

3.66

13.38

4.54

9.09

46.8

FPUF-5OP-

10CAS-0.1CuO

77.27

0.54

3.66

4.54

0.13

9.09

46.8

FPUF-5OP-

10CAS-10EG-

0.1CuO

77.27

0.54

3.66

13.38

4.54

0.13

9.09

46.8

*Surfactant and catalysts represent the mixture of 0.18 g of dimethylethanolamine, 0.18 g of

tin-II-isooctotate and 0.18 g of polyether polysiloxane

2.3 Measurements and characterization

2.3.1 Morphology

A scanning electron microscope (SEM) Zeiss EVO 10 (Oberkochen, Germany) was used to

observe the core structure of the foams at a position perpendicular to the foam rising direction,

and the char residues after the cone calorimeter measurement. SEM was performed with a

constant electron high tension (EHT) voltage of 10 kV in a chamber under high vacuum. The

specimens were sputter-coated with a 15 nm layer of gold on the surface to improve

conductivity and prevent charging prior to investigation.

2.3.2 Physical and mechanical properties

The apparent density of the samples was measured according to ISO 845. The mechanical

properties were determined by measuring the tensile strength and compression strength with a

Zwick Z010 (Ulm, Germany) universal testing machine in accordance with ISO 1798 and ISO

3386-1, respectively. In the compression test, three cycles were performed per specimen. In

tensile and compression tests, four and three specimens were measured for each sample,

respectively.

2.3.3 Thermal decomposition

Thermogravimetric analysis (TGA) was used to measure the change in the sample’s mass over

time as a function of temperature under a continuous constant flow of nitrogen gas. The thermal

decomposition behavior of the samples was measured by a TG 209 F1 Iris from Netzsch

Instruments (Selb, Germany). The samples were first milled into fine powder using a CryoMill

from Retsch (Haan, Germany). 10 mg of powdered sample was placed in the aluminum oxide

crucible. After placing the crucible in the furnace, a heating program at a heating rate of 10

K/min from 40 °C to 650 °C was started under 30 mL/min nitrogen. TGA, in combination with

Fourier Transform Infrared Spectroscopy (FTIR) Tensor 27 FT-IR produced by Brucker

(Ettlingen, Germany), allowed the gaseous products evolved from the samples to be examined

during thermal decomposition.

2.3.4 Fire behavior

All specimens were placed in an environment of 23 °C and 50% relative humidity for at least

48 h before measurement. The limiting oxygen index (LOI) was measured at room temperature

according to ISO 4589-2 with a test specimen of foam II 150 mm x 10 mm x 10 mm in size

(length x width x thickness), using equipment from Fire Testing Technology Limited (West

Sussex, United Kingdom). The horizontal UL 94 burning test (UL 94 HBF) was carried out in

accordance with ISO 9772. A specimen 150 mm x 50 mm x 10 mm in size was used in this test.

The foam was placed horizontally. The measurement determines the spread speed of flame and

the dripping condition of the samples. Cone calorimeter tests were performed according to ISO

5660. The device was manufactured by Fire Testing Technology Limited (West Sussex, United

Kingdom). The specimen 100 mm x 100 mm x 50 mm in size (length x width x thickness) was

fitted in an aluminum foil tray and exposed horizontally to an external heat flux of 25 kW m-2

with a distance of 25 mm between the cone heater and the surface of the specimen. All samples

were measured twice. If any key result deviated by more than 10%, a third measurement was

performed.

2.3.5 Smoke and toxic gas measurements

A smoke density chamber (SDC) produced by Fire Testing Technology Limited (West Sussex,

United Kingdom) was used to determine the smoke production of flammable specimens in a

static airflow chamber. The measurement was performed following the ISO 5959 standard on

a specimen 75 mm x 75 mm x 25 mm in size (length x width x thickness) fitted in a sample

holder and exposed to an external heat flux of 25 kW m-2 under forced flaming conditions. SDC

was used in combination with an FTIR spectrometer, which can collect qualitative and

quantitative data on the gas composition of the smoke during the burning process.

3. Results and discussion

3.1 Morphological characterization and measurement of mechanical properties

The analysis of morphology is critical because the mechanical properties of the FPUFs are

substantially affected by cell wall thickness and cell size. FPUFs typically contain closed, open,

and partially open cells. [33] The foam structure of the specimen is homogeneous and without

any different surface morphologies, because they are all cut from the inner part of the produced

foam. The images of a cross section of the sample were taken at 100 times magnification. All

samples show very similar, complete, and intact open-cell structures of foam as shown in Figure

1. Table 2 shows that the density and the morphology of all samples is highly comparable. Even

samples with different kinds of additive added exhibit no significant differences in terms of

morphology. This means the foam preparation process did not affect these properties of the

foams.

Figure 1. SEM images of (a) FPUF, (b) FPUF-10EG, (c) FPUF-0.1CuO, (d) FPUF-5OP, (e)

FPUF-5OP-10EG, (f) FPUF-5OP-0.1CuO, (g) FPUF-5OP-10EG-0.1CuO, (h) FPUF-10CAS,

(i) FPUF-5OP-10CAS, (j) FPUF-5OP-10CAS-10EG, (k) FPUF-5OP-10CAS-0.1CuO and (l)

FPUF-5OP-10CAS-10EG-01CuO

Table 2. Density and mechanical properties of samples

Sample

Density

(kg m-3)

Tensile

strength

(kPa)

Elongation

at break

(%)

Compressio

n stress at

40%

compression

(kPa)

Compressio

n stress at

60%

compression

(kPa)

FPUF

33.9 ± 1.9

133.4 ± 6.6

119.5 ± 8.9

5.91 ± 0.64

11.53 ±1.26

FPUF-10EG

33.7 ± 1.9

154.8 ± 3.2

146.1 ± 2.2

14.8 ±0.6

35.97 ± 4.09

FPUF-0.1CuO

35.8 ± 3.4

142.8 ± 15.9

119.4 ±

21.9

7.55 ± 0.94

16.69 ± 3.74

FPUF-5OP

32.2 ± 2.8

100.3 ± 7.2

102.3 ±

16.9

6.77 ± 1.69

17.04 ±5.97

FPUF-5OP-10EG

34.5 ± 1.1

131.7 ± 11.2

123.2 ± 17

4.9 ± 0.34

10.19 ± 0.89

FPUF-5OP-0.1CuO

38.9 ± 1.3

114.4 ± 5.2

119.8 ± 8

4.88 ± 0.73

12.07 ± 3.29

FPUF-5OP-10EG-0.1CuO

34.5 ± 0.5

145.9 ± 6.8

128.8 ± 7.9

5.37 ± 0.22

11.69 ± 0.44

FPUF-10CAS

38.3 ± 2

163.2 ± 40.8

130.3 ±

30.1

10.82 ± 3.24

28.93 ± 11.52

FPUF-5OP-10CAS

37.6 ± 2.3

120.6 ± 6.04

128.4 ± 7.9

4.47 ± 1.24

9.6 ± 3.29

FPUF-5OP-10CAS-10EG

33.2 ± 2.7

125.5 ± 9.8

106.1 ± 4

4.67 ± 0.37

10.3 ±1.4

FPUF-5OP-10CAS-

0.1CuO

36.1 ± 2.6

138.3 ± 7.5

136.1 ± 6.1

5.55 ± 1.04

12.95 ± 2.95

FPUF-5OP-10CAS-10EG-

0.1CuO

37.3 ± 2.4

133.5 ± 4.4

125.6 ± 6.4

4.26 ± 0.19

8.92 ± 0.53

According to the results of mechanical tests in Table 2, in general all samples show similar

tensile strength and elongation at break. Moreover, most of the samples show comparable

compression stress at 40% and 60% compression. The replacement of petrochemical polyol

with 5 wt.% OP in FPUF reduced the tensile strength and elongation at break. This is because

FPUF-5OP has a higher crosslinking density than FPUF, since OP has a higher OH number.

The results show that FPUF-10EG and FPUF-10CAS improved all the parameters of

mechanical properties: tensile strength, elongation at break, and compression stress at 40% and

60% compression.

3.2 Pyrolysis: Thermogravimetric analysis (TGA)

The thermal decomposition behavior of FPUFs was studied. Mass loss during pyrolysis under

a nitrogen atmosphere was measured using TGA. Table S1 summarizes the results obtained

from the TGA measurement. Figure 2 shows the mass and DTG (the first derivative of mass (T)

curve) curves of the samples. The mass curve of all samples consists of two decomposition

stages. The first decomposition stage is attributed to the breaking of the polyurethane bonds

(namely hard segments), while the second decomposition stage is related to the decomposition

of soft segments, mainly the polyols. [34,35] All samples had a similar decomposition

temperature in the first decomposition stage. The mass curve of the FPUFs containing EG

shifted to higher temperatures in the second decomposition stage, implying that the samples

containing EG decomposed at higher temperatures. The FPUF containing EG had the highest

amount of residue at 600 °C because the graphite could not evaporate during thermal

decomposition and remained in the crucible. The amount of residue of FPUF-5OP at 600 °C

was double that of FPUF due to phosphorus acting as a charring agent in the condensed phase.

[36,37]

Figure 2. (a) Mass curves and (b) DTG curves of FPUF-OP-EG-CuOs; (c) Mass curves and

(d) DTG curves of FPUF-OP-CAS-EG-CuOs

3.3 Fire behavior: Reaction to the small flame

Table S2 exhibits the LOI and UL 94 horizontal test results, addressing the flammability at the

beginning of a fire. Aside from the used flame retardants, the flammability of PU foams depends

on their morphology; fire properties such as the LOI can usually be described as function of the

density. [38, 39] It should be noted that the foam production and specimen preparation yield

such a similar foam morphology that any significant impact of morphology or density on the

LOI and UL 94 results was ruled out. FPUF and FPUF-0.1CuO showed the lowest value, 18

vol.%. This means that 0.1 wt.% CuO alone did not influence flammability. FPUF-5OP-

10CAS-10EG-0.1CuO reached the highest value among the samples: 24.4 vol.%. FPUF-10EG

and FPUF-5OP had 21 vol.% and 21.2 vol.%, respectively. Generally, the presence of CAS

slightly increased the LOI value of the FPUFs. There was a slight synergistic effect between 5

wt.% OP and 0.1 wt.% CuO in FPUF. There was a clear synergistic effect between 5 wt.% OP

and 10 wt.% of CAS in FPUF. The fire behavior of the samples in reaction to small flames was

also tested by UL 94 HBF. Fire phenomena such as self-extinction and dripping were observed

and recorded. Except for FPUF, none of the samples dripped burning drops that ignited the

underlying cotton. This proved that all additives had an anti-dripping effect on the foams. FPUF

and FPUF-0.1CuO burned at a burning rate of 113 mm/min. The burning rate of FPUF-CAS

was slightly reduced to 100 mm/min. Apart from FPUF, FPUF-0.1CuO and FPUF-10CAS, the

rest of the samples exhibited self-extinguishing behavior before the 25 mm mark of the

specimens was reached. The UL 94 HBF results were consistent with the LOI results.

3.4 Fire behavior: Cone calorimeter

Various parameters such as peak heat release rate (PHRR), total heat release (THR), average

effective heat of combustion (Av-EHC), residue (wt.%), total smoke release (TSR) and

maximum average rate of heat emission (MARHE) were collected in the test and tabulated in

Table 3. The fire behavior of foams is strongly influenced by the foam morphology, for instance

higher density yields increasing fire loads (THR) and longer burning times. [39] Comparing

foams with the same morphology and density ensures that the impact of the flame retardant

components on the fire behavior is investigated, as the additives’ impact on morphology and

foaming is excluded. The HRR and THR curves of the samples are displayed in Figure 3. FPUF

consists of two peaks of HRR. The first peak is associated with the rupture of the urethane bond

in the hard segment and is accompanied by the collapse of the foam structure. The second peak

corresponds to the decomposition of the soft segment. [34, 40-42] The second peak (470 kW

m-2) is much higher than the first peak (273 kW m-2) because a pool fire was created during the

second stage of burning, which dominates the burning process. [41, 43] 10 wt.% of EG reduced

the PHRR substantially, making it even lower than the first peak of FPUF. All samples with

EG have a reduced PHRR, higher char yield, lower total smoke release and a lower MARHE.

Nano-CuO did not reduce the smoke release in FPUF. EG reduced the MAHRE and EHC

considerably. EHC is expressed as the total heat release per total mass loss. The sample with 5

wt.% of OP had a lower EHC due to flame inhibition, the gas phase activity of the phosphorus

compounds released. Notably, the samples with both 5 wt.% of OP and 10 wt.% of EG further

reduced the EHC value to less than 20 MJ kg-1. The reduction in the EHC of FPUF with the

combination of OP and EG was supported by an interplay of flame inhibition, charring, and a

protective layer. It is also notable that the char yield of FPUF with EG and OP was much higher

than that with only EG, increasing from 17.7 wt.% (FPUF-10EG) up to 51.6 wt.% (FPUF-5OP-

10EG), an increase of nearly three times. This was due to a synergistic effect of the phosphorus

compound and EG on the char yield. [4, 44, 45] In addition to the enhanced charring of the PU,

pyrolysis was largely incomplete under the protective residue functioning as heat shield. During

burning, OP increased charring efficiency and became glassy polyphosphate, lending a gluing

effect to EG which strengthened the expanded graphite to form an excellent protective barrier

against heat flux. Phosphorus was pyrolyzed into phosphoric acid derivates to catalyze the

carbonization of polymers at elevated temperatures. Bourbigot et al. [36] showed that

polyaromatic species are crosslinked with phosphohydrocarbonaceous bridges to form a

voluminous thermal insulation layer. As a result, the underlying material undergoes incomplete

pyrolysis. [16, 17, 46] This behavior led to reductions in PHRR, burning time and THR. Thus,

it is confirmed that the flame retardancy modes of action work in both the gas phase and the

condensed phase. Small, insignificant HRR peaks can be observed in FPUF-10EG throughout

the burning process. This is because cracks appeared between the EG layers, causing side

burning by exposing the inner unburned material to the heating source. The FPUF sample was

extinguished by complete consumption of the fuel. FPUF-5OP-10EG and FPUF-5OP-10EG-

0.1CuO exhibit fire behavior different from FPUF, and were extinguished by the excellent

thermal barrier that formed before everything was consumed. The difference in char yield

between them (51.6 wt.% and 20.9 wt.%) due to CuO deteriorated the thermal protective layer,

thus prolonging the burning time. Therefore, FPUF-5OP-10EG-0.1CuO had a longer burning

time and less char yield than FPUF-5OP-10EG. FPUF-5OP-10CAS-10EG-0.1CuO had the

lowest value of THR (15 MJ m-2). FPUF-5OP-10EG reduced the total burning time compared

to the sample with only 10 wt.% EG. The THR of FPUF-5OP was slightly reduced, but the

PHRR was not reduced by the phosphorous polyol (OP).

Figure 3. (a) Heat release rate and (b) total heat release rate of FPUF-OP-EG-CuOs; (c) heat

release rate and (d) total heat release rate of FPUF-OP-CAS-EG-CuOs

Table 3. Table of cone calorimeter results

Sample

PHRR

(kW m-2)

THR

(MJ m-2)

Av-EHC

(MJ kg-1)

Residue

(wt.%)

TSR

(m2 m-

MARHE

(kW m-

FPUF

452 ± 19

42.2 ±

2.3

27.5 ± 2.6

0.1 ±0.1

278 ±

296 ± 1

FPUF-10EG

191 ± 9

30.7 ±

2.6

22.1 ± 0.5

17.7 ± 0.8

42 ± 10

122 ± 4

FPUF-0.1CuO

456 ± 24

44.5 ±

4.3

24.9 ±0.1

0.1 ± 0.1

303 ±

304 ± 19

FPUF-5OP

525 ± 9

41.5 ±

2.1

22.9 ± 0.2

2.7 ± 0.2

494 ±

284 ± 10

FPUF-5OP-

10EG

166 ± 1

16.3 ±

1.3

18.9 ± 1.5

51.6 ± 0.2

67 ± 2

105 ± 5

FPUF-5OP-

0.1CuO

477 ± 32

43.5 ±

1.4

23 ± 0.1

2.9 ± 0.4

550 ± 5

294 ± 5

FPUF-5OP-

10EG-0.1CuO

180 ± 3

26.7 ±

0.4

19.6 ± 0.3

20.9 ± 1.2

68.7 ±

114

FPUF-10CAS

603 ± 14

48.1 ±

2.1

25.1 ± 0.1

0.2 ± 0.2

352 ±

330 ± 11

FPUF-5OP-

10CAS

503 ± 24

43.7 ±

2.9

23.4 ± 0.2

1.7 ± 0.4

570 ±

295 ± 13

FPUF-5OP-

10CAS-10EG

159 ± 6

16.9 ±

2.1

19.9 ± 0.2

49.3 ± 1.7

80 ± 1

100 ± 5

FPUF-5OP-

10CAS-0.1CuO

510 ± 21

41.6 ±

2.5

23.3 ± 0.2

1.6 ± 0.2

518 ±

292 ± 7

FPUF-5OP-

10CAS-10EG-

0.1CuO

165 ± 1

15 ± 0.1

19.8 ± 1

59.3 ± 0.7

68 ± 3

103 ± 2

3.5 Fire residues

The representative photographs and SEM images of the fire residue from the samples after cone

calorimeter measurements are shown in Figure S1 and Figure 4, respectively. Studying the

morphology of fire residue is important to understand the burning behavior. FPUF collapsed

into a pool fire which consumed nearly all the material during burning. [43, 47] The residue of

FPUF was minimal, as seen in Figure S1(a). Figure 4(a) shows the small fragments of the

polyurethane residue of FPUF. Figure 4(c) (FPUF-0.1CuO) also shows the fragments of residue,

but with many tiny holes. Figure 4(h) (FPUF-10CAS) shows that a part of continuous residue

was split apart. The conclusion is that these kinds of residue are fragile and not suitable to work

as a protective layer for underlying materials. Figure 4(d) (FPUF-5OP) shows a continuous flat

surface of residue without large holes. As shown in Figure 4(e), Figure 4 (g), Figure 4 (j), and

Figure 4 (l) (FPUF-5OP-10EG, FPUF-5OP-10EG-0.1CuO, FPUF-5OP-10CAS-10EG, and

FPUF-5OP-10CAS-10EG-0.1CuO), EG was agglomerated by the glassy polyphosphate mixed

with FPUF residue. Figure 5 shows a higher magnification SEM micrograph of FPUF-5OP-

10EG to explain the synergistic effect between phosphorus and EG on char yield. [46] EG

expanded several times, becoming voluminous and “worm-like” char during burning. There are

many fibrous residues containing polyphosphate, which formed a continuous network with

additional expanded graphite. They enhanced the adhesion between the expanded graphite,

thereby improving the integrity and continuity of the protective char layer and preventing

further pyrolysis of the underlying material. Less heat flux is transferred to the pyrolysis front,

eventually leading to incomplete pyrolysis, which maintains the high apparent char yield of the

specimens in cone calorimeter measurements. Without the presence of phosphorous flame

retardant, Figure 4(b) (FPUF-10EG) shows that the fibrous residue was brittle, so that most of

the fibers did not connect to other areas of expanded graphite. In Figure 4(f) (FPUF-5OP-

0.1CuO), cracks and holes appear on the surface, which indicate that the residue was fragile.

The surface of FPUF-5OP-10CAS (Figure 4(i)) is uneven, with many holes. The residue of

FPUF-5OP-10CAS-0.1CuO (Figure 4(k)) had irregular holes on the surface. The high weight

percentage of residue for FPUF-5OP-10EG, FPUF-5OP-10CAS-10EG and FPUF-5OP-

10CAS-10EG-0.1CuO is visible in the photographs of residue in Figure S1(e), Figure S1(j) and

Figure S1(l), respectively. These pictures show the structural integrity of the carbonaceous char

without any curving. The more complete structural integrity of the carbonaceous char, the

greater amount of material underneath and therefore the higher weight percentage of residue

was recorded. The curving of the carbonaceous char in FPUF-10EG and FPUF-5OP-10EG-

0.1CuO resulted in the exposure of the bottom to the flame. Therefore, a lower weight

percentage of residue remained. Hence, OP and CAS improved the structural integrity of

carbonaceous char, but CuO deteriorated it.

Figure 4. SEM images of fire residue of (a) FPUF, (b) FPUF-10EG, (c) FPUF-0.1CuO, (d)

FPUF-5OP, (e) FPUF-5OP-10EG, (f) FPUF-5OP-0.1CuO, (g) FPUF-5OP-10EG-0.1CuO, (h)

FPUF-10CAS, (i) FPUF-5OP-10CAS, (j) FPUF-5OP-10CAS-10EG, (k) FPUF-5OP-10CAS-

0.1CuO and (l) FPUF-5OP-10CAS-10EG-0.1CuO after cone calorimeter measurement

Figure 5. SEM picture of fire residue of FPUF-5OP-10EG after cone calorimeter

measurement

3.6 Smoke and toxic gas measurements

It is crucial to assess the quantity and quality of smoke emitted by materials, as smoke is always

the cause of suffocation in fires. However, the yield of smoke released depends on various

factors such as ambient temperature, the availability of oxygen, type of ignition, and air flow.

Therefore, detailed smoke measurements were performed under two different conditions, good

ventilation and controlled ventilation. In addition to smoke, toxic gases cannot be ignored

during fires. Carbon monoxide is the most common asphyxiant in fires due to incomplete

combustion. Hydrogen cyanide is one of the gaseous products released by nitrogenous materials

during burning. [48] Moreover, hydrogen cyanide is about 25 times more toxic than carbon

monoxide because it forms cyanide ions in the blood, preventing cellular respiration. [49]

Therefore, these two toxic gases were analyzed in the smoke measurements of the FPUF

samples.

3.6.1 Smoke measurement: Cone calorimeter

The smoke release was measured with a cone calorimeter under well ventilated conditions.

According to Figure 6(a), 10 wt.% EG in FPUF released a remarkably small amount of smoke

because the expanded graphite prolonged the residence time of smoke particles in the pyrolysis

zone. [17,19] More aromatics were charred, and fewer light smoke particles were eventually

released. 0.1 wt.% CuO released slightly more smoke. Compared with FPUF, FPUF-5OP and

FPUF-10CAS released more smoke. It is worth noting that 5 wt.% OP substantially increased

the smoke release of FPUF (66% higher than FPUF), because the phosphorus worked in the gas

phase to cause incomplete combustion and released more and larger particles, such as smoke

and soot. Figure 6(b) shows the graph of the carbon monoxide production rate (COP) of selected

samples. FPUF-10EG exhibited the lowest COP. The COP of FPUF-5OP was twice that of

FPUF. This is due to incomplete combustion caused by phosphorous radicals, resulting in

inhibited conversion of carbon monoxide into carbon dioxide, even in the presence of sufficient

ambient oxygen. [37] The COP of FPUF-0.1CuO and FPUF-10CAS was similar to that of

FPUF. EG suppressed the smoke and the amount of carbon monoxide released significantly.

100

Figure 6(c) and Figure 6(d) show that the TSR and COP in mixed systems were reduced by EG

to very low levels even in the presence of OP.

Figure 6. (a) Total smoke release, (b) carbon monoxide production rate curves of selected

FPUFs with single additives, (c) total smoke release and (d) carbon monoxide production rate

curves of selected FPUFs with mixed additives

3.6.2 Smoke measurement: Smoke density chamber

The specimens were burned under ventilation-controlled conditions in the smoke density

chamber (SDC). According to the results of the specific optical density (Ds) measurements in

Figure 7(a), 10 wt.% of EG in FPUF suppressed 65% of the smoke released by FPUF during

the burning process. 0.1 wt.% of CuO in FPUF inhibited more than 30% of smoke release.

However, in this case, 5 wt.% OP and 10 wt.% CAS did not play a role in suppressing smoke

due to incomplete combustion. Figure 7(b) shows the light transmission of the smoke, which

represents visibility during fire evacuation. The Ds value and transmission percentage of PFUF-

101

5OP and FPUF-10CAS are similar to those of FPUF. From the above analysis, 10 wt.% of EG

is sufficient to reduce the smoke emission for FPUF under controlled ventilation conditions.

Figure 7. Smoke density chamber coupled with FTIR gas analysis: (a) Specific optical

density, (b) transmission of the selected samples with single additives, (c) specific optical

density, (d) transmission of the selected FPUFs with mixed additives, (e) carbon monoxide,

and (f) hydrogen cyanide released from the selected samples burned with a pilot flame

102

To study the smoke release from FPUF combined with EG, OP and/or CuO, some samples were

selected for comparison in Figure 7(c) and Figure 7(d). The Ds of FPUF-5OP-10EG-0.1CuO

was slightly lower than that of FPUF-5OP-10EG at the beginning of burning, but they were the

same at the end of the measurement. This means CuO initially suppressed the smoke release.

Comparing the Ds of FPUF-5OP-10EG-0.1CuO and FPUF-5OP-0.1CuO, the addition of 10

wt.% effectively suppressed the smoke overall, especially apparent in the large reduction in the

initial stage.

3.6.3 Toxic gas measurement

Besides smoke, toxic gases, especially carbon monoxide (CO) and hydrogen cyanide (HCN),

are considered to be major hazards for people escaping fires. CO and HCN can cause

asphyxiation because they block the uptake of oxygen by cells. [48] Therefore, CO and HCN

were selected for toxic gas analysis. Figure 7(e) and Figure 7(f) show the emission of CO and

HCN, respectively, from the selected samples burned with a pilot flame. The data were obtained

from the FTIR coupled with SDC. Figure 7(e) shows that 5 wt.% of OP increased the emission

of CO significantly at the beginning of the test. The CAS significantly reduced the emission of

CO, which is more than half of the emissions from FPUF. EG significantly reduced the release

of CO at the beginning of the test. In Figure 7(f), FPUF-5OP doubled the emissions of HCN.

At the beginning of the test, the concentration of HCN surged to the highest peak, which means

the 5 wt.% OP significantly increased the emission of HCN at the beginning of pyrolysis. The

average EHC of FPUF-5OP from the cone calorimeter measurement decreased, while the

amounts of CO and HCN increased. This means flame inhibition took place in the gas phase.

EG decreased the emission of HCN effectively at the beginning of burning. Both CuO and CAS

reduced the concentration of HCN during burning. The CuO particles in FPUF were exposed

to HCN; the surface of CuO nanoparticles were reported to absorb and destruct HCN efficiently.

[50, 51]

103

4. Conclusions

The results on flexible polyurethane foams flame retarded with different flame retardants

obtained in this study revealed that EG is an outstanding flame retardant and smoke suppressant

for FPUF according to the results from the cone calorimeter, smoke density chamber and toxic

gases analysis. EG works mainly in the condensed phase to act as a protective char. EG

significantly reduced the emission of CO and HCN at the beginning of burning under

ventilation-controlled conditions. Furthermore, EG considerably reduced the concentration of

HCN throughout the smoke density chamber test. The synergistic effect between phosphorus

and EG resulted in a significant increase in char yield in cone calorimeter measurements. The

combination of phosphorus and EG serves as an excellent option for dual flame retardant

systems in FPUF. Although OP created a great deal of smoke during burning, the presence of

EG suppressed the smoke effectively. The replacement of 10 wt.% by polyol with CAS not

only increased the bio-based content in FPUF, but also maintained the physical and mechanical

properties, and the fire behavior was similar to that without CAS. It is concluded that the

presence of 0.1 wt.% of CuO additive effectively reduced the release of HCN from FPUF. An

appropriate multi-component additive approach in FPUF combined with flame retardants,

smoke suppressant and biomass substantially improved flame retardancy, smoke suppression,

and bio-based content while reducing toxic gas emissions. This study provided insight into

comprehensively improving the required properties of FPUF for the future.

Associated Content

Supporting Information Available:

Table S1 TGA results, Table S2: LOI & UL 94 results, and Figure S1: Photographs of the fire

residues.

Acknowledgements

104

This project was funded by the DFG (Deutsche Forschungsgemeinschaft) (SCHA 730/19-1).

The authors thank Dietmar Schulze for conducting the mechanical tests on the materials. The

assistance of Tina Raspe and Dr. Simone Krüger with the smoke density chamber coupled with

FTIR is highly appreciated.

References

[1] Engels, H. W.; Pirkl, H. G.; Albers, R.; Albach, R. W.; Krause, J.; Hoffmann, A.;

Casselmann, H.; Dormish, J. Polyurethanes: Versatile Materials and Sustainable Problem

Solvers for Today’s Challenges. Angew. Chem., Int. Ed. 2013, 52(36), 9422-9441. DOI:

10.1002/anie.201302766

[2] Wang, C. Q.; Lv, H. N.; Sun, J.; Cai, Z. S. Flame Retardant and Thermal Decomposition

Properties of Flexible Polyurethane Foams Filled with Several Halogen-Free Flame Retardants.

Polym. Eng. Sci. 2014, 54(11), 2497−2507. DOI: 10.1002/pen.23794

[3] Gharehbaghi, A.; Bashirzadeh, R.; Ahmadi, Z. Polyurethane Flexible Foam Fire

Resisting by Melamine and Expandable Graphite: Industrial Approach. J. Cell. Plast. 2011,

47(6), 549−565. DOI: 10.1177/0021955x11414789

[4] Chan, Y. Y.; Ma, C.; Zhou, F.; Hu, Y.; Schartel, B. A Liquid Phosphorous Flame

Retardant Combined with Expandable Graphite or Melamine in Flexible Polyurethane Foam.

Polym. Adv. Technol. 2022, 33(1), 326−229. DOI: 10.1002/pat.5519

[5] Wolska, A.; Gozdzikiewicz, M.; Ryszkowska, J. Influence of Graphite and Wood-based

Fillers on the Flammability of Flexible Polyurethane Foams. J. Mater. Sci. 2012, 47(15),

5693−5700. DOI: 10.1007/s10853-012-6394-2

[6] Bashirzadeh, R.; Gharehbaghi, A. An Investigation on Reactivity, Mechanical and Fire

Properties of PU Flexible Foam. J. Cell. Plast. 2010, 46(2), 129−158. DOI:

10.1177/0021955x09350805

105

[7] Weil, E. D.; Levchik, S. V. Commercial Flame Retardancy of Polyurethanes. J. Fire.

Sci. 2004, 22(3), 183−210. DOI: 10.1177/0734904104040259

[8] Duquesne, S.; Le Bras, Bourbigot, S.; M.; Delobel, Camino, G.; R.; Eling, B.; Lindsay,

C.; Roles, T. Thermal Degradation of Polyurethane and Polyurethane/Expandable Graphite

Coatings. Polym. Degrad. Stab. 2001, 74(3), 493-499. DOI: 10.1016/S0141-3910(01)00177-X

[9] Jin, J.; Dong, Q. X.; Shu, Z. J.; Wang, W. J.; He, K. Flame Retardant Properties of

Polyurethane/Expandable Praphite Composites. 2013 International Conference on

Performance-Based Fire and Fire Protection Engineering (Icpffpe 2013) 2014, 71, 304−309.

DOI: 10.1016/j.proeng.2014.04.044

[10] Chen, Y. J.; Luo, Y. F.; Guo, X. H.; Chen, L. J.; Xu, T. W.; Jia, D. M. Structure and

Flame-Retardant Actions of Rigid Polyurethane Foams with Expandable Graphite. Polymers.

2019, 11(4), 668. DOI: 10.3390/polym11040686

[11] Zhou, F.; Ma, C.; Zhang, K.; Chan, Y. Y.; Xiao, Y. L.; Schartel, B.; Döring, M.; Wang,

B. B.; Hu, W. Z.; Hu, Y. Synthesis of Ethyl (Diethoxymethyl)phosphinate Derivatives and

Their Flame Retardancy in Flexible Polyurethane Foam: Structure-flame Retardancy

Relationships. Polym. Degrad. Stab. 2021, 188, 109557. DOI:

10.1016/j.polymdegradstab.2021.109557

[12] Liu, H.; Zhang, B.; Han, J. Flame Retardancy and Smoke Suppression Properties of

Flexible Polyurethane Foams Containing an Aluminum Phosphate Microcapsule. RSC Adv.

2017, 7(56), 35320−35329. DOI: 10.1039/c7ra04323f

[13] Zhou, F.; Zhang, T.; Zou, B.; Hu, W. Z.; Wang, B. B.; Zhan, J.; Ma, C.; Hu, Y. Synthesis

of a Novel Liquid Phosphorus-Containing Flame Retardant for Flexible Polyurethane Foam:

Combustion Behaviors and Thermal Properties. Polym. Degrad. Stab. 2020, 171, 109029. DOI:

10.1016/j.polymdegradstab.2019.109029

106

[14] Gaan, S.; Liang S. Y.; Mispreuve, H.; Perler, H.; Naescher, R.; Neisius, M. Flame

Retardant Flexible Polyurethane Foams from Novel DOPO-Phosphonamidate Additives.

Polym. Degrad. Stab. 2015, 113, 180−188. DOI: 10.1016/j.polymdegradstab.2015.01.007

[15] Gómez-Fernández, S.; Günther, M.; Schartel, B.; Corcuera, M. A.; Eceiza, A. Impact

of the Combined Use of Layered Double Hydroxides, Lignin and Phosphorous Polyol on the

Fire Behavior of Flexible Polyurethane Foams. Ind. Crops. Prod. 2018, 125, 346−359. DOI:

10.1016/j.indcrop.2018.09.018

[16] Rao, W. H.; Liao, W.; Wang, H.; Zhao, H. B.; Wang, Y. Z. Flame-Retardant and Smoke-

Suppressant Flexible Polyurethane Foams Based on Reactive Phosphorus-Containing Polyol

and Expandable Graphite. J. Hazard. Mater. 2018, 360, 651−660. DOI:

10.1016/j.jhazmat.2018.08.053

[17] Chan, Y. Y.; Ma, C.; Zhou, F.; Hu, Y.; Schartel, B. Flame Retardant Flexible

Polyurethane Foams Based on Phosphorous Soybean-Oil Polyol and Expandable Graphite.

Polym. Degrad. Stab. 2021, 191. 109656. DOI: 10.1016/j.polymdegradstab.2021.109656

[18] Wang, C. Q.; Ge, F. Y.; Sun, J.; Cai, Z. S. Effects of Expandable Graphite and Dimethyl

Methylphosphonate on Mechanical, Thermal, and Flame-Retardant Properties of Flexible

Polyurethane Foams. J. Appl. Polym. Sci. 2013, 130(2), 916−926. DOI: 10.1002/app.39252

[19] Chan, Y. Y.; Schartel, B. It Takes Two to Tango: Synergistic Expandable Graphite-

Phosphorus Flame Retardant Combinations in Polyurethane Foams. Polymers. 2022, 14(13),

2562. DOI: 10.3390/polym14132562

[20] Levin, B. C. New Approaches to Toxicity: A Seven-Gas Predictive Model and Toxicant

Suppressants. Drug. Chem. Toxicol. 1997, 20(4), 271−280. DOI: 10.3109/01480549709003885

[21] Yuan, Y.; Wang, W.; Xiao, Y.; Yuen, A. C. Y.; Mao, L.; Pan, H. F.; Yu, B.; Hu, Y.

Surface Modification of Multi-Scale Cuprous Oxide with Tunable Catalytic Activity Towards

107

Toxic Fumes and Smoke Suppression of Rigid Polyurethane Foam. Appl. Surf. Sci. 2021, 556,

149792. DOI: 10.1016/j.apsusc.2021.149792

[22] Yuan, Y.; Yu, B.; Shi, Y. Q.; Ma, C.; Song, L.; Hu, W. Z.; Hu, Y. Highly Efficient

Catalysts for Reducing Toxic Gases Generation Change with Temperature of Rigid

Polyurethane Foam Nanocomposites: A Comparative Investigation. Composites, Part A 2018,

112, 142−154. DOI: 10.1016/j.compositesa.2018.05.028

[23] Cifarelli, A.; Boggioni, L.; Vignali, A.; Tritto, I.; Bertini, F.; Losio, S. Flexible

Polyurethane Foams from Epoxidized Vegetable Oils and a Bio-Based Diisocyanate. Polymers.

2021, 13(4), 612. DOI: 10.3390/polym13040612

[24] Petrović, Z. S. Polyurethanes from Vegetable Oils. Polym. Rev. 2008, 48(1), 109−155.

DOI: 10.1080/15583720701834224

[25] Tu, Y. C.; Suppes, G. J.; Hsieh, F. H. Water-Blown Rigid and Flexible Polyurethane

Foams Containing Epoxidized Soybean Oil Triglycerides. J. Appl. Polym. Sci. 2008, 109(1),

537−544. DOI: 10.1002/app.281.53

[26] Zhang, L.; Jeon, H. K.; Malsam, J.; Herrington, R.; Macosko, C. W. Substituting

Soybean Oil-Based Polyol into Polyurethane Flexible Foams. Polymer. 2007, 48(22),

6656−6667. DOI: 10.1016/j.polymer.2007.09.016

[27] Sonnenschein, M. F.; Wendt, B. L. Design and Formulation of Soybean Oil Derived

Flexible Polyurethane Foams and Their Underlying Polymer Structure/Property Relationships.

Polymer. 2013, 54(10), 2511−2520. DOI: 10.1016/j.polymer.2013.03.020

[28] Rojek, P.; Prociak, A. Effect of Different Rapeseed-Oil-Based Polyols on Mechanical

Properties of Flexible Polyurethane Foams. J. Appl. Polym. Sci. 2012, 125(4), 2936−2945. DOI:

10.1002/app.36500

108

[29] Panda, S. S.; Samal, S. K.; Mohanty, S.; Nayak, S. K. Preparation, Characterization, and

Properties of Castor Oil-Based Flexible Polyurethane/Cloisite 30B Nanocomposites Foam. J.

Compos. Mater. 2018, 52(4), 531−542. DOI: 10.1177/0021998317710707

[30] Sharma, C.; Kumar, S.; Unni, A. R.; Aswal, V. K.; Rath, S. K.; Harikrishnan, G. Foam

Stability and Polymer Phase Morphology of Flexible Polyurethane Foams Synthesized from

Castor Oil. J. Appl. Polym. Sci. 2014, 131(17), 40668. DOI: 10.1002/app.40668

[31] Ogunfeyitimi, O. S.; Okewale, A. O.; Igbokew, P. K. The Use of Castor Oil as a

Reactive Monomer in Synthesis of Flexible Polyurethane Foam. Int. J. Multidisciplinary Sci.

Eng. 2012, 3(10), 10-14. DOI: 10.13140/RG.2.2.24190.33608

[32] Lorenzetti, A.; Dittrich, B.; Schartel, B.; Roso, M.; Modesti, M. Expandable Graphite in

Polyurethane Foams: The effect of Expansion Volume and Intercalants on Flame Retardancy.

J. Appl. Polym Sci. 2017, 134(31), 45173. DOI: 10.1002/app.45173

[33] Zammarano, M.; Matko, S.; Krämer, R. H.; Davis, R. D.; Gilman, J. W.; Sung, L. P.;

Fox, D. M.; Mehta, S. Smoldering in Flexible Polyurethane Foams: The Effect of Foam

Morphology. In: Morgan A. B.; Wilkie, C. A.; Nelson, G. L. Fire and Polymers VI: New

Advances in Flame Retardant Chemistry and Science. ACS, Washington, 2012, ACS

Symposium Series 1118, 459−479. DOI: 10.1021/bk-2012-1118.ch029

[34] Sut, A.; Metzsch-Zilligen, E.; Großhauer, M.; Pfaendner, R.; Schartel, B. Synergy

between Melamine Cyanurate, Melamine Polyphosphate and Aluminum Diethylphosphinate in

Flame Retarded Thermoplastic Polyurethane. Polym. Test. 2019, 74, 196−204. DOI:

10.1016/j.polymertesting.2019.01.001

[35] Sut, A.; Metzsch-Zilligen, E.; Großhauer, M.; Pfaendner, R.; Schartel, B. Rapid Mass

Calorimeter as a High-Throughput Screening Method for the Development of Flame-Retarded

TPU. Polym. Degrad. Stab. 2018, 156, 43−58. DOI: 10.1016/j.polymdegradstab.2018.08.004

109

[36] Bourbigot, S.; Lebras, M.; Delobel, R.; Bréant, P.; Trémillon, J. M. Carbonization

Mechanisms Resulting from Intumescence—Part II. Association with an Ethylene Terpolymer

and the Ammonium Polyphosphate-Pentaerythritol Fire Retardant System. Carbon. 1995, 33(3),

283−294. DOI: 10.1016/0008-6223(94)00131-I

[37] Schartel, B. Phosphorus-Based Flame Retardancy Mechanisms—Old Hat or a Starting

Point for Future Development? Materials. 2010, 3(10), 4710−4745. DOI: 10.3390/ma3104710

[38] Bian, X.-C.; Tang, J.-H., Li, Z.-M.; Lu, Z.-Y.; Lu A. Dependence of Flame-retardant

Properties on Density of Expandable Graphite Filled Rigid Polyurethane Foam. J. Appl. Polym.

Sci. 2007 104(5), 3347-3355. DOI: 10.1002/app.25933

[39] Günther, M.; Lorenzetti, A.; Schartel, B. Fire Phenomena of Rigid Polyurethane Foams.

Polymers 2018, 10(10), 1166 DOI: 10.3390/polym10101166

[40] Günther, M.; Levchik, S. V.; Schartel, B. Bubbles and Collapses: Fire Phenomena of

Flame-Retarded Flexible Polyurethane Foams. Polym. Adv. Technol. 2020, 31(10), 2185−2198.

DOI: 10.1002/pat.4939

[41] Morgan, A. B. Revisiting Flexible Polyurethane Foam Flammability in Furniture and

Bedding in the United States. Fire Mater. 2021, 45(1), 68−80. DOI: 10.1002/fam.2848

[42] Pitts, W. M. Role of Two Stage Pyrolysis in Fire Growth on Flexible Polyurethane Foam

Slabs. Fire Mater. 2014, 38(3), 323−338. DOI: 10.1002/fam.2183

[43] Krämer, R. H.; Zammarano, M.; Linteris, G. T.; Gedde, U. W.; Gilman, J. W. Heat

Release and Structural Collaspe of Flexible Polyurethane Foam. Polym. Degrad. Stab. 2010,

95(6), 1115−1122. DOI: 10.1016/j.polymdegradstab.2010.02.019

110

[44] Ming, G.; Chen, S.; Sun, Y. J.; Wang, Y. X. Flame Retardancy and Thermal Properties

of Flexible Polyurethane Foam Containing Expanded Graphite. Combust. Sci. Technol. 2017,

189(5), 793−805. DOI: 10.1080/00102202.2016.1251910

[45] Yao, W. G.; Zhang, D. Y.; Zhang, Y. X.; Fu, T.; Guan, D. B.; Dou, Y. L. Synthesis

Flame Retardant Effects of Expandable Graphite and Ammonium Polyphosphate in Water-

Blow Polyurethane Foam. Adv. Mater. Sci. Eng. 2019, 2019, 6921474. DOI:

10.1155/2019/6921474

[46] Wilke, A.; Langfeld, K.; Ulmer, B.; Andrievici, V; Hörold, A.; Limbach, P.; Bastian,

M.; Schartel, B. Halogen-Free Multicomponent Flame Retardant Thermoplastic Styrene-

Ethylene-Butylene-Styrene Elastomers Based on Ammonium Polyphosphate-Expandable

Graphite Synergy. Ind. Eng. Chem. Res. 2017, 56(29), 8251−8263. DOI:

10.1021/acs.iecr.7b01177

[47] Baguian, A. F.; Ouiminga, S. K.; Longuet, C.; Caro-Bretelle, A.-S.; Corn, S.; Bere, A.;

Sonnier, R. Influence of Density on Foam Collapse under Burning. Polymers. 2021, 13(1), 13.

DOI: 10.3390/polym13010013

[48] McKenna, S. T.; Hull, T. R. The Fire Toxicity of Polyurethane Foams. Fire Sci. Rev.

2016, 5, 3. DOI: 10.1186/s40038-016-0012-3

[49] Hull, T. R.; Stec, A. A. Introduction to Fire Toxicity. In: Stec, A. A.; Hull, T. R. Fire

Toxicity, Woodhead Publ. Cambridge. 2010, chap 1: 3−25. DOI: 10.1533/9781845698072.1.3

[50] Yang, M. Q.; He, J. H.; Hu, X. C.; Yan, C. X.; Cheng, Z. X.; Zhao, Y. Q,; Zuo, G. M.

Copper Oxide Nanoparticle Sensors for Hydrogen Cyanide Detection: Unprecedented

Selectivity and Sensitivity. Sensor Actuat. B-Chem. 2011, 155(2), 692−698. DOI:

10.1016/j.snb.2011.01.031

111

[51] Liang, S. H.; Wang, S. L.; Pedersen, D. B. Adsorption of HCN onto Copper@Copper-

Oxide Core-Shell Nanoparticle Systems. Adsorpt. Sci. Technol. 2009, 27(4), 349−361. DOI:

10.1260/026361709790252632

112

TOC graphic

113

Supporting Information

Flame retardant combinations with expandable

graphite/phosphorus/CuO/castor oil in flexible

polyurethane foams

Yin Yam Chana, Andreas Korwitzb, Doris Pospiechb, Bernhard Schartela*

aBundesanstalt für Materialforschung und -prüfung (BAM), Unter den Eichen 87, 12205

Berlin, Germany

bLeibniz Institute for Polymer Research Dresden, Hohe Str. 6, 01069, Dresden, Germany

*Corresponding author. E-mail address: [email protected] (B. Schartel); Tel.: +49-

30-8104-1021

114

Table S1. Selected data obtained from TGA (mass and DTG curves) of the FPUF samples

Sample

T 5% *

(°C)

Tmax1#

(°C)

Tmax2#

(°C)

Tmax1

and Tmax2

(°C)

Residue

at 600 °C

(%)

Mass

1^ (%)

Mass

2^ (%)

FPUF

265.0

±0.6

302.6

±0.3

383.2

±0.5

80.6

1.6 ± 0.4

36.9 ± 0.1

60.6 ± 0.1

FPUF-

10EG

267.9

±1.0

299.3

±0.4

384.5

±0.2

85.2

9.4 ± 0.4

30.3 ± 0.1

59.3 ± 0.2

FPUF-

0.1CuO

269.0

±1.4

301.6

±0.6

382.1

±1.6

80.5

1.3 ± 0.1

33.9 ± 0.3

63.8 ± 0.3

FPUF-

5OP

260.5

±0.5

295.2

±0.1

382.4

±0.3

87.2

3.32 ± 0.2

32.9 ± 0.1

63.0 ± 0.1

FPUF-

5OP-

10EG

260.3

±0.7

292.2

±1.0

387.7

±0.5

95.5

12.7 ± 0.1

28.6 ± 0.3

57.9 ± 0.4

FPUF-

5OP-

0.1CuO

262.9

±1.0

297.2

±1.0

382.7

±0.2

85.5

2.4 ± 0.1

32.4 ± 0.4

64.3 ± 0.5

FPUF-

5OP-

10EG-

0.1CuO

260.7

±0.2

293.4

±0.3

387.4

±0.4

12.9 ± 0.2

28.6 ± 0.2

57.5 ± 0.1

FPUF-

10CAS

270.0

±0.2

302.0

±0.6

387.7

±0.8

85.7

1.9 ± 0.1

33.6 ± 0.1

63.9 ± 0.1

FPUF-

5OP-

10CAS

256.9

±0.1

293.4

±0.4

389.1

±0.1

95.7

2.6 ± 0.1

33.3 ± 0.1

63.1 ± 0.1

115

FPUF-

5OP-

10CAS-

10EG

259.3

±1.3

291.9

±0.8

393.4

±0.1

101.5

13.2 ± 0.6

28.8 ± 0.1

57.4 ± 0.2

FPUF-

5OP-

10CAS-

0.1CuO

260.7

±0.9

293.9

±0.7

389.6

±0.5

95.7

3.0 ± 0.1

33.1 ± 0.1

63.5 ± 0.2

FPUF-

5OP-

10CAS-

10EG-

0.1CuO

259.4

±0.6

292.6

±0.5

392.1

±0.8

99.5

12.8 ± 0.2

29.1 ± 0.1

57 ± 0.2

* T5%: Temperature at 5% mass loss

# Tmax1: First maximum of thermal decomposition rate; Tmax2: Second maximum.

Mass 1: Mass loss of the 1st decomposition stage;

Mass 2: Mass loss of the 2nd

decomposition stage.

116

Table S2. LOI and UL 94 horizontal test results

Sample

LOI (vol.%)

Self-extinguished

before the 25 mm

mark

Burning

drops

Burning

rate

(mm/min)

FPUF

18.0 ± 0.4

Yes

113

FPUF-10EG

21.0 ± 0.2

Yes

FPUF-0.1CuO

18.0 ± 0.2

113

FPUF-5OP

21.2 ± 0.1

Yes

FPUF-5OP-10EG

23.6 ± 0.2

Yes

FPUF-5OP-0.1CuO

22.0 ± 0.1

Yes

FPUF-5OP-10EG-

0.1CuO

22.8 ± 0.1

Yes

FPUF-10CAS

18.4 ± 0.2

100

FPUF-5OP-10CAS

22.2 ± 0.2

Yes

FPUF-5OP-10CAS-

10EG

24.0 ± 0.2

Yes

FPUF-5OP-10CAS-

0.1CuO

21.4 ± 0.2

Yes

FPUF-5OP-10CAS-

10EG-0.1CuO

24.4 ± 0.2

Yes

117

Figure S1. Photographs of the fire residue of (a) FPUF, (b) FPUF-10EG, (c) FPUF-0.1CuO,

(d) FPUF-5OP, (e) FPUF-5OP-10EG, (f) FPUF-5OP-0.1CuO, (g) FPUF-5OP-10EG-0.1CuO,

(h) FPUF-10CAS, (i) FPUF-5OP-10CAS, (j) FPUF-5OP-10CAS-10EG, (k) FPUF-5OP-

10CAS-0.1CuO and (l) FPUF-5OP-10CAS-10EG-0.1CuO after cone calorimeter measurement

118

4.6 It takes two to tango: Industrial benchmark PU-foams with expandable graphite/P-flame

retardant combinations

Yin Yam Chan and Bernhard Schartel. KGK. 2022, 75(6), 39-46.

Status: This article was accepted and published.

First author contribution:

• Conceptualization

• Methodology

• Validation

• Investigation

• Writing – Original draft, review and editing

• Visualization

• Project administration

Contributions from other authors:

• Bernhard Schartel

o Conceptualization

o Methodology

o Resources

o Writing – Original draft, review and editing

o Supervision

o Project administration

o Funding acquisition

Abstract

Polyurethane foams (PUF) are generally flammable, so they are limited in some applications due to

strict fire safety requirements. In this study, three distinct industrial benchmark polyurethane foams

containing synergistic combinations of expandable graphite (EG) and phosphorous flame retardants (P-

FR) were investigated one by one for their fire performance and smoke behavior. This paper aims to

substantiate the hypothesis that the combination of EG and P-FR used in polyurethane foams yields a

top-notch composite in terms of flame retardancy and smoke behavior by meeting the demanding

requirement of low maximum average heat emission (MARHE) and smoke emission in a variety of

applications, like advanced materials in construction, lightweight materials for railways, and more.

119

Accepted Manuscript

It takes two to tango: Industrial benchmark PU-foams with

expandable graphite/P-flame retardant combinations

Keywords: polyurethane foam, expandable graphite, phosphorus flame retardant

Short summary:

Polyurethane foams (PUF) are generally flammable, so they are limited in some applications

due to strict fire safety requirements. In this study, three distinct industrial benchmark

polyurethane foams containing synergistic combinations of expandable graphite (EG) and

phosphorous flame retardants (P-FR) were investigated one by one for their fire performance

and smoke behavior. This paper aims to substantiate the hypothesis that the combination of EG

and P-FR used in polyurethane foams yields a top-notch composite in terms of flame retardancy

and smoke behavior by meeting the demanding requirement of low maximum average heat

emission (MARHE) and smoke emission in a variety of applications, like advanced materials

in construction, lightweight materials for railways, and more.

Es gehören immer zwei dazu: Industrie-Benchmark PU-

Schäume mit expandierbarem Graphit / P-Flammschutz-

Kombinationen

120

Schlagworte: Polyurethanschaum, Blähgraphit, Phosphor, Flammschutzmittel

Kurze Zusammenfassung:

Polyurethanschaumstoffe (PUF) sind im Allgemeinen entflammbar, so dass sie in einigen

Anwendungen aufgrund strenger Brandschutzanforderungen eingeschränkt sind. In dieser

Studie wurden drei verschiedene industrielle Benchmark-Polyurethanschaumstoffe, die

synergistische Kombinationen von expandierbarem Graphit (EG) und phosphorhaltigen

Flammschutzmitteln (P-FR) enthalten, auf ihr Brandverhalten und Rauchverhalten untersucht.

Ziel dieser Arbeit ist es, die Hypothese zu untermauern, dass die Kombination von EG und P-

FR in Polyurethanschaumstoffen einen erstklassigen Lösungsansatz in Bezug auf Flammschutz

und Rauchverhalten ergibt, der die anspruchsvollen Anforderungen an eine niedrige maximale

durchschnittliche Wärmeemission (MARHE) und Rauchemission in einer Vielzahl von

Anwendungen erfüllt, wie z. B. fortschrittliche Materialien im Bauwesen, Leichtbaumaterialien

für den Schienenverkehr und vieles mehr.

Authors:

Yin Yam Chan1, Bernhard Schartel1*

1 Bundesanstalt für Materialforschung und -prüfung (BAM), Unter den Eichen 87, 12205 Berlin,

Germany

Corresponding author:

Bernhard Schartel

121

Bundesanstalt für Materialforschung und -prüfung (BAM), Unter den Eichen 87, 12205 Berlin,

E-mail: [email protected]

1. Introduction

Flame retardancy mechanisms and thus flame-retardant modes of action and efficiency are

usually quite specific, due to the distinct reactions between flame retardant, additives, fillers,

adjuvants, and polymeric material. [1-4] Considering the application of the product, like

furniture, building material, materials used in transportation, or electrical engineering, flame

retardancy is also specific with respect to the desired protection goal, fire scenario, and

characteristic test specimen. [5-7] Preventing the ignition of housings in electrical engineering

demands for different approaches than reducing the flame spread of a flooring. Further,

different flame retardants are favored for the same polymer depending on whether bulk,

composite, fibers or foam is addressed. As an alternative to replace highly efficient halogenated

flame retardants, current flame retardancy in commercially successful systems is done with

specifically tailored, often multicomponent solutions. [8,9] For this task, a multitude of different

non-halogenated flame retardants usually applied in multicomponent approaches are used to

cover the large number of different products and applications. Rather general approaches such

as using flame retardants containing Br together with Sb2O3 or ammonium polyphosphate with

pentaerythritol are rare. [10] One of these rare and extraordinary champions in currently

commercial flame retardant materials is the synergistic combination of expandable graphite

(EG) and phosphorous flame retardant (P-FR) in polyurethane foams (PUF). [11-13] The

combination of EG and P-FR effectively addresses the challenges of obtaining a low heat

release rate, and low smoke and toxic products emission during burning for PUF. This

combination of EG and P-FR is reported by several studies to show a pronounced synergistic

effect of EG and P-FR. [14-16] This article tries to turn the spotlight on the concept by

122

highlighting three very different industrial benchmarks for flame retarded PUFs. Because the

examples were chosen to cover the breadth of the field, they complement each other to form a

comprehensive picture. They act together strongly, underlining the superior combination of EG

and P-FR.

2. Experimental

2.1 Materials

Sylomer® SR and Sylomer® FR products were purchased from Getzner Werkstoffe GmbH

(Bürs, Austria). Sylomer® is a flexible polyurethane elastomeric foam used in construction,

mechanical engineering, shipbuilding, and rail vehicles. [17] The damping characteristics of

Sylomer® SR and Sylomer® FR are optimized mainly for ships and railway applications and

cover a large range. Sylomer® FR series is a flame retardant product family which can be used

to tackle specific fire safety requirements. Halogen-free flame retardant rigid foam (PB-165)

was purchased from Polymerics GmbH (Berlin, Germany). PB-165 contains flame retardants

in a combination with EG and recycled ABC extinguishing powder (composed of

monoammonium phosphate and ammonium sulfate). [18] PB-165 is used as a filling material

for cable and pipe penetrations in walls, floors, and ceilings. In the case of fire, the foam ensured

fire resistance against fully developed fires. PU high-performance foam (EP4311) was obtained

from Rühl PUROMER GmbH (Friedrichsdorf, Germany). EP4311 is used in lightweight

construction applications like rail vehicles, with EG and P-FR as flame retardants. The samples

of PB-165 were prepared by pressing the 2K cartridge into a metal box and allowing it to rise

freely in the mold. Other samples were already structured as foam when received from the

companies.

123

2.2 Physical and mechanical properties measurements of foam

The apparent density of the samples was measured following ISO 845. The compression stress

of Sylomer® SR and Sylomer® FR products was determined in accordance with ISO 3386-1.

The tensile strength and elongation at break of EP4311 were measured according to ISO 1798.

The compression stress, tensile strength and elongation at break were measured by a Zwick

Z010 universal testing machine (Ulm, Germany). Four cycles per specimen were performed for

the compression test. For both the tensile and the compression tests, four test specimens were

measured for each material.

2.3 Fire behavior measurements

All the specimens used were conditioned at a temperature of 23 ± 2 °C and a relative humidity

of 50 ± 5 % for at least 48 h in accordance with ISO 554. The limiting oxygen index (LOI) was

determined at ambient temperature according to ISO 4589-2, using a specimen size of 150 mm

x 10 mm x 10 mm (length x width x thickness). Cone calorimeter tests were performed in

accordance with ISO 5660. The specimen size used in the cone calorimeter was 100 mm x 100

mm x 13 mm (length x width x thickness) for all of the materials. Each specimen was placed in

an aluminum foil tray and exposed horizontally to an external heat flux of 50 kW m-2 with a

distance of 35 mm between the cone heater and the surface of the specimen. This slightly greater

distance was used to give the materials some freedom to show intumescence without disturbing

the heat flux over the test specimen area. [19,20] Both the oxygen index apparatus and cone

calorimeter were manufactured by Fire Testing Technology Limited (West Sussex, United

Kingdom).

2.4 Smoke and toxic gas measurements

124

A smoke density chamber (SDC), manufactured by Fire Testing Technology Limited (West

Sussex, United Kingdom), was used to determine the smoke generation from the samples in a

static airflow chamber. The measurement was performed in accordance with ISO 5959 on

samples 75 mm x 75 mm x 13 mm (length x width x thickness) in size, mounted in a sample

holder and exposed to an external heat flux of 25 kW m-2, with a distance of 35 mm between

the heat source and the specimen surface under forced flaming conditions.

3. Results and discussion

3.1 Example 1: Sylomer® SR and Sylomer® FR products

Sylomer® SR and Sylomer® FR products show a semi-open cellular structure under scanning

electron microscopy (SEM) in Figure 1. The cells vary in size, corresponding to the density of

the foams. The smaller the cell size, the higher the density.

Sylomer® SR and Sylomer® FR samples are available in distinct densities, and they are used

in different damping applications. The compression stress of the products depends on their

apparent density. As shown in Table 1, the higher the apparent density, the higher the

compression stress at both 40% and 65% compression.

In Table 2, the LOI value of the Sylomer® FR series is much higher than that of the Sylomer®

SR series. The Sylomer® SR series has an LOI value of around 23 vol.-%, while the Sylomer®

FR series has remarkably high LOI values ranging around 38 to 40 vol.-%. Sylomer® FR is

more difficult to ignite and the foams are self-extinguishing.

The cone calorimeter was used to determine the fire behavior of the samples. Important

parameters such as time of ignition (tig), peak heat release rate (PHRR), total heat release (THR),

average of apparent effective heat of combustion (av. EHC), weight percentage of residue at

flameout, and the maximum average rate of heat emission (MARHE) are shown in Table 2.

125

Figure 2 shows the HRR curves of the Sylomer® SR and Sylomer® FR for better display of

the flame retardant efficiency on the HRR. The Sylomer® SR series displayed a plateau at the

beginning of burning due to the collapse of the foam structure. The temperature was high

enough to turn the solid material into a molten melt and became a pool fire. [21] As a result,

the HRR surged to its extremely high PHRR. Once the molten melt was burned off, the flame

extinguished. The density of the Sylomer® SR series and thus the mass of the test specimens

generally affect the PHRR and the THR. The higher the density, the higher the PHRR and THR.

Comparing the Sylomer® SR series and the Sylomer® FR series, the PHRR of Sylomer® FR

is only one tenth as high as that of Sylomer® SR. The burning time of Sylomer® FR series is

much longer than that of the SR series due to the thermal protective layer formed in the

Sylomer® FR series. Sylomer® FR has much higher percentage of residue due to incomplete

pyrolysis of the underlying material shielded by protective layer of the expanded graphite. The

THR of the Sylomer® FR series is significantly reduced. To explain the burning behavior of

the Sylomer® FR series in detail, Figure 3 is used to show the heat release rate curve along with

real-time photos of one of the Sylomer® FR samples, FR418, as an example. ①Once the

specimen was exposed horizontally to the cone heater, the material began to smolder and the

EG expanded even without a flame due to the high temperature. ②When the specimen was

ignited, the HRR immediately increased to only around 100 kW m-2 until a sufficient protective

layer was formed. ③Due to the presence of the underlying expanded graphite, the thickness

of the residue kept increasing. ④ Therefore, the HRR kept decreasing due to the good thermal

protection and the flame was even extinguished after 213 s. The protective layer resulted in

incomplete pyrolysis, thus reducing the THR and PHRR. ⑤ Since we used no frame for the

samples in the cone calorimeter test, artificial edge burning occurred afterwards. The HRR rose

again from 213 s to 250 s due to the edge burning. ⑥ Afterward, the HRR climbed to 250 kW

m-2 and the burning was more intense due to the cracks at the edges and volatile flammable

material escaping from the bottom of the specimen through the cracks. ⑦When volatile fuel

126

from the material was burned out, the flame extinguished at around 600 s. The occurrence of

the second burning and thus second PHRR is due to the edge burning we provoked through

measuring without the frame. This second burning proved that the main flame retardant mode

of action is the efficient protective layer preventing complete pyrolysis in the first burning.

Expanded graphite is apparent as the dominant component of the protective fire residue.

Nevertheless, the interaction with P-FR, matrix, and adjuvants is the key to stabilizing the fire

residue and optimizing its performance. [12,15,22] As seen in Figure 2b, edge burning did not

occur in FR428, and the PHRR was only around 100 kW m-2. The PHRR value is consequently

much lower when compared to other FR series samples. In Table 2, two values were given for

PHRR, THE, and MARHE when the artificial second burning was observed. All Sylomer® FR

samples have a MARHE value below 90 kW m-2, which means they meet the demanding fire

requirement of EN 45545.

Sylomer® SR did not form a protective layer, and almost all the material burned off, taking

SR18 as an example as shown in Figure 4a. In Figure 4b, taking FR418 as an example, the

protective layer expanded to more than four times its original height for Sylomer® FR samples.

The smoke behavior under a static environment was measured by the SDC. The specific optical

density of the samples (Ds) and VOF4 is displayed in Figure 5. The Ds of a Sylomer® SR and

a Sylomer® FR are plotted on the same graph because their apparent densities are comparable.

The Ds of the Sylomer® FR series is very low compared to that of the Sylomer® SR series,

amounting to just 1/10. The combination of EG and P-FR in Sylomer® provides the material

with excellent smoke suppression. The thermal protective layer created by expanded graphite

yielded crucially incomplete pyrolysis as the basis for smoke reduction. Sylomer® FR emitted

an extremely low amount of smoke in the first 4 minutes, giving people more time to evacuate

from fires.

127

The investigated products are applied as damping materials; hence they are a kind of flexible

PUF. Reducing both their high PHRR and pronounced smoke emission by a factor of 10 at the

same time to fulfill the demanding EN 45545 requirements is an extraordinary achievement.

Usually, such a challenge in flexible PUF development turns out to be a mission impossible; it

cannot be realized by the common approaches of flame inhibition, charring, and protective layer.

The combination of EG and P-FR is optimized, extinction occurs when 90% of the material

remains not pyrolyzed, which is impressive.

3.2 Example 2: PB-165

Figure 6 shows the morphology of PB-165. The cell structure looks somewhat destroyed, like

ragged spheres with tattered cell walls. Nevertheless, the structure is rather homogeneous and

indicates a rather low density.

Table 3 lists the fire performance and smoke emission behavior data for PB-105. The LOI value

of PB-165 is very high (>50 vol.-%), which means that it is not only self-extinguishing, but also

difficult to burn in an existing fire. In Figure 7a, the HRR spiked to the first peak within 10 s

after ignition, as the material burned without efficient protection at the beginning. The

protective layer was gradually built up by the expanded graphite, forming an excellent thermal

protective layer, and the HRR decreased to around 58 kW m-2. The HRR increased towards the

end of burning due to some cracks in the protective layer. The HRR then reached its second

peak. The HRR curve pattern shows a pronounced two-peak shape, one sharp peak at the

beginning and one broad peak at the end, separated by a clear minimum in HRR in between.

The HRR curve pattern is similar to wood [23], also a porous charring material. For wood, a

second burning stage at the end is proposed, as the destruction of intermediate char may be

accompanied by a second pyrolysis front at high temperatures. The MARHE value of PB-165

is around 80 kW m-2, which is low enough to reach the requirement of European EN 45545

128

standard. Figure 7b shows the Ds and transmission curves from the SDC measurement. The

amount of smoke released was extremely low. At the end of the test, the light transmission

remained high, which means that visibility for escape is high. As shown in Figure 8, after the

cone calorimeter measurement the height of the specimen expanded to 3 times its original height,

from 13 mm to 40 mm. This thick protective layer significantly reduced the heat transfer to the

underlying material and thereby reduced the HRR.

PB-165 was developed for fire resistant applications in building construction. It is a PUF foam

used, for instance, to close openings in fire protection walls. Indeed, it is developed to resist a

fully developed fire, a fire protection goal that transcends the usual domain of flame retardants.

Reduction in HRR and smoke emission is not what the material is designed for. The excellent

flame retardancy of PUF achieved is a byproduct, nevertheless this example underlines the

outstanding and superior effect of combining EG with P-FR and adjuvants.

3.3 Example 3: EP4311

The SEM image of EP4311 is displayed in Figure 9. Most of the cells are closed. Some cell

walls and structures are quite large. The multicellular structure appears to approximate

separated bubbles dissolved in the polymer matrix, as is typical for a foam with higher density.

Table 4 lists the physical and mechanical properties of EP4311.

The very high LOI value shown in Table 5 proved that EP4311 is not flammable. The LOI

value of over 50 vol.-% means it is self-extinguishing and difficult to burn.

Figure 10a shows the HRR curve of EP4311. Since the specimen without protection was

initially exposed to the cone heater, the HRR quickly reached its PHRR (85 kW m-2). After that,

the HRR decreased rapidly because of the formation of thermal protective layers dominated by

the expanded graphite in the system. The HRR remains at around 50 kW m-2 for some time.

129

The HRR of the EP4311 was further decreased because of the accumulation of expanded

graphite after 188 s, which can be interpreted as flameout, and reached a plateau of very low

HRR, marking a kind of afterglow. There is no obvious second PHRR found in the graph. As

shown in Figure 10b, the Ds was still low at the end of the test, which means the concentration

of smoke was low to provide people more time to escape from fires. Figure 11 shows the residue

after the cone calorimeter measurement of EP4311. The residue increased to 4 times its original

height, forming a thermal protective layer to reduce heat transfer from the top to the material

underneath.

EP4311 is a typical material used in lightweight construction. This PUF is used to replace bulk

material, as it weighs only half as much as a technical polymer. Railroad vehicles are a typical

application for this type of product, as they must meet the rigorous demands of the EN 45545

standard. The combination of EG with P-FR is used to achieve the demanding fire safety level.

4 Conclusion

All the examples with flame retardants (EG and P-FR) exhibit a very high LOI value, low HRR

and low Ds, reducing fire and smoke hazards. The combination of EG and P-FR serves as an

excellent way to improve the flame retardancy of PUF products. The synergistic effect of EG

and P-FR creates more char residue during burning, forming an efficient thermal protection

barrier to prevent further pyrolysis and burning of the underlying materials. Expanded graphite

dominates the fire residue. The height of the protective residue is around 3 to 5 times the original

height of the test specimens. The very efficient thermal insulation causes extinction far before

pyrolysis is completed. The results indicate that as little as just 10 wt.-% of the PUF may be

pyrolyzed. The massive reduction in consumption of the PUF reduces the fuel release rate and

fire load. Furthermore, it is a direct and efficient way to reduce the release of smoke and toxic

products.

130

To conclude, the combination of EG and P-FR is used successfully in commercial PUF systems

and achieves the highest standards like MARHE and the smoke requirements of EN 45545. The

combination of EG and P-FR is one of the rare general highly efficient approaches in flame

retardancy.

Acknowledgements

The authors are grateful to Dietmar Schulze for his contribution to the mechanical tests of the

materials and Tina Raspe and Dr. Simone Krüger for their support with the SDC. Dr. Ulrich

Fehrenbacher from RÜHL PUROMER GmbH is acknowledged for providing the material

EP4311.

References

[1] A. B. Morgen, J. W. Gilman. An overview of flame retardancy of polymeric materials:

Application, technology, and future directions. Fire Mater. 37 (2013) 259–279, doi:

10.1002/fam.2128.

[2] M. Lewin. Synergistic and catalytic effects in flame retardancy of polymeric materials

– An overview. J. Fire Sci. 17 (1999) 3–19, doi: 10.1177/073490419901700101.

[3] J. Q. Wang, W. K. Chow. A brief review on fire retardants for polymeric foams. J. Appl.

Polym. Sci. 97 (2005) 366–376, doi: 10.1002/app.21758.

[4] B. Schartel. Multicomponent flame retardants. In: Non-Halogenated Flame Retardant

Handbook, 2nd ed.; A. B. Morgan Ed. Scrivener Publishing LLC, Beverly, MA, 2022, pp. 413–

474.

[5] A. B. Morgan. Revisiting flexible polyurethane foam flammability in furniture and

bedding in the United States. Fire Mater. 45 (2021) 68–80, doi: 10.1002/fam.2848.

[6] E. D. Weil, S. V. Levchik. Flame retardants in commercial use or development for

polyurethanes, polyisocyanurates, and polyureas. In Flame Retardants for Plastics and Textiles:

131

Practical Applications, 2nd ed.; E. D. Weil and S. V. Levchik Eds. Hanser Publishers: Munich,

Germany, 2016, pp. 205–230.

[7] Y. Tan, V. Wachtendorf, T. Kukofka, P. Klack, J. Ruder, X. B. Lin, B. Schartel.

Degradation of flame retardance: A comparison of ethylene-vinyl acetate and low-density

polyethylene cables with two different metal hydroxides. J. Appl. Polym. Sci. 138, (2021)

app50149, doi: 10.1002/app.50149.

[8] E. D. Weil, S. V. Levchik. Commercial flame retardancy of polyurethanes. J. Fire Sci.

22 (2004) 183–210, doi: 10.1177/0734904104040259.

[9] B. Schartel. Phosphorus-based flame retardancy mechanisms – Old hat or a starting

point for future development? Materials 3 (2010) 4710–4745, doi: 10.3390/ma3104710.

[10] M. Lewin. Synergism and catalysis in flame retardancy of polymers. Polym. Adv.

Technol. 12 (2001) 215–222, doi: 10.1002/pat.132.

[11] A. Wolska, M. Gozdzikiewicz, J. Ryszkowska. Thermal and mechanical behavior of

flexible polyurethane foams modified with graphite and phosphorous fillers. J. Mater. Sci. 47

(2012) 5627–5634, doi: 10.1007/s10853-012-6433-z.

[12] W. H. Rao, W. Liao, H. Wang, H. B. Zhao, Y. Z. Wang. Flame-retardant and smoke-

suppressant flexible polyurethane foams based on reactive phosphorus-containing polyol and

expandable graphite. J. Hazard. Mater. 360 (2018) 651–660, doi:

10.1016/j.jhazmat.2018.08.053.

[13] M. Modesti, A. Lorenzetti. Halogen-free flame retardants for polymeric foams. Polym.

Degrad. Stab. 78 (2002) 167–173, doi: 10.1016/S0141-3910(02)00130-1.

[14] Y. X. Hu, Z. J. Zhou, S. S. Li, D. Yang, S. Zhang, Y. K. Hou. Flame retarded rigid

polyurethane foams composites modified by aluminum diethylphosphinate and expanded

graphite. Front. Mater. 7 (2021) 629284, doi: 10.3389/fmats.2020.629284.

[15] Y. Y. Chan, B. Schartel. It takes two to tango: Synergistic expandable graphite-

phosphorus flame retardant combinations in polyurethane foams. Polymers 14 (2022) 2562, doi:

10.3390/polym14132562.

[16] A. Wilke, K. Langfeld, B. Ulmer, V. Andrievici, A. Hörold, P. Limbach, M. Bastian, B.

Schartel. Halogen-free Multicomponent Flame Retardant Thermoplastic Styrene-Ethylene-

Butylene-Styrene Elastomers Based on Ammonium Polyphosphate – Expandable Graphite

Synergy. Ind. Engin. Chem. Res. 56 (2017) 8251–8263, doi: 10.1021/acs.iecr.7b01177.

132

[17] Getzner Werstoffe GmbH. Sylomer® The leading material for vibration isolation.

https://www.getzner.com/en/products/sylomer.

[18] German Patent DE 103 10 006 B4 (2005) Halogenfreier, flammhemmender

Polyurethanhartschaum für brandschutztechnische Anwendungen. Polymerics GmbH, P.

Brusenbach, A. Leistner, A. Leistner, H.-S. Petras.

[19] B. Schartel, M. Bartholmai, U. Knoll. Some comments on the use of cone calorimeter

data. Polym. Degrad. Stabil. 88 (2005) 540–547, doi: 10.1016/j.polymdegradstab.2004.12.016.

[20] B. Schartel, T. R. Hull. Development of fire-retarded materials – Interpretation of cone

calorimeter data. Fire Mater. 31, (2007) 327–354, doi: 10.1002/fam.949.

[21] M. Günther, A. Lorenzetti, B. Schartel. Fire phenomena of rigid polyurethane foams.

Polymers. 10 (2018) 1166, doi: 10.3390/polym10101166.

[22] Y. Y. Chan, C. Ma, F. Zhou, Y. Hu, B. Schartel. Flame retardant flexible polyurethane

foams based on phosphorous soybean-oil polyol and expandable graphite. Polym. Degrad. Stab.

191 (2021) 109656, doi: 10.1016/j.polymdegradstab.2021.109656.

[23] S. Rabe, P. Klack, H. Bahr, B. Schartel. Assessing the fire behavior of woods modified

by N-methylol crosslinking, thermal treatment, and acetylation. Fire Mater. 44 (2020) 530–539,

doi: 10.1002/fam.2809.

133

Table 1 Apparent density and compression stress results of Sylomer® SR and Sylomer® FR

products

Sample

Apparent

density/ kg

m-3

Compression

stress at 40%

compression

/ kPa

Compression

stress at 65%

compression/

kPa

SR18

191.0 ± 9.4

36.4 ± 0.1

124.5 ± 13.9

SR28

218.2 ± 1.5

48.4 ± 0.5

149.7 ± 2.7

SR42

251.1 ± 0.4

72.8 ± 0.3

226.4 ± 1.9

SR55

267.4 ± 0.1

98.7 ± 0.3

332.2 ± 3.4

SR110

407.3 ± 0.4

232.5 ± 1.9

1064 ± 7.4

FR418

262 ± 0.6

38.1 ± 0.5

132.2 ± 1.5

FR428

293.6 ± 1.6

65.4 ± 0.8

267.7 ± 5.7

FR442

332.7 ± 10

87.4 ± 1.2

345.8 ± 15

FR455

383.1 ± 0.1

108.2 ± 1.4

463.2 ± 1.7

FR4110

484.8 ± 29

238.7 ± 1.8

1383.7± 55.9

134

Table 2 Fire performance and smoke behavior results of Sylomer® SR and Sylomer® FR

products, with the data in brackets are the overall burning result

* VOF4 is a cumulative value of specific optical densities in the first 4 min. of the test

Sample

LOI

/ vol.-%

tig/ s

PHRR/

kW m-2

THR/ MJ

m-2

Average

EHC/

MJ kg-1

Residue

flameout

/ wt.-%

MARH

E/ kW

m-2

Max.

specific

optical

density

(Ds max)

VOF4*

SR18

26.0±0.2

9±1

1441±21

63.2±3.2

27.0±0.1

1.9±0.1

589±1

386±36

879±37

SR28

22.2±0.1

9±1

1718±4

73.2±0.4

27.4±0.1

2.1±0.1

645±8.1

396±20

775± 3

SR42

22.2±0.2

8±1

1610±104

83.2±0.8

27.2±0.3

2.6±0.2

640±8

468±12

838±34

SR55

23.6±0.2

9±1

1980±80

89.6±1.0

27.6±0.4

2.8±0.2

686±10

467±15

838±2

SR110

22.5±0.1

10±1

2691±190

129±17.9

27±2.7

6.5±3.3

785±74

494±7

789±14

FR418

40.2±0.1

5±1

106±4

(270±17)

15±0.5

(46.2±1.5)

19.6±0.5

28.1±0.5

71±2

(87±9)

42±1

41±1

FR428

38.0±0.1

7±1

104±7

7.5±1.7

18.9±0.8

89.3±2.1

68.3±3.5

49.0±0.3

30.2±1.9

FR442

40.1±0.1

8±1

110±9

(298±23)

22±2

(65±5)

20.8±1.6

27.4±2.1

63.8±4.9

(84±25)

44.8±0.9

22.4±2.1

FR455

39.6±0.2

8±1

103±4

(293±12)

25±1

(71±3)

20.6±0.8

28.2±1.1

62±2

(86±15)

50.2±0.5

18.3±0.9

FR4110

34.4±0.2

12±1

105±5

(248±16)

14±1

(100±6)

20.4±1.5

18.0±5.7

68±4

(113±12)

58±3

11.± 0.6

135

Table 3 Fire performance and smoke behavior data of PB-165

Apparent density/

kg m-3

LOI/

vol.-%

Maximum specific

optical density (Ds max)

VOF4

PB-165

134.7±1.3

51.3±0.1

38.9±6.5

66.4±13.2

tig/ s

PHRR/ kW m-2

THR/

MJ m-2

Average EHC/ MJ kg-1

Residue at flameout/

wt.-%

2.5±0.5

112±10

19.7±0.5

18±0.1

37.3±2.6

Table 4 Apparent density and mechanical test results of EP4311

Apparent density / kg

m-3

Tensile strength /

kPa

Elongation at break / %

EP4311

505.4 ± 3.4

986.5 ± 25.7

36.5 ± 1.8

Table 5 Fire performance and smoke behavior data of EP4311

LOI/

vol.-%

Maximum specific

optical density (Ds max)

VOF4

EP4311

54.4 ± 0.2

65.0 ± 11.8

20.4 ± 0.5

tig/ s

PHRR/

kW m-2

THR/ MJ m-2

Average

EHC/ MJ kg-1

Residue at

flameout/ wt.-%

MARHE/

kW m-2

11± 0.1

86.7 ± 1.8

7.8 ± 1.7

16.1 ± 0.7

92.3 ± 1.9

51.1 ± 0.7

136

Figure 1 SEM images of Sylomer® (a1) SR18, (a2) FR418, (b1) SR28, (b2) FR428, (c1)

SR42, (c2) FR442, (d1) SR55, (d2) FR455, (e1) SR110, and (e2) FR4110

137

Figure 2 Heat release rate curves of (a) SR18 & FR418, (b) SR28 & FR428, (c) SR42 &

FR442, (d) SR55 & FR455, and (e) SR110 & FR4110

138

Figure 3 Heat release rate curve with real-time photos of FR418

Figure 4 Residue after cone calorimeter measurement of (a) SR18 and (b) FR418

139

Figure 5 Specific optical density curves of (a) SR18 & FR418, (b) SR28 & FR428, (c) SR42

& FR442, (d) SR55 & FR 455, and (e) SR110 & FR4110

140

Figure 6 SEM image of PB-165

Figure 7 (a) Heat release rate curve and (b) specific optical density and transmission curves of

PB-165

141

Figure 8 Residue after the cone calorimeter measurement of PB-165

Figure 9 SEM image of EP4311

Figure 10 (a) Heat release rate curve and (b) specific optical density and transmission curves

of EP4311

142

Figure 11 Residue after the cone calorimeter measurement of EP4311

143

4.7 It takes two to tango: Synergistic expandable graphite – phosphorus flame retardant

combinations in polyurethane foams

Yin Yam Chan and Bernhard Schartel. Polymers. 2022, 14, 2562. Doi: 10.3390/polym14132562.

Status: This article was accepted and published. https://doi.org/10.3390/polym14132562

First author contribution:

• Conceptualization

• Methodology

• Validation

• Investigation

• Writing – Original draft, review and editing

• Visualization

• Project administration

Contributions from other authors:

• Bernhard Schartel

o Conceptualization

o Methodology

o Resources

o Writing – Original draft, review and editing

o Supervision

o Project administration

o Funding acquisition

Abstract

Due to the high flammability and smoke toxicity of polyurethane foams (PUFs) during burning, distinct

efficient combinations of flame retardants are demanded to improve the fire safety of PUFs in practical

applications. This feature article focuses on the one of the most impressive halogen-free combinations

in PUFs: expandable graphite (EG) and phosphorus-based flame retardants (P-FRs). The synergistic

effect of EG and P-FRs mainly superimposes the two modes of action, charring and maintaining a

thermally insulating residue morphology, to bring effective flame retardancy to PUFs. Specific

interactions between EG and P-FRs, including the agglutination of the fire residue consisting of

expanded-graphite worms, yields an outstanding synergistic effect, making this approach the latest

champion to fulfill the demanding requirements for flame-retarded PUFs. Current and future topics such

as the increasing use of renewable feedstock are also discussed in this article.

Citation: Chan, Y.Y.; Schartel, B. It

Takes Two to Tango: Synergistic

Expandable Graphite–Phosphorus

Flame Retardant Combinations in

Polyurethane Foams. Polymers 2022,

14, 2562. https://doi.org/10.3390/

polym14132562

Academic Editors: Jelena Vasiljevi´c

and Ivan Jerman

Received: 27 May 2022

Accepted: 21 June 2022

Published: 23 June 2022

Publisher’s Note: MDPI stays neutral

with regard to jurisdictional claims in

published maps and institutional affil-

iations.

Licensee MDPI, Basel, Switzerland.

This article is an open access article

distributed under the terms and

conditions of the Creative Commons

Attribution (CC BY) license (https://

creativecommons.org/licenses/by/

4.0/).

polymers

Article

It Takes Two to Tango: Synergistic Expandable

Graphite–Phosphorus Flame Retardant Combinations in

Polyurethane Foams

Yin Yam Chan and Bernhard Schartel *

Bundesanstalt für Materialforschung und-prüfung (BAM), Unter den Eichen 87, 12205 Berlin, Germany;

[email protected]

*Correspondence: bernhar[email protected]; Tel.: +49-30-8104-1021

Abstract:

Due to the high flammability and smoke toxicity of polyurethane foams (PUFs) during

burning, distinct efficient combinations of flame retardants are demanded to improve the fire safety of

PUFs in practical applications. This feature article focuses on one of the most impressive halogen-free

combinations in PUFs: expandable graphite (EG) and phosphorus-based flame retardants (P-FRs).

The synergistic effect of EG and P-FRs mainly superimposes the two modes of action, charring and

maintaining a thermally insulating residue morphology, to bring effective flame retardancy to PUFs.

Specific interactions between EG and P-FRs, including the agglutination of the fire residue consisting

of expanded-graphite worms, yields an outstanding synergistic effect, making this approach the

latest champion to fulfill the demanding requirements for flame-retarded PUFs. Current and future

topics such as the increasing use of renewable feedstock are also discussed in this article.

Keywords:

synergy; phosphorus-containing flame retardant; expandable graphite; polyurethane

foams

1. Introduction

As fire safety has always been a major concern, fire protection is in high demand.

One of the key approaches to improving fire protection entails adding flame retardants to

polymeric materials, because most synthetic polymers are easily ignited due to their high

content of hydrocarbons, an excellent fuel for fires. Currently, efficient flame retardancy is

achieved through specific solutions tailored to different kinds of polymeric materials, as

they have different properties [

–

]. Flame retardancy is specific with respect to the flame-

retardant mechanisms and to the flame retardant’s reactions with polymeric materials [

–

]

and with other ingredients, such as additional flame retardants, fillers/fibers, additives,

adjuvants, and synergists [8].

Flame retardancy is specific with respect to the protection goal and fire scenario;

ignition scenarios require different approaches from developing fires or fully developed

fires [

]. Different flame retardants are even favored for the same polymer depending on

whether it is applied in bulk, as a composite, or in the form of fibers or foam. Thus, a

multitude of different flame retardants in various combinations are used to protect the

entire spectrum of foam-containing consumer goods [

]. Generalized approaches of

efficient combinations working in different matrices, such as flame retardants containing Br

combined with Sb

or ammonium polyphosphate (APP) with pentaerythritol (PER), are

rare; furthermore, environmental concerns mean that halogen-free systems are preferred.

One of the champions among the currently proposed flame retardants that has found

application in today’s products is the synergistic combination of expandable graphite

(EG) with a phosphorous flame retardant (P-FR). This approach has become legendary for

the excellent flame retardancy it provides to polyurethane foams (PUFs) [

–

]. In this

feature article, we turn the scientific spotlight on the concept, mechanisms, and role of

Polymers 2022,14, 2562. https://doi.org/10.3390/polym14132562 https://www.mdpi.com/journal/polymers

Polymers 2022,14, 2562 2 of 24

the synergistic combinations of EG and P-FRs in PUFs in order to evoke the applause this

approach deserves [15–18].

1.1. Polyurethane Foams (Flexible and Rigid)

Polyurethane foams (PUFs) have been used in a wide range of applications because

their physical and mechanical properties can be customized by changing the chemical

composition. PUFs are divided into two main categories: flexible polyurethane foams

(FPUFs) and rigid polyurethane foams (RPUFs) [

–

]. The main chemicals used in the

formulation for PUFs are polyols, isocyanates, catalysts, surfactants, and blowing agents.

The differences in the physical and mechanical properties of FPUFs and RPUFs depend

mainly on the chemical characteristics of the reactants—the polyols and the isocyanates.

Urethane linkages form in PUFs through a polyaddition reaction between the hydroxyl

group of polyols and the NCO groups of isocyanates [

–

]. Nevertheless, the function-

ality of polyols and the type of isocyanate used are different in RPUFs and FPUFs. The

relationship among the average molecular weight, functionality, and OH value of polyols is

shown in Equation (1). These parameters are key characteristics that define the properties of

polyols and ultimately affect the property profile of polyurethanes when polyols react with

diisocyanates. For example, increasing the OH value leads to a higher crosslink density

in polyurethane.

Mn=

z×56106

OH value (1)

where Mnand zare the average molecular weight and the functionality, respectively.

For RPUFs, polyols with shorter chains exhibiting higher functionality (z= 2.5–5)

are combined with polymeric diphenylmethane diisocyanate (pMDI) in order to gener-

ate more crosslinks, providing more strength, thereby increasing rigidity. As a result,

the apparent density of RPUFs is higher than that of FPUFs due to the former’s higher

crosslink density [

]. Accordingly, RPUFs are more commonly used in construction,

transportation, and refrigeration because of their extremely low thermal conductivity due

to their closed-cell structure, which is superior to that of other commercially available

insulation materials [

]. Because FPUFs are so flexible, long-chained polyols with lower

functionality (z= 2–3) and toluene diisocyanate (TDI) are typically the main components

added [

]. The open-cell structure of FPUFs contains cell window, strut, and strut join.

Such morphology offers different degrees of cushioning, making it a frequent material

choice in furnishings, automotive seating, mattresses, and packaging.

1.2. Polyisocyanurate Foams and Polyurethane Foams

Although polyisocyanurate foams (PIRFs) and polyurethane foams (PUFs) have simi-

lar chemical compositions, PIRFs exhibit considerably better flame retardancy than PUFs.

As PUFs have a more balanced equivalent weight ratio of the isocyanate group and the

hydroxyl group of polyols (NCO/OH ratio

≈

1.05–1.1), they mainly form urethane linkages,

while the much higher excess amount of isocyanates in PIRFs usually yields isocyanurates

via trimerization reaction. Due to the higher content of ring structures, PIRFs produce more

char during burning, which ensures superior fire behavior. Günther et al. [

] compared

the morphology of PIRF and RPUF residues with cone calorimeter measurements; the

better fire behavior of PIRFs was attributed to the dense, thick cellular structure residue

retained to some degree as thermal insulation, while RPUFs showed a thin, brittle residue

layer. Therefore, the residue from PIRFs protects the underlying material better than the

residue from PUFs during burning.

1.3. Flammability and Smoke Toxicity during Burning of Polyurethane Foams

Despite all the advantages enjoyed by PUFs, one of their major problems is high

flammability. Regardless of whether RPUFs or FPUFs are used, both forms of PUFs present

highly porous and cellular-structured material that easily catches fire. Because the cell

walls and struts are thermally thin, they can be heated to the ignition temperature swiftly,

Polymers 2022,14, 2562 3 of 24

causing early ignition [

]. Meanwhile, the low thermal conductivity makes the entire

specimen thermally thick, which concentrates the heat at the surface, resulting in rapid

flame spread. Therefore, the flame retardancy of polyurethane foams needs to be improved

to meet high fire protection standards. Due to the open-cell structure and higher surface-

to-mass ratio, FPUFs have higher flammability than RPUFs [

]. FPUFs show a lower

tendency to char and a higher tendency to yield liquid products and thus collapse and

yield pool fires. Figure 1, which illustrates the chemical structure, shows that PUFs are

predominantly composed of combustible elements such as carbon and hydrogen, which

increase the growth rate of fire. Apart from their flammability, PUFs evolve poisonous

gases during burning, including carbon monoxide, nitrogen oxides, and hydrogen cyanide,

as well as a large amount of smoke particles, all of which threaten human lives. The quality

and quantity of poisonous gas evolved from the same material vary with different oxygen

concentrations and burning temperatures [32,33].

Polymers 2022, 14, x 3 of 26

walls and struts are thermally thin, they can be heated to the ignition temperature swiftly,

causing early ignition [30]. Meanwhile, the low thermal conductivity makes the entire

specimen thermally thick, which concentrates the heat at the surface, resulting in rapid

flame spread. Therefore, the flame retardancy of polyurethane foams needs to be im-

proved to meet high fire protection standards. Due to the open-cell structure and higher

surface-to-mass ratio, FPUFs have higher flammability than RPUFs [31]. FPUFs show a

lower tendency to char and a higher tendency to yield liquid products and thus collapse

and yield pool fires. Figure 1, which illustrates the chemical structure, shows that PUFs

are predominantly composed of combustible elements such as carbon and hydrogen,

which increase the growth rate of fire. Apart from their flammability, PUFs evolve poi-

sonous gases during burning, including carbon monoxide, nitrogen oxides, and hydrogen

cyanide, as well as a large amount of smoke particles, all of which threaten human lives.

The quality and quantity of poisonous gas evolved from the same material vary with dif-

ferent oxygen concentrations and burning temperatures [32,33].

Figure 1. Chemical structure of flexible polyurethane foams.

1.4. Commercial Flame Retardants for PUFs

To cope with the flammability of PUFs, additive and reactive flame retardants are

often introduced into the matrix to delay ignition and reduce heat release in the event of

a fire, thereby slowing flame spread. Compared with RPUFs, it is more difficult to enhance

structural frame retardancy of FPUFs. This is because flame retardants added physically

often increase the viscosity of the polymer system and limit foam growth. The flame re-

tardant is mainly embedded in the thin cell struts, causing the structure to collapse under

the weight of the additive during the foaming process.

In the past, dispersing halogenated flame retardants such as organochlorine and or-

ganobromine compounds in polyurethane foams were very attractive for the industry,

because they work effectively in the gas phase and greatly reduce heat release during

burning [10]. However, the hydrogen halides released from halogenated flame retardants

during burning are highly corrosive, are toxic to human beings, and may pollute the en-

vironment. Due to environmental and biological health concerns, some countries have

already considered legislation restricting the use of flame retardants containing halogen.

As a result, more and more halogen-free and environmentally friendly flame retardants

have been developed and used in recent decades [34]. Today, dimethyl methylphospho-

nate (DMMP) [35], triaryl phosphates [36], melamine [31,37], aluminum hydroxide (ATH),

Figure 1. Chemical structure of flexible polyurethane foams.

1.4. Commercial Flame Retardants for PUFs

To cope with the flammability of PUFs, additive and reactive flame retardants are

often introduced into the matrix to delay ignition and reduce heat release in the event

of a fire, thereby slowing flame spread. Compared with RPUFs, it is more difficult to

enhance structural frame retardancy of FPUFs. This is because flame retardants added

physically often increase the viscosity of the polymer system and limit foam growth. The

flame retardant is mainly embedded in the thin cell struts, causing the structure to collapse

under the weight of the additive during the foaming process.

In the past, dispersing halogenated flame retardants such as organochlorine and

organobromine compounds in polyurethane foams were very attractive for the industry,

because they work effectively in the gas phase and greatly reduce heat release during

burning [

]. However, the hydrogen halides released from halogenated flame retardants

during burning are highly corrosive, are toxic to human beings, and may pollute the en-

vironment. Due to environmental and biological health concerns, some countries have

already considered legislation restricting the use of flame retardants containing halogen.

As a result, more and more halogen-free and environmentally friendly flame retardants

have been developed and used in recent decades [

]. Today, dimethyl methylphosphonate

(DMMP) [

], triaryl phosphates [

], melamine [

], aluminum hydroxide (ATH), ex-

pandable graphite (EG) [

–

], and ammonium polyphosphate (APP) [

] are common

Polymers 2022,14, 2562 4 of 24

halogen-free additive flame retardants for polyurethane foams. Many studies have found

that mixing two flame retardants or combining two or more flame-retardant elements

in a single compound can increase the flame-retardant efficiency. This phenomenon is

called synergism [

Li et al. [44]

investigated the flame retardancy of RPUFs combined

with DMMP and modified APP. They found that DMMP and modified APP enhanced

flame retardancy through good coordination in the gas phase and the condensed phase.

Wang et al. [45]

synthesized a flame retardant containing phosphorus and nitrogen in

RPUFs. The foam with the flame retardant formed a protective char layer, which enhanced

flame retardancy. Tris(1-chloro-2-propyl) phosphate (TCPP) and tris(1,3-dichloro-2-propyl)

phosphate (TDCP), both with two flame-retardant chemical elements (i.e., halogen and

phosphorus), are still common additive flame retardants for polyurethane foams [

]. Most

additive flame retardants deteriorate the morphology and mechanical properties of poly-

mers. Hence, reactive flame retardants are an alternative to improve flame retardancy

by chemically bonding to the polyurethane structure without excessively damaging the

PUF structure. Phosphorous polyols are used as reactive flame retardants to replace petro-

chemical polyols in the formulation. Commercial non-halogenated phosphorous polyols

such as Exolit

OP 550 and Exolit

OP 560 from Clariant AG (Muttenz, Switzerland) are

successfully used in the industry. However, it is worth noting that FPUFs are sensitive to

the hydroxyl number of polyols [

]. Higher hydroxyl values of polyols may cause the

structure of FPUFs to collapse. Therefore, determining the appropriate amounts and types

of polyols is the key to successful foaming.

2. Task

2.1. Burning Behavior of Rigid and Flexible Polyurethane Foams

In terms of burning behavior, FPUFs can be ignited more easily than RPUFs, and fire

propagates more quickly because of their lower density and open-cell structure [

while RPUFs have a higher density and a closed-cell structure [

]. The curves of the

heat release rate (HRR) of FPUFs from cone calorimeter measurements are displayed in

Figure 2. For FPUFs, the curve exhibits three stages. In the first step (i), the surface of the

foam is heated up; then, decomposition is initiated, and the foam ignites. According to

the two-step decomposition of polyurethane, mainly, urethane bonds decompose, and the

volatile pyrolysis products of the hard segments feed the flame. After this ignition

stage (i)

in stage (ii), the foam is covered by a molten layer of pyrolyzing polyurethane, such that

the foam burns, collapses, and forms a pool of intermediate liquid pyrolysis products.

After the first peak or plateau-like burning in stage (ii), the heat release rate surges to

another, higher peak in stage (iii), because the remaining material burns in a violent pool

fire [

]. The differences between the different burning stages can be described by

the temperature–thickness relationship. Figures 2and 3show the temperature–thickness

relationship of FPUFs and RPUFs during different stages of burning, respectively. At the

beginning of burning, individual cell walls or struts behave as thermally thin materials,

such as a film or fiber. After ignition in stage (i), the very top layer of FPUFs at d

consumed under the influence of thermal radiation, forming a thin pyrolysis zone at the

first pyrolysis temperature (T

). The yielded liquid pyrolysis products mainly belong

to the soft segments, and the volatiles released mainly belong to the hard segments. The

excellent thermal insulation of the foam results in a rapid decrease in temperature across the

intact foam, as the entire unmolten part is thermally thick. As heating continues,

stage (ii)

reached, with the next few layers from the top of the FPUF also collapsing and melting,

forming a thicker pyrolysis zone. Due to the good convection of the molten melt from d

, the melt is considered to reach the same pyrolysis temperature at T

. As the remaining

unburned material is still thermally thick, its temperature decreases inversely toward the

bottom of the material. The cellular structure melts and collapses in stage (ii), resulting

in a pool fire (iii), generally at the second pyrolysis temperature (T

) from d

to d. The

high fluidity of the melt under high temperatures exhibits a constant temperature due to

Polymers 2022,14, 2562 5 of 24

convection. Almost no heat flux is attributed to further heating in stage (iii), but the heat

flux is completely transferred to pronounced pyrolysis.

Polymers 2022, 14, x 5 of 26

high fluidity of the melt under high temperatures exhibits a constant temperature due to

convection. Almost no heat flux is attributed to further heating in stage (iii), but the heat

flux is completely transferred to pronounced pyrolysis.

RPUFs behave quite differently from FPUFs during burning, showing the typical

HRR curve for residue-forming materials in Figure 3 [9,26,27]. After ignition in stage (i),

they reach the PHRR immediately, undergoing distinct charring at Tp1, with no structural

collapse and no formation of a pool fire because of their higher crosslink density. The char

on the top acts as a protective layer, shielding the material underneath. The PHRR is sub-

sequently followed by steady burning in stage (ii) at a lower HRR. The pyrolysis front at

Tp2 continuously consumes the material downward from the top (d0) to d1. Due to the

effective protective layer formed, the temperature of the unburned material from d1 to d

decreases inversely toward the bottom of the material. The length of the steady-burning

phase in the HRR curve depends on the amount of combustible material [26]. Therefore,

less heat is released by RPUFs as the char yield increases [9,27]. In stage (iii), the pyrolysis

front at Tp2 moves to the bottom of the material, and the flame is finally extinguished.

Figure 2. Temperature–thickness relationship of flexible polyurethane foams during different burn-

ing stages.

Figure 2.

Temperature–thickness relationship of flexible polyurethane foams during different burn-

ing stages.

Polymers 2022, 14, x 6 of 26

Figure 3. Temperature–thickness relationship of rigid polyurethane foams during different burning

stages.

2.2. Role of Selecting Contents of Isocyanate, Polyol, Foaming Agent, and Flame Retardants

Polyurethane chemistry is based on the high reactivity of isocyanates. Diisocyanates

are organic compounds with two isocyanate groups, which are widely used to link polyols

together through an exothermic reaction between isocyanates and hydroxyl groups in or-

der to build crosslinked polyurethane. The content of diisocyanates in the formulation

influences the thermal stability, rigidity, and fire behavior of PUFs. For instance, any ex-

cess isocyanates are converted into trimers by trimerization (see Figure 4), called isocy-

anurate rings [52]. Isocyanurates improve flame retardancy because the presence of a ring

structure facilitates charring, forming a protective layer in the condensed phase. Apart

from isocyanurates, side products such as polyurea with urea linkages are formed

through the reaction of isocyanates with amine-terminated compounds. Polyurea pro-

vides the foam with rigidity and thermal stability.

Figure 4. Trimerization of isocyanates.

Polyester polyols and polyether polyols are the two main types of polyols. The dif-

ference in chemical structure between ester and ether is shown in Figure 5. Polyether pol-

yols have more resistance to hydrolysis but are less stable to oxidation; the inverse is true

for polyester polyols. Polyurethane foams based on polyether polyols have a lower de-

composition temperature in air than those based on polyester polyols. To improve the fire

behavior of PUFs, flame-retardant polyols such as VORAGUARD TM Polyol from The Dow

Chemical Company (Midland, MI, USA) and Exolit® OP 560 from Clariant AG (Muttenz,

Switzerland) are used. Another way to enhance the flame retardancy of PUFs is to use

aromatic polyols to promote char yield during burning [53].

Figure 3.

Temperature–thickness relationship of rigid polyurethane foams during different burn-

ing stages.

RPUFs behave quite differently from FPUFs during burning, showing the typical

HRR curve for residue-forming materials in Figure 3[

]. After ignition in stage (i),

they reach the PHRR immediately, undergoing distinct charring at T

, with no structural

collapse and no formation of a pool fire because of their higher crosslink density. The

char on the top acts as a protective layer, shielding the material underneath. The PHRR is

subsequently followed by steady burning in stage (ii) at a lower HRR. The pyrolysis front

at T

continuously consumes the material downward from the top (d

) to d

. Due to the

effective protective layer formed, the temperature of the unburned material from d

to d

decreases inversely toward the bottom of the material. The length of the steady-burning

phase in the HRR curve depends on the amount of combustible material [

]. Therefore,

less heat is released by RPUFs as the char yield increases [

]. In stage (iii), the pyrolysis

front at Tp2 moves to the bottom of the material, and the flame is finally extinguished.

Polymers 2022,14, 2562 6 of 24

2.2. Role of Selecting Contents of Isocyanate, Polyol, Foaming Agent, and Flame Retardants

Polyurethane chemistry is based on the high reactivity of isocyanates. Diisocyanates

are organic compounds with two isocyanate groups, which are widely used to link polyols

together through an exothermic reaction between isocyanates and hydroxyl groups in

order to build crosslinked polyurethane. The content of diisocyanates in the formulation

influences the thermal stability, rigidity, and fire behavior of PUFs. For instance, any excess

isocyanates are converted into trimers by trimerization (see Figure 4), called isocyanurate

rings [

]. Isocyanurates improve flame retardancy because the presence of a ring struc-

ture facilitates charring, forming a protective layer in the condensed phase. Apart from

isocyanurates, side products such as polyurea with urea linkages are formed through the

reaction of isocyanates with amine-terminated compounds. Polyurea provides the foam

with rigidity and thermal stability.